Progress in Materials Science 57 (2012) 911–946

Contents lists available at SciVerse ScienceDirect

Progress in Materials Science

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

Manufacturing and processing of NiTi implants: A review

Mohammad H. Elahinia a,b,⇑, Mahdi Hashemi a,b,1, Majid Tabesh a,b,
Sarit B. Bhaduri a,c
a
Mechanical Engineering, The University of Toledo, 2801 West Bancroft St., Toledo, OH 43606, USA
b
Nitinol Commercialization Accelerator, The University of Toledo, 2801 West Bancroft St., Toledo, OH 43606, USA
c
Multifunctional Materials Laboratory, The University of Toledo, 2801 West Bancroft St., Toledo, OH 43606, USA

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

Article history: NiTi is categorized as a shape memory alloy that found interesting
Received 20 June 2011 applications in vast areas of engineering from aerospace to bio-
Received in revised form 13 September 2011 medical; the latter applications are due to its biocompatibility in
Accepted 2 November 2011
addition to its unique properties. The unique properties such as
Available online 17 November 2011
shape memory and pseudoelasticity make NiTi an excellent candi-
date in many functional designs. However, the manufacturing and
processing complications of this alloy pose impediments to
widespread applications. This paper discusses challenges and
opportunities in making NiTi parts for biomedical applications
such as implants. To this end, common manufacturing processes
for NiTi from casting and powder metallurgy to machining are
discussed. Also, new opportunities in additive manufacturing
processes such as laser and electron beam techniques towards
making 3D components from NiTi are described. Finally, the
challenges in heat treatment and shape-setting of NiTi parts in
order to attain desired shape memory properties are reviewed.
Ó 2011 Elsevier Ltd. All rights reserved.

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 912
2. NiTi implants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 914
3. Metallurgy and processing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 915
4. Casting processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 917
5. Post-treatments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 919
6. Manufacturing challenges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 921

⇑ Corresponding author at: Mechanical Engineering, The University of Toledo, 2801 West Bancroft St., Toledo, OH 43606, USA.
Tel.: +1 419 530 8224; fax: +1 410 530 8206.
E-mail address: [email protected] (M.H. Elahinia).
1
PhD student focusing on materials science & engineering.

0079-6425/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.pmatsci.2011.11.001
912 M.H. Elahinia et al. / Progress in Materials Science 57 (2012) 911–946

7. Powder metallurgy methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 921

7.1. Conventional/non-conventional powder metallurgy processes . . . . . . . . . . . . . . . . . . . . . . . . . 922
7.2. Additive manufacturing powder metallurgy processes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 927
8. Heat treating and shape setting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 940
9. Conclusions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 942
Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 942
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 943

1. Introduction

Shape memory alloys (SMAs) such as NiTi have enabled technology development in various areas,
such as microrobotics and manipulation for instance [1,2,8]. These alloys undergo a reversible solid-
sate displacive crystalline phase transformation dominated by shear between a high symmetry parent
phase (austenite in the form of ordered BCC superlattice b phase in the case of Ni–50.0% Ti) and a low
symmetry product phase (martensite in the form of monoclinic distortion of a B19 lattice) [2]. The
deformation can also occur under martensite variant reorientation or detwinning of twins, where vari-
ants favorably oriented towards the applied load can form. There are 24 habit plane variants for stress-
induced martensitic transformation. This temperature and/or stress induced phase transformation (as
opposed to conventional diffusion induced transformations) is the basis for the unique properties in
these alloys, namely shape memory effect and pseudoelasticity. For an introduction, these important
concepts are briefly described in the following.
One way shape memory effect (1WE) is the recovery of large strains (up to 8%) mechanically cre-
ated in the low temperature range through reorientation/detwinning. This can be accomplished by
raising the temperature to a pre-specified higher temperature called austenite finish (Af) temperature.
Other important temperatures associated with this behavior are austenite start (As), at which the
recovery starts, and martensite start (Ms), and martensite finish (Mf), at which transformation to mar-
tensite phase starts and completes, respectively. These transition temperatures can be manipulated by
altering the chemical composition of the material at the time of manufacturing or through performing
thermomechanical treatments. The reorientation/detwinning is directional and in polycrystalline
materials can be affected by the texture.
In 1WE, only the shape of the parent phase is recovered. However, it is possible to stabilize the con-
figuration of martensite through the introduction of irreversible slip. The SMA specimen then remem-
bers the shape in both parent and product phases and can reconfigure between the two phases by
heating and cooling the specimen. This can be accomplished without application of an external load.
This phenomenon is known as two way shape memory effect (2WE). The stability of the 2WE can be
improved by an appropriate microstructure in the SMA. The SMAs with higher strength (i.e. finer grain
size), are harder to train and show smaller two way shape memory effect but with higher stability
against functional fatigue [3].
Pseudoelasticity (PE) is the other distinct behavior of these alloys. PE allows for a reversible stress–
strain behavior with strain values significantly higher than those of the classic metals or alloys. The
recoverable strain for a single crystalline sample of SMA can reach up to 10% [1]. PE describes the
non-linear recoverable deformation behavior of SMAs at temperatures above the Af temperature,
including the stress-induced martensitic (forward) transformation in loading and the spontaneously
thermally induced austenitic (reverse) transformation upon unloading. While the inelastic deforma-
tion caused by martensite variant reorientation/detwinning can be recovered through a reverse trans-
formation, some unrecovered deformation can be developed due to dislocation motion and glide
during loading. Factors such as temperature, annealing condition, grain size, stress mode, and micro-
structure (especially texture) can affect the subsequent shape recovery.
The three commercially important SMAs are NiTi, CuZnAl, and CuAlNi. Among them, NiTi is the most
commonly studied and used alloy due to its better functional fatigue and biocompatibility. As an inter-
metallic, this alloy has good ductility (which is related to the martensitic transformation with different
deformation modes), low anisotropy and relatively small grain size. Binary NiTi alloys have transforma-
 M.H. Elahinia et al. / Progress in Materials Science 57 (2012) 911–946 913

Nitinol material
fabrication with
desired shape and
transformation
temperatures.

Nitinol Nitinol device

Nitinol device Commercialization fabrication:
characterization: Laser cutting, heat
Accelerator
thermo-mechanical treatment, shape
testing, functional setting, surface
evaluation, lifetime treatment, and
and fatigue testing. welding.

Fig. 1. Development cycle at the Nitinol (NiTi) Commercialization Accelerator indicates the unique challenges and
opportunities in developing NiTi implants.

tion temperatures (Af) typically between 0 °C and 100 °C and show a temperature hysteresis of 25–40 °C
[4]. The Ni50.8–Ti (atomic%) with Af transformation temperature 8 °C is vastly used in pseudoelastic
SMA components. The transformation temperatures decrease drastically with increasing the Ni content
such that a 1% shift in the amount of either Ni or Ti in the alloy will result in a 100 °C change in the
alloy’s transformation temperature. For Ni-rich NiTi alloys, the solubility of Ni increases with increasing
temperature. Therefore, the thermomechanical treatment of these alloys involves precipitation of
metastable Ni4Ti3 and Ni3Ti2 phases. The formation of precipitates gives rise to the coherency stress
fields in the matrix and also decreases the Ni content of the matrix, hence it affects some features of
martensite transformation such as changing the transformation temperatures and strengthening the
parent phase (increasing the stress for martensite reorientation) [5–7]. It is also noteworthy that the
concentrations of point defects in shape memory intermetallics, which can be varied by composition
or heat treatment, have an effect on the Ms temperature and transformation hysteresis [6]. Generally,
any factor such as lattice defects, internal stress/strain fields, and precipitates, that influences the
mobility of the martensite interfaces can affect the shape memory properties of the alloy.
In addition to the distinctive properties of 1WE and PE, NiTi alloys have been recognized as desir-
able materials for bone implants because of their excellent corrosion, wear resistance, biocompatibil-
ity, mechanical properties, and high strength to weight ratio. NiTi has been used for orthopedics,
orthodontics, cardiovascular, and minimally invasive surgical instruments [23]. Elahinia et al. have
used NiTi to realize and control the behavior of systems and devices such as assistive rehabilitation
devices, minimally invasive surgery tools, and automotive actuators [9–22]. This paper covers the fun-
damental aspects of the behavior of these alloys, highlights biomedical application of SMAs, and re-
views the state of the art of manufacturing of these systems. This review is based on the experience
in device development by the mentioned group, at the Nitinol Commercialization Accelerator as illus-
trated in Fig. 1, in developing NiTi biomedical devices.
914 M.H. Elahinia et al. / Progress in Materials Science 57 (2012) 911–946

Fig. 2. Schematic stress–strain curves of stainless steel, NiTi, and bone [24].

2. NiTi implants

Fig. 2 [24], schematically compares stress–strain diagrams of three different materials. In the case
of stainless steel, the recoverable strain is less than 0.5%. On the other hand, up to 8% of deformation is
recoverable in NiTi. Similarly, bone exhibits more than 1% recoverable strain as well as hysteresis in
the loading–unloading cycles. This similarity between deformation behavior of NiTi and bone illus-
trates the biomimetic behavior of load bearing SMA implants under loading–unloading conditions
in the body. This property highlights another aspect of the excellent biomechanical compatibility of
NiTi. Moreover, NiTi exhibits unique corrosion behavior after deformation in superelastic and plastic
regimes in different deformation modes [24].
Two main drawbacks of NiTi are weak interfacial bonds, and mismatch of Young’s modulus be-
tween bone and implants for using these metallic implants in orthopedic surgery. A potential solution
to overcome these problems is to use porous NiTi. Mechanical properties of porous materials can be
altered and optimized by controlling porosity, pore size, shape, and pore distribution to better match
the properties of natural bone [25,26]. Porous NiTi has excellent mechanical properties that can be
engineered to match those of bones or tissues. The porous structure results in reduced density and
greater permeability that allows the in-growth of new bone tissue as well as the transport of body
fluids.
Many potential users of NiTi implants fear that in corrosive environments (such as human fluids) a
high dosage of Ni could be released. Thus, some prefer the use of less dangerous, low Ni-containing
materials such as stainless steels, pure titanium, and titanium or cobalt based alloys. It is however
known that equiatomic NiTi is an intermetallic compound with well-defined bulk and surface proper-
ties. The surface properties are largely defined by the fact that Ti is more readily oxidized than Ni. Gen-
erally, NiTi devices exhibit an outermost protective Ti-based oxide layer, which improves corrosion

Table 1
Physical and mechanical properties of NiTi vs. stainless steel [28].

Property NiTi Stainless steel

Recovered Elongation 8% 0.8%
Biocompatibility Excellent Fair
Effective modulus Approx. 48 (GPa) 193 GPa
Torqueability Excellent Poor
Density 6.45 (g/cm3) 8.03 (g/cm3)
Magnetic No Yes
Ultimate tensile Approx. Approx.
Strength (UTS) 1240 (MPa) 760 (MPa)
Coefficient of Thermal Expansion (CTE) Martensite – 6.6  106 (cm/cm/°C) 17.3  106 (cm/cm/°C)
Austenite – 11.0  106 (cm/cm/°C)
Resistivity 80–100 (lO cm) 72 (lO cm)
 M.H. Elahinia et al. / Progress in Materials Science 57 (2012) 911–946 915

resistance of this material, and acts as an effective barrier to Ni ion diffusion/release. Es-Souni et al.
have shown through in vitro cytotoxicity testing that NiTi alloys have good biocompatibility, albeit
varying with cell types, materials processing, and testing conditions [27]. SMAs have been used in
medical applications. Shape memory effect and pseudoelasticity can be employed to activate the med-
ical devices into operation through body heat or external sources of heat. Such designs could not be
realized with conventional alloys (see Table 1).
One of the advantageous features of the NiTi implants is that they can change shape after implan-
tation. This is possible by initiating the shape memory effect of this material using body temperature.
These shape changes give two primary benefits: (i) enhanced bone fixation and (ii) minimization of
invasive surgery [29].

3. Metallurgy and processing

The shape memory effect, pseudoelastictiy, damping, impact absorbing, and thermo-mechanical
properties of near equiatomic composition of Ni and Ti is strongly dependent on the stoichiometry
and thermal/mechanical treatments. A variation of 0.1 atomic% in the Ni content is shown to shift
the transformation temperatures by nearly 10 °C. Impurities such as oxygen, nitrogen, and carbon
should also be avoided since the transformation temperatures, hysteresis loop, strength, and ductility
of the material are very sensitive to these impurities.
Fig. 3 illustrates the Ni–Ti binary equilibrium phase diagram according to which thermodynami-
cally stable phases exist in the proximity of equiatomic percentages of Ni and Ti. Applying heat treat-
ments such as annealing, solution treatment, and aging can have significant effects on types of
microstructural phases in the final products and also on thermomechanical properties of NiTi devices.
The precipitation reaction of Ti11Ni14, during a 300–500 °C aging treatment of 50Ti–50Ni alloy,
causes migration of Ni atoms to participate and also causes diffusion of Ti atoms into the Ni–Ti matrix.

Fig. 3. Ni–Ti phase diagram [30].

916 M.H. Elahinia et al. / Progress in Materials Science 57 (2012) 911–946

Fig. 4. Conventional time–temperature–transformation (TTT) diagram for NiTi wire with initial Af = 11 °C [31].

This ultimately leads to an increase in transformation temperatures. However, at temperatures above

500 °C, the Ti11Ni14 precipitates dissolve and Ni atoms diffuse back in the matrix. This causes the
transformation temperatures decrease correspondingly [31]. Fig. 4 depicts the effect of aging temper-
ature and time on the phase transformation of Ti–50.8% Ni wire with an initial Af temperature of 11 °C.
To achieve the balance between driving force and diffusion rate required for phase transformation, for
all curves with different Af temperatures, the maximum precipitation rate occurs at about 400 °C,
which is the best temperature for aging. On the other hand, the higher Af temperature, the wider
the area of temperature hysteresis between austenite and martensite phases. Based on this fact, it
is possible to obtain totally martensite microstructure in samples with high Af temperatures easier
and in lower cooling rates. Nevertheless, above 500 °C, higher diffusion rate takes place and therefore
the time required for transformation to be finished decreases.
There are abundant amounts of processing data available in the industry to process the SMA mate-
rial. However, they are often kept proprietary and not released in the public domain. More so than

Table 2
Summary of the NiTi manufacturing methods.

Manufacturing of NiTi Components

Casting
i Powder Metallurgy

VAR
A VIM
M EBM Conventional Processes Additive Manufacturing

CS
C SHS
H HIP SPS MIM SLS SLM LENS
E EBM

Method Description Method Description

VAR Vacuum Arc Remelting SHS Self-propagating High Temperature Synthesis (combustion) Synthesis
VIM Vacuum Induction Melting HIP Hot Isostatic Pressing
EBM Electron Beam Melting SPS Spark Plasma Sintering
CS Conventional Sintering MIM Metal Injection Molding
SLS Selective Laser Sintering LENS Laser Engineered Net Shaping
SLM Selective Laser Melting
 M.H. Elahinia et al. / Progress in Materials Science 57 (2012) 911–946 917

other materials, its manufacturing processes significantly affect properties of NiTi. Fabrication tech-
niques can be classified according to the schematic illustration as summarized in Table 2.
SMAs are an excellent candidate for application in biomedical devices [17,21–23]. The one way
shape memory effect can be used to activate devices via either Joule heating or body heat. The pseudo-
elasticity (PE) of these alloys provides a non-linear reversible stress–strain behavior with nearly con-
stant stresses over a wide range of deformation; this is a very desirable feature for use in devices such
as orthodontic wires, self-expandable stents, or in minimally invasive approaches.

4. Casting processes

Production of SMA components conventionally entails arc or induction melting followed by a hot
working process and machining to the final shape. To minimize possibility of contamination during
melting, an inert gas working atmosphere is used. Pure raw materials are essential to achieve good
mixing of the constituent elements for making alloys with homogeneity and uniformity in the prop-
erties [32–34].
Two major methods of melting commonly used are Vacuum Induction Melting (VIM) and Vacuum
Arc Remelting (VAR). The cost of production by either method is similar and they both provide suitable
material for current medical device requirements (ASTM F2063).
VIM involves melting in a graphite crucible under vacuum or an inert gas atmosphere for the simul-
taneous melting of all raw materials. Electrical eddy currents are induced in the graphite crucible and
in the metallic charges. The alternating magnetic field produced by an induction coil is the source for
the eddy current. Also, electrodynamic forces result in stirring and mixing the melt. Graphite crucibles
are generally preferred for this process due to their easy handling, inexpensiveness compared to other
crucibles, and chemical homogeneity of the melt product. On the other hand during vacuum induction
melting in graphite crucibles, NiTi melts dissolve carbon and TiC particles form during solidification.
This increases the Ni concentration in the NiTi alloy, which in turn depresses the phase transformation
temperatures. Carbon content in the molten Ni–Ti alloy largely depends on the melt temperature. If
the melt temperature exceeds (939.4 °C) 1723 K, the use of a graphite crucible is impractical. Fortu-
nately, the melting point of the stoichiometric Ni–Ti alloy is (821.1 °C) 1510 K so that the melting pro-
cedure can be carried out at relatively low temperatures. The carbon content in the ingot prepared
under a pertinent operation lies between 200 and 500 ppm. Such a small amount of carbon does
not affect the shape memory characteristics of the alloy. Another advantage of induction melting is

Fig. 5. (a) Laboratory setup for VIM processing of binary NiTi (1 – graphite crucible; 2 – Ti rods; 3 – Ni pellets; 4 – isolation; 5 –
water cooled copper coil; 6 – mold); (b) schematic illustration of the crucible filling with Ni-pellets in contact with the graphite
crucible; (c) crucible filling with Ti-disk cladding preventing direct contact between Ni and graphite [33] .
918 M.H. Elahinia et al. / Progress in Materials Science 57 (2012) 911–946

the controllability of the chemical composition. If the operation proceeds carefully, the Ms tempera-
ture of the ingot can be controlled to within ±5 °C. Other crucible materials have problems such as sen-
sitivity to thermal cracking, higher prices, and thermodynamic instability as for oxygen evolution.
SAES Smart Materials and FURUKAWA, two major suppliers of NiTi, utilize this method [33,110,1].
Fig. 5 illustrates a laboratory setup for VIM processing of binary NiTi ingot metallurgy in a graphite
crucible and starting with pure elemental Ni pellets and Ti rods. The process has the ability to yield
1 kg of high quality low carbon and chemically homogeneous ingots. The process can achieve a signif-
icant reduction in the carbon contact by minimizing the direct contact between Ni and graphite uti-
lizing Ti disk cladding. The protective TiC layer on the inside surface of the crucible could partially but
not completely diminish the transfer of C from the crucible to the melt [33].
The VAR procedure does not need any crucible. It gives very high purity yields and obtains high
homogeneity through multiple melting cycles. Small amounts of inclusions could find their way into
the product but they are not evenly distributed. Wah Chang, an important supplier of NiTi, uses the
VAR method.
The VAR method is classified into two types with regard to the heating system. The first type
uses a non-consumable electrode and the second a consumable electrode consisting of materials
to be melted. The first method is preferred in laboratories because it is applicable to many kinds
of alloys. In this method, raw metals are installed on a copper mold and irradiated by the argon
arc from an electrode made of a tungsten rod. When the alloy is melted down, its shape resembles
a button due to the surface tension effect. The solidified button shaped ingot is turned over and re-
melted repeatedly to improve the homogeneity of the composition. A single arc-melting step is gen-
erally not sufficient to provide a homogeneous ingot, because only the upper section of the button
melts, while a small layer of the lower section (in direct contact with the water-cooled Cu hearth)
remains solid. To promote thorough mixing, remelting is required. In the second method, the fur-
nace uses a consumable electrode that consists of raw materials. The electrode has two roles: a heat-
ing source and a material source. The electrode is heated by the argon arc and the molten alloy
drops down onto the mold and forms a cylindrical ingot. The productivity of the second method
is higher than the first [35,111,1].
High-quality NiTi alloys can be produced by VAR if proper attention is paid to the required number
of melting cycles, ingot homogeneity, and oxygen pick-up. Frenzel et al. have previously showed the
effect of an increasing number of remelting cycles on ingot homogeneity. After melting, all ingots were
subjected to a homogenization anneal at (661.6 °C) 1223 K for 24 h. After five melting steps followed
by drop casting, the ingot showed good chemical homogeneity, which deviated only slightly from the
desired target compositions. A drawback of melting multiple times is the resulting carbon and oxygen
pick-up if any vacuum leak is present. Oxygen is more likely to be picked up during such an event. The
presence of carbides and oxygen-rich phases leads to a Ti depletion of the alloy, and in turn lowers the
phase transition temperatures [111].
Double melting process using VIM primary melting followed by VAR remelting is often used to get
further refining. Thirty-five and a half centimeter diameter ingots weighing 1000 kg are normally pro-
duced using the VIM/VAR double melt process.
The raw materials used for these processes are 99.99% pure titanium sponge and electrolytic nickel
with 99.94% purity. Inclusions and impurities can affect the martensitic transformation characteristics
such as transformation temperatures and hysteresis loop. For instance, ceramic-like inclusions could
depress the transformation temperatures.
Both VIM and VAR processes have major drawbacks including extreme reactivity of the melt and
segregation possibility due to the density difference of the reacting melts. Also, rapid grain growth
occurs as a result of subsequent high-temperature working which leads to poor fatigue properties.
Despite the fundamental differences between the VAR and VIM melting processes, VAR and VIM/
VAR double melt products appear to have similar mechanical and fatigue properties [35]. General
requirements on NiTi chemistry and trace elements are defined in the ASTM specification, F2063-
00 [36].
As a source of contamination in VIM, carbon combines with titanium precipitating TiC particles
with a much richer matrix in nickel content than the elementary composition. This contamination
consequently decreases the martensitic transformation temperatures [37,38].
 M.H. Elahinia et al. / Progress in Materials Science 57 (2012) 911–946 919

Fig. 6. EBM furnace model EMO 80, with 80 kW EB power: (A) lateral feeding mechanism; (B) melting chamber [37].

In addition to the conventional casting processes, an electron beam melting (EBM) process can be
used to make NiTi alloys. In this process, carbon and oxygen contamination is minimized because
melting is performed in a water-cooled copper crucible and under a high vacuum (pressure less than
102 Pa). The disadvantage of the EBM process in alloy production comes from the fact that during
melting and remelting, it is difficult to control the nominal chemical composition due to the high vac-
uum operation and heating temperature. This causes some constituent element evaporation, thus
changing the martensitic transformation temperatures. This effect is more pronounced on the nick-
el-rich side of the phase diagram. Meanwhile, composition homogeneity in the ingot is insufficient be-
cause the alloy solidifies uni-directionally from the bottom. In spite of these shortcomings, this
method is used to prepare NiTi SMAs, which do not require precise control of the transformation tem-
perature [37–40,1].
In earlier works, NiTi ingots were produced with good composition homogeneity using EBM fur-
naces. In a dynamic process, the raw material (nickel and titanium) is laterally fed continuously into
path of a vertical electron beam that melts the charge. Simultaneously, the molten product drops into
water-cooled copper extractor that is mounted inside the water-cooled copper mold. Fig. 6 shows an
EBM furnace, which has been already used to produce NiTi ingots [37].
In the EBM process, carbon content in final products is very low and ranges from 0.007% to 0.016%
compared to 0.04% to 0.06% of NiTi alloy commercially processed by VIM [37]. The oxygen content de-
pends on the internal pressure of the melting chamber and on the quality of the raw material. Oxygen
impurity is lower than that in the ingot produced by VIM [39,40]. Forty millimeter diameter by
270 mm long ingots weighing approximately 2.2 kg have been formerly manufactured using continu-
ous charge feeding and casting [39]. In general, it is possible to produce NiTi ingots using EBM pro-
cesses and the operation in theory permits scale-up [37]. A scale-up program of NiTi production is
now underway by analyzing operational aspects of EBM [40]. There is also research in progress to pro-
duce much cleaner materials for medical applications [39].

5. Post-treatments

As-cast microstructure and surface properties of NiTi products are not acceptable for medical appli-
cations and further processing is required. These post-processes can include hot working, cold work-
ing, machining, surface treatments, joining, and heat treatments. Hot working procedures include
press forging, hot rolling, and rotary forging. Final product shapes such as wire, tubing, and sheet
can be achieved via cold working. The average ductility of NiTi allows 30–50% of cold work. NiTi wires,
the most widely available form of this material, are produced via drawing. It is also possible to draw
seamless tubing (mostly available in superelastic form). Rolled sheets and strips of NiTi have been
used for photochemically etched and stamped devices [41,42].
920 M.H. Elahinia et al. / Progress in Materials Science 57 (2012) 911–946

Surface treatments are useful in enhancing the biocompatibility of the alloy. Titanium is a biocom-
patible element; however excessive intake amount of nickel may cause local and systemic toxicity,
carcinogenic effects, and immune responses. Nickel in NiTi is chemically joined to the titanium with
a strong intermetallic bond, so the risk of reaction even in patients with nickel-sensitivity, is extremely
low. One type of the surface treatments consists of a thermal oxidation, performed under low oxygen
pressure to avoid Ni oxidation, which leads to the formation of a pure titanium dioxide (TiO2) on NiTi
surface. This TiO2 oxide has been shown to efficiently protect the NiTi surface from release of Ni ions
into the exterior medium. Therefore, this new surface treatment is expected to improve NiTi cytocom-
patibility by decreasing the risks of toxic reactions associated to Ni. Also, TiO2 oxide on NiTi surfaces
has similar electrochemical corrosion resistance properties to native pure titanium oxide. This could
be of paramount importance when applying the oxidization treatment to NiTi devices for biomedical
applications. It is worth noting that pure titanium is a highly biocompatible metallic material widely
used in medicine because of the appropriate properties of its surface oxide [27,43].
Moreover, nitrogen or oxygen plasma immersion ion implantation (PIII) leads to dramatically im-
proved corrosion resistance and tribological properties such as surface hardness. The leaching of near-
surface Ni concentration in NiTi alloys has been significantly suppressed by implanting atoms on the
surface (either with N or O) using PIII. The effects can be attributed to the formation of a barrier layer
consisting of TiN and TiOx, respectively. Carbon plasma immersion ion implantation and deposition
(PIII&D) has been also proved to increase the corrosion resistance and other surface and biological
properties of NiTi. The ion-mixed amorphous carbon coating produced via PIII&D or direct carbon PIII
can improve the corrosion resistance and block the leakage of Ni and lead to enhanced surface
mechanical and biomechanical properties [44].

Fig. 7. Major drawbacks in machining NiTi shape memory alloys: (a) high tool wear; (b) undesirable chip formation; (c)
formation of burrs after turning (d) and grinding [41].
 M.H. Elahinia et al. / Progress in Materials Science 57 (2012) 911–946 921

6. Manufacturing challenges

Machining of NiTi is difficult due to its specific properties and resistance to deformation. As shown
in Fig. 7, machining causes severe tool wear. Abrasive methods such as grinding and abrasive saws are
therefore the preferred cutting methods for this material.
The machinability of NiTi significantly depends on the cutting speed and feed rate, which should be
chosen high enough. Poor chip breaking and the formation of burrs is another problem that can be
attributed to the high ductility as well as unconventional stress–strain behavior. Despite the optimi-
zation of machining parameters, tool wear still remains a problem in machining of these alloys. Dril-
ling of NiTi components with high cutting feeds and speeds helps extending the tool life and improve
work piece quality although it results in an increased microhardness in the subsurface zones of the
part [41].
Laser cutting, electric discharge machining (EDM), photo-chemical etching, and water jet cutting
have been proven to be better alternative processes to manufacture final products. Laser cutting
has the advantage of causing no mechanical stresses, no significant heat dissipation, no tool wear,
and high lateral resolution. However, laser-cut parts must undergo further processing to remove
the heat affected zone (HAZ). Modern laser cutting machines, using a pulsed Nd:YAG laser and
equipped with a CNC motion control system, offer high speed, high accuracy, and the capability for
rapid prototyping [45].
EDM works well with most NiTi compositions. A recast surface layer consisting of oxides and con-
taminants from Cu electrode and dissolved dielectric medium is generally present and may need to be
removed depending on the application [46].
The presence of surface oxide of NiTi makes it difficult for joining methods. Welding, brazing, and
soldering could be effective if carried out properly. Welding NiTi to itself has been successfully per-
formed using CO2 laser, Nd:YAG laser, tungsten inert gas (TIG), and resistance welding under an Ar
or He protective atmosphere. Pseudoelasticity and shape memory effect are generally well preserved
in these welding processes. Degradation in tensile strength and the resistance to permanent deforma-
tion in fusion zone and HAZ, however, was noted in particular during CO2 laser and TIG welding. Ti-
rich alloys are more susceptible to weld cracking. Using consumable filler metal in resistance welding
helps in reducing the risk and significantly increasing joint strength [45].
Joining NiTi to dissimilar metals is significantly more challenging. Joining NiTi to stainless steel can
be accomplished using proper interlayer materials. With proper design and tooling, mechanical join-
ing techniques such as crimping and swaging offer alternatives that can provide reliable joints be-
tween NiTi and dissimilar metals. Soldering NiTi alloys using halogen-based fluxes is described in
the US patent 5,242,759 [45].

7. Powder metallurgy methods

Ingot metallurgy (IM) processing involves casting molten metal into molds, solidifying at rela-
tively slow cooling rates (e.g. 0.1 °C/s), and subsequently converting to the final product form by
thermomechanical processing. The as-cast ingot exhibits chemical compositional variations, poros-
ity, and a non-uniform microstructure containing equiaxed and columnar grains. Manufacturing a
consolidated, nearly fully dense product involves ingot melting and casting, powder manufacturing,
and consolidation.
Powder can be made by various rapid solidification (RS) processes including gas atomization and
rotating electrode to form particulates. Gas-atomized powders generally have spherical particle shape
and require consolidation techniques such as hot isostatic pressing or metal injection molding. The
properties of parts produced by these methods are generally equal to and sometimes superior to their
IM counterparts. Alternatively, powder can be made by non-RS processes such as chemical reactions
including precipitation, or grinding and machining ingots.
The powder metallurgy (PM) processing route enables materials to achieve higher alloyed com-
positions without encountering segregation. This homogeneous microstructure with a consistent
composition leads to isotropic mechanical and physical properties and near-net shape parts in
922 M.H. Elahinia et al. / Progress in Materials Science 57 (2012) 911–946

comparison to IM. Not only do the properties have less variation than those properties found in IM
products, but the physical and mechanical properties are often enhanced due to the RS inherent in
the powder-making processes and consolidation. Better properties extend product life cycles or al-
low lesser grade materials to be used in more stringent applications. The ductility afforded by PM
processing also facilitates forging, rolling, extruding, and other hot-working processes. In addition,
machinability is enhanced due to the uniformity and fine microstructure of PM materials [112].
There are two kinds of powder metallurgy techniques for Ni–Ti alloys: raw metal powder sintering
and alloy powder sintering. In the former method, pure metal powders are blended, pressed and sin-
tered so that inhomogeneity in the alloy composition is inevitable. The latter method uses a prealloyed
powder and homogeneity of the sintered alloy is further improved. The processed alloy exhibits a good
1WE that is comparable to the vacuum melted alloy. Since powder metallurgy does not involve a melt-
ing process that causes composition inaccuracy, it can be applied to precise control of the transforma-
tion temperature [1].
Considerable research work has been devoted to PM processing of NiTi during the last two decades
with the aim of fabricating near net-shape components. However, there are several important prob-
lems in practice. The brittle oxide (Ti4Ni2Ox:0 < x 6 1) content in sintered alloys is generally high com-
pared to melt-cast alloys. Also, dense materials are difficult to obtain, particularly when starting from
elemental Ni–Ti mixed powder. This is due to the difference in diffusivity between Ti and Ni and cap-
illary effects resulting from the presence of Ti2Ni and Ni3Ti liquid eutectics and the high exothermic
NiTi formation reaction [24,27,47–49].
Considering production of NiTi components for industrial purposes, near net shape fabrication pro-
cesses are preferred due to the limited machinability of the alloy. PM is a well-known process for its
ability of providing semi-finished and net-shaped products. Furthermore, material savings, energy
savings, process automation, as well as precise control of the chemical compositions can be achieved
by PM while avoiding problems associated with the melting procedures like segregation.
Various PM processes have been experimentally developed for NiTi in which both prealloyed pow-
ders and elemental powders can be used as starting materials. Inert gas atomization [50], hydriding
and pulverization [51], or mechanical alloying [52] can be used to produce prealloyed NiTi powders
[45]. In this paper the powder-based methods are divided into: (1) conventional PM processes, and
(2) additive manufacturing PM processes.

7.1. Conventional/non-conventional powder metallurgy processes

Five methods have been used for the production of dense and porous NiTi from elemental and pre-
alloyed powders. These include conventional sintering (CS) [53], self-propagating high temperature
synthesis (SHS) [54,55], sintering at elevated pressure via a hot isostatic pressing (HIP) [30], spark
plasma sintering (SPS) [56], and metal injection molding (MIM) [57,58].

Fig. 8. Schematic representation of the SHS process: Ni + Ti ? NiTi + 67 kJ/mol.

 M.H. Elahinia et al. / Progress in Materials Science 57 (2012) 911–946 923

In conventional sintering (CS), a green compact of elemental Ni and Ti powders is prepared and fur-
ther sintered at near melting temperatures to yield in binary NiTi from diffusion of Ni and Ti elements.
Conventional sintering requires long heating times and samples are limited in shape and pore size. The
porous structure is shown to be of small size and irregular shape. Maximum porosity of 40% has been
achieved with this procedure [53,59].
Self-propagating high temperature (combustion) synthesis (SHS) is initiated by a thermal explosion
ignited at one end of the specimen which then propagates through the specimen in a self-sustaining
manner (due to the exothermic reaction between Ni and Ti). One of the difficulties with SHS is the
inability to control the intermetallic phases [55]. A schematic representation of the SHS process is
shown in Fig. 8.
The porosity of the SHS product depends on the original porosity of the green compact and synthe-
sis parameters such as the change in the molar volume, the combustion front thermal gradients, and
the gas evolution as a result of volatile impurity expulsion. The porosity and the mean pore size and
the distribution of that can, to some extent, be controlled in SHS. Performing the reaction under a re-
duced pressure or adding a gasifying agent increases the porosity. Manipulating the reaction temper-
ature by adding diluents to the initial mixture can be a way to control the pore size. Uniformity of the
temperature profile within the sample greatly affects the homogeneity of the product. Nevertheless,
Ti2Ni, Ni3Ti, and Ni4Ti3 stable phases are usually present in the SHS product matrix that will lead to
a corresponding embrittlement effects [60,61]. Porosities above 50% can be achieved by SHS [62].
The synthesis of NiTi with this method requires preheating of the sample to achieve self-sustained
combustion since the exothermicity of the reaction is relatively low [63]. The preheating temperature
affects the amount of transient liquid phase present at the combustion front. Excessive pre-heating
has been shown to have detrimental effects such as anisotropy in the pore structure [62]. The SHS
reaction can be performed in two different approaches; first by locally initiating the reaction which
will further propagate along the sample [55,62] and second by volume combustion [63] i.e., heating
the entire sample up to the reaction temperature which will let the reaction to take place simulta-
neously throughout the whole sample [62].
Unlike the conventional metallic SHS reactions, the ignition temperature is not around the lowest
melting point of the constituents and is shown to be in the range of 870–1070 °C, which is well below
the melting point of both Ni and Ti. Ignition temperatures, times, and energies for different NiTi com-
positions are significantly affected by the heating cycle of the sample and less affected by the reactant
particle’s size and composition. Moreover, it was shown that controlling the preheating phase in the
SHS sample via an external furnace while triggering the ignition in a very short period of time (in this

Fig. 9. Porosity with a banded structure in a specimen prepared by SHS method [55].
924 M.H. Elahinia et al. / Progress in Materials Science 57 (2012) 911–946

case via laser power) would alleviate the temperature gradients inside the sample leading to a much
more homogeneous product [54].
A bulk porous NiTi SMA has been produced via SHS from Ni and Ti green compact with a pre-heat-
ing temperature of 550 °C. The porosity of the green compact, the transient liquid phase, and the vol-
atilization of the impurities followed by the escape of absorbed gases left three dimensionally
interconnected pores in the product which formed banded channel structures along the propagation
direction of the reaction as shown in Fig. 9. The bulk products with high desired porosity (to reach
mechanical properties close to those of bones) containing 100% intermetallic NiTi compounds were
produced [55].
Hot isostatic pressing (HIP) is a pressure enhanced sintering technique that can be utilized to man-
ufacture theoretically dense products (0.0% porosity). The stress-induced martensite transformation in
dense NiTi occurs throughout the entire medium at nearly the same stress level, commonly resulting
in easily identifiable regions of phase transformations in stress–strain data charts obtained from
mechanical tensile tests. The mixture of elemental powder particles is encapsulated in an evacuated
gas-tight welded canister and undergoes simultaneous isostatic pressure and elevated temperature.
Other than that, Argon gas can be used as an inert environment without the need for the airtight
chamber. In this case, HIP can be used to compress and trap Ar gas bubbles in between the elemental
powders. A subsequent high-pressure diffusion stage leads to Ar-filled pores. Sintering the product at
reduced pressures causes the gas to expand and bring about near-spherical pores in the final product.
Elemental powders can be used in this technique leading to relatively homogenous specimens, as
shown in Fig. 10. However, presence of NiTi2/Ni3Ti precipitates in the matrix cannot be avoided [64].
The advantages of HIP processes include shorter solid state diffusion time, good control over the
pore size, and ability to manipulate various geometries, resulting in a thermodynamically stable
and manageable reaction compared to SHS process. A typical heating and pressurizing procedure
for HIP is shown in Fig. 11 [30].
The capsule-free HIP process was shown to produce homogenous porous NiTi (30–40 vol.%) with
near-spherical pores. The structure could show acceptable pseudoelastictiy due to the removal of
stress concentration in near-spherical pores. Controlling the porosity characteristics in a porous mate-
rial has a great influence on the mechanical properties of the products [59].
Spark plasma sintering (SPS) method, also known as Pulsed Electric Current Sintering (PECS) is used
to produce porous NiTi components starting with prealloyed powders. Fig. 12 is a schematic drawing
of the SPS device. In this procedure, prealloyed powders are loaded and pressed in a graphite die. A
strong step current is applied to the compact. The pulsed current introduces high energy to the com-
pact, which causes a joint formation in the particles at relatively low temperatures and in a very short

Fig. 10. Formation of non-binary phases in HIP process as a result of non-complete diffusion [64].
M.H. Elahinia et al. / Progress in Materials Science 57 (2012) 911–946 925

Fig. 11. Typical procedure for an HIP cycle [30].

Fig. 12. Schematic drawing of the SPS device.

Fig. 13. Schematic process of the metal injection molding [65].

926 M.H. Elahinia et al. / Progress in Materials Science 57 (2012) 911–946

period of time compared to sintering processes such as CS, HIP, or SHS. The spark discharge also results
in the purification of the particles’ surface leading to a high quality sintered product [56].
Metal injection molding (MIM) is another powder-based procedure for manufacturing near-net-
shape NiTi products. MIM operates on the basis of the injection molding of plastics. This process
was developed for massive production of small sized (normally below 400 lm) and complex shaped
parts in a cost-effective manner. Furthermore, the process can also be used for high-density metallic,
intermetallic, ceramic, and composite components. Advantages of this technique include geometrical
precision of the parts and low costs for high production volumes [57,58].
As shown in Fig. 13 [65], the MIM process entails four major steps: feedstock fabrication, injection
molding, debinding, and sintering [66]. To reach the best results, each step must be optimized based
on its effective parameters. In the feed stock preparation, the alloy powder particles are mixed with a
binder whose composition and volume percentages are critical. The feed stock is later injected into the
mold; the pressure and temperature of the feed as well as the mold are the parameters that can be
manipulated. It is worth mentioning that special attention should be directed to the design of the
mold such that it gives proper pressure and temperature distribution of the melting front and ensures
damage free removal of the green product out of the mold. Debinding is performed in a chemical bath
and also under a vacuum with an elevated temperature environment to vaporize the binder from the
product. Finally, the sintering stage is performed at high temperature levels (approximately 1200 °C),
which is sought to yield a product with a density close to the material’s theoretical density [30].
Due to the low sintering activity of the NiTi powders and the lack of pre-compaction, high sintering
temperatures are usually needed in MIM process [30]. Furthermore, preparing the suitable powder in
terms of particle size and composition, as well as protecting the process from external contamination,
guarantees a successful MIM process.
In order to produce the structural aerospace components and medical implants, ASTM F 167 spec-
ification recommends that the oxygen impurity levels must be kept below 300 ppm. Hence the afore-
mentioned procedures should be performed under a protective environment and with minimal
interaction with external contaminants.
In an attempt to manufacture medical devices from NiTi, Krone et al. [66] utilized MIM technique
to produce a staple implant (see Fig. 14). To avoid the exothermic reaction problems with elemental Ni
and Ti powders, NiTi prealloyed powders (diameter <21 lm) were used. It was previously established
that prealloyed powders are required for NiTi MIM process because of the anisotropic swelling of the
compounds made of elemental powders in the sintering stage [30].
An increase in the impurity level after processing was due to oxygen impurities. The level of oxygen
in the sample changed from 0.11 to 0.23 wt.%. The density of the MIM product was 95% of NiTi theo-
retical density. In the course of the MIM procedure, Ti-rich impurity phases (TiC or TiO2) appeared in
the binary NiTi matrix that changed the composition in favor of an Ni-rich phase. This shifted the
phase transformation properties compared to the initial alloyed NiTi powders. A tensile test showed
that the product can be elongated up to 7.3% of strain prior to its brittle fracture and can undergo a

Fig. 14. Dimensions of the manufactured staples as well as photographs of the green and sintered specimens [66].
 M.H. Elahinia et al. / Progress in Materials Science 57 (2012) 911–946 927

Table 3
Advantages and disadvantages of conventional/non-conventional PM processes.

PM Advantages Disadvantages
Process
CS Low cost, good dimensional precision, high production Products contain large amount of pores (as high as
rate, availability of wide variety of materials, %40), no good control on pore size and amounts,
elimination of the need for secondary machining creation of secondary phases is probable [69],
operations [53,68] presence of residual porosity, size and shape
limitations, long heating time [53,68]
SHS Low energy requirement, relative simplicity of the Products contain large size and amount of porosities
process and equipment, higher purity of the products, (as high as %65) [69,71], samples often not fully
low cost [54,55,70] reacted or contain precipitates due to the short
forming time and high heating rates, incomplete
reactions between elemental powder particles,
creation of secondary phases probable [69]
SPS Low sintering temperature and short processing time Expensive pulsed DC generator is required [56],
avoiding any undesired reaction products, different samples with simple shapes [74]
materials (metals, ceramics, composites) could be
processed, high energy efficiency, precision control
over heat, cooling and pressure, uniform sintering,
ease of operation [56,72,73]
HIP Dense products with no pores and good mechanical Inert gas may cause porosities up to %40, creation of
properties possible [69,75], good control on pore size secondary phases is probable [69], costly equipment,
[69], low sintering temperature [75], nearly complete low production rate [68]
reaction while taking low process time [69,1], able to
produce large size and/or complex shape products,
high efficient in materials use [68]
MIM Good shape complexity, high production rate [68], Products with residual pores, high sintering
products with mechanical properties nearly temperature (large amount of impurity phases [75],
equivalent to wrought materials,good dimensional costly tooling [68], part size limitations [57,58]
tolerance control [57,58]

level of stress around 780 MPa. With increasing the number of cycles in a fatigue test, the critical
stress for the start of the transformation reduced, the residual strain increased, and the hysteresis loop
diminished. This fatigue behavior is typical to NiTi shape memory alloys. Furthermore, the thermal cy-
cling was shown to have negligible effects on the transformational properties of the material. Overall,
NiTi parts produced with the MIM process showed stable behavior in terms of cyclic mechanical load-
ing and was suitable for the application in medical devices [67]. Advantages and disadvantages of all
the conventional/non-conventional PM processes are summarized in Table 3.
It is important to know that in many cases (e.g. manufacturing porous bone implants), some dis-
advantages such as large amount of porosities might be assigned as advantages to some extent based
on PM products usability [71].

7.2. Additive manufacturing powder metallurgy processes

The general term Additive Manufacturing (AM), also known as Rapid Prototyping (RP) or Rapid
Manufacturing (RM), describes the process of making a part by adding successive layers of the mate-
rial, rather than removing the material, such that there is little or no waste. Each layer is melted
according to an exact geometry defined by a three-dimensional computer aided design (3D CAD) mod-
el. Additive manufacturing has advantages of building parts with very complex geometries without
any sort of cutting tools or fixtures. It is also a fast production route from CAD to physical part with
very high material utilization and does not require expensive castings or forgings. Therefore, it is a
very cost-effective, energy efficient, and environmentally friendly manufacturing process. The most
common and known processes for making shape from metal powders are Selective Laser Sintering
(SLS), Selective Laser Melting (SLM), Laser Engineered Net Shaping (LENS) and Electron Beam Melting
928 M.H. Elahinia et al. / Progress in Materials Science 57 (2012) 911–946

(EBM). Additive manufacturing technologies lead to today’s freeform fabrication machines that build
parts using a wide variety of metal powder particles [76,77].
In AM, process parameters can be lumped into four categories: (1) beam-related parameters (laser/
electron beam power, spot size, pulse duration, pulse frequency, etc.), (2) scan-related parameters
(scan speed, spacing, and pattern), (3) powder-related parameters (particle shape, size, density, distri-
bution, layer thickness, etc.), and (4) temperature-related parameters (powder bed temperature, pow-
der feeder temperature, temperature uniformity, etc.). The product quality requires the development
of a set of optimized processing conditions or parameters, which assure uniformity and control of
microstructure in the associated mechanical properties and performance. Accuracy and surface finish
of powder-based AM processes are typically inferior to liquid-based (casting) processes. However,
accuracy and surface finish are strongly influenced by the operating conditions and the powder par-
ticle size. Materials with low thermal conductivity result in better accuracy as melt pool and solidifi-
cation are more controllable and grain growth is minimized when heat conduction is minimized
[24,77].
Biomedical implants made though AM processes have resulted in microstructures which give rise
to a mechanical behavior similar to or even superior to wrought or cast products (hardness variations
and tensile strengths ranging 37–57 HRC and 0.9–1.45 GPa respectively) [78]. As a recognized advan-
tage of layered manufacturing, AM processes are very suitable for fabricating porous products [78–80].
Dense titanium alloy implants, for example, are approximately twice as heavy as natural dense corti-
cal bone equivalent. More importantly, the elastic modulus of titanium alloy is about 114 GPa while
that of bone ranges from 0.5 GPa to a maximum of 20 GPa. It is also imperative to build implants with
weights closer to that of the replaced bone. In addition, in order to have long-life products the effective
stiffness of the implant must be comparable to that of healthy surrounding bone tissue. To reduce the
weight and effective stiffness of the titanium implants, they must be made porous. This can help to
reduce the mismatch of properties between the implant and host bone leading to a reduction in the
stress shielding effect and its associated bone necrosis. Porous structures also help as conditions for
material for bone growth.
Selective laser sintering (SLS) is the first commercialized AM process that uses a high power laser to
fuse small powder particles of materials into a mass which has a desired 3-dimentional (3D) shape.
SLS process, a computer-controlled technique, was originally developed for producing plastic proto-
types using a point-wise laser scanning technique and it has subsequently been extended to metals,
ceramic, and composite powders. In fact, the SLS process can produce parts from a relatively wide
range of commercially available powder materials [77]. SLS is an expensive process because it uses
a high-powered laser. Additionally, SLS products need to be finish-machined or polished to achieve
the required accuracy and surface finish [81].

Fig. 15. Schematic of the selective laser sintering process [77].

 M.H. Elahinia et al. / Progress in Materials Science 57 (2012) 911–946 929

Fig. 16. SEM images of the Nitinol implant: (a) morphology (b) microstructure [84].

Table 4
EDX analysis data (wt.%) of areas S1 and S2 shown in Fig. 16a [84].

C O Ti Ni
S1 11.93 22.18 41.95 23.94
S2 8.97 31.29 32.93 26.81

As shown in Fig. 15, powder layers (typically 0.1 mm thick) are built by a rotating roller. All the
process is done in an enclosed chamber filled with nitrogen gas to minimize oxidation. A focused
CO2 laser is directed onto the powder bed to fuse and form a slice cross-section. Subsequently, a
new layer of powder is laid and the process is repeated [77].
930 M.H. Elahinia et al. / Progress in Materials Science 57 (2012) 911–946

A pre-heating treatment is necessary in the SLS process, to minimize the laser powder require-
ments and also to prevent warping during thermal expansion and shrinkage (curling). Finally, to avoid
warping and oxidation, a controlled cool-down cycle is necessary to handle the cooling rate of build
from working temperature to ambient temperature [77].
NiTi implants have been manufactured using SLS process [82]. SLS/SHS combined processes have
also been used to synthesize NiTi [83–87]. The combined process results in implant with higher homo-
geneity in chemical composition, better biocompatibility, and more porosity as needed. Under the
optimum regimes of the combined process, X-ray diffraction (XRD) patterns showed traces NiTi, Ti2Ni
and Ti3Ni, NiTi was the main phase detected with compositions ranging from 60% to 80% by mass [84].
The NiTi implants showed no adverse tissue reactions and exhibited a martensitic transformation in
the temperature range of 50–0 °C, which can be interpreted as a 1WE [87].
Nitinol samples produced by SLS/SHS process were characterized using a scanning electron micro-
scope (SEM) equipped with energy-dispersive X-ray spectroscopy (EDX). Fig. 16a shows morphologi-
cal images of rough and porous surface of products. It has been found that the resulting porosity in the
combined process is suitable for bio-integration. Fig. 16b shows microstructure of Nitinol after SLS.
Results of the EDX analysis are listed in Table 4 [84]. As seen, this novel method allows layer-by-layer
synthesis of biocompatible implants from NiTi powders, which do not possess measured free nickel
content and toxicity. It is important to mention that the presence of carbon and especially oxygen
is related to the residual air content in the pores of initial powder mixture. [84,87].
It was discovered that the surface of porous Nitinol products made by SHS/SLS possessed a signif-
icantly favorable structure to the mechanical interlocking with bone and soft tissues. Regarding both
the surface condition and porosity of implants (porosity 40–50%) produced by SHS/SLS, porous NiTi
implants were previously embedded in the living tissues (see Fig. 17). NiTi will completely be inte-
grated after a 3-month period with the organotypic structures and soldered to bone tissues [84–87].
Selective laser melting (SLM) uses high-powered laser beam (usually a yttrium fiber laser) to create
3D metal parts by fusing fine metallic powders together. The use of lasers with wavelengths better
tuned to the absorptivity of metal powders is one enabling feature of SLM of metals. Today, manufac-
turers use Nd:YAG laser instead of the CO2 laser used in SLS, which results in a much better absorp-
tivity for metal powders (see Fig. 18). Subsequently, almost all SLM machines today have transitioned
to fiber lasers, which are generally cheaper to purchase and maintain, more compact, energy efficient,
and have better beam quality than Nd:YAG lasers. The other key advantage of SLM, compared to SLS,
are different laser scan patterns, the use of f-theta lenses to minimize beam distortion during scan-
ning, and low oxygen inert atmosphere control [77]. The major benefits and drawbacks of SLM for me-
tal powders are summarized in Table 5 [88].
Differential Scanning Calorimetry (DSC) results of feedstock NiTi powder (Ni54.4–Ti) and of the so-
lid SLM processed parts (Fig. 19) indicate that martensitic transition occurs between 32 °C and 58 °C

Fig. 17. Microstructure of the bone tissue in-growth into Nitinol implant after 3 months of biointegration [86].
 M.H. Elahinia et al. / Progress in Materials Science 57 (2012) 911–946 931

Fig. 18. Optical absorption% (absorptivity) of selected metals vs. wavelength (courtesy of Optomec) [77].

Table 5
Benefits and drawbacks of SLM process [88].

SLM Features Benefits Drawbacks

Material No distinct binder and melt phases; hence, the process can Not suited for well controlled composite
choice produce ‘‘single material’’ parts (e.g. steel, Ti or Al alloys), materials (e.g. WC-Co)
rather than producing a composite green parts which
might not be desired
Production Elimination of time consuming and costly furnace post- The laser powder processing needs higher
steps treatments for debinding (in case of polymer binder energy level: i.e. high laser power, good
(time, phase), infiltration or post-sintering beam quality (more expensive laser) and
cost) smaller scan velocities (longer build times)
Part quality Better suited to produce full dense parts (even over 99.9%) SLM suffers more from melt pool
in a direct way, without post-infiltration, sintering or instabilities (surface roughness, risk of
HIPing internal pores) and higher residual stresses
(risk of delamination, distortion when
removing base plate)

Fig. 19. DSC analysis of NiTi powder and solid SLM components [48].
932 M.H. Elahinia et al. / Progress in Materials Science 57 (2012) 911–946

Fig. 20. (a) Surface roughness results as a function of sloping angle, layer thickness (LT); (b) influence of sloping angle on stair
effect [89].

and the austenitic transition is observed between 58 °C and 90 °C. Both the austenite (As and Af) and
martensite (Ms and Mf) transformation temperatures and phase changes in both powder and solid part
correspond. However, the phase transition within SLM NiTi parts spans a wider temperature range and
is far more gradual than that within conventionally processed parts [48].
This gradual phase transition could result in a slow macroscopic shape transformation and present
many potential advantages especially within applications related to microelectromechanical systems
(MEMSs). This gradual behavior allows one to control 1WE changes more easily by optimizing the pro-
cess parameters [48].
Medical applications are very amenable to production by SLM due to their complex geometry,
strong customization and high-aggregate price. To develop SLM into a viable manufacturing technique
for implants, the laser-melted parts have to meet strict material requirements regarding mechanical
and chemical properties and the process must guarantee high accuracy and appropriate surface rough-
ness [89].
The surface roughness depends on many factors: material, powder particle size, layer thickness, la-
ser parameters, scan parameters, scan strategy, and surface post-treatment. Fig. 20 shows the influ-
ence of sloping angle and layer thickness on surface roughness. The gradient effect can be reduced
by decreasing the layer thickness or by increasing the sloping angle. In both cases, more stairs appear,
but the size of the stairs becomes smaller, leading to lower surface roughness. Simple surface post-

Fig. 21. Schematic illustration of the LENS process.

 M.H. Elahinia et al. / Progress in Materials Science 57 (2012) 911–946 933

treatments such as glass blasting and ultrasonic ceramic filing can remove partial molten particles on
the surface, leading to strong reduction of the roughness [89].
Laser Engineered Net Shaping (LENS) was developed by Sandia National Laboratories, and commer-
cialized by Optomec Design Company. Optomec’s machines originally used an Nd:YAG laser, but more
recent machines utilize fiber laser. In a LENS machine, heat from a laser beam forms a small melt-pool.
Simultaneously, metallic powder is injected into the molten pool, building up a feature. 3D CAD soft-
ware is used to manipulate the marking head or the X/Y table holding the part to deposit the data ma-
trix identification symbol (see Fig. 21). The head is moved up vertically as each layer is completed.
Metal powders are delivered and distributed around the circumference of the head either by gravity
or using a pressurized carrier gas. An inert gas is often used to shield the melt pool from atmospheric
oxygen. The injected metallic material can be added to a structure made of different but compatible
material [77].
Typically LENS is performed as a full melting process. It is able to produce fully dense products con-
taining strongly bonded particles. To fabricate a porous implant, the full melting of powders should be
avoided. In this case, the particles are partially melted to create porosity. These surface melted parti-
cles join together due to the presence of liquid metal at the particle interfaces and bond well with the
previous layers. However, there will always be some entrapped powder and loosely attached powder,
especially at the surface of the implant. For the loose powders at the surface or in open pores, a simple
chemical or mechanical treatment such as grit blasting can be used to remove them. For loose pow-
ders in the closed pores, powder removal is not possible. However, these powders do not cause any
irritation, as they are not in direct contact with bone tissue [79].
It is common for laser deposited parts to exhibit superior yield and tensile strengths because of
their fine grain structure. Ductility of parts, however, is generally considered to be inferior to wrought
or cast equivalents. Layer orientation can have a great influence on percent elongation by minimizing
residual stress build-up. Residual stresses are generated as a result of extremely high solidification in
LENS, which can lead to cracking during or after part construction. However, in many alloys ductility
can be recovered and anisotropy minimized by heat treatment without significant loss of strength
[77].
The LENS system can produce parts in a wide variety of metals including titanium, nickel-base
super alloys, stainless steels, and tool steels. In LENS, for instance, the nano/microstructure can be tai-
lored (for example by locally modifying the laser power and scan rate) in a particular location by con-
trolling the size and cooling rate of the melt pool. The results from LENS consistently demonstrate
better metallurgical and mechanical properties than other conventional processes due to an improved
microstructure. For example, LENS-deposited Ti–6Al–4V has a yield strength, ultimate tensile
strength, and elongation higher than those of conventionally processed and annealed samples [77,90].
Any metallic, ceramic, or composite powder that is stable in a molten pool can be used for con-
struction of parts using LENS. In general, metals with high reflectivities and thermal conductivities
(such as gold and some alloys of aluminum and copper) are difficult to process. Most other metals
are quite straightforward to process, provided that there is proper atmospheric preparation to avoid
oxide formation. Generally, metallic materials that exhibit reasonably good weldability are easy to
process [77].
LENS process has some typical drawbacks. The surface finish is very poor and requires both polish-
ing and finish machining. It is a relatively expensive, inherently hazardous, and slow process because
of the laser. Argon consumption could be high as it is used for masking the weld zone from atmo-
sphere as well as for the powder delivery [81].
LENS has been used for fabricating fully dense chemically homogeneous equiatomic NiTi alloy sam-
ples showing higher Vickers microhardness compared to conventionally processed ones without any
undesirable phases [24]. As shown in Fig. 22, the grain size was not uniform over the entire sample,
especially in high-density samples. Samples with high porosity such as Fig. 22b, showed relatively
coarser grains when compared to samples with low porosity. It was found that the specific energy in-
put has a major influence on grain sizes [91]. It was already demonstrated that with lower laser power,
finer grain structures can be created which result in higher compressive strength (in the range of 890–
1050 MPa) [24].
934 M.H. Elahinia et al. / Progress in Materials Science 57 (2012) 911–946

Fig. 22. SEM microstructures showing grain size variation (a) 200 W, 10 mm/s, 15 g/min, 92% dense; (b) 200 W, 10 mm/s, 30 g/
min, 85% dense; and (c) 150 W, 20 mm/s, 30 g/min, 72% dense [91].

Porous NiTi alloys have already been fabricated using LENS with laser powers between 150 and
300 W, scan speeds between 5 and 25 mm/s, powder feed rates in the range of 15 and 38 g/min,
and scan spacing between 0.76 and 1.27 mm [92]. Laser processing did not induce any intermetallic
compound formation such as Ti2Ni, Ni4Ti3, and Ni3Ti, as shown in Fig. 23. Laser processed samples
showed both B2 and B190 peaks at room temperature with relatively higher amount of high temper-
ature B2 phase in products than that in as-received powder particles. In fact, high cooling rate (103–
105 K/s) of laser processing resulted in retention of more high-temperature B2 phase at room temper-
ature in laser processed samples. It is obvious that these high cooling rates are responsible for the fine
scale of dendritic/grain structure. However, partial melting of the powders near the pores can decrease
the solidification rates due to shallow temperature gradients and poor heat conduction, resulting in
coarser grain size. On the other hand, it is well known that rapid cooling can promote locking of dis-
location, and generation of point defects such as vacancies. These crystal defects along with residual
stresses may inhibit the reverse martensitic transformation by imposing frictional stresses on the
martensite/parent interface. Therefore to have 1WE in the final products transformation temperatures
 M.H. Elahinia et al. / Progress in Materials Science 57 (2012) 911–946 935

Fig. 23. XRD patterns of as-received powder and LENS-processed porous NiTi sample [92].

Fig. 24. The modulus of porous samples processed using LENS [93].

As and Af must increase. This is because the presence of thermally induced defects increases the As and
Af while decreasing the martensite transformation temperature upon cooling. However, it was ob-
served that Ms temperature increased. Although the reason for the increase in Ms temperature is
not clear it could be due to incomplete reverse transformation (B190 ? B2) during heating. The trans-
formation sequence in laser-processed samples was one-stage (B190 ? B2) on heating and (B2 ? B190 )
on cooling. However, the transformation sequence for as-received powder showed R phase on heating
(B190 ? R ? B2) and cooling (B2 ? R ? B190 ) [24,91,92].
LENS provides more flexibility for designers to tailor the modulus of porous samples without
changing their bulk density or total pore volume [92]. NiTi samples with lower Young’s modulus of
approximately 11 GPa with total porosity of 16% has been made along with sample with modulus
of approximately 15 GPa and identical porosity and equiaxed pores.
The experimental data shown in Fig. 24 indicate that the modulus of LENS-processed samples can
be tailored by changing the LENS process parameters. Since the open porosity is important for tissue
in-growth, the amount of open pore volume in the NiTi samples was evaluated. Maximum open pore
volume was observed at a scan speed of 15 mm/s with a powder feed rate of 20 g/min and 200 W laser
power. No increase in the open pore volume was observed beyond these values. As shown in Fig. 25d,
highly interconnected porosity with low density can be obtained in samples processed at low laser
power, high powder feed rate and high scan speed. Porous NiTi alloy samples with 12–36% porosity
936 M.H. Elahinia et al. / Progress in Materials Science 57 (2012) 911–946

Fig. 25. Typical as-polished micrographs of porous Nitinol samples showing variation in pore connectivity (a) 200 W, 10 mm/s,
15 g/min; (b) 200 W, 10 mm/s, 30 g/min; (c) 200 W, 20 mm/s, 15 g/min; and (d) 150 W, 20 mm/s, 30 g/min [91].

Fig. 26. Different units of the EBM machine [94].

exhibited low Young’s modulus between 2 and 18 GPa as well as high compressive strength and up to
4% recoverable strain. The modulus of these porous NiTi samples, having high open porosities between
53% and 72% of total volume fraction porosities, is almost the same as that of human cortical bone.
 M.H. Elahinia et al. / Progress in Materials Science 57 (2012) 911–946 937

Fig. 27. Schematic of an EBM apparatus (courtesy of Arcam).

Table 6
Differences between EBM and SLM [77].

Characteristic Electron beam melting Selective laser melting

Thermal source Electron beam Laser
Atmosphere Vacuum Inert gas
Scanning Deflection coils Galvanometers
Energy absorption Conductivity-limited Absorptivity-limited
Powder pre-heating Use electron beam Use infrared heaters
Scan speeds Very fast, magnetically-driven Limited by galvanometer inertia
Energy costs Moderate High
Surface finish Moderate to poor Excellent to moderate
Feature resolution Moderate Excellent
Materials Metals (conductors) Polymers, metals and ceramics

Therefore, not only is the stress-shielding reduced but also the osteoblast/osteoclast-implant interac-
tion is significantly improved [91,93].
At an applied total strain of 2%, a value not high enough to produce permanent slip, no perma-
nent deformation was observed in majority of porous samples. Therefore, the 1WE behavior is
completely preserved. Increasing the total strain from 2% to 6% decreased the recoverable deforma-
tion, a minimum of 70% of applied deformation. Samples with higher density showed high recov-
erable deformation. NiTi samples with density >90% showed a maximum recoverable strain of 6%.
The decrease in the recoverable strain with increasing porosity could be due to concentration and
localization of strains at the pore walls, which can be significantly greater than macroscopically
applied total strain [91].
In Electron Beam Melting (EBM) a part is manufactured by melting metal powders layer by layer
using an electron beam. The entire process takes place at temperatures around 1000 °C and in a high
vacuum at 104 mbar in the chamber and 106 mbar in the electron beam producing gun. This makes
EBM suitable for materials with a very high oxygen affinity. Arcam AB (Mölndal, Sweden) developed
938 M.H. Elahinia et al. / Progress in Materials Science 57 (2012) 911–946

the EBM approach. The Arcam A1 machine, shown in Fig. 26, is designated for additive manufacturing
of orthopedic implants. It offers a beam power of 5–3000 W with a scan speed of up to 8000 m/s. Cur-
rently, EBM process is capable of producing parts up to 200  200  200 mm with a dimensional accu-
racy of 0.4 mm. The device uses a thermionic emission gun that utilizes a Tungsten filament to make
an electron beam, which selectively melts the metal powders with a thickness of 0.07–0.25 mm [94–
97].
Similar to SLM, the EBM process scans a focused beam (in this case and electron beam) across a
thin layer of pre-laid powder, causing localized melting and solidification as per the slice cross-
section (Fig. 27). However, there are some differences between these processes as summarized
in Table 6 [77].
A noticeable feature of both EBM and SLM layer manufacture is the significant reduction in waste
material since the unused powder can be recycled. The lightly sintered powder surrounds the fabri-
cated part and helps support its downward facing surfaces during the construction process. This pow-
der is then broken down during the post-construction sifting process. This allows for most of the
unmelted powder to be recovered and reused. The elevated temperature involved in this operation
also helps reduce residual stresses between the cooling melt pool and previously solidified layers.
In addition, parts are fabricated in a vacuum chamber during the EBM process, which assures impu-
rity-free parts unaffected by oxygen and other chemical species available in the atmosphere. The
residual stresses and distortion are minimized due to vacuum processing. This leads to stress-relieved
parts with hardness and tensile strength higher than cast and comparable to wrought products. The
electron beam heats up the powder bed to an optimal temperature. Therefore, the parts are free from
residual stresses and do not suffer from distortion. Hardness, yield strength and ultimate tensile
strength for SLM products are higher than those for EBM products. Conversely, products made by
EBM have higher elongations than those obtained with SLM. The EBM process involves more signifi-
cant melting than the SLM process [78,96].
The electron beam as an energy source shows advantages compared to the laser beam; especially
high deflection speed realized by electromagnetic lenses, high energy input easily controlled beam fo-
cus and energy, as well as electrically controlled scan. This allows for variations in powder layer build-
ing, including a range of liquid phase sintering to complete particle melting or layer melting
[77,98,99]. Compared to laser beams, electron beams have a greater energy density which leads to re-
duced building times and consequently reduced manufacturing costs. The high energy of electron
beams cause higher temperatures of EBM powder bed and therefore more particle diffusion overlaying
fully melting the metal powder. This results in dense parts with better control of the mechanical prop-
erties and with lesser porosity. Unlike other additive manufacturing (AM) methods, which rely on la-
ser sintering, the EBM process is capable of producing fully melted, void-free, and fully-dense parts. All
other laser sintering-based methods struggle with density because of the relatively low power of the
laser as compared to the electron beam and the fact that they are lower temperature processes
[97,100]. With laser-based systems, 95% of the light energy is reflected by the powder rather than
being absorbed which significantly reduces the efficiency. The higher efficiency of EBM results in high-
er build rate, 3–5 times faster than other AM methods due to increased penetration depth and ele-
vated scanning velocities [77,79,101].
The mechanical properties of EBM parts showed that the EBM process is suitable for the fabrication
of porous parts with mechanical strength equivalent to parts built from conventional casting pro-
cesses. With EBM, it may be possible to selectively optimize the alloy strength, fatigue resistance,
and elastic modulus for greater bone compatibility. Successful use of this technology for implant fab-
rication will result in customized features for faster rehabilitation, increased longevity, better func-
tionality, and cosmesis apart from reducing surgical time [79,97,98].
The power from an electron beam can also be used to produce NiTi ingots. As mentioned earlier, the
conventional Vacuum Induction Melting (VIM) process suffers from a high contamination of carbon
which comes from the graphite crucible and presence of oxygen from the remnant air in the melting
chamber. With electron beam melting, a water-cooled copper crucible is used which eliminates the
carbon contamination and oxygen contamination is minimized because of operation in high vacuum.
It was shown that the carbon contamination is four to ten times lower in EBM than VIM and the oxy-
gen content is dependent on the starting raw material [40,102].
 M.H. Elahinia et al. / Progress in Materials Science 57 (2012) 911–946 939

Fig. 28. Examples of biomedical Ti6Al4V implants produced by EBM: (a) solid (left) and meshed dental implants [76,68], (b)
polished knee implant [78], (c) hip stems [103].

The fabrication of implants from patient specific data with high precision (Accuracy of ±0.025 mm),
controlled mechanical properties, and precise adaptation to the region of implantation is made possi-
ble with EBM, eliminating expensive secondary processing as machining, forging, swaging, or forming
and related lead times, reducing pre-insertion procedures of implants and promoting faster healing
time [76,79].
Benefits in terms of speed, cost, and design complexity all promote this technology as a method
that will find continued application in implant manufacturing [97]. The EBM process is currently in
use for low-volume production of medical components in both Europe and the United States [96].
Harrysson et al. [103] used the EBM method to fabricate custom designed orthopedic components
(hip stems) from Ti6Al4V alloy such that they can tailor its mechanical properties and structural
strength to minimize the stress shielding effect and bone remodeling. As mentioned before, the mis-
match between the stiffness of the prosthetic implant and the patient’s bone results in stress shielding
which leads to bone resorption and loosening of the implant. The EBM method was shown to be a good
940 M.H. Elahinia et al. / Progress in Materials Science 57 (2012) 911–946

Table 7
Advantages and disadvantages of additive manufacturing (AM) processes.

AM Advantages Disadvantages
Process
SLS In case of using mixed powders low melting point Materials must be in powders, weak surface finish
powders act as a binder, capability of having versatile (powdery surface), post-processing is required (to
desgine for porous products [47,105] increase product density and improve mechanical
properties) [47,105]
SLM No distinct binder and melt phase (able to produce Not suited for well controlled compoite materials (e.g.
single material parts of polymers, metals and WC-CO), small scan velocity (long build time), low
ceramics) [47,77,88,105], no need to time consuming surface quiality of products, high amount of residual
and costly furnace post-processes, infiltration or post- stress (delamintion, distortion) [88], high energy level
sintering, manufacturing fully dense samples [88], needs for laser power (costly), difficulty of unbounded
high materials efficiency [78,96] powder removal from porous internal architecture
[47,105]
LENS Low processing time, low cost for modules and dies High energy level needs for laser power (costly),
[47,105], manufacturing products with good materials must be in powder [47,105]
mechanical properties [90]
EBM Fast scan speed, easily controlled electron beam, low Only for producing metallic product, poor surface
processing time [47,77,105], lower energy cost finish, low dimensional accuracy, costly (vaccuum
(compare to laser) [77], high materials efficiency, costs) [47,77,101,105], reple between powder
vacuum chamber assures stress-relieved and particles and incoming negatively charged electrons
impurity-free parts [78,96], high energy of electron themselves leads to sudden spreading of the powder
beam results in dense parts with low porosity and and also need to more diffuse beam [77,99]
good mechanical properties [97,100], precisely
controlled process eliminates expensive secondary
processing such as machining, forging, swaging, or
forming [68,79]

candidate in manufacturing custom designed implants such as holed or grooved hip stems, hip stems
with trabecular structure, or solid hip stems (see Fig. 28). Fully dense Ti6Al4V alloy with higher hard-
ness and tensile strength compare to wrought samples has been manufactured [76,98,104].
There are also many finishing methods used in producing the final implant product from these near
net shape raw material configurations. These manufacturing methods include (1) finish machining, (2)
finish grinding, (3) wire or sink EDM, (4) finish hand polishing, (5) finish grit blasting, (6) implant iden-
tification marking, and (7) cleaning and passivation processing [97].
As already mentioned, NiTi alloys have been previously produced using EB (furnace) melting meth-
od. It could be predicted that there will be many opportunities in area of producing error-free NiTi
products using EBM-AM process. Due to the previous experience and expertise in making Ti alloy im-
plants, it seems there will be increasing possibility in applying EBM-AM process as one of the methods
for producing biomedical NiTi implants in the future.
Advantages and disadvantages of all of the additive manufacturing processes covered in the paper
are summarized in Table 7.

8. Heat treating and shape setting

NiTi and other shape memory alloy products (e.g. bars, wires, ribbons, and sheets) are normally fin-
ished by cold working to achieve dimensional control and enhanced surface quality. Cold working
suppresses the shape memory response of these alloys. It also raises the strength and decreases the
ductility. However, cold working does not raise the stiffness of the material. Heat-treating after cold
working diminishes the effects of cold working and restores the shape memory response of SMAs.
Therefore, in order to optimize the physical and mechanical properties of a NiTi product and to achieve
desired shape memory and/or pseudoelasticity properties, the material is cold worked and heat-
treated.
 M.H. Elahinia et al. / Progress in Materials Science 57 (2012) 911–946 941

It is hypothesized that the increase in transformation temperatures (As and Af) of NiTi alloy sample
manufactured using AM technologies, comes from thermally induced stresses and defects, and high Ms
temperature is due to incomplete transformation during heating. Application of post-fabrication heat
treatments can tailor these temperatures to suit a desired application [24].
The product suppliers normally provide the material in the cold worked condition. The maximum
practical level of the cold work is limited by the alloy components and by the product section size.
Binary superelastic NiTi alloy fine wires with As in the range of 25 to +95 °C are typically supplied
with cold drawing reduction (after the final annealing) in the range of 30–50%. Larger drawing reduc-
tions are sometimes used for very fine wires. These same alloys are limited to about 30% maximum
cold reduction in the larger diameter bars. Binary NiTi alloys with very low As in the range of 50
to 60 °C do not sustain the higher levels of cold work without cracking.
Both superelastic and shape memory properties are optimized by cold working and heat treatment.
This thermomechanical process is applied to all NiTi alloys although different amounts of cold work
and different heat treatments may be used for different alloys and property requirements [106].
Shape setting is accomplished by deforming the NiTi part to the shape of the desired component,
constraining the NiTi by clamping, and then heat-treating. This is normally done on the materials in
the cold worked condition, for example in cold drawn wires. However, annealed wires may be
shape-set with a subsequent lower temperature heat treatment.
In shape-setting cold worked material, care must be taken to limit the deformation strain to pre-
vent cracking of the material. Another approach is to partially anneal the wire prior to shape setting.
Yet, another option is to shape set in incremental steps.
The electrical resistance of NiTi makes it a good candidate for heating by electric current. NiTi will
be oxidized when heat-treated in air. Therefore, surface requirements and atmosphere control are
important considerations.
A wide range of temperature from 300 to 900 °C can be chosen for the shape-setting of NiTi. How-
ever, in order to optimize the combination of physical and mechanical properties, heat treating tem-
peratures for binary NiTi alloys are usually chosen in the narrower range of 325–525 °C. Heat treating
times are typically 5–30 min. Consideration must be given to the mass of the heat-treating fixture as
well as that of the product. Sufficient time must be allowed in the furnace to get the entire mass to the
desired temperature.
In the case of shape-setting SMA wires, a tooling fixture made of stainless steel can be used to hold
the wire in a taut condition (see Fig. 29). It is customary to restrict the amount of the strain in the wire

Fig. 29. The shape setting mold for preparing a NiTi helix [34].
942 M.H. Elahinia et al. / Progress in Materials Science 57 (2012) 911–946

to 2% during shape setting so that the fatigue properties (e.g., crack initiation resistance) do not dimin-
ish. After heating, the wire contracts and gets tightened. Hence bolts and nuts with washers may be
used to clamp the wire. Tying or coiling the wire is not suggested. The protocol calls for placing the
fixture in a furnace and once it has reached the desired temperature, taking it out after the required
amount of time and quenching it [34].
The effects of time and temperature of shape setting on the shape recovery quality, austenite finish
temperature, and non-linear mechanical behavior of NiTi shape memory alloy was investigated by Liu
et al. [107]. The shape-setting process showed stable results where the NiTi wire was constrained and
treated on a mold at 500 °C for aging times of longer than 10 min. Af temperature increased with the
aging time and the peak values were obtained after aging at beyond 400 °C. The upper and lower pla-
teau stresses decreased with the aging time. The increase of Af with the aging time was consistent
with the expected decrease in the Ni content of the NiTi matrix (due to precipitation). At high temper-
atures, there is sufficient thermal energy to permit rapid diffusion of Ni and Ti atoms in the matrix; but
it becomes more difficult for the same atoms to form precipitate nuclei as the temperature increases.
Higher nucleation rates will occur at lower temperature levels but diffusion rates will be slower.
As mentioned earlier, the phase transformation temperatures of NiTi SMA depend on the chemical
composition, the amount of cold working, and the heat treatment processing parameters. The heat
treatment temperature and time as well as cooling rate can be significant parameters altering forward
and reverse phase transformations between austenite and martensite. Yeung et al. [108] investigated
the effects of these parameters for three stages of treatment processes: (1) solid solution, (2) solid
solution followed by aging at high temperatures, and (3) solid solution followed by aging at low tem-
peratures. Longer treatment times gave higher Ap (the peak temperature on the differential scanning
calorimeter reverse transformation plot, As < Ap < Af). The solid solution at different temperatures did
not have significant effect on Ap but the cooling rate and the secondary aging could considerably
change Ap. Secondary aging treatment between 400 and 480 °C could increase Ap but the time dura-
tion had a lesser effect. Water quenching generally gave the highest transformation temperature
change among other cooling methods.
Smith and Hodgson [109] reviewed the types of furnaces and fixture hardware or mandrels that
have been used for the heat treatment of NiTi. Many types of furnaces were used including box fur-
naces, continuous belt hearth furnaces, tube furnaces, heated platen presses, vacuum furnaces, induc-
tion heaters, salt baths, and fluidized bath furnaces. NiTi will be oxidized in a non-protected
environment. Therefore, care must be taken in selecting the type of the heat treatment furnace.

9. Conclusions

In this paper, manufacturing and processing of NiTi implants was reviewed. These processes range
from casting and powder metallurgy processes to machining and post-treatments. Casting processes
including VAR, VIM and EBM processes and their pros and cons were described. It was shown that,
NiTi samples could be produced containing the lowest amount of impurities using EBM process.
Among powder metallurgy processes, the most versatile and practical ones to make NiTi implants
were also discussed. Both conventional and additive manufacturing processes were reviewed. All the
advantages and disadvantages of the processes were described in comparison with each other.
Additionally, new opportunities in additive manufacturing processes towards making 3D compo-
nents form NiTi, using Laser or Electron beam were critiqued in detail, quoting practical examples. Fi-
nally, the challenges in heat treatment and shape-setting of NiTi parts in order to attain desired shape
memory properties were reviewed.

Acknowledgements

The authors would like to acknowledge the financial support provided for the project ‘‘Nitinol Com-
mercialization Accelerator’’ by the Ohio Department of Development through Grant WP 10-010. NSF
support though award 0731087 Research to Aid Person with Disability is also appreciated.
 M.H. Elahinia et al. / Progress in Materials Science 57 (2012) 911–946 943

References

[1] Patoor E, Lagoudas DC, Entchev PB, Brinson LC, Gao X. Shape memory alloys, Part I: General properties and modeling of
single crystals. Mech Mater 2006;38:391–429.
[2] Lagoudas DC, Entchev PB, Popov P, Patoor E, Brinson LC, Gao X. Shape memory alloys, Part II: Modeling of polycrystals.
Mech Mater 2006;38:430–62.
[3] Otsuka K, Wayman CM. Shape memory materials. Cambridge University Press; 1998. p. 72–78.D [chapter 3].
[4] Stoeckel. The shape memory effect: phenomenon, alloys and applications. In: Proceedings of shape memory alloys for
power systems. EPRI; 1995. p. 1–13.
[5] Khalil-Allafi J, Dlouhy A, Eggeler G. Ni4Ti3-precipitation during aging of NiTi shape memory alloys and its influence on
martensitic phase transformations. Acta Mater 2002;50:4255–74.
[6] Otsuka K, Ren X. Recent developments in the research of shape memory alloys. Intermetallics 1999;7:511–28.
[7] Liu Y, Humbeeck JV, Stalmans R, Delaey L. Some aspects of the properties of NiTi shape memory alloy. J Alloys Compd
1997;247:115–21.
[8] Elahinia MH, Ashrafiuon H. Nonlinear control of a shape memory alloy actuated manipulator. J Vib Acoust – Trans ASME
2002;124:566–75.
[9] Elahinia MH, Seigler T, Michael Leo DJ, Ahmadian M. Nonlinear stress-based control of a rotary SMA-actuated
manipulator. J Intell Mater Syst Struct 2004;15:495–508.
[10] Elahinia MH, Ahmadian M, Ashrafiuon H. Design of Kalman filter for rotary shape memory alloy actuators. J Smart Mater
Struct 2004;13:691–7.
[11] Elahinia MH, Koo J, Ahmadian M, Woolsey C. Backstepping control of an SMA-actuated robotic arm. J Vib Control
2005;11:407–29.
[12] Elahinia MH, Ashrafiuon H, Ahmadian M, Tan H. A temperature-based controller for a shape memory alloy actuator. J Vib
Acoust – Trans ASME 2005;127:285–91.
[13] Elahinia MH, Ahmadian M. An enhanced SMA phenomenological model. Part I. The shortcomings of the existing models. J
Smart Mater Struct 2005;14:1297–308.
[14] Elahinia MH, Ahmadian M. An enhanced SMA phenomenological model. Part II. The experimental study. J Smart Mater
Struct 2005;14:1309–19.
[15] Eshraghi M, Ashrafiuon H, Elahinia MH. Position control of a three-link shape memory alloy actuated arm. J Intell Mater
Syst Struct 2006;17:381–92.
[16] Elahinia MH, Ahmadian M. Application of the extended Kalman filter to control of a shape memory alloy arm. J Smart
Mater Struct 2006;15:1370–84.
[17] Tarkesh Esfahani E, Elahinia MH. Stable walking pattern for an SMA-actuated biped. IEEE/ASME Trans Mechatron
2007;12:534–41.
[18] Williams E, Elahinia M. An automotive SMA mirror actuator: modeling, design and experimental evaluation. J Intell Mater
Syst Struct 2008;19:425–1434.
[19] Williams E, Shaw G, Elahinia M. Control of an automotive shape memory alloy mirror actuator. Mechatronics
2010;20:527–34.
[20] Hadi A, Yousefi-Koma A, Moghadam M, Elahinia MH, Ghazavi A. Developing a novel SMA-actuated robotic module.
Sensors Actuat A 2010;162:72–81.
[21] Tarkesh Esfahani E, Elahinia MH. Developing an adaptive controller for a shape memory alloy walking assistive device. J
Vib Control 2010;16:1897–914.
[22] Kanjwal K, Yeasting R, Baptista C, Elsamaloty H, Sheikh M, Elahinia M, et al. Retro-cardiac esophageal mobility and
deflection to prevent thermal injury during atrial fibrillation ablation: an anatomic feasibility study. J Interv Cardiac
Electrophysiol 2011;30:45–53.
[23] M. Tabesh, Finite element analysis of shape memory alloy biomedical devices. M.Sc. thesis, University of Toledo, Toledo,
Ohio; 2010.
[24] Krishna BV, Bose S, Bandyopadhyay A. Laser processing of net-shape NiTi shape memory alloy. J Metal Mater Trans A
2007;38:1096–103.
[25] Krishna BV, Bose S, Bandyopadhyay A. Fabrication and characterization of porous Ti6Al4V parts for biomedical
applications using electron beam melting process. J Acta Biomater 2007;3:997–1006.
[26] Shishkovsky IV, Volova LT, Kuznetsov MV, Morozov YuG, Parkin IP. Porous biocompatible implants and tissue scaffolds
synthesized by selective laser sintering from Ti and NiTi. J Mater Chem 2008;18:1309–17.
[27] Es-Souni Mo, Es-Souni Ma, Fischer-Brandies H. Assessing the biocompatibility of NiTi shape memory alloys used for
medical applications. J Anal Bioanal Chem 2005;381:557–67.
[28] http://jmmedical.com/resources/231/Comparison-of-Properties-of-NiTi-and-Stainless-Steel.html.
[29] Shishkovsky I, Morozov Yu, Smurov I. Nanofractal surface structure under laser sintering of titanium and nitinol for bone
tissue engineering. J Appl Surf Sci 2007;254:1145–9.
[30] Bram M, Ahmad-Khanloua A, Heckmannb A, Fuchsa B, Buchkremera HP, Stovera D, et al. Powder metallurgical fabrication
processes for NiTi shape memory alloy parts. Mater Sci Eng A 2002;337:254–63.
[31] Pelton AR, DiCello J, Miyazaki S. Optimization of processing and properties of medical-grade NiTi wire. Min Invas Ther
Allied Technol 2000;9:107–18.
[32] Funakubo H, editor. Shape memory alloys. New York: Gordon and Breach Publishers; 1987.
[33] Frenzel J, Zhang Z, Neuking K, Eggeler G. High quality vacuum induction melting of small quantities of NiTi shape memory
alloys in graphite crucibles. J Alloys Compd 2004;28:214–23.
[34] Morgan NB, Broadley M. Taking the art out of smart!—forming processes and durability issues for the application of NiTi
shape memory alloys in medical devices. In: Shrivastava S, editor. Proceedings of the materials and processes for medical
devices conference. ASM International; 2003. p. 247–52.
944 M.H. Elahinia et al. / Progress in Materials Science 57 (2012) 911–946

[35] Reinoehl M, Bradley D, Bouthot R, Proft J. The influence of melt practice on final fatigue properties of superelastic NiTi
Wires. In: Proceedings of international conference on shape memory and superelasticity technologies, Pacific Grove,
California; 2000.
[36] ASTM standard specification F2063-00. American society for testing and materials; 2000.
[37] Otubo J, Rigo OD, Neto CDM, Kaufman MJ, Mei PR. NiTi shape memory alloy ingot production by EBM. J Phys IV
2003;112:813–20.
[38] Otubo J, Mei PR, Koshimizu S, Martinez LG. Niti shape memory alloys produced by electron bean melting: preliminary
results. JOM, USA 1998;1:1063–8.
[39] Otubo J, Rigo OD, Neto CDM, Kaufman MJ, Mei PR. Low carbon content NiTi shape memory alloy produced by electron
beam melting. J Mater Res 2004;7:263–7.
[40] Otubo J, Rigo OD, Neto CDM, Kaufman MJ, Mei PR. Scale up of NiTi shape memory alloy production by EBM. J Phys IV
2003;112:873–6.
[41] Weinert K, Petzoldt V. Machining of NiTi based shape memory alloys. Mater Sci Eng A 2004;378:180–4.
[42] Stoeckel D. Forming NiTi – a challenge. In: Siegert K, editor. New developments in forging technology; 2001. p. 119–34.
[43] Michiardi A, Aparicio C, Planell JA, Gil FJ. Electrochemical behaviour of oxidized NiTi shape memory alloys for biomedical
applications. Surf Coat Technol 2007;201:6484–8.
[44] Poona RWY, Hoa JPY, et al. Improvements of anti-corrosion and mechanical properties of NiTi orthopedic materials by
acetylene, nitrogen and oxygen plasma immersion ion implantation. Nucl Instrum Methods Phys Res Sect B: Beam
Interact Mater Atoms 2005;237:411–6.
[45] Wu MH. Fabrication of NiTi materials and components. In: Proceedings of the International conference on shape memory
and superelastic technologies, Kunming, China; 2001. p. 285–92.
[46] Lin HC, Lin KM, Cheng IS. The electro-discharge machining characteristics of NiTi shape memory alloys. J Mater Sci
2001;36:399–404.
[47] Alvarez K, Nakajima H. Metallic scaffolds for bone regeneration. J Mater 2009;2:790–832.
[48] Clare AT, Chalker PR, Davies S, Sutcliffe ChJ, Tsopanos S. Selective laser melting of high aspect ratio 3D nickel–titanium
structures two way trained for MEMS applications. Int J Mech Mater Des 2008;4:181–7.
[49] Hang NZ, Khosrovabadi PB, Lindenhovius JH, Olster BHK. Assessment of the duration of time required for conventional Ti–
Ni sintering. J Phys IV 1991;1:163–8.
[50] Johnson WA, Domingue JA, Reichman SH. P/M processing and characterization of controlled transformation temperature
NiTi. J Phys Colloques 1982;43. C4-285–90.
[51] Wipf H, Dietz M, Aslanisdis D, Serneels A, Mooreleghem WV. In: Proceedings of international conference on shape
memory and superelasticity technologies, Antwerp, Belgium; 1999. p. 375.
[52] da Silva MCA, da Silva LC, Filho FA. High energy milling and hot extrusion of equiatomic NiTi shape memory alloy. Adv Sci
Technol 2008;59:18–23.
[53] Li BY, Rong LJ, Li YY. Porous NiTi alloy prepared from elemental powder sintering. J Mater Res 1998;13:2847–51.
[54] Zanottia C, Giuliania P, Terrosua A, Gennarib S, Magliab F. Porous Ni–Ti ignition and combustion synthesis. Intermetallics
2007;15:404–12.
[55] Li B, Ronga L, Lia Y, Gjunterb VE. A recent development in producing porous Ni–Ti shape memory alloys. Intermetallics
2000;8:881–4.
[56] Ying Z, Minoru T, Yansheng K, Akira K. Compression behavior of porous NiTi shape memory alloy. Acta Mater
2005;53:337–43.
[57] Aust E, Limberg W, Gerling R, Oger B, Ebel T. Advanced TiAl6Nb7 bone screw implant fabricated by metal injection
moulding. Adv Eng Mater 2006;8:365–70.
[58] Benson JM, Chikwanda HK. Challenges of titanium metal injection moulding. In: Transportation weight reduction: 8th
annual international RAPDASA conference, Tshwane University of Technology and Pilanesburg, South Africa; 7–9
November 2007. p. 1–11.
[59] Yuan B, Chung CY, Zhang XP, Zeng MQ, Zhu M. Control of porosity and superelasticity of porous Niti shape memory alloys
prepared by hot isostatic pressing. Smart Mater Struct 2005;14:S201–6.
[60] Li BY, Rong LJ, Gjunter VE, Li YY. Porous Ni–Ti shape memory alloys produced by two different methods. Z Metallkd
2000;91:291–5.
[61] Greiner C, Oppenheimer SM, Dunand DC. High strength, low stiffness, porous NiTi with superelastic properties. Acta
Biomater 2005;1:705–16.
[62] Li BY, Rong LJ, Li YY, Gjunter VE. Synthesis of porous Ni–Ti shape-memory alloys by self-propagating high-temperature
synthesis: reaction mechanism and anisotropy in pore structure. Acta Mater 2000;48:3895–904.
[63] Biswas A. Porous NiTi by thermal explosion mode of SHS: processing, mechanism and generation of single phase
microstructure. Acta Mater 2005;53:4015–25.
[64] Lagoudas DC, Vandygriff EL. Processing and characterization of NiTi porous SMA by elevated pressure sintering. J Intell
Mater Syst Struct 2002;13:837–50.
[65] Krone L, Mentz J, et al. 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 2005;7:613–9.
[66] Krone L, Schuller E, Bram M, Hamed O, Buchkremer H, Stover D. Mechanical behaviour of Niti parts prepared by powder
metallurgical methods. Mater Sci Eng A 2004;378:185–90.
[67] http://www.arburg.com/com/COM/en/products/machines/standard/index.jsp.
[68] Upadhyaya GS. Powder metallurgy technology. Cambridge International Science Publishing; 2002. p. 68–117 [chapters 6
and 7].
[69] L.E. Penrod, Fabrication and characterization of porous shape memory alloys. Master thesis, Texas A&M University; 2003.
[70] Xinghong Z, Chuncheng Z, Wei Q, Xiaodong H, Kvanin VL. Selp-propagating high temperature synthesis of TiC/TiB2
ceramic–matrix composites. Compos Sci Technol 2002;62:2037–41.
[71] Bansiddhi A, Sargeant TD, Stupp SI, Dunand DC. Porous NiTi for bone implants: a review. Acta Biomater 2008;4:773–82.
 M.H. Elahinia et al. / Progress in Materials Science 57 (2012) 911–946 945

[72] Li J, Chen F, Shen Q, Jiang H, Zhang L. Fabrication and dielectric properties of Si3N4–MgO–Al2O3 by spark plasma sintering
technique. Mater Sci – Poland 2007;25:699–707.
[73] http://www.mps-us.com/SPS.html.
[74] Dobedoe RS, West GD, Lewis MH. Spark plasma sintering of ceramics. Bull Eur Ceram Soc 2003;1:19–24.
[75] Mentz J, Bram M, Buchkremer HP, Stover D. Improvement of mechanical properties of powder metallurgical NiTi shape
memory alloys. Adv Eng Mater 2006;8:247–52.
[76] Chahine G, Koike M, Okabe T, Smith P, Kovacevic R. The design and production of Ti–6A1–4V ELI customized dental
implants. JOM 2008;60:50–5.
[77] Gibson I, Rosen DW, Stucker B. Additive manufacturing technology: rapid prototyping to direct digital manufacturing,
Springer Science + Business Media, Inc.; 2010, p. 103–42 [chapter 5].
[78] Murr LE, Quinones SA, Gaytan SM, Lopez MI, Rodela A, Martinez EY, et al. Microstructure and mechanical behavior of Ti–
6Al–4V for biomedical applications produced by rapid-layer-based manufacturing. J Mech. Behav Biomed Mater
2009;2:20–32.
[79] Parthasarathy J, Starly B, Raman Sh, Christensen A. Mechanical evaluation of porous titanium (Ti6Al4V) Structures with
electron beam melting (EBM). J Mech Behav Biomed Mater 2010;3:249–59.
[80] Mullen L, Stamp RC, Brooks WK, Jones E, Sutcliffe ChJ. Selective laser melting: a regular unit cell approach for the
manufacture of porous titanium, bone in-growth constructs. J Biomed Mater Res Part B 2009;89:325–34.
[81] Bernard A, Taillandier G, Karunakaran KP. Evolutions of rapid product development with rapid manufacturing: concepts
and applications. Int J Rapid Manuf 2009;1:3–18.
[82] Shishkovsky I, Morozov Y, Smurov I. Nanostructural self-organization under selective laser sintering of exothermic
powder mixtures. Appl Surf Sci 2009;255:5565–8.
[83] Shishkovsky I. Shape memory effect in porous volume NiTi articles fabricated by selective laser sintering. Tech Phys Lett
2005;31:186–8.
[84] Shishkovsky I, Kuznetsov MV, Morozov Y. Porous titanium and nitinol implants synthesized by SHS/SLS: microstructural
and histomorphological analyses of tissue reactions. Int J Self Propag High Temp Synth 2010;19:157–67.
[85] Shishkovsky I, Tarasova EY, Zhuravel LV, Petrov AL. The synthesis of a biocomposite based on nickel titanium and
hydroxyapatite under selective laser sintering conditions. Tech Phys Lett 2001;27:211–3.
[86] Shishkovsky I, Morozov Y, Smurov I. Nanofractal surface structure under laser sintering of titanium and nitinol for bone
tissue engineering. Appl Surf Sci 2007;254:1145–9.
[87] Shishkovsky IV, Volova LT, Kuznetsov MV, Morozovc YG, Parkin IP. Porous biocompatible implants and tissue scaffolds
synthesized by selective laser sintering from Ti and NiTi. J Mater Chem 2008;18:1309–17.
[88] Kruth JP, Levy G, Klocke F, Childs THC. Consolidation phenomena in laser and powder-bed based layered manufacturing. J
CIRP 2007;56:730–59.
[89] Vandenbroucke B, Kruth JP. Selective laser melting of biocompatible metals for rapid manufacturing of medical parts.
Rapid Prototyp J 2007;13:196–203.
[90] http://www.optomec.com/site/technology_lens.
[91] Krishna BV, Bose S, Bandyopadhyay A. Fabrication of porous NiTi shape memory alloy structures using laser engineered
net shaping. J Biomed Mater Res B: Appl Biomater 2009;89B:481–90.
[92] Bandyopadhyay A, Krishna BV, Xue W, Bose S. Application of laser engineered net shaping (LENS) to manufacture porous
and functionally graded structures for load bearing implants. J Mater Sci Mater Med 2009;20:S29–34.
[93] Krishna BV, Xue W, Bose S, Bandyopadhyay A. Developments in titanium: engineered porous metals for implants. JOM
2008;60:45–8.
[94] http://www.arcam.com/technology/ebm-hardware.aspx.
[95] Wohlers T, Grimm T. A year filled with promising R&D. Time-Compres Technol 2002.
[96] Mazzoli A, Germani M, Raffaeli R. Direct fabrication through electron beam melting technology of custom cranial implants
designed in a PHANToM-based haptic environment. J Mater Des 2009;30:3186–92.
[97] Christensen A, Lippincott A, Kircher R. Qualification of electron beam melted (EBM) Ti6Al4V-ELI for orthopaedic implant
applications. In: Medical device materials IV: proceedings from the materials & processes for medical devices conference,
Palm Desert, CA/USA; 2007. p. 48–53.
[98] Murr LE, Esquivel EV, Quinones SA, Gaytan SM, Lopez MI, Martinez EY, et al. Microstructures and mechanical properties of
electron beam-rapid manufactured Ti–6Al–4V biomedical prototypes compared to wrought Ti–6Al–4V. J Mater Charact
2009;60:96–105.
[99] Kahnert M, Lutzmann S, Zaeh MF. Layer formations in electron beam sintering. In: 18th Solid freeform fabrication
symposium proceedings, The University of Texas at Austin/USA; 2007.
[100] Xie OH. Design, simulation and experimental study of shape memory alloy and micro-motor activated high pressure
optical cell for bio-physical studies. PhD dissertation, Drexel University; 2007.
[101] Zah MF, Lutzmann S. Modelling and simulation of electron beam melting. Prod Eng Res Dev 2010;4:15–23.
[102] Otubo J, Rigo OD, Moura NC, Meib PR. The effects of vacuum induction melting and electron beam melting techniques on
the purity of NiTi shape memory alloys. Mater Sci Eng A 2006;438–440:679–82.
[103] Harrysson O, Cansizoglu O, Marcellin-Little D, Cormier D, West H. Direct metal fabrication of titanium implants with
tailored materials and mechanical properties using electron beam melting technology. Mater Sci Eng C 2008;28:366–73.
[104] Facchini L, Magalini E, Robotti P, Molinari A. Microstructure and mechanical properties of Ti–6Al–4V produced by
electron beam melting of prealloyed powders. Rapid Prototyp J 2009;15:171–8.
[105] Eyers DR, Dotchev KD. Rapid manufacturing for mass customization. In: IPROMS 2009 conference, Cardiff University,
Wales, UK; 2009
[106] Christensen A, Lippincott A, Kircher R, Shape memory alloys: modeling and engineering applications. In: Lagoudas
Dimitris C, editor. Springer; 2008. http://www.medicalmodeling.com/ebm/.
[107] Liu X, Wang Y, Yang D, Qi M. The effect of ageing treatment on shape-setting and superelasticity of a NiTi stent. Mater
Charact 2008;59:402–6.
946 M.H. Elahinia et al. / Progress in Materials Science 57 (2012) 911–946

[108] Yeung KWK, Cheung KMC, Lu WW, Chung CY. Optimization of thermal treatment parameters to alter austenitic phase
transition temperature of NiTi alloy for medical implant. Mater Sci Eng A 2004;383:213–8.
[109] Smith SA, Hodgson DE. Shape setting nitinol. In: Medical device materials: proceedings from the materials & processes for
medical devices conference. Anaheim (CA/USA): ASM International; 2003. p. 266–70.
[110] Frenzel J, Zhang Z, Somsen Ch, Neuking K, Eggeler G. Influence of carbon on martensitic phase transformations in NiTi
shape memory alloys. Acta Mater 2007;55:1331–41.
[111] Frenzel J, George EP, Dlouhy A, Somsen Ch, Wagner MF-X, Eggeler G. Influence of Ni on martensitic phase transformations
in NiTi shape memory alloys. Acta Mater 2010;58:3444–58.
[112] ASM Handbook. Powder metal technologies and applications, vol. 7. ASM International; 1998. p. 308–2096.