Acta Materialia 127 (2017) 143e152

Contents lists available at ScienceDirect

Acta Materialia
journal homepage: www.elsevier.com/locate/actamat

Full length article

Effect of temperature and texture on the reorientation of martensite

variants in NiTi shape memory alloys
G. Laplanche*, T. Birk, S. Schneider, J. Frenzel, G. Eggeler
€t Bochum, 44801 Bochum, Germany
Institut für Werkstoffe, Ruhr-Universita

a r t i c l e i n f o a b s t r a c t

Article history: Martensitic Ni50Ti50 wires and sheets with different textures were tensile tested in the temperature
Received 4 October 2016 range between
100  C and 60  C. The effect of texture and temperature on reorientation of martensite
Received in revised form variants was investigated. After deformation, all material states were heated into the austenite regime to
9 January 2017
study their shape memory behavior. During room temperature tensile testing, in-situ digital image
Accepted 10 January 2017
Available online 11 January 2017
correlation revealed that the reorientation of martensite variants is associated with the nucleation and
propagation of a macroscopic Lüders band. A comparison between the mechanical data obtained for wire
and sheet specimens revealed a strong effect of texture. The plateau stresses of sheets were found to be
Keywords:
Shape memory
25e33% larger and their recoverable strains were 30% lower than for wires. However, the product of
Digital image correlation plateau stress and recoverable strain, which represents the external work per unit volume required for
Martensitic phase transformation martensite variants reorientation does not depend on texture. The tensile tests performed at different
Texture temperatures revealed that in the temperature range considered the recoverable strain does not depend
significantly on temperature. In contrast, the plateau stress as well as the external work required to
reorient martensite decrease with increasing testing temperature. We use a thermodynamic approach
involving the elastic strain energy associated with the growth of reoriented martensite variants to
rationalize these temperature dependencies.
© 2017 Acta Materialia Inc. Published by Elsevier Ltd. This is an open access article under the CC BY
license (http://creativecommons.org/licenses/by/4.0/).

1. Introduction and objectives important: the martensite start and finish temperatures (Ms and
Mf) which describe the transformation from austenite to martensite
NiTi alloys are the most successful shape memory alloys (SMAs) when cooling from the high temperature regime and the austenite
in the medical [1] and engineering [2] sectors due to their superior start and finish temperatures (As and Af) which characterize the
mechanical and functional properties compared with other sys- reverse transformation from martensite to austenite when heating
tems such as Cu- [3,4], Fe- [5,6], Ru- [7,8] and Ti-based [9e11] SMAs. from the low temperature regime. The mechanical and functional
Various shape memory properties show an orientation dependence behaviors of shape memory alloys have been shown to strongly
[12e16]. Therefore the texture which develops during thermo- depend on temperature and two temperature regimes can be
mechanical processing of SMAs has a strong influence on their distinguished.
mechanical and functional properties [16e22]. Textures of NiTi In the high temperature regime, not too far away from Af [31],
drawn wires and rolled sheets have received considerable attention the material is austenitic and exhibits a B2 (CsCl type) crystal
in the literature. It is now well-documented that drawn NiTi wires structure. During tensile testing, a stress induced phase trans-
after recrystallization exhibit a strong 〈111〉B2 fiber texture along formation is observed which accounts for up to 10% strain, i.e. the
the wire axis in the high temperature austenitic state [23,24]. In austenite B2 transforms into martensite (B190 monoclinic crystal
contrast, NiTi rolled sheets are characterized by 〈011〉B2 austenite structure) [32]. As the stress induced martensite is metastable in
fiber texture parallel to the rolling direction whose amplitude de- the high temperature regime, it transforms back to austenite upon
pends on the rolling temperature [25e30]. unloading and the material recovers its original shape. This me-
For shape memory alloys, four specific temperatures T are chanical memory is usually referred to as superelasticity or pseu-
doelasticity. In the high temperature regime, the critical stress for
the formation of martensite strongly increases with increasing
* Corresponding author. temperature and this phenomenon can be rationalized by a
E-mail address: [email protected] (G. Laplanche).

http://dx.doi.org/10.1016/j.actamat.2017.01.023
1359-6454/© 2017 Acta Materialia Inc. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
144 G. Laplanche et al. / Acta Materialia 127 (2017) 143e152

Clausius-Clapeyron type of relationship, e.g. Ref. [33]. 4 ¼ 2 ln(di/diþ1) (1)

In the low temperature regime below the martensite finish
temperature Mf [32], the material is fully martensitic and exhibits a where di and diþ1 are the rod diameters before and after swaging
self-accommodated microstructure. During straining, favorably step i, respectively. Swaging was performed at 800  C and the
oriented martensite variants grow at the expense of less favorably diameter of the cylindrical NiTi ingot was reduced from 40 mm to
oriented ones and this can provide up to 10% pseudoplastic strain 5.5 mm (total cross-sectional reduction of 2.0) in 14 successive
[12]. After unloading, the martensitic alloy remains deformed but steps with a cross-sectional reduction of about 0.3 in each step [45].
the material can recover its original shape on heating above Af Before each swaging step, the NiTi rod was annealed at 800  C for
(where martensite fully transforms to austenite). The fact that all 10 min. After the final diameter reduction, the cylindrical specimen
martensite variants have to transform to austenite which cannot was annealed at 800  C for 10 min, followed by water quenching.
form variants, results in thermal memory which is often referred to Then the 5.5 mm rod was cold drawn using a HMP machine of type
as the one way effect (1WE) [34]. In the low temperature regime, ZPR 2000 6 down to a final diameter of 1.7 mm (total cross-
the apparent yield stress is shown to increase with decreasing sectional reduction of 1.2). The NiTi rod was cold drawn with a
temperature [35,36]. drawing speed of 2.4 mm/s in several successive steps with
While the temperature dependence of the mechanical proper- 4 ¼ 0.07. After three successive drawing steps, intermediate
ties of pseudoelastic alloys has been extensively studied in the annealing at 800  C for 10 min followed by water quenching was
literature [35,37e40], the reorientation of martensite variants and performed. Finally the as-drawn wire was straight annealed under
the associated temperature dependence of functional properties a tensile stress of 50e150 MPa at 800  C for 10 min. Recrystalliza-
have received less interest [36]. To our knowledge, no attempt has tion occurs during this heat treatment which is followed by water
been undertaken so far to model the effect of temperature on the quenching to freeze the solutionized high temperature state.
reorientation of martensite variants.
In this study, after a brief review of the elementary deformation 2.3. Thermomechanical processing of Ni50Ti50 sheets
mechanisms which govern the reorientation of martensite variants,
an attempt is made to explain the effect of temperature and texture After homogenization, the rectangular ingot (initial thickness of
on the reorientation of martensite variants. To reach this goal, two 20 mm) was rolled at 800  C in 22 steps down to a thickness of
Ni50Ti50 SMAs with different textures were subjected to tensile 1.8 mm which corresponds to a cross-sectional reduction by rolling
testing at temperatures between
100  C and þ60  C. The obtained of 4 ¼ |ln (ti/tf)| z 2.4 where ti and tf are the initial and final
results are rationalized on the basis of a thermodynamic approach thicknesses, respectively. Before each hot rolling step, the material
which accounts for the growth of reoriented martensite variants in was annealed at 800  C for 10 min. The thickness of the hot rolled
a self-accommodated microstructure. sheet was then reduced by rolling at room temperature from 1.8 to
1.2 mm (total cross-sectional reduction of 0.4) in three steps. More
2. Experimental details about each rolling step can be found in the literature [30].
Between each cold rolling step, the material was annealed at 800  C
2.1. Melting and casting for 10 min followed by water quenching. After the final cold rolling
step, the resulting NiTi sheet was recrystallized at 800  C for 10 min
High purity nickel pellets (Ni, > 99.98 wt%) of an average size of and water quenched.
9 mm and titanium slabs with a quadratic shape (Ti, > 99.995 wt%),
8 mm thick and 50 mm large, were purchased from Ampere GmbH 2.4. Thermal and microstructural characterization
(Frankfurt, Germany) and Hauner Metallische Werkstoffe GmbH
€ ttenbach, Germany), respectively. Ti rapidly oxidizes in air and
(Ro The phase transformation temperatures Ms, Mf, As and Af of our
therefore its oxide layer was removed by grinding (80 grid SiC alloys were measured by differential scanning calorimetry (DSC),
paper) prior to melting. The raw materials were melted under a for more details about DSC see Ref. [46]. The transformation tem-
high purity argon atmosphere of 500 mbar (99.998 vol%) by vac- peratures of Ni50Ti50 wires and sheets are similar and are given in
uum induction melting (VIM, type VSG 010 from PVA TePla AG) and Table 1.
subsequently poured into steel molds preheated to 500  C. Two In order to characterize the microstructure of the NiTi wires and
molds were used in the present study, a cylindrical mold (40 mm sheets, longitudinal sections were etched at room temperature
diameter, 120 mm height) and a rectangular mold with dimensions using a stirred solution containing 14 g K2S2O5, 1000 ml distilled
of 20  76  90 mm. More details about the melting parameters water, 200 ml HCl (32 wt%), and 24 g (NH4)HF2. The specimens
used in the present work can be found in the literature [30,41,42]. were subsequently characterized by optical microscopy using
Prior to inserting the molds into the VIM furnace, their walls were polarized light. A representative optical micrograph for a NiTi wire
coated with an yttria slurry. Two 1 kg ingots of the equiatomic is shown in Fig. 1 where the swaging/drawing direction is high-
Ni50Ti50 alloy were produced. An effort was made to keep carbon lighted by a white arrow in the upper right corner of the micro-
and oxygen impurity levels low, because their presence signifi- graph. Etching of the martensitic alloy allows to reveal the previous
cantly affects phase transformation temperatures [43] and me- grains of the parent austenitic phase. Indeed the inset in the lower
chanical properties [44]. The cast ingots were homogenized for left corner of the optical micrograph shows a prior austenitic grain
10 h at 1000  C under an argon atmosphere, followed by water
quenching to achieve a homogeneous microstructure. Table 1
Transformation temperatures (in  C) of Ni50Ti50 wires and sheets deter-
mined by DSC (precision: ± 5  C).
2.2. Thermomechanical processing of Ni50Ti50 wires
Wires Sheets
The homogenized cylindrical ingot was processed by means of Af 100 105
rotary swaging using a four dies swaging machine of type HMP R6- As 70 80
4-120-21S (HMP Umformtechnik GmbH, Pforzheim, Germany). The Ms 60 70
Mf 30 50
cross-sectional reduction by swaging, 4, is given by
 G. Laplanche et al. / Acta Materialia 127 (2017) 143e152 145

software Vic3D (Limess, Pforzheim, Germany).

3. Results

3.1. Digital image correlation (DIC)

A typical stress-strain curve for the Ni50Ti50 wire recorded

during the DIC experiment at room temperature is shown in Fig. 2a
where the circled numbers 1 to 12 indicate at which stage DIC-
measurements (shown in Fig. 2b) were performed. The colors in
the DIC-images represent the local strain distribution as shown in
the color code on the right side of Fig. 2b. The initial linear elastic
part of the stress-strain curve is followed by a pseudo-plastic
regime where inelastic deformations are homogeneously distrib-
uted over the whole wire (compare DIC-images 1 and 2 in Fig. 2b).
However, between stage 2 and 3, a load drop occurs which reflects
the formation of a band of localized deformation near the bottom
grip. This so-called Lüders band propagates towards the upper grip
at a constant stress and a constant velocity, see DIC-images 3 to 9 in
Fig. 2b. For each deformation state (points 3 to 9), the strain dis-
tribution can be divided into three regions. The first region is
Fig. 1. Optical micrograph of an etched Ni50Ti50 wire where the grains of the parent
located at the bottom of the wire (where the Lüders band nucle-
phase were split in several variants forming a self-accommodated microstructure. ated) and exhibits a homogeneous tensile strain of 8.5%. The second
region is at the top of the wire and shows small deformation levels
not exceeding 2%. These two regions are separated by a third
which consists of several martensite variants after cooling. Sizes of transition region (or Lüders band) where the local deformation
the prior austenitic grains were measured with the Heyn linear varies progressively from 2 to 8.5%. At each subsequent data point
intercept method outlined in ASTM E112-10 [47] using ten parallel on the plateau stress (3e9), additional strain increments occur
and equidistant reference lines of identical length. Two sets of in- exclusively by the motion of the transition region which has a finite
tercepts were used, one parallel and another perpendicular to the width of ~ 2 mm, comparable to the diameter of the martensitic
deformation direction (wire axis and rolling direction) in order to wire (1.7 mm). More detailed features regarding distributions of
measure whether grains are equiaxed or elongated along the wire local strains and local strain rates are presented in Fig. 3 for a
axis/rolling direction. One reference line i of length li intersecting a macroscopic strain of 4% (see 5 in Fig. 2b). These strain and strain
number ni of grain boundaries yields a grain size of di ¼ li/ni. Typical rate data are representative for all macroscopic strains along the
grain size measurements involved about 500 intercepts per stress-plateau (points 3 to 9 in Fig. 2b). The distribution of local
metallographic cross-section. Both recrystallized wires and sheets strains εx as a function of the position x along the wire axis exhibits
revealed equiaxed grains with a mean grain size of 15 ± 5 mm. a step-like shape (see Fig. 3b). In the region near the bottom grip
(8.5% local strain, Fig. 3a and b) where martensite variants were
2.5. Mechanical characterization reoriented, local strains correspond to the end of the macroscopic
stress-plateau (Fig. 2a). In contrast, local strains (< 2%) inside the
The NiTi wires and sheets were tensile tested in tension using a region at the top of the wire, where martensite variants still form a
Zwick/Roell test rig of type Z100 equipped with a 100 kN load cell self-accommodated microstructure, correspond to the macroscopic
and a clip-on extensometer [48]. The NiTi sheets were tested along strain at the onset of the stress-plateau. Now regarding local strain
the rolling direction. Specimens with a gauge length of 25 mm were rates (see Fig. 3c), they are nearly equal to zero in the regions at the
used. Displacement-controlled tensile tests were conducted in the bottom and at the top of the wire while a strain rate peak is
temperature range
100 < T < þ60  C using a cross-head speed of observed in the transition region. This distribution of local strain
0.5 mm/min. A climate chamber kept the temperature constant rates shows that all deformation/reorientation takes place in the
during tensile testing. Prior to mechanical testing, the NiTi speci- narrow transition region. The end of the stress-plateau in Fig. 2a
mens were cooled down to
100  C to establish a fully martensitic occurs when the Lüders band reaches the upper loading grip, DIC-
microstructure prior to mechanical loading. The wires and sheets image 10. When continuing testing beyond the stress-plateau,
were deformed up to 9% and 6% strain, respectively, i.e. about 1% stresses sharply increase in an elasto-plastic regime. These results
strain beyond the end of their respective stress-plateau. After are similar to those observed for austenitic NiTi alloys during the
unloading, the shape memory behavior was investigated by heating stress induced austenite to martensite phase transformation.
the material in 10  C per minute up to 150  C (50  C above Af). This Therefore, it is likely that the Lüders-type of deformation mecha-
temperature was held for 30 min and the material was cooled down nisms for martensitic and austenitic alloys have a similar nature.
(10  C/min) to room temperature. The unloading part of the stress strain curve in Fig. 2a is approxi-
Additional tensile tests were carried out at room temperature mately linear elastic up to point 11 which is followed by a nonlinear
and strain fields were documented using digital image correlation unloading part between points 11 and 12 in Fig. 2a. This indicates
(DIC). These tensile tests were performed using a miniature 10 kN that inelastic recovery processes are occurring. During unloading,
tensile test rig from Kammrath & Weiss GmbH, Dortmund, at a the DIC-images 11 and 12 suggest that the NiTi wire exhibits a
displacement rate of 0.5 mm/min. A random speckle pattern was homogeneous deformation.
applied by spraying black and white paints onto the specimen
surface. The system for DIC strain measurements consisted of one 3.2. Effect of temperature and texture on mechanical properties
pair of 5 megapixel cameras that acquired 5 images per second.
Strains were determined from the displacement fields using the Representative stress-strain curves obtained at different
146 G. Laplanche et al. / Acta Materialia 127 (2017) 143e152

Fig. 2. Stress strain characteristics of martensitic Ni50Ti50 wire deformed at room temperature. (a) Stress-strain curve which indicates 12 positions where strain distributions were
measured. (b) Color coded strain fields (DIC measurements) at positions highlighted in (a). (For interpretation of the references to color in this figure legend, the reader is referred to
the web version of this article.)

shows representative stress-strain curves for the wire (blue) and

the sheet (green) materials deformed at 20  C up to 9% and 6%
strain, respectively (~1% strain beyond the end of their respective
stress-plateau). The general shapes of these curves are similar.
However, due to the different textures of both materials the
magnitude of the stress-plateau, its length as well as the recover-
able strain are different. The temperature dependence of the stress-

Fig. 3. Strain and strain rate distributions for position 5 in Fig. 2a. (a) Color coded
strain distribution. (b) Local strains εx plotted as a function of wire position x. (c) Local
strain rates ε_ x as a function of the position x along the wire axis (0 marks a position at
the middle of the wire). For details see text. (For interpretation of the references to
color in this figure legend, the reader is referred to the web version of this article.)

temperatures for NiTi sheets tensile strained along their rolling

direction (RD) are presented in Fig. 4. In addition to the stress-strain
response, recoverable strain which occurs when the 1WE is trig-
gered by heating to 150  C, is shown as a black dashed line. Fig. 5
Fig. 5. Comparison of tensile stress-strain curves and strain recovery (dashed lines)
obtained at 20  C for NiTi wires and sheets.

Fig. 4. Loading/unloading tensile stress-strain curves together with strain recovery associated with the one way effect (black dashed line) obtained at different temperatures for
Ni50Ti50 sheets strained along their rolling direction.
 G. Laplanche et al. / Acta Materialia 127 (2017) 143e152 147

plateau and recoverable strain are shown in Fig. 6a and b, respec- εrec of the NiTi wires and sheets are 5.6 ± 0.5% and 4 ± 0.5%,
tively. In the investigated temperature range, the plateau stresses respectively, i.e. they are always larger for wires. Interestingly,
monotonically decrease with increasing temperature (Fig. 6a) while when the product of the stress-plateau and the recoverable strain is
the recoverable strains remain constant (Fig. 6b). The magnitude of plotted as a function of temperature in Fig. 6c, all data points fall
the stress-plateau of the sheets is always larger than that of the onto one common master curve, regardless of the texture of wires
wires (between 25 and 33% larger). In contrast, recoverable strains and sheets.

4. Discussion

4.1. Elementary deformation mechanisms associated with the

reorientation of martensite variants

In the present study, martensitic NiTi wires and sheets with

similar grain sizes and different textures were tensile strained at
temperatures between
100  C and þ60  C. Stress-strain curves of
both wires and sheets exhibit a similar shape with a characteristic
stress-plateau which is associated to the propagation of a Lüders
band, Fig. 2. Both wires and sheets were strained 1% beyond the end
of their respective stress-plateau, unloaded and subsequently
heated 50  C above Af to trigger the 1WE. The obtained data
revealed a strong effect of texture on the magnitude of the stress-
plateau, its length, and the recoverable strain obtained after the
1WE. However, the product of the stress-plateau and the recover-
able strain, which represents an energy per unit volume required to
reorient martensite variants, is found to be independent of texture,
Fig. 6. This strongly suggests that this energy is characteristic of a
unique elementary deformation mechanism which includes
texture effects. In the following, after a brief review of the
elementary deformation mechanisms which govern the reor-
ientation of martensite variants, attempts will be made to explain
the fact that the external work per unit volume required to reorient
martensite variants is independent of texture and to rationalize its
temperature dependence.
Although the inhomogeneous nature of the stress induced
austenite to martensite phase transformation has been well docu-
mented by several researchers [49e64], few efforts have been
made so far to study the reorientation of martensite variants by DIC
[65,66]. In this section we briefly summarize the governing
elementary deformation mechanisms, compiled from various
literature sources, responsible for the appearance of tensile stress-
strain curves and the propagation of the Lüders band observed by
DIC. A typical stress-strain curve showing the 1WE is shown in
Fig. 7. This stress-strain curve can be subdivided into four regimes.
The corresponding elementary deformation mechanisms are
illustrated in Fig. 8aee. Prior to mechanical loading, the NiTi SMA
consists of 24 self-accommodated martensite variants. The self-
accommodated microstructure has a multiscale character. At the
nanoscale, the martensite variants in NiTi SMAs are internally
twinned (mostly [011]B190 type II twinning) [67]. For the sake of
simplicity, these internally twinned martensite variants are here-
after referred to as variants. At a larger scale, different neighboring
variants in self-accommodated microstructures can share a
coherent junction plane [68] also termed as macrotwin [69], midrib
[70,71], or conjugation boundary [72,73]. These two different types
of boundaries at the nano and macroscale may be mobile under
stress. Their motion allows martensite variants, which are present
in the initial self-accommodated microstructure, to reorient. For
more details about macrotwins motion, see Refs. [74e76].
We now discuss a scenario which reflects the elementary pro-
cesses governing re-orientation of martensite variants during me-
chanical loading. In Fig. 8a, grain boundaries of the parent
austenitic phase are black, twin planes are blue and macrotwin
Fig. 6. Temperature dependence of shape memory properties for NiTi sheets and
wires. (a) plateau stress, (b) recovery strains, and (c) products of plateau stresses and
boundaries between two neighboring variants are represented by
corresponding recovery strains. The black line in (c) is calculated using the thermo- red lines. Regime I in Fig. 7 defines the initial elastic stage of the
dynamic approach developed in section 4.3, for details see text. material. The elastic stage is followed by a first pseudo-plastic
148 G. Laplanche et al. / Acta Materialia 127 (2017) 143e152

yield point (end of regime II in Fig. 7). Here the critical stress to
nucleate new favorably oriented martensite variants has been
reached, see elliptical nuclei at the bottom of Fig. 8c. Between re-
gimes II and III, a stress drop occurs followed by a stress-plateau.
The stress drop corresponds to the formation of a macroscopic
Lüders band close to the bottom grip [77,78], see orange region in
Fig. 8c. In Fig. 8cee, the region where martensite has been reor-
iented is highlighted in red while the propagating transition layer is
colored in orange. Within this Lüders band, new favorably oriented
martensite variants nucleate and grow within unfavorably oriented
martensite variants [77,79]. The interfaces between the old and
new martensite variants are coherent ((100)B19’/(010)B2 junction
planes [79]). Then the deformation in regime III proceeds in a
catalytic manner which was also termed as domino detwinning
[80], i.e. the interaction of reoriented martensite variants with prior
austenitic grain boundaries in the Lüders band generate stress
concentrations due to deformation incompatibilities between
grains. These internal stresses can be relieved by the nucleation of
new variants in neighboring grains outside the Lüders band. This
behavior provides the basis for localized reorientation of
martensite variants in a Lüders-like manner. The stress drop in
Fig. 7. Typical tensile stress-strain curve, for a fully martensitic Ni50Ti50 wire, for de- Fig. 7 is due to the fact that the nucleation of the Lüders band re-
tails see text.
quires higher stresses than for its propagation [49]. Note that this
localized deformation mechanism is very similar to the stress
induced formation of martensite in austenitic NiTi SMAs [50]. The
end of the stress-plateau in Fig. 7 is reached when the Lüders band
has propagated through the whole specimen, see Fig. 8e. Due to
incompatibilities between grains (former austenitic grains), the
reorientation process is never complete, i.e. grains of the parent
phase cannot transform into a single crystalline domain of
martensite.
Depending on the magnitude of the critical stress, dislocation
slip can accompany the reorientation of martensite variants in
stage III [33]. When the critical stress to reorient martensite vari-
ants is lower than the critical stress to induce dislocation plasticity
(~150 MPa in solution-treated single crystals [81]), the inelastic
strains in stage III can be fully recovered through phase trans-
Fig. 8. Schematic representation of the elementary deformation mechanisms occur- formation upon heating (1WE). However, when the critical stress
ring during martensite variant reorientation. (a) Stress-free self-accommodated
which is required to reorient martensite variants is higher than that
microstructure where grain boundaries are black, twin boundaries are blue and
coherent junction planes between two neighboring twinned martensite plates are for slip, irreversible deformations hamper full strain reversibility
represented by red lines. (b) Martensite variant reorientation by movement of junction during the 1WE. In this latter case, the deformation is partially
planes. (c) Lüders band (orange region) formation where favorably oriented martensite reversible as shown in Fig. 7.
variants nucleate (red ellipses between the old and new martensite variants). (d) Between stages III and IV, a short elastic stage (between 145 and
Lüders band propagation (swept area colored in red). (e) The Lüders band has prop-
agated through the whole sample. Due to incompatibilities between grains, the
160 MPa) is likely followed by yield which indicates a transition to
martensite variant reorientation process is not complete. (For interpretation of the irreversible plastic deformation due to dislocation activities, reor-
references to color in this figure legend, the reader is referred to the web version of this ientation of martensite variants and/or deformation twinning
article.) [82,83].

4.2. Effect of texture on the reorientation of martensite variants

regime II, lying between an apparent yield stress (~100 MPa) and an
upper yield stress (~175 MPa), Fig. 7. At the apparent yield stress,
To rationalize the strong effect of texture shown in Fig. 6 we use
favorably oriented twin boundaries are mobile which allow det-
a Taylor-type model for textured polycrystals. While the orientation
winning, see blue arrows in Fig. 8b. Moreover, macrotwin bound-
dependence of the Taylor factor has been reported in the literature
aries are also mobile and allow favorably oriented twinned
for the stress induced transformation from austenite into
domains to grow at the expense of unfavorably oriented ones, see
martensite [84], to our knowledge nothing equivalent has been
red arrows in Fig. 8b. The motion of these boundaries is the first
reported so far regarding the reorientation of martensite. In other
step of the reorientation of martensite variants. Twin boundaries
words, the involved shear systems may differ between the reor-
favorably oriented with respect to the direction of loading (high
ientation of martensite and the austenite to martensite trans-
resolved shear stresses) move first and less favorably oriented ones
formation. However, if we assume the Taylor factors to be the same,
require higher stresses to become mobile. From a macroscopic
a textured polycrystal, in which the crystallite axes along the
point of view the deformation of the specimen is homogeneous
loading direction are preferentially aligned along the 〈111〉B2 di-
within regime II as shown by digital image correlation, image 2 in
rections, has a Taylor factor of 2.4, 3.2 for a 〈101〉B2 fiber texture, 4.9
Fig. 2b. All the variants cannot reorient by motion of pre-existing
for a 〈001〉B2 fiber texture and 3.2 for a random texture [84]. Based
macro and nanotwin boundaries. As a consequence, the level of
on the expected texture of the wire and sheet, i.e. 〈111〉B2 fiber
internal stresses increases and the material hardens until the upper
texture along the wire axis and 〈011〉B2 fiber texture along the
 G. Laplanche et al. / Acta Materialia 127 (2017) 143e152 149

rolling direction of the sheet, we expect the Taylor factor of the wire
and sheet to be 2.4 and 3.2, respectively. The relationship between p m ð2
nÞ 2 pm
K¼ g þ ε2 (5)
the resolved shear stress t, the stress-plateau splateau and the Taylor 8 ð1
nÞ 13 4 ð1
nÞ 33
factor M is splateau ¼ t  M and the relationship between the shear
magnitude associated with the martensitic transformation where m is the elastic shear modulus, n the Poisson’s ratio, g13 de-
g13 ¼ 0.13 (see section 4.3), the recoverable strain εrec and the Taylor notes the shear and ε33 the volume change associated with the
factor M is εrec ¼ g13/M [85]. According to these equations it is reorientation of martensite variants. As the reorientation of
expected that the plateau stresses of wires are 30% lower and their martensite variants does not involve any volume change, we have
recoverable strain 30% larger than for sheets. These results are in ε33 ¼ 0. Now we develop an expression for g13. It has been reported
excellent agreement with those shown in Fig. 6. It is also worth in the literature that the reorientation of martensite in NiTi shape
mentioning that the product splateau  εrec ¼ t  g13, which is the memory alloys involves twinning on the (100)B190 plane in the
energy per unit volume required to reorient martensite, is inde- [001]B190 direction [79]. An elementary reorientation process is
pendent of the Taylor factor and therefore independent of texture, schematically illustrated in Fig. 9 where a white single variant is
as shown in Fig. 6c. being reoriented into a new grey variant under the action of a shear
stress by twinning on the (100)B190 plane in the [001]B190 direction.
From Fig. 9, it can be seen that the shear angle q associated with the
4.3. Temperature dependence of the interaction energy for the reorientation of martensite is related to the monoclinic angle b of
reorientation of martensite variants the NiTi unit cell by

In previous sections, we have shown that the Lüders band

p
q¼2 b
(6)
propagates at a constant stress splateau corresponding to a constant 2
energy per unit volume (splateau  εrec). As the nucleation and
The deformation gradient F associated with the simple shear
growth of new martensite variants plays a key role in the propa-
shown in Fig. 9 is given by
gation of the Lüders band, it is likely that it is responsible for the
temperature dependence of the external work required to reorient 0 1
1 0 q
martensite variants. To investigate this possibility, we treat an F ¼ @0 1 0A (7)
isolated reoriented martensite nucleus in a self-accommodated 0 0 1
microstructure as an ellipsoidal inclusion embedded in a homo-
geneous effective medium invoking Eshelby’s theory [86]. The total and the strain tensor can be calculated using
Gibbs free energy change DG due to the formation of an ellipsoidal
0 1
reoriented martensite variant with radius r and semi-thickness t is
1 T 0 0 q=2
given by Ref. [87]. ε¼ F þF
I ¼@ 0 0 0 A (8)
2

 q=2 0 0
4 4
DG ¼ pr2 t Dgch þ prt 2 K þ 2pr2 g (2) T
3 3 In Eq. (8), F is the transpose of the matrix F and I is the identity
matrix. Therefore from Eq. (8), it can be seen that
where Dg ch is the chemical free energy change per unit volume
accompanying the reorientation of the martensite variant. The g13 ¼ q/2 ¼ bep/2 (9)
second term in Eq. (2) is the elastic strain energy and the last term
represents the interfacial energy, which shape memory researchers When Eqs (5) and (9) are inserted in Eq. (4), we find
usually neglect [88]. In Eq. (2), K is a strain energy parameter and g
 

is the specific interfacial energy of a fully coherent inclusion. In the p ð2
nÞ t p2
splateau εrec ¼ m b
(10)
case of the reorientation of martensite variants during uniaxial 4 ð1
nÞ r 2
loading, the chemical free energy corresponds to the interaction
energy for twinning into a preferential variant (external work) Assuming that the ratio t/r of fully-grown reoriented domains is
[89,90]. It is expressed as independent of temperature and has a value in the range
[0.01e0.04] [91], Eq. (10) shows that the temperature dependence
Dgch z
splateau εrec (3) of the interaction energy per unit volume can only result from the
evolution of the elastic moduli (m and n) and the evolution of the
Right after nucleation, the nucleus grows radially faster than monoclinic angle b with temperature. Using ultrasonic measure-
along the thickness direction. This continues until the nucleus ments, Brill et al. [93] have measured the temperature dependence
growth is radially stopped at a grain boundary of the parent phase of the Poisson’s ratio and elastic shear modulus m in Ni50.5Ti49.5
[91]. Subsequently, the reoriented martensite variant can only grow single crystal which transformed into a self-accommodated
in the thickness direction in response to the driving force
vDG=vt microstructure upon cooling. The Poisson’s ratio was shown to
where t is the semi-thickness of the reoriented martensite variant. exhibit a step-like behavior from n ¼ 0.43 (austenite) in the high
This thickening does not cease until the driving force decreases to
zero (vDG=vt ¼ 0) [92], which yields

splateau εrec t
¼ K (4)
2 r
Therefore, when a unit volume of self-accommodated
martensite has been swept by the Lüders band, the thickening of
a reoriented martensite variant stops when its elastic strain energy
Fig. 9. Schematic drawings illustrating the reorientation of the monoclinic (B190 ) unit
has reached half the magnitude of the external work [92]. cell (with lattice parameters a, b, c and b) by twinning on the (100)B190 plane in the
Assuming isotropic elasticity, the parameter K in Eq. (4) can be [001]B190 direction under a shear stress. The angle q shown in the right part of the
defined as [87] figure represents the shear angle associated with the reorientation of martensite.
150 G. Laplanche et al. / Acta Materialia 127 (2017) 143e152

temperature regime to n ¼ 0.35 (self-accommodated martensite) in error on recoverable strains is as large as 0.5%, i.e. the recoverable
the low temperature regime [93]. The evolution with temperature strains of the wire and sheet are 5.6 ± 0.5% and 4.0 ± 0.5%,
of the elastic shear modulus m obtained by Brill et al. [93] and the respectively, it is therefore not possible to experimentally detect
evolution of the monoclinic angle b determined using X-ray the slight increase of the recoverable strain with decreasing tem-
diffraction by Prokoshkin et al. [94] are shown in Fig. 10a and b, perature as shown in Fig. 6b. To summarize, while an increase of
respectively. The red trend lines represent best fits to the experi- recoverable strain with decreasing temperature may occur, the
mental data for interpolation purposes. Note that the exact chem- experimental error does not allow us to detect it.
ical composition of the NiTi alloys which were investigated by Brill Inserting the experimental temperature dependence of m and b
et al. [93] and Prokoshkin et al. [94] differs slightly from the as shown in Fig. 10a and b into the energy term presented in Eq. (10)
equiatomic concentration (Ni50.5Ti49.5 and Ni50.26Ti49.74, respec- we obtain the calculated black line shown in Fig. 6c with Ms ¼ 60  C
tively). To avoid any discrepancy related to alloy composition, we and t/r ¼ 0.0112. It is a striking new finding that this calculated line
subtract Ms (martensite start temperature) from the temperature. which is only based on elastic moduli (m) and martensite crystal-
Fig. 10 shows the evolution of m and b with T e Ms. lography (b) rationalizes the experimental data. The excellent
As the shear angle b increases with decreasing temperature (see match between experimental and theoretical data in Fig. 6c clearly
Fig. 10b), the shear magnitude g13 associated with the reorientation shows that the temperature dependence of the plateau stress
of martensite variants increases by 5% when the temperature de- (Fig. 6a) is related to a softening of the elastic shear modulus and a
creases from 0 to
100  C. The recoverable strains should also in- decrease of the monoclinic angle with increasing temperature.
crease by the same amount in this temperature range. The
recoverable strain of the wire and sheet are expected to increase
from 5.6% to 5.88% and 4%e4.2%, respectively, when the tempera- 5. Summary and conclusions
ture decreases from 0 to
100  C. However, since the experimental
In the present study, fully martensitic Ni50Ti50 wires and sheets
with similar transformation temperatures, similar grain sizes but
different textures (〈111〉B2 fiber texture along the wire axis, 〈011〉B2
fiber texture along the rolling direction of the sheet) were pro-
duced. Their stress-strain curves exhibit the well-known plateau
behavior, which is governed by the propagation of a Lüders band.
The experiments were interrupted after a small additional strain
interval (~1%) beyond the end of the stress-plateau. After unload-
ing, the specimens were heated to trigger the one way shape
memory effect and the associated recoverable strains were
measured. From the results obtained in the present work the
following conclusions can be drawn:

1) A stress drop at the beginning of the stress-plateau indicates the

nucleation of a Lüders band. Digital image correlation shows
that the Lüders band nucleates at one grip of the test rig, from
where it propagates through the specimen until it reaches the
other grip.
2) Despite their similar transformation temperatures and grain
sizes, NiTi wires and sheets exhibit strong differences in their
tensile and shape memory behaviors due to their different
textures. The stress-plateau of the sheets is 25e33% larger and
their recoverable strains are 30% lower than those of the wires.
However, the external work per unit volume required for the
reorientation of martensite variants, defined in the present
study as the product of the stress-plateau and the recoverable
strain, is independent of texture. All these results can be ratio-
nalized by a Taylor-type model for polycrystals.
3) The recoverable strain was found to be independent of tem-
perature while the stress-plateau increases monotonically with
decreasing temperature. As a result, the external work per unit
volume required to reorient the martensite variants increases
with decreasing temperature. This is due to the fact that the
elastic strain energy per unit volume which hampers the reor-
ientation of martensite variants is increasing with decreasing
temperature.
4) The decrease of the elastic strain energy per unit volume with
increasing temperature results from two factors. First, the elastic
shear modulus softens when the temperature increases towards
the martensite start temperature (this represents the major
contribution 60e70%). Second, the magnitude of the monoclinic
angle and therefore the magnitude of the shear associated with
Fig. 10. Experimental literature data for martensitic NiTi alloys as a function of tem- the reorientation of martensite variants are decreasing with
perature plotted as T e Ms. (a) shear modulus m [93] and (b) monoclinic angle b [94]. increasing temperature (minor contribution 30e40%).
 G. Laplanche et al. / Acta Materialia 127 (2017) 143e152 151

Acknowledgements [29] S.H. Chang, S.K. Wu, Textures in cold-rolled and annealed Ti50Ni50 shape
memory alloy, Scr. Mater. 50 (2004) 937.
[30] G. Laplanche, A. Kazuch, G. Eggeler, Processing of NiTi shape memory sheets e
G.L. acknowledges funding from the Alexander von Humboldt microstructural heterogeneity and evolution of texture, J. Alloys Compd. 651
(AvH) Foundation. G.E., T.B. and J.F. acknowledge funding provided (2015) 333.
by the German Research Association (DFG: Deutsche For- [31] Y. Chumlyakov, I. Kireeva, E. Panchenko, I. Karaman, H.J. Maier, E. Timofeeva,
Shape memory effect and high-temperature superelasticity in high-strength
schungsgemeinschaft) through project FR 2675/2-1. single crystals, J. Alloys Compd. 577 (2013) S393.
[32] K. Otsuka, C.M. Waymann, in: K. Otsuka, C.M. Waymann (Eds.), Shape Memory
Materials, Cambridge University Press, Cambridge, 1998, p. p.1.
References [33] K. Otsuka, X. Ren, Physical metallurgy of Ti-Ni-based shape memory alloys,
Prog. Mater. Sci. 50 (2005) 511.
[1] T. Duerig, A. Pelton, D. Sto € ckel, An overview of nitinol medical applications, [34] T. Saburi, S. Nenno, The shape memory effect and related phenomena, in:
Mater. Sci. Eng. A 273e275 (1999) 149. H.I. Aaronson, D.E. Laughlin, R.F. Sekerka, C.M. Wayman (Eds.), Proc. Int.
[2] J. Van Humbeeck, Non-medical applications of shape memory alloys, Mater. Conference on Solid-solid Phase Transformations, 1981, p. p.1455. Pittsburgh
Sci. Eng. A 273e275 (1999) 134. AIME.
[3] Y. Sutou, T. Omori, R. Kainuma, K. Ishida, Grain size dependence of pseu- [35] S. Miyazaki, S. Kimura, F. Takei, T. Miura, K. Otsuka, Y. Suzuki, Shape memory
doelasticity in polycrystalline CueAleMn-based shape memory sheets, Acta effect and pseudoelasticity in a Ti-Ni single crystal, Scr. Metall. 17 (1983)
Mater. 61 (2013) 3842. 1057.
[4] M.L. No, A. Ibarra, D. Caillard, J.S. Juan, Quantitative analysis of stress-induced [36] T. Saburi, T. Takagaki, S. Nenno, K. Koshino, Mechanical behavior of shape
martensites by in situ transmission electron microscopy superelastic tests in memory Ti-Ni-Cu alloys, MRS Int. Meet. Adv. Mater. 9 (1989) p.147.
Cu-Al-Ni shape memory alloys, Acta Mater. 58 (2010) 6181. Pittsburgh.
[5] T. Omori, K. Ando, M. Okano, X. Xu, Y. Tanaka, I. Ohnuma, R. Kainuma, [37] P.H. Adler, W. Yu, A.R. Pelton, R. Zadno, T.W. Duerig, R. Barresi, On the tensile
K. Ishida, Superelastic effect in polycrystalline ferrous alloys, Science 333 and torsional properties of pseudoelastic NiTi, Scr. Metall. Mater 24 (1990)
(2011) 68. 943.
[6] M. Vollmer, P. Krooß, I. Karaman, T. Niendorf, On the effect of titanium on [38] H. Sehitoglu, J. Jun, X. Zhang, I. Karaman, Y. Chumlyakov, H.J. Maier, K. Gall,
quenching sensitivity and pseudoelastic response in Fe-Mn-Al-Ni-base shape Shape memory and pseudoelastic behavior of 51.5%NieTi single crystals in
memory alloy, Scr. Mater. 126 (2017) 20. solutionized and overaged state, Acta Mater. 49 (2001) 3609.
[7] K. Chastaing, A. Denquin, R. Portier, P. Vermaut, High-temperature shape [39] Y. Liu, A. Mahmud, F. Kursawe, T.-H. Nam, Effect of pseudoelastic cycling on
memory alloys based on the RuNb system, Mater. Sci. Eng. A 481 (2008) 702. the ClausiuseClapeyron relation for stress-induced martensitic trans-
[8] A.M. Manzoni, A. Denquin, P. Vermaut, F. Prima, I. Puente-Orench, C. Pauly, formation in NiTi, J. Alloys Compd. 449 (2008) 82.
F. Mücklich, R.A. Portier, Constrained hierarchical twinning in Ru-based high [40] H. Kato, K. Sasaki, Transformation-induced plasticity as the origin of serrated
temperature shape memory alloys, Acta Mater. 111 (2016) 283. flow in an NiTi shape memory alloy, Int. J. Plast. 50 (2013) 37.
[9] M. Tahara, H.Y. Kim, T. Inamura, H. Hosoda, S. Miyazaki, Lattice modulation [41] J. Frenzel, Z. Zhang, K. Neuking, G. Eggeler, High quality vacuum induction
and superelasticity in oxygen-added b-Ti alloys, Acta Mater. 59 (2011) 6208. melting of small quantities of NiTi shape memory alloys in graphite crucibles,
[10] E. Bertrand, P. Castany, T. Gloriant, Investigation of the martensitic trans- J. Alloys Compd. 385 (2004) 214.
formation and the damping behavior of a superelastic TieTaeNb alloy, Acta [42] T. Birk, S. Biswas, J. Frenzel, G. Eggeler, Twinning-induced Elasticity in NiTi
Mater. 61 (2013) 511. Shape Memory Alloys, Shape Memory and Superelasticity, 2016, p. 1.
[11] P. Castany, A. Ramarolahy, F. Prima, P. Laheurte, C. Curfs, T. Gloriant, In situ [43] J. Frenzel, E.P. George, A. Dlouhy, C. Somsen, M.F.X. Wagner, G. Eggeler, In-
synchrotron X-ray diffraction study of the martensitic transformation in fluence of Ni on martensitic phase transformations in NiTi shape memory
superelastic Ti-24Nb-0.5N and Ti-24Nb-0.5O alloys, Acta Mater. 88 (2015) alloys, Acta Mater. 58 (2010) 3444.
102. [44] M. Rahim, J. Frenzel, M. Frotscher, J. Pfetzing-Micklich, R. Steegmüller,
[12] S. Miyazaki, S. Kimura, K. Otsuka, Y. Suzuki, The habit plane and trans- M. Wohlschlo € gel, H. Mughrabi, G. Eggeler, Impurity levels and fatigue lives of
formation strains associated with the martensitic transformation in Ti-Ni pseudoelastic NiTi shape memory alloys, Acta Mater. 61 (2013) 3667.
single crystals, Scr. Metall. 18 (1984) 883. [45] C. Grossmann, J. Frenzel, V. Sampath, T. Depka, A. Oppenkowski, C. Somsen,
[13] H. Sehitoglu, I. Karaman, R. Anderson, X. Zhang, K. Gall, H.J. Maier, K. Neuking, W. Theisen, G. Eggeler, Processing and property assessment of
Y. Chumlyakov, Compressive response of NiTi single crystals, Acta Mater. 48 NiTi and NiTiCu shape memory actuator springs, Materwissen. Werksttech 39
(2000) 3311. (2008) 499.
[14] G. Laplanche, J. Pfetzing-Micklich, G. Eggeler, Orientation dependence of [46] J. Frenzel, A. Wieczorek, I. Opahle, B. Maaß, R. Drautz, G. Eggeler, On the effect
stress-induced martensite formation during nanoindentation in NiTi shape of alloy composition on martensite start temperatures and latent heats in
memory alloys, Acta Mater. 68 (2014) 19. NieTi-based shape memory alloys, Acta Mater. 90 (2015) 213.
[15] G. Laplanche, J. Pfetzing-Micklich, G. Eggeler, Sudden stress-induced trans- [47] A. E112e10, Standard Test Methods for Determining Average Grain Size,
formation events during nanoindentation of NiTi shape memory alloys, Acta 2010.
Mater. 78 (2014) 144. [48] J. Olbricht, A. Yawny, J.L. Pelegrina, A. Dlouhy, G. Eggeler, On the stress-
[16] K. Bhattacharya, Microstructure of Martensite: Why it Forms and How it Gives induced formation of R-phase in ultra-fine-grained Ni-rich NiTi shape mem-
Rise to the Shape-memory, Oxford University Press, Oxford, 2003. ory alloys, Metall. Mater. Trans. A 42 (2011) 2556.
[17] Y.C. Shu, K. Bhattacharya, The influence of texture on the shape-memory ef- [49] J.A. Shaw, S. Kyriakides, Thermomechanical aspects of NiTi, J. Mech. Phys.
fect in polycrystals, Acta Mater. 46 (1998) 5457. Solids 43 (1995) 1243.
[18] W.Q. Yuan, S. Yi, Pseudo-elastic strain estimation of textured TiNi shape [50] J.A. Shaw, S. Kyriakides, On the nucleation and propagation of phase trans-
memory alloys, Mater. Sci. Eng. A 271 (1999) 439. formation fronts in a NiTi alloy, Acta Mater. 45 (1997) 683.
[19] S. Miyazaki, V.H. No, K. Kitamura, A. Khantachawana, H. Hosoda, Texture of Ti- [51] Z.Q. Li, Q.P. Sun, The initiation and growth of macroscopic martensite band in
Ni rolled thin plates and sputter-deposited thin films, Int. J. Plast. 16 (2000) nano-grained NiTi microtube under tension, Int. J. Plast. 18 (2002) 1481.
1135. [52] G. Tan, Y. Liu, P. Sittner, M. Saunders, Lüders-like deformation associated with
[20] K. Gall, H. Sehitoglu, R. Anderson, I. Karaman, Y.I. Chumlyakov, I.V. Kireeva, On stress-induced martensitic transformation in NiTi, Scr. Mater. 50 (2004) 193.
the mechanical behavior of single crystal NiTi shape memory alloys and [53] G. Murasawa, S. Yoneyama, T. Sakuma, M. Takashi, Influence of cyclic loading
related polycrystalline phenomenon, Mater. Sci. Eng. A 317 (2001) 85. on inhomogeneous deformation behavior arising in NiTi shape memory alloy
[21] Q.S. Zheng, Y. Liu, Prediction of the detwinning anisotropy in textured NiTi plate, Mater. Trans. 47 (2006) 780.
shape memory alloy, Philos. Mag. A 82 (2002) 665. [54] D. Favier, H. Louche, P. Schlosser, L. Orge as, P. Vacher, L. Debove, Homoge-
[22] S. Daly, G. Ravichandran, K. Bhattacharya, Stress-induced martensitic phase neous and heterogeneous deformation mechanisms in an austenitic poly-
transformation in thin sheets of Nitinol, Acta Mater. 55 (2007) 3593. crystalline Tie50.8 at.% Ni thin tube under tension. Investigation via

[23] P. Sittner, P. Luk as, V. Nov
ak, M.R. Daymond, G.M. Swallowe, In situ neutron temperature and strain fields measurements, Acta Mater. 55 (2007) 5310.
diffraction studies of martensitic transformations in NiTi polycrystals under [55] C. Efstathiou, H. Sehitoglu, Local transformation strain measurements in
tension and compression stress, Mater. Sci. Eng. A 378 (2004) 97. precipitated NiTi single crystals, Scr. Mater. 59 (2008) 1263.
[24] S.C. Mao, J.F. Luo, Z. Zhang, M.H. Wu, Y. Liu, X.D. Han, EBSD studies of the [56] A. Sch€ afer, M.-X. Wagner, Strain mapping at propagating interfaces in pseu-
stress-induced B2-B19’ martensitic transformation in NiTi tubes under uni- doelastic NiTi, in: European Symposium on Martensitic Transformations: EDP
axial tension and compression, Acta Mater. 58 (2010) 3357. Sciences, 2009, p. p.06031.
[25] D.Y. Li, X.F. Wu, T. Ko, The texture of Ti-51.5at.% Ni rolling plate and its effect [57] Y.J. He, Q.P. Sun, Rate-dependent domain spacing in a stretched NiTi strip, Int.
on the all-round shape memory effect, Acta Mater. 38 (1990) 19. J. Solids Struct. 47 (2010) 2775.
[26] J.H. Mulder, P.E. Thoma, J. Beyer, Anisotropy of the shape memory effect in [58] X. Zhang, P. Feng, Y. He, T. Yu, Q. Sun, Experimental study on rate dependence
tension of cold-rolled 50.8 Ti 49.2 Ni (at.%) sheet, Z. Met. 84 (1993) 501. of macroscopic domain and stress hysteresis in NiTi shape memory alloy
[27] H. Inoue, N. Miwa, N. Inakazu, Texture and shape memory strain in TiNi alloy strips, Int. J. Mech. Sci. 52 (2010) 1660.
sheets, Acta Mater. 44 (1996) 4825. [59] M.F.X. Wagner, A. Schaefer, Macroscopic versus local strain rates during
[28] Y. Liu, Z.L. Xie, J. Van Humbeeck, L. Delaey, Effect of texture orientation on the tensile testing of pseudoelastic NiTi, Scr. Mater. 63 (2010) 863.
martensite deformation of NiTi shape memory alloy sheet, Acta Mater. 47 [60] J. Lackmann, T. Niendorf, M. Maxisch, G. Grundmeier, H.J. Maier, High-reso-
(1999) 645. lution in-situ characterization of the surface evolution of a polycrystalline NiTi
152 G. Laplanche et al. / Acta Materialia 127 (2017) 143e152

SMA-alloy under pseudoelastic deformation, Mater. Charact. 62 (2011) 298. [78] P. Sittner, Y. Liu, V. Novak, On the origin of Lüders-like deformation of NiTi
[61] C. Bewerse, K.R. Gall, G.J. McFarland, P. Zhu, L.C. Brinson, Local and global shape memory alloys, J. Mech. Phys. Solids 53 (2005) 1719.
strains and strain ratios in shape memory alloys using digital image correla- [79] Y. Liu, Z. Xie, J. Van Humbeeck, L. Delaey, Asymmetry of stress-strain curves
tion, Mater. Sci. Eng. A 568 (2013) 134. under tension and compression for NiTi shape memory alloys, Acta Mater. 46
[62] B. Reedlunn, C.B. Churchill, E.E. Nelson, J.A. Shaw, S.H. Daly, Tension, (1998) 4325.
compression, and bending of superelastic shape memory alloy tubes, J. Mech. [80] Y. Liu, Z. Xie, J.V. Humbeeck, L. Delaey, Y. Liu, On the deformation of the
Phys. Solids 63 (2014) 506. twinned domain in Niti shape memory alloys, Philos. Mag. A 80 (2000) 1935.
[63] D. Depriester, A. Maynadier, K. Lavernhe-Taillard, O. Hubert, Thermo- [81] F. Takei, T. Miura, S. Miyazaki, S. Kimura, K. Otsuka, Y. Suzuki, Stress-induced
mechanical modelling of a NiTi SMA sample submitted to displacement- martensitic transformation in a Ti-Ni single crystal, Scr. Metall. 17 (1983) 987.
controlled tensile test, Int. J. Solids Struct. 51 (2014) 1901. [82] M. Nishida, S. Ii, K. Kitamura, T. Furukawa, A. Chiba, T. Hara, K. Hiraga, New
[64] C. Elibol, M.F.X. Wagner, Investigation of the stress-induced martensitic deformation twinning mode of B190 martensite in Ti-Ni shape memory alloy,
transformation in pseudoelastic NiTi under uniaxial tension, compression and Scr. Mater. 39 (1998) 1749.
compressioneshear, Mater. Sci. Eng. A 621 (2015) 76. [83] A.P. Stebner, S.C. Vogel, R.D. Noebe, T.A. Sisneros, B. Clausen, D.W. Brown,
[65] T. Ezaz, H. Sehitoglu, W. Abuzaid, H.J. Maier, Higher order twin modes in A. Garg, L.C. Brinson, Micromechanical quantification of elastic, twinning, and
martensitic NiTidthe (20-1) case, Mater. Sci. Eng. A 558 (2012) 422. slip strain partitioning exhibited by polycrystalline, monoclinic nickel-
[66] S. Dilibal, Investigation of nucleation and growth of detwinning mechanism in etitanium during large uniaxial deformations measured via in-situ neutron
martensitic single crystal NiTi using digital image correlation, Metallogr. diffraction, J. Mech. Phys. Solids 61 (2013) 2302.
Microstruct. Anal. 2 (2013) 242. [84] N. Ono, A. Satoh, H. Ohta, A discussion on the mechanical properties of shape
[67] K.M. Knowles, D.A. Smith, The crystallography of the martensitic trans- memory alloys based on a polycrystal model, Mater. trans. JIM 30 (1989) 756.
formation in equiatomic nickel-titanium, Acta Metall. 29 (1981) 101. [85] Y. Sutou, T. Omori, K. Yamauchi, N. Ono, R. Kainuma, K. Ishida, Effect of grain
[68] T. Inamura, T. Nishiura, H. Kawano, H. Hosoda, M. Nishida, Self-accommoda- size and texture on pseudoelasticity in CueAleMn-based shape memory wire,
tion of B190 martensite in TieNi shape memory alloys. Part III. Analysis of Acta Mater. 53 (2005) 4121.
habit plane variant clusters by the geometrically nonlinear theory, Philos. [86] J.D. Eshelby, The determination of the elastic field of an ellipsoidal inclusion,
Mag. 92 (2012) 2247. and related problems, Proc. R. Soc. Lond. A Math. Phys. Eng. Sci. 241 (1957)
[69] P. Boullay, D. Schryvers, J.M. Ball, Nano-structures at martensite macrotwin 376.
interfaces in Ni65Al35, Acta Mater. 51 (2003) 1421. [87] G.B. Olson, M. Cohen, Dislocation theory of martensitic transformations, in:
[70] R.D. James, K.F. Hane, Martensitic transformations and shape-memory ma- F.R.N. Nabarro (Ed.), Dislocations in Solids, vol. 7, Elsevier Science Publishers,
terials, Acta Mater. 48 (2000) 197. 1986, p. p.297.
[71] X. Balandraud, D. Delpueyo, M. Gre diac, G. Zanzotto, Almost compatible mi- [88] E. Patoor, D.C. Lagoudas, P.B. Entchev, L.C. Brinson, X. Gao, Shape memory
crostructures in shape memory alloys, Acta Mater. 58 (2010) 4559. alloys, Part I: general properties and modeling of single crystals, Mech. Mater.
[72] B. Muntifering, R. Pond, L. Kovarik, N.D. Browning, P. Müllner, Intra-variant 38 (2006) 391.
substructure in NieMneGa martensite: conjugation boundaries, Acta Mater. [89] J.R. Patel, M. Cohen, Criterion for the action of applied stress in the martensitic
71 (2014) 255. transformation, Acta Metall. 1 (1953) 531.
[73] M.J. Szczerba, R. Chulist, Detwinning of a non-modulated NieMneGa [90] T. Inamura, Y. Yamamoto, H. Hosoda, H.Y. Kim, S. Miyazaki, Crystallographic
martensite: from self-accommodated microstructure to single crystal, Acta orientation and stress-amplitude dependence of damping in the martensite
Mater. 85 (2015) 67. phase in textured TieNbeAl shape memory alloy, Acta Mater. 58 (2010) 2535.
[74] B. Muntifering, L. Kovarik, N.D. Browning, R.C. Pond, W.B. Knowlton, [91] V. Raghavan, M. Cohen, Growth path of a martensitic particle, Acta Metall. 20
P. Müllner, Stress-assisted removal of conjugation boundaries in non- (1972) 779.
modulated NieMneGa by coordinated secondary twinning, J. Mater. Sci. 51 [92] G. Olson, M. Cohen, Thermoelastic behavior in martensitic transformations,
(2016) 457. Scr. Metall. 9 (1975) 1247.
[75] P. Müllner, A.H. King, Deformation of hierarchically twinned martensite, Acta [93] T.M. Brill, S. Mittelbach, W. Assmus, M. Mullner, B. Luthi, Elastic properties of
Mater. 58 (2010) 5242. NiTi, J. Phys. Condens. Matter 3 (1991) 9621.
[76] R.C. Pond, B. Muntifering, P. Müllner, Deformation twinning in Ni2MnGa, Acta [94] S.D. Prokoshkin, A.V. Korotitskiy, V. Brailovski, K.E. Inaekyan, S.M. Dubinskiy,
Mater. 60 (2012) 3976. Crystal lattice of martensite and the reserve of recoverable strain of thermally
[77] Y. Liu, Z. Xie, J. Van Humbeeck, L. Delaey, Some results on the detwinning and thermomechanically treated Ti-Ni shape-memory alloys, Phys. Met.
process in NiTi shape memory alloys, Scr. Mater. 41 (1999) 1273. Metallogr. 112 (2011) 170.