Progress in Materials Science: J.P. Oliveira, R.M. Miranda, F.M. Braz Fernandes

Progress in Materials Science 88 (2017) 412466
Contents lists available at ScienceDirect
Progress in Materials Science

journal homepage: www.elsevier.com/locate/pmatsci
Welding and Joining of NiTi Shape Memory Alloys: A Review

J.P. Oliveira a,b,, R.M. Miranda c, F.M. Braz Fernandes a
a
CENIMAT/I3N, Departamento de Cincias dos Materiais, Faculdade de Cincias e Tecnologia, Universidade Nova de Lisboa, 2829-516 Caparica, Portugal
b
Department of Materials Science and Engineering, The Ohio State University, 1248 Arthur E. Adams Drive, Columbus, OH 43221, USA
c
UNIDEMI, Departamento de Engenharia Mecnica e Industrial, Faculdade de Cincias e Tecnologia, Universidade Nova de Lisboa, 2829-516 Caparica, Portugal
a r t i c l e i n f o a b s t r a c t
Article history: NiTi is an increasingly applied material in industrial applications. However, the difficulties
Received 15 March 2017 faced when welding and joining is required, limits its broader use in the production of
Accepted 6 April 2017 complex shaped components. The main weldability problems associated with NiTi are:
Available online 19 April 2017
strength reduction, formation of intermetallic compounds, modification of phase transfor-
mation and transformation temperatures, as well as, changes in both superelastic and
Keywords: shape memory effects. Additionally, NiTi is envisaged to be joined to other materials, in dis-
NiTi shape memory alloys
similar joints with more complex problems depending on the other base material. Thus,
Welding
Joining
intensive research in welding and its effects on the joints performance has been conducted
Dissimilar joints since the early stages of NiTi. This paper presents a detailed review of welding and joining
Superelasticity processes applied to NiTi, in similar and dissimilar combinations considering both fusion
Shape memory effect and solid-state processes. Since laser is the most studied and applied welding process, a
special section is devoted to this technique.
2017 Elsevier Ltd. All rights reserved.
Contents
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 413
2. Welding and joining technologies for NiTi . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 418
2.1. Major developments in welding technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 418
2.2. Fusion-based welding processes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 419
2.2.1. Tungsten inert gas (TIG) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 419
2.2.2. High energy density beams . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 420
2.2.3. Resistance welding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 421
2.3. Solid-state processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 422
2.3.1. Explosive welding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 422
2.3.2. Friction welding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 423
2.3.3. Ultrasonic welding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 423
Abbreviations: A, austenite; BM, base material; EDS, energy dispersive spectroscopy; FIB, focused ion beam; FSW, friction stir welding; FZ, fusion zone;
GTAW, gas tungsten arc welding; HAZ, heat affected zone; HI, heat input; M, martensite; MIG, metal inert gas; MAG, metal active gas; MRW, micro-
resistance welding; Nd:YAG, neodymium-doped yttrium aluminium garnet; R, crystal growth rate; T, temperature gradient; SEM, scanning electron
microscopy; SMAs, shape memory alloys; TEM, transmission electron microscopy; TIG, tungsten inert gas; UTS, ultimate tensile strength.
Corresponding author at: Department of Materials Science and Engineering, The Ohio State University, 1248 Arthur E. Adams Drive, Columbus, OH
43221, USA.
E-mail addresses: [email protected], [email protected] (J.P. Oliveira).
http://dx.doi.org/10.1016/j.pmatsci.2017.04.008
0079-6425/ 2017 Elsevier Ltd. All rights reserved.
J.P. Oliveira et al. / Progress in Materials Science 88 (2017) 412466 413
2.3.4. Adhesives joining . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 424

2.3.5. Brazing and soldering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 425
2.3.6. Diffusion and reactive diffusion bonding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 427
3. Laser welding of NiTi . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 428
3.1. Similar laser welding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 431
3.1.1. Microstructure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 431
3.1.2. Mechanical behaviour . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 432
3.1.3. Shape memory effect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 439
3.1.4. Phase transformation temperatures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 442
3.1.5. Corrosion resistance. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 444
3.1.6. Post-weld heat treatments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 446
3.1.7. Filler materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 448
3.2. Laser welding of dissimilar joints involving NiTi . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 450
3.2.1. NiTi/steel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 450
3.2.2. NiTi/Ti6Al4V . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 454
3.2.3. NiTi/other engineering materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 456
4. Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 457
5. Summary and future outlook . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 458
Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 459
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 459
1. Introduction
The demand for shape memory alloys (SMAs) has increased in recent years due to their functional properties, namely
shape memory effect and superelasticity, present unique solutions for state-of-the-art applications. As a consequence of
intensive research activity, shape memory alloys have been used in a wide variety of fields, mainly biomedical, automotive
and aeronautical, taking advantage of their smart functionalities and a good actuation force-to-weight ratio [1].
Shape memory alloys are functional materials that present two very distinct properties: superelasticity and shape mem-
ory effect. Among these alloys, NiTi is the most important one, not only because of its functional properties, but also because
it presents high strength and ductility [2].
The functional properties exhibited by these alloys are due to a reversible martensitic transformation. SMAs have two
phases: a low temperature one, known as martensite (M), and a high temperature one, known as austenite (A) or parent
phase.
Fig. 1. Typical superelastic behaviour of a shape memory alloy [6].

414 J.P. Oliveira et al. / Progress in Materials Science 88 (2017) 412466
Fig. 2. Stress-strain-temperature plot exhibiting superelasticity and shape memory effect [8].
When cooling the material from the austenitic domain, it starts to transform into martensite at a given Ms temperature.
This austenite to martensite transformation (referred in the literature as forward or direct transformation), finishes at the Mf
temperature.
Starting with the material in the low temperature phase, by heating it up to a given As temperature, austenite starts to
form. Such transformation from martensite to austenite by heating (referred in the literature as reverse transformation) fin-
ishes when the Af temperature is reached. Ms, Mf, As and Af, are designated as the transformation temperatures of a given
shape memory alloy, where M and A designate the phase (martensite or austenite) and the subscripts s and f apply
for start and finish of transformation, respectively.
Martensitic transformations, as those taking place in shape memory alloys, are diffusionless transformations [3]. When
such transformations occur only by effect of temperature, the macroscopic shape change is negligible.
In these alloys, each martensitic crystal has a given different orientation, called variant. An assembly of martensitic vari-
ants may exist in two distinct forms: twinned martensite or detwinned martensite. The former is created by the combination
of self-accommodated martensitic variants to keep the overall shape when martensite is thermally induced, while in the lat-
ter a specific variant is dominant in the system, usually as a result of an applied external load [4].
The reversibility of transformation between austenite and martensite, and vice versa, is the basis of the peculiar beha-
viour of this class of alloys.
The martensitic transformation may also be stress-induced, giving rise to superelastic effect. Superelasticity occurs by the
application of stress when the material is in the temperature range of thermally stable austenite. Such property allows for
the material to undergo a significant deformation during loading (up to 10% strain) with full recovery of its shape upon
unloading [5], as shown in Fig. 1. Consider a given fixed temperature, above AF, where the material is tested under an applied
uniaxial tensile load. Firstly, there is elastic deformation of austenite (along the AB path in Fig. 1). At B, for a given stress
rMs, the martensitic transformation starts to occur. Such stress induced transformation from austenite to martensite occurs
Fig. 3. Crystal structures of austenite and martensite.

Fig. 4. Phase diagram of Ti-Ni [2].
at an approximately constant stress level, up to a rMf stress. When rMf is reached, the material is fully martensitic. When the
transformation from austenite to martensite occurs, this martensite may turn instantly into detwinned martensite, depend-
ing on the relative orientation of the crystallographic variants and the imposed external solicitation. From C to D, elastic
deformation of the detwinned martensite occurs. If the material is not strained beyond point D, in theory, full recovery of
the imposed strain can be obtained. From D to E the martensite elastically unloads. At point E, for a rAs stress level the
martensite to austenite transformation starts to occur and proceeds until it is finished (at point F). Complete removal of
the imposed stress allows for the elastic unloading of austenite (along the path FA).
Opposite to superelasticity, the shape memory effect occurs starting from the material in the martensitic state [7]. Con-
sider a shape memory alloy with a given form in the austenitic phase. Cooling it down so that martensite (in the twined
state) is stable, will not promote any shape alteration in the material. However, if a given deformation is imposed to the
material in the martensitic state, it will be retained by the material, as long as it remains in the temperature range where
martensite is the stable phase. As a consequence of the applied deformation, martensite detwinning occurs. It is possible
to recover the original shape when the material is heated above the Af temperature. When the As temperature is crossed,
the shape memory effect starts to occur, that is, the material starts to recover its initial shape, and this proceeds until the
Af temperature is reached, when full recovery of the imposed deformation has occurred. Further cooling into the martensite
domain will occur with no significant macroscopic shape change of the material. Fig. 2 depicts the schema of both supere-
lastic and shape memory effects in a stress-strain-temperature diagram.
Though superelasticity and shape memory effect have been discussed separately, they are closely related. It is expected
that any shape memory alloy exhibits both superelastic and shape memory properties, depending on the temperature at
which they are tested and as long as the critical stress for dislocation slip is not reached.
The critical stress for stress-induced martensite follows a Clausius-Clapeyron relationship [9]: dr=dT DS=e (being r
the applied stress, T the temperature, DS the variation in entropy and e the applied strain). Increasing the temperature, a
higher stress is required to trigger the stress-induced transformation. On the other hand, the critical stress to induce slip,
and therefore plastic deformation, decreases with increasing temperature. As such, it is possible that, at a given temperature,
the critical stress for dislocation slip occurs earlier than the superelastic effect, thus inhibiting this property. The temperature
above which the stress-induced martensitic transformation is no longer favored is designated by Md. However, it is possible
to manipulate the critical stress for inducing dislocations slip at a given temperature: by softening or hardening the material,
the critical stress for slip decreases or increases, respectively.
Another functional property exhibited by these materials is the two-way shape memory effect [4]. When the material is
trained to have this property, it remembers a given geometrical shape in the parent phase and another in the martensitic
phase. These two shapes are obtained without the help of any mechanical loading.
Fig. 5. Variation of the Ms temperature as a function of the Ni content in several NiTi alloys [2].
Besides austenite (with a B2 structure of cubic symmetry) and martensite (with a B190 monoclinic structure), NiTi may
present another phase: R-phase, with a trigonal structure, that may exist between austenite and martensite in special cir-
cumstances, as it will be discussed latter. The crystal structures of both austenite and martensite are depicted in Fig. 3.
The crystal structure of R-phase can be viewed as a stretch of the austenite crystal structure along the h1 1 1i direction.
The Ti-Ni phase diagram is depicted in Fig. 4. It can be seen that NiTi, above a certain temperature, is not a line compound
with a fixed composition, but rather has a certain solubility, that is much larger on the Ni side than on the Ti. Other com-
pounds, namely precipitates Ti2Ni, Ni3Ti, Ni3Ti2 and Ni4Ti3 are also observed for this system.
Extensive studies [2,1013] have been performed studying the precipitation phenomena for NiTi shape memory alloys.
The precipitates Ni4Ti3, Ni3Ti2, Ni3Ti can form when the NiTi alloy is slightly Ni-rich. In opposition, for Ti-rich NiTi alloys just
Ti2Ni precipitates may exist.
As a consequence of these precipitation phenomena, namely of Ni4Ti3, the intermediary phase of the NiTi system, R-
phase, can be formed. When Ni4Ti3 is coherent with the matrix, or in the presence of dislocation networks, a thermoelastic
transformation, which is a transformation that proceeds by changing the stress or temperature between austenite and R-
phase, is observed. Such R-phase transformation, when present, is always an intermediate step for the austenite M marten-
site transformation.
Fig. 6. Number of published papers on joining processes for NiTi shape memory alloys over time.
The matrix composition of a given NiTi alloy influences its transformation temperatures and, thus, the functional prop-
erties at a given temperature. The influence of the alloy composition on the transformation temperature varies significantly
if it is a Ni or Ti-rich NiTi. Fig. 5 depicts this variation. It can be observed that, for the Ni-rich NiTi, the transformation tem-
peratures decrease significantly with increasing of Ni content. On the other hand, increasing the Ti content does not vary
significantly the transformation temperatures. Such major difference between Ni and Ti-rich NiTi alloys is probably due
to the fact that solubility limit of NiTi on the Ti-rich side is almost vertical. The transformation temperatures of either Ti-
rich or equiatomic NiTi are similar. However, for Ni-rich NiTi an increase of 0.1 at.% Ni may decrease the transformation tem-
peratures by 20 K [7].
The poor workability of SMAs by conventional machining processes, and their increasing applications demand for suitable
joining techniques in order to obtain complex shaped components. This poor workability is related to the high ductility and
strong work-hardening of NiTi [14]. In the last two decades, the number of experiments reported in the literature concerning
welding and joining techniques for NiTi shape memory alloys has increased significantly, as shown in Fig. 6. Traditional
welding processes were tested, but the increasing need to perform highly demanding tasks pushed for the use of state-
of-the-art technologies, like laser welding, with promising results. For this, a special section (Section 3) is devoted to laser
welding of both similar and dissimilar joints.
There is a large number of welding and joining processes, and these can be classified into three major groups:
Mechanical joining, when the parts are put together with clamps, bolts, rivets, etc.;
Welding, when the parts are joined by fusion or in the solid state;
Adhesives, when the parts are bonded by an adhesive applied on the surfaces.
Mechanical joining is a simple and easy to use process, especially when disassembly is envisaged. However, it does not
create a continuous joint, so fatigue and corrosion problems can limit its application.
Concerning welding, there are several classifications of the welding processes, but the one in use splits the processes into
three main groups:
Fusion processes, based on melting and solidification of the area to be joined;

Solid state processes, when there is no melting and joining is accomplished by plastic deformation and solid state
diffusion;
Brazing and soldering when one of the materials or the filler melt and solidify, the parts being thus joined by diffusion.
Diffusion reaction-based process can be considered in this group.
Depending on the size of parts and the application, the joining process can be selected.
Fusion welding is the most used process in industry and, thus, a classification exists based on the type of energy source
used to melt the materials: electrical arc, thermo-chemical energy, radiant energy and electrical resistance.
Examples of the most studied fusion welding processes for NiTi are: tungsten inert gas (TIG), laser, plasma and electron
beam welding. Resistance welding has also been studied for micrometric sized components, as wires, or when mechanical
strength of the overall part is not a major request. As far as solid-state processes are concerned, ultrasonic, explosive and
friction welding have been investigated. Brazing and reactive diffusion have been studied especially for joining dissimilar
materials. Joining by adhesives, diffusion and, more recently, reactive diffusion have also been reported by several research
groups.
Laser welding is, by far, the most investigated and used welding process due to its particular characteristics, such as very
low heat input (HI, which is defined as the ratio between the welding power, P and the welding speed, v: HI = P/v) and
reduced extension of the heat affected zone (HAZ) and fusion zone (FZ), just to mention the most important ones.
Often, there is a need to join different materials taking advantage of the individual properties of each material. In this
case, dissimilar welds are envisaged and these are, in many aspects, different from similar welds. Thermo-physical properties
of the base materials are different, leading to different heat transfer rates on both sides of the weld. During the process,
chemical composition varies across the weld pool, and the interface between solid/liquid phases is not clear. Intermetallic
compounds are formed and convection processes occur, like Marangoni effects [15], since base metal densities vary signif-
icantly. Some detailed studies on the solidification mechanisms, were conducted on dissimilar welds reporting the
microstructural evolution in the FZ of dissimilar NiTi welds, as well as, the main metallurgical problems and strategies to
overcome them [1618]. Embrittlement occurs mainly due to high temperature reaction of NiTi with elements such as oxy-
gen, nitrogen and hydrogen. Brittle intermetallics like Ti2Ni and Ni3Ti can also precipitate during solidification, having
adverse effects on the strength and shape memory characteristics. In order to fulfil specific needs of the medical industry,
welding experiments on dissimilar joints like NiTi/stainless steel [19] or Hastealloys [20] were performed and reported.
In the aerospace industry new parts like gas turbine nozzles are being tested where NiTi is welded to Ti6Al4V [21], substi-
tuting the previously used mechanical joining by fasteners.
Mechanical and functional behaviours of these alloys are strongly influenced by the thermal effects and modifications of
the chemical composition associated with the selected welding process. Nevertheless, welded specimens tend to present
lower ultimate tensile strength (UTS) and elongation to fracture. The shape memory effect may prevail depending on
whether the strain applied is sufficient (or not) to induce an irreversible deformation of martensite.
This paper presents a critical revision of existing knowledge on welding and joining NiTi, which is the most studied shape
memory alloy. The effects of the process on the martensitic transformation, responsible for the shape memory effect and
superelastic behaviour are discussed in detail, as well as their effects on the mechanical properties. Due to its industrial rel-
evance, joining of NiTi to other alloys is also addressed.
2. Welding and joining technologies for NiTi
2.1. Major developments in welding technology
Welding is known since very ancient times (the first evidences date back from 8000 BCE), specially brazing and soldering
[22]. In the 18th century there were significant empirical developments that much contributed to the welding development,
but it was just in the 20th century, when some major scientific and technological developments took place, that a wide range
of welding processes were brought to the industrial market place.
Amongst these developments are: the discovery of the electric arc in 1800 and its later mastering, the widespread use of
electricity after the World War I and, more recently, microelectronics and information and communications technology, to
quote the most important ones.
These discovery landmarks brought welding into mass production especially after WWI when, finally, welding was con-
sidered beneficial for industry.
Oxy-fuel welding was the predominant process till the World War I but the shortage of gases in England by that time
pushed for the developments of electric arc based fusion processes, since electricity was already in use [23]. However,
the need to protect the welds from atmospheric environments rich in Hydrogen and Oxygen was of major concern to prevent
the formation of oxides and the embrittlement and, so, shielded metal arc welding with enrobed electrodes started to be
used. Intensive research was performed in those days, aiming to develop new coatings and new metallic rods.
In the following decades advances were seen in welding reactive metals that together with developments in automatic
welding and alternate current expanded the use of electric arc based processes.
Tungsten inert gas (TIG) welding was patented in 1942 and in the following years a new development emerged that was
the use of a continuous feed wire which very much increased productivity and allowed to weld non-ferrous alloys as
Aluminium.
The concepts of metal active/inert gas (MAG/MIG) were eventually set in 1948 and during the World War II construction
and repair of both ships and airplanes using welding operations was in place.
Fusion welding processes based on electric arc dominated most of the developments observed till the 50s when the emer-
gence of other concurrent processes started. The scientific community assisted to the revolution of welding processes driven
by needs that followed by almost a chronological order:
the requests of productivity achieved with developments in flux cored or electro-slag welding in the 50s.
the requirements of joining new soft materials as polymers in the 60s when solid state processes, as ultrasonic welding,
started to be developed.
the difficulties to weld materials as Al and Ti alloys that pushed for electron beam welding in the 60s followed by laser in
the following decade.
the need to meet requirements of transport sectors as aeronautics improved adhesives used nowadays in metals, com-
posites or dissimilar welds with medium to low mechanical strength under shear stress solicitations.
the increasing need to produce components in parts with specific functionalities and in different materials. This was a
driving force for solid state processes to emerge and a wide range exist nowadays.
the international directives to reduce energy consumption and increase environmental conditions and safety in welding.
In 1991, a new solid state process was patented by The Welding Institute in United Kingdom, and this is Friction Stir
Welding (FSW) [24], which is commonly used for materials that are difficult to join by fusion-based processes [25]. In
FSW a non-consumable rotating tool consisting of a pin and a shoulder provides mechanical mixing and viscoplastic defor-
mation of the materials in the area to weld. Heat is generated by internal energy dissipation during the viscoplastic material
deformation and interfacial friction of the material. The axial load exerted by the tool forges the material in a close forging
matrix. Friction Stir Welding prevents most of the problems associated to fusion welding, such as distortions and residual
stresses, and can produce high strength joints especially in materials with good viscoplasticity as Al alloys.
Friction welding was known since the 1950s, after the Russians, and this was a process where no fusion or melting was
achieved even in low melting temperature alloys.
In 2000, the use of electromagnetic fields, first in forming and later in joining, was another advance in solid state welding,
named magnetic pulse welding (MPW).
More recently, after 2008, laser hybrid welding processes were seen an increasing market, as they reduce the disadvan-
tages of both laser and arc welding processes, namely: joint preparation, fit-up and penetration depth. Other hybrid pro-
cesses are being investigated as friction stir welding assisted by electrical joule effect [26] or by ultrasonic welding [27],
to name a few.
As observed in other technologies, there were parallel developments in both materials and manufacturing technologies.
NiTi, the most studied shape memory alloy [2], with its peculiar characteristics previously outlined, put weldability prob-
lems and, thus, welding this material is a challenge.
In similar joints of NiTi no filler material is needed, while in dissimilar ones, between NiTi and stainless steel or Ti based
alloys, interlayers are used to prevent atomic diffusion of the elements and the formation of complex intermetallic
compounds.
This overview will focus mostly on autogenous welding processes in similar joints but dissimilar ones will also be
included whenever applicable.
The discussion on existing knowledge and the challenges faced with NiTi are presented separating fusion-based processes
from solid-state ones. In the first group, TIG, electron beam and laser welding processes are discussed, since the other fusion-
based processes are not applicable to NiTi. Solid-state welding has also been investigated and this review presents an anal-
ysis of studies performed with explosives, ultrasonic, friction and friction stir, adhesives, diffusion and reactive diffusion
bonding, brazing and soldering and resistance.
2.2. Fusion-based welding processes
2.2.1. Tungsten inert gas (TIG)

Tungsten Inert Gas or Gas Tungsten Arc Welding (GTAW) is an electric arc based welding process, in which the arc is
established between a non-consumable electrode in tungsten and the base material under an inert shielding gas atmosphere.
The main function of the electric arc is to produce the heat to melt the material, though it also has a cleaning function of both
the weld pool and the base materials.
The electrode is made of Tungsten, due to its high melting temperature and thermal emissivity, but additions of rare met-
als, such as Rhenium or Thorium, are used to improve emissivity enabling arc stability. Continuous current in either direct or
reverse polarity are used. The first increases penetration, while the second is used for special materials, such as Aluminium
alloys, to break superficial alumina. In reverse polarity, there is a shallow penetration and an overheating of the electrode
that has to be cooled in order to prevent its damage. The electrode is typically 12 mm in diameter so the molten pool
has small dimensions and the total energy per unit of molten volume is smaller than in other arc-based processes.
TIG is a very controllable and precise process with some particular advantages: (i) it can be applicable to any form of cou-
pling; (ii) the arc is easy to stabilize, even at small electric current intensities; (iii) the heat input can be small enough to limit
the extent of the thermal affected zones; (iv) can weld a significant range of material thicknesses, including thin sheets.
Welded joints are of high quality with no defects under appropriately selected operating parameters. Limited welding pro-
ductivity and the need of highly qualified operators are seldom considered the major disadvantages.
TIG was the first welding process tested in similar NiTi joints in 1961 [28]. In this study, though no mechanical analysis
was performed, the authors produced defect-free joints. Interstitial phases (based on H, N and O), however, were observed
due to insufficient protection of the molten pool. Ikai et al. [29] several years later resumed this process and observed a
decrease in the mechanical properties of the welded joints. In particular, the welded material presented a high irrecoverable
strain during cycling tests for a total of 50 cycles at 4% strain. Just welded wires could withstand the entire cycling routine,
while welded plates suffered rupture after 39 cycles. Additionally, wires with different transformation temperatures were
successfully joined and the welded joints exhibited the stress-induced martensite transformation in two steps, where each
plateau corresponded to the original ones in the base materials. However, as later was studied by [30,31], the presence of
contaminants degraded the functional properties of the welded joints.
Recently, Oliveira et al. [32] performed TIG welding of 1.5 mm NiTi plates in butt joint configuration using dedicated fix-
ing devices to allow inert gas to flow on the face and on the root of the welds to minimize oxidation. The tensile tests of the
welds showed a total strain to rupture of 20% and a superelastic plateau 30 MPa below that of the base material. Superelas-
ticity of the welded joints was confirmed by cycling up to 12% of strain in a total of 600 cycles performed. Shape memory
effect was also found to be preserved after welding. An increase in the transformation temperatures of the FZ was observed
by differential scanning calorimetry, due to Ti oxidation, despite the good gas shielding.
Considering the significant advantages of the process, a couple of interesting applications were reported using TIG weld-
ing of NiTi to stainless steel [33,34]. In the first one, actuators were manufactured using tubes of NiTi and AISI304 butt joined.
Since the tubes had 1.9 mm thickness and an outer diameter of 9.3 mm, TIG orbital was performed with an insert of Ni
placed between the tubes at the interface prior to welding. The Ni interlayer acted as a filler material preventing the Ti-
Fe intermetallic formation. Viable joints could be created with a narrow HAZ, a maximum hardness of about 817 HV and
an ultimate torsional strength of 415 MPa, which was satisfactory for the construction of a solid state torsional actuator.
The high hardness in the partially mixed zone was due to the presence of Ni3Ti2 and Ni4Ti3 intermetallics. When compared
to laser welding, the authors concluded that laser was more suitable for thin walled tubes, while TIG could give full pene-
tration in one single pass at the expense of a larger HAZ due to the higher energy introduced and larger weld bead.
The second interesting application of TIG welding was the manufacture of medical occluders [34], where a cluster of Ti-
rich NiTi shape memory alloy wires was welded by TIG and this weld was later welded to a stainless steel AISI 316L pipe by
laser. Mutual diffusion within a short distance near the fusion boundaries was observed and intermetallics such as Ni3Ti and
(Fe, Ni)Ti formed around the fusion boundary.
Another uncommon application was to use TIG to yield a NiTi alloy overlay on AISI 1048 carbon steel substrate aiming to
improve erosion behaviour of the steel under solid/liquid (sand/water) mixture [35]. A NiTi rod was welded to steel as a
deposited overlay. The hardness of the welded overlay was much higher than the initial hardness of the NiTi rod due to
the formation of oxides during welding. However, the austenite phase of the initial rod was maintained, which is the target
phase for superelasticity. Lattice distortions caused by residual stresses, induced by the thermal gradient, also contributed to
the hardness increase. The oxide particles formed during welding and the small fraction of martensite identified, blocked
dislocation movement, so the erosion resistance was improved even at high impact angles of the sand.
2.2.2. High energy density beams

Electron beam welding is performed in a vacuum chamber where the electrons are accelerated by a high voltage from the
cathode to the anode and focused by electromagnetic field lenses onto the material surface. The kinetic energy of the elec-
trons impacting the material is transferred under the form of heat, rising the material temperature well above its melting
point. In the narrow interaction area, a metallic vapour column forms which transmits the heat into depth in a keyhole weld-
ing mode. The beam is focused into a very small spot, so that the energy per unit area of beam/material interaction is very
high. As a high energy density process, the heat input is very small, thus almost no HAZ is produced and the FZ is very nar-
row, so this is a precision welding process. If the material is thin, a good control of the focal point position has to be ensured
in order to have a good weld bead profile and prevent molten metal to drop. The main advantage is the vacuum environment
that prevents oxidation and contamination, but this is also a limitation due to economic reasons, both equipment invest-
ment, running and maintenance costs and productivity, that is, the time to make a weld.
Few studies exist on electron beam welding of NiTi, essentially due to the process cost. Yang et al. [36] conducted a study
on the weldability of austenitic NiTi with electron beam on 4 mm thick plates and some interesting results can be retained.
The welding mode was a mixed one of keyhole and conduction, with a weld pool constituted by coarse columnar grains with
a small quantity of martensite B190 . Both Ti2Ni and Ni4Ti3 were formed in the FZ. A decrease of the Ms temperature was mea-
sured by differential scanning calorimetry and this means that some Ni was lost during welding, which is in agreement with
existing studies, mostly on laser welding as it will be discussed later. An UTS of 85% of that of the base material was mea-
sured and either in the base material or in the FZ a one-step reversible B2 M B190 transformation was observed. After anneal-
ing, grain refinement in the weld metal was observed and grain size decreased with increasing annealing temperature. Ni-
rich particles precipitated and both Rs and Ms temperatures decreased. Additionally, the ultimate tensile stress dropped [37].
As already mentioned, laser was developed a few years after electron beam and replaced this process with significant
advantages. A beam of light radiation is produced by a stimulation induced process wherein an energy source excites the
active species to a high energetic level which, upon decay to the fundamental level, emits a light radiation where all the pho-
tons have the same energy and wavelength. The produced beam is, thus, monochromatic and coherent, that is, all photons
are in phase (both in space and in time) and have limited divergence, being possible to focus into a very small spot.
When the laser beam interacts with the work piece the photon energy is transferred to the material and converted into
heat, vaporizing and ionizing the molten material creating a plasma above the weld. This plasma absorbs the laser beam and
the efficiency of energy transfer is reduced, or can even be suppressed, when a high plasma plume is formed. So a gas is intro-
duced to blow away the plasma facilitating the laser/material continuous interaction.
Fig. 7. Absorption of a number of metals as a function of laser beam radiation wavelength (adapted from [38]).
Depending on the laser power density, two main welding modes may occur, usually referred to as: conduction and key-
hole modes [38]. Conduction mode is predominant when the energy density is below a threshold value depending on the
material thermal and physical properties and the beam wavelength. The heat transfer is mostly controlled by conduction,
as in fusion arc welding processes and the weld aspect ratio is similar to a TIG weld, that is, the face width to height ratio
of the FZ is larger than 1. In keyhole welding, the introduced heat is higher than in the previous mode, and a plasma is formed
in the interaction area and then transferred into depth, forming a cylinder of vapour plasma. A narrow FZ is formed with an
aspect ratio below 1. An intermediate mode is often identified that is a mixture of the previous ones.
Under an appropriate gas shielding atmosphere, laser welding is a very precise process, feasible for SMAs. Since this is the
most studied process and existing applications mostly uses laser, it is analysed in detail in Section 3.
When compared to arc welding processes, laser welding shows several benefits, including narrow and deep FZ, small HAZ,
low residual stresses and a much higher productivity.
In the past ten years lasers have shown improvements and developments mainly to increase energy efficiency and beam
quality. Solid state lasers, as Nd/YAG and, most recently, fibre lasers have been developed to accomplish both characteristics,
as well as, to improve beam delivery and focusability. When processing materials with a high power laser beam, the ability
to focus the beam is critical, because good focusability means a smaller spot size, higher power density and increased weld-
ing depth.
Wavelengths also affect the beam quality as, at shorter wavelengths, the more energetic photons can be absorbed by a
greater number of bound electrons and, so, the reflectivity falls while the absorption of the beam on a work piece surface
increases. Beam divergence and focal spot size needs to be considered in all applications as these play an important role
in the quality of laser welding, e.g., small beam divergence and focal spot diameters lead to welds with higher penetration
and smaller width for a given welding speed.
While in Nd/YAG lasers the active species is Nd3+ dopant in an Yttrium Aluminium Garnet (YAG) disk or rod, fibre lasers
are based on ytterbium, erbium or thallium-doped multi-clad silica fibres (or bundles of these). Depending on the dopant, or
active species, the emitted wavelength varies in the range of 0.81 mm that is in the infrared region of the spectrum, where
metallic materials highly absorb the beam radiation as shown in Fig. 7.
A high melting efficiency is also observed for low wavelengths lasers compared to CO2 ones. The melting efficiency is the
energy required to melt a unit volume of material. So the heat input and the power density are determinant when controlling
the weld bead profile and the depth of penetration [39].
2.2.3. Resistance welding

Resistance welding is a simple process widely used in several industries, especially in automotive. It consists of forcing a
high intensity electric current (AC or DC) to pass through the parts to weld, either plates or tubes. According to Joule Law, the
quantity of heat generated is proportional to the product of electrical resistance of the interfaces between these, the square
of the current intensity and the time.
The resistance between materials is the highest one, so the quantity of heat generated in the interface is higher than in the
bulk materials and thus it melts surrounded by solid state material at high temperature, partially deformed due to the vol-
ume expansion during melting and the decrease of the yield strength with temperature. Finally, welding is achieved by
applying a forging force, so, a weld is produced in a close environment (between parts) with no filler material. The electrical
current is driven to the parts through two highly conductive Copper electrodes, which have the additional function of forging
the weld till it cools down into air.
The major parameters to control in this process are: the current intensity, the time it passes through the interface, the
pressure applied to forge the weld and the electrical resistance which depends on the material resistivity and the surface
conditions (e.g. oxides, roughness, etc.).
The process is simple to control and, thus, it was studied for joining NiTi. However, it is a thermomechanical welding pro-
cess and it is well established the sensitiveness of functional properties of NiTi, so a good heat distribution is a requirement.
Basic research on micro-resistance welding (MRW) was performed by Tam et al. [40] especially in crossed wires of
0.5 mm diameter. The group studied the bonding mechanisms as well as the mechanical properties, the microstructural fea-
tures and phase transformations. Heat generated during welding locally annealed the strain hardened material, altering the
transformation temperatures of the weld metal, as expected, which were similar to solution treated material. The bonding
mechanisms involved are described to be mostly in solid state bonding involving interfacial melting and dynamic recrystal-
lization, amongst other.
Besides the applications of MRW, an interesting study was performed by [41] in resistance welding when joining thin
walled NiTi tubes to create ordered architectural cellular structures as an alternative to other manufacturing technologies
to produce porous structures in NiTi for biomedical and other applications. Two tubes were joined along their generatrix,
using superelastic NiTi with 7 mm outer diameter and a wall thickness of 0.2 mm. The welding process was seen to highly
affect the transformation behaviour, microstructure and mechanical properties of the joint, so annealing treatments in dif-
ferent conditions of maximum temperature and time were conducted in a post weld heat treatment. The authors demon-
strated that RW was a feasible technique for joining NiTi in complex structures with low density and reversible elasticity.
2.3. Solid-state processes
Independently of the welding process, melting always involves solid-liquid-solid transformations, and these are accom-
panied by dimensional variations, distortions and residual stresses. So, solid-state welding processes can prevent these, as
well as, the formation of undesired phases formed during solidification and in solid state upon cooling. For this, several
solid-state processes have been studied to join NiTi.
2.3.1. Explosive welding

Explosive welding is, eventually, the first solid-state process studied to weld NiTi. In this process, detonation of explosives
is used to bring the weld surfaces of the materials to be joined in a very high velocity oblique collision [42]. As a consequence
of the impact, a severe, but localized plastic flow at the interacting surfaces is created. Due to the dynamics of the process
high temperatures may occur at the weld interface. The major advantage of such techniques is for joining dissimilar metal
combinations that are impossible by other methods. Nevertheless, similar joining using explosive welding is also possible.
First studies reported on using explosive welding to join NiTi were for dissimilar welding with steel [4345]. In these
studies a common objective was aimed: take advantage of the high resistance of NiTi to cavitation erosion, when compared
to the alloys used to build pumps and hydroturbines, and join it with low-carbon steel which provided the structural
strength for such application.
Zimmerly et al. [43] welded austenitic NiTi to low carbon steel. After welding, the NiTi interface was observed by X-ray
diffraction and low-intensity martensite peaks were identified due to the formation of shock-induced martensite during
welding. One of the largest challenges for such type of works is, mainly, to maintain an austenitic interface after joining,
as this phase shows a better resistance to cavitation erosion than martensite. As a consequence of the plastic deformation
and defect production during the shock loading, an increase of NiTi hardness was observed. As far as the cavitation erosion
properties of the as-welded material are concerned, a slight decrease was observed when compared to the NiTi base mate-
rial. However, post-weld heat treatments restore most of the lost resistance. The effect of the welding parameters on the
interface resistance could not be assessed, as failure during tensile lap shear tests occurred, either in the NiTi or in the
low-carbon steel, but never at the bond interface. Richman et al. [44] added new results on joining martensitic NiTi to
the same steel by explosives. Those results agreed with Zimmerly et al. [43], as the welded austenitic NiTi required a
post-weld heat treatment in order to recover most of the erosion resistance lost due to the welding technique. On the other
hand, the effect of post-weld heat treatments on the as-welded martensitic NiTi was not evident to be beneficial to the
cavitation-erosion resistance of the welded joint.
More recently, and after the preliminary work of Prummer et al. [46], Belyaev et al. [4749] produced NiTi/stainless steel
composites by explosive welding. In [47,48] it was shown that the plastic deformation induced during welding lead to a par-
tial depression of the martensitic transformation, but with a post-weld heat treatment, the kinetics of this transformation
could be recovered. No brittle intermetallics between NiTi and stainless steel were found after explosive welding. Later,
in [49] NiTi/stainless steel composites were studied under repeated thermal cycling, showing the shape memory effect.
The recoverable strain exceeded the value for the two-way shape memory effect of the NiTi base material, due to the action
of the elastic steel layer. It was concluded that, such a bimetal composite produced via explosive welding, could be used as an
actuator due to its functional properties.
Explosive welding for joining NiTi to itself was also studied by several authors [5055]. In these studies, two different NiTi
alloys, with different compositions and transformation temperatures, were used. The objective was to obtain a NiTi/NiTi
Fig. 8. Optical micrograph of the NiTi/stainless steel joint produced by friction welding using a Ni interlayer (adapted from [74]).
composite with macroscopic heterogeneities in chemical composition. In [5052,54,55], after explosive welding of NiTi to
itself a controlled thickness reduction was performed and the effect of heat treatments on the transformation temperatures
to the welded joints was analysed. As expected, by a proper selection of the heat treatment temperature, the martensitic
transformation could be adjusted so that the hysteresis of the transformation varied depending on the foreseen application.
Xing et al. [54] studied the damping behaviour of NiTi/NiTi explosive joints and found that the internal friction peaks, which
are related to the phase fractions ratio austenite/martensite, of the two base materials had a different role on the damping
behaviour, depending on the testing temperature. This was due to the different transformation temperatures that are known
to influence the damping behaviour of NiTi shape memory alloys [56].
Although, explosive welding for joining NiTi to itself and to steel, both stainless and low-carbon, were already success-
fully performed, some issues still need to be addressed. In applications where very high strain rates are applied, such as
in explosive welding, material models are necessary to describe the dependence of the flow stress on temperature, strain
rate and pressure during the process [57]. For shape memory alloys, their unique mechanical properties of these materials
pose additional theoretical and computational challenges.
2.3.2. Friction welding

Friction welding is achieved due to the frictional heat generated on the surface between two work pieces under rapid
rotation while being pressured together [58]. Due to the increasing heat generated during this process, the surfaces soften,
plasticize and can mix together. As such, it is required that at least one of the materials to be joined can withstand significant
torque [59]. The strength of friction welded joints can overcome that of the base materials to be joined.
Friction welding has several variants which are dependent on the processes and equipment involved: linear friction weld-
ing [60]; rotary friction welding [61]; spin welding [62]; inertia friction welding [63]; and friction stir welding [25]. From
these, friction stir welding is the most recent and, as such, special attention has been devoted to this technique.
In FSW, a non-consumable rotating tool consisting of a pin and a shoulder moves along the joint, rigidly clamped [64]
under an axial forging force. Due to the friction between the moving tool and the work piece, as well as, to internal friction
during plastic deformation, heat is developed which increases the viscoplasticity of the materials and allows mixing of both
materials in order to form a weld in the solid state, that is, without any melting. During stirring, the temperature increases
due to plastic deformation and originates dynamic recrystallization producing an equiaxial refined grain structure in the stir-
red zone. One of the important advantages of friction stir welding is the low energy consumption, compared to other fusion
welding processes [65], and is environmental friendly since no gas, fumes or particles are emitted.
Shinoda et al. [66,67] joined NiTi rods and observed a grain refinement due to dynamic recrystallization in the stirred
zone, as expected. Increasing the applied pressure, the grain size in the dynamically recrystallized zone decreased. The
UTS of the welded joints was similar to that of the base material, but with a lower elongation to fracture [66,68]. Post-
weld heat treatments were seen to improve the functional properties of the welds, which were then similar to those of
the base material.
Friction stir processing is a variant of friction stir welding with the objective to modify the in-volume or surface proper-
ties of a material, though it can also be used for joining thin sheets in overlap configuration. Friction stir processing was stud-
ied to assess its feasibility for joining NiTi [69] and confirmed that superelasticity and shape memory effect were preserved
[70]. Additionally, NiTi particles were added to an Al matrix creating a NiTi/Al composite [71,72]. Good bonding between the
NiTi particles and the Al was observed under SEM and no harmful interfacial products were observed. With appropriate pro-
cessing parameters and suitable size of the introduced NiTi particles, the tensile strength of the NiTi/Al composite could sur-
pass that of the Al base material [72]. More recently, NiTi/Al composites were successfully obtained by FSW assisted by
electrical current [73]. With this hybrid variant of FSW it was possible to improve bonding between NiTi and Al.
Dissimilar joints of NiTi to stainless steel were produced by rotary friction welding with and without a Ni interlayer [74].
Optical microscopy of the joints produced with the Ni interlayer (Fig. 8) revealed a planar interface without intermetallic
phases. In the absence of the interlayer, the UTS of the joint was of 200 MPa due to the presence of Fe2Ti brittle inter-
metallics. However, this phase was avoided and the UTS increased to 510 MPa when the Ni interlayer was used. Although
the welding time was short, a NiTi + Ni3Ti eutectic layer was observed at the NiTi/Ni interface.
Significant research work still needs to be performed on friction based processes to weld NiTi, in particular with friction
stir welding in order to assess the effect of this process on the mechanical and functional properties of welded NiTi.
2.3.3. Ultrasonic welding

In ultrasonic welding, the parts to weld are pressed by a static compressive load while a high frequency mechanical vibra-
tion is induced by a piezoelectric transducer. The piezoelectric transforms electric current into high frequency mechanical
vibration. Ultrasonic transducers drive the transversely vibrating tip of the sonotrode which is pressed against the top work
piece. The high frequency vibration applied locally generates friction between the surfaces, increasing the temperature in the
interface well below the material recrystallization temperature and causes plastic deformation of the faying surfaces in con-
tact. The combination of the static compressive load and the shear forces due to vibration, originates a dynamic stress field
and an elastoplastic deformation.
This relative interface movement cause shear deformation of contacting surface asperities, dispersing interface oxides
and ultimately bringing clean metal-to-metal contact and adhesion between the surfaces. Atomic diffusion along the inter-
face assists bonding and the final weld has a fine isotropic grain structure. The energy required is directly proportional to the
Fig. 9. Optical micrograph of the cross section of the laser gas nitride NiTi joint (adapted from [82]).
material hardness and thickness. Other process parameters are: the contact force, the ultrasonic wave amplitude and the
time to weld. However, ultrasonic welding is limited to small parts and thin materials.
Some authors have attempted to use ultrasonic additive manufacturing for metallic matrix composites reinforced with
micrometric wires of NiTi. The composites envisaged find applications in adaptive or smart structures that can monitor their
condition, control, or even, heal damage and adapt to environments [75,76].
These applications aimed at consolidating embedded NiTi SMAs in an Al matrix around the reinforcing fibres to improve
bonding quality without affecting the functional properties of the alloys. Using low pressure and oscillation amplitude no
material degradation was observed, but it was seen that bonding was stronger when the temperature increased to about
300 C, that is 25% the melting point of NiTi. However, when the applied pressure was high, transformation of austenite
occurred with detrimental effect on the overall properties of the composite. When the NiTi wires undergo martensite to
austenite transformation, an increase in composite stiffness of 6% was measured, as well as, a reduction in resistivity of
3% [77].
2.3.4. Adhesives joining

Adhesive bonding is another joining technique that can be tailored according to the requirements of the materials to be
joined. Bonding by adhesives is of particular importance for the manufacturing of lightweight constructions. Compared to
mechanical joints, it allows for a homogenous stress-distribution when the joint is loaded [78], which is rather beneficial
when fatigue is an issue.
In order to have successful joints by adhesive bonding, the quality of the materials surfaces is of major importance which
implies, very often, a previous surface treatment to maximize the degree of molecular interaction between the substrate and
the adhesive [79]. Additionally, surface treatments can be used to obtain a more suitable microstructure on the substrates in
order to improve adhesion.
Fig. 10. Stress-strain curves for sandblasted and degreased adhesive bonded NiTi joints (adapted from [83]).
Adhesive bonding can be easily integrated into most industrial sequences allowing great flexibility in terms of design and
production [80]. Also, with such technique, large surface areas can be joined together. However, an important disadvantage
of the use of adhesives for materials joining is its limited heat resistance, which may prevent the use of adhesive joints in
aggressive environments.
Several parameters have a key role in order to promote a sound bond between the adhesive and the substrates to be
joined. Some of the most relevant to be addressed are: adhesive chemistry; type of adhesive cure; substrate compatibility;
joint design and its function [81].
In this review on the adhesive bonding of NiTi shape memory alloys, attention is only given to the cases where conven-
tional joining by adhesives (using overlap joint configuration) was performed.
The first experimental work analysing adhesive bonding between two NiTi plates was done by Man et al. [82] to modify
the superficial texture of NiTi in order to enhance the adhesion properties of the surfaces. Bonding between two NiTi plates
was achieved by irradiating the joint with a laser, while high purity nitrogen gas was discharged onto the molten pool. Later,
etching was performed in order to promote a three-dimensional TiN dendritic structure protruding from the NiTi substrate.
These three-dimensional structures can be generated due to the different etching rates of pure TiN and NiTi. Since TiN pos-
sesses a better corrosion resistance, NiTi is preferentially removed during etching leaving behind the TiN dendrites. The TiN
phase is created due to the reaction of the Ti atoms with the N atoms when the NiTi surface melts. Due to the Marangoni
effect [15], TiN is transferred towards the bottom of the weld pool, allowing the formation of a dendritic structure as
observed in Fig. 9. With an appropriate etching time, the adhesive strength could increase up to 150% when compared to
sandblast etched samples used for control. The reason for such an improvement was related to more protruding dendrites
obtained after longer etching times. As such, more locking sites were available resulting in a higher adhesion and strength.
The mechanical and functional properties of single lap adhesive bonded NiTi joints were assessed also by Niccoli et al.
[83]. Different surface treatments were analysed. Chemical etching was disregarded due to significant material loss. Sand-
blasting and degreasing were also considered as surface treatments. Degreasing did not modify the surface roughness, in
opposition to sandblasting which was able to increase it. Tensile tests performed on the bonded joints in the martensitic
state revealed higher UTS when these were produced after sandblasting (up to 330 MPa and 9% strain vs 210 MPa and
5.8% strain for degreased joints as depicted in Fig. 10). After training, the bonded joints were able to exhibit complete shape
memory recovery, indicating that the functional properties were preserved after adhesive bonding.
Although several studies have addressed the adhesion interaction in composites using polymers with embedded NiTi,
there is a significant lack of research on the study and production of NiTi joints (either similar or dissimilar) by adhesive
bonding. Considering that this technique is widely used in the automotive and aerospace industries, understanding the fea-
sibility of its application to NiTi shape memory alloys might boost the use of these joints.
2.3.5. Brazing and soldering

In brazing and soldering, joining is achieved by a low melting temperature filler material [84], lower than 450 C in braz-
ing, or above this in soldering [85], while the base materials are in the solid state. When the filler is melted by arc or a chem-
ical energetic process, it wets the materials surfaces creating a bond by short distance diffusion mechanisms to the
surrounding solid base materials.
The right choice of the braze alloy is of great importance: its mechanical properties, wettability and possible formation of
brittle intermetallics with the materials to be joined must be considered when selecting the filler.
The time to melt the filler depends on its melting temperature [86] and heating method [87,88], which can be performed
on furnace, by microwave or infrared radiation to have fast thermal cycles [89].
Fig. 11. Time-resolved development of the brazed microstructure (adapted from [91]).
Both similar and dissimilar NiTi brazed joints were investigated. For the similar brazed joints, mainly two distinct filler
materials were used: Nb and Ag-based alloys. Dissimilar brazing involving NiTi mostly involve NiTi/stainless steel joints,
despite NiTi/Ti6Al4V, NiTi/NiAl and others have also been studied.
The first work reported on similar joining of NiTi by brazing techniques was presented by Yang et al. [90] using Cu and Ti-
15Cu-15Ni as filler materials heated by infrared laser. The influence of the filler material on the functional properties of the
brazed joint was evident: when using Cu as filler, full recovery by shape memory effect was observed, while with Ti-15Cu-
15Ni filler no bending tests could be performed due to the brittleness exhibited by the joint, resulting from the presence of
Ti2(Ni, Cu).
Later, Grummon et al. [9194] used Nb as a filler material and performed brazing in a vacuum furnace. Cellular NiTi struc-
tures formed and since Nb was placed in close contact to NiTi, an eutectic reaction occurred around 1180 C. Time-resolved
microstructural observations of the brazed joint are presented in Fig. 11. It can be seen that after just 1 s significant progres-
sion of the melting had already occurred. Increasing the brazing time, Nb is consumed and the amount of pro-eutectic NiTi
decreases.
Mechanical properties of these cellular honeycomb structures, which had a relative density of only 5%, revealed very
interesting mechanical properties: superelasticity was observed, with up to 50% compressive strains recovered upon unload-
ing [93].
Wang et al. [95] also took advantage of the eutectic reaction created between NiTi and Nb to braze NiTi wires with Nb
powders deposited on the contact areas of the wires, in a furnace. The advantage of using Nb to promote joining by brazing,
aside from creating an eutectic reaction with NiTi, is its biocompatibility [96] and its influence to widen the hysteresis of the
transformation temperatures of NiTi [97,98].
Ag and Ag-based fillers are also widely used in brazing due to their low melting temperatures. Several works have
addressed the influence of Ag-based filler materials on the joint characteristics of NiTi/NiTi brazed joints [99101]. Different
fillers and heating methods were tested. Shiue et al. [99] used an infrared laser to melt Ag-Cu-Ti braze alloys (two different
compositions) and join equiatomic NiTi. The joints with either of the brazing alloys had Ag-rich, Cu-rich and Ti-rich phases,
but the shear properties were not affected by the braze alloys. However, the exact composition of the braze alloy was found
to have significant influence on the shape memory recovery: higher Ti content created joints with superior shape memory
recovery (92% vs 64% of the applied deformation). This significant difference was attributed to the different microstructures
originated by each brazing alloy. The presence of CuNiTi had a detrimental effect on the shape memory recovery.
Eijk et al. [100], brazed NiTi with Ag-Ti and Ag-Cu-Ti braze alloys in a microwave. With both of these alloys joining was
achieved with a significant difference of the existing phases after the brazing process. Ag-Cu alloy was used for vacuum fur-
nace brazing of NiTi by Zhao et al. [101]. Lap joints were prepared and it was observed that after a given lap length the
strength of the brazed joint was higher than that of the base material. However, when the lap length was below 3 mm,
the diffusion layer created between NiTi and the Ag-Cu braze was found to be the weakest region of the joint. The brazed
joints were found to exhibit full recovery by shape memory effect when the applied strain was below 4%. Above this defor-
mation, even after triggering the shape memory effect, some irrecoverable strain was observed.
Similar brazing of NiTi assisted by infrared laser was also accomplished using pure Cu and Ti-Cu-Ni alloy as filler mate-
rials [90]. It was observed that the brazing parameters had significant influence on the shape memory recover of the joints:
longer brazing times originated decreasing amounts of CuNiTi phase, which was detrimental for the shape memory recovery.
Significant formation of Ti2 (Ni, Cu) resulted in high brittleness of the brazed joints.
Significant research work has been devoted to cope with dissimilar joining of NiTi to other engineering materials. Gale
et al. [102] investigated martensitic NiTi and NiAl furnace brazed using Cu as a filler material. Likewise to NiTi, NiAl alloys
also exhibit shape memory effect [103]. Although no significant microstructural changes were detected in the NiAl substrate,
Ti migration from NiTi lowered the transformation temperatures of the NiTi substrate and thus, austenite was present at
room temperature. In the bond-line between the two base materials, Ni3Ti with a D024 structure was observed. Interestingly,
some of these precipitates adopted a metastable L12 structure due to the high Ni content in this region.
Fig. 12. SEM imaging of the microstructure in a dissimilar stainless steel-NiTi laser-brazed joint. Stainless steel is positioned on the left side, while NiTi on
the right side (adapted from [104]).
From the available literature on dissimilar brazed joints, the NiTi/stainless steel combination using Ag-based braze alloys
is the most studied [100,104106]. Li and his co-workers [104,106] analysed the effect of brazing parameters on the
microstructure and mechanical properties of NiTi/stainless steels joints brazed by laser using Ag-based (Ag-Cu-Zn-Sn) filler
material. Good wettability of the braze alloy with both base materials was observed, as shown in Fig. 12. Within the brazing
joint, a-Ag, AgZn, Ag3Sn and Cu5Zn8 were detected by X-ray diffraction analysis. Diffusion of Ag, Cu, Sn and Zn into the two
base materials was observed due to the formation of a diffusion reaction layer. Similarly, Ni and Ti from NiTi, and Fe, Cr and
Ni from the stainless steel diffused to the braze alloy during the joining process. Hardness measurements revealed that a
significant decrease, compared to both base materials, occurred in the joint centreline due to the formation of soft phases.
By adjusting the brazing parameters, a maximum strength of 360 MPa was achieved [106]. Although the superelastic beha-
viour of the brazed joint was not studied, local analysis of the NiTi HAZ shown a decrease in its superelastic properties due to
a coarse grain structure. Similarly, deterioration of shape memory effect in the thermal affected NiTi region was observed
(91% recovery vs 100% of the base material).
The same authors have also studied the corrosion behaviour of their laser brazed NiTi/stainless steel joints in artificial
saliva [105]. Although the brazed joint had lower corrosion resistance than the stainless steel, it was higher than the filler
material and similar to NiTi. However, the pH value of the artificial saliva influenced the corrosion behaviour: in more acidic
environments (lower pH) the corrosion resistance decreased. No significant change in the mechanical properties, in terms of
the tensile strength and shape memory recovery after bending, of the NiTi HAZ was observed after the corrosion tests. In the
former, a 25 MPa decrease was observed, while in the latter only a 3% change was measured. It must be noticed that the
mechanical properties of the dissimilar brazed joints using Ag-based filler can meet requirements for orthodontic applica-
tions, as the maximum load during orthodontic clinic is significantly below the maximum load sustained by the joints
[107]. Changing the filler material from Ag-based to Cu did not revealed any significant modification of the mechanical prop-
erties of the brazed dissimilar joints.
Successful joining between NiTi and stainless steel was reported by Satoh et al. [108,109] in laser autogenous brazing.
In this process there is no filler material and the laser parameters have to be carefully controlled so that the base material
irradiated by the laser beam does not exceed its melting point. In the NiTi/stainless steel dissimilar joints, the laser was posi-
tioned either on the stainless steel side or on the NiTi side and moved towards the interface between both materials. Thermal
resistance of the interface causes an increase of temperature up to its melting point. When the laser beam reached the inter-
face, it was turned off and the liquid layer quickly quenched due to the contact with adjacent cold base material. Advantages
of this method are the high quench rate, the minimization of mixing between the two materials and localized melting in just
one material. Appropriate brazing parameters decreased the mixed layer formed and maximized the strength up to 500 MPa,
which is higher than the onset for the stress induced transformation. Nevertheless, during the tensile tests, no evidence of
the superelastic plateau was observed.
NiTi was also brazed to Ni-based superalloy (Hastelloy C-276) by microwave using AgTi filler [100]. Excessive porosity
was observed which is detrimental for the mechanical properties of the joints. Multi-diffusion occurs throughout the brazed
joints leading to a complex microstructure. It is apparent that the AgTi filler is not a suitable choice for joining both NiTi and
Hastelloy.
Brazing NiTi to NiTiNb was successfully obtained by Zhao et al. [110] using an Ag-Cu filler in a vacuum furnace. The most
interesting remark from this study was the wide recovery temperature range from 60 to +45 C. Laser brazing NiTi to
Ti6Al4V using Ag nanoparticles was attempted by Quintino et al. [111]. However, poor adherence of the nanoparticles to
the base materials was observed, indicating that joining between NiTi and Ti6Al4V was not efficiently promoted.
NiTi and low carbon steel rod were brazed assisted by oxyacetylene flame, using Ag-based filler material [112]. Sound
joints were obtained and these were mechanically cycled, without fracture, up to 60 load/unload cycles within the supere-
lastic plateau and very close to their UTS. As such, this process can be promising for structural applications in civil engineer-
ing where a not expensive process is required.
NiTi was also used as a braze alloy by Munez et al. [113] for joining W to ferritic-martensitic steel for a divertor appli-
cation [114] in fusion reactors. Good wettability between NiTi and the base materials, due to the existence of Ti, ensured
sound joining. Also, as the thermal expansion of NiTi was between those of the base materials, the thermal stresses origi-
nated upon cooling decreased, and so did the risk of cracking in the brazed joint.
2.3.6. Diffusion and reactive diffusion bonding

In diffusion bonding, the joining mechanism is dependent on the interdiffusion of atoms across the interface between the
surfaces to be joined [115]. For interatomic bonds to be established, the materials must be in close contact, that is, within the
atomic spacing. However, virtually, there is no effective way to ensure such close contact and several factors prevent this
intimate proximity, the most important ones being related to the surface conditions, namely the surface geometry and
roughness and the presence of impurities and oxides. In diffusion bonding, the parts to join are placed under pressure to
eliminate the surface asperities and below the yield strength of the materials to avoid macroscopy deformation. The temper-
ature can range between 0.5 and 0.8 of the absolute melting temperature of involved materials to improve diffusion [116].
The time necessary for diffusion bonding can range between few minutes up to several hours [117].
Although this is a solid-state technique, it is possible to have some liquid material in the joint interface, if a low melting
temperature interlayer is used that may form a low melting eutectic phase (usually Copper) [115]. In this case, joining is
produced by a transient liquid phase diffusion bonding [118].
Fig. 13. SEM imaging of the NiTi/Ti6Al4V joint using Ni/Ti multilayers (adapted from [123]).
Some of the advantages of diffusion bonding are: low deformation of the joints; ability to join large areas; the thermal
gradient tends to be smaller than those in fusion based processes [119]. However, specimen preparation and joint fit-up
require extreme caution has they significantly influence the ability to obtain a sound joint.
Diffusion bonding involving NiTi is recent and the available literature is scarce. Senkevich et al. [120] studied the effect of
bonding temperature and holding times on the mechanical properties of the similar joints: increasing the temperature (rang-
ing from 800 to 1100 C) and/or the holding time (from 5 to 60 min) improved the mechanical properties. High temperatures
and prolonged times were required to obtain high quality diffusion bonded joints. However, it was observed the formation of
brittle intermetallics such as Ti2Ni and Ti4Ni2O, created by the interaction of the bonded surfaces with residual oxygen. The
affinity of Ti to O at high temperatures required high vacuum environments. NiTi base material microstructure was also of
great importance in diffusion bonding of NiTi, as noted by Senkevich [121]. The presence of inclusions, such as Ti2Ni, can
cause formation of a liquid phase by an inverse peritectic reaction: Ti2Ni ? L + NiTi, after holding times above 2 h. By
decreasing, or avoiding, the presence of such intermetallics, melting during the bonding process decreases, and so does
the degradation of the bonded joints.
Senkevich et al. [122] and Cavaleiro et al. [123] produced NiTi/Ti6Al4V joints by diffusion bonding. In [122] the optimum
temperature ranged between 900 and 950 C, with a welding time between 30 and 60 min, using a contact pressure of
5 MPa. For these parameters range, the tensile strength of the joints reached a maximum of 170 MPa, which was higher than
the 107 MPa reported for laser welding of the same dissimilar combination [21].
Cavaleiro et al. [123] deposited Ni/Ti nanometric multilayers by sputtering onto the two base materials. The use of such
layers provides extra energy at the interface, in a reaction-assisted diffusion bonding process, thus reducing the bonding
temperature [124]. In-situ X-ray diffraction analysis, using synchrotron radiation, allowed to observe the formation of brittle
intermetallics, such as Ti2Ni at 750 C [123]. Ti2Ni intermetallics were also observed in [122] giving raise to significant hard-
ness increase when compared to the base materials. Optical microscopy of the reaction-assisted diffusion joints revealed a
high quality joint, with an almost non-detectable bond-line between the two materials (Fig. 13). The influence of the Ni/Ti
multilayers was proven very effective in decreasing the temperature required for joining NiTi to Ti6Al4V: sound joints were
produced above 900 C [122], but also at 750 C [123]. Thus, for industrial applications, this temperature drop may be an
economic advantage.
Despite the fact that just recently diffusion bonding, started to be applied to NiTi, it is clear that this processing is an inter-
esting one either for similar or dissimilar material combinations especially in microjoining or when joining small parts. Addi-
tionally, it is expected that other dissimilar material combinations, relevant for industrial applications, may be studied.
To summarise, Table 1 presents a comparison of the most relevant advantages and disadvantages of each welding process
discussed within this review.
3. Laser welding of NiTi
Laser welding is, by far, the most used joining technique for NiTi shape memory alloys owing to its intrinsic character-
istics, namely: low heat input, high energy density, and monochromaticity, narrow HAZ and reproducibility as described
previously.
As a consequence of the wide variety of situations of NiTi base material to be welded (initial composition and thermo-
mechanical history prior to welding, for example), a comparison to understand the effects of laser welding and its implica-
Table 1
Comparison of the different joining techniques for NiTi shape memory alloys: main characteristics, benefits, limitations and existing works.
Process Main characteristics Benefits Limitations Existing works

TIG Fusion welding by Low heat input Manual process [29,3235]
electric arc Simple welding source Low productivity
Inert shielding gas Flexible and easy to use Possibility of tungsten inclusions
Wide range of material Possibility of element volatilization and mate-
thickness above 1 mm rial oxidation
Environmental problems
Laser beam Fusion welding No contact with heat Complex to use [1719,125
Monochromatic source Fixing devices and dimensional tolerances 171,172191]
light beam Small interaction area Equipment and running costs
radiation High energy density Preferential volatilization of Ni giving rise to
Minimal heat input local changes of functional characteristics
Narrow HAZ and FZ Expensive process
Low thermal damage
and distortions
High precision
Automatic or robotised
High welding speed
Electron beam Fusion welding No contact with heat Complex to use [36,37]
Electron beam source Fixing devices and dimensional tolerances
radiation Small interaction area Time of preparation
High precision Vacuum chamber
High energy density Equipment and running costs
Minimal heat input Expensive process
Narrow HAZ and FZ
Low thermal damage
and distortions
No oxidation or
contamination
Automatic or robotised
Resistance Fusion welding Simple to apply and Joint configuration limited to overlap [40,41]
Joule effect control Size of components
Pressure Simple equipment Non continuous joining
Explosion Solid state process Dissimilar joining Environmental problems [4345,4755]
welding Plastic Simple process Joint configuration limited to overlap
deformation Not expensive
Minimal residual
stresses
Friction Solid state process Dissimilar joining Equipment costs [6672,74]
Plastic No oxidation or Risk of functional properties degradation due
deformation volatilization to high strain hardening
No heat affected zone
Minimal distortion and
residual stresses
Environmental friendly
Ultrasonic Solid state process Dissimilar joining Small thicknesses and part size [7577]
Plastic No liquid/solid Joint configuration limited to overlap
deformation transformations
No heat affected zones
Minimal distortion
Low residual stresses
No risk of inter-
metallics formation
Adhesives Solid state process Dissimilar joining Joint configuration limited to overlap [82,83]
Adhesive bonding Simple process Reduced tensile strength and fatigue strength
Flexibility
Not expensive
Brazing Diffusion joining Dissimilar joining Poor mechanical strength of the joints [90,91,9395,99
No oxidation or Joint configuration limited to overlap 102,104106,108
volatilization 113]
No heat affected zones
residual stresses
Diffusion and Diffusion joining Dissimilar joining Mechanical strength [120123]
reactive No heat affected zones Size of the parts
diffusion No oxidation or
volatilization
(continued on next page)

Table 1 (continued)
Process Main characteristics Benefits Limitations Existing works

residual stresses
tions on the mechanical behaviour and microstructure is not obvious. Also, the use of CO2, Nd:YAG or fibre lasers in either
pulsed or continuous wave mode are known to produce different effects on the welded joints. For this, a systematic review of
the most relevant studies is presented, focusing on the effects of laser welding on the microstructure, mechanical behaviour
(both tensile, cycling tests, one-way and two-way shape memory effect), transformation temperatures, corrosion behaviour
and the effect of post-weld heat treatments as well as, on the addition of filler materials if appropriate.
As in any manufacturing process, the operating conditions largely affect the efficiency of the joints. As far as laser welding
is concerned, laser wavelength, power and mode, welding speed, focus size and position and shielding gas type and flow rate
are the most relevant. The interaction of the power and welding speed were found to be statistically important in the Taguchi
analysis performed by Chan et al. [133], while shielding gas flow rate was considered a weak factor. However, the shielding
gas, both type and flow rate, are of extreme importance to prevent or minimize oxidation and contamination during welding.
So, the gas type is of concern. For example, Argon is an effective shielding gas when welding NiTi [133], but when Helium is
used it facilitates the formation of narrow welds due to its capacity to rapidly extract heat and it also helps to control the
plasma [192]. Additionally, the effect of the shielding gas flow rate on the microstructural and mechanical behaviour must
not be neglected.
Considering the types of lasers, Nd:YAG and fibre lasers are now the main choice when compared to CO2 lasers. The lower
wavelengths of Nd:YAG and fibre lasers enhance the beam absorption thus requiring less energy for welding. From published
results, it is perceived that laser in pulse wave mode is preferred for joining wires or thin foils, while for higher thickness
sheets laser welding operating in continuous wave mode produces joints with better mechanical properties.
Based on their previous work, Chan et al. [132] used Taguchi experiments to determine the best welding conditions,
which are dependent not only on the base material, but also on the laser system, for improving the mechanical properties
of the welded joints. In fact, one of the tested samples was able to reach a strength of about 95% that of the base material,
with just a slightly decrease on the stress onset for martensite detwinning of the welded joint.
More recently, Otibar et al. [152] used design of experiments to evaluate the quality of laser welded NiTi joints. Factors
such as alloy composition, laser power and pulse duration were studied, amongst others. Based on this study the authors
proposed an optimum combination of different factors to ensure the best response in terms of both tensile strength and
strain. This is an interesting approach, especially for industrial environment, providing that all stages of the process can
be monitored accordingly (i. e. have a good control of the initial condition of the base material and on the laser welding
parameters to have the highest reproducibility possible).
Fig. 14. Different solidification microstructures observed in the fusion zone of laser welded NiTi (adapted from [132]).
Fig. 15. Microstructural differences between the base material and the heat affected and fusion zones (adapted from [167]).
Fig. 16. X-ray diffraction measurements on different regions of the welded NiTi joint (adapted from [168]).
3.1. Similar laser welding
Laser welding of NiTi shape memory alloys dates back to 1989 when Araki et al. [125] welded 3 mm thick Ni-rich NiTi
plates, using a CO2 laser. Due to the high sensitivity of NiTi to high temperatures, laser welding was considered to be an ade-
quate process to minimize the thermal affected regions and to preserve its functional properties: shape memory effect and
superelasticity.
Early laser welding studies on NiTi shape memory alloys were performed using CO2 lasers [31,125127] while, nowadays,
the primary choice for these studies are the infrared lasers, such as Nd:YAG [128131,145,165,166] and fibre lasers [132
134] due to their low wavelength and increased material absorptivity and, thus, narrow welds [38,135].
3.1.1. Microstructure
The microstructure in the FZ of the NiTi laser welds can present either a dendrite-like, planar or cellular microstructure or
a combination of these. The solidification structure observed can be explained based on rapid solidification theory [193],
which establishes four parameters to govern the microstructure found in the FZ: temperature gradient (T), undercooling
(DT), crystal growth rate (R) and alloy composition. The growth rate to temperature gradient ratio, G=T, governs the solid-
ification mode, while the product G.R represents the cooling rate and controls the crystal growth. As the constitutional
supercooling increases, the solidification structure changes from planar to cellular and from cellular to dendritic. As such,
it is possible to observe only one, two or even all three structures in the FZ of a laser weld. An example of the coexistence
of the three types of microstructures is presented in Fig. 14.
In the heat affected zone solid state transformations occur, such as: grain growth, formation of intermetallics and recrys-
tallization. Fig. 15 depicts the microstructures observed under optical microscopy of a laser welded NiTi thin plate. In this
case, it is clear a refined equiaxed structure close to the fusion zone, while going into the base material, an anisotropic struc-
ture is clearly revealed since the maximum temperature in this zone was not sufficient to induce microstructural modifica-
tions observable at this scale [146].
The effect of the weld thermal cycle on the hardness of the different regions is not straightforward, as it also depends on
the initial condition of the base material, that is, on the thermo-mechanical history of the material [194,195]. This depen-
dence was studied by Tuissi et al. [137], where two heat treatments were performed on the NiTi base material prior to weld-
ing. One of the heat treatments consisted of a full solution at 900 C for 60 min and the other one was ageing at 400 C during
20 min. In the former, the hardness in the FZ increased compared to the base material, while in the latter, the opposite beha-
viour was registered: lower hardness of the FZ. Nevertheless, [31,138] did not observe a significant change when measuring
the hardness across the different regions of the weld.
In the HAZ there is usually a decrease of hardness values, due to recrystallization and grain size effects [132,139], being
these effects more notorious when the base material is cold-worked. In the FZ the hardness tends to decrease, which can be
explained by a full annealing effect promoted by the laser [140].
Several research groups have been performing structural characterization of existing phases in the different zones of laser
welds, especially by X-ray diffraction [196], due to the microstructural features size and the narrow dimension of the differ-
ent heat affected zones [133,141143,166,168]. Fernandes et al. [142] and Oliveira et al. [166,168170] used synchrotron X-
ray radiation to probe laser welds from the base material into the FZ, in transmission mode, with a reduced spot size of
200 200 mm and good spatial resolution. Fig. 16 depicts measured data [168] and a significant microstructural change is
observed: both the fusion and the heat affected zones show martensite and austenite to coexist, while the base material
was fully austenitic at room temperature. Similar observations were presented by Iijima et al. [141], using conventional
Cu Ka radiation, although not discussed.
The structural characterization is of great importance, as the identification of these phases in the welded region is crucial
to understand the mechanical behaviour of the welded joints.
One of the reasons for the deterioration of the mechanical properties lays on the formation of brittle intermetallics in the
FZ, such as Ti2Ni [127,144,146,147]. Schlobmacher et al. [144] proposed that Ti2Ni precipitation and the presence of oxygen
in the FZ lead to a premature failure of the joints. This assumption reinforces the need to have the NiTi base material as free
of impurities as possible and required good gas protection during welding so that no contaminants (such as hydrogen or oxy-
gen) can be picked-up by the liquid bath. The presence of Ti2Ni is also detrimental for the corrosion behaviour of the laser
joints, as these, when finely dispersed, are known to be preferential sites for pit initiation in both equiatomic and Ni-rich NiTi
[197].
Aside from Ti2Ni, other intermetallics were also found to exist in the FZ of the laser joints, namely: Ni3Ti
[132,133,148,149], and Ni4Ti3 precipitates [150,167]. The presence of Ni-rich precipitates in the HAZ due to laser welding
must be more carefully analysed as the thermo-mechanical history of the base material has a significant role on the onset
for the precipitation to occur. For example, if the base material was cold-rolled, the temperature raise in the HAZ may be
sufficient for precipitation and/or recrystallization phenomena to occur, even though the time available for such processes
is quite short. On the other hand, an annealed base material may not experience any microstructural change in the region
adjacent to the FZ by the sole reason that the heat treatment performed on the base material has already promoted the
microstructural changes. This is the reason why, sometimes, it is reported that no HAZ is observed in NiTi laser welds
[31,146].
The presence of precipitates in the thermal affected regions must be considered as they decrease the Ni or Ti content of
the surrounding matrix, depending on the precipitation of Ni or Ti-rich compounds, respectively. For Ni-rich alloys, the
dependence of the transformation temperatures on the composition is more significant than for Ti-rich NiTi, which is related
with the solubility range on the Ti-rich side of the Ni-Ti phase diagram [2]. As such, for laser welded Ni-rich NiTi alloys, the
composition dependence of the transformation temperatures may bring significant differences in terms of transformation
characteristics, when comparing the FZ and the HAZ with the base material. For example, a 0.1 at.% shift may change the
transformation temperatures by 20 K [7]. As it will be discussed later, the most striking aspect arising from compositional
shifts in the different regions of the weld is related to the transformation temperatures within each zone, which may cause,
for example, the base material to show one stable phase, while the HAZ and/or FZ have another, at a given fixed temperature.
Another possible effect of such precipitation phenomena is the appearance of the R-phase transformation on the HAZ and/
or FZ. R-phase transformation may occur when Ni4Ti3 coherent precipitates are formed due to the formation of stress fields
[198,199]. Due to the low extension of the thermal affected regions, after laser welding, conventional analytical techniques
such as conventional X-ray diffraction using laboratorial sources, differential scanning calorimetry or electrical resistivity,
cannot be used with high accuracy for determining the phase transformations in each region. For such analysis, special ded-
icated techniques, such as synchrotron or micro X-ray diffraction and/or transmission electron microscopy, must be applied.
3.1.2. Mechanical behaviour

3.1.2.1. Tensile properties. Most of existing work on laser welding of NiTi is devoted to Ni-rich alloys, probably because they
can exhibit superelastic properties at room temperature. Ti-rich alloys, on the other hand, tend to be martensitic at the same
temperature.
For Ti-rich NiTi a significant decrease of the UTS (more than 30% that of the base material [127,138]) in the joints is jus-
tified by the grain boundary precipitation of Ti2Ni in the HAZ and the presence of micro-cracks in the FZ [127]. An embrit-
tlement of the welded material, when compared to the base material, justified the lower ductility and tensile strength.
It was already reported that, in pulsed wave mode, dislocations can be introduced in the material, leading to work-
hardening, due to the high thermal impact of multiple laser individual spots [136,146]. These dislocations decrease the duc-
tility of the joint when compared to the base material.
The plateau for the martensite detwinning in martensitic NiTi can reach up to 6% strain [200]. Falvo et al. [138] reported a
decrease in the martensitic plateau after welding, as shown in Fig. 17. It was proposed that this detwinning deformation was
reduced due to dislocation slip in the HAZ. It is apparent that, after the detwinning plateau of martensite, elastic deformation
Fig. 17. Stress-strain curve for martensitic NiTi and its laser welded joint (adapted from [138]).
of the detwinned structure occurs prior to fracture. Although no structural analysis was performed in Falvos work, it is rea-
sonable to conclude that the large microstructural difference, observed by optical microscopy, between FZ (coarse grain size)
and base material (fine grain size) may cause the observed decrease in mechanical properties of the as-welded joint.
The UTS of the welded joints is mostly below that of the base material, though it depends on the initial condition of the
base material. Analysing the stress-strain curve of the base material in Chan et al. work [132], it can be observed that the
tensile stress of NiTi base material and its elongation at fracture are considerably lower than most studies reported in the
literature such as [5,201,202], where NiTi was tested in more demanding conditions, in terms of strain or stress applied,
without fracture. Gugel et al. [128] reported that similar laser welds presented a stress-strain behaviour similar to the base
material, but the UTS of the welded joints was 25% below that of the base material. However, the NiTi joints exhibited the
superelastic plateau up to 7% and fracture occurred at about 1 GPa. Comparing these values with the ones presented by Chan
et al. [132] also for laser welds, a superior mechanical behaviour is reported by Gugel [128], despite the UTS of the joints
being 75% that of the base material. It seems evident that the ultimate tensile strength of a welded joint is not the adequate
parameter to compare the effect of laser welding amongst different works, as for the base material it is strongly depending
on its initial condition.
As such, to assess the possible detrimental effects in NiTi joints, other parameters must be considered to be compared to
the base material, namely: transformation strain, start and finishing stresses of the plateau, superelastic recovery upon load-
ing/unloading cycling and shape memory effect recovery.
The effect of the rolling direction when welding NiTi plates must also be considered as it is a factor that may alter the
mechanical properties of the welded joint. Vieira et al. [140] reported that a higher UTS was measured when the welding
direction was perpendicular to the base material rolling direction. So, this should be taken in consideration for designing
purposes.
The Clausius-Clapeyron relationship expresses the rate of change of the critical stress for the stress-induced transforma-
tion of temperature for the effect of the lattice distortion of martensite on the transformation [203]. The relationship can be
expressed generally as:
dr DS

dT etr
Fig. 18. Temperature dependence of the irrecoverable strain after loading up to 6% and subsequent unloading (adapted from [136]).
Fig. 19. Observation of stress kinks: (a) base material; (b) welded material (adapted from [151]).
where r is the critical stress, T is the testing temperature, DS is the entropy variation and etr is the transformation strain. For
NiTi the dr=dT ratio varies between 5 and 8 MPa K1 [204]. Schlobmacher et al. [136] presented the only study to analyse the
effect of laser welding on the dr=dT ratio. Samples from the base material, welded joint and a so called remolten specimen
were tested by load/unload up to 6%, at different temperatures, and the stress at 3% strain in the superelastic plateau was
measured, while recording the irrecoverable strain after each unload to 0 stress. The remolten specimen consisted on
an array of laser-welds covering the entire gauge length of the specimen under test. Although the mechanical behaviour
of such specimen is not entirely correspondent with the FZ, it is still a good estimation. It was also observed, that the laser
welded specimen followed the Clausius-Clapeyron relation with a gradient of about 7 MPa K1, very close to the value of the
base material, within the expected range for NiTi [204]. The irrecoverable strain, from both base and laser welded materials,
are depicted in Fig. 18. At 10 C the irrecoverable strain is significantly high due to presence of martensite at that temper-
ature. At 50 C, above the Md temperature (where the stress for inducing dislocation slip is lower than for the stress induced
transformation), plastic deformation was observed in both samples. Thus, a working temperature range between 20 and
40 C, was suggested for applications based on superelasticity, where the irrecoverable strain of the welded material is just
slightly higher than the base material. However, it must be noticed that this working range depends of the base material
transformation temperatures and on the laser welding parameters, namely the heat input and the energy density. The re-
molten specimen presented significantly higher values of irrecoverable strain (up to 4.4%), which can be partially attributed
to the high density of glissile dislocations observed in these samples by transmission electron microscopy.
The stress-induced martensite transformation in NiTi proceeds with the propagation of deformation in macroscopic shear
bands [205]. Experimental results suggest that the end of the superelastic plateau (also called Lders-type plateau) may not
be necessarily coincident with completion of the austenite to martensite phase transformation [206]. Micro or macro-
mechanical modelling of laser welded NiTi is scarce. Chan et al. [151] presented the only study on this topic, where they
proposed a constitutive model to describe the mechanical behaviour of the welded joint under cyclic solicitation. The con-
stitutive model was adapted and modified from [207] and the main modifications performed were based on the fact that,
after laser welding, residual stresses exist in the HAZ and FZ of laser welded NiTi joints [169]. As such, the critical stress
for the stress-induced transformation should occur at a different stress level of the base material. This argument was sus-
tained by the fact that, during cyclic tests, stress kinks were observed prior to the superelastic plateau [151] or within
the plateau [128]. Fig. 19 shows the described stress kinks on both base material and welded joint reported by [151]. It must
be noticed that, in austenitic NiTi, prior to the superelastic plateau, an overshoot of stress is observed as reported by Tobushi
et al. [208]. This overshoot can be explained by the fact that an extra force is required to create the martensite nucleus. Then,
the interface between austenite and martensite moves under constant stress, as typical for the superelastic plateau. If the
strain rate during such tests is low, the interface can move freely after the creation of the martensite nucleus, allowing
for the relaxation of the high stress required. But, if the strain rate is high, the interface will move quickly after the creation
of the martensite nucleus and will not have time to relax the stresses in the interface. If the test speed is very low, these
stress kinks are observed [128,151].
Other adaptation of Chans model relied on considering that the HAZ and FZ, after the stress induced transformation of
austenite to martensite, will undergo plastic deformation. This assumption was made considering that the irrecoverable
strain after the load/unload cycle, presented a higher value of deformation in the welded joint than in the base material
(Fig. 19). So, a component was introduced in the mathematical formulation that considered the plastic strain of the thermal
affected regions. Such higher irrecoverable strain, after the stress-induced transformation, may be due to martensite stabi-
lization which can be of thermal (due to the local chemical composition changes induced by the welding procedure [171]) or
mechanical origin (when, during superelastic cycling, the accumulation of defects allows for an increase of the volume frac-
tion of non-transforming austenite [209]).
A fairly good agreement between experimental tests and the developed constitutive model was observed (Fig. 20).
Although in this experimental work Chan et al. [151], assumed that the thermal affected regions were still austenite, as
the base material, other authors have reported the presence of martensite in those regions, while the base material was aus-
Fig. 20. Experimental and simulated stress-strain curves of NiTi joint up to: (a) 4%; (b) 6%. 1st, 2nd, 5th, 10th, 25th and 50th cycles are shown. Experimental
and simulated data are represented by dots and blue line, respectively (adapted from [151]).
Fig. 21. Stress-strain curves for base material and laser welded joint, which is assumed to be a composite of heat affected zone/fusion zone/heat affected
zone (adapted from [31]). Points A and B were used to compute the expected tensile strength of the fusion zone (rFZ) according to the equation:
rcom 0:34rBM 0:66rFZ , where rBM and rcom correspond to points A and B, respectively.
tenitic [142,166,168,170]. As such, other significant modifications (mainly considering the role of martensite on the stress
relaxation and irrecoverable strain after each load/unload cycle) must be incorporated in the proposed constitutive model.
Tensile curves of laser welds usually comprise the full weld including the HAZ and FZ. A more in depth study of the tensile
behaviour of the thermal affected regions of Ni-rich laser welded NiTi, was performed by Hsu et al. [31]. Within the gauge
length of tensile specimens, two-thirds of the volume fraction of the analysed weld corresponded to the FZ, while the
Fig. 22. Schematics of the welded joint (adapted from [139]).

Fig. 23. SEM images of: (a) top of welded joint; (b) bottom of welded joint (adapted from [139]).
remaining was the HAZ. Assuming that the tensile specimens resembled a lamellar composite (HAZ/FZ/HAZ) an expected
ultimate tensile strength of the FZ was computed and was close to 740 MPa. Tensile tests revealed that the composite
had a higher UTS than the base material, despite the elongation at fracture being considerably lower, as observed in
Fig. 21. This value was in good agreement with the tensile curve of the laser weld. However, the specimen presented a brittle
fracture, while the base material had ductile-like features. As such, the mechanical behaviour of the laser joint was limited by
ductility on the FZ, which was found to be lower than the base material.
An example of a more complex welded NiTi joint was studied by Tam et al. [139], where Ni-rich wires were welded as
depicted in Fig. 22. SEM images revealed full penetration welds with no apparent defects (Fig. 23). Of particular interest
was the smooth surfaces presented by the welded joint. This condition was considered very good for biomedical implants
as it should increase the biocompatibility with the human body [210].
The superelastic behaviour of any shape memory alloy can only be assessed after at least one load/unload cycle. When the
material is mechanically tested in the austenitic or martensitic phase a constant-stress plateau is expected. In the former
case, upon unloading, the irrecoverable strain at zero stress should be significantly lower than in the latter. For that reason,
although a constant-stress plateau can be observed in tensile tests of NiTi welds, it is not assured that superelasticity is
preserved.
Several factors are known to influence the superelastic response of NiTi, namely: the testing temperature [208], the effect
of previous thermo-mechanical treatments [211], the maximum applied strain [212] and the grain size [213]. As it can be
easily understood, all these factors play a significant role on the mechanical cycling behaviour of the welds. Thus, the inter-
pretation and discussion of the observed behaviour can be highly complex. Also, the large variety of conditions regarding
similar laser welding of NiTi requires paying close attention to the experimental details of each study so that a clear com-
parison between results can be performed. Different lasers and welding parameters affect the superelastic behaviour of the
NiTi welds, as these significantly alter the microstructure in the thermally affected regions.
As a rule of thumb, laser welds present lower superelastic behaviour (higher irrecoverable strain after unloading) when
compared to the base material. Chan et al. [132] subjected laser welds to a load/unload cycle at different strains (4, 6 and 8%).
It was observed that, with increased applied strain, the irrecoverable strain upon unloading increased. Also, these values
were greater than those of the base material, this being attributed to the microstructural modifications occurred in the
FZ. It was also observed that the critical stress, for the stress-induced transformation, in the welded joint decreased, when
compared to the base material. Similar findings, regarding the decrease of the critical stress for the occurrence of the supere-
lastic plateau, were also reported in [137,140]. It is suggested that this stress reduction occurs due to the microstructural
modifications introduced in both HAZ and FZ, which allows for these regions to reach the constant stress plateau, after
the initial elastic deformation, at lower stress values than those of the original base materials.
3.1.2.2. Mechanical cycling behaviour. In NiTi SMAs fatigue can occur by two distinct modes: functional and structural fatigue
[214]. In the former the functional properties of the material suffer a degradation during cyclic solicitation, while the latter
the material fails, during high cyclic loads like any other engineering material. Several studies report the cycling behaviour of
NiTi welds [31,125,126,132,137,140,144,166]. However, a significant disparity in the reported experiments is observed.
Cycling tests reported are in a range from 1 to 600 cycles and in a strain range between 3 and 10%. The test speeds, also tend
to be very different. As such, a comparison between different results is not straightforward.
Hsu et al. [31] cycled laser welds under 3% imposed strain during 50 mechanical load/unload cycles and observed a per-
manent strain of 0.19% which was higher than that of the base material (0.09% of irrecoverable strain). However, this small
permanent deformation was attributed to the low total strain imposed during the tests. Similar results were reported by
Gugel et al. [128], when cycling both base material and welded specimens for a total of 10 cycles up to 2.5% strain. Nonethe-
less, for applications where conventional materials are not able to present such elastic behaviour, NiTi laser welds can def-
initely be used.
The effect of maximum strain imposed during cycling on accumulated irrecoverable strain was studied by Vieira et al.
[140]. When testing both the base material and the welded joint up to 4 and 8% strain for a total of 30 cycles, significant
Fig. 24. Example of superelastic behaviour presented by the laser welded NiTi joints after a given pre-defined number of cycles (adapted from [166]).
differences were found. While after the cycling tests at 4% the welded material presented an accumulated irrecoverable
strain of nearly 0.7% and this value increased up to about 4.4% for 8% imposed strain tests. In both cases, the welded material
always had a higher accumulated irrecoverable strain than the base material. Interestingly, it was pointed out that the
superelastic behaviour of the joints was observed up to applied stresses that were just 50 MPa below their UTS, which
may be of significant importance for industrial applications.
Chan et al. [153] analysed the cycling behaviour of NiTi welds up to 100 cycles, with a maximum imposed strain of 4%.
The aforementioned stress kinks (Fig. 19 from [151]) which were attributed to a local stress induced transformation of the
thermal affected region were only observed in the first loading cycle. The absence of those features in the following cycles
suggests that local plastic deformation of those regions has occurred, thus, not contributing for the superelastic effect and
leading to an increase in the irrecoverable strain.
Another experimental condition that may influence the analysis of the results, concerning the effect of laser welding on
the superelastic behaviour, is related to the ratio between the volumes of the heat affected regions to the base material dur-
ing these cycling tests. It is expected that, for the tests performed in the same conditions, when this ratio is higher (the rel-
ative importance of the thermal affected region is, thus, more significant), a decrease in the superelastic properties should
occur. This effect may be related to the already reported microstructural heterogeneities observed in those thermal affected
regions when compared to the base material.
The functional fatigue by superelasticity of NiTi laser welded plates was assessed by Oliveira et al. [166] by subjecting the
welded samples to a total of 600 load/unload cycles at maximum imposed strains of 8 and 10%. Although martensite was
observed in the HAZ and FZ, while the base material was fully austenitic, the welded joints still exhibited superelasticity
(Fig. 24). The superelastic recovery was found to be correlated with heat input introduced during welding: higher heat input
led to higher extension of the thermal affected regions and, as a consequence, the superelastic recovery decreased. Aside
from the evidence of thermal stabilization of martensite in the HAZ and FZ of the welds, mechanical stabilization of the
stress-induced martensite during superelastic cycling was also observed under synchrotron radiation [171]. Thermal stabi-
lization of martensite occurred due to a local variation of the chemical composition of both HAZ and FZ due to welding,
which raised the transformation temperatures of those regions. Mechanical stabilization of martensite occurred due to
the introduction of dislocations during cycling of the welded joints. The microstructural differences of the as-welded, after
4 and 600 cycles at 10% strain are depicted in Fig. 25.
Typical applications that take advantage of superelasticity in cyclic solicitations are displayed in Table 2. It is, thus, evi-
dent that understanding the fatigue behaviour and the underlying microstructural processes is of great importance, in order
to safely design NiTi components [214]. Microstructural features, such as grain size, are known to influence the fatigue
strength of NiTi [215]. For example, the presence of coarse grain size in the thermal affected areas after welding, can be
responsible for the decrease on the fatigue life of polycrystalline materials [216].
Conventional fatigue of laser welded NiTi was studied using rotating-bending specimens [154,155] analysing the effect of
the imposed strain amplitude on the number of cycles to failure [154] and the influence of the bending frequency on the
fatigue strength [155]. As expected, it was observed a decrease on the fatigue properties of NiTi after laser welding, which
was attributed to the segregation of solute and impurities upon solidification of the weld metal and to the existence of coarse
grain dendritic structures in this region.
It was observed that the differences between the base material and welded material increased when increasing the strain
amplitude. When the imposed strain amplitude was below 0.4%, the welded joints were able to withstand up to 106 cycles
Fig. 25. Superimposition of the X-ray diffractograms for laser welded NiTi samples: (a) as-welded; (b) after 4 cycles at 10% strain; (c) after 600 cycles up to
10%. Intensity in log scale (from [171]).
Table 2
Typical applications of NiTi shape memory alloys that require
superelastic behaviour and expected fatigue life [214].
Application Fatigue life [number of cycles]

Stents 108
Damping, internal friction 108
Orthodontic wires 105
without rupture. However, when the strain amplitude increased to 1.0%, the fatigue life of the NiTi joint was reduced to
about 103 cycles, while the base material was able to sustain a total of 104 cycles.
Considering the influence of the bending frequency, it was seen a decrease of the fatigue strength when increasing the
bending frequency. During the stress induced transformation of austenite to martensite, the temperature of the sample
increases [217]. As such, when the frequency increased, the generated heat also increased. With increasing temperature,
the plateau for stress induced transformation increases requiring a higher stress to occur. It is even possible to overcome
the Md temperature, at which the stress for dislocation slip is lower than that required for the starting of the stress induced
Fig. 26. Influence of the imposed strain, ep, on the strain recovery by shape memory effect, em, and permanent strain, et, for both base material and laser
welded joint (adapted from [138]).
transformation. Thus, these factors explain the lower fatigue strength of the NiTi base material and laser welds with increas-
ing frequency.
Fracture analysis of the welded samples after fatigue tests, revealed that cracks were originated in the FZ due to higher
roughness, when compared to the base material, coming from solidification after laser melting. In NiTi, it was reported that
increasing surface roughness, as usually found in the FZ, may decrease the fatigue-crack growth resistance [218], justifying
the decrease in the fatigue strength of the welded material. Brittle-like features in the fracture surface were reported in
[154], while in [155] the welded material fractured in a ductile mode.
Nevertheless, laser welded NiTi joints revealed a lower fatigue strength when compared to the base material, though
these joints were still able to withstand a significant number of cycles, which is of significant importance for applications
in micro-electro-mechanical systems [219], for example.
3.1.3. Shape memory effect

Shape memory effect is, eventually, the most important functional property attributed to NiTi and, thus, preserving the
shape memory effect after welding is of major concern, for example, in applications such as in actuators.
The shape memory effect on laser welded NiTi was reported in early studies [125,126]. Full recovery of the imposed strain
was achieved when the samples were heated to the austenitic phase. That is, from the functional properties presented by
NiTi, the shape memory effect appears to be the less affected by laser welding.
Several authors reported that at least 90% of the imposed deformation is recovered when the shape memory effect is trig-
gered [31,127,137,144,168]. Evidence of the shape memory effect in laser welded NiTi was investigated by either tensile
[31,125,127,137,144] or bending tests [168]. In the former, the samples are pulled, in the martensitic phase, up to a certain
Fig. 27. Evolution of the recovery angle over time for a NiTi weld (from [168]). The Global angle and HAZ + FZ angle corresponds to the shape memory
effect observed in the base material of the welded joint, and in the heat affected and fusion zones, respectively.
strain and unloaded to a 0 stress condition, presenting a significant deformation. Then, the welded specimens were heated
above the Af temperature and the shape memory effect occurred, allowing for a significant strain recovery of the imposed
deformation. In bending tests, the procedure was similar but the samples were bent over a given curvature radius.
However, the shape memory recovery is dependent on the applied strain. Schlobmacher et al. [144] tested NiTi joints up
to 6% strain with subsequent unloading to 0 stress and heating above Af. They reported that the irrecoverable strain after
martensite ? austenite transformation increased with the applied strain, as it was expected. Although the irrecoverable
strain of the welded joints was small (bellow 0.1 and 0.3% for strains of 4 and 6%, respectively), these were less performant
than the base material. The effect of imposed strain on the shape memory recovery for Ti-rich laser welded NiTi was also
evidenced by Falvo et al. [138]. Fig. 26 depicts the experimental results of Falvo et al. It was observed that the maximum
strain recovery of the welded joint was 4% for an imposed strain of 5%, while, for the base material, the maximum strain
recovery achieved 5.2% for a total strain of 7.3%. It can be noticed that the curve correspondent to the evolution of the strain
recovery, em, increases up to a certain point and then starts to decrease, being this feature observed earlier in the welded
sample. This was seen to be related to the development of dislocation slip, leading to plastic deformation. This early devel-
opment of slip in the welded material can be due to the lower strength of the joint when compared to that of the base
material.
When reporting the shape memory effect of NiTi welds, researchers are often concerned with the overall behaviour of the
welded joint. However, recently, Oliveira et al. [168], observed the presence of distinct shape memory effects in the base
material and in the region comprising the HAZ and FZ. The evolution of the recovery angle with time measured (Fig. 27),
shows that the thermal affected regions return to their original shape earlier than the base material, which is an interesting
observation. Though both the heat affected zone and the fusion zones were constituted of austenite and martensite at room
temperature (while the base material was fully austenitic), heating the samples to this temperature after bending, the shape
memory effect was clearly evidenced.
The problem of the two-way shape memory effect after welding was also addressed [130]. Ti-rich NiTi, 1.2 mm thick
plates, martensitic at room temperature, were butt welded using a Nd:YAG laser. After welding, the residual stresses result-
ing from previous cold working of the base material and laser welding were reduced by post-weld heat treatment. A load/
unload cycle, up to 3.5% strain was performed to compare the mechanical response of the base material and the welded joint;
the main difference observed was on the stress for the onset of martensite detwinning, which was found to be higher in the
welded specimen. This effect was attributed to the grain growth and formed precipitates due to the weld thermal cycle. Pre-
vious research performed on NiTi showed that increasing the recrystallization temperature of Ti-rich NiTi give rise to
increasing transformation temperatures [220]. So, for the same mechanical testing temperature, the critical stress for the
stress induced transformation should be lower.
Similar load/unload curves for laser welds were recorded by [138,144], which did not showed significant difference on the
stress values for the onset of martensite detwinning, when compared to the base material, in opposition to others observa-
tions [130]. This highlighted that the thermo-mechanical history of the base material prior to welding plays a significant role
on the features presented by the joints after welding. Experimentally, this was confirmed by heat treating cold worked Ti-
rich NiTi laser welded samples after at 400 C for 45 min [138] and at 700 C for 20 min [130].
Training of the samples was also performed for a total of 8 cycles up to a strain of 3.5% to obtain the two-way shape mem-
ory effect [130]. Fig. 28 depicts the stress-strain curves correspondent to a selected number of training cycles of both base
material and welded samples. It can be observed that, increasing the number of training cycles, hardening occurs and, thus, a
large stress level is measured together with a decrease of the stress required for the onset of the martensite detwinning.
As a requirement for the two-way shape memory effect, after each load/unload cycle the material must be heated above
the Af temperature so that martensite transforms into austenite [4]. Thus, Falvo et al. [130] recorded (i) the residual strain, er,
corresponding to the strain at 0 stress after each load/unload cycle, (ii) the recovery deformation, ere, which corresponded to
Fig. 28. Training cycles of: (a) base material; (b) welded joint (adapted from [130]).
Fig. 29. Evolution of the shape memory behaviour and plastic strain with the number of cycles in: (a) base material; (b) welded joint (adapted from [130]. er
residual strain; er recovery deformation; etw two-way shape memory strain.
Fig. 30. Differential scanning calorimetry curves for base material and welded region where the base material had a heat treatment of: (a) 900 C for 1 h; (b)
400 C for 20 min (adapted from [137]).
the strain at 0 stress after transformation into austenite and (iii) the two-way shape memory strain etw, after subsequent
cooling to the martensitic phase. In the base material, the two-way shape memory strain recovery increased continuously
with the number of cycles; despite this behaviour was also observed in the welded joint for the early training cycles, the
two-way shape memory recovery started to decrease with a correspondent increase on the residual strain, as observed in
Fig. 29. This effect may be explained by the development of dislocations slip in the welded region leading to plastic
deformation.
These results show that smart actuators based on laser welded NiTi joints can be produced. However, it must be pointed
out that the performance of these actuators may be reduced when compared to the base material, due to the lower values of
two-way shape memory strain.
A two-stage shape memory effect, which is different from the two-way shape memory effect, was observed [126] in welds
of two NiTi sheets with different compositions, with full recovery of the imposed deformation on the joint. As the base mate-
rials had different transformation temperatures, it was expected that the FZ had a transformation temperature in between
them. Thus, the shape memory effect could be triggered at different times: after deformation in the martensitic phase by
heating to the lower Af temperature of both base materials causing the first-stage shape memory effect; subsequent heating
up to the higher Af temperature of the other base material promoted the second-stage shape memory effect.
Similarly, analysing butt welds of an equiatomic alloy and a Ni-rich NiTi by differential scanning calorimetry, the base
materials showed just one transformation peak upon cooling/heating corresponding to the austenite M martensite transfor-
mation [156]. After welding, two peaks were identified due to mixing of the two materials. However, it is interesting to note
that, upon cooling, a first transformation peak was observed which matched the transformation peak of the equiatomic base
material, while the second peak identified was shifted to lower temperatures, when compared to the Ni-rich base material.
Upon heating, the second transformation peak of the weld matched that of the Ni-rich base material. That is an increase of
the Ni content on the Ni-rich NiTi alloy occurred since laser welding can be comparable to a solution treatment, and as such,
Fig. 31. Differential scanning calorimetry curves of: (a) base material; (b) thermal affected regions (heat affected zone and fusion zone) (adapted from
[138]).
a decrease in the transformation temperature on the thermal affected regions of the Ni-rich alloy happened which is in good
agreement with existing knowledge [2].
3.1.4. Phase transformation temperatures

As known, transformations temperatures can be determined by differential scanning calorimetry [221,222], dilatometry
[223] or electrical resistivity [224,225]. However, transformation temperatures of the thermal affected regions in NiTi welds
have been mostly determined by differential scanning calorimetry, despite their small dimensions.
This problem was reported by [144] and, to overcome it, the minimum required dimension of the sample to be tested was
remolten by the laser to simulate the FZ. However, it is very difficult to ensure that a given sample for thermal analysis has
only one of the thermal affected regions [156].
A clear example of the effect of the base material initial condition, and its influence on the transformation temperatures of
the FZ, was presented by Tuissi et al. [137] and is depicted in Fig. 30. The type 1 material, was heat treated at 900 C for 1 h
prior to welding, and the welded region only had a slight increase in the transformation temperatures. The smooth increase
of the heat flow signal prior to the Ms temperature (upon cooling), observed in the laser welded joint can be related to a
higher density of nucleation sites for martensite formation. For the type 2 material, which, prior to welding, had a heat
treatment at 400 C for 20 min, it was possible to detect a spreading of the transformation temperatures over a range higher
than 80 C. This particular feature shows that a strong modification was introduced by the laser and individual contributions
to a given transformation cannot be recognized properly.
Some of the changes in the transformation temperatures on the welded region were attributed to oxygen pick-up during
welding by Hsu et al. [31] in equiatomic NiTi/NiTi joints, where a shift of Ms to lower temperatures was observed, when com-
pared to the base material (32 vs 38 C, respectively). The oxygen content was measured and it was found to be significantly
higher in the FZ. For NiTi, the increase in oxygen content is known to decrease the transformation temperatures by favouring
the formation of Ti-rich secondary phases [226], and this reinforces the need for a good shielding protection, so that the con-
tamination in the FZ is not detrimental to the joints properties.
Falvo et al. [138] stated that laser welding has the potential to reset the effect of cold-working and heat treatments on the
welded joint. Comparing the differential scanning calorimetry curves of the base material and thermal affected regions (HAZ
and FZ) in the same sample, depicted in Fig. 31, it was observed that the two-stage transformation, corresponding to austen-
ite ? R-phase ? martensite gives place to a one-stage transformation from austenite to martensite. Cold-worked NiTi sub-
jected to annealing heat treatments may show two-step transformation [227,228]. This thermal behaviour may be
completely destroyed by the welding procedure.
Tuissi et al. [127] performed differential scanning calorimetry measurements in both base material, HAZ and FZ. Their
findings show that in the base material and in the HAZ only one transformation peak is identified in both cooling and heat-
ing. However, in the FZ a partially overlapped two-step transformation was observed and was attributed to the presence of
R-phase due to the imposed residual stresses after welding.
The effect of laser welding on the modification the transformation temperatures does not have always the same trend. For
example, the Ms temperature of the welded region can be lower than that of the base material (40.5 vs 2.2 C, respectively
[153]), or higher (10 vs below 60 C [154], respectively). It is thus clear that, other microstructural features in the ther-
mally affected regions may have a role on these modifications. Namely, the possibility of laser-induced defects (as reported
in [136]) or the observed grain size differences, when comparing the FZ to the base material [128,140,153], without exclud-
ing the possibility of precipitation phenomena as noticed in [132,148], or the incorporation of external elements during
welding [31]. All these factors affect the transformation temperatures of NiTi.
Fig. 32. Evolution of the martensite and austenite phase fraction in a laser welding NiTi joint with a fully austenitic base material (from [170]).
Increased defect density is known to cause a broadening of the transformation peaks [229]. This broadening can be over-
turned by annealing. As such, when welding cold-worked NiTi, it was referred that laser welding acts as a solution treatment
[156] and can reset the effects of cold-work and previous heat treatments [138], as stated previously. The dependence of the
transformation temperatures with the grain size is related with the grain boundary area [230]. With a higher grain boundary
area per unit of volume, corresponding to smaller grain size, nucleation sites also increase, leading to a narrower transfor-
mation peak. In opposition, for large grain sizes broadening of transformation peaks is expected to occur. As a general rule of
thumb, increasing the grain size decreases the transformation temperatures.
Precipitation in NiTi modifies the local composition of the matrix. A Ni or Ti depletion occurs when either Ni or Ti-rich
precipitates form. Chan et al. work [132], observed Ni3Ti in the welded region and so, it was expected a raise in the trans-
formation temperatures due to the Ni depletion in the surrounding matrix. However, a decrease in the transformation tem-
peratures was observed and attributed to thermally induced defects in the weld. In fact this is a complex issue and further
research is still needed for a deeper understanding of the phenomena involved. The addition of alloying elements is also
known to change the transformation temperatures of NiTi [2].
Laser processing of NiTi was performed by Khan et al. [231]. Although this is not a laser welding research, some interest-
ing findings can still be applicable. In the laser processed material, full penetration of NiTi with a pulse overlap of 80% was
attained. This can be then compared to bead on plate welding, which is normally used for determining the best range of
parameters for butt welding. Differential scanning calorimetry of the processed region, where fusion occurred, revealed
the presence of extra peaks at high temperatures when compared to the one-stage transformation observed, both in cooling
and heating, in the base material. These extra peaks were due to a variation in the microstructure of the material after laser
processing: while the base material was fully austenitic, the processed region had both austenite and martensite (observed
by X-ray diffraction). As such, a good agreement between the thermal analysis and the microstructural characterization by
X-ray diffraction was obtained. The existence of martensite in the processed region was explained based on preferential
volatilization of Ni in the FZ [232]. In fact, Ni depletion increases the transformation temperatures [2]. So, the austenitic base
material at room temperature prior to laser processing, experienced an increase in its transformation temperatures that
allowed for martensite to be stable at room temperature, replacing austenite as the equilibrium phase. From [232] significant
contributions can then be retrieved for understanding the implications of similar laser welding of NiTi joints. Namely, on
how a material that is fully austenitic at room temperature can present, at room temperature, martensite in the FZ after
welding; this may also justify the increase of transformation temperatures observed in this region when compared to the
base material. Detailed structural characterization using synchrotron X-ray diffraction, performed by other authors, also
showed, at room temperature, the presence of martensite in the fusion after welding, while the base material was austenitic
[142,166,168,170]. In fact, it was observed in [170] that, after welding, in the HAZ and FZ of an originally fully austenitic NiTi
SMA there is a variation of the phase fraction of martensite and austenite, with the martensite content increasing towards
the weld centreline (Fig. 32).
It must be also noticed that for NiTi, differential scanning calorimetry is often used to index the transformation steps
[206]. However, the transformation peaks observed with this technique do not provide direct information on the crystallo-
graphic phases that were formed [233]. As such, multi-step transformations, as reported in the literature for laser welding of
NiTi [127,137], may then be misinterpreted. So, complementary analysis of both X-ray diffraction and differential scanning
calorimetry can provide a solid basis to clearly identify the transformation characteristics in the welded material.
Fig. 33. Cyclic potentiodynamic polarization curves in 0.9% NaCl solution of: (a) base material; (b) welded joint (adapted from [147]).
3.1.5. Corrosion resistance

Combining the shape memory effect and superelasticity to a good corrosion resistance turns NiTi into an ideal candidate
for the biomedical industry, where devices such as stents [234] or archwires [235] are manufactured. It is reported that the
corrosion resistance of NiTi is between that of stainless steel and Ti6Al4V alloy [236].
Due to its use in the biomedical field, corrosion resistance of NiTi is usually tested in fluids that simulate human body
ones [237]. As the welded NiTi joints are intended for being used in such industry, it is, thus, important to understand
how laser welding influences the corrosion resistance of NiTi and to assess this property for the long-term stability and bio-
compatibility of implants and devices [147]. Nickel release may occur when NiTi is in contact with body fluids [238] and can
cause allergic, toxic or carcinogenic effects [239]. Thus, NiTi may require the use of barrier layers to prevent Ni into the
human body [240].
Several works have studied the corrosion behaviour of laser welded NiTi [133,147,157160,165]. The presence of hetero-
geneous microstructures can be detrimental to the corrosion resistance of NiTi [241]. As such, it would be expected that NiTi
laser welded joints would present a lower corrosion resistance than the base material, due to the significant microstructural
differences that can be found in the thermal affected regions. However, Yan et al. [165] showed that the welded joint had a
superior corrosion resistance than the base material. Tests were performed in both NaCl and Hanks solutions with different
concentrations and the breakdown potential measured in the welded joints increased. When analysing the base material, it
was observed the presence of carbides. These were completely removed from the FZ due to the weld thermal cycle, resulting
on an improvement of the corrosion properties of the joints. Aside from removing such carbides, an increasing Ti/Ni ratio was
measured in the FZ [165], which is known to be beneficial for the improvement of the corrosion properties, of NiTi [242]. The
welded specimens did not present a decrease in its corrosion properties, when compared to the base material, which was
Fig. 34. Stress-strain behaviour of the welded material in oil (black curve) and in Hanks solution (red curve). The open circuit potential (OCP) of the weld in
the Hanks solutions is represented by the blue curve (adapted from [159]).
justified by the homogenization of the microstructure and the non-existence of weld defects in the FZ [133]. Similar findings
were also reported by Yan et al. [157].
In another study, Yan et al. [147] showed that, aside from a raise of 700 mV in the breakdown potential of the welded
joint when compared to the base material, the former has a wider passive region, as shown in Fig. 33. Ruhlig et al. [158] also
studied the local electrochemical activity of NiTi welds compared to the base material and did not find any significant dif-
ferences between them.
In opposition to the aforementioned works on corrosion behaviour of laser welded NiTi, Hsu et al. [31] reported a signif-
icant decrease of those properties when tested in artificial saliva where the corrosion rate of the FZ was significantly higher
(274 mm/year) than for the base material (17 mm/year). This decrease was attributed to the formation of intermetallic par-
ticles in the welded region, which were not reported by other authors. So, a good control of the laser welding procedure must
be used to avoid the formation of undesirable intermetallic phases in the FZ.
The combination of both mechanical solicitations and the surrounding medium may significantly change the corrosion
behaviour of NiTi [243]. As such, is of great importance to understand the corrosion performance of NiTi welds upon
mechanical solicitation. Chan et al. performed two distinct works reporting the susceptibility to stress corrosion in Hanks
solution of laser welded NiTi [159,160]. The corrosion resistance of NiTi comes from the formation of Titanium oxide
[234], which forms spontaneously. However, during mechanical solicitation this passivating film may be damaged and
not promptly restored. Localized corrosion in superelastic NiTi is strongly affected by the cracking of the surface oxide, crack
opening/closing and passivation, mutually competing during cyclic mechanical loading in fluids [244]. Also, in chloride based
mediums, such as the Hanks solution, the poor-self healing of the oxide films is intensified. Moreover, embrittlement of NiTi
is more likely to occur due to hydrogen absorption [245]. Stress-induced martensite was found to be more prone to oxidation
than austenite [246], which must be considered when investigating the viability of using NiTi in possible corrosive environ-
ments where superelasticity is required.
In both [159,160] studies, slow strain-rate tests were performed to analyse the mechanical properties of NiTi welds in
Hanks solution. Chan et al. [159] found out that the HAZ was the weakest zone due to its coarse grain size. Stress corrosion
cracking of the welded material occurred at a potential close to Epit of that region, while the base material was immune at
this potential. A small reduction on the UTS of the welds was reported at open circuit potential, as it can be observed in
Fig. 34. Also, there is a clear change in the slopes of the regions II and III.
The analysis of the evolution of the open potential during the slow strain rate test, shown in Fig. 34, gives substantial
information on the cracking and posterior re-formation of the oxide in the welded region. When the open circuit potential
decreases, the oxide film is damaged, while, when it increases, growth of the oxide film occurs. At small strains (region I in
Fig. 34), the oxide film is not damaged, so the open circuit potential remains almost constant. In region II, stress-induced
martensite of the thermal affected regions occurs, as explained previously [151]. For this, the open circuit potential of the
welded sample exhibited a mixed potential of both the base material and the thermal affected regions. Further increase
of the applied strain leads to an almost constant decrease of the open circuit potential (region III). Two mechanisms take
place in this region: (i) competition between film damage and its formation and (ii) increasing amount of stress-induced
martensite. Region IV shows a sharp decrease of the open circuit potential due to necking. The occurrence of necking rapidly
increases oxide damage, promoting the fracture of the welded material.
In [160], Chan et al. studied the cycling behaviour (10 cycles up to 4% strain) of the welded material in the Hanks solution
or in an inert environment. The more significant difference between the cyclic behaviour was a higher accumulated irrecov-
Fig. 35. Evolution of the accumulated irrecoverable strain with the number of cycles (4% strain) for: base material on oil (black); weld in oil (red); base
material in Hanks solution (green); weld in Hanks solution (blue) (adapted from [160]).
Fig. 36. Differential scanning calorimetry curves for the base material and fusion zone of NiTi welds with and without post-weld heat treatments (PHWHT).
The dashed line represents room temperature (adapted from [153]).
erable strain presented in Hanks solution (Fig. 35). This decrease in the superelastic properties of the welded material in the
Hanks solution was caused by presence of hydrogen in the repassivation of the oxide film. This welded sample showed the
presence of hydrogen induced cracking and a brittle fracture.
An in vitro study of the biocompatibility after laser welding of NiTi was reported by Chan et al. [161]. In this study, the
response of mesenchymal stem cells [247] on the effects induced by laser welding on the material (surface roughness, topog-
raphy and Ti/Ni ratio) were evaluated. A good biocompatibility of the welded material was reported based on the successful
cell adhesion and spreading. Usually, cell spreading is a good indicator of the non-toxicity of a given material/medium [248].
Interestingly, it was pointed out that the morphology of those stem cells were modified by the topography of the FZ, namely
its dendritic structure.
It is apparent that the corrosion performance of NiTi welds presents interesting behaviour that should allow their use in
aggressive environments, where such type of joints may be of great use like in biomedical devices.
3.1.6. Post-weld heat treatments

Post-weld heat treatments are usually used for: stress relief of residual stress due to the weld thermal cycle [249] or opti-
mization of the tensile properties of the welded material by obtaining a more suitable microstructure [250,251]. Accordingly,
by choosing appropriate heat treatment parameters (temperature and time) the mechanical properties of the welds can be
improved.
For Ni-rich NiTi shape memory alloys, precipitation of Ni4Ti3, Ni3Ti2 and Ni3Ti was seen to occur. Ni4Ti3 and Ni3Ti2 are
metastable precipitates [252,253] while Ni3Ti is the equilibrium precipitate [2]. For Ti-rich NiTi appropriate heat treatments
can originate Ti2Ni [254].
Fig. 37. Effect of post-weld heat treatments on the mechanical properties of laser welded NiTi joints (adapted from [163]).
Fig. 38. Cycling behaviour of welded joints after heat treatment performed at: (a) 350 C for 60 min; (b) 450 C for 60 min (adapted from [153]).
Fig. 39. Amplitude strain and number of cycles to failure for laser welded NiTi with and without post-weld heat treatments (from [162]).
As a consequence of these post-weld heat treatments, it was observed the presence of Ni4Ti3 in [134,153,162,163]. The
presence of Ni4Ti3 may improve the mechanical properties of NiTi as it can resist the plastic flow by acting as effective bar-
riers to dislocation motion [153].
For Ni-rich NiTi, the heat treatments promote a raise in the transformation temperatures, due to Ni depletion of the
matrix surrounding the Ni-rich precipitates [2]. This was also observed by Chan et al. [153] and is shown in Fig. 36. As such,
it appears that the modification of the transformation temperatures of the welded region, after a post-weld heat treatment, is
enhanced by the removal of structural defects (such as dislocations) and Ni depletion. In opposition, for Ti-rich welded NiTi,
precipitates formed during heat treatments can promote a decrease in the transformation temperatures.
Fig. 36 shows that a two-step transformation took place after two different post-weld heat treatments performed in
[153]. This is an indication that R-phase should exist as an intermediate phase during austenite M martensite transforma-
tion. Similar results were observed by Yan et al. [163]. R-phase formation is justified by the presence of Ni4Ti3 precipitates,
which are known to promote the existence of coherency stress fields favouring the formation of this phase [255].
In Chan et al. work [153], the elongation at fracture could increase by 1%, when compared to the non-treated weld, pro-
viding that the time and temperature chosen for the heat treatment were adequate. The onset for the stress-induced marten-
site can also be modified as shown in Fig. 37 [163].
The effect of post-weld heat treatments on the cycling behaviour of NiTi joints has been studied (Fig. 38) by Chan et al.
[153]. While after the heat treatment, performed at 350 C for 60 min, superelastic behaviour is observed (Fig. 38a), after the
heat treatment performed at 450 C for the same time, a typical martensitic curve is identified (Fig. 38b). The absence of
stress kinks, as reported by the same authors in [151] (shown in Fig. 19), is pointed out in Fig. 38. This gives an indication
that the heat treatments have promoted a homogenization of the existing phases along the welded material.
The irrecoverable strain observed in Fig. 38a is mainly due to martensite detwinning, as this phase is stable at the test
temperature alongside with austenite. When compared to the weld without any post-weld heat treatment, an improvement,
in terms of the irrecoverable strain, was observed. This behaviour was attributed to the fine precipitation of Ni4Ti3. For
Fig. 38b the irrecoverable strain is more significant after the performed post-weld heat treatment, which changed signifi-
Fig. 40. Effect on the post-weld heat treatments on laser welded NiTi: (a) variation of the open circuit potential during slow strain-rate cycling test (10
cycles up to 4% strain) in Hanks solution; (b) evolution of the accumulated irrecoverable strain during the cycling tests in different environments (adapted
from [164]).
cantly the transformation temperatures, allowing for martensite to be stable at the test temperature (room temperature),
together with some austenite. An adequate choice of the post-weld heat treatment parameters is then required in order
to preserve the functional properties of the welded joints.
Regarding the effects of post-weld heat treatments on the rotating-bending fatigue strength of the NiTi welds, Yan et al.
[162] showed that with the appropriate heat treatment a significant improvement can be obtained (Fig. 39). While for both
heat treatments Ni4Ti3 precipitation has occurred, a significant decrease in the fatigue strength was observed for the heat
treatment performed at a higher temperature (500 C vs 400 C, for 60 min). The reason for this behaviour, lays on the coarse
grain size of such precipitates. When coarse Ni4Ti3 precipitates (300 nm vs 10 nm) are found in the material, the stress fields
are attenuated with the introduction of dislocations to relax this misfit.
The corrosion behaviour of the welded joints can also be improved by choosing a suitable post-weld heat treatment,
which depends on the initial condition of the base material. For example, Chan et al. [133,134] reported that a post-weld
heat treatment at 350 C for 60 min could greatly improve the corrosion resistance of the welded joints. In opposition, for
a post-weld heat treatment at 450 C for 60 min, although an improvement was also reported, it was not as significant. This
increase in the corrosion resistance behaviour, after the post-weld heat treatments, was attributed to Ni4Ti3 precipitation,
which altered the surface composition. Moreover, aside from the raise in the Ti/Ni ratio, due to Ni-rich precipitation, the
TiO2/TiO ratio also increases, explaining, thus, the better corrosion resistance after the post-weld heat treatment.
During slow strain-rate tests in Hanks solution, Chan et al. [164] tested the cycling behaviour up to 4% strain for 10
cycles, of a NiTi weld with a post-weld heat treatment at 350 C for 60 min. A comparison was performed with the base
material and with the weld with no post-weld heat treatment. In Fig. 40a the open circuit potential of the welded sample
after heat treatment reveals a more stable pattern than the as-welded sample. The better corrosion resistance after the heat
treatment is ensured by the higher value of the open circuit potential. Nonetheless, the corrosion behaviour of the heat trea-
ted welded sample is still lower than that of the base material in the Hanks solution. For the slow strain-rate cycling tests in
the Hanks solution, the accumulated irrecoverable strain of the welded sample after heat treatment significantly decreases,
when compared to the as-welded sample (Fig. 40b). As such, superior superelastic behaviour was obtained. If the welds are
to be subjected to stress corrosion conditions, the lower residual stresses achieved after adequate post weld heat treatment
have a beneficial effect on the weld performance.
It is, thus, apparent that post-weld heat treatments, can in fact optimize the mechanical and functional properties of the
welded NiTi joints. For this, it is necessary to choose the optimum conditions for the heat treatment.
3.1.7. Filler materials

Though laser welding is usually performed under autogenous conditions, in particular cases a filler material can be used
to improve mechanical properties by suppressing undesired phases or by promoting a more appropriate microstructure
[256]. Since laser welding of NiTi can be easily performed, provided that the welding parameters are carefully selected,
few studies exist on the use of filler materials in similar NiTi joints. However, since the transformation temperatures of NiTi
are also dependent on the alloying elements [2], the use of a filler material can be interesting to tune the transformation
temperatures of the welded joints and their mechanical response.
Zhao et al. [148,149] reported the use of Cerium as a filler in laser welding of Ni-rich NiTi sheets. In [148], the microstruc-
tural differences between Cerium-filler and non-filler laser welds were analysed. Optical microscopy did not reveal any sig-
nificant difference between both joints. By energy-dispersive X-ray spectroscopy only a small variation of the chemical
composition was observed in the fusion boundary that could have resulted from the complex mixtures of different struc-
Fig. 41. Stress/strain behaviour of base material and laser welded NiTi joints with and without Cerium addition (adapted from [149]).
tures. The effect of the Cerium addition was clear by X-ray diffraction, which showed the presence of CeO2 particles uni-
formly dispersed in the FZ.
Concerning the mechanical response of the welded NiTi joint, with and without Cerium, a two-stage superelasticity was
reported in the latter [149], as depicted in Fig. 41.
This peculiar behaviour was attributed to co-existence martensite. The microstructure of Cerium-added welds resembled
a heat treatment microstructure with CeO2 and Ni-rich precipitates uniformly dispersed in the surrounding matrix. As such,
when compared to the base material, the Cerium-added weld shows a two-stage superelasticity with a first plateau at the
same critical stress of the base material and a second plateau at a significantly higher stress level. This feature conveys to the
welded joint a higher apparent rigidity.
Ce and Nb-additives were introduced in NiTi joints by Zhao et al. [257]. For Cerium addition a thin foil of (50.9Ni-49.1Ti)-
2Ce (at.%) was placed between the 50.9Ni-Ti (at.%) base material. Niobium addition into the FZ was achieved by laser welding
the same base material with 47Ni-44Ti-9Nb (at.%). The mechanical properties of the additive-welded joints were superior to
those of the similar joints, as shown in Fig. 42. The improvement of the mechanical properties of the additive joints is
explained through the fine solidification structures with a good grain-linking in the weld centre, in opposition to a coarse
grain structure in the non-additive joints. So, the main effect of the Cerium or Niobium addition was to refine the microstruc-
ture of the weld. Second phases were found to exist in both Ce and Nb-additive joints: Ce-oxides and b-Nb phase, respec-
tively, and were not found to be detrimental to the joints properties.
Li et al. [150] laser butt welded Ti-rich NiTi using a pure Copper interlayer between the interfaces to be joined. Structural
analysis by micro X-ray diffraction revealed the presence of Ni4Ti3 and a significant amount of intermetallics Cu-based com-
pounds: CuTi, Cu4Ti3, Ti4Cu3 and Cu3Ti2. These precipitates were found to increase the hardness of the welded joint com-
pared to the base material. Tensile tests of the welded joints revealed a 50% decrease on the mechanical proprieties when
compared to the as-welded joint. This significant decrease may be related to the formation of Cu-based intermetallics which
contributed to a mixed ductile-brittle morphology on the fracture surface of the FZ. The welded joint presented shape mem-
ory effect after welding (recovery of 99% of the imposed deformation).
In theory, different filler materials can be used to obtain different properties of the welded joints, namely Au, Co, Fe, Hf,
Pd, Pt or Zr, as they are all elements which can create NiTiX shape memory alloys, being X a given element [2,258]. The main
Fig. 42. Tensile tests on base material and both additive and non-additive joints (adapted from [257]).
problem to overcome would be to properly select the thickness of the filler material so that, its amount, in the FZ, could cre-
ate an alloy with the appropriate composition to exhibit shape memory and/or superelastic properties.
3.2. Laser welding of dissimilar joints involving NiTi
Multimaterial components involving different materials are facing increasing applications for both technical and eco-
nomic reasons, and these constitute a challenge as far as welding and joining are concerned. These multimaterial joints
can provide better and more flexible solutions, as they can take benefit of the individual properties exhibited by the mate-
rials involved. For example: heat, wear and corrosion resistances can be optimized [259,260]. However, the different thermo-
physical and chemical properties of the materials pose difficult welding problems related to the formation of brittle phases
and segregations, as well as the existence of significant residual stresses due to physical mismatch of the welded materials.
Although some of these problems can be solved using solid state techniques, such as friction-stir welding or adhesive bond-
ing, others may arise. For example, a specific joint geometry can make friction-stir welding unfeasible; diffusion welding can
have superior technical benefits to obtain dissimilar metals combinations, but such technique is time consuming, leading to
increased costs during the manufacturing process [259]. Laser welding has significant advantages in these cases. In partic-
ular, the high power density of the process, the possibility to have reduced thermal affected regions, easy to control process,
flexibility and reproducibility, makes it a suitable joining technique for promoting dissimilar joining of metals [260]. How-
ever, the effectiveness of laser welding to promote sound dissimilar joints, is highly dependent on the physical properties of
the materials to be welded [261]. Electron beam welding can be a competitor, but the need of a vacuum system is a disad-
vantage towards laser welding, as higher production rates can be obtained with the latter technique.
The transport phenomena, namely heat transfer, fluid flow and elements distribution, can be numerically modelled for
laser welding of dissimilar metals [262]. It must be noticed that the metallurgical phenomena involved (due to Marangoni
effect and convection in the weld pool alongside with melting and mixing) on welding of dissimilar metal combinations can
be extremely complex. Nonetheless, mathematical modelling and simulation can be used to get a better knowledge on the
physical processes which occur in the FZ [263265].
Dissimilar laser welding of NiTi mainly involves three engineering materials: carbon steel, stainless steel and Ti alloys,
particularly, Ti6Al4V. However, other dissimilar combinations have been studied. As such, this section is divided in three
sub-sections. This review aims to discuss NiTi/steel and NiTi/Ti6Al4V welding, due to their industrial importance, but also
other attempts of joining different materials to NiTi are reported. This discussion includes microstructural effects and
mechanical properties of the joints. The corrosion behaviour of the dissimilar welded joints is also discussed (whenever
applicable).
3.2.1. NiTi/steel
Due to the biocompatibility of austenitic stainless steel [266,267] and the functional properties of NiTi, joining these
materials is highly desired for biomedical applications [268] as clamping and/or crimping methods would be easily replaced.
However, due to the alloying elements in steels several intermetallic compounds can occur in fusion based processes. Thus,
highly focused heat sources with reduced thermal affected regions are envisaged for these applications. Some of the possible
intermetallics to be formed in the FZ of dissimilar NiTi/steels joints are Ti2Fe, Ti2Cr, Ti2Ni and Fe2Ti, for example.
By carefully controlling the process parameters, Gugel et al. [128] were able to obtain faultless NiTi/austenitic steel joints.
Optical microscopy of the dissimilar joints did not reveal a coarse grain microstructure, in opposition to the similar NiTi/NiTi
joints described previously. Hardness measurements in the FZ were significantly affected by the presence of brittle inter-
metallics: while the maximum hardness in the NiTi and steel base materials were of 300 and 600 HV, respectively, and
the FZ had a maximum of 950 HV. The dissimilar joints fractured at the end of either the NiTi FZ or the steel HAZ.
Of particular importance in Gugel et al. work [128] is the fact that the joint strength was of about 620 MPa, that is higher
than the onset for the superelastic behaviour of NiTi (400 MPa [269]). So, it was possible to obtain superelastic behaviour of
the NiTi/steel joints opening a wide range of possibilities in different industries. The superelastic behaviour of the dissimilar
joints was studied for a total of 10 load/unload cycles up to 2.5% strain. The superelastic plateau was observed slightly before
1% strain and an almost negligible irrecoverable strain after each unloading was recorded. It must be noticed that the authors
stated that although it is possible to obtain sound joints in very thin wires, as in [128], similar results may be difficult to
obtain when trying to joint bulk material. This is due to the significant differences in the relative importance of the brittle
intermetallics formed in the FZ of thin wires, when compared to plates or wires of higher calibre.
Similarly to carbon steels, dissimilar joining NiTi to stainless steel can originate severe metallurgical problems as a con-
sequence of the formation of brittle intermetallics in the FZ. These metallurgical problems were described by Pouquet et al.
[172] for overlap welding thin foils of NiTi and stainless steel. The joint configuration was found to be of great importance to
promote full penetration: when NiTi was placed on top of stainless steel a higher energy was required. Homogenization of
the weld pool could also be manipulated by controlling the pulse duration, as better mixing was obtained when the pulsed
duration increased. Without the addition of filler material several defects were observed in the FZ due to presence of brittle
intermetallics, such as Fe2Ti. In order to minimize, or suppress, these intermetallic compounds, a Ni interlayer was posi-
tioned between the base materials. Cracking in the FZ was reduced due to the suppression of Fe2Ti, while c (FeNi) and Ni3Ti
were observed. Similar results were presented by Mirshekari et al. [143] using a butt joint configuration. The dissimilar joints
presented poor UTS (200 MPa) and fracture occurred along the weld centreline with brittle characteristics. X-ray diffraction
Fig. 43. SEM image of the NiTi/stainless steel joint deformed up to end of the superelastic plateau (adapted from [17]).
analysis revealed the existence of several brittle intermetallics compounds: Fe2Ti, Cr2Ti, FeTi, Ni3Ti and Ti2Ni. It is evident
that, from the aforementioned results, that a careful control of the process parameters, joint configuration and use of filler
materials determine the ability to obtain sound NiTi/steel joints.
Such a precise control of the conditions to promote a sound joining between NiTi and stainless steel was already achieved
by some authors [17,19,173,174,176,177]. Vannod et al. [17] presented a very in-depth study on the characterization of NiTi/
stainless steels joints using wires. Crack-free joints were obtained and the UTS of the welds was above the superelastic pla-
teau of NiTi. Using in-situ tensile testing it was possible to observe that a crack appeared at the begging of the superelastic
plateau of the welded joint, in the transition zone between the melted region and NiTi. However, it did not evolve during the
plateau. In fact, only at the end of the superelastic plateau, where an increased stress is required to continue deformation, the
crack propagated along the perimeter of the weld. The sudden crack opening has reduced the cross section of the welded
joint and, as such, an increased stress was observed at which the joint failed (620 MPa). SEM imaging of the dissimilar
welded joint just before the end of the superelastic plateau is shown in Fig. 43. The cracks (marked by dashed white circles)
are formed in the transition between the fusion and HAZ on the NiTi side propagating almost parallel to this region.
In order to understand the mechanism that justified the observed cracking and subsequent failure, a model was devel-
oped in order to investigate the stress-strain behaviour of the welded joint. It was observed that the contact of NiTi with
a much harder zone, the FZ, confined the largest stresses to the region where the cracks were initiated. Another fact that
contributed to this increased stress in the neighbourhood of NiTi, was the fact that this region was in tension not only along
the axial component, but also in both hoop and radial components [17].
This peculiar site for crack initiation and propagation was studied in more detail [175] by a new 3D-EDS technique com-
bining both FIB and SEM. With such configuration it was possible to obtain multiple EDS spectra of the region of interest and
different heights (steps of 100 nm, with a total of 352 EDS maps). From the 3D stacks of EDS measurements, based on the Fe-
Ni-Ti ternary phase diagram, it was possible to identify the phases, which had coarse structures, such as dendrites and these
were Fe2Ti and c-(Fe, Ni). For finer structures, as eutectics, 3D scattered electron images were able to resolve these
microstructures (c-(Fe, Ni)/Fe2Ti and (Fe, Ni) Ti/Ni3Ti). From this study it was retained that the region where the cracks ini-
Fig. 44. Torque-angular displacement plot for different samples of 5.08 mm dissimilar welded tubes (adapted from [173]).
Fig. 45. Torque-angular displacement plot for different samples of 9.53 mm dissimilar welded tubes (adapted from [173]).
tiate during tensile tests did not comprise any intermetallics layer, but rather a complex mixture of intermetallic compounds
originated from different solidification conditions during welding.
From the aforementioned studies on laser welding of NiTi to steels (stainless included) the dimensions of the base mate-
rials are rather small (up to 1 mm in thickness or diameter), which limits their use in structural applications. An interesting
approach was made by Hahnlen et al. [173] when welding thick tubes (5.08 and 9.53 mm of outer diameter) of NiTi to 304
stainless steel using Ni filler between both base materials. The welding procedure consisted on rotating the base materials
while a laser was pulsed, creating a continuous FZ. While in similar joints three different regions are usually observed (base
material, heat affected and fusion zone), two extra regions were identified by Hahnlen et al. [173]: partially mixed zone and
unmixed zone. It must be noticed that within the FZ it was possible to identify several partially mixed zones, with different
compositions. The unmixed zone occurred when the melted base material had no mixing with the remaining FZ. An
increase in hardness was observed in the thermal affected regions of the dissimilar joints. However, by diluting the weld
pool, the Ni filler was able to partially suppress the formation of brittle intermetallics. The mechanical behaviour of the
welded joints was evaluated by torsional tests; while the samples were being loaded, the applied torque and the angular
displacement were measured (Fig. 44, corresponding to the 5.08 mm thick tubes). Elastic loading of twinned martensite
was followed by detwinning of this phase, at a low torque-angle slope. Afterwards, elastic deformation of detwinned marten-
site preceded the plastic deformation region, where fracture occurred. The corresponding ultimate shear stresses of the
welded samples ranged from 312 to 481 MPa. Pinhole defects were observed on the joint that presented the lower mechan-
ical properties. These defects acted as stress concentration points causing premature failure [270].
A different torque-angular displacement relationship was observed for 9.53 mm thick laser welded tubes (Fig. 45). The
higher thickness of the tubes promoted a significant stress gradient along the tube. Thus, it was possible that NiTi was com-
pletely detwinned in the outer surface, while the inner surface had not yet reached the critical stress for the onset of detwin-
ning. Nonetheless, it is apparent that some detwinning occurs for these large welded tubes and the shape memory effect can
be triggered with a temperature raise above Af of the NiTi base material. It was also observed that the stainless steel was
plastically deformed during mechanical testing, suggesting that the joint strength was higher than the yield strength of
the stainless steel base material. Fracture of all welded samples occurred with brittle characteristics along the fusion bound-
ary on the NiTi side. The mechanical properties exhibited by the welded joints in Hahnlen et al. study [173] is of significant
importance as far as structural applications are concerned, as these joints were able to overcome the onset for martensite
detwinning and the yield strength of AISI 304 stainless steel.
Similarly to Pouquet et al. [172] the use of filler materials between NiTi and stainless steel in butt joint configuration was
performed by Li et al. in [19,176,177] using Co, Ni and Cu interlayers, respectively. In these studies a similar finding is
observed: interlayers can promote better mechanical properties of the welded joints as they reduce the amount of brittle
precipitates, such as: Fe2Ti or Cr2Ti. However, excess of filler material inside the FZ may promote the formation of other
intermetallics based on the added element such as Ti3Cu, Ti3Co, and Ni3Ti, for Cu, Co and Ni interlayers, respectively.
Although the results are similar, the effect of each element on the welded joints has a particular role.
For the Ni interlayer [176], it was observed that the dilution of the FZ occurred by increasing the interlayer thickness
(ranging from 10 to 100 mm), as depicted in Fig. 46. As Ni is a c-stabilizer, the increase in Ni content promoted the formation
of c-Fe. Confirming the evolution of the weld content with interlayer thickness, structural characterization by means of X-
ray diffraction revealed a decrease in the intensity peaks of Cr2Ti and Ti2Ni. However, excess of Ni content in the FZ, aside
from introducing defects such as shrinkage cavities and increase the susceptibility to gas pores, promoted the formation of
Ti3Ni. The mechanical properties of the welded joints were also related to the interlayer thickness. The maximum strength
was observed for a thickness between 40 and 60 mm (370 MPa). All welded joints fractured along the FZ on the NiTi side in
a brittle-like mode.
Fig. 46. Effect of Ni interlayer thickness on the contents of Ti, Cr, Fe and Ni within the FZ (adapted from [176]).
Addition of Co into the FZ [19] was also investigated, since the analysis of the binary phase diagrams of Fe-Co, Ni-Co, Cr-
Co and Ti-Co, it can be observed that Co and Fe, Ni, Cr and Ti have full solubility in the liquid state, while in the solid state
there is limited solubility [271]. The influence of Co addition on the hardness of the FZ is clearly depicted in Fig. 47. Improve-
ment of the mechanical properties of the joints was achieved after insertion of the Co interlayer (UTS of 345 MPa vs 200 MPa
for the joint without interlayer). These improved mechanical properties were caused by the decrease of intermetallics com-
pounds, such as: Fe2Ti or Cr2Ti, which were partially replaced by TiCo, Ti2Co and Co2Ti as identified by X-ray diffraction.
Regarding the use of Cu interlayer between NiTi and stainless steel [177], it was observed that, above a critical thickness
of the interlayer (80 mm), Cu segregation areas were encountered in the FZ. The hardness of the FZ could also be tuned by
controlling the interlayer thickness: the higher the thickness the lower the hardness. This was due to the formation of
Cu-based intermetallics, which are softer than Ti2Ni or Fe2Ti, for example. The UTS of the welded joints reached a maximum
when the interlayer thickness was of 80 mm: 520 MPa and 4.8% strain (Fig. 48). Fracture of the dissimilar joints also occurred
on the NiTi side on the FZ. The shape memory effect of these dissimilar joints was also tested through bending tests. How-
ever, from the experimental details, it is perceived that the FZ was not subjected to bending, but rather the base material,
preventing to truly assess the shape memory effect of the welds.
It is apparent that, although the insertion of interlayers between NiTi and stainless steel is capable of promoting better
joints with superior mechanical properties in comparison with dissimilar joints without interlayers, a precise control of the
interlayer type and thickness is needed. From the aforementioned Li et al. works [19,176,177], the highest UTS of joints with
Co, Ni and Cu interlayers, was of 345, 370 and 520 MPa, respectively. Thus, despite suppressing or decreasing the amount of
brittle intermetallics such as Ti2Ni and Fe2Ti, other brittle compounds can form within the FZ as a consequence of the inter-
layer. One possibility that was not explored properly yet for NiTi/stainless steel dissimilar joints was the use of high temper-
ature interlayers, which could create a physical barrier between both base materials. However, a thorough study of phase
diagrams of the base materials and the envisaged interlayer must be performed in order to select a filler material that does
not create brittle compounds either in the HAZ and FZ.
Fig. 47. Influence of Co interlayer thickness on hardness of NiTi/stainless steel dissimilar joints (adapted from [19]).
Fig. 48. Influence of Cu interlayer thickness on UTS and elongation of dissimilar laser welded NiTi/stainless steel joints (adapted from [177]).
The only study employing high temperature interlayers was done by Ng et al. [174]. In this work the authors used a Ta
interlayer, which, aside from its high melting point, also possesses good corrosion resistance [272], as well as, high biocom-
patibility [96]. Ta interlayer was positioned between the two base materials, in a butt joint configuration. The influence of
the interlayer thickness, and thus, the amount of Ta that was diluted into the FZ was studied and similar findings to those of
the Cu, Co and Ni interlayers were observed: the Ta interlayer partially suppressed the formation of intermetallics such as
CrFe, giving raise to others Ta-based (Cr2Ta and Ni3Ta); increasing the interlayer thickness resulted in lower hardness values
in the FZ. An optimum interlayer thickness of 50 mm was found to increase the mechanical properties when compared to the
dissimilar joint without filler material (250 MPa vs 125 MPa, respectively). However, the tensile strength of the joint is not
high enough to promote the superelastic behaviour, thus limiting the possible applications of such combination.
Due to the possible use of these dissimilar welded joints for biomedical applications, it is important to understand the
corrosion behaviour of the joints. From the studies of Li et al. [177], using Cu interlayer to achieve the higher tensile prop-
erties reported for NiTi/stainless steel joints, Zhang et al. evaluated its susceptibility to static [178] and stress corrosion [179]
in artificial saliva and they observed that Cu release was dependent on the protein amount in the artificial saliva: the higher
the protein content, the lower the Cu release. The influence of protein content on the corrosion behaviour of the welded
joints is crucial if dissimilar joints using Cu interlayers are envisaged in orthodontic devices. Another major issue is that
the corrosion performance of these joints was seen to be highly susceptible to pH: in more acidic environments, the corro-
sion resistance decreased. With applied stresses, the protective oxide films created can be damaged, leading to a significant
release of Cu [179]. Thus, prior to the use of these NiTi/stainless joints, biomedical studies are still required in order to con-
firm that the patients health is not affected by the orthodontic device.
Aside from the study in artificial saliva, Zhang et al. [180] studied the stress corrosion of the dissimilar joints using Cu
interlayer in fluoride medium. With increasing fluoride concentration the corrosion resistance of the joint is decreased. This
effect is exacerbated when stress is imposed to the weld as it damages the oxide protective film and thus the corrosion sus-
ceptibility is increased. However, it must be noticed that the Cu release is below harmful levels for the human body; as such,
it may, in principle, be used as a biomedical device.
Aiming at biomedical applications, namely orthodontic wires, Watanabe et al. [181] welded NiTi to stainless steel by
laser. When compared to autogenous welding of NiTi/NiTi and stainless steel/stainless steel, the dissimilar joints exhibited
very poor mechanical properties, making impossible its use for medical applications. So, it seems that the insertion of filler
materials between NiTi and stainless steel is the only way to enable the use of these dissimilar joints in the biomedical
industry.
Up to now no studies have address the influence of post-weld heat treatments on dissimilar joints. This may be related to
the fact that, NiTi and stainless steels required specific heat treatments to improve its properties. However, this problem may
be overcome by performing local heat treatments in a controlled procedure.
3.2.2. NiTi/Ti6Al4V
Titanium and its alloys are very attractive for several industrial sectors. The most used Titanium alloys is the biphasic a
+ b Ti6Al4V [273]. In the biomedical industry, they are highly considered due to its high specific strength and excellent bio-
compatibility [274], where can substitute hard tissues [275]. In aerospace structures and propulsion systems, this alloys is
commonly used due to the exceptional combination of mechanical and corrosion properties, associated with low weight
[276]. Ti6Al4V can also be used in both chemical/acid, marine and industrial environments, due to is ability to form a pro-
tective oxide in a significant range of temperatures [277]. It is thus evident that joining NiTi to Ti6Al4V would be of great
interest as it could provide a combination of the functional properties presented by NiTi with the mechanical and corrosion
Fig. 49. Eutectic microstructure on the NiTi/Nb interface observed in TEM (adapted from [18]).
properties of Ti6Al4V. However, Ti6Al4V, as all Titanium alloys, is extremely reactive to oxygen [278] and to hydrogen pick-
up [279]. As such, embrittlement of the alloy may occur when welding this material.
Significant metallurgical problems arise when welding NiTi to Ti6Al4V. Assessment of the weldability of NiTi to Ti6Al4V
was investigated by Miranda et al. [182]. Regardless of the heat input introduced during welding, to control the cooling rate
upon solidification, cracking was always observed in the FZ. SEM analysis of fracture surfaces revealed the presence of brittle
Ti2Ni which caused transgranular cracking upon cooling. The formation of Ti2Ni is promoted by the migration of Nickel, from
NiTi, to the liquid Titanium. As such, it is apparent that the solution for overcoming the embrittlement of NiTi/Ti6Al4V
welded joints can be made by introducing an element that either inhibits or decreases the amount of brittle intermetallics
formed in the FZ.
The insertion of an interlayer was tried by Zoeram et al. [183,184], using Cu. The embrittlement of the welded joint was
reduced, as the amount of Ti2Ni decreased when compared to the joint welded without interlayer. Nonetheless, some other
Cu-based intermetallics were formed due to the presence of the Cu interlayer, namely Ti2Cu, Cu2Ti, Cu3Ti and Cu4Ti. Hard-
ness measurements on the FZ revealed higher values than both base materials, despite a significant decrease when compared
to the FZ of the joint welded without the Cu interlayer. The thickness of the interlayer used, and thus the amount of Cu intro-
duced into the FZ, has been seen to play a significant role on the properties of the joints. As the thickness of the interlayer
increased (up to 75 mm), the Cu-rich areas also increased in the FZ. However, further increase of the interlayer thickness pro-
moted formation of shrinkage cavities and increasing the interlayer thickness was found to be detrimental to the joints prop-
erties due to the formation of brittle Cu-based intermetallics. An optimum interlayer thickness was found to be of 75 mm, and
the correspondent welded joint experienced rupture at 300 MPa and 3.3% strain, without any indication of superelastic
plateau.
Oliveira et al. [18], succeeded to avoid the formation of intermetallic compounds in the FZ, while ensuring a sound joint
between NiTi and Ti6Al4V by using a high temperature Nb interlayer. In order to do so, the laser was positioned on the
Ti6Al4V side, which is less prone to degradation of its properties at high temperatures. Fusion and subsequent solidification
of the Ti6Al4V base material ensured joining with the Nb interlayer, without any intermetallic compounds formed. Nb was
incorporated in the Ti alloy by diffusion into the liquid bath. Nb forms an eutectic with NiTi [91] and so, a precise control of
the laser positioning on the Ti side was required so that the heat transfer was enough to promote the eutectic reaction
between NiTi and Nb. The eutectic microstructure observed in the NiTi/Nb interface is depicted in Fig. 49. The use of the high
temperature interlayer ensured a physical barrier to the mixing of both NiTi and Ti6Al4V preventing the formation of brittle
intermetallics which are detrimental to the joint properties. Mechanical characterization of the welded joint revealed an UTS
that matched that of Nb (about 300 MPa). Although no superelastic plateau was observed, this kind of approach may allow
Fig. 50. NiTi stent welded with: Tantalum (left); Gold (middle); Platinum (right) (adapted from [185]).
the use of other high temperature interlayers, which must have an UTS higher than the onset for the superelastic plateau of
the NiTi base material.
Other materials can be used as possible interlayer to promote a favourable joining between NiTi and Ti6Al4V, such as Fe
or Co as they are all known to form shape memory alloys in the form of NiTiX, being X one of those elements [2]. Ideally,
preventing Nickel migration into the liquid Titanium would be preferred as, this way, formation of brittle Ti2Ni would be
inhibited. Thus, in theory, it may be expected that superelastic properties can be presented by the welded joints, allowing
new applications in both aerospace and biomedical industries. The use of high temperature materials, such as Ta or W
may also be considered an interesting option for this dissimilar combination, but it was not yet studied.
3.2.3. NiTi/other engineering materials

Despite most of the work on dissimilar laser welding involving NiTi aims at obtaining NiTi/Stainless steel and NiTi/
Ti6Al4V sound joints, other combinations have been also tested either by laser welding NiTi to pure metals or to specific
metallic alloys.
Haas et al. [129] reported laser welding of NiTi to both Niobium and Tantalum. In the NiTi/Niobium welds, hardness mea-
surements revealed that the FZ had a hardening effect, which was not observed in the NiTi/Tantalum joints. In the latter case,
the FZ revealed a fine grained microstructure, which can be interpreted as a favourable metallurgical compatibility between
both materials. Nonetheless, due to the very high melting points of both Niobium (2477 C) [280] and Tantalum (3017 C)
[280] compared to NiTi (1310 C) [2], process optimization was required.
Tuissi et al. [127] performed butt joints of NiTi to Monel, Inconel and Ni, without filler. In the NiTi/Monel case, transversal
cracks along the thickness of the joint were observed due to the very different thermal conductivities of both base materials.
These differences give raise to thermal stresses during solidification, which cause the initiation of cracks. Cu-based segrega-
tions were observed in the FZ due to the existence of Copper in Monel-400. Tensile test of the welded joint revealed, as
expected, poor mechanical properties (150 MPa) due to the aforementioned cracking phenomena and Cu-based phases.
NiTi/Inconel joints, similarly to the NiTi/Monel ones, exhibited macrofractures in the FZ. Hardness measurements
revealed very high hardness (1000 HV vs 300 HV for NiTi and Inconel), which explained the poor mechanical properties
of the welded joint.
Although in NiTi/Ni dissimilar welds no cracks were observed by optical microscopy, no tensile tests were performed.
Nevertheless, hardness measurements revealed, once more, hardness values above 1000 HV in the FZ. As such, a significant
decrease in the ductility of the dissimilar joint was expected.
NiTi, aside from its functional properties, presents also good radiopacity [234]. Nonetheless, the methods to enhance the
visibility under X-ray radiation may include the addition of noble metals, such as gold or platinum [281]. Tantalum is used in
biomedical devices and can improve the radiopacity of other materials [282]. Steegmuller et al. [185] presented NiTi stents
welded with Au, Pt or Ta to improve the radiopacity of the welded material. Smooth surfaces were obtained and no gaps
between the NiTi and the other elements were discernible, as shown in Fig. 50. Increased radiopacity of the welded stents
was confirmed for all dissimilar combinations.
The corrosion behaviour of the welded stents was found to be intimately related to the dissimilar combination chosen.
While the NiTi stent welded with Ta presented a corrosion behaviour similar to the NiTi base material, for the other com-
binations pitting corrosion occurred at a low breakdown potential. As such, it appeared that only NiTi/Ta welded stents
Table 3
Overview of dissimilar laser welded combinations involving NiTi.
Dissimilar combination Base material shape Joint configuration Filler material/interlayer Reference
NiTi/steel Wire Butt n.a. [128]
NiTi/stainless steel Wire Butt n.a. [17,143,175,181]
Wire Butt Co [19]
Plate Overlap Ni [172]
Wire Butt Ni [176]
Wire Butt Cu [177180]
Wire Butt Ta [174]
NiTi/Ti6Al4V Plate Butt n.a. [182]
Plate Butt Cu [183,184]
Plate Butt Nb [18]
NiTi/Monel Plate Butt n.a. [127]
NiTi/Inconel Plate Butt n.a.
NiTi/Ni Plate Butt n.a.
NiTi/Ta Plate Butt n.a. [129]
NiTi/Co-Cr Wire Butt n.a. [181]
NiTi/X Stents Overlap n.a. [185]
X = Ta, Au, Pt
NiTi/MP35N Wire Butt n.a. [186]
NiTi/Cu Wire Overlap n.a. [187,191]
can be used to improve of the radiopacity of such devices. The fatigue behaviour of these welded joints is of great interest to
be studied as it can give a good indication on the viability of using such dissimilar combination for applications in stents.
In the biomedical industry orthodontic wires can be made of NiTi, stainless steel, b-Titanium and based on Cobalt-
Chromium [283]. Each of these wires has different characteristics, such as Youngs modulus and its ability to be manipulated.
Within the mouth of a patient, sometimes it is required that the orthodontic wires can exert different actuating forces along
wires [284]. As such, it would be beneficial to use the different mechanical properties of the orthodontic wires in order to
obtain a biomedical device that could be more suitable for the patient requirements. Watanabe et al. [181] reported such
study. Aside from welding NiTi orthodontic wires to stainless steel, dissimilar laser welded joints of NiTi/Co-Cr and NiTi/
b-Ti were also performed. When compared to the homogenous similar combinations (NiTi/NiTi, Co-Cr/Co-Cr, stainless
steel/stainless steel and b-Ti/b-Ti), the dissimilar joints presented a decrease in mechanical properties, measured by
three-point bending tests. This decrease was related to the low affinity between NiTi and stainless steel and Co-Cr wires,
thus, not promoting a suitable microstructure. Although NiTi/b-Ti joints presented better mechanical properties than the
other dissimilar combinations, these were not enough to ensure its appropriate use in clinical environments.
NiTi was also laser welded to MP35N (an age hardened Ni-Co-based alloy) by Panton et al. [186]. The composition of the
FZ was seen to be affected by laser offsetting. Solid solution strengthening in the FZ and formation of Ti-based intermetallics
(such as Ti2Ni, Ti2Co and/or Cr2Ti) were pointed out as responsible for a significant increase of the hardness in the FZ, when
compared to the base materials. Titanium mixing with the alloying element of MP35N was minimized when the laser was
positioned on the MP35N base material, resulting in lower hardness values when compared to the joint made with the laser
positioned on the materials interface or offset to the NiTi side. Macroscopic cracks were observed in the FZ, under optical
microscopy, when the laser was positioned on the NiTi side or at the contact interface between both materials. However,
when the laser was positioned on the MP35N side, micro-cracks could only be observed by SEM imaging. These solidification
cracks, together with the formation of brittle precipitates and increased hardness of the FZ, explained the decrease of the
mechanical properties of the welded joint (66% of the base material).
Recently, Zheng et al. reported the first successful study on laser welding of NiTi to Copper [187]. By selecting suitable
welding parameters, the authors were able to improve the mechanical properties of the joint. The tensile strength was above
that of the superelastic plateau of NiTi. As such, superelasticity can be used in this dissimilar combination. As far as the pos-
sible applications are concerned, such ability to promote sound joining between these two materials is of major relevance,
especially for integration of NiTi into electromechanical systems. Non-conventional geometries for sensors or actuators can
thus be designed taking profit of the freedom that this welding technology permits in project and design. Table 3 summa-
rizes the existing studies on dissimilar laser welding of NiTi.
4. Applications
Owing to the interesting properties of the laser welded NiTi joints, applications have been developed while others are
currently under study and development.
Although laser welding of NiTi to itself and to other engineering materials has been studied for several years now, only
few applications have been developed. However, over the last years an increasing trend on using laser welding for NiTi has
widespread the use of the welded joints (both similar and dissimilar), especially for the biomedical industry. As such, some of
the more relevant applications (already developed but still under development) are briefly presented.
A microforcep for endoscopic brain surgery was presented in [129,188]. Fatigue testing of the laser welded microforcep
was compared to the monolithic forcep and no significant differences were found: the life cycle of the laser welded device
was not affected by the welding procedure. Haas et al. [129] also presented laser welded NiTi tubes with different diameters
(0.8 and 0.9 mm) to be used as catheters. The ability to obtain catheters with different diameters along its length may
decrease the difficulty to obtain these devices with variable diameters [285], while reducing production costs by decreasing
the manufacturing time provided by laser welding.
Although NiTi has good radiopacity properties [234], researchers have tried to improve those properties by joining NiTi to
Ta, Au and Pt as mentioned and shown before (Fig. 50) [185,189]. The aim was to produce NiTi stents that could be more
easily seen during medical examinations. Only Tantalum has proven to be a suitable choice as, despite the other elements
increase the radiopacity of the welded joints, its corrosion behaviour in the human body was poor, which was not the case
with NiTi/Ta joints.
Several studies have been focused on laser welding orthodontic wires, either in similar or dissimilar joints. The reason for
these attempts lies on the fact that different types of orthodontic wires have different mechanical characteristics, and joining
such distinct materials may improve the treatment of the patient, as to correct different tooth movements requires different
actuation forces [284]. Such distinct actuation was observed by Sevilla et al. [145], when welding NiTi to NiTiCu. The joints
presented two superelastic plateaus in the stress-strain curve which corresponded to the different base materials used.
Although the presence of Cu in the welded joint decreased its corrosion properties, the Ni release was considered to be safe
for the human health. Similarly, two step plateaus were evidence by the same authors when joining NiTi with different com-
positions and hence with different onsets for the stress induced transformation. It was thus proved that laser welding can be
efficiently used for joining different orthodontic wires. However, the choice of the dissimilar combination greatly affects the
mechanical properties of these joints: while for NiTi/NiTiCu good mechanical properties were obtained, the same did not
occur for NiTi/stainless steel, NiTi/Co-Cr and NiTi/b-Ti [181] due to low affinity of NiTi to the other elements. The character-
istics of laser welded NiTi orthodontics wires was also studied and considered to be acceptable for clinical use [141].
The problem of laser welding NiTi to stainless steel suffered a great turn around with the results from Li et al. [177] using
a Cu interlayer. The superior mechanical properties and the fact that the joint strength was above the superelastic plateau for
the stress induced transformation are of great importance. In order to understand the viability of such dissimilar joints for
biomedical applications, corrosion tests were performed. Despite the corrosion properties of the welded joints decreased
with increasing fluoride content and in more acidic environments, the Cu release was below harmful levels than can be dan-
gerous for human beings. However, prior to being able to use such welded joints in orthodontic applications (or others), a
better mechanical characterization, namely fatigue strength, is still required.
Some interesting properties presented by similar NiTi joints can be used in the future for sensors or actuators. In partic-
ular, when joining NiTi with different compositions [126,156], the shape memory effect of the welds can be depicted in two
steps, as the different compositions of the base materials give raise to different transformation temperatures. The distinct
shape memory effect presented by the thermal affected regions when compared to the original base material [168] may also
be used for such applications.
One final note regarding the use of laser in NiTi: Hagiwara et al. [190] presented a novel method to remove debris from
broken endodontic NiTi files. Basically, the authors had an extractor tube, with an optical fibre inside. The broken endodontic
files were positioned inside the extractor and the laser was fired promoting joining between the NiTi file and the extractor.
The retention force, which is required to remove the debris inside the root canals, was measured and considered to be suf-
ficient to be used in patients. An interesting issue about this study was the fact that the oral temperature only increase by
4 C due to the laser; as such, no harm would be done to the patient. This innovative technique may soon start being used in
dental clinics, as fracture of such NiTi files inside root canals is considered a problem of major importance for the industry,
the clinics and the patients [286,287].
5. Summary and future outlook
In this paper, a comprehensive review on welding NiTi is presented addressing the following aspects: welding processes,
influence of process parameters, microstructural transformations, mechanical performance, functional properties, corrosion
behaviour and industrial applications.
NiTi has a set of weldability problems mostly related to: grain size heterogeneities along the joint; formation of brittle
intermetallics; and compositional variations in the thermal affected zones due to preferential volatilization (in the fusion
zone) and precipitation phenomena (in the heat affected zone).
Laser welding is, by far, the most studied process due to minimal thermal effect on the base materials, and thus a large
number of experimental studies exist and have been described. So, a detailed discussion was devoted to this process, pre-
senting the characteristics of the microstructural changes observed, under a materials science perspective, and its influence
on the mechanical properties of the welds, evaluated by tensile, bending and cycling tests and shape memory effect.
Concerning dissimilar welding of NiTi, there is still limited research work on this topic. Since both mechanical and
thermo-physical properties of materials involved are quite different, successful joining mostly requires the use of interlayers.
These interlayers aim at reducing or prevent mixing of the liquids of the base materials involved in the fusion zone and for-
mation of brittle intermetallics. For this purpose, high melting temperature materials have to be used. However, few types of
interlayers have, so far, been tested and the existing knowledge is mainly experimental.
Despite the ongoing research on welding and joining shape memory alloys, some areas still need to be addressed and
these are:
Similar joints
As discussed in Section 3, limited experiments have been performed on in-situ analysis of the mechanical behaviour of
welded joints. This would be beneficial to provide further understanding on the load transfer mechanisms in the dif-
ferent regions of the joints, based on micromechanical models that take into account the microstructural changes
induced by the welding process.
The authors agree that, in the future, joining can be used to simultaneously produce structures with a controlled func-
tional gradient of transformation characteristics.
Dissimilar joints
It was observed that joining NiTi to other materials is not effective in the absence of a filler material. So, new filler
materials/interlayers have to be studied to increase joining efficiency and both mechanical and corrosion behaviours.
A better knowledge on the joining mechanisms, namely through the use of thermodynamic calculations, will allow to
optimize materials selection for the interlayer.
The ability to join dissimilar combinations by solid state processes could be a viable alternative to fusion-based pro-
cesses, as long as, the functional properties of NiTi can be preserved.
Joining NiTi to Cu-based shape memory alloys, may also be of great interest in order to get an interesting combination
of functional characteristics, alongside with high thermal and electrical conductivity, typical of Cu-based systems.
Further research on new base material combinations, involving distinct shape memory alloys, to produce parts for
high added-value applications, is also a challenge.
Acknowledgements
JPO and FMBF acknowledge funding by National Funds through the Portuguese Foundation for Science and Technology
(FCT-MCTES), Reference UID/CTM/50025/2013 and FEDER funds through the COMPETE 2020 Programme under the project
number POCI-01-0145-FEDER-007688. RMM acknowledges FCT-MCTES for its financial support via the project PEst-OE/
EME/UI0667/2014 (UNIDEMI). JPO acknowledges FCT-MCTES for funding PhD grant SFRH/BD/85047/2012.
References
[1] Van Humbeeck J. Non-medical applications of shape memory alloys. Mater Sci Eng, A 1999;273275:13448. doi: http://dx.doi.org/10.1016/S0921-
5093(99)00293-2.
[2] Otsuka K, Ren X. Physical metallurgy of Ti-Ni-based shape memory alloys. Prog Mater Sci 2005;50:511678. doi: http://dx.doi.org/10.1016/j.
pmatsci.2004.10.001.
[3] Bram M, Ahmad-Khanlou A, Heckmann A, Fuchs B, Buchkremer HP, Stver D. Powder metallurgical fabrication processes for NiTi shape memory alloy
parts. Mater Sci Eng, A 2002;337:25463. doi: http://dx.doi.org/10.1016/S0921-5093(02)00028-X.
[4] Liu YY, Liu YY, Van Humbeeck J, Humbeeck JVAN. Two-way shape memory effect developed by martensite deformation in NiTi. Acta Mater
1998;47:199209. doi: http://dx.doi.org/10.1016/S1359-6454(98)00325-5.
[5] Dolce M, Cardone D. Mechanical behaviour of shape memory alloys for seismic applications 2. Austenite NiTi wires subjected to tension. Int J Mech Sci
[6] Lagoudas DC. Shape memory alloys: modeling and engineering applications. New York: Springer; 2008. http://dx.doi.org/10.1007/978-0-387-47685-
8.
[7] Allafi JK, Ren X, Eggeler G. The mechanism of multistage martensitic transformations in aged Ni-rich NiTi shape memory alloys. Acta Mater
[8] Seo J, Kim Y, Hu J. Pilot study for investigating the cyclic behavior of slit damper systems with recentering shape memory alloy (SMA) bending bars
used for seismic restrainers. Appl Sci 2015;5:187208. doi: http://dx.doi.org/10.3390/app5030187.
[9] Liu Y, Mahmud A, Kursawe F, Nam TH. Effect of pseudoelastic cycling on the Clausius-Clapeyron relation for stress-induced martensitic
transformation in NiTi. J Alloys Compd 2008;449:827. doi: http://dx.doi.org/10.1016/j.jallcom.2006.02.080.
[10] 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;50:425574. doi: http://dx.doi.org/10.1016/s1359-6454(02)00257-4.
[11] Wagner MFX, Dey SR, Gugel H, Frenzel J, Somsen C, Eggeler G. Effect of low-temperature precipitation on the transformation characteristics of Ni-rich
NiTi shape memory alloys during thermal cycling. Intermetallics 2010;18:11729. doi: http://dx.doi.org/10.1016/j.intermet.2010.02.048.
[12] Nishida M, Wayman CM, Honma T. Precipitation processes in near-equiatomic TiNi shape memory alloys. Metall Trans A 1986;17:150515. doi:
http://dx.doi.org/10.1007/BF02650086.
[13] Khalil-Allafi J, Eggeler G, Dlouhy A, Schmahl WW, Somsen C. On the influence of heterogeneous precipitation on martensitic transformations in a Ni-
rich NiTi shape memory alloy. Mater Sci Eng, A 2004;378:14851. doi: http://dx.doi.org/10.1016/j.msea.2003.10.335.
[14] Weinert K, Petzoldt V. Machining NiTi micro-parts by micro-milling. Mater Sci Eng, A 2008;481482:6725. doi: http://dx.doi.org/10.1016/j.
msea.2006.10.220.
[15] Mills KC, Keene BJ, Brooks RF, Shirali A. Marangoni effects in welding. Philos Trans R Soc A: Math Phys Eng Sci 1998;356:91125. doi: http://dx.doi.
org/10.1098/rsta.1998.0196.
[16] Hall PC. Laser welding nitinol to stainless steel. In: Wang G, editor. Proc SMST 1997.
[17] Vannod J, Bornert M, Bidaux J-E, Bataillard L, Karimi A, Drezet J-M, et al. Mechanical and microstructural integrity of nickeltitanium and stainless
steel laser joined wires. Acta Mater 2011;59:653846. doi: http://dx.doi.org/10.1016/j.actamat.2011.06.031.
[18] Oliveira JP, Panton B, Zeng Z, Andrei CM, Zhou Y, Miranda RM, et al. Laser joining of NiTi to Ti6Al4V using a niobium interlayer. Acta Mater
2016;105:915. doi: http://dx.doi.org/10.1016/j.actamat.2015.12.021.
[19] Li H, Sun D, Cai X, Dong P, Gu X. Laser welding of TiNi shape memory alloy and stainless steel using Co filler metal. Opt Laser Technol 2013;45:45360.
doi: http://dx.doi.org/10.1016/j.optlastec.2012.06.010.
[20] Van Der Eijk C, Fostervoll H, Sallom ZK, Akselsen OM. Plasma welding of NiTi to NiTi, stainless steel and hastelloy C276. Solutions 2003:135.
[21] Chau ETF, Friend CM, Allen DM, Hora J, Webster JR. A technical and economic appraisal of shape memory alloys for aerospace applications. Mater Sci
Eng, A 2006;438440:58992. doi: http://dx.doi.org/10.1016/j.msea.2006.02.201.
[22] Kutz M. Mechanical engineers handbook. 3rd ed. Hoboken: John Wiley & Sons Inc.; 2014. doi: http://dx.doi.org/10.1002/9781118985960.
[23] Lancaster JF. Metallurgy of welding. 6th ed. Cambridge: Woodhead Publishing Limited; 1999.
[24] Thomas M, Nicholas E, Needham J, Murch M, Templesmith M, Dawes C. Improvements relating to friction welding; 1991.
[25] He X, Gu F, Ball A. A review of numerical analysis of friction stir welding. Prog Mater Sci 2014;65:166. doi: http://dx.doi.org/10.1016/j.
pmatsci.2014.03.003.
[26] Santos TG, Miranda RM, Vilaa P. Friction stir welding assisted by electrical Joule effect. J Mater Process Technol 2014;214:212733. doi: http://dx.
doi.org/10.1016/j.jmatprotec.2014.03.012.
[27] Liu X, Wu C, Padhy GK. Characterization of plastic deformation and material flow in ultrasonic vibration enhanced friction stir welding. Scripta Mater
2015;102:958. doi: http://dx.doi.org/10.1016/j.scriptamat.2015.02.022.
[28] Buehler J, Wiley R. The properties of TiNi and associated phases; 1961. <http://oai.dtic.mil/oai/oai?verb=getRecord&metadataPrefix=html&identifier=
AD0266607>.
[29] Ikai A, Kimura K, Tobushi H. TIG welding and shape memory effect of TiNi shape memory alloy. J Intell Mater Syst Struct 1996;7:64655. doi: http://
dx.doi.org/10.1177/1045389X9600700604.
[30] Yunlian Q, Ju D, Quan H, Liying Z. Electron beam welding, laser beam welding and gas tungsten arc welding of titanium sheet. Mater Sci Eng, A
[31] Hsu YT, Wang YR, Wu SK, Chen C. Effect of CO2 laser welding on the shape-memory and corrosion characteristics of TiNi alloys. Metall Mater Trans A
2001;32:56976. doi: http://dx.doi.org/10.1007/s11661-001-0073-2.
[32] Oliveira JP, Barbosa D, Fernandes FMB, Miranda RM. Tungsten inert gas (TIG) welding of Ni-rich NiTi plates: functional behavior. Smart Mater Struct
2016;25:03LT01. doi: http://dx.doi.org/10.1088/0964-1726/25/3/03LT01.
[33] Fox G, Hahnlen R, Dapino MJ. Fusion welding of nickel-titanium and 304 stainless steel tubes: part II: tungsten inert gas welding. J Intell Mater Syst
Struct 2013;24:96272. doi: http://dx.doi.org/10.1177/1045389X12461076.
[34] L SX, Yang ZL, Dong HG. Welding of shape memory alloy to stainless steel for medical occluder. Trans Nonferr Met Soc Chin Engl Ed 2013;23:15660.
doi: http://dx.doi.org/10.1016/S1003-6326(13)62442-5.
[35] Weng J, Chang J, Chen K, He J. Solid/liquid erosion behavior of gas tungsten arc welded TiNi overlay. Wear 2003;255:21924. doi: http://dx.doi.org/
10.1016/S0043-1648(03)00120-0.
[36] Yang D, Jiang HC, Zhao MJ, Rong LJ. Microstructure and mechanical behaviors of electron beam welded NiTi shape memory alloys. Mater Des
2014;57:215. doi: http://dx.doi.org/10.1016/j.matdes.2013.12.039.
[37] Yang D, Jiang HC, Zhao MJ, Rong LJ. Effect of post-weld annealing on microstructure and properties of NiTi welding joints. Mater Res Innov 2014;18.
doi: http://dx.doi.org/10.1179/1432891714Z.000000000751. S4-588S4-591.
[38] Steen WM, Mazumder J. Laser material processing. London: Springer; 2010. http://dx.doi.org/10.1007/978-1-84996-062-5.
[39] Quintino L, Costa A, Miranda R, Yapp D, Kumar V, Kong CJ. Welding with high power fiber lasers a preliminary study. Mater Des 2007;28:12317.
doi: http://dx.doi.org/10.1016/j.matdes.2006.01.009.
[40] Tam B, Pequegnat A, Khan MI, Zhou Y. Resistance microwelding of Ti-55.8 wt pct Ni nitinol wires and the effects of pseudoelasticity. Metall Mater
Trans A: Phys Metall Mater Sci 2012;43:296978. doi: http://dx.doi.org/10.1007/s11661-012-1115-7.
[41] Delobelle V, Delobelle P, Liu Y, Favier D, Louche H. Resistance welding of NiTi shape memory alloy tubes. J Mater Process Technol 2013;213:113945.
doi: http://dx.doi.org/10.1016/j.jmatprotec.2013.01.013.
[42] Blazynski T. Explosive welding, forming and compaction. Netherlands, Dordrecht: Springer; 1983. http://dx.doi.org/10.1007/978-94-011-9751-9.
[43] Zimmerly CA, Inal OT, Richman RH. Explosive welding of a near-equiatomic nickel-titanium alloy to low-carbon steel. Mater Sci Eng, A
1994;188:2514. doi: http://dx.doi.org/10.1016/0921-5093(94)90379-4.
[44] Richman R, Associates D, Kew M. Cavitation erosion of NiTi explosively welded to steel. Wear 1995;181183:805. doi: http://dx.doi.org/10.1016/
0043-1648(94)07029-6.
[45] Richman RH, McNaughton WP. A metallurgical approach to improved cavitation-erosion resistance. J Mater Eng Perform 1997;6:63341. doi: http://
dx.doi.org/10.1007/s11665-997-0057-5.
[46] Prummer R, Stockel D. Nitinol stainless steel compound materials made by explosive welding. In: Staudhammer K, Murr L, editors. Proc 2000 int
conf fundam issues appl shock-wave high-strain-rate phenom. Elsevier; 2001. p. 5815.
[47] Belyaev S, Rubanik V, Resnina N, Rubanik V, Rubanik O, Borisov V. Martensitic transformation and physical properties of steelTiNi bimetal
composite, produced by explosion welding. Phase Transitions 2010;83:27683. doi: http://dx.doi.org/10.1080/01411591003656757.
[48] Belyaev SP, Rubanik VV, Resnina NN, Rubanik VV, Rubanik OE. Effect of annealing on martensitic transformations in steel TiNi alloy explosion
welded bimetallic composite. Met Sci Heat Treat 2011;52:4326. doi: http://dx.doi.org/10.1007/s11041-010-9310-x.
[49] Belyaev S, Rubanik V, Resnina N, Rubanik Jr V, Rubanik O, Borisov V, et al. Functional properties of bimetal composite of stainless steelTiNi alloy
produced by explosion welding. Phys Proc 2010;10:527. doi: http://dx.doi.org/10.1016/j.phpro.2010.11.074.
[50] Xing T, Zheng Y, Cui L. Transformation and damping characteristics of NiTi/NiTi alloys synthesized by explosive welding. Mater Trans
2006;47:65860. doi: http://dx.doi.org/10.2320/matertrans.47.658.
[51] Li J, Zheng Y, Cui L. Transformation characteristics of TiNi/TiNi alloys synthesized by explosive welding. Front Mater Sci Chin 2007;1:3515. doi:
http://dx.doi.org/10.1007/s11706-007-0065-2.
[52] Juntao L, Yanjun Z, Lishan C. Effects of severe plastic deformation and heat treatment on transformation behavior of explosively welded duplex
TiNiTiNi. Pet Sci 2007;4:10712. doi: http://dx.doi.org/10.1007/BF03187464.
[53] Yan Z, Cui L, Zhen Y. Microstructure and martensitic transformation behaviors of explosively welded NiTi/NiTi laminates. Chin J Aeronaut
[54] Xing T, Zheng Y, Cui L, Mi X. Influence of aging on damping behavior of TiNi/TiNi alloys synthesized by explosive welding. Trans Nonferr Met Soc Chin
[55] Jiang X, Jiang D, Zheng Y, Cui L. Transformation behavior of explosively welded TiNi/TiNi laminate after diffusion annealing and aging. Mater Res Bull
2013;48:50335. doi: http://dx.doi.org/10.1016/j.materresbull.2013.04.031.
[56] No A, De Montre EP, Supe EDT, Piedboeuf MC, Gauvin R, Thomas M. Damping behaviour of shape memory alloys: strain amplitude, frequency and
temperature effects. J Sound Vib 1998;214:885901. doi: http://dx.doi.org/10.1006/jsvi.1998.1578.
[57] Clifton RJ. Response of materials under dynamic loading. Int J Solids Struct 2000;37:10513. doi: http://dx.doi.org/10.1016/S0020-7683(99)00082-7.
[58] Yilbas BS, Sahin AZ. Friction welding thermal and metallurgical characteristics. Berlin, Heidelberg: Springer; 2014. http://dx.doi.org/10.1007/978-3-
642-54607-5.
[59] Blau P. Friction science and technology from concepts to applications, 2nd ed. CRC Press; 2008. http://dx.doi.org/10.1201/9781420054101.
[60] Vairis A, Frost M. High frequency linear friction welding of a titanium alloy. Wear 1998;217:11731. doi: http://dx.doi.org/10.1016/S0043-1648(98)
00145-8.
[61] Sketchley P, Threadgill P, Wright I. Rotary friction welding of an Fe3Al based ODS alloy. Mater Sci Eng, A 2002;329331:75662. doi: http://dx.doi.org/
10.1016/S0921-5093(01)01656-2.
[62] Potente H, Uebbing M. Friction welding of polyamides. Polym Eng Sci 1997;37:72637. doi: http://dx.doi.org/10.1002/pen.11716.
[63] DAlvise L, Massoni E, Walloe SJ, DAlvise L, Walle SJ. Finite element modelling of the inertia friction welding process between dissimilar materials. J
Mater Process Technol 2002;125:38791. doi: http://dx.doi.org/10.1016/S0924-0136(02)00349-7.
[64] Nandan R, Debroy T, Bhadeshia HKDH. Recent advances in friction-stir welding process, weldment structure and properties. Prog Mater Sci
2008;53:9801023. doi: http://dx.doi.org/10.1016/j.pmatsci.2008.05.001.
[65] Mishra RS, Ma ZY. Friction stir welding and processing. Mater Sci Eng R Rep 2005;50:178. doi: http://dx.doi.org/10.1016/j.mser.2005.07.001.
[66] Shinoda T, Tsuchiyta T, Takahashi H. Functional characteristics of friction welded near-equiatomic TiNi shape memory alloy. Trans Jpn Weld Soc
1991;22:1028. , <http://ci.nii.ac.jp/naid/110003380338/>.
[67] Shinoda T, Tsuchiya T, Takahashi H. Friction welding of shape memory alloy. Weld Int 1992;6:205. doi: http://dx.doi.org/10.1080/
09507119209548136.
[68] Shinoda T, Owa T, Magula V. Microstructural analysis of friction welded joints in TiNi alloy. Weld Int 1999;13:1805. doi: http://dx.doi.org/10.1080/
09507119909447361.
[69] London B, Fino J, Pelton A, Fuller C, Mahoney M. Friction stir processing of nitinol. In: Jata K, Mahoney M, Mishra R, Lienert T, editors. Frict stir weld
process III. Dover; 2005. p. 6774.
[70] Barcellona A, Fratini L, Palmeri D, Maletta C, Brandizzi M. Friction stir processing of NiTi shape memory alloy: microstructural characterization. Int J
Mater Form 2010;3:104750. doi: http://dx.doi.org/10.1007/s12289-010-0950-6.
[71] Dixit M, Newkirk JW, Mishra RS. Properties of friction stir-processed Al 1100NiTi composite. Scripta Mater 2007;56:5414. doi: http://dx.doi.org/
10.1016/j.scriptamat.2006.11.006.
[72] Ni DR, Wang JJ, Zhou ZN, Ma ZY. Fabrication and mechanical properties of bulk NiTip/Al composites prepared by friction stir processing. J Alloys
Compd 2014;586:36874. doi: http://dx.doi.org/10.1016/j.jallcom.2013.10.013.
[73] Oliveira JP, Duarte JF, Incio P, Schell N, Miranda RM, Santos TG. Production of Al/NiTi composites by friction stir welding assisted by electrical
current. Mater Des 2017;113:3118. doi: http://dx.doi.org/10.1016/j.matdes.2016.10.038.
[74] Fukumoto S, Inoue T, Mizuno S, Okita K, Tomita T, Yamamoto A. Friction welding of TiNi alloy to stainless steel using Ni interlayer. Sci Technol Weld
Join 2010;15:12430. doi: http://dx.doi.org/10.1179/136217109X12577814486692.
[75] Kong CY, Soar RC, Dickens PM. Ultrasonic consolidation for embedding SMA fibres within aluminium matrices. Compos Struct 2004;66:4217. doi:
http://dx.doi.org/10.1016/j.compstruct.2004.04.064.
[76] Friel RJJ, Harris RAA. Ultrasonic additive manufacturing a hybrid production process for novel functional products. Proc CIRP 2013;6:3540. doi:
http://dx.doi.org/10.1016/j.procir.2013.03.004.
[77] Hahnlen R, Dapino MJ, Short M, Graff K. Aluminum-matrix composites with embedded Ni-Ti wires by ultrasonic consolidation. Proc SPIE 2009;7290.
doi: http://dx.doi.org/10.1117/12.817036. 729009729009-12.
[78] Habenicht G. Applied adhesive bonding. 1st ed. Weinheim: Wiley-VCH Verlag GmbH & Co. KGaA; 2008. doi: http://dx.doi.org/10.1002/
9783527626458.
[79] Ebnesajjad S. Surface treatment of materials for adhesive bonding. 2nd ed. Amsterdam: Elsevier; 2014. doi: http://dx.doi.org/10.1016/B978-0-323-
26435-8.00004-6.
[80] Brockmann W, Gei PL, Klingen J, Schrder B. Adhesive bonding, 1st ed. Weinheim: Wiley-VCH Verlag GmbH & Co. KGaA. http://dx.doi.org/10.1002/
9783527623921.
[81] Dillard D. Advances in structural adhesive bonding. 1st ed. Cambridge: Woodhead Publishing Limited; 2010.
[82] Man HC, Zhao NQ. Enhancing the adhesive bonding strength of NiTi shape memory alloys by laser gas nitriding and selective etching. Appl Surf Sci
2006;253:1595600. doi: http://dx.doi.org/10.1016/j.apsusc.2006.02.057.
[83] Niccoli F, Alfano M, Bruno L, Furgiuele F, Maletta C. Mechanical and functional properties of nickel titanium adhesively bonded joints. J Mater Eng
Perform 2014;23:238590. doi: http://dx.doi.org/10.1007/s11665-014-0890-2.
[84] Jacobson D, Humpston G. Principles of brazing. 1st ed. Ohio: ASM International; 2005.
[85] Schwartz M. Brazing. 2nd ed. Ohio: ASM International; 2003. doi: http://dx.doi.org/10.1036/1097-8542.094100.
[86] Tetsui T. Effects of brazing filler on properties of brazed joints between TiAl and metallic materials. Intermetallics 2001;9:25360. doi: http://dx.doi.
org/10.1016/S0966-9795(00)00129-1.
[87] Nakao Y, Shinozaki K, Hamada M. Diffusion bonding of intermetallic compound TiAl. ISIJ Int 1991;31:12606. doi: http://dx.doi.org/10.2355/
isijinternational.31.1260.
[88] Shiue R, Wu S, Chen S. Infrared brazing of TiAl intermetallic using BAg-8 braze alloy. Acta Mater 2003;51:19912004. doi: http://dx.doi.org/10.1016/
S1359-6454(02)00606-7.
[89] Wang RR, Welsch GE. Joining titanium materials with tungsten inert gas welding, laser welding, and infrared brazing. J Prosthet Dent
[90] Yang TY, Shiue RK, Wu SK. Infrared brazing of Ti50Ni50 shape memory alloy using pure Cu and Ti15Cu15Ni foils. Intermetallics 2004;12:128592.
doi: http://dx.doi.org/10.1016/j.intermet.2004.03.020.
[91] Grummon DS, Shaw JA, Foltz J. Fabrication of cellular shape memory alloy materials by reactive eutectic brazing using niobium. Mater Sci Eng, A
2006;438440:11138. doi: http://dx.doi.org/10.1016/j.msea.2006.03.113.
[92] Grummon DS, Low K-B, Foltz J, Shaw Ja. A new method for brazing nitinol based on the quasibinary TiNi-Nb system. In: 48th AIAA/ASME/ASCE/AHS/
ASC struct struct dyn mater conf; 2007. p. 17.
[93] Shaw JA, Grummon DS, Foltz J. Superelastic NiTi honeycombs: fabrication and experiments. Smart Mater Struct 2007;16:S1708. doi: http://dx.doi.
org/10.1088/0964-1726/16/1/S17.
[94] Shaw JA, Churchill C, Triantafyllidis N, Michailidis P, Grummon D, Foltz J. Shape memory alloy honeycombs: experiments & simulation. In: 48th AIAA/
ASME/ASCE/AHS/ASC struct struct dyn mater conf, vol. 1; 2007. p. 42836. <http://pdf.aiaa.org/preview/CDReadyMSDM07_1473/PV2007_1739.pdf>.
[95] Wang L, Wang C, Dunand DC. Microstructure and strength of NiTi-Nb eutectic braze joining NiTi wires. Metall Mater Trans A 2015;46:14336. doi:
http://dx.doi.org/10.1007/s11661-015-2781-z.
[96] Matsuno H, Yokoyama A, Watari F, Uo M, Kawasaki T. Biocompatibility and osteogenesis of refractory metal implants, titanium, hafnium, niobium,
tantalum and rhenium. Biomaterials 2001;22:125362. doi: http://dx.doi.org/10.1016/S0142-9612(00)00275-1.
[97] Siegert W, Neuking K, Mertmann M, Eggeler GF. Influence of Nb content and processing conditions on microstructure and functional properties of
NiTiNb shape-memory alloys. Mater Sci Forum 2002;394395:3614. doi: http://dx.doi.org/10.4028/www.scientific.net/MSF.394-395.361.
[98] Zhao X, Yan X, Yang Y, Xu H. Wide hysteresis NiTi(Nb) shape memory alloys with low Nb content (4.5 at.%). Mater Sci Eng, A 2006;438440:5758.
doi: http://dx.doi.org/10.1016/j.msea.2006.03.110.
[99] Shiue RH, Wu SK. Infrared brazing of Ti50Ni50 shape memory alloy using two AgCuTi active braze alloys. Intermetallics 2006;14:6308. doi: http://
dx.doi.org/10.1016/j.intermet.2005.10.012.
[100] van der Eijk C, Sallom ZK, Akselsen OM. Microwave brazing of NiTi shape memory alloy with AgTi and AgCuTi alloys. Scripta Mater
[101] Zhao XK, Tang JW, Lan L, Haung JH, Zhang H, Wang Y. Vacuum brazing of NiTi alloy by AgCu eutectic filler. Mater Sci Technol 2009;25:14957. doi:
http://dx.doi.org/10.1179/174328409X405625.
[102] Gale WF, Guan Y. Microstructural development in copper-interlayer transient liquid phase bonds between martensitic NiAl and NiTi. J Mater Sci
1997;32:35764.
[103] Oikawa K, Wulff L, Iijima T, Gejima F, Ohmori T, Fujita A, et al. Promising ferromagnetic NiCoAl shape memory alloy system. Appl Phys Lett
2001;79:32902. doi: http://dx.doi.org/10.1063/1.1418259.
[104] Li MG, Sun DXQ, Qiu XM, Sun DXQ, Yin SQ. Effects of laser brazing parameters on microstructure and properties of TiNi shape memory alloy and
stainless steel joint. Mater Sci Eng, A 2006;424:1722. doi: http://dx.doi.org/10.1016/j.msea.2006.01.054.
[105] Li MG, Sun DQ, Qiu XM, Yin SQ. Corrosion behavior of the laser-brazed joint of TiNi shape memory alloy and stainless steel in artificial saliva. Mater
Sci Eng, A 2006;441:2717. doi: http://dx.doi.org/10.1016/j.msea.2006.08.052.
[106] Qiu XM, Li MG, Sun DQ, Liu WH. Study on brazing of TiNi shape memory alloy with stainless steels. J Mater Process Technol 2006;176:812. doi:
http://dx.doi.org/10.1016/j.jmatprotec.2005.06.004.
[107] Newman GV. Epoxy adhesives for orthodontic attachments: progress report. Am J Orthod 1965;51:90112. doi: http://dx.doi.org/10.1016/0002-9416
(65)90203-4.
[108] Satoh G, Yao L. Laser autogeneous brazing a new method for joining dissimilar metals. Proc Int Congr Appl Lasers Electro-Optics 2011:31524.
[109] Satoh G, Brandal GB, Naveed S, Yao YL. Laser autogenous brazing of biocompatible, dissimilar metals in tubular geometries. In: Proceeding north am
manuf res inst; 2013. p. 110.
[110] Zhao XK, Lan L, Sun HB, Huang JH, Zhang H. Preparation of NiTi/NiTiNb laminated alloys by vacuum brazing. Adv Mater Res 2010;97101:16536.
doi: http://dx.doi.org/10.4028/www.scientific.net/AMR.97-101.1653.
[111] Quintino L, Liu L, Hu A, Miranda RM, Zhou Y. Fracture analysis of Ag nanobrazing of NiTi to Ti alloy. Soldag Inspeo 2013;18:2816. doi: http://dx.
doi.org/10.1590/S0104-92242013000300010.
[112] Santos TG, Fernandes FB, Bernardo G, Miranda RM. Analyzing mechanical properties and nondestructive characteristics of brazed joints of NiTi shape
memory alloys to carbon steel rods. Int J Adv Manuf Technol 2013;66:78793. doi: http://dx.doi.org/10.1007/s00170-012-4366-y.
[113] Munez CJ, Garrido MA, Rams J, Urea A. Experimental study of W-Eurofer laser brazing for divertor application. J Nucl Mater 2011;418:23948. doi:
http://dx.doi.org/10.1016/j.jnucmat.2011.07.008.
[114] Janeschitz G, Borrass K, Federici G, Igitkhanov Y, Kukushkin A, Pacher HD, et al. The ITER divertor concept. J Nucl Mater 1995;220222:7388. doi:
http://dx.doi.org/10.1016/0022-3115(94)00447-1.
[115] Kazakov NF. Diffusion bonding of materials. 1st ed. Oxford: Pergamon Press; 1985.
[116] Hill A, Wallach ER. Modelling solid-state diffusion bonding. Acta Metall 1989;37:242537. doi: http://dx.doi.org/10.1016/0001-6160(89)90040-0.
[117] Orhan N, Khan TI, Erolu M. Diffusion bonding of a microduplex stainless steel to Ti-6Al-4V. Scripta Mater 2001;45:4416. doi: http://dx.doi.org/
10.1016/S1359-6462(01)01041-7.
[118] Kejanli H, Taskin M, Kolukisa S, Topuz P. Transient liquid phase (TLP) diffusion bonding of Ti45Ni49Cu6 P/M components using Cu interlayer. Int J Adv
Manuf Technol 2009;44:6959. doi: http://dx.doi.org/10.1007/s00170-008-1860-3.
[119] Derby B, Wallach ER. Theoretical model for diffusion bonding. Met Sci 1982;16:4956. doi: http://dx.doi.org/10.1179/030634582790427028.
[120] Senkevich KS, Shlyapin SD. Investigation of the process of diffusion bonding of alloys based on titanium nickelide. Weld Int 2012;26:7368. doi:
http://dx.doi.org/10.1080/09507116.2011.653156.
[121] Senkevich KS. A study of the microstructure of diffusion joints of TiNi-base alloys. Met Sci Heat Treat 2014;55:6759. doi: http://dx.doi.org/10.1007/
s11041-014-9689-x.
[122] Senkevich KS, Knyazev MI, Runova YE, Shlyapin SD. Special features of formation of a TiNi VT6 diffusion joint. Met Sci Heat Treat 2013;55:41922.
doi: http://dx.doi.org/10.1007/s11041-013-9647-z.
[123] Cavaleiro AJ, Ramos AS, Braz Fernandes FM, Schell N, Vieira MT. In situ characterization of NiTi/Ti6Al4V joints during reaction-assisted diffusion
bonding using Ni/Ti multilayers. J Mater Eng Perform 2014;23:16259. doi: http://dx.doi.org/10.1007/s11665-014-0930-y.
[124] Cao J, Feng JC, Li ZR. Microstructure and fracture properties of reaction-assisted diffusion bonding of TiAl intermetallic with Al/Ni multilayer foils. J
Alloys Compd 2008;466:3637. doi: http://dx.doi.org/10.1016/j.jallcom.2007.11.033.
[125] Araki T, Hirose A, Uchihara M, Kohno W, Honda K, Kondoh M. Characteristics and fracture morphology of Ti-Ni type shape memory alloy and its laser
weld joint. J Soc Mater Sci Jpn 1989;38:47883. doi: http://dx.doi.org/10.2472/jsms.38.478.
[126] Hirose A, Araki T, Uchihara M, Honda K, Kondoh M. Laser welding of Ti-Ni type shape memory alloy. J Jpn Inst Met Mater 1990;54:2629.
[127] Tuissi A, Bassani P, Gerosa M, Mauri D, Pini M, Capello E, et al. CO2 laser welding of NiTi/Ni-based alloys. In: Pelton A, Duerig T, editors. Proc int conf
shape mem superelastic technol. p. 22938. doi: http://dx.doi.org/10.1361/cp2003smst229.
[128] Gugel H, Schuermann A, Theisen W. Laser welding of NiTi wires. Mater Sci Eng, A 2008;481482:66871. doi: http://dx.doi.org/10.1016/j.
msea.2006.11.179.
[129] Haas T, Schuessler A. Welding and joining of TiNi shape memory alloys: engineering aspects and medical applications. In: Proc int conf shape mem
superelastic technol; 1999. p. 111.
[130] Falvo A, Furgiuele FMM, Maletta C. Functional behaviour of a NiTi-welded joint: two-way shape memory effect. Mater Sci Eng, A 2008;481
482:64750. doi: http://dx.doi.org/10.1016/j.msea.2006.11.178.
[131] Panton B, Oliveira JP, Zeng Z, Zhou YN, Khan MI. Thermomechanical fatigue of post-weld heat treated NiTi shape memory alloy wires. Int J Fatigue
2016;92:17. doi: http://dx.doi.org/10.1016/j.ijfatigue.2016.06.012.
[132] Chan CW, Man HC, Yue TM. Effects of process parameters upon the shape memory and pseudo-elastic behaviors of laser-welded NiTi thin foil. Metall
Mater Trans A Phys Metall 2011;42:226470. doi: http://dx.doi.org/10.1007/s11661-011-0623-1.
[133] Chan CW, Man HC. Laser welding of thin foil nickeltitanium shape memory alloy. Opt Lasers Eng 2011;49:1216. doi: http://dx.doi.org/10.1016/j.
optlaseng.2010.08.007.
[134] Chan CW, Man HC, Yue TM. Effect of post-weld heat-treatment on the oxide film and corrosion behaviour of laser-welded shape memory NiTi wires.
Corros Sci 2012;56:15867. doi: http://dx.doi.org/10.1016/j.corsci.2011.11.020.
[135] Schuessler A. Laser processing of nitinol materials. In: Proc int conf shape mem superelastic technol; 2001. p. 2532.
[136] Schlossmacher P. Laser welding of Ni-rich TiNi shape memory alloy pseudoelastic properties. In: Proc int conf shape mem superelastic technol;
1997. p. 13742.
[137] Tuissi A, Besseghini S, Ranucci T, Squatrito F, Pozzi M. Effect of Nd-YAG laser welding on the functional properties of the Ni49.6 at.%Ti. Mater Sci Eng,
A 1999;273275:8137. doi: http://dx.doi.org/10.1016/S0921-5093(99)00422-0.
[138] Falvo A, Furgiuele FM, Maletta C. Laser welding of a NiTi alloy: mechanical and shape memory behaviour. Mater Sci Eng, A 2005;412:23540. doi:
http://dx.doi.org/10.1016/j.msea.2005.08.209.
[139] Tam B, Khan MI, Zhou Y. Mechanical and functional properties of laser-welded Ti-55.8 wt pct Ni nitinol wires. Metall Mater Trans A
[140] Vieira LA, Fernandes FMMB, Miranda RMM, Silva RJC, Quintino L, Cuesta A, et al. Mechanical behaviour of Nd:YAG laser welded superelastic NiTi.
Mater Sci Eng, A 2011;528:55605. doi: http://dx.doi.org/10.1016/j.msea.2011.03.089.
[141] Iijima M, Brantley WA, Yuasa T, Muguruma T, Kawashima I, Mizoguchi I. Joining characteristics of orthodontic wires with laser welding. J Biomed
Mater Res Part B Appl Biomater 2008;84B:14753. doi: http://dx.doi.org/10.1002/jbm.b.30856.
[142] Braz Fernandes FM, Mahesh KK, Craciunescu CM, Oliveira JP, Schell N, Miranda RM, et al. In situ structural characterization of laser welded NiTi shape
memory alloys. Mater Sci Forum 2013;738739:33843. doi: http://dx.doi.org/10.4028/www.scientific.net/MSF.738-739.338.
[143] Mirshekari GR, Saatchi A, Kermanpur A, Sadrnezhaad SK. Laser welding of NiTi shape memory alloy: comparison of the similar and dissimilar joints to
AISI 304 stainless steel. Opt Laser Technol 2013;54:1518. doi: http://dx.doi.org/10.1016/j.optlastec.2013.05.014.
[144] Schlossmacher P, Haas T, Schussler A. Laser welding of Ni-Ti shape memory alloys. In: Proc int conf shape mem superelastic technol; 1994. p. 8590.
[145] Sevilla P, Martorell F, Libenson C, Planell JA, Gil FJ. Laser welding of NiTi orthodontic archwires for selective force application. J Mater Sci - Mater Med
[146] Schlossmacher P, Haas T, Ussler AS. Laser-welding of a Ni-rich TiNi shape memory alloy: mechanical behavior. J Phys IV 1997;7. doi: http://dx.doi.org/
10.1051/jp4:1997539. C5-251C5-256.
[147] Yan XJ, Yang DZ, Liu XP. Corrosion behavior of a laser-welded NiTi shape memory alloy. Mater Charact 2007;58:6238. doi: http://dx.doi.org/10.1016/
j.matchar.2006.07.010.
[148] Zhao X, Wang W, Chen L, Liu F, Huang J, Zhang H. Microstructures of cerium added laser weld of a TiNi alloy. Mater Lett 2008;62:15513. doi: http://
dx.doi.org/10.1016/j.matlet.2007.09.021.
[149] Zhao X, Wang W, Chen L, Liu F, Chen G, Huang J, et al. Two-stage superelasticity of a Ce-added laser-welded TiNi alloy. Mater Lett 2008;62:353941.
doi: http://dx.doi.org/10.1016/j.matlet.2008.03.042.
[150] Li H, Wang W, Dong P. Laser welding of TiNi shape memory alloy using pure Cu as interlayer. J Mater Eng 2010:912. <http://jme.biam.ac.cn/EN/
Y2010/V0/I10/9>.
[151] Chan CW, Chan SHJ, Man HC, Ji P. Constitutive model for localized Lders-like stress-induced martensitic transformation and super-elastic behaviors
of laser-welded NiTi wires. Comput Mater Sci 2012;63:197206. doi: http://dx.doi.org/10.1016/j.commatsci.2012.06.009.
[152] Otibar D, Rathmann C, Lygin K, Kreimeier D, Szymansky P. Analyzing laser-welded NiTi-NiTi-joints for actuator applications using design of
experiments. J Mech Eng Autom 2015;5:7681. doi: http://dx.doi.org/10.17265/2159-5275/2015.02.003.
[153] Chan CW, Man HC, Yue TM. Effect of postweld heat treatment on the microstructure and cyclic deformation behavior of laser-welded NiTi-shape
memory wires. Metall Mater Trans A 2012;43:195665. doi: http://dx.doi.org/10.1007/s11661-011-1062-8.
[154] Yan XJJ, Yang DZZ, Qi M. Rotating-bending fatigue of a laser-welded superelastic NiTi alloy wire. Mater Charact 2006;57:5863. doi: http://dx.doi.org/
10.1016/j.matchar.2005.12.009.
[155] Chan CW, Man HC, Cheng FT. Fatigue behavior of laser-welded NiTi wires in small-strain cyclic bending. Mater Sci Eng, A 2013;559:40715. doi:
http://dx.doi.org/10.1016/j.msea.2012.08.119.
[156] Song YGG, Li WSS, Li L, Zheng YFF. The influence of laser welding parameters on the microstructure and mechanical property of the as-jointed NiTi
alloy wires. Mater Lett 2008;62:23258. doi: http://dx.doi.org/10.1016/j.matlet.2007.11.082.
[157] Yan X, Yang D, Liu X. Electrochemical behavior of YAG laser-welded NiTi shape memory alloy. Trans Nonferr Met Soc Chin 2006;16:5726. doi: http://
dx.doi.org/10.1016/s1003-6326(06)60100-3.
[158] Ruhlig D, Gugel H, Schulte A, Theisen W, Schuhmann W. Visualization of local electrochemical activity and local nickel ion release on laser-welded
NiTi/steel joints using combined alternating current mode and stripping mode SECM. Analyst 2008;133:17006. doi: http://dx.doi.org/10.1039/
b804718a.
[159] Chan CW, Man HC, Yue TM. Susceptibility to stress corrosion cracking of NiTi laser weldment in Hanks solution. Corros Sci 2012;57:2609. doi:
http://dx.doi.org/10.1016/j.corsci.2011.12.010.
[160] Chan CW, Man HC, Yue TM. Susceptibility to environmentally induced cracking of laser-welded NiTi wires in Hanks solution at open-circuit
potential. Mater Sci Eng, A 2012;544:3847. doi: http://dx.doi.org/10.1016/j.msea.2012.02.087.
[161] Chan CW, Hussain I, Waugh DG, Lawrence J, Man HC. In vitro mesenchymal stem cell responses on laser-welded NiTi alloy. Mater Sci Eng, C
2013;33:134454. doi: http://dx.doi.org/10.1016/j.msec.2012.12.035.
[162] Yan XJJ, Yang DZZ, Liu XPP. Influence of heat treatment on the fatigue life of a laser-welded NiTi alloy wire. Mater Charact 2007;58:2626. doi: http://
dx.doi.org/10.1016/j.matchar.2006.05.001.
[163] Yan X, Ge Y. Influence of post-weld annealing on transformation behavior and mechanical properties of laser-welded NiTi alloy wires. J Mater Eng
Perform 2014;23:34749. doi: http://dx.doi.org/10.1007/s11665-014-1156-8.
[164] Chan CW, Man HC. Reduction of environmentally induced cracking of laser-welded shape memory NiTi wires via post-weld heat-treatment. Mater Sci
Eng, A 2013;588:38894. doi: http://dx.doi.org/10.1016/j.msea.2013.09.051.
[165] Yan XJ, Yang DZ. Corrosion resistance of a laser spot-welded joint of NiTi wire in simulated human body fluids. J Biomed Mater Res A
2006;77:97102. doi: http://dx.doi.org/10.1002/jbm.a.30378.
[166] Oliveira JP, Miranda RM, Schell N, Braz Fernandes FM, Schell N, Braz Fernandes FM. High strain and long duration cycling behavior of laser welded
NiTi sheets. Int J Fatigue 2016;83:195200. doi: http://dx.doi.org/10.1016/j.ijfatigue.2015.10.013.
[167] Oliveira JP, Braz Fernandes FM, Miranda RM, Schell N. On the mechanisms for martensite formation in YAG laser welded austenitic NiTi. Shape Mem
Superelast 2016;2:11420. doi: http://dx.doi.org/10.1007/s40830-016-0058-z.
[168] Oliveira JP, Fernandes FMB, Schell N, Miranda RM. Shape memory effect of laser welded NiTi plates. Funct Mater Lett 2015;8:1550069. doi: http://dx.
doi.org/10.1142/S1793604715500691.
[169] Oliveira JP, Fernandes FMB, Miranda RM, Schell N, Ocaa JL. Residual stress analysis in laser welded NiTi sheets using synchrotron X-ray diffraction.
Mater Des 2016;100:1807. doi: http://dx.doi.org/10.1016/j.matdes.2016.03.137.
[170] Oliveira JP, Braz Fernandes FM, Miranda RM, Schell N, Ocaa JL. Effect of laser welding parameters on the austenite and martensite phase fractions of
NiTi. Mater Charact 2016;119:14851. doi: http://dx.doi.org/10.1016/j.matchar.2016.08.001.
[171] Oliveira JP, Fernandes FMB, Schell N, Miranda RM. Martensite stabilization during superelastic cycling of laser welded NiTi plates. Mater Lett
2016;171:2736. doi: http://dx.doi.org/10.1016/j.matlet.2016.02.107.
[172] Pouquet J, Miranda RM, Quintino L, Williams S. Dissimilar laser welding of NiTi to stainless steel. Int J Adv Manuf Technol 2012;61:20512. doi:
http://dx.doi.org/10.1007/s00170-011-3694-7.
[173] Hahnlen R, Fox G, Dapino MJ. Fusion welding of nickel-titanium and 304 stainless steel tubes: part I: laser welding. J Intell Mater Syst Struct
2012;24:94561. doi: http://dx.doi.org/10.1177/1045389X12461075.
[174] Ng CH, Mok ESH, Man HC. Effect of Ta interlayer on laser welding of NiTi to AISI 316L stainless steel. J Mater Process Technol 2015;226:6977. doi:
http://dx.doi.org/10.1016/j.jmatprotec.2015.06.039.
[175] Burdet P, Vannod J, Hessler-Wyser A, Rappaz M, Cantoni M. Three-dimensional chemical analysis of laser-welded NiTistainless steel wires using a
dual-beam FIB. Acta Mater 2013;61:30908. doi: http://dx.doi.org/10.1016/j.actamat.2013.01.069.
[176] Li HM, Sun DQ, Cai XL, Dong P, Wang WQ. Laser welding of TiNi shape memory alloy and stainless steel using Ni interlayer. Mater Des
2012;39:28593. doi: http://dx.doi.org/10.1016/j.matdes.2012.02.031.
[177] Li H, Sun D, Gu X, Dong P, Lv Z. Effects of the thickness of Cu filler metal on the microstructure and properties of laser-welded TiNi alloy and stainless
steel joint. Mater Des 2013;50:34250. doi: http://dx.doi.org/10.1016/j.matdes.2013.03.014.
[178] Zhang C, Sun X, Hou X, Li H, Sun D. The corrosion resistance of composite arch wire laser-welded by NiTi shape memory alloy and stainless steel wires
with Cu interlayer in artificial saliva with protein. Int J Med Sci 2013;10:106872. doi: http://dx.doi.org/10.7150/ijms.5878.
[179] Zhang C, Sun X. Susceptibility to stress corrosion of laser-welded composite arch wire in acid artificial saliva. Adv Mater Sci Eng 2013;3:18. doi:
http://dx.doi.org/10.1155/2013/738954.
[180] Zhang C, Zhao S, Sun X, Sun D, Sun X. Corrosion of laser-welded NiTi shape memory alloy and stainless steel composite wires with a copper interlayer
upon exposure to fluoride and mechanical stress. Corros Sci 2014;82:4049. doi: http://dx.doi.org/10.1016/j.corsci.2014.01.040.
[181] Watanabe E, Stigall G, Elshahawy W, Watanabe I. Deflection load characteristics of laser-welded orthodontic wires. Angle Orthod 2012;82:698702.
doi: http://dx.doi.org/10.2319/061411-393.1.
[182] Miranda RM, Assuno E, Silva RJC, Oliveira JP, Quintino L. Fiber laser welding of NiTi to Ti-6Al-4V. Int J Adv Manuf Technol 2015;81:15338. doi:
http://dx.doi.org/10.1007/s00170-015-7307-8.
[183] Shojaei A, Zoeram SAA, Mousavi Akbari. Laser welding of Ti6Al4V to nitinol. Mater Des 2014;61:18590. doi: http://dx.doi.org/10.1016/
j.matdes.2014.04.078.
[184] Shojaei A, Zoeram SAA, Mousavi Akbari. Effect of interlayer thickness on microstructure and mechanical properties of as welded Ti6Al4V/Cu/NiTi
joints. Mater Lett 2014;133:58. doi: http://dx.doi.org/10.1016/j.matlet.2014.06.141.
[185] Steegmller R, Strobel M, Flaxmeier E, Schler A. Micro-welding for improved radiopacity of nitinol-stents. Proc Int Conf Shape Mem Superelastic
Technol 2004:5916.
[186] Panton B, Pequegnat A, Zhou YN. Dissimilar laser joining of NiTi SMA and MP35N wires. Metall Mater Trans A Phys Metall Mater Sci
[187] Zeng Z, Panton B, Oliveira JP, Han A, Zhou YN. Dissimilar laser welding of NiTi shape memory alloy and copper. Smart Mater Struct 2015;24:125036.
doi: http://dx.doi.org/10.1088/0964-1726/24/12/125036.
[188] Guber AE, Giordano N, Schussler A, Baldinus O, Loser M, Wieneke P. Nitinol-based microinstruments for endoscopic neurosurgery. In: Proceedings
actuator 96 5th int conf new actuators; 1996. p. 3758.
[189] Siekmeyer G, Schrader B, Hegel A, Steegmller R, Strobel M, Schuessler A. Novel micro-joining techniques to improve stent radiopacity: a comparison
of welding and riveting processes. Adv Mater Process 2005;163:72.
[190] Hagiwara R, Suehara M, Fujii R, Kato H, Nakagawa K, Oda Y. Laser welding method for removal of instruments debris from root canals. Bull Tokyo
Dent Coll 2013;54:818. , <http://www.ncbi.nlm.nih.gov/pubmed/23903578>.
[191] Zeng Z, Yang M, Oliveira JP, Song D, Peng B. Laser welding of NiTi shape memory alloy wires and tubes for multi-functional design applications. Smart
Mater Struct 2016;25:85001. doi: http://dx.doi.org/10.1088/0964-1726/25/8/085001.
[192] Ion J. Laser processing of engineering materials. 1st ed. Amsterdam: Elsevier; 2005. doi: http://dx.doi.org/10.1007/s13398-014-0173-7.2.
[193] Kou S. Welding metallurgy. 2nd ed. Hoboken: John Wiley & Sons Inc.; 2002. doi: http://dx.doi.org/10.1002/0471434027.
[194] Weinert K, Petzoldt V. Machining of NiTi based shape memory alloys. Mater Sci Eng, A 2004;378:1804. doi: http://dx.doi.org/10.1016/j.
msea.2003.10.344.
[195] Frick CP, Ortega AM, Tyber J, Maksound AEMEM, Maier HJ, Liu Y, et al. Thermal processing of polycrystalline NiTi shape memory alloys. Mater Sci Eng,
A 2005;405:3449. doi: http://dx.doi.org/10.1016/j.msea.2005.05.102.
[196] Dinnebier R, Billinge S. Powder diffraction: theory and practice. 1st ed. Cambridge: Royal Society of Chemistry; 2008. doi: http://dx.doi.org/10.1039/
9781847558237.
[197] Dutta RS, Madangopal K, Gadiyar HS, Banerjee S. Biocompatibility of NiTi shape memory alloy. Br Corros J 1993;28:21721. doi: http://dx.doi.org/
10.1179/000705993798318560.
[198] Tirry W, Schryvers D. Quantitative determination of strain fields around Ni4Ti3 precipitates in NiTi. Acta Mater 2005;53:10419. doi: http://dx.doi.
org/10.1016/j.actamat.2004.10.049.
[199] Schryvers D, Tirry W, Yang ZQ. Measuring strain fields and concentration gradients around Ni4Ti3 precipitates. Mater Sci Eng, A 2006;438440:4858.
[200] Wada K, Liu Y. Shape recovery of NiTi shape memory alloy under various pre-strain and constraint conditions. Smart Mater Struct 2005;14:S27386.
doi: http://dx.doi.org/10.1088/0964-1726/14/5/016.
[201] Liu YO, Galvin SP. Criteria for pseudoelasticity in near-equiatomic NiTi shape memory alloys. Acta Mater 1997;45:44319. doi: http://dx.doi.org/
10.1016/S1359-6454(97)00144-4.
[202] Liu Y, Xiang H. Apparent modulus of elasticity of near-equiatomic NiTi. J Alloys Compd 1998;270:1549. doi: http://dx.doi.org/10.1016/S0925-8388
(98)00500-3.
[203] Liu Y, Yang H. The concern of elasticity in stress-induced martensitic transformation in NiTi. Mater Sci Eng, A 1999;260:2405. doi: http://dx.doi.org/
10.1016/S0921-5093(98)00959-9.
[204] Liu Y, Yang H. Strain dependence of the Clausius-Clapeyron relation for thermoelastic martensitic transformations in NiTi. Smart Mater Struct
2007;16:S227. doi: http://dx.doi.org/10.1088/0964-1726/16/1/S03.
[205] Sittner P, Liu Y, Novak V. On the origin of Lders-like deformation of NiTi shape memory alloys. J Mech Phys Solids 2005;53:171946. doi: http://dx.
doi.org/10.1016/j.jmps.2005.03.005.
[206] Tan G, Liu Y, Sittner P, Saunders M. Lders-like deformation associated with stress-induced martensitic transformation in NiTi. Scripta Mater
[207] Chan CW, Chan SHJ, Man HC, Ji P. 1-D constitutive model for evolution of stress-induced R-phase and localized Lders-like stress-induced martensitic
transformation of super-elastic NiTi wires. Int J Plast 2012;3233:85105. doi: http://dx.doi.org/10.1016/j.ijplas.2011.12.003.
[208] Tobushi H, Shimeno Y, Hachisuka T, Tanaka K. Influence of strain rate on superelastic properties of TiNi shape memory alloy. Mech Mater
[209] Sedmk P, ittner P, Pilch J, Curfs C. Instability of cyclic superelastic deformation of NiTi investigated by synchrotron X-ray diffraction. Acta Mater
[210] Chrzanowski W, Szade J, Hart AD, Knowles JC, Dalby MJ. Biocompatible, smooth, plasma-treated nickel-titanium surface an adequate platform for
cell growth. J Biomater Appl 2012;26:70731. doi: http://dx.doi.org/10.1177/0885328211416023.
[211] Saburi T, Nenno S, Nishimoto Y, Zeniya M. Effects of thermo-mechanical treatment on the shape memory effect and the pseudoelasticity of Ti-50.2Ni
and Ti-47.5Ni-2.5Fe alloys. Iron Steel Inst Jpn 1986;72:5718. <https://www.jstage.jst.go.jp/article/tetsutohagane1955/72/6/72_6_571/_article>.
[212] Lin PH, Tobushi H, Tanaka K, Hattori T, Makita M. Pseudoelastic behaviour of TiNi shape memory alloy subjected to strain variations. J Intell Mater
Syst Struct 1994;5:694701. doi: http://dx.doi.org/10.1177/1045389X9400500514.
[213] Saburi T, Yoshida M, Nenno S. Deformation behavior of shape memory Ti-Ni alloy crystals. Scr Metall 1984;18:3636. doi: http://dx.doi.org/10.1016/
0036-9748(84)90453-8.
[214] Eggeler G, Hornbogen E, Yawny A, Heckmann A, Wagner M. Structural and functional fatigue of NiTi shape memory alloys. Mater Sci Eng, A
2004;378:2433. doi: http://dx.doi.org/10.1016/j.msea.2003.10.327.
[215] Gall K, Tyber J, Wilkesanders G, Robertson SW, Ritchie RO, Maier HJ. Effect of microstructure on the fatigue of hot-rolled and cold-drawn NiTi shape
memory alloys. Mater Sci Eng, A 2008;486:389403. doi: http://dx.doi.org/10.1016/j.msea.2007.11.033.
[216] Schaffer J. Structure-property relationships in conventional and nanocrystalline NiTi intermetallic alloy wire. J Mater Eng Perform 2009;18:5827.
doi: http://dx.doi.org/10.1007/s11665-009-9369-y.
[217] Mukherjee K, Sircar S, Dahotre NB. Thermal effects associated with stress-induced martensitic transformation in a Ti-Ni alloy. Mater Sci Eng
1985;74:7584. doi: http://dx.doi.org/10.1016/0025-5416(85)90111-9.
[218] McKelvey AL, Ritchie RO. Fatigue-crack growth behavior in the superelastic and shape-memory alloy nitinol. Metall Mater Trans A: Phys Metall Mater
Sci 2001;32:73143. doi: http://dx.doi.org/10.1007/s11661-001-0089-7.
[219] Pan Q, Cho C. The investigation of a shape memory alloy micro-damper for MEMS applications. Sensors 2007;7:1887900. doi: http://dx.doi.org/
10.3390/s7091887.
[220] Miyazaki S, Ohmi Y, Otsuka K, Suzuki Y. Characteristics of deformation and transformation pseudoelasticity in Ti-Ni Alloys. J Phys 1982;43:C422560.
doi: http://dx.doi.org/10.1007/s00338-002-0215-z.
[221] Bradley TG, Brantley Wa, Culbertson BM. Differential scanning calorimetry (DSC) analyses of superelastic and nonsuperelastic nickel-titanium
orthodontic wires. Am J Orthod Dentofac Orthop 1996;109:58997. doi: http://dx.doi.org/10.1016/S0889-5406(96)70070-7.
[222] Shaw JA, Churchill CB, Iadicola MA. Tips and tricks for characterizing shape memory alloy wire: part 1-differential scanning calorimetry and basic
phenomena. Exp Tech 2008;32:5562. doi: http://dx.doi.org/10.1111/j.1747-1567.2008.00410.x.
[223] Uchil J, Mahesh KK, Kumara KG. Dilatometric study of martensitic transformation in NiTiCu and NiTi shape memory alloys. J Mater Sci
2001;36:58237. doi: http://dx.doi.org/10.1023/A:1012908222409.
[224] Uchil J, Mahesh KK, Kumara KG. Electrical resistivity and strain recovery studies on the effect of thermal cycling under constant stress on R-phase in
NiTi shape memory alloy. Phys B Condensed Matter 2002;324:41928. doi: http://dx.doi.org/10.1016/S0921-4526(02)01462-X.
[225] Novk V, ittner P, Dayananda GN, Braz-Fernandes FM, Mahesh KK. Electric resistance variation of NiTi shape memory alloy wires in
thermomechanical tests: experiments and simulation. Mater Sci Eng, A 2008;481482:12733. doi: http://dx.doi.org/10.1016/j.msea.2007.02.162.
[226] Frenzel J, George EP, Dlouhy A, Somsen C, Wagner MF-X, Eggeler G. Influence of Ni on martensitic phase transformations in NiTi shape memory alloys.
Acta Mater 2010;58:344458. doi: http://dx.doi.org/10.1016/j.actamat.2010.02.019.
[227] Miller DA, Lagoudas DC. Influence of cold work and heat treatment on the shape memory effect and plastic strain development of NiTi. Mater Sci Eng,
A 2001;308:16175. doi: http://dx.doi.org/10.1016/S0921-5093(00)01982-1.
[228] ittner P, Landa M, Luk P, Novk V. R-phase transformation phenomena in thermomechanically loaded NiTi polycrystals. Mech Mater
2006;38:47592. doi: http://dx.doi.org/10.1016/j.mechmat.2005.05.025.
[229] Kurita T, Matsumoto H, Abe H. Transformation behavior in rolled NiTi. J Alloys Compd 2004;381:15861. doi: http://dx.doi.org/10.1016/
j.jallcom.2004.03.108.
[230] Gil FJ, Manero JM, Planell JA. Effect of grain size on the martensitic transformation in NiTi alloy. J Mater Sci 1995;30:252630. doi: http://dx.doi.org/
10.1007/BF00362129.
[231] Khan MI, Zhou Y. Effects of local phase conversion on the tensile loading of pulsed Nd:YAG laser processed Nitinol. Mater Sci Eng, A 2010;527:62358.
[232] Khan MI, Pequegnat A, Zhou YN. Multiple memory shape memory alloys. Adv Eng Mater 2013;15:38693. doi: http://dx.doi.org/10.1002/
adem.201200246.
[233] Carroll MC, Somsen C, Eggeler G. Multiple-step martensitic transformations in Ni-rich NiTi shape memory alloys. Scripta Mater 2004;50:18792. doi:
http://dx.doi.org/10.1016/j.scriptamat.2003.09.020.
[234] Trepanier C, Tabrizian M, Yahia L, Bilodeau L, Piron DL. Effect of modification of oxide layer on NiTi stent corrosion resistance. J Biomed Mater Res
1999;48:43340. doi: http://dx.doi.org/10.1002/(SICI)1097-4636(199824)43:4<433::AID-JBM11>3.0.CO;2-#.
[235] Lee T-H, Huang T-K, Lin S-Y, Chen L-K, Chou M-Y, Huang H-H. Corrosion resistance of different nickel-titanium archwires in acidic fluoride-containing
artificial saliva. Angle Orthod 2010;80:54753. doi: http://dx.doi.org/10.2319/042909-235.1.
[236] Figueira N, Silva TM, Carmezim MJ, Fernandes JCS. Corrosion behaviour of NiTi alloy. Electrochim Acta 2009;54:9216. doi: http://dx.doi.org/10.1016/
j.electacta.2008.08.001.
[237] Shabalovskaya SA. Surface, corrosion and biocompatibility aspects of nitinol as an implant material. Biomed Mater Eng 2002;12:69109. , <http://
www.ncbi.nlm.nih.gov/pubmed/11847410>.
[238] Stoeckel D, Pelton A, Duerig T. Self-expanding nitinol stents: material and design considerations. Eur Radiol 2004;14:292301. doi: http://dx.doi.org/
10.1007/s00330-003-2022-5.
[239] Eliades T, Athanasiou AE. In vivo aging of orthodontic alloys: implications for corrosion potential, nickel release, and biocompatibility. Angle Orthod
2002;72:22237. doi: http://dx.doi.org/10.1043/0003-3219(2002) 072<0222:IVAOOA>2.0.CO;2.
[240] Poon RWY, Ho JPY, Liu X, Chung CY, Chu PK, Yeung KWK, 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;237:4116. doi: http://dx.doi.org/10.1016/j.nimb.2005.05.030.
[241] Es-Souni M, Es-Souni M, Fischer-Brandies H. On the properties of two binary NiTi shape memory alloys. Effects of surface finish on the corrosion
behaviour and in vitro biocompatibility. Biomaterials 2002;23:288794. doi: http://dx.doi.org/10.1016/S0142-9612(01)00416-1.
[242] Man H, Cui Z, Yue T. Corrosion properties of laser surface melted NiTi shape memory alloy. Scripta Mater 2001;45:144753. doi: http://dx.doi.org/
10.1016/S1359-6462(01)01182-4.
[243] Wang J, Li N, Rao G, Han EH, Ke W. Stress corrosion cracking of NiTi in artificial saliva. Dent Mater 2007;23:1337. doi: http://dx.doi.org/10.1016/
j.dental.2006.01.001.
[244] Racek J, ittner P, Heller L, Pilch J, Sedlk P, Kadervek L. Electrochemistry of NiTi wires/springs subjected to static/cyclic loadings. Mater Today Proc
2015;2:S9659. doi: http://dx.doi.org/10.1016/j.matpr.2015.07.442.
[245] Yokoyama K, Kaneko K, Moriyama K, Asaoka K, Sakai J, Nagumo M. Hydrogen embrittlement of Ni-Ti superelastic alloy in fluoride solution. J Biomed
Mater Res 2003;65A:1827. doi: http://dx.doi.org/10.1002/jbm.a.10457.
[246] Yokoyama K, Ogawa T, Fujita A, Asaoka K, Sakai J. Fracture of Ni-Ti superelastic alloy under sustained tensile load in physiological saline solution
containing hydrogen peroxide. J Biomed Mater Res, Part A 2007;82A:55867. doi: http://dx.doi.org/10.1002/jbm.a.31173.
[247] Caplan AI. Mesenchymal stem cells. J Orthop Res 1991;9:64150. doi: http://dx.doi.org/10.1002/jor.1100090504.
[248] Altankov G, Grinnell F, Groth T. Studies on the biocompatibility of materials: fibroblast reorganization of substratum-bound fibronectin on surfaces
varying in wettability. J Biomed Mater Res 1996;30:38591. doi: http://dx.doi.org/10.1002/(SICI)1097-4636(199603)30:3<385::AID-JBM13>3.0.
CO;2-J.
[249] Cho JR, Lee BY, Moon YH, Van Tyne CJ. Investigation of residual stress and post weld heat treatment of multi-pass welds by finite element method and
experiments. J Mater Process Technol 2004;155156:16905. doi: http://dx.doi.org/10.1016/j.jmatprotec.2004.04.325.
[250] Krishnan KN. The effect of post weld heat treatment on the properties of 6061 friction stir welded joints. J Mater Sci 2002;37:47380. doi: http://dx.
doi.org/10.1023/A:1013701104029.
[251] Elangovan K, Balasubramanian V. Influences of post-weld heat treatment on tensile properties of friction stir-welded AA6061 aluminum alloy joints.
Mater Charact 2008;59:116877. doi: http://dx.doi.org/10.1016/j.matchar.2007.09.006.
[252] Otsuka K, Ren XB. Recent developments in the research of shape memory alloys. Intermetallics 1999;7:51128. doi: http://dx.doi.org/10.1016/s0966-
9795(98)00070-3.
[253] Ke CB, Cao SS, Ma X, Zhang XP. Modeling of Ni4Ti3 precipitation during stress-free and stress-assisted aging of bi-crystalline NiTi shape memory
alloys. Trans Nonferr Met Soc Chin 2012;22:257885. doi: http://dx.doi.org/10.1016/S1003-6326(11)61503-3.
[254] Paula AS, Mahesh KK, dos Santos CML, Braz Fernandes FM, da Costa Viana CS. Thermomechanical behavior of Ti-rich NiTi shape memory alloys. Mater
Sci Eng, A 2008;481482:14650. doi: http://dx.doi.org/10.1016/j.msea.2007.02.142.
[255] Sitepu H, Schmahl W, Allafi J, Eggeler G, Dlouhy A, Toebbens D, et al. Neutron diffraction phase analysis during thermal cycling of a Ni-rich NiTi shape
memory alloy using the Rietveld method. Scripta Mater 2002;46:5438. doi: http://dx.doi.org/10.1016/S1359-6462(02)00032-5.
[256] Dawes CT. Laser welding: a practical guide. 1st ed. Cambridge: Abington Publishing; 1992.
[257] Zhao X, Lan L, Sun H, Huang J, Zhang H. Mechanical properties of additive laser-welded NiTi alloy. Mater Lett 2010;64:62831. doi: http://dx.doi.org/
10.1016/j.matlet.2009.12.025.
[258] Cai W, Meng XL, Zhao LC. Recent development of TiNi-based shape memory alloys. Curr Opin Solid State Mater Sci 2005;9:296302. doi: http://dx.
doi.org/10.1016/j.cossms.2006.07.002.
[259] Sun Z, Karppi R. The application of electron beam welding for the joining of dissimilar metals: an overview. J Mater Process Technol 1996;59:25767.
doi: http://dx.doi.org/10.1016/0924-0136(95)02150-7.
[260] Sun Z, Ion JC. Review laser welding of dissimilar metal combinations. J Mater Sci 1995;30:4025214. doi: http://dx.doi.org/10.1007/BF00361499.
[261] Cao X, Jahazi M, Immarigeon JP, Wallace W. A review of laser welding techniques for magnesium alloys. J Mater Process Technol 2006;171:188204.
doi: http://dx.doi.org/10.1016/j.jmatprotec.2005.06.068.
[262] Phanikumar G, Chattopadhyay K, Dutta P. Modelling of transport phenomena in laser welding of dissimilar metals. Int J Numer Methods Heat Fluid
Flow 2001;11:15674. doi: http://dx.doi.org/10.1108/09615530110381575.
[263] Phanikumar G, Chattopadhyay K, Dutta P. Computational modeling of laser welding of Cu-Ni dissimilar couple. Metall Mater Trans B
[264] Mackwood AP, Crafer RC. Thermal modelling of laser welding and related processes: a literature review. Opt Laser Technol 2005;37:99115. doi:
http://dx.doi.org/10.1016/j.optlastec.2004.02.017.
[265] Chakraborty N, Chakraborty S. Modelling of turbulent molten pool convection in laser welding of a copper-nickel dissimilar couple. Int J Heat Mass
Transf 2007;50:180522. doi: http://dx.doi.org/10.1016/j.ijheatmasstransfer.2006.10.030.
[266] Bordjih K, Jouzeau J-Y, Mainard D, Payan E, Delagoutte J-P, Netter P. Evaluation of the effect of three surface treatments on the biocompatibility of
316L stainless steel using human differentiated cells. Biomaterials 1996;17:491500. doi: http://dx.doi.org/10.1016/0142-9612(96)82723-2.
[267] Nie FL, Wang SG, Wang YB, Wei SC, Zheng YF. Comparative study on corrosion resistance and in vitro biocompatibility of bulk nanocrystalline and
microcrystalline biomedical 304 stainless steel. Dent Mater 2011;27:67783. doi: http://dx.doi.org/10.1016/j.dental.2011.03.009.
[268] Duerig T, Pelton A, Stckel D. An overview of nitinol medical applications. Mater Sci Eng, A 1999;273275:14960. doi: http://dx.doi.org/10.1016/
S0921-5093(99)00294-4.
[269] Nemat-Nasser S, Guo W-GG. Superelastic and cyclic response of NiTi SMA at various strain rates and temperatures. Mech Mater 2006;38:46374. doi:
http://dx.doi.org/10.1016/j.mechmat.2005.07.004.
[270] Kinsey B, Wu X. Tailor welded blanks for advanced manufacturing. 1st ed. Oxford: Woodhead Publishing Limited; 2011. doi: http://dx.doi.org/
10.1533/9780857093851.2.118.
[271] Davis J. Nickel, cobalt, and their alloys. 1st ed. Ohio: ASM International; 2000.
[272] Robin A, Rosa JL. Corrosion behavior of niobium, tantalum and their alloys in hot hydrochloric and phosphoric acid solutions. Int J Refract Met Hard
Mater 2000;18:1321. doi: http://dx.doi.org/10.1016/S0263-4368(99)00034-7.
[273] Arrazola P-J, Garay A, Iriarte L-M, Armendia M, Marya S, Le Matre F. Machinability of titanium alloys (Ti6Al4V and Ti555.3). J Mater Process Technol
2009;209:222330. doi: http://dx.doi.org/10.1016/j.jmatprotec.2008.06.020.
[274] Sidambe A. Biocompatibility of advanced manufactured titanium implantsa review. Materials (Basel) 2014;7:816888. doi: http://dx.doi.org/
10.3390/ma7128168.
[275] Niinomi M. Mechanical properties of biomedical titanium alloys. Mater Sci Eng, A 1998;243:2316. doi: http://dx.doi.org/10.1016/S0921-5093(97)
00806-X.
[276] Sen I, Tamirisakandala S, Miracle DB, Ramamurty U. Microstructural effects on the mechanical behavior of B-modified Ti-6Al-4V alloys. Acta Mater
[277] Gurrappa I. Characterization of titanium alloy Ti-6Al-4V for chemical, marine and industrial applications. Mater Charact 2003;51:1319. doi: http://
dx.doi.org/10.1016/j.matchar.2003.10.006.
[278] Poondla N, Srivatsan TS, Patnaik A, Petraroli M. A study of the microstructure and hardness of two titanium alloys: commercially pure and Ti6Al4V.
J Alloys Compd 2009;486:1627. doi: http://dx.doi.org/10.1016/j.jallcom.2009.06.172.
[279] Tal-Gutelmacher E, Eliezer D. The hydrogen embrittlement of titanium-based alloys. J Miner Met Mater Soc 2005;57:469. doi: http://dx.doi.org/
10.1007/s11837-005-0115-0.
[280] Gale W, Totemeir T. Smithells metals reference book. 8th ed. Oxford: Elsevier; 2004. doi: http://dx.doi.org/10.1017/CBO9781107415324.004.
[281] Cheng Y, Cai W, Li HT, Zheng YF. Surface modification of NiTi alloy with tantalum to improve its biocompatibility and radiopacity. J Mater Sci
[282] Stoeckel D, Bonsignore C, Duda S. A survey of stent designs. Minim Invasive Ther Allied Technol 2002;11:13747. doi: http://dx.doi.org/10.1080/
136457002760273340.
[283] Kapila S, Angolkar PV, Duncanson MG, Nanda RS. Evaluation of friction between edgewise stainless steel brackets and orthodontic wires of four alloys.
Am J Orthod Dentofac Orthop 1990;98:11726. doi: http://dx.doi.org/10.1016/0889-5406(90)70005-W.
[284] Proffit W, Fields H. Orthodontics contemporary. 4th ed. St. Louis: Elsevier; 2007. <http://www.americanbanker.com/issues/179_124/which-city-is-
the-next-big-fintech-hub-new-york-stakes-its-claim-1068345-1.html>.
[285] Crocker M, Hodgson J, Kebede G. Variable diameter balloon dilatation catheter, 5470313; 1995.
[286] Peters O. Current challenges and concepts in the preparation of root canal systems: a review. J Endod 2004;30:55967. doi: http://dx.doi.org/10.1097/
01.DON.0000129039.59003.9D.
[287] Spili P, Parashos P, Messer HH. The impact of instrument fracture on outcome of endodontic treatment. J Endod 2005;31:84550. doi: http://dx.doi.
org/10.1097/01.don.0000164127.62864.7c.

Progress in Materials Science: J.P. Oliveira, R.M. Miranda, F.M. Braz Fernandes

Uploaded by

Copyright:

Available Formats

Progress in Materials Science: J.P. Oliveira, R.M. Miranda, F.M. Braz Fernandes

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Progress in Materials Science: J.P. Oliveira, R.M. Miranda, F.M. Braz Fernandes

Uploaded by

Copyright:

Available Formats

Progress in Materials Science 88 (2017) 412466

Contents lists available at ScienceDirect

Progress in Materials Science

Welding and Joining of NiTi Shape Memory Alloys: A Review

2.3.4. Adhesives joining . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 424

Fig. 1. Typical superelastic behaviour of a shape memory alloy [6].

Fig. 3. Crystal structures of austenite and martensite.

Fig. 4. Phase diagram of Ti-Ni [2].

Fusion processes, based on melting and solidification of the area to be joined;

2. Welding and joining technologies for NiTi

2.1. Major developments in welding technology

2.2. Fusion-based welding processes

2.2.1. Tungsten inert gas (TIG)

2.2.2. High energy density beams

2.2.3. Resistance welding

2.3. Solid-state processes

2.3.1. Explosive welding

2.3.2. Friction welding

2.3.3. Ultrasonic welding

2.3.4. Adhesives joining

2.3.5. Brazing and soldering

2.3.6. Diffusion and reactive diffusion bonding

3. Laser welding of NiTi

Process Main characteristics Benefits Limitations Existing works

(continued on next page)

Process Main characteristics Benefits Limitations Existing works

3.1. Similar laser welding

3.1.2. Mechanical behaviour

Fig. 22. Schematics of the welded joint (adapted from [139]).

Application Fatigue life [number of cycles]

3.1.3. Shape memory effect

3.1.4. Phase transformation temperatures

3.1.5. Corrosion resistance

3.1.6. Post-weld heat treatments

3.1.7. Filler materials

3.2. Laser welding of dissimilar joints involving NiTi

3.2.3. NiTi/other engineering materials

5. Summary and future outlook

You might also like