Materials and Design 32 (2011) 29232933

Contents lists available at ScienceDirect

Materials and Design

journal homepage: www.elsevier.com/locate/matdes

Characterization of a shape memory alloy hybrid composite plate subject

to static loading
Fathollah Taheri-Behrooz, Farid Taheri , Ramin Hosseinzadeh
Department of Civil and Resource Engineering, Dalhousie University, Halifax, NS, Canada

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

Article history: In this manuscript the design process of a SMA wire-reinforced hybrid composite (SMAHC) plate is intro-
Received 9 July 2010 duced in detail. The inuence of various parameters on the overall behavior of the SMAHC plate is dis-
Accepted 28 November 2010 cussed. These parameters include: the mechanical behavior of the constituents (i.e. the host composite
Available online 21 December 2010
and SMA), the SMA volume fraction, the temperature dependent loading effects and the fabrication pro-
cess. For that, a series of SMAHC plates were fabricated and tested under monotonic loading condition.
Keywords: The characteristics curves and formula for assessing the effect of the SMA at different temperatures
A. Polymer matrix composites
are presented. It would be demonstrated that the embedment of SMA wires could improve the overall
G. Coupon testing
H. Selection of components
structural response of the host material in terms of stiffness and strength at elevated temperatures. Also
capability of the one dimensional constitutive models in predicting the macroscopic stressstrain behav-
ior of SMAHC plates is veried experimentally.
2010 Elsevier Ltd All rights reserved.

1. Introduction
Shape memory alloys (SMAs) are gaining interest in the aero-
space, automotive, and medical industries due to their unique
characteristics, which include the shape memory effect, pseudo-
elasticity, and improved material behavior upon heating. Unlike
the extensive research into widely used applications of SMA as
an actuator muscle, the investigation into structural enhancement
via the use of SMA is relatively scarce. It should be noted that in all
of the aforementioned applications, it is the shape memory effect
of SMA that is utilized. This unique property that occurs by the
crystalline transformation of the alloy from martensite phase
(low stiffness) into austenite phase (high stiffness) upon heating
induces high tensile force in SMA. However, the type of application
addressed here, in the context of structural reinforcement by SMA,
is different from the conventional applications that often use SMA
as a miniature actuation muscle. The latter application takes
advantage of SMA retraction (i.e. the shape memory effect). In con-
trast, through transformation between its two crystalline phases,
the SMA reinforcement uses its higher thermally-induced stiffness
as embedded within a separate host substrate to enhance the over-
all response. Knowing that SMA retraction is affected by the mate-
rial deformation of the host substrate, the structural reinforcement
would be achieved by the shape memory effect of SMA.
Characterization of shape memory alloys dates back to 1969,
when the National Aeronautics and Space Administrative (NASA,
Corresponding author. Address: 1360 Barrington Street, Halifax, Nova Scotia,
Canada B3J 1Z1. Tel.: +1 902 494 3935; fax: +1 902 484 6635.
E-mail address: farid.taheri@dal.ca (F. Taheri).
USA) released a comprehensive report [1] that presented the con-
stitutive thermo-mechanical behavior of Nickel Titanium Novel
Ordinance Laboratory (NiTiNOL) wire. Since then, many research-
ers have addressed and formulated the SMA behavior based on
the materials introduced within the aforementioned report [26].
A systematic study of thermo-mechanical behavior of NiTiNOL
was conducted by Liang and Rogers [7], through which they pre-
sented a constitutive relationship for the formulation of crystalline
transformation from martensite to austenite phases as well as the
reverse process. Their proposed formulation is considered a basic
mathematical model of the transient behavior of NiTiNOL upon
heating or cooling between its two crystalline phases. Since then,
the application of SMA in structural reinforcement via using its
mitigating effect has been an extensively researched subject.
Among the various types of structural reinforcement, the use of
SMA within composite structures is an interesting subject due to
the many applications of composite materials in various industries.
Numerous applications use the characteristic of greater stiffness
induced in NiTiNOL alloys upon their transformation into the aus-
tenite phase. Many researchers [812] have proposed theoretical
and numerical models to investigate the effect of SMA reinforce-
ment in composite structures subject to different loadings; how-
ever, few of them have addressed this issue from a practical
perspective.
The research into experimental characterization of shape mem-
ory alloy reinforced hybrid composite (SMAHC) is still continuing.
Turner et al. [13], Turner [14,15] and Davis et al. [16], developed a
manufacturing method for fabrication of SMAHC beams and
characterized their behavior under thermo-mechanical loading.
Their manufacturing method is very similar to the manufacturing

0261-3069/$ - see front matter 2010 Elsevier Ltd All rights reserved.
doi:10.1016/j.matdes.2010.11.068
2924 F. Taheri-Behrooz et al. / Materials and Design 32 (2011) 29232933
method presented herein. Han et al. [17] used SMA springs for the
reinforcement of epoxy columns under buckling. They proposed
analytical, numerical and experimental methods to show how
SMA springs could be used to enhance the buckling behavior of
epoxy columns. Subsequently, Han et al. [18] presented their
method on enhancing the stability of composite plates by embed-
ding SMA wires within the plate. In their process, SMA wires were
activated by applying DC current.
2. Motivation
Although there are several works dedicated to the development
of formulation for structural behavior of SMA and SMA reinforced
composites, only a few of them address the parameters that one
should consider in the design and manufacturing of shape memory
alloy reinforced structures to achieve the best performance under
various loading conditions. In other words, only a few of the manu-
scripts have detailed the effect of design and manufacturing
parameters, such as the mechanical behavior of the constituents,
their volume contents, and the fabrication method on the overall
behavior SMAHC systems when subject to static loading. This void
in the literature motivated this investigation.
To address this issue, detailed steps, which should be followed
for the design and fabrication of a SMA reinforced composite are
presented in this paper. Next, the material behavior of the SMAHC
plates under monotonic loading are described and compared under
the ambient and elevated temperatures. The one dimensional con-
stitutive equation for SMAHC structures and rule of mixture (ROM)
formula, are presented and veried experimentally. Finally, modi-
cations for an improved performance design and production pro-
cess are recommended.
3. Scope and objectives
The design process for a shape memory alloy hybrid composite
is discussed in depth. Shape memory alloys are assumed to serve as
structural reinforcement for those applications wherein heating
would be a source of material degradation. In turn, the martens-
iteaustenite transformation of embedded SMA in a composite
substrate is postulated to mitigate the reduced stiffness of the
composite at elevated temperatures. The verication of this
assumption is the main objective of this study. Therefore, in the
rst experiment, the mechanical properties of the commercial
SMA wires referred to as NiTiNOL (Memry, Cop., USA) were evalu-
ated by a series of tensile tests. These tests were conducted on the
wires at different temperatures and under different pre-strains.
Recovery stress of 4% pre-strained of SMA wires were measured
experimentally at elevated temperatures and used as input data
to estimate the stressstrain response of the SMAHC plates. As sta-
ted, the embedded SMA wires are assumed to provide the reinforc-
ing attribute to the composite plates at elevated temperatures.
Subsequently, the manufacturing process for pre-straining the
SMA wires and sandwiching them between two composite lami-
nates is described. A series of monotonic tensile tests on the
SMAHC coupons were conducted at various temperatures and
the mechanical properties, as a function of temperature, were gen-
erated. The results were compared with the experimental data
from a series of similar tests on the non-reinforced composite cou-
pons. The ROM was used for evaluating the mechanical properties
of the hybrid composite, at the ambient and elevated tempera-
tures. The integrity and accuracy of the ROM in evaluating the
properties were veried by comparing the results to those ob-
tained experimentally. In addition, the capability of the earlier
developed constitutive stressstrain relationship for SMAHC struc-
tures were veried experimentally.
The focus of this research is therefore more on the reinforcing
attribute of SMA wires when incorporated within woven glass/
epoxy composites.
4. Macroscopic behavior of SMA hybrid composites
It has been shown experimentally that SMA wires in martensite
phase under monotonic tensile loading experience large deforma-
tion (typically as high as 7%) and recover to their original shape
when heated up to temperatures more than their activation tem-
perature. If SMA wires are restrained to retain their free shape
recovery by imposition of external constrains, a recovery stress
would be created by the deformed SMA against the constraint.
The amount of this recovery stress depends on the pre-strain im-
posed on the SMA, heating temperature and mechanical loads
[19,20]. In addition, as stated earlier, the stiffness of SMA wires is
a function of the applied temperature; for example, the stiffness
at the austenite phase (at elevated temperature) is almost three
times of the stiffness at the martensite phase (at ambient temper-
ature). The shape recovery and stiffness enhancement characteris-
tic of SMA wires due to change in temperature could be effectively
used to mitigate stiffness and strength degradation of structures
when subject to higher temperatures. One may incorporate SMA
wires in polymeric composite laminates to compensate material
degradation at elevated temperatures.
Fig. 1 schematically illustrates the macroscopic behavior of a
SMAHC plate in terms of its stressstrain behavior under mechan-
ical loads and isothermal conditions, as a function of its boundary
conditions.
In Fig. 1a a SMAHC is shown at ambient temperature. In this
case the SMA wires are elongated in the martensite phase before
being embedded in the matrix (the word matrix hereafter refers
to the host laminate composite). Fig. 1b and c show the same
SMAHC at an elevated temperature (i.e. at fully austenite phase)
under free and xed boundary conditions, respectively.
If the nal SMAHC plate is subjected to any mechanical or thermal
load, then the strain developed in the SMA wires could be estimated
by the following equation, as explained by Liang and Rogers [7]:
es e0 e 1
where es is the strain in SMA wire, e0 is the pre-strain applied to the
SMA wires before embedding them in the matrix, and e is the total
applied strain due to external tensile force (positive) or the strain
produced by SMAs activation recovery stress on the SMAHC speci-
men under unrestricted boundary conditions (negative).
The stress in SMA wires is a function of the level of pre-strain
and temperature, and could be expressed as follows:
rs es  Es 2
where rs and Es are the stress and modulus of elasticity of the SMA
wires, respectively. The asterisk sign in the equation denotes the
dependency of a parameter to the temperature.
A simple one dimensional constitutive equation can be devel-
oped, based on the mechanics-of-materials approach, to explain
the macroscopic behavior of SMA hybrid composites under
mechanical and thermal loads. Following assumptions are made
in developing the constitutive equation [7].
 The bond between the wires and matrix is perfect, resulting in
strain compatibility across the interface.
 The elastic modulus, strength, diameters of the wires and spac-
ing between the wires are uniform.
 The wires are continuous and parallel.
 The wires and matrix follow Hookes law (linearly elastic).
 The effect of thermal expansion is neglected.
 F. Taheri-Behrooz et al. / Materials and Design 32 (2011) 29232933
Fig. 1. Structure of a SMAHC plate and its constituents.
Assuming a SMAHC subjected to a uni-axial tensile load (F), the
force equilibrium on the cross-section could be written as:
F Fs Fm
where Fs and Fm are the loads taken by SMA wires and the matrix,
respectively.
By rewriting Eq. (3) in terms of stress, strain and cross-sectional
area of components, one obtains:
As Am
E  e Es  es  Em  em 
A A
By replacing the strain of the SMA wires from Eq. (1) into Eq. (4),
the macroscopic stressstrain relation of SMAHC could be obtained
as:
r Es  V s Em  V m  e Es  V s  e0
where E; Es ;
Em ; e; es ; em and A; As ; Am are the moduli, strains and
cross-sections of SMAHC, SMA wires and matrix, respectively. Sec-
ond term in right hand of Eq. (5) is the residual tensile stress in-
duced due to phase transformation in SMAHC specimens.
Eq. (5) has been used by different researchers to describe the
stressstrain relation of SMAHC plates. In this research the macro-
scopic behavior of SMAHC are investigated by using Eq. (5) under
two important phases of SMA wires (i.e., fully martensite and aus-
tenite). Due to the bilinear stressstrain behavior of SMA wires at
ambient and elevated temperatures, the overall behavior of
SMAHC would also take a bilinear form as schematically illustrated
in Fig. 2. In drawing this schematic curve it was assumed that: (a)
the stiffness of SMA wires and matrix are almost the same under
ambient temperature, (b) Isothermal condition is considered under
the ambient and elevated temperature, (c) strain to failure of ma-
trix is less than the failure strain of SMA wires.
In Fig. 2 curves 0AB and 0A0 B0 are representative stress
strain curves of SMAHC plates under ambient and elevated tem-
peratures, respectively.
4.1. SMAHC behavior under the ambient temperature
Under the ambient temperature, as illustrated by curve 0AB
in Fig. 2, the slope of the stressstrain curve could be constant
up to the detwinning stress of SMA wire (point A). After this point,
because of the martensite detwinning under almost constant loads,
Youngs modulus of the SMA wires reduce to their lowest value
(say, to 5% of their elastic portion), which results in reduction in
the stiffness of the SMAHC plate. After point A, the stress in both
matrix and SMAHC increases until failure occurs at point B.
Because of the strain compatibility, the strain in both the SMA
wires and matrix at points A and B would be equal and can be ob-
tained experimentally. Eq. (5) could be used to calculate the resul-
tant stress at points A and B.
2925
3
4
5
Fig. 2. Schematic stressstrain curve of the SMAHC plate and its components at
ambient and elevated temperatures.
It should be noted that under the ambient temperature, re-
moval of the mechanical load at any stage between points A and
B would leave some residual strain in the SMA wires and matrix.
This residual tensile strain would in return decrease the load
capacity of the SMAHC plate under repeated loadingunloading
applications. Based on the magnitude of the pre-strain, a suitable
heating and cooling procedure could be adapted in order to remove
some part of the induced residual strain from the system for the fu-
ture loading applications. In this case, any value of pre-strain with
a magnitude falling between points A and C shown in Fig. 2 would
not affect the global behavior or load capacity of the SMAHC plate,
but for pre-strain values close to the strain at point C, the stress
strain behavior would be closer to a linear curve, and the resulting
load capacity would be increased somewhat in comparison to the
lower pre-strain conditions. In such a case large pre-strain values
would remain on the structure during unloading.
4.2. SMAHC behavior under elevated temperatures
Under elevated temperatures, since SMA wires were elongated
before embedding them into the host material, a compressive
strain will be developed within the host material, which in return
will enhance the load capacity of the SMAHC plate when subjected
to a tensile load.
It is obvious that the pre-strain created by detwinning of SMA
wires, before they are embedded in the matrix, would not cause
any residual strain in the host material under the ambient temper-
ature; however, under elevated temperatures SMA wires activation
2926 F. Taheri-Behrooz et al. / Materials and Design 32 (2011) 29232933
would cause a residual compressive strain on the host materials.
The value of the induced residual strain is a function of the value
of pre-strain applied onto the SMA wires, geometrical and mechan-
ical property of the host materials and the applied temperature. By
using the magnitude of the recovery stress of individual SMA wires
which should be measured by experimental methods, one may ob-
tain the value of the residual strain on the host materials.
In order to calculate the magnitude of the compressive strain on
the host material, the SMAHC plate could be heated, while held in a
freefree boundary condition as shown in Fig. 1b. By considering
the equilibrium of forces across the cross section along the SMA
wires direction, one would obtain:
Fm Fs 0
 em  Em  Am es  Es  As 0
By substituting Eq. (1) into Eq. (6), one would get:
em  Em  1  V s rr  Es em  V s 0
and
rr  V s
em
Em  1  V s Es  V s
where rr and Vs are the recovery stress and volume fraction of the
SMA wires, respectively, em is the induced compressive strain in the
matrix which would enhance the load capacity of the system at ele-
vated temperatures.
It is worth mentioning that during the application of monotonic
loading under displacement control condition the abovementioned
strain would impose a tensile force on the specimen. The value of
this force could be estimated by using the following equation:
F E1-SMAHC  em  A
where E1-SMAHC and em are Youngs modulus and the total strain of
SMAHC, respectively.
Moreover, Eqs. (5) and (8) could be used to calculate the stres-
ses related to points A0 and B0 . It should be noted that under ele-
vated temperature the A0 B0 portion of stressstrain curve could
be very short, very long or even vanished based on the pre-strain
value and heating temperature of SMA wires.
5. Application of the rule of mixture (ROM)
In order to evaluate the mechanical properties of a SMAHC
based on its constitutes, the standard rule of mixtures microme-
chanics relation is used in this work [21]. It should be mentioned
that ROM is valid only if the stressstrain relationship of the con-
stituents is linear. In this research, as explained earlier, a bilinear
stressstrain relation is considered for the SMA wires in both the
fully martensite and austenite phases; therefore the ROM should
be applied separately for each phase of the material. On the other
hand the SMAHC plate would have a bilinear stressstrain behav-
ior. The ROM equation for the modulus is given as:
E1-SMAHC Es  V s Em  V m
The mechanical properties of the constituents used in Eq. (10)
should be selected from the data provided in Table 1 at both ambi-
ent and elevated temperatures.
Note that for more accurate evaluation of the modulus
(E1-SMAHC) by the ROM, the amount of resin lling among the wires
during SMAHC fabrication was also taken into account.
6. Experimental investigation
A systematic experimental investigation was performed in two
stages. In the rst stage the mechanical properties of the SMA
wires and prepreg composite plates were evaluated at the ambient
and elevated temperatures. In the second stage, an experiment was
designed and performed on SMAHC plates, also at ambient and ele-
vated temperatures, to establish: (a) the accuracy of the ROM for
predicting the longitudinal modulus of elasticity of the SMAHC;
(b) the integrity and accuracy of Eq. (5) in predicting the stress
strain relationship of the SMAHC; and (c) the recovery stress in-
duced by SMA wires on the host composites.
6.1. SMA wire testing
SMA may be found in different shapes, such as wire, springs,
ribbons and sheets; herein, SMA was chosen in wire form, because
6 SMA wires can produce a greater homogenized distribution of rein-
forcement in the composite, thereby producing more consistent re-
sponse of the SMAHC. In contrast, the sheet form has the benet of
providing higher SMA activation attribute in SMA reinforced sys-
7 tems, though it would require much higher activation energy
(either by current or heating).
The NITI SMA wires used was obtained by Memry Corp. (Bethel,
CT), composed of Nickel and Titanium (as are most of the commer-
8
cial NiTiNOL wires); the vendor recommends the material for high
endurance applications such as cyclic loading. The diameter of the
SMA wire used in the current work was 0.381 mm (0.01500 ). This
diameter was chosen in order to provide a relatively high SMA vol-
ume fraction to ensure the best performance possible. The austen-
ite start temperature (Tas) of the wire was 69 C and the austenite
nish temperature (Taf) was 84 C, according to the vendor data.
This is below the glass-transition temperature of the glass/epoxy
composites used in this research. However, mechanical constrain
would increase activation temperatures of the wires imbedded in
9 the host material, so a higher heating temperature such as 100 C
was selected intentionally to ensure full transformation into the
austenite phase. There are two different methods for activating
the SMA wires: (i) by heating or (ii) by the application of direct cur-
rent (DC). The latter actuation method was adapted in this re-
search. A DC current between 1.1 and 1.5 A was found adequate
to produce the desired temperatures between 64 and 105 C on
40 cm long SMA wires used in the test specimens.
In order to establish the mechanical behavior of the SMA wire,
isothermal tensile tests were conducted on individual wires at the
ambient (28 C) and elevated (100 C) temperatures according to
ASTM F2516 [22].
Testing at these temperatures simulates fully martensite and
austenite phases of the SMA, respectively. For comparison pur-
poses, the SMA wires tested at the ambient and elevated tempera-
tures were elongated by 0% and 4% strains, respectively. This was
done due to the fact that SMA thermo-mechanical properties are
dependent on variations in loading and temperature. The obtained
results are consistent with the test data available in literature [1].
Representative plots are shown for the cases of 0% pre-strain at
ambient and elevated temperatures in Fig. 3. During this test no
difference in the global stressstrain behavior of SMA wires was
10 found at the different pre-strain values tested at the ambient tem-
perature; only the large inelastic deformation capacity of SMA
wires was reduced when 4% pre-strain was applied at the ambient
temperatures. However, the stress induced martensite plateau that
was observed in the wires tested at the ambient temperatures
could not be observed in the stressstrain curve of the wires tested
at elevated temperatures of 100 C (see Fig. 3). Moreover the initial
martensite shape was not restored after removing the thermal and
mechanical loads. This phenomenon indicates that the induced
martensite stress at elevated temperature for this specic SMA
wire is larger than the yield stress of the wire; as a result the
SMA wire yields due to the high stress level before entering to its
martensite phase.
 F. Taheri-Behrooz et al. / Materials and Design 32 (2011) 29232933 2927

Fig. 3. Stressstrain response of NiTiNOL wires as received at ambient (28 C) and elevated (100 C) temperature (without pre-strain).

Youngs moduli were calculated from the initial linear portion

of the stressstrain curves. The summary of the isothermal test re-
sults for SMA wires are shown in Table 1. The data shown in the
table are averaged over 10 specimens tested under each condition.
A laser extensometer was used to measure strain during tensile
tests of SMA wires. Length of the SMA wires was approximately
400 mm and special gripping xtures, as shown in Fig. 4, were used
to perform the isothermal test on SMA wires.
The evaluation of the recovery stress in SMA wires is important
for the effective design of a SMAHC system. Tests were conducted
under strain control mode, in order to measure the recovery stress
under 4% pre-strain. The specimens were loaded up to 4% strain
and then test was stopped and the wires were activated (heated)
by DC current to 100 C. The heating and cooling process was re-
peated to assure the accuracy and consistency of measured recov-
ery stress for each wire. The test was continued to SMA wires
failure point. A representative curve showing the wires stress var-
iation as a function of time is shown in Fig. 5.

6.2. Laminated composite testing

The composite material used in this research (also referred to as

the matrix in this manuscript), was 7781 woven E-glass/NB321
epoxy prepreg produced by Newport Adhesives and Composites
Inc. (Irvine, CA). In order to obtain its mechanical properties, com-
posite laminate plates, with the dimensions of (300 mm  300mm)
were fabricated (hand-laid), using eight layers of woven prepregs,
by vacuum bagging technique and cured under heat, in an oven un-
der full vacuum per the instruction provided by the vendor. The
Fig. 4. Tensile test setup of SMA wires at activated temperatures (i.e., resistive
layup was subjected to approximately 2500 Hg vacuum. The vacu- heating).
umed laminate was heated to 135 C at an applied heat rate of
2 C/min and was held at that temperature for 120 min, and then
cooled down to room temperature. The plates were cut into appro- 6.3. SMAHC specimen preparation
priate specimen size according to ASTM D 3039 [23] (i.e.,
250 mm  22 mm  1.95 mm (L  W  T)). The specimens were The SMAHC specimen fabrication procedure used in this study
subsequently tested under a tensile load applied by an Instron consists of the following steps; pre-straining the SMA wires, sand-
8500+ universal testing machine, equipped with an extensometer wiching the SMA wires in between composite laminate, curing the
with 25 mm gage length, at the ambient (28 C) and elevated hybrid-laminate while the wires are under the pre-strain, cooling
(100 C) temperatures. A total of ve specimens were tested at down the assembly and cutting the wires.
each temperature, with the results summarized in Table 1. Fiber The SMA wires were pre-strained using an in-house designed ri-
volume fraction of the specimens was evaluated by matrix burn- gid steel frame with movable rods as shown in Fig. 6. On the
off procedure according to ASTM D 3171 [24] and was determined adjustable and xed rods, threads with appropriate pitch size were
to be between 46% and 50%. Typical intact and failed specimens are machined to produce the maximum SMA volume fraction after
shown in Figs. 7 and 13, respectively. winding the SMA wires around them. Moreover, due to constant
2928 F. Taheri-Behrooz et al. / Materials and Design 32 (2011) 29232933

Fig. 5. Recovery stress measuring in NiTiNOL wires at 4% pre-strain and elevated temperature (100 C).

Fig. 6. The rigid adjustable frame and SMAHC plate after curing.

pitch size in the rods, an equal distance between any two wires
could be attained. This ensured that the wires would be homoge-
nously distributed within SMAHC plates. The movable horizontal
rods could be adjusted by the two nuts mounted on each of the
side rods to create the desired pre-strain. In order to adjust the va-
lue of the wire pre-strain precisely, a digital deectometer was
used to measure the deection in the side rods.
After winding the SMA wires around the horizontal rods, the
wires were stretched by the adjustable frame to a certain level.
Then the tension on the wires was released, and the wires were
sandwiched between glass/epoxy composite prepregs to produce
the SMAHC plates. At this stage, the entire assembly was vacuum
bagged to approximately 2500 Hg vacuum. The assembly was sub-
jected to the oven cure cycle recommended by the vendor (heating
rate of 2 C/min, ramped to 135 C, held for 120 min at that
temperature and cooled down to room temperature) and the
resulting consolidated plate produced within the manufacturing
frame is shown in Fig. 6.
The consolidated SMAHC plates were machined into appropri-
ate size test specimens (i.e., 220 mm  19 mm  2.4 mm
(L  W  T)) by using water jet machine to eliminate any possible
edge delamination which might be produced by diamond saw cut-
ting. A two part heat resistant adhesive from Huntsman, Inc. (TX,
USA) with the commercial name of Araldite 2014 was used for tab-
bing the tensile specimens according to ASTM D 3039, as shown in
Figs. 7 and 9. A total of ve specimens were tested for each temper-
ature under monotonic tensile loading.
By using four layers of prepreg laminate composite on each side
of the SMA wires, SMAHC plates with volume fractions of 67% of
SMA wires were obtained. It is worth noting that the thickness of
 F. Taheri-Behrooz et al. / Materials and Design 32 (2011) 29232933
Fig. 7. A consolidated SMAHC plate and typical specimens machined from plates with and without SMA wires.
the composite layer could potentially have a signicant adverse
inuence on the SMA reinforcement due to two different parame-
ters. Firstly, the larger the plate thickness, the lower the volume
fraction of the SMA wires, and thus the lower the stiffening effect
of SMA within the SMAHC plate. Secondly, the use of thicker com-
posite plates would lower the residual compressive stress induced
by the SMA wires; this would in turn result in higher residual ten-
sile stress and strain in SMA wires, thus lowering the strain to fail-
ure limit of the SMA wires to less than that of the hosting
composite plate, and nally reducing the load capacity of the
SMAHC.
SMAHC plates were also produced with 4% pre-strain to study
their macroscopic mechanical behavior under tensile loading at
both the ambient and elevated temperatures.
The glass-transition temperature for the prepregs was 149 C
according to the vendor data, which was higher than the elevated
temperature adopted in the tests (i.e., 90100 C). All coupons
were tested three minutes after their temperature had reached
90 C to ensure thermal steadiness of the coupons.
6.4. SMAHC specimen testing and discussion
In order to evaluate the reinforcing effect of the SMA wires on
composite materials, three sets of tests were performed on the fab-
ricated SMAHC specimens. These were (i) scanning electron micro-
scope analysis; (ii) monotonic tensile testing; and (iii) induced
strain measurement.
6.4.1. Scanning electron microscopy analysis
Scanning electron microscopy (SEM) was conducted on the
SMAHC specimens in order to evaluate: (a) quality of the bond be-
tween the SMA wires and host material; and (b) whether the SMA
wires were uniformly distributed within the host composite.
To facilitate good adhesion between the wires and the resin of
the host material, the wires surface was thoroughly cleaned by
Acetone (three times) before embedding them in the host material.
The good bond quality between the wires and host matrix of the
intact specimens can be observed through the SEM images illus-
trated in Figs. 8 and 9. The good adhesion is evident by the pres-
ence of large resin clusters around the wires (see Fig. 8a). Fig. 8b
further illustrates the perfect bond and lack of any voids in the
interface of the SMA wire and matrix. Fig. 9 is a representative
micrograph, showing a typical longitudinal cross section of one
of the SMAHC specimens. As seen in the magnied image of
Fig. 9b, there is no evidence of any gap or crack within the longitu-
dinal interface region. In contrast to the intact specimens, the SEM
micrograph (shown in Fig. 10), prepared from the surface of a failed
specimen under monotonic tensile loading, reveals the presence of
2929
a micro-crack between the SMA wires and host matrix. Fig. 11
shows the uniform spacing of the SMA wires within the matrix.
In addition, the SEM image illustrated in Fig. 12 shows a rela-
tively lower magnication of the transverse cross-section of the
SMAHC; the uniform distribution of the SMA wires within the host
matrix is evident. In the image, the darker regions surrounding the
wires indicate the presence of pure resin deposited between the
SMA wires during manufacturing of the SMAHC plates. The volume
fraction of the SMA wires in the SMAHC specimen is evaluated to
be between 6% and 7%.
Obtained results from the SEM observations exhibited uniform
distribution of SMA wires inside the host material and veried
existence of appropriate bonding qualitatively in SMAcomposite
interface.
6.4.2. Monotonic tensile testing of the SMAHC specimens
As stated earlier, the SMAHC specimens were tested at two dif-
ferent temperatures (i.e. ambient (28 C) and elevated (90100 C)
temperatures), enabling us to evaluate the inuence of SMA rein-
forcement on the host composites response under different tem-
peratures. As stated earlier, the mechanical tests were carried out
within an Instron servo-hydraulic universal testing machine, con-
trolled by the Instron 8500+ electronics. A DC power supply was
used to activate (heat) the SAM wires of the SMAHC specimens.
The loading rate for the monotonic testing was 1 mm/min. The
strain was measured using an Instron extensometer with 25 mm
gage length. Fig. 12 shows the tensile response of the SMA rein-
forced and un-reinforced specimens at the two different tempera-
tures. As seen, the SMA reinforcement improved both the stiffness
and the ultimate strength of the SMAHC coupons tested at the ele-
vated temperature compared to those at the ambient temperature.
The comparison of the experimental results (i.e. at ambient and
elevated temperatures) and the theoretical results obtained by
using Eqs. (5), (8), and (10) are also illustrated in Fig. 12.
The results illustrated in Fig. 12 indicates that under the ambi-
ent temperature, once the strain of SMA wires approached the det-
winning strain, the stiffness of SMAHC plates decreased. Matrix
failure (here, referring to the host laminated composite) did not re-
sult in the failure of SMA wires; SMA wires just exhibited larger
deformation due to martensite detwinning.
Under the elevated temperature, the initial stiffness of the
SMAHC plates increased marginally because of the high austenite
modulus of the wires. However, after the strain reached to the
maximum elastic strain of the wires at the elevated temperature,
the stiffness of the SMAHC plates reduced. At the elevated temper-
ature, failure of the SMA wires and composite sheets occurred al-
most simultaneously, indicating that the temperature of wires
during testing exceeded the austenite nished temperature (e.g.
2930 F. Taheri-Behrooz et al. / Materials and Design 32 (2011) 29232933

Fig. 8. Representative micrographs of an intact specimen with different magnications.

Fig. 9. SEM micrographs showing a plane cut along the axial direction.

Fig. 10. SEM image of a specimen failed under tensile loading.

84100 C as indicated by the vendor), as a result, the wires fol-

lowed the regime shown in Fig. 3. The stiffness and strength in this Fig. 11. Uniform distribution of the SMA wires in the transverse cross section of the
SMAHC specimen.
case was observed to be higher than those observed at the ambient
temperature, indicating the positive attribute resulting from the
activation of SMA wires at the elevated temperature. Fig. 13 shows and SMAHC specimens under the ambient and elevated
a typical failed laminate composite (i.e. un-reinforced with SMA) temperatures.
 F. Taheri-Behrooz et al. / Materials and Design 32 (2011) 29232933
Fig. 12. Stressstrain behavior of SMAHC plates under ambient and elevated temperatures.
Fig. 13. Typical failed non-SMA reinforced composite specimen, and SMAHC specimens tested under ambient and elevated temperatures.
The data reported in Table 1 and Fig. 3, along with Eqs. (5), (8),
and (10) were used to predict the stressstrain curve of SMAHC
specimens. Under the ambient temperature, the differences be-
tween the experimental and theoretical stressstrain curves
shown in Fig. 12 could be due to the higher elastic modulus pre-
dicted by the ROM. A better correlation of the results is observed
for specimens tested under the higher temperature. For the sake
of comparison, the initial stress induced in the specimens by
SMA wires is removed from the theoretical and the experimental
results obtained at the elevated temperatures. In other words, at
the elevated temperatures, the theoretical and experimental
curves should be starting from 26 MPa to 13.6 MPa, respectively,
instead of zero (i.e. the magnitudes of the initial stress induced
by the SMA wires onto SMAHC plates).
The monotonic tensile test results also generated another inter-
esting outcome; it could be veried that the simple ROM could pre-
dict the material properties of the SMA reinforced composite
plates, using the properties of the constituents, with a reasonable
accuracy. The experimentally obtained material properties of the
SMAHC plates, as well as those of its constituents, are tabulated
in Tables 1 and 2, respectively. The reported modulus of elasticity
of the SMAHC and composite plates are the average of ve tests in
each case. In calculating the theoretical modulus of the SMAHC, as
stated, the contribution of the extra resin surrounding the wires
was accounted for by assuming a modulus value of 2 GPa for that
resin volume. In addition, the stressstrain behavior of the host
composite (without SMA wires) was nonlinear; however, in this re-
search it was assumed as a bilinear curve, with the modulus of
2931
elasticity of the rst part being 24 GPa and that of the second linear
part being 20 GPa. The summary of the data are reported in Table 2.
6.4.3. Induced strain measurement
Pre-strained SMA wires, once activated, should recover their
initial detwinning martensite structure. If any external constrains
are involved (in our SMA-reinforced specimens case, the SMA are
restrained by the host composite laminate material), then they
can only recover only a portion of their initial strain. The recovered
part of their pre-strain would therefore impose a compressive
strain on the host material; while the unrecovered part would re-
main as a pre-strain in the activated SMA wires.
A set of unconstrained SMAHC specimens were tested at the
elevated temperature in order to measure the experimentally in-
duced strain on the SMAHC plates. During this test a laser exten-
someter was used to monitor the induced strain. The mean value
of the recorded strain was 0.052% during the resistive thermal acti-
vation. The induced strain calculated by Eq. (8) was determined to
be 0.1%. The discrepancy is believed to be due to the thermal
expansion of the SMAHC specimens, a phenomenon not considered
by Eq. (8). The level of the induced strain in the host material is an
indicator of the quality of the bond between the SMA wires and
surrounding matrix.
Finally, as it has been shown in Fig. 12 and Table 2, the stiffness
and load capacity of the SMAHC could be enhanced by appropriate
activation of the SMA wires at elevated temperatures. By increas-
ing the volume fraction of the SMA wires to more than 7%, addi-
tional enhancement could be expected. In addition, based on our
2932 F. Taheri-Behrooz et al. / Materials and Design 32 (2011) 29232933

Table 1
Material properties of the glassepoxy composite and the SMA wires evaluated at ambient (28 C) and elevated temperatures (100 C) under monotonic tensile testing.

SMA Ambient Es 28 GPa 1; ms 0:33; rmd 200 MPa 5; emd 0:7%; emf
s 5:57%; X s 1450 MPa 30
s s
(martensite)
y y
Elevated Es 80 GPa 5; rms  
s 600800 MPa; X s 1100 MPa 20, es 1:1%; T as 69 C; T af 84 C, recovery
(austenite) stress at 4% pre-strain (rr) = 380 MPa (10)
Composite Ambient Em 24 GPa 0:8; mm 0:03; X m 540 MPa 8; eut
m 2:5%; V f 4650%
(pseudo-matrix)
Elevated For this woven glass/epoxy materials the same mechanical properties are obtained at elevated and the ambient temperatures

Letters, m, d, s, X,m, t, c, u, y, Vf, ms, md, mf, refer to matrix, detwinning, SMA wire, tensile strength, Poissons ratio, ultimate, yield, ber volume fraction, martensite (start,
detwinning and nished) respectively. Provided data in bracket refer to the standard deviation of the mean values.

Table 2
SMAHC plates mechanical properties at the ambient (28 C) and elevated temperatures (90100 C).

Property Test condition Theorya Experiment (mean) Standard deviation

E1-SMAHC (GPa) Ambient 22.304 20.03 1.4
Elevated 25.92 26.3 1.29
Ultimate stress (MPa) Ambient 413.7 379 10
Elevated 455.54 452 32
Induced strain (mm/mm) Elevated 0.001 0.00052
a
These data were calculated by using Eqs. (5), (8), and (10).

experimental results, the ROM (Eq. (10)), along with the constitu-
tive equations derived based on the one-dimensional model (Eqs.
(5) and (8)), could be used effectively, as a rule of thumb, in
designing SMAHCs. One may employ these equations for establish-
ing the appropriate pre-strain value, activation temperature, in-
duced strain and the proportioning of the constituent materials
for a given application.
7. Conclusions
As seen from the results presented earlier, the inclusion of the
SMA wires increased the stiffness and strength of the reinforced
composite plates at elevated temperatures. The enhancement
was found to be a function of SMAs volume fraction, the magni-
tude of the pre-strain and activation temperature of the SMA wires
incorporated within the composite.
In summary, the following conclusions are made:
(1) Considering the observations made in this research, the use
of SMAHC could be considered viable in applications where
elevated temperatures could signicantly degrade the stiff-
ness of the host conventional polymeric composites, thus
signicantly inuencing their intended performance.
(2) The results showed that one can safely use the ROM for
estimating other mechanical properties by knowing the
characteristics of the constituents. It was also observed that
the stress and strain predicted by the use constitutive
equation derived based on the one-dimensional model
could produce results with acceptable accuracy in compar-
ison to the experimental data. Therefore, this equation,
along with the ROM, could be effectively used as a rule
of thumb for predicting the macroscopic behavior of
SMAHC systems.
(3) SMA wires can be incorporated as reinforcement in host
materials like composites and polymers that have relatively
lower stiffness/strength than SMA (those measured at the
austenite phase). Such reinforcement could signicantly
compensate the host materials degradation due to exposure
to elevated temperatures. However, their application in
structures that have inherently comparable stiffness/
strength would not lead to appreciable enhancement.
(4) The shape memory effect (martensiteaustenite transforma-
tion) is believed to be the main reason for mitigating stiff-
ness and strength degradation in SMAHC plates at elevated
temperature.
(5) The level of pre-strain of SMA played as important role in
enhancing the strength of the SMAHC; in fact, in the austen-
ite phase, higher pre-strains produced higher induced com-
pressive stress in the matrix, which in return produced
higher strength for the SMAHC. However, it should be noted
that the higher pre-strain may also induce an excessively
large magnitude of initial tensile stress on SMA wires, thus
adversely inuencing the overall strength and ultimate
strain capacity of the SMAHC plates.
(6) A residual tensile strain remained in the SMA wires because
of the constraints provided by the host material (e.g. during
manufacturing), which could in turn cause an increase in the
activation temperature of SMA wires. Therefore, a readjust-
ment of the vendor suggested activation temperature is
recommended.
(7) It is concluded that selecting a suitable pre-strain activation
temperature is very important; this temperature would dif-
fer from one application to another.
Acknowledgement
The nancial support of the National Sciences and Engineering
Research Council of Canada (NSERC) in support of this work is
gratefully acknowledged.
References
[1] Cross WB, Kariotis AH, Stimler FJ. NiTiNOL characterization study. Goodyear
Aerospace Corporation (OH, USA) for Langley Research Center (NASA-CR-
1433); 1969.
[2] Perkins J. Shape memory effects in alloys. NY, London: Plenum Press; 1976.
[3] Funakubo H. Shape memory alloys. Gordon and Breach Science Publishers;
1987 [translated by Kennedy JB].
[4] Jackson CM, Wagner HJ, Wasilewski RJ. 55-Nitinol-the alloy with a memory: its
physical metallurgy, properties, and applications, NASA-SP-5110; 1972. p. 91.
[5] Srinivasan AV, McFarland MD. Smart structures, analysis and
design. Cambridge University Press; 2001.
[6] Culshaw B. Smart structures and materials. Artech House; 1996.
 F. Taheri-Behrooz et al. / Materials and Design 32 (2011) 29232933
[7] Liang C, Rogers CA. One-dimensional thermomechanical constitutive relations
for shape memory. Mater J Intell Mater Syst Struct 1990;1:20734.
[8] Jia J, Rogers CA. Formulation of a mechanical model for composites with
embedded SMA actuators. J Mech Des 1992;114(4):6706.
[9] Jarali CS, Raja S, Upadhya AR. Micromechanical behaviour of SMA composite
materials under hygro-thermo-elastic strain elds. Int J Solids Struct
2008;45:2399419.
[10] Mara S, Sacco E. Micromechanics and homogenization of SMA-wire-
reinforced materials. J Appl Mech 2005;72(2):25968.
[11] Tsai XY, Chen LW. Dynamic stability of a shape memory alloy wire reinforced
composite beam. J Compos Struct 2002;56:23541.
[12] Gao X, Burton D, Turner TL, Brinson LC. Finite element analysis of adaptive-
stiffening and shape-control SMA hybrid composites. J Eng Mater Technol
2006;128(3):28593.
[13] Turner TL, Lach CL, Cano RJ. Fabrication and characterization of SMA hybrid
composites. In: Lynch CS, editors. Proceeding of SPIE 8th annual symposium on
smart structures and materials, vol. 4333; 2001. p. 34354.
[14] Turner TL. SMA hybrid composites for dynamic response abatement
applications. J Vib Acoust 2005;127(June):2739.
[15] Turner TL. A new thermoelastic model for analysis of shape memory alloy
hybrid composites. J Intell Mater Syst Struct 2000;11(May):38294.
2933
[16] Davis B, Turner TL, Seelecke S. Measurement and prediction of the
thermomechanical response of shape memory alloy hybrid composite
beams. J Intell Mater Syst Struct 2008;19(February):12943.
[17] Han H, Ang KK, Wang Q, Taheri F. Buckling enhancement of epoxy columns
using embedded shape memory alloy spring actuators. J Compos Struct
2006;72:20011.
[18] Han H, Taheri F, Pegg N, Zhang Z. Local buckling mitigation and stress analysis
of a shape memory alloy hybrid composite plate with and without a cutout. J
Smart Mater Struct 2007;16(3):589604.
[19] Lagoudas DC. Shape memory alloys, modeling and engineering
applications. Springer; 2008.
[20] Bollas D, Pappas P, Parthenios J, Galiotis C. Stress generation by shape memory
alloy wires embedded in polymer composites. Acta Mater 2007;55:548999.
[21] Hyer MW, Scott RW. Stress analysis of ber reinforced composites
materials. WCB: McGraw-Hill; 1997.
[22] ASTM-F2516. Standard test method for tension testing of nickeltitanium
superelastic materials; 2007.
[23] ASTM-D3039. Standard test method for tensile properties of polymer matrix
composite materials; 2008.
[24] ASTM-D3171. Standard test methods for constituent content of composite
materials; 2009.