DOI: 10.1002/adem.

201200246
COMMUNICATION

Multiple Memory Shape Memory Alloys**

By Mohammad Ibraheem Khan, Andrew Pequegnat* and Y. Norman Zhou*

Until now, shape memory alloys (SMAs) have been largely limited to ‘‘remembering’’ a single memory.
In other words, monolithic components only possess a single set of functional properties. The current
work describes how theorized change to local chemical composition induced through laser processing
enables controlled augmentation of transformation temperatures. Proof of concept was demonstrated
by locally embedding multiple shape memories into a monolithic NiTi component. This novel technique
overcomes traditional fabrication challenges and promises to enhance SMA functionality and facilitate
novel applications through producing a new class of smart materials; namely multiple memory
materials (MMMs).

The unique capabilities of shape memory alloys (SMAs) mation. As a result, these functional properties are closely
have made them ideal engineering materials for novel devices linked to phase transformation temperatures.[4,5] The high
currently implemented in numerous industries including sensitivity of transformation temperatures to alloy composi-
medical, automotive, aerospace, and micro-electronics. Their tion and processing history[6–9] has made efforts to create
potential in high profile applications such as non-invasive application-specific alloys problematic. Furthermore, tradi-
surgical tools, biomedical implants, and energy harvesting tional SMA fabrication technologies are performance limiting
clean technologies are driving innovation and application since monolithic components consist of only a single set
of these smart materials.[1–3] Among all SMAs, NiTi alloys of phase transformation characteristics. A novel material
are the most commonly applied due to superior strength, processing protocol is therefore required to overcome these
ductility, recoverable strains, corrosion resistance, biocompat- challenges to enable further innovation and application.
ibility, and stability of transformation temperatures. In the past, various techniques have been applied to SMAs
The functional properties of SMAs, namely the shape in attempts to control local functionality. Specific methods
memory effect and pseudoelasticity, are attributed to a previously attempted included, direct electric resistance heat
reversible thermoelastic austenite-martensite phase transfor- treatment (DERHT),[10,11] direct hot air heat treatment
(DHAHT),[10] powder metallurgy (PM),[12–15] gradient anneal-
ing,[16] laser annealing,[17] and joining of multiple alloys.[18]
[*] Dr. M. I. Khan
All of these techniques aimed to enable designs that utilize
Smarter Alloys, Inc., MaRS Centre, South Tower, 101 College
more than one set of functional properties in a monolithic
Street, Suite 200, Toronto, ON M5G 1L7, Canada
component. However, inherent processing issues restricted
E-mail: ibraheem@smarteralloys.com
their widespread application. For example, local heat treat-
A. Pequegnat, Prof. Y. N. Zhou
Centre for Advanced Materials Joining (CAMJ), University of ment methods can only be applied to simple geometries, are
Waterloo, 200 University Avenue West, Waterloo, ON N2L time intensive, and provide a lack of resolution in functional
3G1, Canada properties.[10,11,16,17] PM methods exhibited issues with
E-mail: apequegn@uwaterloo.ca; nzhou@uwaterloo.ca porosity, poor surface finish, and composition inhomogeneity
A. Pequegnat, Prof. Y. N. Zhou leading to undesirable intermetallic compound (IMC) forma-
Centre for Bioengineering and Biotechnology (CBB), University tion.[13–15] In contrast, joining of multiple alloys showed to be
of Waterloo, 200 University Avenue West, Waterloo, ON N2L relatively effective for simple devices.[18] However, challenges
3G1, Canada associated with joining dissimilar SMAs (i.e., IMC formation)
[**] The Authors would like to acknowledge the support of the and more complex designs again limit application. Hence,
Natural Sciences and Engineering Research Council of Canada past attempts at achieving multiple functional properties in
(NSERC). Also, a special thanks to Siu Kei Tang, Billy Tam, a single component have not been successful. A rapid
Matt Daly, and Jeff Wang of the microwelding subgroup of the processing protocol is still required that is capable of high
Centre for Advanced Materials Joining (CAMJ). resolution and avoids the aforementioned issues.
Supporting Information is available from the Wiley Online The current article presents a novel laser processing
Library or from the author. technique that locally augments transformation temperatures

386 wileyonlinelibrary.com ß 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ADVANCED ENGINEERING MATERIALS 2013, 15, No. 5
M. I. Khan et al./Multiple Memory Shape Memory Alloys

COMMUNICATION
of SMAs. Specifically, theorized change to local chemical vapor pressures of each element. Assuming ideal conditions at
composition via controlled vaporization was implemented to high temperatures, equilibrium vapor pressure of both Ni and
modify the Ni:Ti ratio in the binary NiTi system. Remarkably, Ti can be calculated over the alloy mixture of the molten pool
this work resulted in proof of concept by successfully via Equation (1):[25,26]
embedding multiple memories in a monolithic NiTi compo-
Pi ¼ Xi P0i (1)
nent. This was achieved with the necessary resolution, while
avoiding detrimental defects observed in competing technol- where Xi is the mole fraction of the element i and is P0i
ogies. This unprecedented technique overcomes traditional the equilibrium vapor pressure of the element i over pure i.
fabrication challenges while enabling the realization of a new The equilibrium pressure P0i of each alloying element was
class of smart materials known as multiple memory materials calculated with respect to temperature using Equation (2):[37]
(MMMs).
B
log10 ðP0i Þ ¼ A þ þ Clog10 ðTÞ þ DT þ ET 2 (2)
1. Results and Discussion T
where A,B,C,D, and E are constant coefficients and T refers to
1.1. The Vaporization Concept
temperature. The corresponding coefficients for Ni and Ti that
Composition changes induced by vaporization using
were used in this study are provided in Table 1.[37]
high power density processing methods (e.g., laser beam,
Calculated vapor pressures of Ni and Ti over an equiatomic
electron beam, etc.) is well documented in literature.[19–27]
NiTi alloy as a function of temperature are presented in
Prior studies have concluded that local loss of elemental
Figure 1. Significant differences in vapor pressures were
constituents can be detrimental to the properties of a bulk
observed between the two elemental constituents, with Ni
material. For example, it has been reported that the loss of
being substantially larger over the plotted temperature range.
certain alloying elements during laser processing was
responsible for reduction in solid solution strengthening in
Al alloys (i.e., due to loss of Mg)[20,23] or decreased corrosion Table 1. Constant coefficients for calculation of equilibrium vapor pressure (P0i ) of Ni
and Ti.[37]
resistance in stainless steels (i.e., due to loss of Cr).[22,24–27]
The concept of utilizing laser induced vaporization as an
advantage has not yet been investigated. However, by A B C D E

controlling elemental loss during laser processing, the local Ni
214.3
3.52E þ 03 7.49E þ 01
1.80E-02 1.51E-06
properties of an alloy can theoretically be precisely tailored Ti
194.8742
8.27E þ 03 6.83E þ 01
1.73E-02 1.55E-06

for that of a specific application. NiTi SMAs are a prime

example of a group of materials that are extremely sensitive to
change in composition, where the advantages of elemental
loss during processing can be realized. Slight modifications
to composition have been shown to substantially alter
transformation temperatures which play a critical role in
their functional properties. For example, composition changes
of 1 at.% have been reported to result in an 80–100 K change in
transformation temperatures of a binary NiTi SMA.[8,9,28]
Similarly, transformation temperatures and temperature/
stress hysteresis in ternary (or quaternary, etc.) NiTi systems
and other SMAs, such as Cu
Al alloys, were also found to
be highly composition dependent.[29–36] A laser induced
controlled vaporization protocol is therefore proposed
for quickly modifying the local functional properties of
monolithic SMA components.

1.2. Selective Vaporization of Elemental Constituents

To apply this novel vaporization protocol to NiTi SMAs
it is necessary to determine the relative vaporization flux of
each elemental constituent (i.e., Ni and Ti) to aid in predicting
the change in composition. The relative loss of elemental
constituents during laser processing is determined largely by
their volatility.[19–27] More specifically, the pressure gradient
has been shown to have the largest effect on vaporization
Fig. 1. Nickel and Titanium vapor pressures over an equiatomic NiTi mixture.
flux during laser processing.[26,27] Therefore, relative loss of Significantly higher vapor pressure of Nickel corresponds to a comparatively larger
elemental constituents was approximated by calculating the vaporization flux.

ADVANCED ENGINEERING MATERIALS 2013, 15, No. 5 ß 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim http://www.aem-journal.com 387
 M. I. Khan et al./Multiple Memory Shape Memory Alloys
COMMUNICATION

Hence, the vaporization flux of Ni can be expected to be larger Relative composition change of laser processed materials is
than that of Ti during laser processing of a near equiatomic highly dependent on a balance between the vaporization
NiTi SMA; the basis of the controlled constituent vaporization flux of elemental constituents and molten pool dilution.[26,27]
protocol developed in this work. Since vaporization flux and molten pool dilution are closely
linked to laser processing parameters, it was of interest to
1.3. Augmentation of Transformation Temperatures investigate the effects of peak power and pulse duration on the
A decrease in the Ni:Ti ratio translates to an increase compositionally sensitive transformation temperatures. Peak
in phase transformation temperatures.[8,9,38–41] Therefore, power and pulse duration, however, were found to have only
the anticipated loss of Ni following laser processing was a minor effect compared to the number of applied laser pulses.
expected to increase the phase transformation temperatures; Effects of pulse duration on transformation temperatures
as was found by Meier et al., studying selective laser are shown in Figure 3. Using a laser power of 0.6 kW,
melting of NiTi.[42] This hypothesis agrees with results increasing pulse duration increased DSC peak intensity. The
from the differential scanning calorimetry (DSC) analysis more prominent transformation peaks were a result of an
presented in Figure 2. After processing the NiTi alloy, a increase in the processed material to base metal volume ratio
second set of transformation peaks appeared at higher with increased pulse duration. Transformation temperatures
temperatures compared to those of the solutionized base of processed material decreased slightly (i.e., by <20 K) with
metal. Furthermore, during pulsed Nd:YAG laser processing increasing pulse duration, as shown in Figure 4. Since
it has been shown that vaporization primarily occurs during vaporization flux is highest during the first few milliseconds
the first few milliseconds of laser application, when utilizing of the laser pulse and decreases drastically with pulse
the keyhole laser interaction mode.[27] Therefore, a compound duration,[27] dilution of the molten pool caused by an increase
effect resulted when using multiple pulses that maximized in volume was likely the mechanism behind decreasing
vaporization. Accordingly, transformation temperatures transformation temperatures with longer pulse durations.
were found to increase with each additional laser pulse, Hence, vaporization flux decreased and the pool volume
identifying the possibility for high resolution through this increased with increased pulse duration, leading to smaller
laser processing protocol (Figure 2). changes in final composition; consistent with literature.[27]

Fig. 2. Thermo-analytical DSC analysis. Following laser processing with 0.6 kW peak power and 30 ms pulse duration, a second set of phase transformation peaks were identified at
higher temperatures. During heating, the endothermic B2 austenite peaks were measured. Similarly, the exothermic B190 martensite peaks were measured during cooling. Increasing
the number of pulses caused a compound vaporization effect, raising the transformation temperatures with increasing number of pulses.

388 http://www.aem-journal.com ß 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ADVANCED ENGINEERING MATERIALS 2013, 15, No. 5
M. I. Khan et al./Multiple Memory Shape Memory Alloys

COMMUNICATION
Fig. 3. The effect of pulse duration on transformation temperatures as measured by DSC analysis. Less significant changes in transformation temperature of the processed region
were observed with increasing pulse duration and a laser power of 0.6 kW compared to number of laser pulses.

The effects of laser power on transformation temperatures

are shown in Figure 4. Only a small decrease in transformation
temperature (i.e., <10 K) was observed when increasing laser
power to 0.9 kW from 0.6 kW. Similar to the decrease in
transformation temperatures with increasing pulse duration,
increasing the laser power can also cause an increase in
dilution of the molten pool leading to lower transformation
temperatures.
The observed correlations between laser parameters and
transformation temperatures accentuate the controllability of
the process. Even though the number of pulses has the greatest
effect on changes in transformation temperatures, the laser
power and pulse duration are necessary to achieve application
specific properties and the highest possible resolution of the
process. Using multiple pulses having different power and
pulse duration parameters or by using unique temporal
profiles, local phase transformation characteristics can be
precisely selected.

1.4. Verification of the Vaporization Mechanism

Thus far thermoanalytical DSC analysis was shown to be a
useful tool in identifying the effects of laser parameters on
Fig. 4. Effect of laser power on transformation temperatures. Ms temperature versus transformation temperatures. However, it is still necessary to
pulse duration for the 0.6 and 0.9 kW laser power conditions. Less than 10 K difference in
transformation temperatures was observed between the 0.6 and 0.9 kW laser power
definitively confirm that vaporization was in fact the primary
conditions. mechanism at play.

ADVANCED ENGINEERING MATERIALS 2013, 15, No. 5 ß 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim http://www.aem-journal.com 389
 M. I. Khan et al./Multiple Memory Shape Memory Alloys

found to be consistent with monoclinic B190 martensite,

COMMUNICATION

There have been some inconsistencies in prior studies that

investigated the effects of composition on transformation whereas the adjacent base metal possessed the anticipated
temperatures.[38–40] However, investigators have consistently cubic B2 austenite crystal structure. This phenomenon
reported that when the composition of solutionized NiTi resulted from the austenite phase being thermodynamically
alloys was near or below equiatomic, the maximum Ms stable in the base metal and the martensite phase being
temperature is approximately 338 K (65 8C).[8,9] When Ni stable in the processed region at room temperature; consistent
composition is below 49 at.%, transformation temperatures with DSC analysis in Figure 2. Transmission electron
become independent of composition due to Ti saturation. microscopy (TEM) revealed nodular particles surrounded
The Ms transformation temperatures plotted with respect by a twinned martensite matrix in the processed material
to the number of laser pulses in Figure 5 shows a steady [Figure 6(c)]. EDS analysis identified the particles as the
increase until the 338 K (65 8C) maximum Ms temperature Ti2Ni intermetallic; detecting Ti and Ni atomic compositions
was achieved. This suggests that with increasing number of 66.3 and 33.6 at.%, respectively. The presence of the Ti-rich
of pulses a threshold exists where the processed material Ti2Ni intermetallic was further evidence of the compositional
becomes Ti-rich. Therefore, these DSC results strongly change of the processed alloy, since Ti was no longer
correlate the augmentation of transformation temperature soluble in NiTi. Significant amounts of Ti-rich IMCs were
to a decrease in Ni:Ti ratio. only found in over-processed samples where Ti saturation
Compositional and phase analysis further confirmed that had occurred.
changes in local chemistry were the primary mechanism Other possible mechanisms responsible for changes
behind the altered transformation temperatures. An energy- in transformation temperatures in NiTi alloys include
dispersive X-ray spectroscopy (EDS) line scan including the changes in dislocation density and growth of Ni-rich
base metal and processed region (i.e., after five laser pulses) precipitates.[1,5,29,43] However, lower energy density laser
was performed in order to detect Ni depletion as shown in processes that induce melting but limit vaporization lead to
Figure 6(a). Average base metal Ni content (50.07  0.53 at.% decreases in the transformation temperatures of binary NiTi
Ni) was significantly higher when compared to the processed alloys.[44] The melting of the NiTi alloy eliminates the effects
material (48.36  0.84 at.% Ni). Furthermore, EDS analysis of all previous processing by annihilation of dislocations
performed on the captured plume produced during laser and dissolution of Ni-rich precipitates. The presence of high
processing showed 75.5 at.% Ni and 24.5 at.% Ti; confirming density dislocations and Ni-rich precipitates were also not
the expected higher Ni vaporization flux compared to that of observed during TEM analysis in this work. Hence, vaporiza-
Ti. The thermo-analytical DSC results were also verified using tion was the primary mechanism behind the augmentation
room temperature micro-X-ray diffraction (XRD); shown in of transformation temperatures observed in this study.
Figure 6(b). Diffraction patterns of the processed region were
1.5. Multiple Memory Material Proof of Concept
MMMs fabricated using this controlled constituent vapor-
ization protocol have great potential in aiding in the
realization and development of novel SMA devices. Proof
of concept was demonstrated through the local modification
of the transformation temperatures in the NiTi component
shown in Figure 7. Commercially available material that
exhibits a single set of transformation temperatures (i.e., single
shape memory) was locally processed yielding processed
regions with surface morphology similar to that of a pulsed
laser weld, as illustrated in Figure 7(a). Pulsed Nd:YAG laser
processing at the annotated sites caused a localized increase
in phase transformation temperatures while the phase
transformation dynamics in the untreated regions remained
unchanged. This increase in transformation temperatures
allowed for a two-stage (i.e., two-memory) actuation through
sequential heating of the specimen beyond the austenite finish
temperatures of the base metal and the processed material,
respectively. A schematic and photograph of the actuation
process are shown in Figure 7(b) and (c). At low temperatures
the entire sample was in the martensitic phase and easily
Fig. 5. Martensite start (Ms) temperature plotted against number of laser pulses. A deformed (i.e., by detwinning) into a ‘‘C-like’’ shape. Upon
maximum Ms of approximately 338 K (65 8C) was achieved with increasing pulses. This heating to a temperature greater than T1, the bulk unprocessed
maximum Ms temperature was a result of the limited solubility of Ti in NiTi below
equiatomic composition. The error bars represent one standard deviation of a 6 sample
base metal transforms to austenite and recovers its original
population. shape, while the processed regions remain in the martensite

390 http://www.aem-journal.com ß 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ADVANCED ENGINEERING MATERIALS 2013, 15, No. 5
M. I. Khan et al./Multiple Memory Shape Memory Alloys

COMMUNICATION
Fig. 6. Compositional and phase analysis of laser processed NiTi. (a) EDS line scan across laser processed region. A decrease in the Ni:Ti ratio was measured in the processed region
(i.e., after five laser pulses) compared to the adjacent base metal resulting from the preferential vaporization of Nickel. (b) Room temperature micro-XRD analysis identified the B190
martensite phase in the processed region and the B2 austenite phase in the adjacent base metal. As a result of the local change in composition, both martensite in the processed region
and austenite in the base metal were stable at room temperature in the monolithic component. (c) TEM analysis identified Ti-rich Ti2Ni intermetallics in heavily processed samples
(i.e., >5 pulses). The presence of Ti2Ni was a result of Ti no longer being soluble in NiTi below equiatomic composition.

phase. Subsequently, heating to above T2 induces the of additional alloying elements can also be achieved using
transformation of the processed regions to austenite and laser processing on more complex alloy systems.[45] MMMs
the recovery of the second shape memory (see Supporting have the potential to greatly enable the engineering of
Information Video). complex SMA devices. SMA functionality has been immen-
In addition to the demonstrated increased functionality of sely improved where concepts such as, but not limited to,
MMMs, there are numerous benefits compared to competing dynamic actuation via passive atmospheric control, self-
technologies. MMMs benefit from the speed and accuracy biasing monolithic actuators, and local tuning of functional
accredited to laser processing. Also, defects such as poor properties are now feasible.
surface finish, porosity, and detrimental IMC formation
were not found to be issues in this study. Parallel studies
are underway to further characterize and understand the 3. Experimental
performance of this new class of smart materials. Commercially available 0.37 mm thick SE508 Nitinol strip manu-
factured by Nitinol Devices and Components, Inc. was used in this
study. The chemistry for this particular NiTi alloy was 55.8 wt%
2. Conclusions
(50.7 at.%) Ni and 44.2 wt% (49.3 at.%) Ti with maximum oxygen
In conclusion, local augmentation of transformation and carbon contents of 0.05 wt.% and 0.02 wt.%, respectively. The
temperatures creating MMM NiTi SMAs can be achieved as-received cold-rolled material was heat treated for 3.6 ks at 1073 K
through controlled constituent vaporization. With an under- (800 8C) followed by rapid quenching to remove the effects of prior
thermo-mechanical processing and ensure no R-phase was present.
standing of the effects of pulsed laser processing on
A dilute solution of hydrofluoric and nitric acid was used to remove a
vaporization of elemental constituents, the local transforma-
thick black oxide before laser processing.
tion temperatures of SMA components can be tailored to that An LW50 A Miyachi Unitek pulsed Nd:YAG laser system with a
of a specific application. This study focused on the binary NiTi 1.06 mm wavelength, 400 mm nominal spot diameter, and a Gaussian
system however, the MMM technology has potential for use spatial profile was used in this work. This particular system was
on other SMAs or any other material that may benefit from equipped with a power monitor allowing for accurate in situ
constituent vaporization. Synergistic vaporization and mixing assessment of incident power output. Processing was conducted on

ADVANCED ENGINEERING MATERIALS 2013, 15, No. 5 ß 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim http://www.aem-journal.com 391
 M. I. Khan et al./Multiple Memory Shape Memory Alloys
COMMUNICATION

Fig. 7. MMM proof of concept. (a) Four select regions of a commercially available NiTi alloy were laser processed, increasing the phase transformation temperatures (not to scale).
(b) Schematic demonstration of multiple actuations of the laser processed component. (c) Actual images of processed component during different stages of actuation. After cooling
below the martensite finish temperatures of both the processed region (T2) and base metal (T1) the component was deformed by detwinning in the martensite phase to a ‘‘C-like’’ shape.
Upon heating past the austenite phase transformation temperatures of the base metal, the base metal recovered its shape while the processed region remains strained and in the
martensite phase. Further heating past the austenite transformation temperatures of the processed regions lead to a second transformation and complete recovery of the original
shape of the processed NiTi component. Hence, the monolithic component possessed two distinct shape memories after processing.

monolithic NiTi sheets. The top and bottom were shielded with argon Metallographic cross-sections were prepared for EDS and micro
to avoid oxidation; a flow rate of 14.2 L min
1 (30 CFH) was found XRD analysis by polishing down to 1 mm diamond followed by
to be optimum. The minimum peak power criteria included full 0.03 mm colloidal silica. Samples were etched in a 14 mL HNO3, 3 mL
penetration through the NiTi strip. Peak power parameters of 0.6 and HF, and 82 mL H2O solution for 25 s to reveal microstructure.
0.9 kW with a 1 to 30 ms square pulse were used in this work. In order to measure Ni depletion in the processed samples and
DSC analysis was conducted using a Thermal Analysis Q2000 analyze the composition of the vapor plume, a Jeol JSM-6460 scanning
system equipped with a refrigerated cooling system. DSC curves were electron microscope (SEM) equipped with an INCA X-Sight 350 EDS
recorded in a temperature range from 198 K (
75 8C) to 423 K (150 8C) analysis aperture was used for chemical analysis. Chemical measure-
using a controlled heating and cooling rate of 5 K min
1. The ments were collected at a 20 keV acquisition energy using a nominal
martensite start (Ms), martensite finish (Mf), austenite start (As), 1 mm interaction spot. NiTi SMA samples were polished and etched
and austenite finish (Af) phase transformation temperatures were following the same procedure outlined for optical microscopy
measured using the TA Universal Analysis software (version 4.5) by analysis. Prior to SEM analysis, the samples were carbon coated for
calculating the intercept of the tangents to the phase transformation improved conductivity. For analysis of the vapor plume, vaporized
peak slopes and the base line; as per the ASTM F2004-05 standard. For elements were collected by placing a quartz tube co-axial to the
simplification, the Ms temperature was used as a representative incident laser beam where vapor condensed on the inner wall. A laser
measure of overall transformation changes; consistent with literature. peak power of 0.6 kW with 30 ms pulse duration was used for this test.
Samples were carefully extracted using a Struers, Ltd., Acutome 50 Crystallographic phases in laser processed samples were identified
precision saw with a 0.01 mm tolerance. Due to the small specimen using XRD analysis. Room temperature XRD patterns were captured
dimensions, remnant base metal material remained on the cut using a Rigaku SA-HF3 (1.54 Å Copper-Ka) X-ray source equipped with
samples. Hence, transformation peaks from the base material and a 500 mm collimator operating at an excitation voltage of 50 kV and
processed regions were measured. 40 mA current.

392 http://www.aem-journal.com ß 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ADVANCED ENGINEERING MATERIALS 2013, 15, No. 5
M. I. Khan et al./Multiple Memory Shape Memory Alloys

COMMUNICATION
A Philips CM12 TEM equipped with EDS and an accelerating [23] A. Blake, J. Mazumder, J. Eng. Ind. 1985, 107, 275.
voltage of 120 keV was used in this study. The TEM sample [24] M. M. Collur, A. Paul, T. Debroy, Metall. Trans. B 1987,
preparation was conducting using a Fischione ion milling system at 18B, 733.
183 K (
90 8C) to minimize heat induced artifacts.
[25] X. He, T. Debroy, P. W. Fuerschbach, J. Phys. D: Appl.
Received: July 13, 2012 Phys. 2003, 36, 3079.
Final Version: December 20, 2012 [26] X. He, T. Debroy, P. W. Fuerschbach, J. Appl. Phys. 2004,
Published online: February 15, 2013 96, 4547.
[27] M. Jandaghi, P. Parvin, M. J. Torkamany, J. Sabbaghza-
[1] T. G. Frank, W. Xu, A. Cuschieri, Minim. Invasive Ther. deh, J. Phys. D: Appl. Phys. 2008, 41, 235503.
2000, 9, 89. [28] X. Ren, K. Otsuka, Mater. Sci. Forum 2000, 327-328, 429.
[2] C. Smith, A. Villanueva, K. Joshi, Y. Tadesse, S. Priya, [29] H. Funakubo, Shape Memory Materials, Gordon and
Smart Mater. Struct. 2011, 20, 012001. Breach Scientific Publishers, New York, NY 1987.
[3] A. C. Keefe, G. P. McKnight, G. A. Herrera, P. A. Bedegi, [30] K. H. Eckelmeyer, Scr. Metall. 1976, 10, 667.
C. B. Churchill, A. L. Browne, J. Brown, in Proceedings of [31] H. C. Donkersloot, J. H. N. Van Vucht, J. Less-Common
the ASME 2011 Conference on Smart Materials, Adaptive Met. 1970, 20, 83.
Structures and Intelligent Systems, Phoenix, Arizona 2011, [32] P. G. Lindquist, C. M. Wayman, Proc. MRS Int. Mtg. Adv.
p. 669. Mater. 1989, 9, 123.
[4] X. Huang, Y. Liu, Scr. Mater. 2001, 45, 153. [33] R. Zarnetta, R. Takahashi, M. L. Young, A. Savan,
[5] C. P. Frick, A. M. Ortega, J. Tyber, A. El. M. Maksound, Y. Furuya, S. Thienhaus, B. Maaß, M. Rahim, J. Frenzel,
H. J. Maier, Y. Liu, K. Gall, Mater. Sci. Eng., A 2005, 405, H. Brunken, Y. A. Chu, V. Srivastava, R. D. James,
34. I. Takeuchi, G. Eggeler, A. Ludwig, Adv. Funct. Mater.
[6] M. H. Wu, Mater. Sci. Forum 2002, 394, 285. 2010, 20, 1917.
[7] D. Miller, D. Lagoudas, Mater. Sci. Eng., A 2001, 308, 161. [34] M. Zarinejad, Y. Liu, Adv. Funct. Mater. 2008, 18, 2789.
[8] J. Frenzel, E. P. George, A. Dlouhy, Ch. Somsen, [35] J. Cui, Y. S. Chu, O. O. Famodu, Y. Furuya, J. Hattrick-
M. F.-X. Wagner, G. Eggeler, Acta Mater. 2010, 58, 3444. Simpers, R. D. James, A. Ludwig, S. Thienhaus,
[9] W. Tang, Metall. Mater. Trans., A 1997, 28A, 537. M. Wuttig, Z. Zhang, I. Takeuchi, Nat. Mater. 2006, 5,
[10] F. Miura, M. Mogi, Y. Ohura, Eur. J. Orthod. 1988, 10, 187. 286.
[11] G. Airoldi, G. Riva, J. Physique IV, Colloque C2, sup. J. [36] V. Recarte, R. B. Pérez-sáez, J. San Juan, E. H. Bocanegra,
Physique III 1995, 5, 397. M. L. Nó, Metall. Mater. Trans., A 2002, 33A, 2581.
[12] D. Cluff, S. Corbin, R. Gorbet, in Proceedings of Cans- [37] C. L. Yaws, Handbook of Vapor Pressure, Vol. 4, Gulf
mart 2009, the 12th Cansmart Meeting – International Publishing Co, Houston, TX 1995.
Workshop on Smart Materials and Structures, (Ed: G. [38] F. W. Wang, W. J. Beuhler, S. J. Pickart, J. Appl. Phys.
Akhras), Montreal, Quebec 2009, p. 293. 1965, 36, 3232.
[13] J. C. Hey, A. P. Jardine, Mater. Sci. Eng., A 1994, 188, 291. [39] S. R. Butler, J. E. Hanlon, R. J. Wasilewski, Met. Sci. J.
[14] M. Whitney, S. F. Corbin, R. B. Gorbet, Acta Mater. 2008, 1967, 1, 104.
56, 559. [40] D. P. Dautovich, Z. Melkvi, G. R. Purdy, C. V. Stager, J.
[15] B. Vamsi Krishna, S. Bose, A. Bandyopadhyay, Metall. Appl. Phys. 1966, 37, 2513.
Mater. Trans., A 2007, 38A, 1096. [41] R. J. Wasilewski, S. R. Butluer, J. E. Hanlon, D. Wordon,
[16] A. S. Mahmud, Y. Liu, T. H. Nam, Smart Mater. Struct. Metall. Trans. 1971, 2, 229.
2008, 17, 015031. [42] M. Meier, C. Haberland, J. Frenzel, R. Zarnetta, in
[17] Q. Meng, Y. Liu, H. Yang, B. S. Shariat, T. H. Nam, Acta Innovative Developments in Design and Manufacturing:
Mater. 2012, 60, 1658. Advanced Research in Virtual and Rapid Prototyping
[18] P. Sevilla, F. Martorell, C. Libenson, J. A. Planell, F. J. Gil, (Eds: P. J. Bártolo, M. A. Jorge, H. A. Almeida, J. M.
J. Mater. Sci.: Mater. Med. 2008, 19, 525. Matias, J. C. Vasco, J. B. Gaspar, M. A. Correia, N. C.
[19] A. Block-Bolten, T. W. Eager, in Trends in Welding Andre, N. F. Alves, P. P. Novo, P. G. Martinho, R. A.
Research in the United States (Ed: S. A. David), ASM Carvalho), CRC Press, Taylor & Francis Group, Boca
International, Novelty, OH 1982, p. 55. Raton, FL 2009, p. 233.
[20] D. W. Moon, E. A. Metzbower, Weld. J. Res. Sup. 1983, 62, [43] P. Filip, K. Mazanec, Scr. Mater. 2001, 45, 701.
53. [44] B. Tam, M. I. Khan, Y. Zhou, Metall. Mater. Trans., A
[21] A. Block-Bolten, T. W. Eager, Metall. Trans. B 1984, 15B, 2011, 42, 2166.
463. [45] M. Daly, A. Pequegnat, Y. Zhou, M. I. Khan, Smart Mater.
[22] P. A. A. Khan, T. Debroy, Metall. Trans. B 1984, 15B, 641. Struct. 2012, 21, 045018.

ADVANCED ENGINEERING MATERIALS 2013, 15, No. 5 ß 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim http://www.aem-journal.com 393