Edoc Design of Shape Memory Alloy Sma Actuators

SPRINGER BRIE FS IN APPLIED SC IENCES AND
TECHNOLOGY  COMPUT ATIONA L MECHA NICS
Ashwin Rao
A.R. Srinivasa
J.N. Reddy
Design of Shape
Memo
(SMA) rAycAllo
tuatyors
13
SpringerBriefs in Applied Sciences
and Technology
Computational Mechanics
Series editors
Holm Altenbach, Magdeburg, Germany
Lucas F.M. da Silva, Porto, Portugal
Andreas Öchsner, Southport, Australia
More information about this series at http://www.springer.com/series/8886
Ashwin Rao A.R. Srinivasa
•
J.N. Reddy
Design of Shape Memory

Alloy (SMA) Actuators
13
Ashwin Rao J.N. Reddy
Department of Mechanical Engineering Department of Mechanical Engineering
Texas A&M University Texas A&M University
College Station, TX College Station, TX
USA USA
A.R. Srinivasa
Department of Mechanical Engineering
Texas A&M University
College Station, TX
USA
ISSN 2191-530X ISSN 2191-5318 (electronic)

SpringerBriefs in Applied Sciences and Technology
ISSN 2191-5342 ISSN 2191-5350 (electronic)
SpringerBriefs in Computational Mechanics
ISBN 978-3-319-03187-3 ISBN 978-3-319-03188-0 (eBook)
DOI 10.1007/978-3-319-03188-0
Library of Congress Control Number: 2015938741
Springer Cham Heidelberg New York Dordrecht London

© The Author(s) 2015
This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part
of the material is concerned, speci cally the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on micro lms or in any other physical way, and trans mission
or informationnow
methodology storage andorretrieval,
known hereafterelectronic
developed.adaptation, computer software, or by similar or dissimilar
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a speci c statement, that such names are exempt from
the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this
book are belie ved to be true and accur ate at the date of publication. Neith er the publisher nor the
authors or the editors give a warranty, express or implied, with respect to the material contained herein or
for any errors or omissions that may have been made.
Printed on acid-free paper
Springer International Publishing AG Switzerland is part of Springer Science+Business Media

(www.springer.com)
This book is dedicated to my parents
Raghavendra Rao and Jayashree R. Rao, who
have raised me, motivated me, and have
guided me through all ups and downs of my
life.
Ashwin Rao
This book is dedicated to my parents

Ramaswamy and Saroja and to my wife
Prabha and children Vishnu and Divya, who
sacri ced their time so that I could nish this
book.
Arun Srinivasa
This book is dedicated to my baava, Tummala

Janardhan Reddy, a constant admirer of the
highest evidences of the human spirit, for his
friendship and for introducing me to
“ rational thinking.
”
J.N. Reddy
Preface
Shape memory alloys have been fascinating to designers, architects, and researchers
in the past decade. There is something about the uncanny ability of seemingly
inanimate wires suddenly reacting to external nonmechanical stimulus that evokes
curiosity and childlike fascination in everyone. However, commercial applications
(other than in the medical eld) has been slow. Part of the reason is the lack of
accessible explanations that allow people with only basic exposure to such mate-
rials to carry out designs that are viable.
Tooaspects
arcane often, papers
of SMAand books written
behavior are not (many
meant by
for the authorsThis
designers. themselves) about
leads to exas-
peration from a designer who wants us to “tell me how do I design with this? ”
This book seeks to provide an accessible account of SMA behavior together with
examples of preliminary design methodology to students with a basic undergrad-
uate background. The aim is to provide an “on ramp ” to explore the unique
properties of these devices, and so the book only deals with the “bare necessities”
and ignores many nuances including important issues of functional fatigue. Rather
the design recommendations are based on being conservative and making design
decisions that will eliminate the need for considering such issues at the expense of
not being optimal. Our philosophy in designing with SMA is “robust, repeatable,
guaranteed behavior” over “optimal” response.

This monograph is by no means extensive but just an introduction and an
invitation to the readers to explore the behavior of these materials. It grew out of a
National Science Foundation Grant to develop “strength of materials-like ”
approaches to shape memory wires and springs. While the current applications at
the cutting edge have moved on to tubes, plates, and so on, the commercial
availability of wires and springs for common devices has grown quite a bit and this
is what we wish to emphasize.
vii
viii Preface
If a reader gains a qualitative understanding of SMA response together with the

ability to do a “ rst-cut” design by reading this book and is able to go on to explore
SMA better, then we have succeeded in achieving our purpose of writing this book.
CollegeStation,Texas AshwinRao
A.R. Srinivasa
J.N. Reddy
Acknowledgments
The authors acknowledge the help of a large group of people in creating this book.
In particular, the authors acknowledge the invaluable help, advice, discussions, and
guidance of Dr. S. Chandrasekaran (SMA specialist and visiting Scientist, IIT
Madras), Prof. Sivakumar Srinivasan (Department of Mechanical Engineering, IIT
Madras), and Dr. Vidya Shankar Buravalla (GE Global Research) in the writing of
this book. The authors also gratefully acknowledge the support of the research by
the National Science Foundation through research Grant No. CMMI-1000790 (with
A.R.
them Srinivasa andthe
to undertake J.N. Reddyofasthis
writing co-PIs)
book.to Texas A&M University, which inspired
Ashwin Rao thanks his colleagues at Evident Thermoelectrics for their support
and encouragement during the completion of this book.
ix
Contents
1 Introduction to Shape Memory Alloys . . . . . . . . . . . . . . . . . . . . . 1

1.1 Smart Materials –An Overview . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Smart Structures —System Level Response . . . . . . . . . . . . . . . 2
1.3 Shape Memory Alloys: Temperature Induced Phase
Transformations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.4 Shape Memory Effect and Superelasticity/Pseudoelasticity . . . . 8
1.5 Commonly Used Shape Memory Alloys . . . . . . . . . . . . . . . . 12
1.6 SMA
1.6.1 Applications:
Biomedical Overview Applications . . . . .. .. .. .. .. . ................... . .. .. . . . .. . 15
15
1.6.2 Civil Engineering Applications . . . . . . . . . . . . . . . . . 19
1.6.3 Aerospace and Automotive Applications . . . . . . . . . . 23
1.6.4 Miscellaneous Applications . . . . . . . . . . . . . . . . . . . 25
1.7 Chapter Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
2 Need and Functionality Analysis . . . . . . . . . . . . . . . . . . . . . . . . . 33

2.1 The System Design Process . . . . . . . . . . . . . . . . . . . . . . . . . 35
2.1.1 Design Methodology: Structure and Guidelines . . . . . . 36
2.2 The Five Major Subsystems . . . . . . . . . . . . . . . . . . . . . . . . . 38
2.3 How Do We Identify Need and Functionality for SMAs . . . . . 40
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
3 Manufacturing and Post Treatment o f SM A Components . . . . . . . 43

3.1 Different Manufacturing Techniques . . . . . . . . . . . . . . . . . . . 43
3.1.1 Vacuum Induction Melting (VIM) Technique . . . . . . . 44
3.1.2 Vacuum Arc Remelting (VAR) Technique . . . . . . . . . 45
3.1.3 Electronic Beam Melting (EBM) Technique . . . . . . . . 46
3.1.4 Conventional/Normal Sintering Technique . . . . . . . . . 46
3.1.5 Selective Laser Sintering (SLS) . . . . . . . . . . . . . . . . 47
3.1.6 Hot Isotactic Pressing (HIP) . . . . . . . . . . . . . . . . . . . 47
xi
xii Contents
3.1.7 Spark Plasma Sintering (SPS) . . . . . . . . . . . . . . . . . . 47

3.1.8 Selective Laser Melting (SLM) . . . . . . . . . . . . . . . . . 48
3.1.9 Metal Injection Molding (MLM) . . . . . . . . . . . . . . . . 49
3.2 Post Treatment of SMAs . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
3.2.1 Machining of SMA Components . . . . . . . . . . . . . . . . 50
3.2.2 Surface Treatment of SMA Components . . . . . . . . . . 50
3.2.3 Annealing and Coldworking of SMA . . . . . . . . . . . . 52
3.2.4 Joining of SMA to Itself and Other Materials
Like Stainless Steel . . . . . . . . . . . . . . . . . . . . . . . . . 53
3.2.5 Shape Setting of Nitinol . . . . . . . . . . . . . . . . . . . .. 55
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
4 Basic SMA Component Geometries and Responses . . . . . . . . . . . . 61

4.1 SMA Wire Response —Tensile Loading. . . . . . . . . . . . . . . . . 61
4.2 SMA Wire Response —Torsional Loading . . . . . . . . . . . . . . . 65
4.3 SMA Spring Response —Torsional Loading . . . . . . . . . . . . . . 67
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
5 Factors In fluencing Design of SMA Actuators . . . . . . . . . . . . . . . 73

5.1 Geometry Factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
5.2 Effect of Alloy Composition . . . . . . . . . . . . . . . . . . . . . . . . 75
5.3 Effect of Shape Setting Conditions for Custom Shape

(Making SMA Springs) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
5.4 Effect of Operating Temperature on Mechanical Response . . . . 76
5.5 Effect of Loading Rates . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
5.6 Wire Training/Hysteresis Stabilization . . . . . . . . . . . . . . . . . . 78
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
6 Graphical Description of Temperature Controlled Actuation

of SMA Wires . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
6.1 SMA Wire + Bias Spring Arrangement . . . . . . . . . . . . . . . . . 81
6.1.1 SMA Wire Selection . . . . . . . . . . . . . . . . . . . . . . . . 81
6.1.2 Operating Temperature of SMA Wire . . . . . . . . . . . . 83
6.2 Graphical Design Approach for Stroke Estimation . . . . . . . . . . 83
6.2.1 Graphical Design Approach for Stroke
Estimation—Load and Displacement
Controlled Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
6.2.2 SMA Wire + Bias Spring: Graphical Design
Approach for Stroke Estimation Using Linearized
Loading Response Only . . . . . . . . . . . . . . . . . . . . . . 87
6.3 Case Study 2: Linear to Rotary Arrangement Using a SMA
Wire + Bias Spring Arrangement Using Linearized
Loading Response Only . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
Contents xiii
6.4 Case Study 3: SMA Wire + Bias Spring Arrangement

Using Linearized Loading —Unloading Response . . . . . . . . . . 93
6.5 Case Study 4: SMA Wire + Bias Spring Arrangement
Using Complete Hysteretic Loading —Unloading Response . . . 96
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
7 Case St udies in th e Preliminary De sign of SMA A ctuators . . . . . . 99

7.1 Different Modes of Operation . . . . . . . . . . . . . . . . . . . . . . . . 101
7.1.1 Constant Force Mode . . . . . . . . . . . . . . . . . . . . . .. 101
7.1.2 Constant De flection Mode . . . . . . . . . . . . . . . . . . . . 102
7.1.3 Simultaneous Force-De flection Mode . . . . . . . . . . . . 102
7.2 Design of SMA Wires Under Constant Force . . . . . . . . . . . . . 103
7.3 Case Study II: Design Procedure for Ti –Ni (SMA) Springs . . . 105
7.4 Spring Design Case Study . . . . . . . . . . . . . . . . . . . . . . . . . . 105
7.4.1 Design Model and Assumptions . . . . . . . . . . . . . . . . 105
7.4.2 Terms Used in Design of SMA Springs . . . . . . . . . . . 107
7.5 Example: Design of a R emote Controlled Flow Control
Valve Using an SMA Compression Spring . . . . . . . . . . . . . . . 108
7.5.1 Statement of Requirement . . . . . . . . . . . . . . . . . . . . 108
7.6 Extensional Spring Design . . . . . . . . . . . . . . . . . . . . . . . . . . 112
7.7 Heating and Cooling of Shape Memory Wires . . . . . . . . . . . . 113
7.7.1 Time Taken to Heat Up and Cool Down . . . . . . . . . . 114

References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
8 Coupling SMA Actuators wi th M echanisms: Principle

of Virtual Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
8.1 The Need for Mechanisms . . . . . . . . . . . . . . . . . . . . . . . . . . 117
8.2 The Loading Curve and the SMA Response . . . . . . . . . . . . . . 120
8.3 3-D Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
8.4 Bias Forces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
Reference . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
9 Fatigue of SMAs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125

9.1 Structural and Functional Fatigue in SMAs . . . . . . . . . . . . . . 125
9.2 Reporting Fatigue Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
Chapter 1
Introduction to Shape Memory Alloys
1.1 Smart Materials–An Overview
Classical materials like metals and alloys have played a significant role as
structural materials for many centuries [1]. Engineers have designed components
and selected alloys by employing the classical engineering approach of understand-
ing the macroscopic properties of the material and selecting the appropriate one to
match the desired functionality based on the application [2]. With advancements in
material science and with increasing space and logistical limitations, scientists have
been constantly developing high performing materials for various applications [2].
The everlasting goal for engineers in many cases is to improve product efficiency and
reduce its weight without comprising on either its cost or performance. To achieve
this goal, replacing multi-component and multi-material systems with fewer multi-
functional light weight, high performing materials has been an attractive alternative
[2, 3]. Such advanced materials have played a leading role in the development of
many engineering innovations and achievements like the Airbus A380, Boeing 787
Dreamliner, reliable fuel efficient cars, superior drug delivery devices, and so on, to
list a few. The ingenious commercial products across various engineering disciplines
are meeting all requirements by encompassing many of the latest technologies and
meeting the challenges of tomorrow’s needs.
In pursuit of this, material scientists over the last few decades have focused on
the possibility of tailoring the microstructure of the material to generate the required
functionality for different applications [2, 4]. Such an effort has resulted in an entire
new area of active or multifunctional materials that posses more than one desirable
property [5, 6]. With the introduction of such materials, researchers are now focusing
on how the combined microstructural changes of such materials are able to perform
multiple functions. The integration of multiple functions like actuation, sensing, and
control into a single structure using one or more material constituent is seen as a
possibility [2, 3]. Mamoda in her recent review of future materials discusses some
application ideas with such materials like: a smart solar panel that can change its
orientation automatically during the day depending on sun’s position; a smart shock
© The Author(s) 2015 1

A. Rao et al., Design of Shape Memory Alloy (SMA) Actuators ,
SpringerBriefs in Computational Mechanics, DOI 10.1007/978-3-319-03188-0_1
2 1 IntroductiontoShapeMemoryAlloys
Table 1.1 Multiphysics coupling

Input energy Output energy
Mechanical Magnetic Thermal Electrical
Mechanical Magnetostriction Piezoelectric,
piezoresistive
Magnetic Magnetostriction Magnetocaloric Magnetoresistance
Thermal thermomechanical Pyroelectric T hermoelectric
Electrical Electrostrictive, Thermoelectric
piezoelectric, ER
fluid
absorber that can alter its damping ability based on the road profile; morphable wings
and blades for aircraft maneuvering during flight; a coating that changes color on
demand, to list a few examples [3, 6].
Smart materials are a subgroup of such active/multifunctional materials that show
a capability of recognizing a non-mechanical external stimuli from its surrounding
environment and reversibly respond to the same [2]. Such materials can judge the
magnitude of this external stimuli (signal ) and react with an optimal response by
either changing its physical or mechanical properties (generally macroscopic shape
change) [2, 3, 7]. The non-mechanical stimuli could be in the form of changes in
temperature, magnetic field, electric potential, light intensity, moisture, changes in
pH (chemical stimuli), and others [8]. Table1.1 and Fig. 1.1 illustrate the various
possible couplings between the the input signal (i.e. magnitude of external stimuli)
that the material is exposed to and its corresponding physical response for the same.
For example, magnetostrictive phenomenon involves changes in magnetic fields that
result corresponding changes in mechanical properties. Table 1.2 provides a summary
of some commonly used smart materials with their corresponding input signals that
show such multiphysics coupling.
1.2 Smart Structures—System Level Response
The combination of one or more smart materials can lead to a system level design
that can combine multiple functions in a single material system. With such a design
one can reduce system complexity and thus lowering system weight/volume [3]. To
illustrate this further, two examples are considered in Fig. 1.2. Lord Corporation has
developed controllable Magnetorheological (MR) dampers that can smartly adapt
to the terrain on which the vehicle is travelling with the goal of improving the
overall passenger ride comfort [ 11, 12]. This is made possible with the use of a
MR fluid1 that can change their viscosity under the action of external magnetic fields
1
A typical MR fluid consists of carbonyl iron particles in oil, petroleum products [13].
1.2 SmartStructures—SystemLevelResponse 3
Mechanical
ion
sha
i ct pe
t o str e
me
g ne
mo
iv
Ma
ry
i ct eff
str ect
ro
ct
le
E Magnetocaloric
Magnetic ic, Thermal
tr
ec
el
zo
c ie
ti P )
e
n li
g u
a m
ts i
P
m sl h
rot
ot
l a
ir
a
o
c ic et
el
el c
as
E tri m a ti
lec e m ci
oe h
rm (c
ty
t he g
e
n
il
g a
Electrical n e Optical
a h
h fl
c
e
H S
p
Chemical
Fig. 1.1 Smart materials can involve multi-physics coupling based on the external stimuli it is sub-
jected to which results in changes of physical/mechanical properties, that is, generally macroscopic
shape change (adapted from [9, 10])
[4, 14]. The ability of such materials to change their viscosities real time based on the
terrain makes them better alternatives compared to conventional dampers employing
standard SAE fluids. Such MR dampers have replaced the classic passive dampers
in many off-road and high end vehicles and are even finding applications as seismic
dampers in earthquake resistant structures [11, 15].
Biodegradable shape memory polymers (SMP) 2 have found many biomedical
applications in the form smart stents or sutures. The sutures used are programmed
to shrink when heated above their glass transition temperature Tg . In order to heal
a wound, a SMP fiber is loosely tied across the wound and then heated (above their
Tg ) for tightening the knot with an optimal force. Since the SMPs are programmed
to shrink upon heating, they remain as a knot till the wound is fully healed.
2
Thermally responsive SMP’s have shown the ability to recover large deformations when subjected
to external thermal stimuli. Above the glass transition temperature Tg , there is an onset of long
range molecular motion i.e. it changes from a glassy solid phase to an unordered rubbery phase and
vice versa [2]. During this process, SMP’s can exhibit large changes in shape and moduli which
can used for many applications [2].
Table 1.2 Commonly used smart materials with their corresponding input signals (i.e., the external
stimuli it is exposed to)
Smartmaterial Inputsignal
Shape memory alloys (SMA) Temperature/magnetic field
Shape memory polymers (SMP) Temperature/moisture/light/pH
Ionic polymer metal composite (IPMC) Electric field
Mangnetorheological fluids (MR fluids) Magnetic field
Electrorheological fluids (ER fluids) Electric field
Dielectric elastomer (DE) Electric field
Motor Vehicle System Smart Surgical Tool (Suture)
Entire product
Adaptable Structure
Adaptive Shock Absorber Self Tightening Suture
Sensor + Actautor + Control
(Single Entity)
SMP Fiber
MR Fluid Multifunctional Element +
+
Magnetic Field Capability of performing multiple functions Thermal Signal
(i.e Material as both Sensor + Actuator)
Magneto- Shape
rheological Material with “interesting” properties Memory
Fluid exhibiting “coupled” behaviour Polymer (SMP)
(MR Fluid)
Ex. : Sensing + External Stimuli + Mechanical response
Fig. 1.2 Integration of multiple functions like actuation, sensing, and control into a single structure
using one or more smart material constituents (adapted from [4])
In both of these examples, either the MR fluid or biodegradable SMP acts both
as a sensor and actuator with its ability to recognize non-mechanical stimulus (like
changes in temperature, pH, moisture, magnetic or electric fields etc.) from its sur-
rounding environment and subsequently change its mechanical properties as desired
for the application [2]. This integration of multiple functions like actuation, sensing
and control into a single structure has provided a system level design with the use of a
single material system. These materials particularly have the potential to completely
revolutionize the design of a wide variety of devices in applications ranging in areas
from biomedical, automotive, aerospace, civil engineering to energy harvesting [16].
1.3 Shape Memory Alloys: Temperature Induced Phase Transformations 5
1.3 Shape Memory Alloys: Temperature Induced Phase

Transformations
Shape memory alloys (SMAs) are a subset of a broad class of smart materials where
the functionalities arise from their underlying microstructural changes when sub-
jected to external non-mechanical stimuli like temperature or magnetic field changes
[17, 18]. In thermally responsive SMAs, the reversible solid-solid, diffusionless3
thermoelastic phase transformations between a stable high temperature austenitic
phase and low temperature martensitic phase are responsible for them to demonstrate
interesting phenomenon like shape memory effect (SME) and superelasticity (SE).
Their capability to return to a predetermined shape on heating is referred to as
the shape memory effect (SME). Their ability to recover large strains ( ∼8%) and
associated large stress–strain hysteresis due to mechanical loading–unloading under
isothermal conditions is referred to as superelasticity/pseudoelasticity (SE) [18].
The phase transformations are a result of shear lattice distortions (twinning) rather
than long range diffusion of atoms [1, 19]. The austenitic phase in SMA has a
ordered B2 cubic crystal structure as compared to the martensitic phase which has
either tetragonal, orthorhombic or monoclinic structures [20]. Due to its non-cubic

structure, the martensitic phase (B19 crystal structure) can have different orientations
(variants)4 [1, 19]:
•
“self-accommodated martensite variants” (also referred to as twinned martensite)
or
• “single variant martensite” (also referred to as detwinnned or reoriented marten-
site) with a specific variant being dominant.
Figure 1.3 shows the cubic structure of austenite and different variants of marten-
site. Under stress free conditions, the martensite in SMA’s are usually self-
accommodating in nature [22]. All of these phase transformations occur over
characteristic transformation temperatures namely martensitic start ( Ms ), marten-
sitic finish ( M f ), austenitic start ( As ) and austenitic finish ( A f ). ASTM standard
F2004-05R10 discusses the details of measuring these transformation temperatures
using a “Differential Scanning Calorimetry (DSC)” test for a Ni-Ti SMA [24]. The
test uses small, stress-free annealed SMA samples along with a reference sample
(empty Aluminium pans in some cases) in a DSC setup that is automatically capable
3
The word diffusionless suggests that the atoms do not leave their lattice positions and there is no
“long range diffusion” of atoms or species. Some literature on SMA commonly refer the twinning
events in phase transformations as “short range diffusion” where the atomic displacements are less
than the average interatomic distances.
4
Phase transformations in SMA (responsible for their functionality) under both SE and SME are
between austenite and single variant martensite variants. The “self-accommodated martensite” is
generally a combination of single variant martensite species [21]. For example, under tension and
compression loading cases, two unique single variant martensite species (say M t and M c for tension
and compression loading cases) can exist. In such a case, the “self-accommodated martensite”
version would be 50% M t and 50 % M c .
Austenite
• High temperature phase

• Cubic crystal structure
Ni
Ti
Martensite
Monoclinic (B19’)
Self accomadated Martensite Single variant Martesnite

(Twinned Martesnite) (Detwinned Martensite)
• Low temperature phase
• Monoclinic / Body Centered Tetragonal (BCT) crystal structure
Fig. 1.3 The underlying microstructural changes in shape memory alloys during thermoelastic
phase transformations occur between a stable high temperature austenitic phase and low temperature
martensitic phase. The non-cubic martensite phase can exist in different orientations (variants).
(adapted from [1, 23])
of heating and cooling the samples up to 10 ◦ C/min [24]. The DSC setup automat-
ically records the differential energy inputs between the specimen and reference
by monitoring the heat flow rates. The DSC setup automatically generates a heat
 J
flow gK
versus temperature (in ◦ C) [24]. Figure 1.4 shows an ideal DSC plot
with the tangents drawn on the heating and cooling spikes to determine the four
characteristic transformation temperatures. The peaks observed in the plot are the
latent heat of transformations due to austenite–martensite phase transformations in
SMA. These endothermic and exothermic peaks are characteristic to any first order
1.3 Shape Memory Alloys: Temperature Induced Phase Transformations 7
Stress free thermal cycling in DSC
Twinned Martensite Austenite
During Cooling
M A
←
Mf Ms
Twinned Martensite Austenite
During Heating
M A
→
As Af
Ms : Martesnitic Start Temperature

Mf : Martesnitic Finish Temperature
1 ∆Hheating
As : Austenitic Start Temperature
Af : Austenitic Finish Temperature
0.8 M A→
)
W Heating
(m As Af
w0.6
lo
F Mf Ms Cooling
ta
e
H0.4
M A ←
0.2 ∆Hcooling
-40 -20 0 20 40 60 80
Temperature C o
Fig. 1.4 Crystal structures of self-accommodated martensite and austenite along with associated
forward and backward transformation cases. Transformation temperatures martensitic start ( Ms ),
martensitic finish ( M f ), au stenitic start ( As ) and austenitic finish ( A f ) are obtained from a standard
DSC test as per ASTM standard F2004-05R10 (adapted from [1, 24])
phase transformations like melting of ice (solid–liquid), boiling of water (liquid–gas)

etc. [23]. However, in case of SMA’s, these phase transformations are solid–solid
(austenite–martensite) in nature.
1.4 Shape Memory Effect and

Superelasticity/Pseudoelasticity
The ability of SMA to return to a predetermined shape on heating above the char-
acteristic transformation temperature A f is referred to as the shape memory effect
(SME). A typical SME effect in SMA is depicted as path 1 – 6 on stress–strain–
temperature diagram as shown in Fig. 1.5. The various paths in SME are described
below.
• Path 1 – 2 : Under stress-free conditions, the self accommodated martensite vari-
ant favourably align themselves resulting in deformation at lower stresses without
any change in volume. This part of the response looks like a classic elastic defor-
mation of metal or alloy specimen.
• Path 2 – 3 : With higher loads, the self accommodated martensitic variants twins
into stress preferred single variant martensite (also called detwinned martensite)
typically associated with large (∼6–8 %) macroscopic strains and shape changes
[22]. These macroscopic strains are also referred to as martensite plateaus or the
twinning response of SMA purely due to mechanical loading. If deformed further
then the single variant martensite under higher stresses can undergo permanent
plastic deformation due to slip which cant be recovered.
• Path 3 – 4 : Unloading from point 3 results in elastic unloading of detwinned
martensite as shown
not completely returnin point
back to its4undeformed
[25]. It must be noted
state 1 and that
therethe
is amaterial does
clear residual
strain of about 4–6 % depending on the extent to which it was loaded. This residual
strains seems like classic permanent unrecoverable strains observed when metals
or alloys which are plastically deformed when beyond their yield point. However,
in case of SMA, this residual deformation can be recovered back upon heating and
thus making them a unique compared to classical materials.
• Path 4 – 5 – 6 : Upon heating above A f , residual strains are completely recovered
as the low symmetric martensitic phase (M) is transformed back to stable austenitic
phase (A). This complete strain recovery upon heating is referred to as shape
memory effect [ 22]. During the heating event, as the temperature reaches As
(point 5 ), the SMA starts transforming back to austenite and the transformation
is complete when the temperature reaches or exceeds A f , that is, point 6 .
• Path 6 – 1 : At point 6 , if the SMA is cooled below M f in absence of external
loads, austenite transforms back to self-accommodating martensite twins, that is,
point 1 without any volume change. This event indicates temperature induced
phase transformations in absence of any external loads. This event is commonly
observed as the cooling peak (latent heat of transformation) in a DSC test. It must
be highlight that this particular stress free transformation in quite rare in SMA
applications as the SMA’s are subjected to some form of external loads at all
times.
A simple illustration of SMA wire under external load demonstrating SME is
shown in Fig. 1.6. In this illustration, SMA wire is in its martensitic state A is
1.4 Shape Memory Effect and Superelasticity/Pseudoelasticity 9
Superelastic Effect : Path 7 – 8 – 9 – 10 – 11 – 12 – 13
Austenite Load
10
0.8
8 9
)
a Thermocouple
P 0.6
G
( Detwinned
s 0.4
rest 11
Martensite Extensometer
S 0.2 12
7,13
0
80
Test Specimen
70
6 Load
60 5
T Af
em 50
pe As
ra
t ur 40
e
C 3
30
Ms
20
Mf 2
10
4
1 0 2 4 6 8 10 12
Strain %
Twinned Martensite
Shape Memory Effect : Path 1 – 2 – 3 – 4 – 5 – 6
Fig. 1.5 A three-dimensional stress–strain–temperature plot describing the thermomechanical

experiment showing both shape memory effect (path 1–6) and superelastic effects (path 7–13).
A description of these states are discussed in Sect. 1.4. These reversible effects are manifestations
of solid–solid phase transformations between a stable high temperature austenitic phase and low
temperature martensitic phase (adapted from [23])
deformed to state B under external load. Upon heating above temperatures A f ,

the SMA wire contracts back to its austenitic state denoted by state C or hot state
position. Upon cooling, the SMA wire reaches state D. Cycle A ↔D is termed as
shape memory effect (SME).
On the other hand, these martensitic transformations can be induced purely due
to mechanical loading–unloading in the austenitic phase with SMA. The ability of
SMA to recover large strains ( ∼8 %) with associated stress–strain hysteresis due to
mechanical loading–unloading under isothermal conditions is referred to as super-
elastic/pseudoelastic effect (SE) [22]. These effects are observed at temperatures
State A State B State C State D
Loading Heating ( > Af ) Cooling
SMA wire speci men Hot State Position

Recovery
Mass (Stroke)
Cold State Position
Mass
Mass
Reference Position
Fig. 1.6 A simple example of SMA wire in its martensitic state A is deformed to state B under
external load. Upon heating above temperatures A f , the SMA wire contracts back to its austenitic
state denoted by state C or hot state position. Upon cooling, the SMA wire reaches state D. Complete
cycle A↔D is termed as shape memory effect (SME)
greater than A f as shown by path 7 – 13 in Fig. 1.5. The various paths are described
below.
• Path 7 – 8 : Above A f , under moderate external loads, the austenite elastically
deforms which is represented by path 7 – 8 . This looks like a classic elastic
deformation of metal or alloy specimen.
• Path 8 – 9 : Upon continued loading, at some critical stress, it is thermodynami-
cally stable for austenite to start transforming to a single variant martensite (also
commonly refereed to stress induced martensite (SIM) to differentiate this marten-
site variant from that formed due to twinning in SME). This forward transformation
from a cubic to monoclinic crystal structure results in large macroscopic defor-
mation of about 6–8 % strains. In a classic superelastic response under displace-
ment controlled experiment, this event occurs at relatively constant stress and both
austenite and martensite phases coexist during this event. Towards the end of the
transformation, most of the austenite has transformed to SIM. However, in reality
some pockets of austenite do not transform to martensite.
• Path 9 – 10 : Upon complete transformation (i.e., the end of flat plateau region), if
loaded further, this results in elastic deformation of SIM. However, higher loads can
cause plastic deformation of SIM and they cannot be recovered back completely.
In real world applications, this part of the response is rarely used in designing
SMA components due to poor fatigue life and thus the maximum strains in SMA’s
are restricted to about 6 % strains (i.e., up to point 8 ).
• Path 10 – 11 : Upon unloading from point 11 with the assumption that the stress
level at 11 did not induce any permanent plastic deformation, the SIM unloads
1.4 Shape Memory Effect and Superelasticity/Pseudoelasticity 11
until point 10 at which it is thermodynamically stable for it to transform back to

austenite.
• Path 11 – 12 : This represents the reverse transformation from SIM to austenite.
Similar to path 8 – 9 , it is accompanied by large macroscopic strains (in this
case its shortening) until it transforms completely back to austenite by time it
reaches 12 . This flat plateau is particular important for deigning many biomedical
applications that use SMA (details available in Sect. 1.6.1)
• Path 12 – 13 : In an ideal superelastic response, this path overlaps on path 7 – 8
and it represents the elastic unloading of the austenite. However, in most super-
elastic responses, there is generally some residual strains and points 13 and 7
do not generally overlap. The residual strains could be attributed to pockets of
austenite or SIM that do not transform back during either the forward or backward
transformations. During transformation regions 8 – 9 or 11 – 12 , the deforma-
tion is highly nonhomogeneous as both the phases coexist which might results
some pockets of SMA from not transforming from austenite to SIM or vice-versa.
Further, higher stress levels during the path 9 – 10 can also results in residual
strains due to possible plastic deformation of SIM.
An ASTM standard F2516-07 2 [26] discusses a standard tension test method-
ology on NiTi superelastic materials and explains the salient features of such super-
elastic effects.5 Some important salient features worth noting in a typical superelastic
response (see Fig. 1.7) are outlined here.

• Upper plateau strength (UPS) : The stress at 3% strain during loading of the
sample.
• Lower plateau strength (LPS) : The stress at 2.5% strain during unloading of
sample after loading to 6 %.

• Residual Elongation (Elr ): The difference between strains at a stress of 7 MPa
during loading and unloading operations.

• Uniform Elongation : Elongation determined at maximum force sustained by
specimen prior to necking or fracture or both.

The phase transformations in the superelastic response of SMA results in non-
linear hysteretic response which make them excellent candidates as damping materials
5
Commercially, the NiTi SMA’s are available as actuator wires or superelastic wires. In case of
actuator wires, the SMA is martensitic at room temperature, that is, the M f is slightly above the
room temperature and the A f may be around 60 to 100 ◦ C depending on the alloy composition
and other material processing conditions. Such actuator wires under external mechanical loads
at room temperature causes the self accommodated twins to detwin into more stress preferred
martensite variants and thus demonstrating macroscopic shape change. Such wires would show
superelastic behaviour above A f . On the other hand, the superelastic wires are austenitic at room
temperature, that is, its A f is below room temperature and its M f is far below sub zero temperatures
(around -60 to 100 ◦ C again depending on the alloy composition and other material processing
conditions). Such superelastic wires will show SE behaviour at room temperature and demonstrate
martensite detwinning at temperatures below zero (i.e., below M f ). A more detailed discussion
on the effect of composition and other material processing conditions will be taken up in future
chapters of this book.
Fig. 1.7 a Shows a typical (a)

superelastic response with 1200
internal loops of a NiTi SMA
with some salient features
like upper plateau strength 900 EA
(UPS), lower plateau )
a
strength (LPS), residual P
(M UPS
elongation (Elr ) highlighted ss
er 600 EM
as reported in ASTM t
S
standard F2516-07 [26]. In
addition, the austenitic and
300
martensitic moduli E A and
E M are shown. b Shows the LPS
temperature dependence of
the superelastic response at 2
2.5%
4 6 8
three different temperatures Elr 3%
Strain %
(b)
1000 348 K
373 K
800 398 K
)
a
P 600
M(
ss
e
Str 400
200
2 4 6 8
Strain %
due to their superior energy dissipation capabilities [27]. The ability of SMA to deliver
large plateau strains over relatively constant forces/stresses makes them good can-
didates as actuators [28]. Due to different crystal structures of austenite and marten-
site, SMA’s have different moduli over various parts of the superelastic response
[see Fig. 1.7a]. Further, the superelastic responses are also sensitive to the external
stimuli, that
atures, the is, operating
critical stressestemperature as shown
for transformation getinhigher
Fig. 1.7b. Withmaking
and thus increasing temper-
the material
response sensitive to temperature fluctuations.
1.5 Commonly Used Shap e Memory Alloy s
The martensitic phase transformations were initially observed in many steels (in a
Iron–carbon phase diagram) during the early 20th century. However, the reversible
martensitc phase transformations were underutilized until the discovery of SMA in
1.5 CommonlyUsedShapeMemoryAlloys 13
Ni-Ti-Ta
Ni-Ti-Cu
Au-Cu-Zn
Cu-Zn-Ga
Cu-Sn
Ag-Cd
Cu-Zn-Sn Ti-V-Pd
Cu-Zn-Si Zr-Co Ni-Ti-Co-V
Ti-Nb Fe-Mn Cu-Al-Ag
Cu-Al-Ni Fe- Pt Ni-Mn-Ga Cu-Al-Mn
U-Nb
Cu-Zn Ni-Al Ti-Pd-Ni Zr-Rh
Cu-Zn-Al Au-Cd-Zn Fe-Mn-Si Zr-Cu
In-Tl
Au-Cd Ni-Ti Ni-Ti-Fe Ti-Pd Fe-Ni-C
1950 1960 1970 1980 1990 2000
Fig. 1.8 Various shape memory alloys discovered over the second half of 20th century. The most
important commercially used alloys are highlighted in blue. (adapted from Wayman and Duerig
[29] and Shaw et al. [23])
early 1900s [1]. The SME in Au-Cd alloy was first reported in 1950s by Chang and
Read [29]. Since then, several other alloy systems (intermetallics) demonstrating
SME were discovered and many of them are summarized in Fig. 1.8 [23, 29]. Nearly
equiatomic Ni-Ti, Cu-Al-Ni, Cu-Zn-Al and Ni-Ti-Cu alloys are some of the com-
mercially exploited SMA in different applications. The most popular among these
is the Ni-Ti alloy that was discovered by William J. Buehler and coworkers from
the US Naval Ordnance Laboratory in White Oak, Maryland in 1963 and is often
referred to as NiTiNOL in its honour [1, 29]. A comparison of some important prop-
erties of three commercially available SMA’s with 316 stainless steel is tabulated
in Table1.3. In this book, most of the discussion will be restricted to NiTiNOL and
Cu-Zn-Al alloys.
Some High temperature SMAs (HTSMAs) like Pt-Ti, Ni-Ti-Pt, Ni-Ti-Hf, Ni-Ti-
Pd, Ni-Mn-Pt, Ni-Mn-Al, Ti-Ni-Zr, Cu-Al-Ni-Mn-Ti, Ru-Ta, Ru-Nb, and so on are
also under development where the transformation temperatures vary from 200 ◦ C to
as high as 700 ◦ C [35, 36]. Some ferromagnetic and metamagnetic SMAs like Fe-
Ni-Co-Ti, Fe-Ni-Co-Al, Fe-Ni-Mn-Al, Co-Ni-Al, Co-Ni-Ga, Ni-Mn-Ga, Ni-Mn-In,

Ni-Mn-Co, Ni-Mn-Sn, Ni-Fe-Ga, etc. are some alloys under investigation due to their
ability to work at higher frequencies [36]. However, further work needs to be done for
better understanding of underlying microstructural changes, operational limits, creep
and fatigue studies for HTSMAs, ferromagnetic and metamagnetic SMAs [35, 36].
l
e
e
t
] s
4 s
3 s
– e
l
0 n )c
[3 i
a it
t 2 sa
m S 2 8 - - le -
o rf 6 0 – – 4 3
. a a 7 a 0 say say say say
1 0 5 .8 6 .7 7 n n ( 9 n 0
d 3 5 1 7 2 0 1 - - 2 1 - 7 E E E E
ile
m p
o
c
taa
d (
le 0
0
tse 2
ss i 4 4 0 o
t tl tl
7 2
. 1
. 0 0 ) u u
el lN 5 5 7 0 6 0 0 c c d
– 7 9 – 1 0 fi fi
n A 3 – – – – 0 6 – 0 ira if if
o
ait u 7 0 1
. 0 1
. 7 0 0 – 0 2 0 o
s C 3 3 7 7 0 1 3 7 5 8 (– 3 F D D G
6
1
3
h
it
w
s’
A
M 0
S 5
1 d
el
b
l 0
2
1
. 0 o
t etd o
a A 0 0 8 0 0 0 ) c o
ail n
Z 4
–
0 2
1
– –
4 9 –
7 7
–
0 6 0
1
– 0 0 ri y
sa trise g
y
re
av u 9 4 5
. – 0
. 7 0 0 – 0 2 5 a
y C 3 8 7 7 0 1 4 4 4 7 (– 1 F E R V
lal
i
rec
m
m
co
ere 0
0
th 2
f 0 0 0 o
t lt lt lt
o 2 8 .5 .1 1 0 3 ) u u u
6 2 9 1 3 0 c c c
eis i –
0
1
–
6
– 3
–
1
–
1
– –
0
–
0 8
8
– 0 0 fi fi ri fi
tr iT 5 6
. 4. 9 5. 6
. 0 0 – 8 2
– 0 if if a if
e N 4 8 6 1 0 6 5 1 6 2 ( 4 D D F D
p
o r
p
t l
n
at ) o
)
C rt
r  C n
o 6 ) ( ◦
o
p 
0 K
) ◦(
er c
%
−
emi )1 )a (i A
d g
en tu n
Wm
K n o

) m tn P ) a an ar ar tii
m a r e s
so C Ω e
ic (M P st M re p p
o n
)
ss M n ( u t em
f
◦ 6 o
0 fi m is
o 0 fe er (s o i ) a t
−
2
n
o
)
C
(
y (1 o ts s at aP er g oc u
tr
s it c re m G p n i d x
rai ) y y r ( m at
◦
g v g ti n re ts o n e
p K tci K itv si
o
v g fs u
s et e
h
a
g
,
J(/ u J/ o n l
i n u n er
g n
m iss n n it lil
o ta d
n g3
k( p
a erc rk rat d t
o
i v g
n
y
t
C ta re x o a o acs o i lii
s
ie h
e co  m e la l e o
m w m M rm m (r rk b
,
.3 t c al y h ci a u l u ’s o u g g o
in
a
1 r
e fi
ic m ti t
n trc m
a
im m m i g fs m i n it in w- h
s r ax n
le p
e er en e er ax o u n x el
m d
l ac
b
a
o
r p h at el h o ar a o
r o
T P S T D L E T M N M Y T M M F C M
1.6 SMAApplications:Overview 15
1.6 SMA Applications: Overview
The ability of SMA to reversibly respond to external temperature changes and change
their physical/mechanical properties has enabled them to find many applications. In
a thermomechanical system, SMAs can be used as combined sensors and actuators
where they can sense the changes in external stimuli and monitor certain desired
functions [37]. The unique characteristics of SME and SE discussed earlier have
made SMAs the material system of choice in applications ranging from sensing and
control, vibration damping, biomedical, automotive and aerospace areas [16, 37–
40]. A review of many SMA devices in use across many engineering applications is
detailed in the following sections.
1.6.1 Biomedical Applications
In particular, NiTi SMAs have found many biomedical applications due to their
excellent biocompatibility.6 Studies have shown that Nickel on its own is quite toxic
and any contact with nickel can lead to various medical complications [41]. However,
in case of intermetallic NiTi alloys, the bonding between Ni and Ti is quite strong
(like in ceramic materials) as compared to the Nickel bonding in steel and other
materials [43]. Further, the commercial NiTi alloys uses a passive TiO 2 (titanium
oxide) layer coating on its outer surface that prevents any nickel leakage as it acts
as a physical and chemical barrier in preventing Ni oxidation [43, 44]. The TiO2
layer is harmless to human body and provides high resistance towards corrosion of
Ti alloys [41]. Many clinical studies have shown minimal Nickel contamination due
to use of NiTi SMAs and thus showing good biocompatibility [44].
In addition, NiTi SMA’s show excellent MRI compatibility, kink resistance, corro-
sion resistance and substantial moduli differences between austenitic and martensitic
phases. All of these properties make SMAs a good choice for many biomedical appli-
cations like: drug delivery systems, self-expanding stents, stent delivery systems,
implantable devices, catheters, guide-wires, atrial occlusion devices, and thrombec-
tomy devices [35, 41, 43–48]. In most of these biomedical applications, transforma-
tion temperatures of the SMA are programmed such that the Af is below the body
temperature [35]. The superelastic SMA components are cooled to their martensitic
state (i.e., below Mf ) and deformed to a temporary shape for easy insertion. Upon
6
Biocompatibility is dependent on the allergic reactions between the foreign material and the host
[41]. Material characteristics, patients health’s and several other factors play an important role for
assessing biocompatibility. In case of SMAs, several clinical studies were performed to establish Ni–
Ti alloys biocompatibility and FDA® has cleared many SMA products for medical use [41]. FDA®
is a registered trademark of “The Food and Drug Administration (FDA or USFDA)” affiliated with
the United States Department of Health and Human Services (an United States’ Federal executive
department) [42].
10
0.8
)
a 0.6
P
G 8 9
(
ss
e
tr 0.4
S
0.2 12 11 LPS
7,13 2 4 6 8
Strain %
Stainless steel
50
)
m
m 40
N
(
e β - Ti
u
q
r 30
o
T
20
NiTi - SMA
10
15 30 45 60 75 90
Twist (degrees)
Fig. 1.9 A SMA torsional spring used for tooth movement, space closure. The idea is to design
the spring at the LPS so that it can deliver constant forces over large strokes (Figure adapted from
[58]. Comparison graph adapted from [54])
deployment at the right location in the body, the SMA device is heated above Af (i.e.,
austenitic state) where the NiTi component recovers back to its srcinal shape and
performs the necessary function as desired. Some specific examples are discussed
below to illustrate this point.
SMA springs, wires and braces in many orthodontic applications are designed in
the LPS region so that they can deliver relatively constant forces over large activation
strokes as shown in Fig. 1.9. Their ability to deliver constant forces are commonly
employed for space closure and tooth movement in many orthodontic applications
[49–51]. Further, based on various test results, researchers have suggested that
SMAs provide superior spring-back properties, large recoverable strains thus making
them better alternatives when compared against its counterparts like stainless steel,
β -Ti, Co-Cr for medical applications [52–54]. Though the stainless steel counterparts
can deliver higher forces, however their force delivery rapidly decays over time as
compared to SMA springs for relatively long activation ranges as shown in Fig. 1.9
[55, 56]. In many of these dental applications, by choosing suitable wire diameters
of SMA, the required force can be varied based on application [57].
SMAs have found applications in many “minimal invasive surgery applications”
as they can easily pass through convoluted paths and reach various parts of human
body [ 43]. NiTi wires, tubes that demonstrate good kink resistance 7 are used to
maneuver complex paths and still remain in its srcinal shape at the time of delivery
[43]. Similarly, several occlusion systems uses catheters that generally consist of
NiTi wire loops on small umbrellas that support “microporous polyurethene webs”
at its ends [44]. The use of SMA here provides optimal occlusion in areas of limited
surgical exposure and access as the SMA loops spring back to their srcinal shape
upon release of external loads. Several scissors, biopsy forceps, graspers and other
endoscopic instruments use NiTi rods for actuation as they can undergo large bending
deformations without any permanent deformation (no kinks or buckling) [44].
SMAs are used in the neurosurgical field as coils, stents or microguidewires
mainly to treat cerebral aneurysms [59]. To prevent aneurysm rupture, coils are
positioned into the aneurysm to facilitate “clotting or thrombotic reaction within
the aneurysm” [59]. The superelastic effects of SMA are utilized here which allows
large deformations and prevent crushing of the coil. Microguidewires made of NiTi
are
[44,employed for stent
59]. Similarly, withpositioning
the goal ofdue to superior
trapping kink and
blood clots resistance and flexibility
dissolving them, many
cardiovascular devices have been developed to tackle pulmonary embolism and one
of the first SMA filter called “Simon Filter” was developed [41, 60].
Self expanding stents named in the honour of dentist C.T. Stent find applications
in cardiovascular applications with the goal of preventing collapse of blood vessels
[41, 45]. As shown in Fig. 1.10, NiTinol Stents are shape set in their deployed con-
figuration (generally expanded diameter in its austenitic state) and then compressed
into a catheter at lower temperatures below M f . Stents employed in arteries may
be subjected to continuously varying external pressures and collapsing or crushing
of deployed stents could result in serious medical complications [44, 46]. SMA
stents with their superior flexibility and spring back properties prevent the stents
from collapsing when compared to its counterparts [44]. The stents are designed to
work in the LPS region such that even higher external pressures does not allow them
to transform to their martensitic phase as the difference between plateau stresses
are quite significant. This ensures that the stents are in their austenitic state without
undergoing any permanent deformation under constant external pressures at all times
during their deployment.
7
These applications indicate the use of SMA like a “Metallic rubber band” without the component
loosing its shape. Superelastic SMA devices in their austenitic state provide high resistance to
deformation/kinks and keeping them in desired shape without any permanent deformation even
under considerable loads.
Sheath is removed and stent is

heated above A f to recover back to
its srcinal shape
Exapnded Ballon
Af
Ms
Compressed
Plaque
(a)
HT Phase
Widened artery
Mf As
Stiff
(Austentic phase) (e) Increased blood
BT : Body Temperature flow
RT : Room Temperature
LT : Low Temperature
cool Ballon
HT : High Temperature
deployed and
Heated
Artery
(c) deformed at LT
(b) LT shape Plaque
Compressed
deform (d) Stent
Soft Undeployed
(Martensitic phase) Ballon
Compressed
on a SheathStent insert
delivery Catheters
system
Compressed Cold Stent +
delivery system with steath
to reduce drag
Fig. 1.10 Smart Nitinol stents are deployed to improve blood flow in blocked arteries. a SMA stent
shape at high temperature (≈ID of artery); b SMA stent at room temperature (martensitic state)
under no external force; c SMA stent deformed at room temperature to a temporary shape that is
easy for insertion with the delivery system (typically balloon deployment; d Deformed SMA stent
inserted into required position using catheters; e SMA stent heated to high temperature (i.e., above
A f ) for it to recover back to its srcinal shape and unblock the artery (figure adapted from [61])
SMA spacers have found applications in many orthopaedic applications with the
intention of applying constant forces on fractured bones to accelerate the bone healing
process [41]. As shown in Fig. 1.11, spacers (sometimes addressed as staples) in its
opened (deformed) shape are deployed at the site and then heated so that the SMA in
its austentic shape can apply the required compressive forces to facilitate rebuilding
fracture bones [41, 47, 61].
All of these unique features makes the use of SMA components a feasible option
in many biomedical application with many of them already having FDA® approvals.
BT : Body Temperature
RT : Room Temperature
LT : Low Temperature
HT : High Temperature
Temperature HT Clamping due to deformation prevention
(e)
HT Phase
Stiff
BT
(a) (Austentic phase)
RT cool Heat
deformed at LT Zero Force insertion at LT

LT shape
(b) deform (c) Insert
Soft (d)
(Martensitic phase)
Break to be clamped
Fig. 1.11 A smart bone clamping technique using smart SMA clamps (staples) that provide enough
compressive forces to clamp small separations. a SMA clamp shape at high temperature; b SMA
clamp at room temperature (martensitic state) under no external force; c SMA clamp deformed at
room temperature to a temporary shape that is easy for insertion; d Deformed SMA clamp inserted
into required position; e SMA clamp heated to high temperature (i.e., above A f ) for it to recover
back to its srcinal shape and clamp due to deformation prevention (figure adapted from [61])
1.6.2 Civil Engineering Applications
In many civil engineering applications, large SMA wires, ropes, springs and beams
are being used as damping elements in bridges, buildings and also in seismic resisting
systems due to their excellent energy dissipation and recentering capabilities [39, 62–
65]. Figures 1.12 and 1.13 show different SMA reinforcements commonly used for
energy dissipation in civil engineering structures.
In seismically active areas, buildings and bridges can be prone to damage due to
lateral displacements during an earthquake event [66]. Building earthquake resistant
structures have been an intense area of interest lately and study of SMA rebars
as possible reinforcements in this pursuit has received significant attention (see
Fig. 1.13)(RC)
concrete [64, have
66]. The use of SMA
performed bettercomponents
in confiningaslateral
reinforcements in reinforced
column displacements
even at higher amplitudes of dynamic loading (i.e., simulating large earthquake
events) compared with their counterparts like steel-RC columns [66]. With supe-
rior damping capacity and ability to deliver large plateau strains at relatively con-
stant stresses, SMAs make excellent candidates for re-centering applications in the
form of bracings or dampers [66]. Many different types of SMA braces for framed
structures have been designed for energy dissipation devices [39] (see Fig . 1.13).
)
a 10
SMA Rods 0.8 P
G
(
s
s
e
SMA Springs r
t
S
0.6
8 9
0.4 Hysteresis - Energy

Dissipation
0.2 11
Typical SMA reinforcements 12
7,13 2 4 6 8
Strain %
Typical SMA Reinforcement locations

High rise Structures
Dampers
Bridges
Fig. 1.12 Common SMA reinforcements in many civil structures for recentering and damping
applications.
curve) The with
associated largesuperelastic
loading–unloading hysteresis
responses (shaded
in SMA makes region
them under
a very the entire
efficient energystress–strain
dissipation
system (photographs reproduced with permission from Elsevier [39, 67–69])
In many cases, use of bundles of post-tensioned superelastic SMA wires as rein-

forcements have been employed for re-centering braces [39].
In addition, SMA’s have also been effective contenders as beam–column or
column–foundation joints in many retrofit structures as the large plateau strains in SE
can be used to provide desired clamping forces to hold the members together [66].
This helps in controlling relative hinge displacements when used as restrainers [66].
By connecting the bearings and bridge deck with superelastic bars, the “position
stability” of simply supported bridges could be improved [68, 70] (see Fig. 1.13).
If a SMA reinforcement in its martensitic form is deformed (also referred to as
pre-tensioning) and embedded in concrete and then electrically activated such that it
reaches temperatures above Af , sufficient constraining forces can be generated as the
SMA tries to recover back to its original shape [66, 70]. The extent of prestressing can
be increased or decreased by controlling the amount of initial deformation of SMA
reinforcement [66]. To illustrate this point, Mo and co-workers looked at developing
a “smart concrete structure” that can potentially “self heal” after a damaging earth-
quake event [71]. In their study, as shown in Fig. 1.14, a concrete block with many
reinforced SMA ropes was considered [71]. The SMA reinforcements were initially
pre-deformed (pre-tensioning) in their martensitic state (i.e., below their M f ) and
reinforced in the concrete block with an initial crack as shown in Fig. 1.14. On a test
SMA Damper for Cable Bridge SMA Anchorage for Column
Concrete
Column
SMA Rods
SMA Damper Coupler
Steel Anchor
Bolts
SMA Restrainer SMA Connector
SMA bar
SMA rods
SMA Connector
SMA braces for frame structure
Beam
Steel Cables
n
m
lu
o
C
SMA Cables
SMA + Steel
reinforcements
Fig. 1.13 SMAcomponents are used as energy dissipation devices as dampers for simply supported
cable bridges, anchorages and connectors for columns, braces for framed structures (photographs
reproduced with permission from Elsevier [39])
Fig. 1.14 a Some

(a) Concrete Block with deformed SMA wire reinforcements
pretensioned SMA
reinforcements in a concrete
block with an initial crack.
b Shows a classic 3-point
bend setup. c Shows the
crack propagation during
loading. d Shows the crack Initial Crack
closure upon unloading and
activating (heating) SMA
reinforcements. The SMA Three point bend setup
wires shorten causing the
crack to close (figures
adapted from [39, 71, 72])
(b)
Initial Crack
Crack Propogation during loading test
(c)
Crack closure upon unloading and SMA activation

(Notice shortening of SMA reinforcements)
(d)
bench, dynamic 3-point bend tests were conducted which resulted in crack propa-
gation with increasing

The cracked structure isloads/frequencies
then unloaded andmimicking a damaging
the SMA ropes earthquake
were activated event.
by heating
them above their A f . Upon heating, the SMA ropes transform back to austenite and
shorten in length back to its srcinal shape. This results in large recovery forces that
results in crack closure (crack healing) as the concrete structures are pulled inwards
and thus minimizing the chances of building collapse after an earthquake event. Many
such “self-restoration” options have been proposed with almost complete crack clo-
sure upon re-centering [39]. On similar lines, many active confinement techniques
where concrete columns are helically wrapped with SMA bands that can provide
necessary tensioning upon heating have also been proposed [68].
In one of the first reported retrofit applications using SMA devices, the San Giorgio
chruch in Italy used bundles of SMA wires to refurbish the church structure after
an earthquake event [64, 69]. As shown in Fig . 1.12, SMA devices were arranged
in series in order to limit the forces to the masonry under the required limits and
further seismic analysis were performed to analyze the performance of SMA devices
[64, 69].
However, given the size of civil engineering structures, the use of NiTi SMA
components have been limited due to their high initial material and processing costs
when compared to biomedical or automotive/aerospace applications [68]. Going for-
ward, developing more cheaper copper or iron based SMAs that could replace NiTi
SMA components would greatly benefit the civil engineering community which can
potentially use many more smart structures with active control abilities . Develop-
ing such copper or iron based SMAs and further understanding their underlying
microstructural changes that influence their functionality is still an active area of
research [35].
1.6.3 Aerospace and Automotive Applications
SMA components are used as thermal actuators in different temperature regimes
depending onfinding
actuators are the kindapplications
of applications in automotive
in many industry [16].
pre-commercialized The use
concepts of smart
like SMA
automotive lighting systems, fuel management, climate control, mirrors adjustments,
locking systems, suspension adjustments, and so on [16, 73, 74]. Similarly, SMA
cables, beams, torque tubes are used in tailoring inlet geometry and orientation,
increase/decrease fan nozzle in various aerospace applications [37, 75–77].
In many European cars, air intake manifolds along with tumple flaps are mounted
on the crankshaft. The crank reciprocates between two angular positions to open
and close series of tumble flaps upon demand. The traditional arrangements uses
electromagnetic or pneumatic actuators for achieving this reciprocating motion [74].
In one proposed case study, sets of SMA springs were used to generate the required
actuation forces (push and pull out forces) to switch the tumple flap’s angular posi-
tions against existing traditional electromagnetic or pneumatic devices [74]. Battery
power supplies were used to actuate the SMA springs upon demand.
In some aircraft applications, smart SMA couplings have shown the possibil-
ity of substitute socket welds or compression fittings [78]. As shown in Fig. 1.15,
at temperatures below M f , the couplings are expanded to a larger diameter such
that it can easily slide through the pipes to be fastened [78]. Once the SMA cou-
pling transforms back to austenite (i.e., above Af which is generally programmed
to be lower than room temperature), it shrinks to its srcinal dimension and thus
securely holding the pipe ends as the SMA coupling enforces sufficient compressive
forces [78].
Chopra [ 76] reviews many possible applications of using smart materials in
aerospace systems. For example, adjusting the “trailing edge tab of a helicopter
Inactive Coupler Active Coupler
Pipe A Pipe B
Fig. 1.15 SMAs can be used as couplers that can substitute socket welds or compression fittings
in aerospace, civil, oil exploration industries. The idea is to deform the couplers in its martensitic
state to larger diameters for easy sliding along the pipes and then heated to temperatures above A f
to hold the ends of pipes firmly with appropriate compressive forces (figure adapted from [78, 79])
Insulation Material SMA rods

Top wire
Rotor Assembly
actuated
Bottom wire
actuated
Fig. 1.16 The use of pretwisted SMA rods to torsionally actuate blade tabs between two angular
positions is possibility by differentially heating of top and lower SMA rods. This way the angular
positioning of “trailing edge tab of a helicopter rotor” can be adjusted and thus allowing for user in-
flight tracking of its position (figure reproduced from [76] with permission from AHS International)
rotor for in-flight tracking” using SMA components is one of the key proposed ideas
[76]. Pretwisted SMA rods employed to torsionally actuate blade tabs between two
angular positions is possibility by differentially actuating the SMA rods as shown in
Fig. 1.16. By heating the upper rod, the SMA rod shortens and causes the rotor assem-
bly to actuate in one direction. By heating the lower rod, the SMA rod enables the
rotor assembly to actuate in a different direction. The actuation of the rotor assembly
in different directions causes the trailing edge to occupy multiple angular positions
during flight maneuver. The SMA rods clearly perform both sensing and actuating
functions and thereby reducing the use of separate working elements for sensors and
actuators to perform the same task.
On similar lines, “Chevrons” in aircraft engines are employed for flow mixing
and their configurations can be altered by using SMA beams based on changes in
flow temperature (normally temperature changes with different altitudes) [37]. At
low altitudes and speeds, with higher engine temperatures, the chevrons fold inward
for improved gas mixing and thus lowering noise [ 37]. These chevrons relax to a
different configuration at lower temperatures (i.e., at higher altitudes and speeds) to
improve engine performance [37].
Japanese bullet trains came up with some smart automatic oil adjustment devices
that use SMA coil springs along with a bias spring arrangement with the intention
of preventing temperature rise in gearbox (see Fig. 1.3 in [80]). The oil temperature
can be used to actuate the SMA spring and open and close bypass holes [80].
1.6.4 Miscellaneous Applications
The oil and gas industry uses many hydraulic and electro-hydraulic systems along
with additional signal conditioning components which makes the overall system
quite bulky [79]. Many SMA components along with bias systems have been con-
sidered to replace these traditional hydraulic and electro-hydraulic systems [79]. As
shown in Fig. 1.17, a combination of SMA and steel bias springs were employed
to actuate pistons in “blowout preventer’s” that either block or unblock drill pipes
(a) Inactive Position
Fixed
SMA spring Moving Piston Sliding Supports with
Fluid Bias Spring Plugging Unit Drill Casing (Pipe)
Guide hard stops
External Heating
(b) Active Position (Actuated)

Fig. 1.17 A combination of SMA springs and bias springs are employ ed to actuate pistons in
“blowout preventer’s” that either block or unblock drill pipes. a Shows the the inactive position of
blocking piston where the bias springs keeps the SMA spring compressed at lower temperatures
(i.e., below M f ). b Shows the SMA spring in its austenitic state (at higher temperatures, i.e. above
A f ) overpowers the bias spring forces and actuates the blocking piston to prevent any flow in drill
pipe to prevent any accidents (figure adapted from [79])
Fig. 1.18 SMA rods used as High Pressure

smart valves to close/open
valves. SMA rods in the core
are surrounded by Nichrome
wires for resistive heating.
At lower temperatures, SMA Heating
rods are inactive and upon
heating of SMA rods results
in enough recovery force
generation to close/open
valves depending on design
Thermal Nicrohme wire

Insulation
SMA Rod
[79]. With reference to Fig. 1.17, position (a) indicates the inactive position of the
blocking piston as at lower temperatures (i.e., below M f ), the bias spring keeps the
SMA spring compressed and thus allowing free flow through the drill pipes during
operation. At higher temperatures (i.e., above A f ), the SMA spring in its austenitic
state overpowers the bias spring forces and actuates the blocking piston to block any
flow in drill pipe to prevent any blowout situation. On similar lines, another example
of a smart valve is discussed in Fig. 1.18 that can open or close the valve when the
SMA is actuated by changes in temperature.
Due to NiTi alloys excellent corrosion resistance properties, many underwater
couplers/connector applications are considered [79]. SMA springs are cryogenically
cooled to temperatures below M f so that they can easily be deformed into a temporary
shape for deployment (generally expanded to larger diameter for easy deployment)
[79]. On exposure to actual operating temperatures (above A f ) they can revert back
to austenitic phase and tightly hold the pipe ends [79]. The couplers in their austenitic
state can provide the required compressive forces to hold the pipe ends in place. On
similar lines, some smart couplings have also been tested for submarine applications
and underwater gas pipelines as shown in Fig. 1.15 earlier. Such couplings are easy
to work with as it does not require any initial pipe surface preparation or subsequent
welding or safety hazards [78].
In some fire protection devices, the ability of SMA to reversibly respond to changes
in temperature have enabled them to find applications like safety valves [35]. Use of
Cu-Zn-Al actuators are responsible for shutting off toxic flammable gases [35]. The
idea here is that when the high temperature toxic flammable gases are sensed by the
safety valves, they transform to their austenitic state which forces the safety valve to
shift to its srcinal shape that shuts the flow of toxic gases and force them through
another exit.
Normal Using SMA

Arrangement SUN
SUN
Sensor
SMA
Controlling Wire or Spring
Actuator
Unit
Mechanical
Electrical Energy Arrangements
Fig. 1.19 Comparing the number of working elements with existing sun tracking system ( left )
versus number of working elements for a proposed smart sun tracking system ( right ) using SMA
components that play the role of actuator and sensor simultaneously (figure adapted from [81])
On a similar note, SMA compression springs along with bias steel springs are find-
ing applications in sprinkler systems that can automatically turn on or off based on
the water temperature [35]. Ni-Ti-Cu springs are finding applications in thermostats,
electrical circuit breakers where, for example, door positions in an self-cleaning
oven are monitored using SMA components [35]. NiTi discs located between elec-
trical contacts in Lithium ion batteries can be used to break the circuits at critical
temperatures [35].
Ganesh and co-workers [81] have attempted to design a simple “sun tracking
mechanism (SSTM)” to illustrate the dual sensing and actuating capability of SMA
components. The solar receptor position is automatically changed over the course of
the day by using the solar radiation to actuate the SMA springs [81]. As described
in Fig. 1.19, the use of such smart SMA springs considerably reduces the number
of working elements as compared against a convectional sun tracking system that
employs separate sensor and actuator for providing system level response [81].
1.7 Chapter Summary
The following are the main points of this chapter.
• Smart materials are a subgroup of advanced materials that can recognize non-
mechanical external stimuli like changes in temperature, magnetic field, electric
field, pH changes, and so on, and reversibly respond to the same by either changing
its physical or mechanical properties.
• The use of smart materials can provide a system level design by integrating multiple
functions like actuation, sensing and control into a single structure with the use
of one or more smart materials that can replace separate sensing and actuator
components.
• Shape memory alloys (SMAs) is one such smart material where the functionalities
arise from their underlying microstructural changes when subjected to external

non-mechanical stimuli like temperature or magnetic field changes.
• In thermally responsive SMAs, the reversible solid-solid, diffusionless thermoelas-
tic phase transformations between a stable high temperature austenitic phase and
low temperature martensitic phase are responsible for them to demonstrate inter-
esting phenomenon like shape memory effect (SME) and superelasticity (SE).
• Phase transformations occur over characteristic transformation temperatures
namely martensitic start ( Ms ), martensitic finish ( M f ), austenitic start ( As ) and
austenitic finish (Af ).
• SME is the ability of SMA to return to a predetermined shape upon heating above
the characteristic transformation temperature A f ; SE is the ability of SMA to
recover large strains (∼8 %) with associated stress-strain hysteresis due to mechan-
ical loading-unloading under isothermal conditions above A f .
• The unique characteristics SME and SE enabled SMAs to find applications across
many areas like vibration damping, biomedical, automotive and aerospace areas.
Smart structures with active control abilities is now a possibility with the use of
SMA.
References
1. Lagoudas D (2008) Shape memory alloys: modeling and engineering applications. Springer,
New York
2. Ghosh P, Rao A, Srinivasa AR (2013) Design of multi-state and smart-bias components using
shape memory alloy and shape memory polymer composites. Mater Des 44:164–171
3. Momoda LA (2004) The future of engineering materials: multifunction for performance-
tailored structures, vol 34. National Academy of Engineering, The Bridge
4. Janocha H (1999) Adaptronics and smart structures: basics, materials, design, and applications.
Springer, Heidelberg
5. Matic P (2003) Overview of multifunctional materials. In: Proceed ings of SPIE, vol 5053,
p 61 P, Rao A, Srinivasa A (2012) Multifunctional smart materia l system (msms) using

6. Ghosh
shape memory alloys and shape memory polymers. In: SPIE smart structures and materi-
als+nondestructive evaluation and health monitoring. International society for Optics and Pho-
tonics, pp 83411O–83411O
7. Okano T (1993) Molecular design of temperature-responsive polymers as intelligent materials.
Responsive Gels Vol Transit II 110:179–197
8. Leng J, Lan X, Liub Y, Dua S (2011) Shape-memory polymers and their composites: stimulus
methods and applications. Prog Mater Sci 56(7):1077–1135
9. Noor AK (1999) Computational structures technology: leap frogging into the twenty-first
century. Comput Struct 73(1):1–31
10. Naraghi M (2012) Multifunctional materials—aero 606 course notes, Fall 2012
References 29
11. Corporation L (2012) Company website, 2012. http://www.lord.com

12. Leventon W (1993) Fluid damper may make truck seats safer. Des News 4:185–186
13. Carlson JD (2002) What makes a good Mr fluid? J Intell Mater Syst Struct 13(7–8):431–435
14. Carlson JD (2005) Mr fluids and devices in the real world. Int J Mod Phys B 19(07n09):1463–
1470
15. Carlson J, Catanzarite D, Clair KS (1996) Commercial magneto-rheological fluid devices. Int
J Mod Phys B 10(23n24):2857–2865
16. Stoeckel D (1990) Shape memory actuators for automotive applications. Mater Des 11(6):302–
307
17. Huang W, Ding Z, Wang C, Wei J, Zhao Y, Purnawali H (2010) Shape memory materials.
Mater Today 13(7–8):54–61
18. Doraiswamy S, Rao A, Srinivasa A (2011) Combining thermodynamic principles with preisach
models for superelastic shape memory alloy wires. Smart Mater Struct 20(8):085032
19. Van Humbeeck J (2001) Shape memory alloys: a material and a technology. Adv Eng Mater
3(11):837–850
20. McNichols J, Cory J (1987) Thermodynamics of nitinol. J Appl Phys 61(3):972–984
21. Rao A (2013) Modeling bending response of shape memory alloy wires/beams under super-
elastic conditions-a two species thermodynamic preisach approach. Int J Struct Changes Solids
5:1–26
22. Wayman C (1992) Shape memory and related phenomena. Prog Mater Sci (UK) 36:203–224
23. Shaw J, Churchill C, Iadicola M (2008) Tips and tricks for characterizing shape memory alloy
wire: part 1 differential scanning calorimetry and basic phenomena. Exp Techn 32(5):55–62
24. A. S. F2004–05R10 (2010) Standard test method for transformation temperature of nickel-
titanium alloys by thermal analysis, ASTM International, West Conshohocken, PA
25. Shaw JA, Kyriakides S (1995) Thermomechanical aspects of niti. J Mechan Phys Solids
43(8):1243–1281
26. A. S. F. 2007e2 (2007) Standard test method for tension testing of nickel-titanium superelastic
materials, ASTM International, West Conshohocken, PA
27. Rao A, Srinivasa AR (2014) A three-species model for simulating torsional response of shape
memory alloy components using thermodynamic principles and discrete preisach models. Math
Mech Solids doi:10.1177/1081286514545917
28. Rao A, Srinivasa A (2013) A two species thermodynamic preisach model for the torsional
response of shape memory alloy wires and springs under superelastic conditions. Int J Solids
Struct 50(6):887–898
29. Wayman C, Duerig T (1990) An introduction to martensite and shape memory. Butterworth-
Heinemann, Engineering Aspects of Shape Memory Alloys (UK), 1990, pp 3–20
30. Otsuka K, Wayman CM (1999) Shape memory materials. Cambridge University Press, Cam-
bridge
31. Duerig T, Pelton A (1994) Ti-ni shape memory alloys. Materials properties handbook: titanium
alloys, pp 1035–1048
32. Huang W (2002) On the selection of shape memory alloys for actuators. Mater Des 23(1):11–19
33. Baumann MA (2004) Nickel-titanium: options and challenge s. Dent Clin North Am 48(1):
55–68
34. Leng J, Lan X, Liu Y, Du S (2011) Shape-memory polymers and their composites: stimulus
methods and applications. Prog Mater Sci 56(7):1077–1135
35. Bogue R (2009) Shape-memory materials: a review of technology and applications. Assembly
Autom 29(3):214–219
36. Ma J, Karaman I, Noebe RD (2010) High temperature shape memory alloys. Int Mater Rev
55(5):257–315
37. Hartl D, Lagoudas D (2007) Aerospace applications of shape memory alloys. Proc Inst Mech
Eng Part G J Aerosp Eng 221(4):535
38. Sutapun B, Tabib-Azar M, Huff M (1998) Applications of shape memory alloys in optics. Appl
Opt 37(28):6811–6815
39. Song G, Ma N, Li H (2006) Applications of shape memory alloys in civil structures. Eng Struct
28(9):1266–1274
40. Van Humbeeck J (1999) Non-medical applications of shape memory alloys. Mater Sci Eng A
273:134–148
41. Machado L, Savi M (2003) Medical applications of shape memory alloys. Braz J Med Biol
Res 36(6):683–691
42. FDA (2012) United states food and drug administartion, 2012. http://www.fda.gov
43. Lagoudas D, Rediniotis OK, Khan MM (2000) Applications of shape memory alloys to bio-
engineering and biomedical technology. In Proceedings of the 4th international workshop on
scattering theory and biomedical applications, Perdika, Greece, pp 195–207 Oct 1999
44. Duerig T, Pelton A, Stöckel D (1999) An overview of nitinol medical applications. Mater Sci
Eng A 273:149–160
45. Mantovani D (2000) Shape memory alloys: properties and biomedical applications. JOM J
Miner Met Mater Soc 52(10):36–44
46. Morgan N (2004) Medical shape memory alloy applicationsthe market and its products. Mater
Sci Eng A 378(1):16–23
47. El Feninat F, Laroche G, Fiset M, Mantovani D (2002) Shape memory materials for biomedical
applications. Adv Eng Mater 4(3):91–104
48. Gil F, Manero J, Planell J (1996) Relevant aspects in the clinical applications of niti shape
memory alloys. J Mater Sci Mater Med 7(7):403–406
49. Manhartsberger C, Seidenbusch W (1996) Force delivery of ni-ti coil springs. Am J Orthod
Dentofac Orthop 109(1):8–21
50. Miura F, Mogi M, Ohura Y, Karibe M (1988) The super-elastic japanese niti alloy wire for use
in orthodontics part iii. studies on the japanese niti alloy coil springs. Am J Orthod Dentofac
Orthop 94(2):89–96
51. Spinella I, Dragoni E, Stortiero F (2010) Modeling, prototyping, and testing of helical shape
memory compression springs with hollow cross section. J Mech Des 132:061008
52. Drake S, Wayne D, Powers J, Asgar K (1982) Mechanical properties of orthodontic wires in
tension, bending, and torsion. Am J Orthod 82(3):206–210
53. Miura F, Mogi M, Ohura Y, Hamanaka H (1986) The super-elastic property of the japanese
niti alloy wire for use in orthodontics. Am J Orthod Dentofac Orthop 90(1):1–10
54. Kapila S, Sachdeva R (1989) Mechanical properties and clinical applications of orthodontic
wires. Am J Orthod Dentofac Orthop 96(2):100–109
55. Von Fraunhofer J, Bonds P, Johnson B (1993) Force generation by orthodontic coil springs.
Angle Orthod 63(2):145
56. Schneevoigt R, Haase A, Eckardt V, Harzer W, Bourauel C (1999) Laboratory analysis of
superelastic niti compression springs. Med Eng Phys 21(2):119–125
57. Thompson S (2000) An overview of nickel-titanium alloys used in dentistry. Int Endod J
33(4):297–310
58. Maganzini AL, Wong AM, Ahmed MK (2010) Forces of various nickel titanium closed coil
springs. Angle Orthod 80(1):182–187
59. Petrini L, Migliavacca F (2011) Biomedical applications of shape memory alloy s. J Metal
2011:15, Article ID 501483. doi:10.1155/2011/501483
60. Ryhänen J (1999) Biocompatibility evolution of nickel-titanium shape memory alloy. Academic
Dissertation, Faculty of Medicine, University of Oulu, Oulu, Finland.[Links]
61. Poncet PP (2000) Nitinol medical device design considerations. Strain 2(4):6
62. Speicher M, Hodgson D, DesRoches R, Leon R (2009) Shape memory alloy ten-
sion/compression device for seismic retrofit of buildings. J Mater Eng Perform 18(5):746–753
63. Saadat S, Salichs J, Noori M, Hou Z, Davoodi H, Bar-On I, Suzuki Y, Masuda A (2002) An
overview of vibration and seismic applications of niti shape memory alloy. Smart Mater Struct
11:218
64. DesRoches R, Smith B (2004) Shape memory alloys in seismic resistant design and retrofit: a
critical review of their potential and limitations. J Earthq Eng 8(3):415–429
References 31
65. Wilson J, Wesolowsky M (2005) Shape memory alloys for seismic response modification: a
state-of-the-art review. Earthq Spectra 21:569
66. Alam M, Youssef M, Nehdi M (2007) Utilizing shape memory alloys to enhance the perfor-
mance and safety of civil infrastructure: a review. Can J Civ Eng 34(9):1075–1086
67. Mirzaeifar R, DesRoches R, Yavari A (2011) A combined analytical, numerical, and experi-
mental study of shape-memory-alloy helical springs. Int J Solids Struct 48(3–4):611–624
68. Janke L, Czaderski C, Motavalli M, Ruth J (2005) Applications of shape memory alloys in
civil engineering structuresoverview, limits and new ideas. Mater Struct 38(5):578–592
69. Indirli M, Castellano MG, Clemente P, Martelli A (2001) Demo-application of shape memory
alloy devices: the rehabilitation of the s. giorgio church bell tower. In: SPIE’s 8th annual
international symposium on smart structures and materials, International Society for Optics
and Photonics, pp 262–272
70. DesRoches R, Delemont M (2002) Seismic retrofit of simply supported bridges using shape
memory alloys. Eng Struct 24(3):325–332
71. Mo Y, Song G, Otero K (2004) Development and testing of a proof-of-concept smart concrete
structure. In: Proceeding of smart structures technologies and earthquake engineering
72. Tamai H, Kitagawa Y (2002) Pseudoe lastic behavior of shape memory alloy wire and its
application to seismic resistance member for building. Comput Mater Sci 25(1):218–227
73. Williams K, Chiu G, Bernhard R (2002) Adaptive-passive absorbers using shape-me mory
alloys. J Sound Vib 249(5):835–848
74. Bellini A, Colli M, Dragoni E (2009) Mechatronic design of a shape memory alloy actuator
for automotive tumble flaps: a case study. IEEE Trans Ind Electron 56(7):2644–2656
75. Loewy R (1997) Recent developments in smart structures with aeronautical applications. Smart
Mater Struct 6:R11
76. Chopra I (2000) Status of application of smart structures technology to rotorcraft systems.
J Am Helicopter Soc 45(4):228–252
77. Razov A, Cherniavsky A (1999) Applications of shape memory alloys in space engineering:
past and future. Eur Space Agency Publ ESA SP 438:141–146
78. Brook G (1983) Applications of titanium-nickel shape memory alloys. Mater Des 4(4):835–840
79. Song G, Patil D, Kocurek C, Bartos J (2010) Applications of shape memory alloys in offshore
oil and gas industry: a review. In: Proceedings of earth and space 2010: engineering, science,
construction, and operations in challenging environments, vol 366, Honolulu, HI, USA, pp
14–17 March 2010
80. Otsuka K, Kakeshita T (2002) Science and technology of shape-memory alloys: new develop-
ments. MRS Bull 27(02):91–100
81. Ganesh N, Maniprakash S, Chandrasekaran L, Srinivasan S, Srinivasa A (2011) Design
and development of a sun tracking mechanism using the direct sma actuation. J Mech Des
133:075001
Chapter 2
Need and Functionality Analysis
Generally the word “smart” is used synonymously with “intelligent,” but in reality
the dictionary definition of “smartness” is “marked by often sharp forceful activity
or vigorous strength,” that is, it is active. Thus, a “smartphone” can only wake you
up with an alarm, but in reality it is smart only if it can “slap” you if you don’t wake
up! Fig. 2.1 shows the hierarchy of systems ranging from “passive” to “intelligent”.
YourIn fact there is

computer orayour
fundamental
cell phonedivide betweenYou
is intelligent. being
dointelligent and
what it tells being
you smart.
to do. Think
about this for a minute. Your computer is your “boss”. You wake up and look to it
for instructions, (email, calendar, etc.) you do what it says and then you report back
to it. If that is not a definition of a “boss,” we don’t know what is? On the other hand,
your car is like your valet or butler: you tell it what to do and it actually carries out
the actions. These are “actuators” that perform actions (typically mechanical) based
on our command. It is the latter category that we need to focus on.
The capabilities of your car came from the assembly or configuration of its parts,
just like the first “valves” were made by assembling parts. Today, most actuators
(think solenoids, electric motors, etc.) are in this stage of development. Over time, the
valves gave way to transistors (where the configuration is microscopic) and finally
to very large scale integrated circuits (VLSI) where the functionality is from its
“microscopic” configuration and no further assembly is required. We are gradually
heading to this stage in actuators too and the challenges are immense due to the
need for motion. Just as information technology is based on the ability to store,
manipulate and retrieve information (bits), actuator technology is based on the ability
to store, retrieve and manipulate shapes or geometry. For cars, and electric motors, this
ability to change geometry is based on the assembly of parts. Shape memory alloys
(SMA) have this ability because of their “microstructure.” They are the equivalent
of transistors.
In order to have the ability to store shapes and retrieve them on demand, we need
to supply two things: Energy to transform the shape and a means for transmitting a
“signal” to initiate the shape change. For a typical SMA, the energy is in the form
34 2 NeedandFunctionalityAnalysis
Structure
PASSIVE
Responds, no quantification, health status not known
Structure + Sensor
SENSORY
Responds, quantification, health status known
Structure + Sensor + Actuator

ACTIVE
Open Control : Corrective measures for a desired response
Structure + Sensor + Processor + Actuator

SMART
Extrinsic Control : Response for an enviornmental stimulus
Structure + Sensor + Processor with AI + Actuator

INTELLIGENT
Integrated response & adaptation to the enviornment changes
Fig. 2.1 Smartness in systems—a hierarchy
The System LEVEL IV
Smart Component/Structure LEVEL III
Multifunctional element LEVEL II
Material with“interesting”properties LEVEL I
Fig. 2.2 Smart system: an integrated use of components/structures, actuators, sensors and control
systems that can reversibly respond to changes in external conditions
of heat and the signal is a temperature signal. The core idea here is to find or create
materials with “interesting” properties and create such systems (see Fig. 2.2). The
question is how best to design components that can exploit this ability? We turn to
this in the next section.
2.1 TheSystemDesignProcess 35
2.1 The System Design Process
One of the key activities that distinguishes humans from other animals is our ability
to design, that is, create or synthesize something that we can use that wasn’t there
before. This activity of designing is something that we take so much for granted that
we automatically do it without paying much attention.
So what is this activity called designing and how do we do it? Design is the act
of identifying societal needs (who is the end user? and what do they need?) and
constraints; quantifying them and using a combination of scientific principles and
experience to synthesize products and processes to satisfy the needs. This combina-
tion of products or devices and processes that they can undergo is called a “system.”
Thus we have the electric grid system, the sanitation system, the transport system,
the computer network system, etc.
The above quote shows the essential difference between a scientist and an engi-
neer: a scientist typically creates “explanations”. An engineer creates products and
processes. In other words, when you are explaining how nature works, you are under-
taking a scientific activity. On the other hand, when you are building a product or a
process, you are undertaking an engineering activity.
Thus all desig ns start with the the idea of “I/we want...,” “I/we need...,” or
“wouldn’t it be nice if we could...” and ends with a specific product (cell phone) or
process (wireless transmission). For a design example, if one looks at a cell phone, it
is a product that allows to communicate and record information (need) in a portable
way (constraint). However, the task of designing is not a simple linear process start-
ing from a need and ending in a product. It is fundamentally iterative: We don’t quite
know what we want when we start but as we begin building we discover things and
improve our design. There are thus three major activities that occur in a cycle (see
Fig. 2.3) involved in designing things:
1. identification of needs and constra ints,
2. embodiment or synthesis of product or proce ss, and
3. evaluation of the product or process to see if it meets the need or constrain ts.
Typically, if you are a novice designer, (remember your high school project), you
will do this iterative process at a subconscious level, without thinking deeply about
it. In particular, you probably built and rebuilt your product until it worked (probably
many times). However, this is not a sensible way to carry out large projects. Imagine
building a bridge this way: You “eye ball” the river and begin building; at the end, if
your bridge doesn’t reach the other end, or if the first truck that goes across it plunges
into the river, you say “oops ... let us rebuild it,” and start over. This would not be a
viable process. Similarly, if you are building a rocket going to the moon, you don’t
say, “oops ... missed again,” and then rebuild the rocket. The lesson to learn from
these admittedly extreme examples is that avoid a simple trial and error approach
when designing.
Creation + Abstraction + Evaluation
SPECIFICATION
NEED ANALYSIS
How do we
What is needed?
recognize/measure success?
ABSTRACTION/Functionality REDESIGN
What tasks need to be done?
EVALUATION
IDEAS
How well did we do
How do we do it?
the tasks?
Fig. 2.3 Designmethodology involves creation, abstraction and evaluation: structure and guidelines
A much better way to design something is to follow a systematic design process.

There are many ways to organize this process and one such way is shown in the
Fig. 2.3. Note the number of cycles or iterative loops in the process. The essential
process is described in the subsection.
2.1.1 Design Methodology: Structure and Guidelines
The list below (see Fig. 2.3) is a very abbreviated/bowlderized version of Nam Suh’s
Axiomatic Systems Design Approach [1].
1. Know your clients, not only users but all thos e who are affected by your design.
2. Develop a need statement. A precise, abstract, qualitative high level statement
that incorporates the principal need(s) and the principal constraint(s).
3. Develop a function hiera rchy or function struct ure. Each function should be an
action that has to be performed on an object producing a measurable outcome.
The hierarchy is arranged in a tree so that asking “why should this function be
performed?” leads up the hierarchy into the root of the tree. Asking “how this
function is to be performed?” leads down the hierarchy into the “leaves.”
4. Using the outcomes of the function hierarchy, develop specifications that will be
used to measure the degree of success of the design. This is typically done using
a “Quality Function Deployment Technique” or “house of quality” [2].
2.1 TheSystemDesignProcess 37
5. Develop at least three embodiments that can be used to satisfy the functions. The
guidelines for developing embodiments are given below:
a. Separate functions. Don’t have more than one optimization criteria for a
single component.
b. Provide short and direct paths. Don’t make complicated conduits; Don’t make
too many transformations of energy etc.
c. Exploit symmetry. This is the idea of “balance ” and “harmony” in art and
design.
d. Avoid sharp gradients. Don’t make abrupt changes in properties or geometry
and try not to connect dissimilar components together.
e. Minimize interfaces. Pay close attention to transitions or connections (most
designs fail at the connections)
f. Do not over constrain.
g. Minimize information content. More parts, more items to specify the less
elegant is the design
6. For each function, evaluate how well it meets the need and specifications.
Each of the above items requires an extensive discussion which is outside the scope
of this book. In many cases, the specifications will change as we design (since we
will know more about the product than when we started, and people will gradually
realize what they actually wanted) hence this is an iterative process. Most design
books suggest that about 50–80 % of the cost of the product will be set at the time
when we are designing. In other words, it is well worth your time and effort to
go through many “paper” iterations before actually building the product. This is a
hard thing to do, since it appears that we have made no progress, but in reality, it
will considerably shorten the total design time since it will eliminate many costly
mistakes or misunderstandings
It is clear that if we could predict the way your product or process is going to
behave before you build it then you wouldn’t have to actually build the product to see
if it works. You could save time, money, and even the lives of many people, if you
could mathematically simulate the process or product under a variety of conditions
and test its performance virtually. Use science and mathematics.
One can predict the behavior and evaluate its performance, only if we have scien-
tific (reproducible) explanations for its behavior and mathematical quantification of
its performance. For example, if one wants to predict whether a bridge or a building
is strong, one needs a scientific explanation of how bridges or buildings fail and a
quantification of how much load this particular bridge or building has to withstand.
Hence the essence of engineering is quantitative decision making, that is, measure
or calculate something and then make decisions on whether it is sufficient or not.
This point is so important that we will restate it.
The purpose of all engineering calculations is to make decisions in other words, for EVAL-
UATING options. Thus, engineering is the art of quantitative decision making,
In your statics and dynamics class, when you solve truss problems and were
asked to find forces in a truss, why did you need to do that calculation? You need
to do it in order to decide whether the truss is safe (i.e., whether it will collapse or
cease to function) or not. In the same manner, EVERY calculation that we make in
engineering has a decision attached to it. This is one of the reasons that engineers are
in such demand: they have been taught the process of quantitative decision making
that is essential for so many disciplines. The purpose of the course on mechanics
of materials is to help you master this capability of decision making in structural
engineering.
2.2 The Five Major Subsystems
In the previous section we introduced the notion of the design of a “system” to satisfy
a societal need. In order to gain a better understanding of such systems, let us consider
an example system such as a car (which is itself a part of a transport system).
Can you think of what societal need a car satisfies? it is a system to transport
by land, a small group of individuals and objects (luggage) safely and at a reason-
able speed between any two chosen locations. This is what a car does. The above
description of what a car does is called a need statement. Notice how it is an abstract
statement (i.e. it does not directly reference what a car looks like (its embodime nt)
or even how it does what it does. It describes what happens to the user when they
useAnother
a car, somewhat
example like
that amight
“before”
giveand
you“after” statement.
some insight into a need statement is that
for a pen. What societal need does a pen satisfy? Before you read further, just pause
and write down what you think is an answer. As a hint, imagine why you wanted
to use a pen and what happens when you finish using a pen. A pen is a device that
records information in a visually retrievable form on a two dimensional surface. You
might have had something like “a pen writes on paper.” What a need statement does
is asks, “give me a physics description of “write” and “paper.” In the need statement
here, we have converted “write” into “record information in a visually retrievable
form” and “paper” as “two dimensional surface.” You might quibble with this but
that is quite acceptable; your definition may be at variance with ours.
The point about a need statement is that it allows for other possible solutions than
just the one that we are currently utilizing. For example, we now have electronic
versions of this, namely the iPad® and other tablet computers that can do the same
thing. Thus a pen, a pencil, and a tablet PC are all different “embodiments” that
satisfy the same need statement in different ways. Seen in this way, a car is a specific
embodiment. A scooter, a bicycle, and (if you are in India) an auto rickshaw are other
embodiments that offer personal land transportation.
In spite of the different embodiments looking completely different, all embodi-
ments share some common features from a scientific point of view. They are all sys-
tems that can transmit, convert store or dissipate some basic physical quantities.
2.2 TheFiveMajorSubsystems 39
Different embodiments differ in the extent to which they carry out this activity.
Specifically, every embodiment can have the following features:
1. Transmit and store momentum . Systems whose primary task is to tra nsmit
momentum are called structures. If a system can also store momentum then it
is said to have inertia. Classical physics tells us that we cannot convert momen-
tum.
2. Transmit, store and convert mass . Systems that are primarily designed to transmit
mass are called

chemical conduits
reactions pipelines,
to other etc.
kinds of Mass
mass andofsystems
one kindthat
cancan
be converted
do this arethrough
called
chemical plants or reactors.
3. Transmit store and convert motion. Systems that are primarily designed to trans-
mit and convert motion are called mechanisms. By motion conversion, we mean
converting rotational to translational motion, etc. springs and solenoids and SMAs
are designed to “store” and convert motion.
4. Transmit, store and convert power (specially mechanical and electrical). Systems
that are primarily devoted to transmit power are called power transmissions or
simply “transmission.” We emphasize that you cannot transmit mechanical power
without simultaneously transmitting force and motion. Power conversion devices
are generally referred to as “transducers.” If a system converts chemical or thermal
power to mechanical power it is called a “heat engine” or simply engine. If
it converts electrical power it is called a “motor.” Power storage devices are
generally referred to as “batteries.”
5. Transmit, store and convert signals/information . Systems that ar e primarily
devoted to transmission, storage and conversion of signals are called informa-
tion systems. If they are also used in conjunction with the other systems they are
referred to as “control systems.”
How does this classification come into play with a car? Some comments are in
order.
1. The body of the car and its suspension are primarily involved in the transmission
of forces and momentum and so they are the structural system of the car.
2. The engine and the AC system is naturally the power system or transducer in the
car.
3. The fuel and electri cal conduits, the heating and AC conduits and the doors and
windows are the system involved with transmission of mass.
4. The sensors, speedometer, indicators, steering wheel, brake1 and so on, form the
information system in the car.
1
You might be thinking, “wait, the brake transmits forces. If you think about it a little bit, you will
realize that the brake only transmits the braking signal to the hydraulic system that actually supplies
the braking force.” You cannot stop the car with the force generated by your foot.
2.3 How Do We Identify Need and Functionality for SMAs
A SMA actuator is indicated when one needs motion with a large strokes and forces
in a confined area and with a small footprint. SMA is like a solid-state micro scale
hydraulic actuator. Due to its very inefficient energy conversion capabilities (less
than 5 % conversion efficiency) it is not yet competitive with piezoelectric vibrators
and motors for rapid repeatable action. However, due to its large stroke and wire-like
form factor, it has its niche (which is growing rapidly).
NiTi SMA wires being very corrosion resistant and thus find suitable applications
for use in harsh environments or for deployment in situations where maintenance is
not easy. For example, it could be used for various applications in solar plants where
heat is readily available and slow large motion is required. Since, SMAs performance
can be affected by operating temperature, loading condition, loading rate, compo-
nent geometry, material composition and thermomechanical history, it is essential to
design a proper feedback control system if precision motion is required. However,
in many applications, it is sufficient to develop a “two state” or two configuration
system, that is, a simple “on-off” type mechanism such as for an emergency relief
valve. In these cases, we can obtain guaranteed motion by incorporating a “slot” or
a “cam” in the system so that motion of the wire is guided and the “on” and “off”
positions are precisely located.
In general, the following questions need to be asked before deciding on the use
of an SMA actuator:
1. Is the motion needed primarily linear “back and forth ” rather than cyclic
“rotary?” If it is a full rotatory motion, it might be better to use an electric
motor.
2. Is the intended motion fast or slow, i.e., what is the cycle duration ? If the frequency
is in the range of 1–10Hz then it would be reasonable to use an SMA. Above that,
there might not be sufficient cooling time for the SMA to retract and complex
multiplexing or cooling systems may be necessary.
3. Is it a two state system ? SMAs perform very well in such situations where precise
location is required only at the start and the end.
4. Is it going to be used primarily for power conversion (i.e. as a heat engine)? If so,
then other technologies may be more competitive due to their higher efficiency.
SMAs could be useful only in cases where there is limited space.
5. What kind of heat source is available? Electrically heated systems perform best
due to the relative ease of heating, however cooling is always a problem. If direct
heating is not possible, then circulating a heated fluid may be very effective ( e.g.,
an SMA tube with internal heating/cooling fluid flow may be feasible). There have
been designs where an external Nichrome wire was wrapped round the SMA to
heat it by passing current to the Nichrome wire.
6. What is the operating temperature? Are there temperature fluctuations? Typical
SMAs work very well between room temperature and about 150 C. While there
◦
are high temperature SMAs they tend to have brittleness and stability issues and
may be very expensive. Also ambient temperature excursions may cause either
2.3 How Do We Identify Need and Functionality for SMAs 41
premature or delayed actuation and hence it is better if SMAs are in enclosures

that are reasonably well protected unless great precision is not needed.
7. How many cycles should it be able to run for ? SMAs are remarkably tolerant
to corrosive environments but repeated cycling causes “memory loss” ( a phe-
nomenon that is common to laptop batteries and in some sense similar to it) and
eventual degradation. Thus, if a large number of thermomechanical cycles are
needed, it is best to restrict strains to about 2–4 % even though it may be able
to be deformed and recover to nearly 8 % strains. Commercial vendors typica lly
suggest about 4% strain
Once these questions are answered, the feasibility of using a SMA may be determined.
References
1. Website W (2014) Nam suhs axiomatic systems design approach. http://en.wikipedia.org/wiki/

Axiomatic_design
2. H. of Quality Matrix (2014) Company website. http://www.webducate.net/qfd/qfd-hoq-tutorial.
swf
Chapter 3
Manufacturing and Post Treatment
of SMA Components
We would like to briefly review the different manufacturing techniques and post-
treatment techniques employed in the SMA community. This is by no means an
exhaustive discussion but a broad overview of some important items. As designers,
it is important to having an understanding of these details especially as many of these
techniques/treatments
immensely. discussed here influence the performance of the component
We start with reviewing some of the major techniques used in manufacturing NiTi
SMAs and later discuss some key issues about post processing these manufactured
SMAs.
3.1 Different Manufacturing Techniques
Much of the SMA manufacturing and the processing literature is propriety and only a
small part of it is published. A broad classification of many production techniques are
summarized in Fig. 3.1. The two primary manufacturing techniques can be classified
under casting or powder metallurgy techniques.
The casting techniques are popular with large scale production of NiTi alloys that
generally involves melting followed by hot working and final machining [1].Themost
popular commercial casting techniques are using Vacuum induction melting (VIM),
Vacuum Arc remelting (VAR) and Electronic Beam Melting (EBM) techniques.
Several powder metallurgy or additive manufacturing techniques have been pro-
posed for small scale NiTi component production. These techniques are mainly used
by the biomedical community that involves manufacturing of porous NiTi compo-
nents or intricate components due to their expensive initial costs.

44 3 Manufacturing and Post Treatment of SMA Components
NiTi SMA components - Manufacturing
Casting Powder Metallurgy
VAR VIM EBM
Conventional Additive
Process Manufacturing
CS SHS HIP SPS MIM DMD SLS SLM LENS EBM
VAR – Vacuum Arc Remelting SPS – Spark Plasma Sintering

VIM – Vacuum Induction Melting MIM – Metal Injection Molding
EBM – Electron Beam Melting DMD – Direct Metal Deposition
CS – Conventional Sintering
SLS – Selective Laser Sintering
SHS – Self propogating high temper-
ature synthesis SLM – Selective Laser Melting
HIP – Hot Isostatic Pressing LENS – Laser Engineered net shaping
Fig. 3.1 Different manufacturing techniques employed for NiTi production—a broad classification
(figure reproduced from [1] with permission from Elsevier)
3.1.1 Vacuum Induction Melting (VIM) Technique
The VIM process involves melting of metals/alloys in inert gas or vacuum conditions
[2]. This technique is particularly beneficial with metals that show a strong affinity
for oxidation [2]. By using eddy currents produced by electromagnetic induction,
the metallic charge consisting of Ni pellets and Ti rods housed in a graphite crucible
are melted and the electrodynamic forces provide a good homogeneous mixing of
the melt [ 1, 2] (see Fig . 3.2). Due to different melting points of Ni and Ti, the Ni
pellets with lower melting point melt first followed by Ti rods that slowly dissolve
to form a homogeneous melt over time [2]. The use of additional Titanium cladding
elements is to prevent direct contact between Nickel and graphite crucible. The
molten charge is then poured into a “shaped steel bowl with Y 2 O3 coating” [ 2].
Frenzel and co-workers [2] discuss the use of graphite crucibles along with VIM
3.1 DifferentManufacturingTechniques 45
Fig. 3.2 A schematic of a

vacuum induction melting
Titanium
setup for processing of NiTi
Rods
alloys. The use of additional
Titanium cladding elements
is to prevent direct contact
between Nickel and graphite Nickel Pellets
crucible (figure adapted from
[1, 2]) Titanium
Disk cladding
Graphite
Crucible
Insulation
Cooling
Jacket
due to various characteristics like better electrical properties, good resistance against
thermal cracking, good chemical homogeneity, minimal carbon dissolution into melt
along with lower costs and easy handling. However, presence of large amounts of
carbon in NiTi alloys can form TiC which can increase Ni concentration and thus
affecting transformation temperatures [1, 2]. Hence, the carbon content in molten
NiTi charges must be under permissible limits (between 300–700 ppm) so that they
do not affect any shape memory characteristics of fatigue life [ 1, 2]. The use of
low porosity graphite and using crucibles with lower contact area can significantly
reduce extent of carbon contamination [2, 3]. VIM also allows good control on the
resulting chemical composition, allows easy process control and degassing options
of the final alloy [1, 3].
3.1.2 Vacuum Arc Remelting (VAR) Technique
In case of VAR, the procedure is quite similar to VIM except that this process doesn’t
use a crucible but uses either consumable or non-consumable electrodes made of
materials desired to be melted [1]. In the non-consumable version, the raw materials
are sued in a copper mold using a tungsten rod for irradiation by argon arc [ 1].
In the consumable version, the consumable electrodes themselves are made of the
desired raw materials that serves both as heater and raw material which result in
higher yields compared to the non-consumable versions [1]. Again, using an argon
arc, the electrode is heated and the molten contents are collected in cylindrical ingot
[1]. VAR technique can result in pure NiTi alloys if the “melting cycles, material
impurity, ingot homogeneity and oxidation” issues are carefully handled [ 1]. The
density difference between the two melts and their extreme reactivity of the melt
constituents can be a major drawback in VAR and VIM techniques.
3.1.3 Electronic Beam Melting (EBM) Technique
Some conventional casting techniques like electronic beam melting (EBM) have been
employed to produce NiTi since their successful use in refining refractory metals such
as Mo, Ta, Nb, and W with high melting points since the 1950s [ 1, 3, 4]. In such
techniques, the use of high vacuum environments and water cooled copper crucibles
results in almost zero carbon contamination or oxidation depending on oxygen levels
in the initial raw material charges [1, 4]. The Ni and Ti billets are continuously fed
under a electron beam path that melts the charge and the molten mixture drops into
a “copper extractor” that is water cooled externally [1] (see Fig . 3.3).
However, the control of chemical composition under high vacuum conditions is
quite difficult as compared with VIM or VAR techniques and might also result in
component evaporation (especially, Ni which has higher vapor pressure than Ti)
resulting in lower yields [1, 3, 4].
3.1.4 Conventional/Normal Sintering Technique
NiTi alloys have been manufactured using the popular powder metallurgy techniques
either by conventional sintering or combustion sintering [6]. In conventional sintering,
Continous Laser and

Scanning system
Focus Lens &

Laser Beam Scanning System
Shielding Gas
Roller
Powder bed
Final Product
Piston movement
(direction)
Powder delivery system Fabrication sytem

(Piston) (Piston)
Fig. 3.3 Selective laser melting (figure reproduced from [1] with permission from Elsevier) [5]
the “solid state interdiffusion” of the two constituents is responsible for them to
develop intermetallic phases at the interface [6]. The extreme reactivity of the con-
stituents results in formation of melt pool with temperatures reaching 1300 C in an
◦
inert vacuum environment [6]. In case of combustion sintering, the initial compacted
mixture is combusted which results in violent combustion of the constituent powders
to result in NiTi alloy [6]. The reactions generally occurs at the interface of connecting
particles around the eutectic temperature for desired alloy properties [6]. Generally
these techniques are time consuming with issues of dimensional stabilities [6].
3.1.5 Selective Laser Sintering (SLS)
Gureev and co-workers have presented a technique of Selective Laser Sintering (SLS)
of powder compositions [ 7]. The setup uses YAG:Nd laser with a focusing lens
arrangement to melt the powdered mixture in argon environment [7]. This technique
uses the SLS process in cognizance with SHS technique that avoids sudden heating of
individual constituents and thus resulting in low exothermic synthesis of NiTi alloys
[7]. With such techniques, the surface finish and porosity is uncontrollable [7, 8].
3.1.6 Hot Isotactic Pressing (HIP)
HIP is one of the powder metallurgy techniques for NiTi alloy production from
their respective elemental powders [9]. The elemental powders in required propor-
tion are packed in cylindrical gas-tight stainless steel containers subjected to simul-
taneous hydraulic pressures at elevated temperatures [ 9]. In some cases, an inert
atmosphere like Argon gas is employed in the absence of gas-tight containers which
might result in trapped Argon gas bubbles. The use of higher temperatures (generally
joule heating) and pressures causes better diffusion of constituents due to partial melt-
ing of components [9, 10]. A follow up sintering procedure on the resultant product
at reduced pressures can force the trapped Air/Argon bubbles thus resulting in near
spherical pores [11]. HIP technique allows superior control over surface finish, final
geometry, microstructure homogeneity and porosity of the NiTi samples compared
to conventional cintering techniques [9]. However, NiTi2 or Ni 3 Ti precipitates in the
resultant product cannot be avoided with such a technique [1, 11].
3.1.7 Spark Plasma Sintering (SPS)
SPS similar to HIP uses elemental constituent powders of Cu, Zn, and Al in a graphite
mold for manufacturing copper based SMA [ 10]. As shown in Fig. 3.4, the mixture
Fig. 3.4 Selective plasma Compressive

sintering (figure reproduced Force
from [1] with permission Upper
from Elsevier) Electrode
Graphite
Punch
Powder Pulsed DC
power supply
Graphite Die
Vacuum
Chamber
Lower
Electrode
Compressive
Force
is heated by repeated DC pulses between electrodes positioned at the top and bottom
of die punch. The entire process was conducted in Argon atmosphere with relatively
lower mechanical pressures compared to HIP [10].
3.1.8 Selective Laser Melting (SLM)
Shishkovsky et al. [12] looked at using direct SLM of NiTi powder to prepare non-
porous samples for medical applications. Nitinol powder containi ng 99 wt% inter-
metallic NiTi phase was use to prepare samples with 55 % Ni – 45 % Ti samples
[12]. Upon granulometric analysis of particles, the powders were dried upon heating
up to 80 C for 12 h and then cooled in a cooling cham ber [ 12]. The processing was
◦
carried out in a SLM machine that employs continuous wave Ytterbium fiber laser
in a nitrogen controlled environment [12]. A 50 W power was focused on a 70 µm
of the powdered layer and using scan velocit ies of 100–160 mm/s resulted in SLM
samples with 97 % bulk density [12].
In some cases, powdered mixture of individual Ni and Ti powders are milled
together with stainless balls in an argon atmosphere under various rotational speeds
[13, 14]. As shown in Fig. 3.5, a CO 2 laser beam is scanned over rectangular areas.
Using multiple scan and traveling speeds, the width and the length of the scan regions
can be achieved [13, 15].
Continous Laser System
Focus Lens
Laser Beam
Shielding Gas
NiTi SMA
Dire
ct i
NiTi powder bed o n o f m e lt
in g
Fig. 3.5 Selective laser melting (figure adapted from [13])
3.1.9 Metal Injection Molding (MLM)
Injection molding has been a popular manufacturing process in plastic industry that
injects materials into pre-designed molds. Such an injection molding technique has
been used manufacture some “near net shape NiTi products” mostly from powdered
constituents [1]. As shown in Fig. 3.6, MIM consists of four major steps that starts
with binding of NiTi feedstock, followed by injection molding, then debonding and
followed by a sintering process [1]. Such techniques are mainly utilized for cost effec-
tive large scale production of smaller parts for the micro-manufacturing industries
as they allow for high degree of geometric precision [1].
3.2 Post Treatment of SMAs
NiTi microstructure and surface properties in as cast/manufactured conditions using

any of the above conditions is not acceptable to many industries especially the bio-
medical fraternity [1]. Numerous post-treatment options like cold and hot working,
joining, heat treatments, surface treatments are considered to tailor the properties of
NiTi as per requirements. The NiTi phase diagram shown in Fig. 3.7 clearly shows
that temperatures can greatly influence formation of various precipitates and can
greatly influence thermomechanical treatments [1].
Fig. 3.6 Metal injection molding [16] (figure reproduced from [1] with permission from Elsevier)
3.2.1 Machining of SMA Components
Components such as wires, tubes and sheets are commonly processed by cold work-
ing as the ductility of NiTi allows up to 30–50 % of cold work [ 1]. However, studies
have shown that machining of SMAs is quite difficult due to its unique properties
especially superior resistance to deformation. Tool wear is a major concern while
using conventional machining techniques, hence, techniques like laser/plasma cut-
ting, electric discharge machining (EDM), water jet machining and other abrasive
methods like girding and abrasive saws are preferred [1, 18]. Laser cutting techniques
like Nd:YAG laser with CNC controls are preffered due to good precision/control
over heat input, smaller energy density, minimal/no mechanical stresses, insignificant
heat dissipation, no tool wear and good resolutions [1, 18].
3.2.2 Surface Treatment of SMA Components
For many biomedical applications, in order to prevent any Nickel leakage, a pure
titanium dioxide (TiO2 ) surface coating is added to the component surface. This is
achieved by thermal oxidation in low oxygen environments to avoid any Nickel reac-
tion/oxidation [1]. This TiO2 coa ting prevent release of any Nickel ions in actual use.
3.2 PostTreatmentofSMAs 51
(a) 1800
1670o C
1600 Liquid 1455o C
1310o C
o
C 1400
er
tu
ar 1200
e (Ni)
pm TiNi
e 1000
T (β Ti)
i 3
Ni
T
800 )i
N
(α Ti) i 2
T
(
600
0 20 40 60 80 100
Ti cN i [atomic %] Ni
(b) 1400
1310 Co
1300 Liquid
o
C 1200
e o
utr
1100 1118 C
TiNi
ar 1000 984 Co
e
p 900 )i
m N
e i 2 Ti2 Ni + TiNi TiNi + TiNi3
T 800 T
(
700
630 C
o
600
30 35 40 45 50 55 60 65
cN i [atomic %]
Fig. 3.7 a Shows a simplified NiTi phase diagram with the melting points of individual constituents
Nickel (Ni) and Titanium (Ti). b Shows the central portion of Ni-Ti phase diagram that is of
importance to metallurgists dealing with nearly equiatomic NiTi SMA’s (both figures are adapted
from [2, 17])
However, this oxide layer can cause difficulty in joining especially to dissimilar met-
als [1] (to be discussed in Sect. 3.2.4 later). Also, fatigue studies have shown that
as the bulk SMA undergoes large strains (order of 1–4 % depending on the applied
external load), the strain incompatibility between the bulk SMA and outer TiO2 layer
(generally brittle) is quite significant [19]. With repeated thermomechanical cycling,
the surface cracks reaching a critical length could lead to final failure [ 19].
Depending on the application chosen, Nitinol surface finish can vary and the
choice of finish can significantly influence fatigue, corrosion and other properties
[20]. Many Nitinol surface conditions techniques have been looked at and some
important ones are briefly highlighted below [20]:
1. Oxide surface. Oxide surface with different colors are achieved by annealing in
controlled environments that vary from light blue to dark black coatings.
2. Mechanically polishing. Use of abrasive grinding tools to polish surface s for a
aesthetic look [20]. Typically employed to provide a shiny surface on tubes [20].
3. Chemical Etching. Use of acid baths provides improved coating adhesion
but rougher finish compared to pure oxide surfaces [20]. Such techniques are
employed in etching shape set devices after heat treatment process (to be dis-
cussed later in Sect. 3.2.5 [20]).
4. Sandblasted surface. Use of sandblasting apparatus prior to straightening results
in better adhesion properties due to rough finish [20].
5. Electropolishing. Many biomedical devices like filters, stents using such polish-
ing techniques to provide notably smooth oxide surfaces that enhances corrosion
resistance and biocompatibility [20].
6. Coatings. Many polymer coatings like teflon, polyurethane, Polytetrafluo-
roethylene [ 20] (PTFE) are available but significantly affect the superelastic
responses at higher temperatures [20].
7. Platings. Many metals like Copper, Nickel, Silver, Gold, Cobalt, etc. have been
plated with Nitinol [20]. However, at higher strains their performance is question-
able and issues of hydrogen embrittlement exist during plating operation [20].
3.2.3 Annealing and Coldworking of SMA
One way to change the material properties of the SMA including the yield strengths
and the A f temperature is through aging or annealing of the SMA subsequent to the
shape setting operation. In general, annealing will tend to make the A f temperature
approach that of the original ingot A f . For “high Nickel” alloys (55.8 to 56 % Nickel),
annealing at around 300 to 500 C will cause the A f to rise (longer the duration higher
◦
it rises) but will then decrease above 500 C. For low Nickel alloys, annealing above
◦
500 C will cause the A f to rise towards the ingot A f . Once, the material is shape
set annealed, it demonstrate superelastic and shape memory properties [20].
Commercial Nitnol components are typically subjected to 30–40% cold work
reduction during the last drawing steps [20]. Amount of cold work also influences
the ultimate strength of the material. Most commercial SMA components like wires,
tubes are shape set annealed condition.
3.2.4 Joining of SMA to Itself and Other Materials

Like Sta inless Steel
With growing applications of SMAs, the issue of “connecting” SMA components to

itself and other materials is of significant importance for designers and application
developers. Studies have shown that Nitinol is very difficult to join/attach with other
common dissimilar materials with most of the reliable options being mechanical in
nature like crimping, swaging or staking [21]. Researchers over the past two decades
have focused on the use of more traditional/reliable joining techniques like welding,
brazing, soldering or adhesive bonding with SMA materials [22].
Given the ubiquitous use of NiTi among all the commercially available SMAs,
NiTi has been the focus of attention for joining research with itself and other materials.
Joining NiTi to itself using CO 2 laser, Nd:YAG laser, tungsten inert gas (TIG), and
resistance and friction welding under inert atmospheres have been successful with
their unique properties being intact even almost after joining.
However, connecting NiTi to dissimilar materials like Stainless steel (SS) has
been a challenge using traditional joining techniques like welding or brazing. Such
traditional joining techniques have localized effects on the characteristic properties
of SMA due to different mechanical and thermal effects [22].
Laser welding is one of the popular welding techniques as they can provide good
precision/control over heat input and smaller energy density that results in smaller
heat affected zones resulting in minimal distortion and residual stresses [18]. Laser
welding is generally accompanied by extremely high heating and cooling rates and
this disparity in cooling rates also influences the homogenization process of chemical
composition in weld pool that results in non-uniform microstructure and distribution
of elements [23].
The disparity of physical and chemical properties between NiTi and SS makes
welding challenging [24]. A direct fusion welding of Nitinol to stainless steel results
in the formation NiTi, TiFe2 , NiTi2 , TiNi3 and γ -Fe phases with many brittle inter-
metallic phases between Ti and Fe namely TiFe and TiFe 2 [22, 24]. These phases
makes the weld very brittle and often cracks during welding process or shows poor
tensile strength and brittle fracture at the weld interface [ 22, 25]. Prior studies on
welding of Ti alloys with other metals such as Fe, Cu, Ni have shown formation of
many brittle intermetallics that make Ti extremely unforgiving when it come to weld-
ing/brazing as a preferred joining technique [26]. The presence of Ti in NiTi makes
chemically very active and readily react with commonly used metals/alloys [26].
To overcome these issues, researchers have tried different filler materials that
can potentially change the weld chemistry using different alloy additions and this
has been an active area of interest in joining SMA components [ 25]. The common
approach of deciding appropriate filler materials is by studying binary/ternary phase
diagrams of Ni, Ti and filler materials [ 26, 27]. This provides a good indication of
element interactions at elevated temperatures and solubility levels that can provide
important information on the formation of “brittle intermetallic” phases during the
actual welding process [27]. The fact that other common materials don’t exhibit
unique characteristics like shape memory effect and superelasticity makes it very
challenging to find the appropriate filer material. It is important for the weld fusion
zone to have good mechanical strength close to parent NiTi material and good ductil-
ity properties that reduce internal stresses during thermal cycling to replicate shape
memory effects.
The first choice was to use to Ni filler material to suppress formation of “brittle
intermetallic” phases during welding of Nitinol to stainless steel. Li and co-workers
[23] suggested an increase of Ni interlayer thickness can suppress the formation of
intermetallic phases with Ti like TiFe2 and TiCr2 [23]. Study of weld microstructures
contained γ -Fe, B2, B19, TiFe 2 , TiCr2 , TiNi3 and Ti2 Ni phases with dendrite and
needle like structures [23]. An increase in Ni content causes increase in γ -Fe, TiNi3
and B19 phases with suppression of brittle intermetallic phases (like TiFe2 and TiCr2 )
and reduction of gas pores and shrinkage cavities [ 23]. Generally the concentration
of Fe and Cr decreases from the stainless steel side to NiTi side and the concentration
of Ni and Ti increased correspondingly as expected [23]. The increase in Ni content
at the weld zone results in superior tensile strength of the joint. An increase in γ -Fe
also results in superior tensile strength and deformation characteristics [23]. Addition
of Ni filler material of the order 30 wt% ensured the weld joint were crack free in
spite of extensive plastic deformation during tensile tests [25]. However, large Ni
concentration (>60 wt%) can have reverse effect of forming more TiNi 3 , TiFe and
TiFe2 brittle intermetallic phases, pores and shrinkage cavities. Hence an appropriate
choice
line forof
theNi is important
welds with the[23,
use of25].
Ni Failures are generally observed at Nitinol fusion
filler material.
Li and co-workers [24] have looked at investigating the effect of Copper filler
material on the chemical composition, microstructure and mechanical properties of
a dissimilar NiTi alloy and SS laser welded joint [ 24]. The choice of Copper is
because it has lower melting point, lower yield strength with improved ductility that
could potentially improve the weld pool and brittleness [24]. Further, the binary alloy
phase diagrams indicate that Cu doesn’t form intermetallic phases with Fe, Cr and
Ni elements. However, Cu tends to form intermetallics with Ti which can be brittle in
nature but are not as brittle as those formed with Fe, Cr and Ni [24]. The increase of
Cu metal thickness results in higher Cu content at the weld interface that renders the
soft Cu solid solutions to be coarser and further results in the suppression of brittle
intermetallic phases like TiFe2 , TiNi3 etc. and improving the tensile strength of the
joints [24]. There is however an upper limit of Cu addition (5.1 %) as too much Cu
content could result in additional Cu-Ti intermetallic phases at the weld interface that
along with existing intermetallic brittle phases like TiFe 2 , TiNi 3 could reduce the
tensile strength of the joints [ 24]. Microstructural investigations also revealed that
the cracks mostly srcinate from the weld interface which is the weak zone due to
presence of brittle intermetallic compound and propagate during tensile testing [24].
Li and co-workers [27] have looked at another filler material possibility of using
Cobalt (Co) by studying binary phase diagrams Fe–Co, Cr–Co, Ni–Co and Ti–Co.
Phase diagrams give an indication of interactions and solubility of Co with Fe, Cr,
Ni, Ti in liquid and solid phases. This information can give an indication of possible
compounds that can results in the HAZ. Ti has good solubility in Co that can result
in superior mechanical properties of the joint. Further, Co has a linear expansion

between NiTi and SS with good ductility (deformation) and corrosion resistance
capabilities. Microstructure investigations of the fusion zone indicates use of Co filler
material results in a ductile fracture (combination of cleavage and dimples) against
a brittle fracture observed in other NiTi SS welding cases. This can be attributed
to the decrease in brittle intermetallic compounds like TiFe 2 , TiCr 2 , etc. with Co
addition. Too much Co addition can result in Co-Ti intermetallics that can reduce
joint strengths.
Though the use of these filler materials have shown improved joint strength and
suppressing the formation of brittle intermetaalics with Ti, the important concern
is the performance of these joints under superelastic and shape memory conditions
[25]. Fatigue life of these joints would also be of grave concern. Since parent NiTi
has excellent biocompatibility and corrosion resistance, it is imperative to understand
the joint performance under these conditions [25]. A more in depth investigation on
these issues and study of the microstructure of the joint (weld metallurgy, phase
identification, and so on) is required and these issues are an open area of research
with the SMA community [25].
3.2.5 Shape Setting of Nitinol
As we have seen in the previous chapters, Nitinol actuators need to be “formed”

or shape set into desired shapes for actuation. While commercial vendors supply
wires, springs, tubes and other simple shapes, sometimes it is necessary to buy
wires/tubes/sheets and form the appropriate shape yourself. Be warned that “shape
setting” of Nitinol is somewhat of a dark art and vendors are secretive about the
process.
3.2.5.1 How to Shap e Set Nitinol? Some Gu idelines
In this section, we describe briefly, how Nitinol can be shape set. The recommenda-
tions are based on our own experience with successes and failures in shape setting of
springs and available information in the literature. The basic idea is relatively simple.
Using a jig or fixture, “lock” the Nitinol into the desired shape at room temperature.
Typical jigs ranges from mandrels for winding SMA springs from wires to pin
and plate forms and other such ideas for the shape setting (see some shape setting
fixture examples in Figs. 3.8 and 3.9). The jig is placed in a furnace whose stabilized
temperature is about 400–500 C for about 30 min, then remove and water or air cool
◦
it. The Nitinol is now shape set or so the theory says. In practice, there are a number
of variables to consider.
Pin and Plate jig for shape setting
Guide Pins
SMA wire
End Lock
Plate profile matching SMA

the shape of SMA to set ‘C’ Clamps
Fig. 3.8 Pin and Plate test rigs to shape set SMA components in desired shapes. Notice the impor-
tance of constraining the wire ends to prevent spring back of “martensitic wire” when introduced
to an oven (figure adapted from [28])
1. Condition of the starting sample of Nitinol Actuator wires (martensitic at room

temperature) are commercially available in “straight annealed" or “as drawn” con-
ditions [29]1 . Further, heat treatment may cause the transformation temperatures
and other characteristics to be significantly affected. So a trial and error method
for shape setting may become necessary (as we found in our experiments).
2. Note on shape setting fixtures As shown in Figs. 3.8 and 3.9, fixtures are designed
to confine SMA wires into desired shapes. It is important to constrain the ends of
1
Straight annealed implies that the wire is preset into a straight length as it finishes the drawing
operation [29]. They are annealed in a tube furnace so that they are straight. Subsequently they
may be cold worked to “adjust” their transformation temperatures typically, cold working raises the
transformation temperature. Studies have shown that any heat treatment after cold-working nullifies
the effects of cold working by restoring shape memory response in SMAs [1]. Thus a good strategy
might be to first shape set at 500 C and then use annealing to obtain the right A f . This is still an
◦
art and there are no clear guidelines that will guarantee good results. On the other hand “as drawn"
wires are directly coiled onto rolls from the final drawing operation [29].
Shape Setting Fixture
Extension Spring
Fig. 3.9 A simple jig used to shape set actuator SMA Nitinol wires into extension springs. Notice
the use of heavy metal clamps to constrain the wire ends to prevent unwinding back to initial state
when placed in oven
wires (especially actuator wires that are martensitic at room temperature) as they
tend to return back to their austenitic state (which could be straight if straight
annealed) when introduced in oven maintained at high temperature. It is also
advised to avoid sharp spools (i.e., avoid sharp gradients) of SMA wires and
restrict the amount of strain to <3% as they lead to pockets of material that do
not transform in actual use (for example hooks in extension springs) [1]. This can
also improve the fatigue life for shape set samples [1].
3. Shape setting temperature and soaking times Generally the temperature range for
shape setting/heat treatment is chosen between 400 and 550 C [29]. The soaking
◦
times in ovens vary between 5 and 30 min. Longer soaking times results in increase
in transformation temperatures (especially A f ) [20]. The literature recommends
steels such as 304, 440, D2, H13, and 17-4PH for the shape setting. Also lock-
ing the wire ends surely is crucial. We have found that the use of heavy metal
clamps for locking the springs works well (see Fig. 3.9). Temperatures and times
involved with the heat treatment/shape setting conditions significantly influence
the final microstructure of the shape set product and its mechanical response,
transformation temperatures, energy dissipation capabilities and plateau strains
[29]. Liu and co-workers [30] have looked at studying the effects of temperatures
and times in shape setting and their results showed the decrease of plateau stresses
with aging time and temperature. For example, Fig. 3.10 shows the surface plots
of variation of upper plateau stresses and ultimate tensile strengths with different
heat treatment temperatures and soaking times [29]. It is evident from these plots
that heat treatment temperature and time have significant effect on the mechanical
response of Nitinol being shape set [29].
The choice of times and temperature vary depending upon the wire diameter
and the mass of the jig being utilized for shape setting. There is unfortunately
no reliable guidance for choosing ideal shape setting treatment protocol to
ensure the shape set product’s reproducible and predictable repeatable mechanical
) )
a a
P P
G(
(G
ht 0.59
g 1.7 th
n g
n
re
t rte 0.51
S S
e 1.5
il
s a
u 0.43
e
n te 25
T 25 al
et 1.3 P 0.35 15 )
ns) i ns
15
r 450
m 450
490 0 ( mi e
p
490
550
0
e (m
ia
ime
550
lt T p Temperat o
Ti m
U emperature ( C) T U ure ( C)
o
Fig. 3.10 Variation of upper plateau stresses and ultimate tensile strengths with different heat
treatment temperatures and soaking times. Graphs adapted from [29]
performance. With appropriate experimentation, the heat treatment tempera-

ture/time is finalized for a given application [29]. For larger fixtures, use of ther-
mocouples in contact with material and fixture is recommended to get a better
understanding of temperature distribution [20].
4. Type of furnace When the Nitinol is shape set, oxide layers may be formed on the
surface. To eliminate this, vacuum furnaces (flushed with inert gases as necessary)
may have to be used especially when cooling. In this regard, water quenching
may be a suitable option. Also care must be taken to ensure uniform heating since
otherwise different sections may get different properties. Smith and Hodgson
[31] have looked at a variety of different furnaces and fixtures/mandrels for heat
treatment purposes that include box and tube furnaces, induction heaters, bath
furnaces—fluidized and salt baths, heated presses and vacuum furnaces [1, 31].
The growing conclusion was the oxidation of NiTi in a non-protected environment
[1, 31].
5. Quenching technique Upon heat treatment in the oven, the entire jig can be either
oil or water quenched. Rapid cooling is preferred to avoid aging effects. Studies
suggest that the SMA properties are relatively unaffected with the quenching
technique followed as long as rapid cooling is ensured.
6. Functional Fatigue One of the more vex ing probl ems with the literature on
shape setting is that too often the shape set wire is not subjected to repeated
cyclic loading. We have found that shape set pieces “forget” their set shape when
subject to thousands of thermomechanical cycles unless they have been properly
done. So a major task is to ensure that the shape setting is done properly and
particularly minimize sharp gradients and large strains during shape setting.
References 59
References
1. Elahinia MH, Hashemi M, Tabesh M, Bhaduri SB (2012) Manufacturing and processing of

niti implants: a review. Prog Mater Sci 57(5):911–946
2. Frenzel J, Zhang Z, Neuking K, Eggeler G (2004) High quality vac uum induction melting of
small quantities of niti shape memory alloys in graphite crucibles. J Alloy Compd 385(1):214–
223
3. Otubo J, Rigo O, Neto CM, Mei P (2006) The effects of vacuum induction melting and electron
beam melting techniques on the purity of niti shape memory alloys. Mater Sci Eng A 438:679–
682
4. Otubo J, Rigo O, Moura Net o C, Kaufman M, Mei P (2003) Scale up of niti shape memo ry
alloy production by ebm. Journal de Physique IV (Proceedings) 112:873–876
5. Gibson I, Rosen DW, Stucker B (2010) Additive manufacturing technologies: rapid prototyping
to direct digital manufacturing. Springer, New York
6. Zhang N, Babayan Khosrovabadi P, Lindenhovius J, Kolster B (1992) Tini shape memory
alloys prepared by normal sintering. Mater Sci Eng A 150(2):263–270
7. Gureev D, Petrov AL, Shishkovsky IV (1999) Formation of intermetallic phases under laser
sintering of powdered shs compositions. In: 6th International conference on industrial lasers
and laser applications’ 98, pp 237–242. International Society for Optics and Photonics
8. Yang Y, Huang Y, Wu W (2008) One-step shaping of niti biomaterial by selective laser melting.
Proc SPIE 6825:68250C–1
9. McNeese MD, Lagoudas DC, Pollock TC (2000) Processing of tini from element al powders
by hot isostatic pressing. Mater Sci Eng A 280(2):334–348
10. Park NJ, Lee SJ, Lee IS, Cho KS, Kim SJ (2004) Manufa cturing of cu-15.0 zn-8 .1 al shape
memory alloy using spark plasma sintering. Mater Sci Forum 449:1109–1112 (Trans Tech
Publ)
11. Lagoudas DC, Vandygriff EL (2002) Processing and characterization of niti porous sma by
elevated pressure sintering. J Intell Mater Syst Struct 13(12):837–850
12. Shishkovsky I, Yadroitsev I, Smurov I (2012) Direct selective laser melting of nitinol powder.
Phys Procedia 39:447–454
13. Kyogoku H, Ramos J, Bourell D (2002) Laser melting of Ti-Ni shape memory alloy. Proceed-
ings SFF symposium, Austin, pp 668–675
14. Bormann T, Schumacher R, Müller B, de Wild M (2012) From powder to complex-shaped niti
structures by selective laser melting. Euro PM2012 Proc 1:193–197
15. Bormann T, Schumacher R, Müller B, Mertmann M, de Wild M (2012) Tailoring selective
laser melting process parameters for niti implants. J Mater Eng Perform 21(12):2519–2524
16. Krone L, Mentz J, Bram M, Buchkremer H-P, Stöver D, Wagner M, Eggeler G, Christ D,
Reese S, Bogdanski D et al (2005) The potential of powder metallurgy for the fabrication of
biomaterials on the basis of nickel-titanium: a case study with a staple showing shape memory
behaviour. Adv Eng Mater 7(7):613–619
17. Duerig T, Pelton A (1994) Ti-Ni shape memory alloys. Materials properties handbook: titanium
alloys, pp 1035–1048
18. Pouquet J, Miranda R, Quintino L, Williams S (2012)Dissim ilar laser welding of niti to stainless
steel. Inter J Adv Manuf Technol 61(1–4):205–212
19. Ramaiah K, Saikrishna C, Ranganath V, Buravalla V, Bhaumik S (2011) Fracture of thermally
activated niti shape memory alloy wires. Mater Sci Eng A 528(16):5502–5510
20. NiTinol (2014) Nitinol university—nitinol devices & components, inc. http://www.nitinol.
com/nitinol-university/nitinol-facts
21. Poncet PP (2000) Nitinol medical device design considerations. Strain 2(4):6
22. Gugel H, Theisen W (2009) Microstructural investigations of laser welded dissimilar nickel-
titanium-steel joints. In: Proceedings 8th European symposium on martensitic transformations,
pp 7–11
23. Li H, Sun D, Cai X, Dong P, Wang W (2012) Laser weldi ng of tini shape memor y alloy and
stainless steel using ni interlayer. Mater Des 39:285–293
24. Li H, Sun D, Gu X, Dong P, Lv Z (2013) Effects of the thick ness of cu filler met al on the
microstructure and properties of laser-welded tini alloy and stainless steel joint. Mater Des
50:342–350
25. Hall PC (2002) Laser welding nitino l to stainless steel. In: SMST-2003: Proceedings of the
international conference on shape memory and superelastic technologies, pp 5–8
26. Masunaga S, Murata K, Nakamur a M, Shiroyama K (1994) Joine d parts of ni-ti alloys with
different metals and joining method therefore. 25 Oct 1994. US Patent 5,358,796
27. Li H, Sun D, Cai X, Dong P, Gu X (2013) Laser welding of tini shap e memory alloy and
stainless steel using co filler metal. Opt Laser Technol 45:453–460
28. Case L, Kreiner Z, Trease B (2004) Shape memory alloy shape traini ng tutorial—a teacher’s
guide to teaching sma shape training. http://www-personal.umich.edu/btrease/share/SMA-
Shape-Training-Tutorial.pdf
29. Morgan N, Broadley M (2004) Taking the art out of smart!-fo rming processes and durabil ity
issues for the application of niti shape memory alloys in medical devices. In: Medical device
materials: proceedings from the materials & processes for medical devices conference 2003,
Anaheim, California, p 247, ASM International, 8–10 Sept 2003
30. Liu X, Wang Y, Yang D, Qi M (2008) The effect of ageing treatment on shape-se tting and
superelasticity of a nitinol stent. Mater Charact 59(4):402–406
31. Smith S, Hodgson D (2004) Shape setting nitinol. In: Medical devi ce materials: proceedings
from the materials & processes for medical devices conference 2003, Anaheim, California,
p 266, American Society for Metals, 8–10 Sept 2003
Chapter 4
Basic SMA Component Geometries
and Responses
SMA components must allow for a large surface area compared to their volume
in order for them to be able to be cooled rapidly and repeated use at reasonable
actuation frequencies. Thus SMAs are commonly used in the form of wires/rods,
springs, tubes or beams under different loading conditions (tension, torsion or bend-
ing) for exploiting their unique characteristics in many practical applications. All
these geometries
a review are governed
of different by their commonly
SMA geometries high surfaceused
to volume ratios.
in different To begin with,
applications and
their complex responses are discussed below.
4.1 SMA Wire Response—Tensile Loading
SMA wires under tension are used widely in civil, aerospace and biomedical appli-
cations. The phase transformations in the superelastic response of SMA results in
non-linear hysteretic response which make them excellent candidates as damping
materials due to their superior energy dissipation capabilities [1].
As discussed previously in Sect. 1.4, ASTM standard F2516-07  2 [2] discusses
a standard tension test methodology on NiTi superelastic materials and explains the
salient features of such superelastic effects. Some important salient features worth
noting in a typical superelastic response (see Fig. 1.7) are listed here.
• Upper plateau strength (UPS) : Stress at 3% strain during loading of the sample.
• Lower plateau strength (LPS) : Stress at 2.5 % strain during unloading of sample
after loading to 6%.
• Residual Elongation (Elr ): The difference between strains at a stress of 7 MPa
during loading and unloading operations.
• Uniform Elongation: Elongation determined at maximum force sustained by
specimen prior to necking or fracture or both.

62 4 Basic SMA Component Geometries and Responses
Fig. 4.1 a The tensile (a) 12

extension–time, b load–time,
and c load–extension 10
responses for a superelastic
response with three loading ) 8
m
internal loops. The nonlinear (m
nature of the material n
io
s
6
response in observed in the n
et
transformation regimes (i.e. x
E 4
during loading and unloading
stages of the response). 2
(Graphs reproduced from [5]
with permission from IOP ©) 0
0 100 200 300 400 500 00
6 700
Time (s)
(b) 450
400
350
300
)
N
( 250
d
a
o 200
L
150
100
50
0
0 100 020 300 400 500 600 700
Time (s)
(c) 450
400
350
300
)
(N 250
d
a
o 200
L
150
100
50
0
0 2 4 6 8 10 12
Extension (mm)
Superelastic responses discussed thus far (as shown in Fig. 1.7 earlier) and the
ASTM standard F2516-07 2 [2] focus purely on the outer loop response. However,
this is not the case in many real world scenarios. If the SMA undergoes intermediate
loading and unloading prior to complete transformation, then it can result in smaller
hysteretic responses referred commonly as “internal loops” which closely mimic the
outer fully transformed loop. A typical SMA wire response is shown in Fig. 4.1.
In applications, understanding SMA component responses under partial or fully
4.1 SMAWireResponse—TensileLoading 63
transformed cases with internal loops is of particular importance as the entire response
might not be utilized always and only a portion of the entire response (internal loop)
might be of significance to designers [ 3, 4]. The area of these smaller hysteretic
internal loops depends on the extent of the loading and unloading levels that the
component is subjected to during service within the transformation region (plateau
region) [4]. A few examples will be discussed in the following sections to highlight
these features and illustrate them better.
Consider a SMA wire subjected to a loading sequence described in Fig. 4.1a.
The three steps observed here correspond to intermediate loading and unload-
ing sequences that mimic partially transformed (internal loops) responses. In this
case, the wire 120mm long was loaded up to 5mm extension ( ∼4.1% strain)
and then unloaded back to 3mm extension that results in the first triangular step
(internal loop 1). Further the sample is loaded up to 7mm extension and unloaded
back to 5 mm extension to observe the second triangular step (internal loop 2). Sim-
ilarly, sample loaded up to 9mm extension and unloading up to 7mm extension
results in the third triangular step (internal loop 3). The final loading up to 11 mm
extension and unloading back to initial state completes the superelastic response.
The corresponding load–time response for the triangular forcing function discussed
above is shown in Fig. 4.1b with the same time stamp. The nonlinear nature of the
material response is clearly visible during the internal loops (stepped regions) and
the unloading part of the response where it is phase transforming between the two
phases.
The “outer” loop corresponds to the stress–strain response when the wire is started
from a purely austenitic state and is fully transformed to a martensitic phase and is
then completely unloaded back to the austenitic phase. The shape of the outer loop
is of great importance in the response of the wire. The “inner loop” is the result of
partially transforming the SMA wire and returning it back to the austenitic state.
The corresponding load–extension and stress–strain plots are shown in Fig. 4.2.
Two important aspects to be noted in this plot is the concept of return point memory
(RPM ), that is, the tendency of SMA to return back to the point of unloading on the
outer loop and the other is residual elongation, i.e. the elongation remaining upon
complete removal of loads (representing some untransformed martensite).
RPM provides important information on the ability of SMA components to return
back to its original unloading point upon completion of a smaller hysteretic loop (par-
tial transformation) and is of particular importance to designers. SMA components
showing good RPM is a desirable feature that indicates minimal residual/irreversible
deformations after repeated complete or partial transformations. Figure 4.2b shows
the RPM aspect of this wire and it is evident that the smaller hysteretic loop does not
completely return back to its srcinal unload ing point which eventually results in a
residual elongation (inability to completely recover strains). Designers must account
for RPM characteristics before using them in actual component design.
Figure 4.3 shows typical response of SMA wires under tension at different tem-
perature regimes. Starting from the bottom and moving upwards, the first curve
represents the detwinning of SMA wires at temperatures below M f followed
by unloading to stress free condition and heating to temperature above A f tocomplete
(a) 1000
900
800
700
)
a 600
P
M
( 500
s
rte 400
S
300
200
100
0
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1
Strain (mm/mm)
(b) 450
400
350
300 Return Point Memory
)
(N 250
d
a
o 200
L
150
100
50
0
0 2 4 6 8 10 12
Residual Elongation
Extension (mm)
Fig. 4.2 Figures shows the corresponding stress–strain plot for data discussed in Fig. 4.1. (Experi-
mental data reproduced from [5] with permission from IOP ©). a Stress–strain plot. b Return point
memory and residual elongation
the shape memory effect. The second curve represents SMA wires responses at tem-
peratures M f < T < A f , that is, closer to As . There is only partial recovery of
strains. The third curve represents a classic superelastic response above Af . The fourth
curve represents temperature dependence of superelastic response. With increase
4.1 SMAWireResponse—TensileLoading 65
T >> A f
T >(Af + 25)
s
s
e
r T > Af
t
S
Mf < T < Af (may be

closer to As )
T < Mf (Loading → Unloading

→ Heating to T >A f
Strain
Fig. 4.3 Typical response of SMA wires under tension at different temperature regimes. Starting
from the bottom—First curve represents the detwinning pf SMA wires at temperatures below M f
followed by unloading to stress free condition and heating to temperature above A f to complete
the shape memory effect. Second curve represents SMA wires responses at temperatures M f <
T < A f , that is, closer to As . Third curve represents a classic superelastic response above A f .
Fourth curve represents temperature dependence of superelastic response. Final curve represents
responses at temperatures far greater than A f (figure adapted from [6])
in temperature, the critical stresses for phase transformation increase causing shift
of response upwards. Generally, shift in plateau stresses is accompanied by stiffer
austenitic moduli. If one were to separate the thermoelastic and dissipative part of
the response at different temperatures above Af , then the purely dissipative part of the
response overlap over each other as they would roughly be the same. This suggests
that the temperature dependence in superelastic response is solely due to the ther-
moelastic effects and dissipative part of the response is “athermal.” The final curve
represents responses at temperatures far greater than A f . One can observe residual
strains at these high temperatures due to plastic deformation of martensitic phase
and at very high stress levels and some residual pockets of austensite untransformed
(more details on this aspect will be discussed in Sect. 5.4 later).
4.2 SMA Wire Response—Torsional Loading
The study of SMA characteristics under torsional loading is of great importance to the
design of SMA wires/tubes/springs since these are primarily loaded in torsion. We
have carried out torsional experiments on SMA wires at room temperature. The proce-
dure to obtain the response of a typical superelastic torsional loading and unloading
of a superelastic SMA wire under isothermal conditions (say, room temperature)
consists of twisting the wire up to 1800 twist (loading cycle) and untwisting back
◦
to initial state of 0 twist (unloading cycle) as shown in Fig. 4.4a.

◦
Fig. 4.4 Figures shows the (a)

twist–time, torque–time, and 1800
torque–twist responses for a
simple loading and 1500
)s
unloading test up to 1800
◦
e
e 1200
maximum twist. The r
g
e
nonlinear nature of the (d 900
material response in ts
i
observed in the w
T 600
transformation regime
(experimental data 300
reproduced from [4] with
permission from Elsevier ©). 0
0 500 1000 1500
a Twist–time plot.
Time (s)
b Torque–time plot.
(b) 35
c Torque–twist plot
30
) 25
m
m
20
(N
e
u 15
q
r
o
T 10
0
0 500 1000 1500
Time (s)
(c) 35
30
) 25
m
m
N 20
(
e
u 15
q
r
o
T
10
0
0 300 600 900 1200 1500 1800
Angle (degrees)
The standard commercial torsional test rigs (microtorsion test rigs) are equipped to
continuously record torque values for the triangular forcing functions programmed in
by the user. The corresponding torque versus time plot is shown on Fig. 4.4b which
clearly exhibits a non-linear behavior especially during the phase transformation
event. The complete response (i.e., torque versus twist) is shown in Fig. 4.4c. The
4.2 SMAWireResponse—TorsionalLoading 67
near perfect pseudoelastic response indicates minimum residual deformation upon

unloading with well defined plateau stresses (UPS and LPS).
SMA components like torque tubes, wires could be subjected to partial twisting–
untwisting scenarios similar to tension loading case discussed earlier. Figure 4.5
shows an example of internal loops during the loading and unloading stages of the
superelastic response. As shown in Fig. 4.5a, a SMA wire is twisted up to 900 ◦
twist (1st loading cycle) and then untwisted back to 400 twist (1st unloading cycle)
◦
that leads to a small hysteretic response (internal loop during the loading cycle).
◦
The wire sample is further loaded up to 1800 twist (2nd loading cycle) and then
unloaded (untwisted) up to 800 twist (2nd unloading cycle) which resembles a
◦
classic superelastic response of flat plateaus and large strains/twists as seen previously
in Fig. 1.7 [4]. The sample is then loaded up to 1300 twist (3rd loading cycle) and
◦
unloaded (untwisted) back to 0 twist (3rd unloading cycle) or the initial state. The
◦
nonlinear nature of the response of the material response for the triangular force
function (see Fig. 4.5a) during the phase transformation event can be clearly observed
in torque–time response as shown in Fig. 4.5b [4]. A much more complex situation is
shown in Fig. 4.6 where complex loading–unloading scenarios can be observed that
result more complex internal loops of smaller hysteretic responses [4].
4.3 SMA Spring Response—Torsional Loading
The primary role of a spring is to store energy. Springs made of conventional mate-
rials like steel, Copper, etc. obey Hooke’s law and demonstrate a linear force deflec-
tion relationship. With large deformations, conventional springs can undergo plastic
deformation that can result in permanent strains. However, SMA springs under tor-
sion have shown the ability to deliver near constant forces over large working strokes
thus enabling them to find various engineering applications. In addition, SMA springs
that dissipative energy are used to perform useful work. Their response to temper-
ature make them ideal candidates for thermal actuators. Compression or extension
SMA springs are the most commonly used in applications.
If a SMA spring is subjected to an axial load P , then the torque applied on any
cross section of the spring wire can be computed using the expression
Dm
 
T = P (4.1)
2
where Dm is the mean diameter of the spring with wire diamete r d . If one assumed
the angle of twist is uniform over the entire length of the active coils and further
ignoring curvature effects, the angle of twist per unit length ( φ ) can be evaluated in
terms of the spring axial displacement as
2δ
φ =
2n
(4.2)
π Dm
Fig. 4.5 Figures show the (a)

torque–twist responses for a
SMA wire that show internal 1500
loops under loading and s)e
unloading stages of the er 1200
g
e
response. The overall d
( 900
superelastic response ts
i
indicates a 1800 maximum
◦
w
T 600
twist (experimental data
300
permission from Elsevier ©).
a Twist–time plot.
0
b Torque–time plot. 0 500 1000 1500 2000 2500 3000
c Torque–twist plot Time (s)
(b) 35
30
25
)
m
m20
(N
e
u 15
q
r
o
T 10
0
0 500 1000 1500 2000 2500 3000
Time (s)
(c) 35
30
25
)
m
m20
(N
e
u 15
rq
o
T 10
0
0 300 600 900 1200 1500 1800
Twist (degrees)
4.3 SMASpringResponse—TorsionalLoading 69
Fig. 4.6 Figures shows the (a)

torque–twist responses for )s 1500
test up to 1800 maximum eer
◦
twist with complex internal g e 1200

d (
loops during both loading ts 900
i
and unloading stages of the w
T
response (experimental data 600
300
permission from Elsevier ©).
a Twist–time plot. 0
0 500 1000 1500 2000 2500 3000 3500 4000
b Torque–time plot.
Time (s)
c Torque–twist plot
(b) 35
30
)
m 25
m
(N 20
e
u
rq 15
o
T
10
0
0 500 1000 1500 2000 2500 3000 3500 4000
Time (s)
(c) 35
30
) 25
m
m
(N
20
e
u
qr 15
o
T
10
0
0 300 600 900 1200 1500 1800
Twist (degrees)
The two results [i.e., Eqs. ( 4.1) and ( 4.2)] are independent of each other with
the former being purely kinetic and the latter being purely geometrical in nature
and are based on elementary strength of materials considerations (i.e., that the cross
sections rotate as rigid planes around the central line of the wire). No constitutive
theory relating the torque to the angle of twist is necessary for these results to be
established.
Figure 4.7 shows experiments performed at different temperatures to mimic
detwinning and superelastic responses. Tests performed at room temperature (298 K)
to simulate shape memory effect and three other temperatures above Af viz. 348, 373,
and 398 K to observe supere lastic behavior of SMA springs.
The force-extension results as obtained from the test setup are shown in Fig. 4.7a
with 75 mm being the maximum axial displacement of the spring. The correspon d-
ing torque–twist plot using Eqs. (4.1) and ( 4.2) respectively. The high temperature
trials (348, 373, and 398 K) demonstrating near perfect pseudo-elasticity are par-
tially transformed cases where the springs are partially transformed to stress induced
martensite (SIM) and the extent of transformation from austenite to SIM is different
in each temperature trial. This is due to the fact that the higher the temperature above
Af , the higher is the critical stress required for transformation and thus lesser is the
transformation from austenite to stress induced martensite (SIM) when compared
against the same maximum deformation (75 mm) of the spring. This is also evident
in Fig. 4.7a where the stiffness increases and the hysteresis area decreases with the
increase in the working temperature above A f . The low temperature trail (298 K)
discussed in Fig. 4.7a shows a residual elongation of 50 mm which can be completely
recovered upon heating to temperatures above As .
Fig. 4.7 NiTi SMA springs (a) 25

tested at temperatures 298, 298K
348, 373, and 398K to 348K
20
mimic detwinning and 373K
superelastic responses at )
398K
different temperatures. The N 15
(
extent of transformation in cer
each high temperature trial is o
F 10
different when compared
against the same maximum
spring elongation (75 mm) or 5
corresponding twist chosen
for each trial (experimental 0
data reproduced from [3] 0 20 40 60 80
with permission from Axial Extension (mm)
Elsevier ©). a Isothermal
force versus axial extension (b) 60
298 K
results as measured in the
50 348 K
test rig. b Force versus axial
373 K
extension of spring
) 398 K
converted to torque versus m 40
angle of twist using m
N( 30
Eqs.( 4.1) and (4.2). e
u
rq
o 20
T
10
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07

Angle of Twist (rad/mm)
References 71
References
1. Rao A, Srinivasa AR (2014) A three-species model for simulatin g torsional response of shape
Mech Solids. doi:10.1177/1081286514545917
2. A. S. F. 2007e2 (2007) Standard test method for tension testing of nickel-titanium superelastic
materials. ASTM International, West Conshohocken
3. Rao A, Srinivasa A (2013) A two speci es thermodynamic preisach model for the torsional
response of shape memory alloy wires and springs under superelastic conditions. Int J Solids
Struct 50(6):887–898
4. Rao A, Ruimi A, Srinivasa AR (2014) Internal loops in superelastic shape memory alloy wires
under torsion-experiments and simulations/predictions. Int J Solids Struct 51(25):4554–4571
6. Stoeckel D (1995) The shape memory effect—phenomenon, alloys and applications. In: Shape
memory alloys for power systems EPR. Nitinol Devices and Components Inc, 47533 Westing-
house Drive Fremont, California 94539, pp 1–13
Chapter 5
Factors Influencing Design of SMA Actuators
Given the complex nonlinear nature of SMA responses, understanding their coupled
thermomechanical responses of different components has been of significant interest
for both researchers and application developers [ 1]. SMA component responses in
different applications can be significantly affected by several factors like material
compositions, component geometries, operating temperature, loading conditions,
loading rates,ofthermomechanical
combination history, processing
several of these parameters conditions
in many cases. etc. and
Since these mostly
materials
demonstrate complex non-linear responses due to phase transformations resulting in
microstructural changes, it is imperative to understand their influence on the overall
mechanical response of SMA components. The factors listed below and their affect
on the mechanical response is taken up in this chapter.
• Geometry factors
– Tube/Solid wire
– Wire diameter
– Spring index
– Number of coils
• Material/Alloy composition
•
Spring shape setting conditions
• Spring response at different operating temperatures
• Loading rate
• Wire training/Hysteresis stabilization
5.1 Geometry Factors
Commercial NiTi SMA wires are available in solid or hollow cross sections with
diameters ranging from 0.1 to 6 mm. When a SMA wire is twisted under torsional

74 5 FactorsInfluencingDesignofSMAActuators
Shear Stress (τ )
Shear Stress Variation
rm
ro ri
Radius (r)
Austenitic Core (r < ri )
Phase Transition Region (ri < r < r m )
Transformed Martensite (rm < r < r o)
Fig. 5.1 Nonlinear shear distribution across the wire cross section under torsional loading with an
inner austenitic core, a phase transition region and an outer transformed martensitic layer (figure
reproduced from [4] with permission from SAGE ©)
loading, the phase transformation front gradually moves from the outer fibers towards
the neutral axis. The shear strain tends to zero at the core of the specimen cross-section
as the wire twists [2]. This implies that the possibility of having a fully transformed
case can only be possible if the angle of twist asymptotically reaches infinity [ 3].
Figure 5.1 is typical partially transformed case seen in wires/springs under torsion.
One would observe three zones—an inner austenitic core that remains untransformed
due to really small shear strains at the core; a transition region that is undergoing
phase transformation from austenite to martensite; a outer martensitic layer that is
already transformed from austenite to martensite. The shear stress distribution is
highly non-linear across the wire cross section as shown in Fig. 5.1.
The geometry of the wire (solid versus hollow) would play an important role
affecting the overall mechanical response. If one compares the response of a hollow
and a solid wire with “equal polar moment of inertia” to the same maximum torque,
the hollow wire can undergo complete transformation as an untransformed austenitic
core (due to really small shear strains at the core) is no longer an issue at higher torque
levels. However this is not the case in solid wires which might have an untransformed
core at higher torque levels leading to partially transformed cases.
The diameter of the wire also influences the mechanical response. Especially in
case of springs, where spring index (ratio of mean coil diameter to wire diameter)
can affect the mechanical response greatly. Figure 5.2 shows the effect of Spring
Index on the response of SMA springs. Figure to the left shows the response of
springs with different spring indexes in its austenitic state (i.e., above A f ). Figure
to the right shows detwinning response of springs with same spring indexes (i.e.,
above M f ). Springs with lower spring index show a stiffer response compared to
those with higher spring index as expected. The effect of spring is quite evident from
these responses and they play an important role in spring design especially as the
5.1Geometry
Factors 75
Spring Index 6.2 Spring Index 7.6 Spring Index 10.2
1.5 1.5
1.2 1.2
) )
N N
( 0.9 ( 0.9
e e
c c
r 0.6 r 0.6
o o
F F
0.3 0.3
0 0
3 6 9 12 15 30 45 60
Deflection (mm) Deflection (mm)
Fig. 5.2 Effect of Spring Index on the respons e of SMA springs. Figure to the left shows the
response of springs with different spring indexes in its austenitic state (i.e., above A f ). Figure to the
right shows detwinning response of springs with same spring indexes (i.e., aboveM f ). Experimental
data reproduced from [5] with permission from IOP Publishing ©
commercial availability of SMA wire in different diameters is quite limited. Hence, a

wide range of spring indexes are chosen in SMA spring design (discussed in greater
detail in Sect. 7.4 later).
5.2 Effect of Alloy Composition
The most widely used SMA is NiTi due to its superior mechanical properties, cor-
rosion resistance and biocompatibility. The commonly used equiatomic NiTi alloy
◦
is martensitic at room temperature with an A f of around 120 C [6, 7]. The Nickel
content in NiTi SMA’s plays a very significant role in determining the characteristic
transformation temperatures and slight fluctuations can result in large changes [ 8].
An increase in atomic percentage of Ni above the equiatomic composition results
in a decrease of transformation temperature [ 6, 7, 9]. Studies have shown that the
changes in transformation temperatures with variation in Ni composition is quite
◦
sensitive with as much as 10 C decrease with increase of 0.1 atomic percent from
its equiatomic composition [ 7, 9]. In addition, the presence of elements like Oxy-
gen, Carbon and Nitrogen can affect hysteresis loops, strength and transformation
temperatures [8]. Generally decrease of Ni percentage below the equiatomic compo-
sition doesn’t affect the transformation temperatures significantly [6]. Prior knowl-
edge of alloy composition thus becomes an important parameter for designers as
the transformation temperatures play an important role while designing SMA actu-
ators. Engineers designing thermal actuators are mostly concerned with M f and A f
transformation temperatures for design purposes.
5.3 Effect of Shape Setting Conditions for Custom Shape

(Making SMA Springs)
As discussed in Sect. 3.2.5, SMAs can be shape set into particular shapes by
certain heat treatment conditions. Temperatures and times involved with the heat
treatment/shape setting conditions significantly influence the final microstructure
of the shape set product and its mechanical response, transformation temperatures,
energy dissipation capabilities and plateau strains [9]. Hence, the initial condition
and thermomechanical treatments of SMA components influence the design process
and must be accounted for.
5.4 Effect of Operating Temperature on Mechanical

Response
SMA components are also sensitive to the working temperature at which they operate.
With increasing temperature the plateau stresses keep increasing as seen in Fig. 5.3.
This is quite contrary to classical materials like steel, aluminum where one would
expect a decrease in yield stress as the dislocations become much more mobile at
higher temperatures. However in case of SMAs, the critical stresses for forward and
backward transformations increase linearly with increasing temperature. This can
be explained by Clausius–Clapeyron equation [see Eq. (5.1)] where σ A M are the
↔
critical stresses during loading cycle, θ being the absolute temperature, ∆η being the
entropy change, ∆ H being the enthalpy change, υo is the molar volume and ma x
being the transfor mation strain which is roughly around 6–8 % for NiTi alloys [10]
(see Fig. 5.3).
∆σ A M
↔
∆η −∆ H
= = (5.1)
∆θ υo ma x o θ
It must be highlighted that there is a limited range of temperatures above Af

where superelasticity can be observed. This is related to the thermodynamic sta-
bility of the two phases under different temperatures and the ease of conventional
dislocation glide at higher loads and temperatures. As shown in Fig. 5.4, superelastic
effects occur above A f and the plateau stresses must be less than stresses required
to induce slip. Hence superelastic effects are observed only under limited range of
temperatures above A f (highlighted triangular region in the figure) [12]. This can be
clearly observed in discussion we had earlier in Sect. 4.1, Fig. 4.3, where at higher
temperatures above A f the critical stress for transformation exceeds the stress levels
inducing slip and thus the resultant remnant plastic strain (irreversible permanent
changes) at the end the loading–unloading cycle.
5.4 Effect of Operating Temperature on Mechanical Response 77
Fig. 5.3 Figure shows the

effect of operating
temperature on the upper 1000 348 K
plateau stresses in SMAs. 373 K
Data is for an SMA wire 398 K
) 800
under tension loading a
conditions [11] P
(M 600
se s
rt
S
400
200
2 4 6 8
Strain %
Critical Stress for slip

s
s
e
r
t
S
Superelasticity region
Critical Stress for martensitic

transformation
Mf Ms As Af Temperature
Fig. 5.4 Superelastic effects occur above A f and the plateau stresses must be less than stresses
required to induce slip. Hence, superelastic effects are observed only under limited range of tem-
peratures above A f , as highlighted by the triangular region (figure adapted from [ 12])
5.5 Effect of Loading Rates
At lower strain rates, the plateau stresses are relatively flat as there is larger time for
martensitic bands to nucleate and propagate during the phase transformation event at
lower loading rates [13, 14]. Further, the latent heat at different strain rates induces
macroscopic and localized temperature changes with a significant effect of plateau
stresses for phase transformation. Figure 5.5 shows the effect of loading rates (order
of magnitude effect) on the plateau stresses, overall energy dissipation (hysteresis)
and amount of plastic strain accumulation (over cycles).
At faster loading rates, the phase transformation changes from a diffusionless or
shear-like mechanism to a more classical plasticity-like dislocation slip [13, 14]. This
means an increase in plateau stresses with an increase in loading rate, and the response
looks similar to a hardening response as observed with other classical materials. At
0.8 ̇ = 10−4 /s
1.4
̇ = 10−4 /s
̇ = 10−3 /s
) ̇ = 10−3 /s
̇ = 10−2 /s
% ̇ = 10−2 /s
(
p

ni
a
r 0.7
0.4 t
S
ci
t
s
al
P
0 4 8 0 25 50
Strain (%) Cycles
Fig. 5.5 Effect of loading rate on plateau stresses and accumulation of plastic strains over cycles
with increase in strain rates. Experimental data adapted from [13]
higher strain rates, more martensitic bands nucleate and propagate simultaneously
leading to higher thermal energies and plastic strains due to mismatch of these bands
as seen in Fig. 5.5 [13, 14]. Also, studies have shown that with an increase in number
of cycles at higher loading rates, the net dissipated energy (hysteresis) significantly
decreases [ 13, 15]. The plateau stresses and dissipation energy show a stronger
sensitivity towards higher temperature than higher strain rates [15].
5.6 Wire Training/Hysteresis Stabilization
The martensitic phase transformations seen in SMAs result in a hysteretic loading

and unloading response. The hysteresis is not a result of dislocation glide seen in
plastic deformation of classical materials but is due to the twinning deformation of
martensite that occurs as a result of nucleation, rotation, and shrinking of individual
martensite variants (24 of them) [16]. All of these variants transform back to a single
parent austenite phase upon reverse transformation. During repeated mechanical or
thermal cycling of SMAs, one can see changes in macroscopic material response
due to these changes in microstructures [6]. This is generally observed as rem-
nant plastic strain (irreversible permanent changes) at the end of each loading and
unloading that can accumulate over several loading cycles (also referred to as ratch-
eting response in plasticity literature 1 ) and also in “non-closure of hysteresis loops”
[6, 17]. The material response is considered stable if the hysteresis for several subse-
quent load cycles are stable (i.e., show no drift) as shown in Fig. 5.6. This “stability”
1
In classical plasticity literature, ratcheting refers to failure to close/stabilize a hysteretic loop with
measured strains creeping (or ratcheting) to the right with every cycle. If one were to plot maximum
strain values against the number of cycles, it would represent an increasing trend. This is generally
due to accumulation of plastic strains as the dislocation density increases.
5.6 WireTraining/HysteresisStabilization 79
Fig. 5.6 A typical

stabilization process Unloading Level
0.5
observed in superelastic
responses of SMAs. One can Stabilized response
clearly observe accumulation (no more drift)
) 0.4
of plastic strains over initial a
P
cycles and reaches a G
(
“stabilized state” where there s 0.3 #1
s #5
e
is no more drift in hysteresis r
t # 25
and accumulation of plastic S
strains. Notice the same 0.2 # 30

unloading level for each
loading cycle [18] 0.1
1 234 5 6
Strain (%)
in response is critical from an application standpoint and deliberate introduction

of defects/dislocations during the thermomechanical processing helps in stabilized
superelastic responses [ 17]. Stabilization can be affected by various factors such
as processing history, effect of loading conditions, and number of service/loading
cycles [17]. Studies have shown that stability is achieved if a material is repeatedly
loaded to same maximum load levels. This is due to inherent history dependency of
the material response, that is, the crytallographic changes from austenite to different
martensite variants over cycles and then eventually “stabilizing” [17] (see Fig . 5.6).
If one were to change the loading level (say use higher loads), then additional cycles
are required to achieve stabilized hysteretic responses at these load levels.
Training of SMA component (typically in actuator components that are marten-
sitic at room temperature) refers to repeated loading and unloading in a well defined
thermomechanical cycle that results in saturated hysteretic response [6]. Generally
SMA wires or springs are subjected to known thermal cycles under constant dead
loads that allows the material to stabilize and saturate the amount of permanent
strains [6]. This process of SMA components undergoing repeated thermomechancial
cycling in often termed as two-way shape memory effect (TWSME) [6]. This stabi-
lized hysteretic response is often used to design actuators to avoid any drift during its
operation and is often interpreted as “imperfect memory” requiring multiple cycles

to become “perfect” [17].
References
1. Shaw JA, Kyriakides S (1995) Thermomechanical aspects of niti. J Mech Phys Solids
43(8):1243–1281
2. Aguiar R, Savi M, Pacheco P (2010) Experimental and numerical investigations of shape
memory alloy helical springs. Smart Mater Struct 19:025008
3. Doaré O, Sbarra A, Touzé C, Moussa M, Moumni Z (2012) Experimental analysis of the quasi-
static and dynamic torsional behaviour of shape memory alloys. Int J Solids Struct 49(1):32–42
4. Rao A, Srinivasa AR (2014) A three-species model for simulating torsional response of shape
Mech Solids. doi:10.1177/1081286514545917
5. An S-M, Ryu J, Cho M, Cho K-J (2012) Engine ering design framework for a shape memor y
alloy coil spring actuator using a static two-state model. Smart Mater Struct 21(5):055009
6. Lagoudas D (2008) Shape memory alloys: modeling and engineering applications. Springer,
Berlin
7. Buehler WJ, Gilfrich J, WileyR (1963) Effect of low-temperature phase changes on the mechan-
ical properties of alloys near composition tini. J Appl Phys 34(5):1475–1477
8. Elahinia MH, Hashemi M, Tabesh M, Bhaduri SB (2012) Manufacturing and processing of
niti implants: a review. Prog Mater Sci 57(5):911–946
9. Morgan N, Broadley M (2004) Taking the art out of smart!-fo rming processes and durabil ity
issues for the application of niti shape memory alloys in medical devices. In: Medical device
materials: proceedings from the materials and processes for medical devices conference 2003,
ASM International, Anaheim, California, p 247, 8–10 Sept 2003
10. Van Humbeeck J (2003) Damping capacity of thermoelastic martensite in shape memory alloys.
J Alloy Compd 355(1):58–64
12. University of Cambridge (2014) Dissemination of it for the promotion of material s science
(doitpoms). http://www.doitpoms.ac.uk/tlplib/superelasticity/limits.php
13. Kim K, Daly S (2011) Experimental studies of phase transformation in shape memory alloys.
In: Mechanics of time-dependent materials and processes in conventional and multifunctional
materials, vol 3, Springer, New York, pp 81–87
14. Pieczyska E, Gadaj S, Nowacki W, Hoshio K, Makino Y, Tobushi H (2005) Characteristics of
energy storage and dissipation in tini shape memory alloy. Sci Technol Adv Mater 6(8):889–894
15. Nemat-Nasser S, Guo W-G (2006) Superelastic and cyclic response of niti sma at various strain
rates and temperatures. Mech Mater 38(5):463–474
16. Graesser E, Cozzarelli F (1991) Shape-memory alloys as new materials for aseismic isolation.
J Eng Mech 117(11):2590–2608
17. Khandelwal A, Buravalla V (2009) Models for shape memory alloy behavior: an overview of
modeling approaches. Int J Struct Changes Solids 1(1):111–148
18. Eggeler G, Hornbogen E, Yawny A, Heckmann A, Wagner M (2004) Structural and functional
fatigue of niti shape memory alloys. Mater Sci Eng A 378(1):24–33
Chapter 6
Graphical Description of Temperature
Controlled Actuation of SMA Wires
In this chapter, we will demonstrate how the SMA wires and springs are actuated
and the way a given stroke is achieved by means of temperature changes. We wish
to emphasize that the approach presented here is purely graphical and provides a
good insight into how the SMA actuator operates. The aim here is to give a the
reader a “feel” for how the actuation occurs. In subsequent chapters we will discuss
quantitative examples in more detail. Without the understanding provided in this
chapter it will be hard to follow calculations.
6.1 SMA Wire + Bias Spring Arrangement
To illustrate this graphical design approach, let us consider a simple case study using
a combination of SMA wire and bias spring arrangement as shown in Fig. 6.1. In
order to design such a setup, it involves finalizing the following important items:
1. SMA wire geometry—wire diameter; wire length; number of wires
2. Bias spring selection
3. Actuation temperature—control system
We will be discussing the selection of SMA wire geometry in greater detail in the
next chapter. In this chapter, we will focus on the effect of temperature on the SMA
response and how can one estimate the strokes graphically.
6.1.1 SMA Wire Selection
Commercial SMA wires are available in specific diameters with limited thermome-
chanical processing and surface finish options due to manufacturing and processing
difficulties. Dynalloy® , one of the leading suppliers of SMA actuator wires, provide
a rough estimate/guideline of the recovery force estimates against different electrical
currents [1]. Dynalloy claims that the mechanical load and the temperature heavily
82 6 Graphical Description of Temperature Controlled Actuation of SMA Wires
Bias Spring
SMA wire (Cold State)
SMA wire Stroke

(Hot State)
Fig. 6.1 Figure shows an arrangement consisting of SMA wire in series with a steel bias sprin g.
SMA wire here acts as an sensor and actuator with changes in temperature. Bias spring resets the
mechanism back to its srcinal position for next cycling
influence the permanent set in the wire and hence influencing the material influence
greatly [1]. Table6.1 provides a reasonable estimate of recovery/pull force for dif-
ferent wire diameters. According to the web site, the heating pull forces were based
on constant “safe stress” of 172MPa and similarly the cooling pull force based on
constant “safe stress” of 70 MPa.
A useful plot of heating pull force versus wire diameter is shown in Fig. 6.2. One
can get a reasonable estimate of either the wire diameter or corresponding pull force
if one ofunder
strains themideal
is fixed based onbut
conditions thewe
endlimit
application. SMA
it to under wires
3–4% can recover
strains 6 or 8 %
due to fatigue
constraints.
Table 6.1 SMA wire selection—wire diameter versus pull force estimate. Data reproduced from [ 1]
Ω
Diameter ( mm ) Resistance
 
m
Heating pull Current estimate for Cooling pull
force ( N) 1s ( mA) force ( N)
0.025 1425 0.09 45 0.04
0.038 890 0.2 55 0.08
0.05 500 0.35 85 0.14
0.076 232 0.78 150 0.31
0.1 126 1.4 200 0.56
0.13 75 2.186 320 0.87
0.15 55 3.14 410 1.26
0.2 29 5.58 660 2.24
0.25 18.5 8.73 1050 3.49
0.31 12.2 12.55 1500 5.02
0.38 8.3 22 2250 8.83
0.51 4.3 34.91 4000 13.96
6.1 SMAWire+BiasSpringArrangement 83
Fig. 6.2 Graph showing pull 40

force (heating cycle) versus
wire diameter. Data plotted
from Table6.1 [1] 30
)
N(
e
c
r 20
o
F
ll
u
P
10
0
0.1 0.2 0.3 0.4 0.5 0.6
Wire diameter (mm)
6.1.2 Operating Temperature of SMA Wire
The operating temperature of SMA wire also plays an important role as SMA com-
ponents tend to show a temperature dependence in their mechanical responses. As
shown in Figs. 1.7 or 5.3, above temperatures of Af , the plateau stresses increase
linearly with temperature. If one were to consider the UPS using the ASTM standa rd
discussed in Fig. 1.7 at different temperatures, a linearized version of the material
response at different temperatures can be constructed. In subsequent sections, the
effect of operating temperature on stroke estimation will shown graphically under
different case studies.
6.2 Graphical Design Approach for Stroke Estimation
A typical SMA wire response in its martensitic or austenitic state showing the detwin-
ning (shape memory effect, that is, at temperatures < M f ) and superelastic responses
(at temperatures > A f ) is shown in top part of Fig. 6.3. Now, using the ASTM stan-
dard F2516-07 [2] discussed earlier in Fig. 1.7 some “salient features” of the super-
elastic response can be extracted for design studies. For a linearized approximation
limited to loading part of the response, one needs the following information
• Austenitic moduli E A that represents the initial slope of the superelastic response
• The critical stress at which the transformation occurs in order to account of flat
plateaus in superelastic responses.
• Martensitic moduli E M that represents the final slope of the superelastic response
equivalent to elastic deformation of detwinned martensite or stress induced marten-
site (SIM).
(a)
)
a
P
G
(
s Superelastic response
s
e
r
t Shape Memory effect
S
0.8
EA EM
UPS
0.4
LPS
DTPS
0 2.5 3
2 4 6 Strain%
(b)
)
a
P
G
(
s
e
r
t
S Shape Memory effect
0.8
EA EM
UPS
0.4
DTPS
0 2.5 3 4
2 6 Strain%
Fig. 6.3 a The upper part of the figure shows SMA wire response at temperatures < Mf showing
the shape memory effect and superelastic responses at temperatures > Af . b The lower part of the
figure shows the linearized approximation of experimental responses (considering only the loading
part) based on ASTM standard F2516-07 [2] DTPS is Detwinning Plateau stress corresponding to
2.5 % Loading. Also refer Fig. 1.7 for more details. Experimental data is taken from [3]
6.2 Graphical Design Approach for Stroke Estimation 85
ASTM standard F2516-07  [2] provides information in measuring all these three
parameters described above.
• E A is the slope of the initial response of the superelastic response starting from
the srcin.
• The critical stress or the Upper plateau strength (UPS) is calculated using the
stress level corresponding to 3 % strain during loading response. These stress levels
serves as the linear approximation for the start of flat plateaus (loading level).
• Martensitic moduli E M is the final slope of the superelastic response (after 6 % flat
plateaus). This is generally not important as the strains are restricted to < (3–4%)
due to poor fatigue life.
Unfortunately, there is no ASTM standard for the detwinning/shape memory effect

(SME) responses. We propose using stress levels corresponding to 2.5 % strain during
loading as a linear approximation measure to determine the corresponding martesnitic
plateau as shown in the bottom half of Fig. 6.3. The martensitic moduli is the initial
slope of the detwinning response similar to the superelastic response. Using this
information, the linearized approximate plots for the superelastic (limited to loading
response) and the detwinning response (shape memory effect) are described in the
bottom half of Fig. 6.3.
6.2.1 Graphical Design Approach for Stroke

Estimation—Load and Displacement Controlled Tests
SMA wire can be used under constant displacement or constant load or in a simul-
taneous load–displacement case. Figure 6.4 discusses linearized plots for a constant
load or constant displacement situation. The linearized approximation here represents
only the loading cycle (single switch event between austenite and martensite) and
not the entire hysteretic response that typically includes both loading and unloading
aspects.
The top half of Fig. 6.4 shows SMA wire loading responses (linearized approx-
imations) at three different temperatures above Af and a detwinning response at
temperature below Mf . If a SMA wire under constant external load ( F ) is heated
from its martensitic state, the SMA wire contracts back to its austenitic state and
delivering a large stroke. The net stroke under the same external load can be varied
by changing the temperature up to which the wire is heated. The circular points on
the responses between the cold martensitic state and hot austenitic state illustrates
this point of variable stroke effects with changes in operating temperature. The net
stroke (strain) is the difference between the cold and high states (temperatures),
∆ =|  c − h | (6.1)
s
s
Load Controlled
e
r
t
S
/
e
c
r
o
F
θ = Af + 60
θ = Af + 40
0
2
θ = Af +20 −
f
M
Dead Weight ← Heating θ =
(F )
→ Cooling
Extension/Strain
Stroke
s
s
e
r Displacement/StrainControlled
t
S
/
e
c
r
o
F
θ A
= f + 60
θ = Af + 40
M
F
θ = Af + 20
>
A → ←
F
C H θ = Mf − 20
o e
o a
li t
n
i
n
g g
Constant Extension/Strain
Extension/strain
(M =  A )
Fig. 6.4 Figure shows the linearized version of SMA wire response in its austenite and martensite
state. The red lines depict the SMA wire response at different temperature above Af . The blue line
depicts the martensite detwinning response at temperature below Mf . By changing temperature,
one can achieve large strokes or recovery/pull forces depending on the loading state (i.e., constant
load or displacement state)
Similarly, in a displacement controlled setup, a SMA wire under constant strain/def-

ormation can provide a differential recovery/pull force under different operating
temperatures. As shown in Fig. 6.4, the net recovery (pull) force between the marten-
sitic and austenitic state can be varied by changing the temperature up to which the
wire is heated.
6.2 Graphical Design Approach for Stroke Estimation 87
6.2.2 SMA Wire + Bias Spring: Graphical Design Approach

for Stroke Estimation Using Linearized Loading
Response only
SMA wires are rarely used in isolation and are generally accompanied by other
passive structural members to reset the SMA wire back to its srcinal state when
cooled. For example, a simple arrangement of SMA wire and bias spring combination
is
theshown
systeminto
Fig. 6.1 where the SMA wire must overcome the bias spring stiffness for
actuate.
State A in Fig. 6.5 shows the initial state with SMA wire in its cold state (i.e., at
temperatures < M f ) with the bias spring partially deformed to balance the marten-
sitic SMA wire stiffness. The martensitic SMA wire stiffness can be evaluated using
 
EM A
the relation k = L , where k is the martensitic stiffness, E M is the martensitic
moduli around 30–40 GPa, A is the wire cross section, and L is the length of SMA
wire.
s
s SMA wire response
e
r
t (> A F i.e hot state)
S
/
e
c
r
o
θ = Af + 60
θ = Af + 40 SMA wire response

0
2 (< MF i.e cold state)
−
→
Co θ = Af + 20 f
oli
ng M
=
←
He
a ti θ Bias spring
ng
response
Extension/Strain
D C B
SMA wire Stroke
(Hot State)
A
SMA wire (Cold State) Bias Spring
Fig. 6.5 Figure shows an arrangement consisting of SMA wire and steel bias spring with their
response overlapped on a force/stress—extension/strain plot. The linearized SMA wires responses
at both hot and cold states along with the bias spring response are overlapped for stroke estimation
When the temperature of the system is raised above Af , the system will acquire a
position B where the SMA wire is in its austenitic state causing a recovery pull force
which forces the extension of bias spring. State B can either be C or D depending
on the temperature to which the SMA wire is heated to above Af . Upon cooling
to temperatures below Mf , the system returns back to position B (i.e. srcinal state)
where the extended bias spring overpowers the stiffness of the martensitic SMA wire.
The response of the bias spring is linear (obeying Hooke’s law) and depicted by the
brown slanting line that intersects the SMA wire responses at different points. The
stiffness of the bias spring determines the slope of brown line and the starting point
of the line on the deformation/strain axis. The net stroke is difference between state
A and state B/ C/ D depending on the temperature of the SMA wire above Af . Such a
graphical representation is a convenient way of reasonably estimating the stroke in
such (SMA wire + bias spring) arrangements.
6.3 Case Study 2: Linear to Rotary Arrangement Using a

SMA Wire + Bias Spring Arrangement Using Linearized
Loading Response Only
Using the information discussed in the above sections, a simple linear to rotary
mechanism using SMA components is discussed in this section. The goal here is to
rotate the link by 90 ◦ to demonstrate a simple linear to rotary motion.
This design would involve identifying the following steps:
1. Step 1 (Component Design/Selection): Identifying mechanism(s) for linear to
rotary motion; appropriate SMA wire selection; Bias spring selection
2. Step 2 (Configurati on): Finalize mechanism using choices from step 1
3. Step 3 (Geometry): Fixing geometry—SMA wire details (i.e., wire diameter,
length and number of wires); Bias spring design; Mechanism parts design
Several different linear to rotary mechanisms are available and in this case the
simplest linear–rotary crank mechanism that one can start with is shown in Fig. 6.6.
It consists of a SMA wire and bias spring opposing each other connected to one end
of the link. The other end of the link is connected to the spindle/shaft of the device
◦
that needs rotation (for example, rotating a door spindle by 90 ).
The given data would be the torque required to rotate the spindle/shaft. Lets
consider 160 N mm as the motor torque rating used for this applica tion. This torque
value will be used as a guideline for designing link, SMA wire diameter and bias
spring stiffness. As shown in Table 6.2, different combinations of link lengths and
force (i.e. the torque arm and force) delivering the required torque can be picked.
The net stroke S = r × rotation depends on the initial selection of link. The goal
would be to minimize the stroke for packaging restrictions. The advantage of SMAs
in general is that they can provide relatively large recovery/pull forces. However,
for superior performance and fatigue/degradation considerations, SMA are designed
6.3 Case Study 2: Linear to Rotary Arrangement Using a SMA Wire … 89
Bias Spring
Link
SMA wire Stroke

(Hot State)
o
90
Fig. 6.6 An arrangement consisting of SMA wire(s) and steel bias spring along with the link form a
simple linear to rotary crank mechanism. With changes in temperature, the recovery force generated
by SMA wire causes the link to rotate by 90 ◦ . Bias spring here again resets the mechanism back to
its srcinal position for next cycling
Table 6.2 Choice of link dimensions, force and required stroke for fixing SMA wire lengths
Torque ( N mm ) Link length (r in mm) Force ( N) Stroke mm

( S = r × rotation)
160 2 80 3.14
160 4 40 6.28
160 6 26.67 9.42
160 8 20 12.57
160 10 16 15.71
160 12 13.33 18.85
160 14 11.43 21.99
160 16 10 25.13
with 3 % strains as limitations. Given the packaging and fatigue limitations, values
italicized in Table 6.2 are selected for fixing the link length (torque arm of 6 mm),
force of 27N and a net stroke of 9.5 mm.
Fixing these, one would now have to select a bias spring or a torsion spring for
returning back the (SMA wire + link) arrangement to its low temperature position.
Considering a linear bias spring as the return arrangement, one needs to select the
spring deformation and stiffness for this arrangement. With the torque/recovery force
values and the link dimensions fixed, one can assume different spring deformations
and pick an appropriate spring stiffness. Since the bias spring is a commodity, differ-
ent deformation–stiffness combinations can be selected for the same torque/recovery
force values. In this case, as shown in the italicized values of Table 6.3, the bias spring
Table 6.3 Bias spring design—stiffness estimation

N
Torque ( N mm ) Link length Force ( N) Deformation Stiffness
 
mm
(r in mm) (assumed) mm
60 2 80 2 40
160 4 40 4 10
160 6 26.67 6 4.44
160 8 20 8 2.5
160 10 16 10 1.6
160 12 13.33 10 1.6
160 14 11.43 12 1.11
160 16 10 14 0.82
deformation was fixed at 6 mm and hence the approp riate stiffness is 4.44 N/mm.
Instead of using a linear spring, one can also also select a torsion spring of the
T
required stiffness calculated using 2× π
.
 
Once the stroke and bias spring stiffness are selected an appropriate SMA wire(s)
and length meeting these force–stroke requirements must be picked. Using Fig. 6.7,
depending on the combined recovery pull force and spring stiffness, a suitable wire
diameter is selected. The combined recovery pull force and bias spring stiffness is
around 55 N. For the SMA wire to deliver 55N pull force, one would require a higher
diameter SMA wire (outside the scope of Fig. 6.7). Due to manufacturing difficulties,
SMA wire diameters are available in limited quantity. With higher wire diameters,
one would need higher power to heat the wire to desired temperatures (i.e., above
Af ). Considering these factors, a wire diameter of 0.4mm corresponding to 26N is
selected. This would mean that two wires of this diameter are to be used in parallel
for this mechanism to function.
Fig. 6.7 Graph showing pull 40

force (heating cycle) versus
wire diameter. Dotted black
line represents the force 30
value selected for Linear to Force Value
N
rotary arrangement using a (
SMA wire + bias spring

arrangement. Corresponding 20
wire diameter can be
selected from this plot. Data
plotted from Table 6.1 [1] 10
0
0.1 0.2 0.3 0.4 0.5 0.6
Wire diameter ( mm)
Once the wire diameter is fixed, the length of the SMA wire needs to be determined.
Due to fatigue/degradation considerations, one limits the use of SMA applications
under 3 % strains. Based on the stroke estimates for the required applications, appro-
priate wire lengths are selecte d using 3 % strain as a limiting strain guideline (i.e., a
100mm wire can recover 3mm repeatedly for 10 6 cycles). Using this conservative
3 % recoverable strain estimate, one can estimate length of SMA wire using
Stroke
L=  (6.2)
0.03

This is equivalent to about 350 mm of SMA wire. Conside ring the packaging con-
straints in accommodating long wires, the SMA wires along with the bias spring
can be connected in different ways as shown in Fig. 6.8.1 This completes fixing all
the variables of the system (SMA component and bias springs) using a graphical
technique. Such a technique can be extended to analyze more complex scenarios of
linear to rotary arrangement like a scissor mechanism shown in Fig. 6.9. By connect-
ing the SMA wires opposing each other (antagonistic) and heating them can result
in a rotary motion. Such a mechanism also allows an automatic return back to ini-
tial state due to the antagonistic nature and further allows better packagi ng. This is
also an improvement upon the spooling options discussed in Fig. 6.8 as it can avoid
complications arising due to sharp bends around guide pins that can result in localized
plastic strains
In most and strokeapplications,
commercial degradation external
over thermal cycling.
DC power supplies are employed for
joule heating of the wires. Given the intermetallic nature and metallurgical com-
positions of commercial SMA’s, they generally have high resistivity which suggests
that passing direct current could subsequently internally heat such alloys. Depending
on the power supply rating and the selected wire diameter, the corresponding wire
resistance and required current estimates for resistive heating of these wires can be
1
Figure 6.8 shows two ways of arranging the 350mm SMA wire. Top option looks at using a long
SMA wire that makes the mechanism unusually long and also aesthetically/practically inconvenient
from an application standpoint. Another option shown in bottom half involves bending wires around
guide pins to better package the long wires (referred commonly as “Spooling” technique). Based
on the application and packaging constraints, one can choose either of the two options or use
shorter wires with lever arrangements for stroke amplifications. One can find several mechanisms
for stroke/force amplifications in any “Mechanism design” texts that could potentially be very useful
in designing SMA actuators especially from a packaging perspective (more on this in Chap. 8 later).
However, it must be highlighted that such spooling technique can result in localized pockets of
untransformed regions at the locations where the wires is bent (especially over guide pins). These
sharp bends can result in accumulation of local plastic/residual strains over thermal cycling (phase
transformation event) that can result in stroke degradation over repeated thermal cycling. It is thus
advised to avoid sharp gradients especially during spooling around guide pins to minimize these
pockets of untransformed regions.
Use of a single long wire (350mm in this example)
Bias Spring
Link
Wire Spooling : To accommodate long wires in a more aesthetic/ pratical manner
Link
Guide Pins
Fig. 6.8 Example showing the arrangements consisting of SMA wire(s) and steel bias spring along
with the link form a simple linear to rotary crank mechanism as shown in Fig. 6.6 earlier. Two
possible examples of arranging SMA wires are discussed here
obtained from Table 6.1. The following equations could help estimate current and
power estimates. The corresponding resistance and current values for different wire
diameters are tabulated in Table 6.1.
ρL
Resistance of the wire R = (6.3)
A
where ρ is the resistivity, L is the SMA wire length, and A is the area of cross-
sectional area.
Power Requirement W = I × V = I 2 × R (6.4)
where I is the current, V is the voltage, and R is the resistance. Once the wire geom-
etry (i.e., diameter, length and number) details are fixed, then the wire is repeatedly
cycled for 50–100 cycles to obtain stabilized hysteretic responses of the material.
It must be highlighted that the resistance/current estimates from these tables are
quite conservative and approximate in nature. Studies have shown that the wire resis-
tance changes during repeated cycling due to microscopic changes with repeated
phase transformations [4]. However, due to limited literature on this aspect, one

A Bias
Sp ring
Bias
S pring
B SMA wire (Cold State)
Stroke
SMA wire
(Hot State)
C
SMA wire
D (Hot State)
Fig. 6.9 An antagonistic scissor mechanism with two SMA wires and bias springs attached oppos-
ing each other. By heating the SMA wires, one can achieve rotary motion
might have to perform experimental trial runs using these estimated values to finalize
the required current values. We will show how to estimate current values and heat-
ing/cooling rates for resistive/joule heating of SMA wires of known resistivity and
dimensions later in Sect. 7.7.
6.4 Case Study 3: SMA Wire + Bias Sprin g Arrangement

Using Linearized Loading—Unloading Response
Until now, the stroke estimation was based on the linearized loading response of
SMA wire. This was the simplest approximation that one could consider to estimate
strokes. However, in reality SMA wire responses are much more complex in nature
which results in hysteretic loading–unloading responses due to inherent phase trans-
formations. Top part of Fig. 6.10 shows a typical SMA wire response at temperatures
< Mf depicting the shape memory effect and superelastic responses at temperatures
> Af . The superelastic response is clipped at 6 % strain due to poor fatigue life’s at
large strokes.
(a)
)
a
P
G
(
s
e
r
t Detwinning response
S
0.8
EA EM
UPS
0.4
LPS
DTPS
0 2.5 3
2 4 6 Strain %
(b)
)
a
P
G
(
s
e
r
t
S Detwinning response
0.8
EA EM
UPS
0.4
LPS
DTPS
0 2.5 3
2 4 6 Strain %
Fig. 6.10 a The upper part of the figure shows SMA wire response at temperatures < Mf showing
the shape memory effect and superelastic responses at temperatures > Af . b The lower part of
the figure shows the linearized approximation of experimental responses based on ASTM standard
F2516-07 [2]. Refer Fig. 1.7 for more details. Experimental data from [3]
In order to approximate the complex nonlinear hysteretic response of SMAs, one

can invoke to the ASTM standard F2516-07  [2] to estimate some “salient features”
of the superelastic response that can serve as a premise for linear approximation. In
Sect. 6.2, Upper plateau strength (UPS) was employed to approximate the loading
plateau and the two moduli’s E A and E M . Now, in order to approximate the unloading
part of the response, ASTM standard F2516-07 [2] provides information on Lower
6.4 Case Study 3: SMA Wire + Bias Spring Arrangement … 95
plateau strength (LPS), which corresponds to stress levels at 2.5% strain during
unloading stage of the superelastic response. These stress levels serve as inputs for
the linear approximations of loading and unloading plateaus as shown in the bottom
half of Fig . 6.10.
Unfortunately, there is no ASTM standard for the detwinning/shape memory
effect (SME) responses. As discussed in Sect. 6.2, we propose using stress levels
corresponding to 2.5 % strain during loading as a linear approximation measure to
determine the corresponding martensitic plateau (referred to as DTPS i.e. Detwin-
ning Plateau stress corresponding to 2.5 % Loading). The martensitic moduli is the
initial slope of the detwinning response similar to the superelastic response as shown
in the bottom half of Fig. 6.10. Figure 6.11 uses this salient points of linearized
loading–unloading response discussed above along with the bias spring information.
As discussed in the previous sections, one estimate the stroke by marking the inter-
section points (shown in orange) denoted by points A, B and C. AC would be the
maximum stroke at the high temperature as shown in Fig. 6.11.
s
s
e
SMA wire response
r
t (> AF i.e hot state)
S
/
e
c
r
o
F
θ = Af + 20
UPS → Loading
SMA wire response

LPS ← Unloading (< MF i.e cold state)
θ = Mf − 20 Bias spring
DTPS
response
6% Extension/Strain
C B
SMA wire Stroke
(Hot State)
A
Fig. 6.11 Figure shows linearized SMA loading–unloading responses at temperatures above Af
and ones below Mf . The bias spring response is overlapped to establish the stroke again using this
information
6.5 Case Study 4: SMA Wire + Bias Sprin g Arrangement

Using Complete Hysteretic Loading—Unloading
Response
So far, in the previous three case studies discussed in this chapter, we have discussed
linearized options from the simplest case of loading response only to a linearized
loading–unloading responses. Now, we can look at a case that considers a complete
hysteretic response with the bias spring response as shown in the Fig. 6.12. This case
s
s
e
r
t
S
/
e
c
r
o
F
SMA wire response
(> AF i.e hot state)
θ = Af + 20
→ Loading
SMA wire response

(< MF i.e cold state)
ng
Loadi
→
Bias spring
θ = Mf − 20 response
Extension/Strain
C
B
SMA wire
(Hot State)
A
Fig. 6.12 Figure shows SMA wire responses at two different temperatures along with the bias
spring response overlapped for stroke estimation
6.5 Case Study 4: SMA Wire + Bias Spring Arrangement … 97
is very similar to the example discussed in case study 3 (see Fig. 6.11) but for the
complete response considered here against the linearized version in case study 3.
One can expect more accurate/refined stroke estimates compared to case study 3.
In the next chapter we will show how we use ”stroke or strain limiters” to guarantee
that:
1. the required stroke is obtained and
2. the low temperature stress does not exceed 70 MPa.
References
1. Website DIC (2014) Dynalloy —makers of dyanmic alloys. http://www.dynalloy.com/tech_

data_wire.php
2. 2007e2 ASF (2007) Standard test method for tension testing of nickel-ti tanium superelastic
materials. ASTM International, West Conshohocken, PA
3. Shaw JA, Kyriakides S (1995) Thermomechanical aspects of niti. J Mech Phys Solid 43(8):1243–
1281
Chapter 7
Case Studies in the Preliminary Design
of SMA Actuators
In spite of the rather complex and hysteretic behavior of SMA components (wires
and springs), and the complex changes that occur in their microstructures, the pre-
liminary design of SMA wires and springs turns out to be reasonably simple provided
we utilize their shape memory characteristics in specific ways. Designing SMA actu-
ators requires defining actuation intervals (temperature profile), load levels and/or
displacement outputs over

tates the transformation its intended
temperatures designed
and life. The alloy
the component’s composition
operating dic-
range. SMA
components could be functional under load controlled or displacement controlled
or simultaneous load and displacement controlled setups. Broadly, if we follow the
two key principles explained below, we will be able to design a “two-state” actuator
rather simply.
1. We need to ensure that at the high temperature, the loaded SMA actuator is entirely
austenitic, that is, we do not load it with sufficient load to cause stress induced
martensite (SIM) transformation. Said differently, we need to heat the SMA wire
high enough so that the upper plateau stress is more than the maximum stress that
is applied to the wire. This will ensure that complete transformation occurs during
the heating process. Also this will considerable ease computations. As discussed
in Chap. 6, Dynalloy® [1] one of the leading suppliers of SMA actuator wires
recommends the safe maximum stress for their wires at the high temperature is
172 MPa and at lower temperatures is 70 MPa.
2. We use a “stroke limiter” or “hard stop” for the wire or spring so that at low
temperature, the elongation of the wire is limited by us and is NOT based on the
low temperature SMA response to the load. The hard stop is chosen such that the
maximum strain in the wire (in the cold condition) is about 3–4 % and for springs
the maximum shear strain is limite d to about 1.5 %. Again, this will ensur e that
the SMA does not degrade with cycling and also will considerably simplify the
device.

100 7 Case Studies in the Preliminary Design of SMA Actuators
These two conditions will ensure that the required geometrical dimensions of the
wire are easily computed for a preliminary design and that the wire will last millions
of cycles.
The core elements of decision making when using a SMA wire are as described
next.
1. Find the stroke needed. As a rule of thumb, a typical SMA wire will provide a
stroke of abou t 3–4 mm per 10 cm of wire (st rain of about 3–4 %) for long ter m
reliable performance. So, if you want a stroke of 10mm and you choose to use
an SMA wire directly, you will need approximately 30 cm of wire. You may not
have space to accommodate such a long wire. You have two options :
a. “Winding” the wire aroun d a system of pulleys i.e. the spool ing syste m
discussed in Chap. 6 to package long length of wires or
b. Use a shorter wire but connect it to a “lever” mechanisms or pulley that will
amplify the stroke.
For example, you may choose to use a 10cm wire connected to a 1:3 lever arm
and amplify the stroke. However, as you know by now, “there is no free lunch.” If
you amplify the stroke, you will actually decrease the force exerted by the same
ratio. That is, if the 10cm wire is capable of exerting a force of 10 N, then the
1:3 stroke amplification will cause the force exerted to drop to about 3.33 N. We
will discuss stroke amplification and other matters in the next chapter.
2. Find the diameter of the wire to use. The previous calculation gives the maximum
force that the wire is capable of generating for the required stroke. The force scales
as the cross-sectional area (and hence scales with the square of the diameter of the
wire). Using recommendations from Dynalloy® [1], safe stresses of 172 and 70
MPa can be selected for heating and cooling pull forces. Dynalloy offers different
wire diameters and as discussed in Chap. 6, plots of heating pull force v/s wire
diameters is useful information to start with (see Fig. 6.2 and Table6.1). One can
get a reasonable estimate of either the wire diameter or corresponding pull force
if one of them is fixed based on the end application. SMA wires can recover 6 or
8 % strains under idea l conditions but we limit it to under 3–4 % strains due to
fatigue constraints.
Roughly, a stress of about 170 MPa =24.66 ksi can be sustained by the wire
when it is heated, so that the load that can be sustained by the wire during heating
2 2
is about 133 .5d N or about 19364 d lbs. This will provide a initial estimate
of the wire diameter needed, of the maximum force that needs to be lifted is
known. For example, to lift a weight of 5 lb through a distance of 1  , we will
 
need approximately 40 of wire of diameter 0 .016 . If the wire diameter needed
exceeds the available wires, then we need to design a set of parallel wires to share
the load.
The above calculations is an indication of whether an SMA wire is even feasible to
use. In this chapter, we will go a little bit deeper to demonstrate how to make design
decisions regarding SMA wires and springs and how to estimate their geometry as
well as how much current is needed if we have to electrically heat it (joule heating).
7.1 DifferentModesofOperation 101
(a)
Loading Heating ( >A f ) Cooling
SMA wirespecimen Hot StatePosition

Recovery
Mass (Stroke)
Cold StatePosition
Mass
Mass
Reference Position
(b)
Austenitic
e (Hot response)
c
r
o
F
Martensitic
(Cold response)
Dead Weight( F )
Austenitic
Deflection
Martensitic
Deflection Stroke
Mass
Extension/Deflection
Stroke
Mass Mass
Fig. 7.1 SMA wire/spring in its martensitc state A is deformed to state B under external load. Upon
heating above temperatures A f , the SMA wire/spring contracts back to its austenitic state denoted
by state C or hot state position. Upon cooling, the SMA wire/spring reaches state D. Removal
of external load in state D results in SMA wire/spring returning back to state A in an ideal case.
Complete cycle A D is termed as shape memory effect (SME). a SMA wire-shape memory effect.
↔
b SMA spring-shape memory effect (figure adapted from [2])
7.1 Different Modes of Operation
We begin by considering the common modes of operation of SMA actuators (see

Fig. 7.1).
7.1.1 Constant Force Mode
Here a shape memory wire/spring lifts, pulls or pushes against a constant force and
the length of the spring varies with temperature. Under external loads, the spring in
its martensitic state ( T < M f ) deforms to a temporary shape as shown in state B

of Fig. 7.1b. Once the temperature is raised to ( T > A f ), the wire/spring returns
back to a intermediate configuration between the fully compressed length under no
load (initial configuration of state A ) and the free length under external load in
its marte nsistc state B . During this event, the SMA wire/spring lifts the weigh t
and behave like a thermal actuator as show in state D . Once cooled, the SMA
spring returns back to state D . Notice that the actual stroke obtained depends upon
the difference in the elongations between the cold state and the hot state. In most
practical situations, the Cold state elongation ( B of Fig. 7.1b) is externally limited
by a Hard stop or a stroke limiter.
If the force is applied by a weight W to an martensitic SMA wire/spring in its
high temperature state (i.e., at T > A f ) and the spring temperature is brought down
to below M f , the weight will deform the wire/spring until motion is prevented by
the hard stop and the load is transferred from the wire to the hard stop. For any shape
memory spring the transformation temperatures increase in a linear fashion with the
applied force and the slope of such a line is called the “stress rate.” For Ni–Ti SMA
springs it is of the order of 0.145 ◦ C/MPa. It also follows that the stroke obtained
decreases with an increase in applied force.
7.1.2 Constant Deflection Mode
Here the spring is constrained to a constant length and the force output from the
spring varies with temperature. Consi der a martensitic wire/SMA spring constrained
at constant deformation/length subjected to temperature cycling. With increase in
temperature above A f , the wire/spring tries to return back to its srcinal shape and
results in a recovery pull force due to external constraints. Under constant deflection
mode, the recovery force can be tweaked by changing the operating temperature
above A f . In this configuration, the internal stresses are continuously changing and
thus spreading the transformation over a “range” of temperatures. The recovery pull
force is responsible for it to perform as a smart actuator in different applications.
Constant deflection modes are commonly used in “locking mechanisms” that are
of interest to the bio-medical community (e.g., a three step bone stapling scheme
utilizing a body temperature SMA, called biotinol as shown in Fig. 1.11). These
applications are not related to the use of SMAs as actuators but only as thermally
activated clamps and so we will not discuss them further in this book.
7.1.3 Simultaneous Force-Deflection Mode
In many real world cases, applications typically requires variation in force and dis-
placement simultaneously with temperature. In such cases it is common practice
to design an assembly consisting of an SMA spring working against a steel “bias”
spring. At low temperatures ( < Mf ), the bias spring compresses the SMA spring to
the hard stop (a coil bound state) and as temperature increases (> A f ), the stiffness of
7.1 DifferentModesofOperation 103
the SMA spring increases where it can exerts enough recovery force to compress the
steel spring (bias spring) giving a net “stroke”. SMA springs are generally used with
bias springs where the bias springs overpowers the SMA springs at lower tempera-
tures and the SMA spring take over in their austenitic state, that is, at temperatures
T > Af.
7.2 Design of SMA Wires Under Constant Force
Our first example is to design a SMA wire actuator to carry a load of 100N and
provide a stroke of 10 mm between the hot and cold states. This is an extremely
simple example and serves to illustrate the process. We will assume that the wire is
capable of withstanding 172 MPa at high temperature and that the strain in the wire
cannot exceed 3 % for long life. Further, we will use a “stroke limiter” to ensure that
the strain in the wire does not exceed 3 %. This strain limiter has an added advantage:
we don’t have to worry about the low temperature response of the SMA at all—we
simply have to ensure that the load is high enough to stretch the wire so that it reaches
the stroke limiter, that is, as long as the stress in the wire is above about 100 MPa, the
wire is guaranteed to reach the stroke limiting condition (see Fig. 7.2). Under these
constraints, we consider the steps described next.
1. Choice of number of wires and wire diameter. The wire diameter is chosen based
on the maximum allowable stress of 172 MPa. Then, since the load is 100 N, the
required wire cross sectional area for the wires is
A = σ F = 100
172
= 0.58 mm2 (7.1)
ma x
If we go for a single wire, we will need a wire diameter of d =√ 4 A /π 0.86

=
mm. Large diameter wires might be difficult to procure commercially and would
requires higher currents for heating due to larger resistances. The requirements
of higher currents for heating and associated longer cooling times with larger
diameter wires, we might be better off opting for multiple smaller diameter wires
(for example, 3 wires of diameter 0.45/0.5 mm, which gives a total cross sectional
area of 0.59mm 2 or something similar). After the preliminary design is carried
≈
out, we will have to calculate its heating and cooling rates to ensure that the wire(s)
operate(s) under suitable conditions. The choice of wire diameter and number of
wires is a balance between competing requirements (trade-off). The need for a
simple design implies that we use a single wire, but then the cooling rate will
be very low (as we shall see later) and it will take a longer time for the wire to
complete an actuation cycle (thereby decreasing the actuation frequency for use
in multiple cycles repeatedly).
2. Choice of the wire length. We now have to compute the length of the wire needed.
This dimension is controlled by the required stroke. We compute the length as
External Heating Stroke or Strain Limiter

(Hard Stop)
Load (P)
Fluid
Sliding Guide
SMA wire(s)/ Spring(s)
Stroke Limiter
Load (P)
Fig. 7.2 Stroke limiter system to limit maximum strain on SMA component (wire/spring) and also
to limit the stress on the SMA at low temperatures
follows. First, we calculate the strain in the wire at high temperature (using
a Young’s modulus of E = 75 GPa) as
F 100
H = AE
= (0.59 × 75000)
= 0.2 % (7.2)
Now we will require that the maximum low temperature strain should be equal
to  L
= 4 % so that the strain change due to actua tion is  L  H − 3.8 % This
=
should provide a stroke S = 10 mm. Hence the wire length L f is calculated as
Lf =  −S  = 265mm (7.3)
L H
We note that in order to obtain a 10 mm stroke, we need a wire of length 26.5 cm!
This is a fairly long wire and may be need to be spooled or wound around a set
of pulleys. Its hot length will be L f (1  H )
× + 265.5 mm and its cold length
=
7.2 Design ofSMA Wires UnderConstantForce 105
will be L f× +(1  L ) = 275.6 mm giving the required stroke of about 10mm.

Note that the since the load is constant and the axial stress in the wire is about
171 MPa which is well above the “plateau stress” of the low temperature phase,
we will need a “hard stop” or “stroke limiter” that will stop the stretching of the
wire and placed at a distance of 275.6 mm from the grip of the wire to stop the
wire from stretching further at the low temperature. In other words, our design
will always be such that the wire stress is above the plateau stress for the “soft
martensite,” and we will use a “stroke limiter” to get the required stroke.
We note that in general we will need quite long wires to get a decent stroke in an
SMA. If we want to shorten the wire length, we have two options:
1. Choose a spring configuration or
2. Select some sort of leverage to obtain the required mechanical advantage.
The second option implies that the load will be multiplied while the stroke decreases,
but we may be able to tolerate that in a design.
7.3 Case Study II: Design Procedure for Ti-Ni (SMA)

Springs1
The first step in the design of SMA actuator springs is to define the actuation tem-
perature interval, force and/or motion output expected and the number of oper-
ational cycles required. The transformation temperatures necessary to realize the
required temperature interval of operation usually dictate the alloy composition and
the processing route to be chosen.
7.4 Spring Design Case Study
Unlike the design of the wire, spring design may require several iterations, since the
stresses are a function of the spring index which are in turn dependent upon wire
dimensions.
7.4.1 Design Model and Assumptions

The following approach provides a “quick and dirty” estimation of the SMA design.
The approach is based on the work of Waram [ 3]. Several factors are noteworthy
here:
1. The design is based on replacing the actual stres s response with simple line ar
approximation at the high temperature. This works well for springs since the
1
We thank Dr. S. Chandrasekaran, SMA specialist at IIT Madras and formerly at Qinetiq
Corporation, for this chapter.
hysteresis in the torque versus angle of twist or the shear stress versus shear strain
curve is very small in the range of use.
2. For compression springs, the design is based on using the “coil bound” condition
as a stroke limiter. For extension springs, stroke limiters are to be used to prevent
excessive deformations.
3. In cases where the load varies so that there is not enough force at low temperature
to cause the necessary stroke, a “bias spring” is may be necessary to increase the
low temperature load to a sufficient value to create the needed stroke.
4. The formulae are based on linear elasticity and can be found in spring design
books such as Shigley [ 4].
5. Actual results may vary but this first cut allow s for a preliminary design.
When the preliminary design is complete and the spring is made, it is perhaps a
good practice to obtain plots of a number of isothermal shear stress–shear strain
curves at various temperatures by measuring the force–deflection behavior of a test
spring at several temperatures and then convert to stress–strain data (see Fig. 4.7
discussed earlier). It must be highlighted that this approach is a simple linearized
approximation of SMA spring responses as only the initial linear part of the material
response is considered. At higher force/torque—stroke/twist levels, the response
enters the nonlinear regime i.e. at higher extents of phase transformation regime. For
now, to avoid mathematical complexities, the normal design is to restrict the strains to
values below the onset of nonlinearity. Under such an assumption of linear response,
the equation for shear stress reduces to the standard linear elastic expression used
for conventional spring design,
Shear Stress = τ = π8PD

D3
(7.4)
The linearity assumption greatly simplifies the design process; however, it brings
two important limitations due to this design simplification. They are listed below.
1. The ability to predict the rese tting forces for the design of the bias springs are
somewhat limited as the nonlinear phase transformation region is not captured
2. The above formula is based on isothermal stress-strain plots obtained at con-
stant temperatures. In reality, however, an SMA spring operates under changing
temperature conditions. Since this model ignores any dynamic effects of cooling
and heating the inaccuracies that result from this limitation are more pronounced
at lower temperatures and affect the resetting behavior on cooling. The above
limitation is rectified by
a. first testing the prototype springs and measuri ng the resetting forces under
dynamic cooling conditions, and then
b. manufacturing/selecting the required bias spring to either extend (for ten-
sion springs) or compress (for compression spring) the spring at the low
temperature so that we can obtain the necessary stroke.
7.4 SpringDesignCaseStudy 107
Table 7.1 Symbols and their meaning

Symbol Meaning
δ Deflection
d Wire diameter
n Active coils
c Spring index
OD Springouterdiameter
D Spring average diameter
τ Shear stress
γ Shear strain
τc Corrected shear stress
P Applied force
W Wahl correction factor
K Spring Rate
G Shear modulus
7.4.2 Terms Used in Design of SMA Springs
For a quick reference, we list the symbols (see Table 7.1) and classical design equa-
tions and for springs (directly from spring design methodology [4])
D = OD − d (7.5)
D
c= (7.6)
d
τ = 8πPdD3 = 8πPc
d2
(7.7)
(4c − 1)
W =
(4c − 4)
+ 0.615
c
(7.8)
8W P c
τc = (7.9)
π d2
Gd 4
K = 8n D 3 (7.10)
dδ
γ = (7.11)
π n D2
Armed with these formulae we will now consider an example application which
goes through the full design estimation for a particular application.
7.5 Example: Design of a Remote Controlled Flow Control

Valve Using an SMA Compression Spring
7.5.1 Statement of Require ment
Petroleum crude is extracted from oil bearing sub-soil rock formations by positioning
hollow shafts in the oil-bed and forcing the oil to rise through them using pressurised
steam. In order to regulate the flow of oil from the oil bed into the hollow shaft
as required, a sliding steel sleeve valve is used. This valve has entry holes on its
sides matching the entry holes on the side of the hollow vertical tube shaft over
which it can slide. This assembly is buried well below the earth surface and stands
immersed in an oil bed. The petroleum fluid surrounding the assembly is extracted
out through the hollow shaft under normal circumstances by positioning the sleeve
in such a way that the holes in the annular shaft and on the sliding sleeve are in exact
alignment. However, once the petroleum fluid is exhausted water from surrounding
regions begins to rush in. This needs to be arrested by shifting the position of the
sleeve such that it slides on the hollow shaft and shuts the entry holes completely.
In the specific application taken up for design the sleeve was to be lifted to a height
of 100mm by remote operation. The lifting action should be triggered only above
55 ◦ C and should be achievable within 5 min. It was proposed to design a compression
SMA spring for this purpose. The sleeve (see Table 7.2 for dimensions) would sit on
the spring in such a way that the entry holes are in perfect alignment to allow oil
flow under normal circumstances. When oil bed is exhausted and water ingress is
suspected the SMA spring is operated remotely by sending a signal that will trigger
electrical heating and cause the spring to lift the sleeve by 100 mm shutting the entry
holes.   −  ×
π 132 92
The mass of this sleeve is 4
15
× 7.8 = 8086 g ≈ 8.1 kg (taken
as 10 kg for convenience). The force exerted by the 10 kg mass is 10 kgf ≈ 100 N.
Thus, our design requirements are as follows.
1. required load = 100 N dead load (weight of the sle eve);
2. required stroke (i.e., motion when the wire is hea ted) = 100 mm;
3. available space = 130 mm diameter (since the outer diameter of the sleeve is
130mm).
Table 7.2 Dimensions of sleeve

ODofsleeve 130 mm
Wallthickness 20 mm
Length 150 mm
g
Density 7.8 cc
7.5 Example: Design of a Remote Controlled Flow Control Valve Using an SMA … 109
For the shape memory alloy, the material constraints are that we have only 7 mm
diameter wire (from which to make the spring) and we will restrict the shear strain in
the wire to <1.5 %. In spite of the fact that the SMA is reported to provide strai ns of
8 %, we will almost never use the full transformation since there will be rapid stroke
degradation of the wire as the cycles increase. The SMA taken up for the design
was a Ti–Ni alloy with an As 55 ◦ C. The effective high temperature austenite
=
shear modulus G H of the spring made of this alloy was taken as 20.7 GPa while the
effective shear modulus G L for springs in the low temperature martensitic state was
taken as 2.75 GPa (see page 52 of [ 3]).2
The spring we are about to design is a compression spring. Such springs tend to
have a longer life and further they have the advantage that they can be made to be
“coil bound,” that is, in a state where the coils are touching. Further deflection is
prevented and therefore there is no question of an overload.
Thus, we will design the spring so that when it is in the cold, soft coil bound
condition, the shear strain is 1.5%. and when it is fully actuated, that is, when it is
in the hot, stiff deployed condition, its stress does not exceed 140 MPa. This makes
the design really easy since in the hot condition the response of the spring is purely
elastic so that the spring can be designed like a traditional elastic spring in the hot
condition. Then everything else is actually only a matter of kinematics. For such
springs, we have to consider the elastic response of only the high temperature state
(i.e., the austenite) and ensure that as long as the martensite is soft enough there will
be
weno further
will deformation
proceed since the spring will be coil bound. Given this initial data,
with the design.
1. The spring index and the choice of wire diameter d. For a given value of force
P to be exerted by the SMA spring when it expands with rising temperature, the
first step is to calculate the required diameter d of the SMA wire and the spring
major diameter D . The wire diameter is determined from the maximum allowed
shear stress, which results in the following equation
 8W P c
d = (7.12)
π τc
where P is the force to be exerted by the spring, c is the spring index, and W is
the Wahl correction factor.
We use an iterative scheme for the geometry design as follows: (1) pick a
spring index “c” and choose a suitable working stress τc ; (2) Find a suitable wire
diameter from the equation above; and (3) find the coil mean diameter D using
the wire diameter and the assumed spring index. If it is not satisfactory go back
to (1). For most of the SMA springs it is usual to choose a value for the spring
index c between 6 and 10. However, there seems to be no stringent requirement
that the maximum value should not exceed 10. In the case of SMA wires one
2
In reality, this is the just the slope of the stress strain curve for the twinning deformation and is not
really an elastic modulus.
is sometimes forced to choose higher values of c because of the difficulty in

commercially procuring appropriate wire diameters.
The value of the shear stress τ c is kept relatively low to ensure a low level of
functional fatigue on cycling and it is usually taken to be between 100 and 140
MPa for Ti–Ni SMA. In the present case the memory spring the mean diameter is
taken as 130 mm (to match the OD of the sleeve). Only 7 mm diameter wire was
130−7
available. Thus the spring index c was estimated as c 7
=  =
18.1 18. ≈
Assuming a critical shear stress τc of 140 MPa for austenite SMA, the wire
diameter d for the spring is recalculated with these updated numbers as explained
next.
a. First, calculate the Wahl correction factor,
W =
 (4 × 18 − 1) +   =
0.615
1.078 (7.13)
(4 × 18 − 4) 18
b. Next, estimate the wire diameter,

 8PW c
× 8 100 × 1.078 × 18 ≈ 6.0 mm
d = = (7.14)
π τc π × 140
It turns out to be not very far away from the available diameter d 7 mm of the
SMA wire. So, if we choose a 7 mm diameter wire and a spring inde x of 18, we =
will end up with a coil diameter of 126mm.
2. Calculation of active number of coils n. Determination of the number of active
coils in the spring is carried out through imposing a constraint on the shear
strain γ . Since a linear stress–strain behavior is assumed, at at high temperature
we have γ H τc / G H so that the high temperature shear strain γ H is then
=
calculated as
140
γH = 0.0068 0.68 %
= = (7.15)
20700
For a good cyclic life it is customary to limit the low temperature shear strain
γL = 0.015 = 1.5 % We will prevent further deform ation in spring by ensuring
that this shear strain will cause the spring to become coil bound.
The difference between the low temperature and high temperature shear
strains is
γ = γL − γ H = 0.015 − 0.0068 = 0.0082 (7.16)
This shear strain difference must deliver the required stroke.

7.5 Example: Design of a Remote Controlled Flow Control Valve Using an SMA … 111
We can now calculate he number of active coils n using the expression
n = π(Dd 2×(γ)
S)
(7.17)
where S is the stroke. This gives
(7.0
× 100)
n 1.6 2 coils (7.18)
= π × 126 × 0.0082 ≈ ≈
2
Thus we are looking for a wide shallow spring with only two active coils. Since
it is a compression spring, assuming that the spring has unground, closed ends
the total number of coils N is given by the expression N n 2, so in this case
= +
the number of coils is 2 2 4.
+ =
3. Calculation of total length of coil. The first step is to determine the high temper-
ature spring rates K H using the formula
4 4
= G8nHDd3 = ((820700
KH
×7 )
× 4 × 1263 ) ≈ 0.758 N/mm (7.19)
Thus the high temperature deflection δ H = P / K H = 100/0.758 = 131 mm in

compression, and hence the low temperature deflection is3 δ L δH S 131
= = + =
100 231 mm. Now, the length of the spring L L when it is fully compressed at +
low temperature is given by
LL = d (n + 3) = 7 × (4 + 3) = 49mm (7.20)
It follows that the length of the spring when it is at high temperature is given by
LH = LL S
+ = + 49 100 =149 mm . Finally, the manufactured length of the
spring, also called the free length L F , is
LF = L H + δ H = 149 + 131 = 280mm (7.21)
We finally have to ensure that with the 100 N force, the spring will become coil
bound at the low temperature, For this reason, we make an approximate estimate
of the slope of the martensitic twinning transformation line at low temperature
and obtain the resul t that for the spring to deflect 231 mm.
We first calculate the spring rate at low temperature as
4 4
KL = G8nLDd3 = (8(2750 × 7 ) ≈ 0.1 N/mm
× 4 × 1263 ) (7.22)
3
Note we do not use the spring constant at low temperature because the spring will be “plastic” at
low temperature and we are designing it to be coil bound.
Thus the force required to compress by 231 mm is R F ( K L δ L ) 0.1 231

= × = × =
23.1 N. It can easily be seen that this spring would go coil-bound un der the action
of the constant force of 100N acting on it.
The key point here is that the we cleverly use the “hard stop” provided by the
coil bound condition to no only guarantee that there will not be excessive defor-
mation but also dramatically simplify the calculations of the spring dimensions
based on only the high temperature elastic response.
7.6 Extensional Spring Design
The design of an SMA extensional springs uses exactly the same formulae as were
used in the design of an SMA compression spring for the wire diameter, spring
diameter, and active number of coils , provided we make use of a slot or hard stop
to limit the stretch at low temperature. In other words, we need to design the spring
with a “hard stop.” Unlike the compression spring where this is provided by the coil
bound condition, for tension springs we need to add hard stops as shown in Fig. 7.2.
The only difference in the design is the determination of the various spring lengths
L H , L L and L F .
First, the spring is “shape set” such that it is close coiled (coil-bound) at high
temperature with no force applied (as can be seen in commercial SMA springs). The
length of the spring in this condition is called the body length of the spring.
The body length L B of the close coiled section of the extension spring (i.e.,
excluding hook lengths) is given by L B d (n 1), where d is the wire diameter
= +
and n is the number of active coils. If it is assumed that end hooks are added on
both ends that are each equal in length to the coil inner diameter (ID), then, for the
example above, the free length, L F of the spring is give n by
LF = L B + 2 (ID) = d (n + 1) + 2 (ID) (7.23)

= 7(4 + 1) + 2(127 − 2 × 7) = 261mm (7.24)
Now, the high temperature length L H is equal to the free length L F plus the high
temperature deflection δ H
L L δ 261 131 392mm (7.25)

H = F +
+ = H =
and the low temperature length L L = L H S = 392 + 100 = 492 mm. We need to
+
provide a hard stop at this length so that we get repeatable response and so that the
stroke is appropriate.
Thus, there are three essential features of the spring design:
1. The spring diameter is obtained by limiting the high temperature elastic response
stress to be about 140 MPa with an assumed spring index. This is an iterative
process to obtain a satisfactory spring index and inner and outer diameters.
2. The low temperature strain is fixed by consid erations of life.
7.6 ExtensionalSpringDesign 113
3. The strain difference combined with the required stroke will give the number of
coils.
We will ensure that the low temperature deflection is limited by a hard stop (either
externally for tension springs or with the coil bound condition for tensile springs).
7.7 Heating and Cooling of Shape Memory Wires
In order for the material to be activated, it must be heated and cooled. Of these, the
cooling is the more problematic issue since, while it is easy to heat the wire, cooling
depends upon ambient conditions. In general there are two means of heating an SMA
wire or spring:
1. Internal heating through the passa ge of an electric current or
2. Contact heating through ambient.
Since the resistivity of SMA is rather high (approximately 10 µ mm = × 1
10−6  cm), it is quite simple to heat the wire by passing an electric current, I .
For such a situation, the steady state temperature that can be reached is given by
the balance between the internal heat generation due to joule heating and the heat
transfer to the surroundings (we will ignore the radiative heat transfer, but it would be
important in situations where the wire is quite hot and in situations such as vacuum
where convective effects are small). Based on this, the steady state temperature would
be
2
T = Tam b + π42ρdI3 h (7.26)
where Tam b is the ambient temperature, ρ is the resistivity ( 10 µ) mm, I is the
≈
current and h is the heat transfer coefficient and d is the wire diameter. Of these, the
most uncertain parameter is the heat transfer coefficient and only very approximate
values are available. The value depend upon:
1. Whether the wire is vertical or horizontal or
2. Whether there is free or forced con vection
Correlations and calculations are available for a variety of cases and should be avail-
able in any heat transfer books.

To illustrate the vast difference that the diameter of the wire makes consider the
current required to heat either a 5 mm wire or a 0.5 mm wire. Assume that the length
of the wire is 200 mm, the ambient temperature is 25 ◦ C with the target temperature
to be 90 ◦ C. We can estimate from the “thermal wizard” web site [ 5] that the heat
transfer coefficient is 12.4 W/m2 C. Using this value, the estimated current would be
 (T − Tam b ) × (π 2 dAmps
I = 3h) (7.27)
4ρ
To obtain a final temperature of 90 ◦ C, we will need a current of approximately 14

Amps! One can see that heating a 5 mm diameter wire can be unfeasible. However,
if one were to consider a wire of 0.5 mm diameter, then the heat transfer coefficient
is 42 W/m 2 C and the current required will be
 (T − Tam b ) × (π 2 d 3 h) = 0.92Amps
I = (7.28)
4ρ
If a very specific final temperature is desired, it is better to mount a thermocouple on

the wire and build a feedback control system to control the current.
If the wire is very thick (or for SMA rods under compression) the current required
may become excessive. In these cases, heating through an external hot body becomes
feasible. In this case the small surface area becomes a disadvantage. It is better to
enclose the SMA wire or spring in a tube with a conducting fluid and heat the tube.
This has been shown to be very effective with solar heating where the tube can be
painted black. Of course, in this case estimating the final temperature may be quite
complicated .
7.7.1 Time Taken to Heat Up and Cool Down
While it is a simple matter to estimate the current required to heat up the wire, it is
actually quite different to estimate how long would it take to heat up or cool it down,
this is because of the fact that the wire undergoes a phase transition during heating 4
this is not correct.
A rough estimate of the time taken to heat or cool the wire is given by considering
the total heat supplied by joule heating and that due to convective heat transfer
and the increase in the internal energy due to changing temperature 5 . Thus, a crude
approximation is based on the following equation:
mC p ∆T + m L = [−h(Tavg − Tam b ) + I 2 R ]∆t (7.29)
where m is the mass of the wire and C p is the heat capacity per unit mass (assumed
constant), L is the latent heat per unit mass for the conversion, and Ta vg is the average
temperature at which the heat transfer takes place. In arriving at this simple expression
we caution that we have made a number of simplifying assumptions: (a) We have
ignored the work done by the wire. It is a reflection of the low conversion efficiency of
the SMA (order of 5–10 %) that we can ignore this and still obtain reasonable answers.
4
Beware that many papers routinely ignore this and try to estimate the heat transfer coefficient using
a simplified model. The answers, not surprising show wide divergence and indicate that the heat
transfer coefficient varies during the process. Actually these papers are attributing the temperature
stagnation due to phase transitions to variations in the heat transfer coefficient.
5
This would be inappropriate for situations were repeated rapid cycling takes place as in the case
of a heat engine or a flapping wing system or in vibration damping applications. For these systems
it is important to consider the heating due to hysteresis also.
7.7 Heating and Cooling of Shape Memory Wires 115
Furthermore, we have ignored the hysteresis losses and we have approximated the
heat transfer by assuming that it occurs at the average temperature rather than at the
actual temperature of the wire. This is indeed a serious approximation but this is
sufficient to provide an order of magnitude estimate.
For the 200 mm long 0.5 mm diameter wire for which ∆T 65 ◦ C, its specific
=
heat C p is about 0.832 kJ/kg C, its latent heat L is about 24.2 kJ/kg, its density ρm is
6540 kg/m 3 and the other quantities have already been estimated. Hence, we obtain
∆t =
− 4h ( T
avg −Tam b ) ++L16) I 2 ρ
ρm (C p ∆T (7.30)
d π 2d4
When we compute the time taken using the above expression, we obtain an estimate
of about 46s for heating and cooling. Again, we caution the reader that these are
very crude order of magnitude numbers and are to be used only to establish the
order of magnitude of the time taken in order to establish feasibility. Due to the
myriad of factors that influence the cooling, it is better to verify this with further
experimentation.
References
1. D. I. C. Website (2014) Dynalloy—makers of dynamic alloys. http://www.dynalloy.com/tech_

data_wire.php
2. Ganesh N, Maniprakash S, Chandrasekaran L, Srinivasan S, Srinivasa A (2011) Design and
development of a suntrack ingmecha nism using the direct sma actuation. J Mech Des 133:075001
3. Waram T (1990) Design principles for ni-ti actuators. Butterworth-Heinemann, Engineering
Aspects of Shape Memory Alloys (UK), 1990, pp 234–244
4. Shigley JE, Budynas RG, Mischke CR (2004) Mechanical engineering design. McGraw-Hill,
New York
5. M. Simulation (2014) Website: thermal wizard. http://www.thermal-wizard.com/tmwiz/
convect/natural/horiz-cylinder/horiz-cylinder.htm
Chapter 8
Coupling SMA Actuators with Mechanisms:
Principle of Virtual Work
In Chap. 7, we covered the basics of SMA wire and spring design for dead loading
conditions. We noted that if the stroke required is very large, a simple SMA wire
may be unfeasible.Thus, successful design of a shape memory wire actuator depends
upon the ability to combine it with a mechanism to achieve the desired force and
stroke when actuated. As a rough guide, a SMA wire has characteristics similar to a
hydraulic cylinder and their traits include the following:

1. Large force when compar ed to other componen ts with similar size
2. Somewhat slow action compared to piezoelectric actuators
3. Inability to return to its starting shape by itself if the low temperature force is not
sufficient
A challenge with the direct use of SMA wires is the fact that excessively long
wires may be necessary to obtain the necessary stroke. For example, to directly lift
a weight of 5 lb through a distance of 1  , we will need approximately 40  of wire of
diameter 0.016 .
8.1 The Need for Mechanisms
The main difficulty is with the length of the wires. A 40 long wire may not be feasible
and hence at this point, we need to see if we can use a simple lever mechanism to
deliver the required stroke. Such lever mechanisms can range from simple levers, bell
crank type mechanisms, all the way to various four bar links and other sophisticated
setups. Another reason for going with mechanisms is to convert motion, that is, for
heating and cooling purposes, it may not be feasible to deploy the wire directly where
needed but place it somewhere more convenient and transmit the motion through
a system of linkages. Alternatively, it may be desirable to generate a flapping or

118 8 Coupling SMA Actuators with Mechanisms: Principle of Virtual Work
rotational motion whereas the SMA produces only rectilinear motion. For all these
reasons, it is necessary to consider linkages and mechanisms as part of design.
A critical challenge with the use of linkages is the fact that the force needed at
the actuator location may vary with the configuration of the linkage even though the
force exerted by the external load on the mechanism may be constant. For example,
consider two configurations of the lever. In case 1:– The force at the SMA wire is
equal to the applied load. However, in Case 2, the force exerted by the lever on the
wire is a multiple of the load and it varies with the orientation of the arm.
For a complicated mechanism such as case 2, estimation of the force(s) required
to keep the system at equilibrium is quite complicated if we simply use Euler’s laws
for rigid bodies, since the calculations will produce all the forces in the body and not
just the forces that are needed. However, as an alternative method, dating back to the
work of Archimedes, termed as the principle of virtual displacement is extremely
useful for these calculations.
The principle is stated (without proof) as follows (see [1]):
Consider a system that is in equilibrium and acted upon by several external forces, F iex t . If
we now subject the system to an arbitrarily small allowable virtual displacement δ xi at each
location where the force is applied such that the displacements must be “compatible” with
each other, not causing the system to be torn apart or violating other constraints), then the
net work done by all the forces in moving through their respective virtual displacements is
zero:
n
δW =
 Fiex t · δ xi = 0
i
(8.1)
This seemingly simple result actually hides a major simplification, namely that ONLY
the external forces are included, not the internal forces. In other words, the virtual
work of all the internal forces are zero . This result dramatically simplifies the calcu-
lations of the force relationships of mechanisms and has found wide use in a variety
of fields. The process for using this result in computing forces in mechanisms that
are in equilibrium are as follows:
1. First identify the degrees of freedom of the mechanism How many ways is the
mechanism allowed to move. Equivalently what is the minimum number of actu-
ators needed to keep the mechanism locked in any particular orientation? As an
illustration, consider the lever mechanism shown in Fig. 8.1, with the load on one
side and the actuator on the other. Assuming that the lever arm AB is rigid, there
is only one allowable motion, that is, the rotation of the arm about the fulcrum O.
2. Choose kinematical parameters One for each degree of freedom q α that can be
used (together with the fixed geometrical parameters) that will represent all the
allowable motions of the mechanism. For the lever mechanism, we can choose
the angle made by AB with the horizontal ( q 1 = θ ) as the kinematical parameter.
Notice that, this choice is by no means unique. We might as well have used the
height of A or the height of B as variables. However we cannot use both the
heights as variables since they cannot be independently varied.
This is the key step in generating compatible displacements, since the q α are
chosen represent every allowed configuration of the body, every variation of q α
8.1 TheNeedforMechanisms 119
Fig. 8.1 Simple lever

mechanism with the load on
one side and the actuator
(spring) on the other O
B q A
denoted as d q α will generate an allowed virtual displacement, without having to

worry about whether the displacements are allowed or not.
3. Using basic princ iples of geometry and trigonome try, find the locations of the
point where the external load is applied and the point of action of the SMA wire
in terms of the chosen kinematical parameters. That is, we want the relations
xi = f i (q α ), i = 1,... Nforces , α = 1,... N dofs (8.2)
where N f orces is the number of forces and N dof s is the number of degrees of
freedom. This is the hard part and requires much ingenuity. For the simple lever
problem, the relationship is
xA = OA (− cos θ i − sin θ j), x B = OB (cos θ i + sin θ j), (8.3)
4. Using calculus find the incremental changes in the locations of the point of action
of the external load in terms of the chosen kinematical variables.
 ∂fi
d xi = dq α (8.4)
α
∂qα
For the example in question, we have
dxA = OA (sin θ i − cos θ j)d θ, x B = OB (− sin θ i + cos θ j)d θ (8.5)
5. Now compute the virtual work done by each of the forces and set the sum equal
to zero
  ∂fi
Fi · d xi = Fi · dq α = 0 (8.6)
α
∂qα
i i
For the lever, since the force are −W j at B and − Fact j at A, we obtain
− (OA) F cos θ d θ − (OB)W cos θ d θ = 0 (8.7)
6. Since the result is true for ALL compatible virtual displacements, that is, for all
values of d q α , we obtain
 ∂fi
Fi · = 0 (8.8)
∂qα
i
as the equilibrium equations of the body. For the lever, this condition means that
OB
− (OA) F = −(OB) W =⇒ F = W (8.9)
OA
The above seems like an awful lot of trouble to get at the principle of the lever,
which is known to every high-schooler. However, the real power of the approach
is seen as the mechanism gets more complicated as shown in the next example.
Before proceeding with this example, we note that a key idea is to get the the virtual
work and to avoid vectorial representations as much as possible. Therefore, if we
choose our coordinates and what we calculate judiciously, we can use the fact that
only the component of the virtual displacement in the direction of the force actually
contributes to the virtual work and this dramatically simplify the calculations.
8.2 The Loading Curve and the SMA Response
Consider the scissor mechanism shown in Fig. 8.2, which is to be actuated by a SMA
spring actuator : As with the previous example, let us assume that the load W is 100 N
and the stroke S needed is 100 mm. In order to design the spring actuator the main
idea is to (a) find the stroke of the spring when the load moves through a distance of
100 mm and (b) find the load on the spring at the beginni ng and the end of the stroke
We will use this information to design the SMA actuator. If we want to compute
these using a conventional Newton’s laws, it will be very complicated. However, in
the current instance, the use of the virtual work principle makes our life significantly
simpler, as explained next.
• DOF First we note that the mechanism is a single degree of freedom mechanism
and the most convenient parameter q 1 = θ as shown in Fig. 8.2.
• KINEMATICS we note that the key motions are the vertical motion y of the
platform and the horizontal motion x of the point P, both of which we can express
in terms of the geometrical parameter L and the kinematical parameter θ . Thus,
from the geometry of the mechanism, we have
x = 2 L sin θ =⇒ dx = 2 L cos θ d θ (8.10)

y = 4 L cos θ =⇒ dy = −4 L sin θ d θ (8.11)
8.2 TheLoadingCurveandtheSMAResponse 121
L Y
SMA Spring Stroke Limiters
Anchor
X
Fig. 8.2 A scissor mechanism actuated by SMA springs with hard stops
• STROKE We next have to decide what length L must be to obtain a desired

stroke of 100mm. Based on experience, we expect that if θ is too close to 90 ◦
(fully flattened) or 0◦ fully deployed, the mechanism will “jam” and require very
large forces to operate. It is thus better to choose to operate around the 45◦ position.
We will choose an angle range of 60◦ to provide the stroke. So we set the operating
angles to range from 15 to 60 ◦ . With this choice. we see that the stroke is given by
Sl = ∆y = y (θ = 15◦ ) − y (θ = 60◦ ) = 1.863 L (8.12)

◦
Ss = ∆x = x (θ = 15 ) − x (θ = 60◦ ) = −1.21 L (8.13)
where Sl is the stroke of the load and Ss is the stroke of the spring. Since we want
a stroke of 100 mm for the load, we can therefore find the mechani sm arm length
L as
L = 100/1.863 = 53.7 mm =⇒ Ss = −1.21 × 53.7 = −65.2 mm (8.14)
Thus, the spring should be designed for a stroke of approximately 74.3mm. To

guarantee this, we will design a hard stop at x = 2 L si n 15◦ = 27.80 mm and
another stop at 2 L si n 60◦ = 93.01 mm. This provide s a stroke of 65.2 mm.
• FORCE While the external applied force is constant, due to the changing geometry
of the mechanism, the force at the spring will change. We need to estimate the
force needed at the beginning and the end of the stroke.
For this, we will use the principle of virtual work and note that
W d y + Fspr d x = 0 =⇒ −W × 4 L cos θ d θ + Fspr × 2 L sin θ d θ = 0 (8.15)

Fig. 8.3 Effect of using a 400

bias spring results in
smoother response—force
) 350
versus stroke plot N
(
g 300
n
i
r
p
s With bias spring
n 250
o Without bias spring
e
c
r 200
fo
l
a
n
r 150
e
t
x
E100
50
0 10 20 30 40 50
60 70 80
Spring Stoke (mm)
Since this is true for all virtual displacements d θ we obtain
2×W
Fspr = (8.16)
tan θ
For W = 100 N, we obtain

2 × 100 2 × 100
Fma x = = 746 N , Fmi n = = 115.47 N (8.17)
tan(15◦ ) tan(60◦ )
A plot of the load on the spring versus the spring stroke is shown by the solid
line in Fig. 8.3. Notice from the previous calculations, that a very large force is
required at the 15◦ angle and that the force drops sharply as the stroke progresses.
We note that the maximum load on the spring increased three fold while the stroke
decreased only slightly from 100 to 74 mm. This is the inefficiency as a result of
not following the design rules that were listed in Chap. 2. Specifically, the design
shown above does not provide a “short and direct load path”, that is, the load
exerted by the horizontally mounted spring is perpendicular the actual load which
is vertical and is therefore not a direct path. The need to change the load direction
is the cause of inefficiency. In fact, every violation of the design rules will cause
new difficulties in the actual embodiment of our ideas. Nevertheless, sometimes
it is not possible to satisfy all of them due to design constraints and we must be
prepared for such inefficiencies.
• SPRING DESIGN We can now design the spring. We will use the required stroke
of 65.2 mm and the max force of 746 N for the design and then ve rify whether the
minimum force of 115 N is sufficient to cause the spring to revert back to the cold
extended position. We will not detail the design procedure here but simply follow
Chap. 7.
8.2 TheLoadingCurveandtheSMAResponse 123
We quickly find that a single spring would be unsuitable. However, if we use

two springs in parallel we get more reasonable answers. From the design rules
that we have quoted in Chap. 2, this might be reasonable to maintain symmetry
of the device since the scissor mechanism will have two sides that have to be
actuated and we could have a spring on each side. However, it violates the design
principle of not over-constraining because we are using two actuators for a one
degree of freedom system. This means that we will have to be ready to deal with
the problems of having to synchronize the two springs to act together. Given the
highly non-linear nature of the response, we should be aware of the fact that this
would not be trivial task especially for larger strokes.
Nevertheless, we will proceed by designing springs for half of the loads but
with the same stroke. We thus have to design a spring to carry a maximum load
of 373 N and a minimum load of 58 N with a stroke of abou t 65 mm. Following
the approach presented in Chap. 7, we find that a preliminary design with a spring
index of 4 seems suitable with a wire diameter of 6 mm an OD of 30 mm and 17
active coils with a coil bound length of 168 mm. It is also apparent tha t the 57.7 N
force appears to be sufficient to fully extend the spring.
8.3 3-D Design
Please don’t confuse the mechanism shown in Fig. 8.2 as the real design. Real designs
live in a 3-D world and we have to consider a variety of factors to do this: A fully
featured design is shown in Fig. 8.4.
Notice the fact that the design exploits symmetry. Note also how the SMA actu-
ators are integrated into the base. Notice also how the “stroke limiter” is actually
Fig. 8.4 Real world scissor

mechanism
embodied in the slots. By moving the strain limiter to one platform and the SMA
springs into the other platform, the design achieves separation of functions in a very
efficient way.
8.4 Bias Forces
In the previous example we note that there is a huge excursion in the external force
on the spring. The force starts out at 373 N when the spring is fully ext ended and
decreases to just 57.7 N when the sprin g is fully retracted. This will give rise to
a “jerky” motion and sometimes will cause the spring to cease contracting (due to
friction) especially when it gets to the fully contracted mode. We can even this out by
attaching a bias spring to the mechanism . The bias spring is to resist the contraction
of the SMA spring. Thus, it is designed to be inactive when the SMA is fully extended
and gradually increases its resistive force as the SMA spring contracts.
We choose the bias spring as follows: When the SMA spring is fully extended,
the bias spring is free (i.e. coil bound) as the SMA contracts the bias spring extends.
So as the SMA spring goes through its strok e of 65 mm, the bias spring is designed
to exert a “compensatory force” of 373 − 58 = 315 N thus the spring rate of the bias
spring would be 315/65 = 4.85 N/mm. When this is added, the SMA works against
an almost constant force of 373 N (see the dotted line in Fig. 8.3) and its response
will be much smoother. Thus, bias springs can be used to change the response and
provide a reasonably “flat” or constant force against which the the SMA spring would
work.
Reference
1. Reddy J (2002) Energy principles and variational methods in applied mechanics, 2nd edn. Wiley,
New York
Chapter 9
Fatigue of SMAs
With growing applications of SMA components in different engineering applica-

tions, the issue of material performance over its designed life is of great concern to
researchers lately. Fatigue studies in SMAs is still an unsolved puzzle and we wish
to highlight some important items that affect the designers and some open questions
in SMA fatigue areas.
As pointed
of SMAs. out by
Factors Eggeler
include et al. [1],
material, numerous
chemical factors canmaterial
composition, influence the fatigue
impurity life
(i.e.,
particle size distribution), type of loading, applied load magnitude, extent of defor-
mation or plateau strains, processing conditions, defect accumulation, accumulated
plastic deformation, microstructure and operating temperature being a few to list [1].
In the following sections, we wish to highlight a high level classification of fatigue
in SMAs and theories employe d for analyzing fatigu e data.
9.1 Structural and Functional Fatigue in SMAs
Eggeler et al. [ 1] suggest subdividing fatigue in SMAs into two categories namely
“structural fatigue” and “functional fatigue” (see Fig. 9.1). SMAs failing like other
engineering materials due to repeated high cyclic mechanical loads is classified under
“structural fatigue” which is generally accompanied by defect accumulation, crack
initiation (normally micro-structure controlled and growth–geometry dependent),
crack propagation to critical lengths and final fracture [ 1, 2]. As seen with fatigue
in classical materials, one can expect cracks to initiate from specimen surface due
to surface irregularities (local surface conditions), scratches or inclusions that can
act as stress raisers. Once the cracks reaches critical crack lengths, final failure is
observed as the component can no longer handle external loads. In SMAs, the brittle
TiO2layer on the NiTi bulk material acts a stress raiser in the presence of any surface
irregularities or scratches that results in crack initiation and propagation [2].
126 9Fatigue
of
SMAs
Fatigue in Shape Memory Alloys (SMA)
Structural Fatigue Functional Fatigue

(Similar to coventional materials) (Degradation – Functional properties)
•
Defect Accumulation •
Reduced Damping capacity
• Crack Initiation • Lower Dissipated Energy
• Crack Propogation • Lower Plateau strains
• Final Failure • Poor shape recovery characteristics
Reporting Fatigue Data : Three main Approaches
S–N Theory Energy Approach

• Wöhler’s diagram • Superelastic behaviour compared to
•
Applied stress v/s cycles to failure plastic shakedown
• Dissipated energy v/s number of cycles
 -N Theory
• Coffin–Manson Equation
• Plastic strain accumul ation v/s cycles to failure
• Strain amplitude v/s cycles to failure
Fig. 9.1 Figure describes classification of SMA fatigue into two categories namely “struct ural
fatigue” and “functional fatigue” by Eggeler et al. [1, 2]. It also lists the three traditional approaches
(theories) of reporting fatigue data in SMA literature primarily devoted for capturing superelastic
effects (mostly mechanical loading effects with temperature being an external control parameter)
However, unlike classical material systems, SMAs can also suffer a degradation of
functional properties during cyclic thermomechanical loading i.e. either superelastic
effect or shape memory effect. Such degradation is classified under “functional
fatigue” and is predominately due to various intrinsic microstructural changes that
result in changes of component transformation temperatures, loss of memory, reduc-
tion in damping capacity (i.e. area of hysteresis), lower transformation strains etc.
being a few to list. [ 1]. In short, “functional fatigue” is the inability of the material
to perform its intended functionalities.
Structural fatigue is well documented in classical fatigue theories on metals,
ceramics and composites available in open literature and the same knowledge can be
9.1 Structuraland FunctionalFatigueinSMAs 127
extended to study structural fatigue of SMA’s. Hence, here, in this section, we would
like to focus more on “functional fatigue” aspects as they affect the performance of
actuators immensely especially those subjected to thermal cycling and it plays a vital
role from designers perspective.
Studies suggest that the degradation of these functional properties (especially in
thermal cycling) is primarily attributed to an increase in dislocation density during
repeated thermal cycling [ 1]. Initial cycles result in creation of many dislocations
and their formation progressively decrease with increasing cycles. This is mainly
due to “pockets/islands” of material that form stabilized martensite variants (plas-
tically deformed martensite) that fail to transform back to parent austentic phase
during thermal cycling [1, 2]. With repeated cycles, these “islands/pockets” of plas-
tically deformed martensite variants not transforming leads to accumulation of plastic
strains over cycles [1, 2]. With the diffusion processes in NiTi SMAs being practi-
cally insignificant as the the operating conditions for shape memory effect are much
much lower than the melting temperature of NiTi, it makes it even harder for these
“islands/pockets” to re-transform or not affect the SMA performance [1].
These “islands/pockets” are also observed in superelastic cycling where some aus-
tentite is untransformed even under higher stresses and also some martensite lattices
that slips at higher load levels leading to permanent strains. Fracture studies have
shown that crack growth in stress induced martensite (SIM) and these pockets/islands
are much slower than that of austenite as the stress intensity factor of SIM/pockets
is much higher
Ramaiah andthan austenite which
co-workers makes
[2] in their it harder
study for cracks
of thermal to propagate
cycling [1, 2].
of SMA wires
observe presence of multiple internal cracks at the core of the fracture region linking
to each other. They proposed that during thermal cycling, there exists a temperature
gradient across the wire cross-section (highest at the core and reducing radially) that
causes the volume fraction of phases to change radially along the temperature gradient
[1, 2]. This temperature gradient could be attributed to changes in local electrical
resistance due to crack generation and propagation during thermal cycling. Such a
temperature gradient indicates higher austenite at the core with a decrease radially to
the surface and vice versa for martensite volume fractio n [1, 2]. In events of forced
external cooling (forced convection scenario), this could be exaggerated as the outer
layers are being maintained at lower temperatures than the core and thus forcing
the outer sections to be martensite rich as they are not “hot enough” to transform
back to austenite instantaneously [1, 2]. This would imply that the softer martensite
phase is susceptible to larger deformation compared to inner austenitic core leading
to higher crack density and propensity to failure. Further, with the existence of the
temperature gradient across the specimen, complete recovery of bulk material is
constrained across adjace nt layers through the specimen thickness. This could lead
to development of shear stress/strains inhomogeneity along the longitudinal loading
direction of the wire and strain inhomogeneity between adjacent layers [1, 2]. These
internal stresses can act as stress risers and can cause faster crack propagations
leading to premature failure internally [1, 2].
Figure 9.2 shows a description of these phases across an SMA wire crossection
subjected to thermal cycling. The cross section can be treated as a composite of
many phases with strain inhomogeneity between phases that lead to formation of
128 9Fatigue
of
SMAs
Brittle Coating ( T iO2 )
Propagating Crack
Outer Martensite region
Transformation region (A ↔ M)
Plastically deformed Martensite
Locked Austenite core
Fig. 9.2 Figure shows a cross section of actuator SMA wire/spring after several thermal cycles.
The cross section can be conceived as a composite of many phases that affect the performance of
the SMA actuator—loss in functionality termed as functional fatigue
these “islands/pockets” locked in the material. This eventually leads to functional

degradation of its characteristic properties like loss of memory, stroke degradation,
shift in transformation temperatures etc. Though, the material might not have com-
pletely fractured, such a degradation of characteristic properties of SMA can render
them unproductive in their desired application . Further, the issue of crack propaga-
tion due to internal stresses and cracks srcinating due to surface irregularities (local
surface conditions), scratches or inclusions along with the inherent sensitiveness to
operating temperature fluctuations, material composition/finish/impurity, process-
ing conditions, and so on make fatigue in SMA’s a very tricky subject to understand
[1, 2]. We believe better understanding of these issues experimentally and develop-
ment of appropriate theories and tools for studying functional fatigue is still an open
area of research.
9.2 Reporting Fatigue Data
Over the years, numerous experimental data concerning SMA fatigue under super-
elastic conditions and thermal cycling (mimicking shape memory effects) have been
reported. Most of the focus has been in understanding fatigue behaviour of SMA
wires under tension with some attention towards bending—rotating tests of SMA
wire and functional fatigue of actuator springs [ 3–9].
The traditional approach of reporting fatigue data in most of the SMA literature
(see Fig. 9.1) is using any of the traditional fatigue theories listed below :
1. S-N theory or Wöhler’s diagram to capture “applied stress verses cycles to failure”
plots [10–15].
2.  -N theory that plot “strain amplitude verses cycles to failure” or employ “Coffin–
Manson (modified Coffin– Manson approaches in some cases)” under low cycle
fatigue cases to obtain “plastic strain accumulation verses cycles to failure” rela-
tionships [16–21] .
9.2 ReportingFatigueData 129
3. Energy approach that develop relationships between dissipated energy verses

number of cycles under stabilized superelastic effects. Superelastic behaviour is
compared to plastic shakedown for these analysis (see [22–27] for examples).
Such empirical theories solely focus on mechanical loading effects with temper-
ature being a external control parameter. Generally with such theories, relationships
are developed for each temperature case and are limited to analyzing superelastic
effects [27, 28]. Given the fact that thermomechanical responses of SMA are sensitive
to operating temperature fluctuations under superelastic conditions, such traditional
theories are incapable of these effects. In addition, these theories don’t throw much
light in capturing functional degradation charecterisitcs discussed in the previous
section under coupled thermomechanical cycling of SMA components.
To capture such complex thermomechanical responses, we believe theories con-
sidering both mechanical and thermal effects in a single framework must be employed
rather than modifying classical fatigue theories developed for pure mechanical load-
ing cycles [ 27, 28]. This is an open area of research and the authors are working in
this regard.
References
1. Eggeler G, Hornbogen E, Yawny A, Heckmann A, Wagner M (2004) Structural and functional

fatigue of niti shape memory alloys. Mater Sci Eng A 378(1):24–33
3. Tobushi H, Ohashi Y, Hori T, Yamamoto H (1992) Cyclic deformation of tini shape-memory
alloy helical spring. Exp Mech 32(4):304–308
4. Tamura H, Mitose K, Suzuki Y (1995) Fatigue properties of ti-ni shape memory alloy springs.
J Phys IV 5(8):C8–617
5. Grossmann C, Frenzel J, Sampath V, Depka T, Oppenkowski A, Somsen C, Neuking K,
Theisen W, Eggeler G (2008) Processing and property assessment of niti and niticu shape
memory actuator springs. Materialwissenschaft und Werkstofftechnik 39(8):499–510
6. Grossmann C, Frenzel J, Sampath V, Depka T, Eggeler G (2009) Elementary transformation
and deformation processes and the cyclic stability of niti and niticu shape memory spring
actuators. Metall Mater Trans A 40(11):2530–2544
7. Tobushi H, Hachisuka T, Yamada S, Lin P (1997) Rotating bend ing fatigue of a tini shape
memory alloy wire. Mech Mater 26(1):35–42
8. de Azevedo
straining on Bahia M, Fonseca
the behaviour Dias R, Buono
of superelastic niti.V
Int(2006) The28(9):1087–1091
J Fatigue influence of high amplitude cyclic
9. Miyazaki S, Mizukoshi K, Ueki T, Sakuma T, Liu Y (1999) Fatigue life of ti-50 at.% ni and
ti-40ni-10cu (at.%) shape memory alloy wires. Mater Sci Eng A 273:658–663
10. Miyazaki S (1990) Thermal and stress cycling eff ects and fatigue prop erties of ni-ti alloys.
Butterworth-Heinemann, Engineering Aspects of Shape, Memory Alloys(UK), pp 394–413
11. Melton K, Mercier O (1979) Fatigue of niti thermoelastic martensites. Acta Metall 27(1):
137–144
12. Mammano G, Dragoni E (2011) Functional fatigue of shape memory wires under cons tant-
stress and constant-strain loading conditions. Procedia Eng 10:3692–3707
13. Kang G, Kan Q, Yu C, Song D, Liu Y (2012) Whole life transformation ratchetting and fatigue
of superelastic niti alloy under uniaxial stress-controlled cyclic loading. Mater Sci Eng A
535:228–234
130 9Fatigue
of
SMAs
14. Ataalla T, Leary M, Subic A (2012) Functional fatigue of shape memory alloys. Sustain Auto-
mot Technol 2012:39–43
15. Scirè Mammano G, Dragoni E (2012) Functional fatigue of ni-ti shape memory wires unde r
various loading conditions. Int J Fatigue
16. Bertacchini O, Lagoudas D, Calkins F, Mabe J (2008) Thermomechanical cyclic loading and
fatigue life characterization of nickel rich niti shape-memory alloy actuators. In: The 15th
international symposium on: smart structures and materials and nondestructive evaluation and
health monitoring, pp 692916–692916, International Society for Optics and Photonics
17. Bertacchini O, Lagoudas D, Patoor E (2009) Thermomechanical transformation fatigue of
tinicu sma actuators under a corrosive environment-part i: experimental results. Int J Fatigue
31(10):1571–1578
18. Lagoudas D, Miller D, Rong L, Kumar P (2009) Thermomechanical fatigue of shape memory
alloys. Smart Mater Struct 18(8):085021
19. Figueiredo A, Modenesi P, Buono V (2009) Low-cycle fatigue life of superelastic niti wires.
Int J Fatigue 31(4):751–758
20. Runciman A, Xu D, Pelton A, Ritchie R (2011) An equivalent strain/coffin-manson approach
to multiaxial fatigue and life prediction in superelastic nitinol medical devices. Biomaterials
32(22):4987–4993
21. Maletta C, Sgambitterra E, Furgiuele F, Casati R, Tuissi A (2012) Fatigue of pseudoelastic niti
within the stress-induced transformation regime: a modified coffin-manson approach. Smart
Mater Struct 21(11):112001
22. Moumni Z, Van Herpen A, Riberty P (2005) Fatigue analysis of shape memory alloys: energy
approach. Smart Mater Struct 14(5):S287
23. Moumni Z, Zaki W, Maitournam H et al (2009) Cyclic behavior and energy approach to the
fatigue of shape memory alloys. J Mech Mater Struct 4(2):395–411
24. Soul H, Isalgue A, Yawny A, Torra V, Lovey F (2010) Pseudoelastic fatigue of niti wires:
frequency and size effects on damping capacity. Smart Mater Struct 19(8):085006
25. Dunand-Châtellet C, Moumni Z (2012) Experimental analysis of the fatigue of shape memory
alloys through power-law statistics. Int J Fatigue 36(1):163–170
26. Gloanec A, Bilotta G, Gerland M (2013) Deformation mechanisms in a tini shape memor y
alloy during cyclic loading. Mater Sci Eng A 564: 351–358
27. Rao A, Srinivasa A (2013) Experiments on functional fatigue of thermally activated shape
memory alloy springs andcorre lations with driving force intensity. In: SPIE smart structures and
materials + nondestructive evaluation and health monitoring, pp 86890T–86890T, International
Society for Optics and Photonics
28. Doraiswamy S, Rao A, Srinivasa A (2013) A two species thermodynamic preisach approach
for simulating superelastic responses of shape memory alloys under tension and bending load-
ing conditions. In: SPIE smart structures and materials+ nondestructive evaluation and health
monitoring, pp 86890X–86890X, International Society for Optics and Photonics

Edoc Design of Shape Memory Alloy Sma Actuators

Uploaded by

Copyright:

Available Formats

Edoc Design of Shape Memory Alloy Sma Actuators

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Edoc Design of Shape Memory Alloy Sma Actuators

Uploaded by

Copyright:

Available Formats

SPRINGER BRIE FS IN APPLIED SC IENCES AND

TECHNOLOGY  COMPUT ATIONA L MECHA NICS

Design of Shape Memory

ISSN 2191-530X ISSN 2191-5318 (electronic)

Library of Congress Control Number: 2015938741

Springer Cham Heidelberg New York Dordrecht London

Printed on acid-free paper

Springer International Publishing AG Switzerland is part of Springer Science+Business Media

This book is dedicated to my parents

This book is dedicated to my baava, Tummala

guaranteed behavior” over “optimal” response.

If a reader gains a qualitative understanding of SMA response together with the

1 Introduction to Shape Memory Alloys . . . . . . . . . . . . . . . . . . . . . 1

2 Need and Functionality Analysis . . . . . . . . . . . . . . . . . . . . . . . . . 33

3 Manufacturing and Post Treatment o f SM A Components . . . . . . . 43

3.1.7 Spark Plasma Sintering (SPS) . . . . . . . . . . . . . . . . . . 47

4 Basic SMA Component Geometries and Responses . . . . . . . . . . . . 61

5 Factors In fluencing Design of SMA Actuators . . . . . . . . . . . . . . . 73

5.3 Effect of Shape Setting Conditions for Custom Shape

6 Graphical Description of Temperature Controlled Actuation

6.4 Case Study 3: SMA Wire + Bias Spring Arrangement

7 Case St udies in th e Preliminary De sign of SMA A ctuators . . . . . . 99

7.7.1 Time Taken to Heat Up and Cool Down . . . . . . . . . . 114

8 Coupling SMA Actuators wi th M echanisms: Principle

9 Fatigue of SMAs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125

1.1 Smart Materials–An Overview

© The Author(s) 2015 1

Table 1.1 Multiphysics coupling

1.2 Smart Structures—System Level Response

Motor Vehicle System Smart Surgical Tool (Suture)

1.3 Shape Memory Alloys: Temperature Induced Phase

• High temperature phase

Self accomadated Martensite Single variant Martesnite

• Low temperature phase

• Monoclinic / Body Centered Tetragonal (BCT) crystal structure

Stress free thermal cycling in DSC

Twinned Martensite Austenite

Twinned Martensite Austenite

Ms : Martesnitic Start Temperature

phase transformations like melting of ice (solid–liquid), boiling of water (liquid–gas)

1.4 Shape Memory Effect and

Superelastic Eﬀect : Path 7 – 8 – 9 – 10 – 11 – 12 – 13

Shape Memory Eﬀect : Path 1 – 2 – 3 – 4 – 5 – 6

Fig. 1.5 A three-dimensional stress–strain–temperature plot describing the thermomechanical

deformed to state B under external load. Upon heating above temperatures A f ,

State A State B State C State D

Loading Heating ( > Af ) Cooling

SMA wire speci men Hot State Position

until point 10 at which it is thermodynamically stable for it to transform back to

response (see Fig. 1.7) are outlined here.

sample after loading to 6 %.

during loading and unloading operations.

specimen prior to necking or fracture or both.

Fig. 1.7 a Shows a typical (a)

1.5 Commonly Used Shap e Memory Alloy s

1950 1960 1970 1980 1990 2000

Ni-Co-Ti, Fe-Ni-Co-Al, Fe-Ni-Mn-Al, Co-Ni-Al, Co-Ni-Ga, Ni-Mn-Ga, Ni-Mn-In,

1.6 SMA Applications: Overview

1.6.1 Biomedical Applications

Sheath is removed and stent is