You are on page 1of 26

Acta Materialia 116 (2016) 382e407

Contents lists available at ScienceDirect

Acta Materialia
journal homepage: www.elsevier.com/locate/actamat

By invitation only: overview article

Cold spraying e A materials perspective


H. Assadi*, H. Kreye, F. Ga
€rtner, T. Klassen
Helmut Schmidt University, Institute of Materials Engineering, Hamburg, Germany

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

Article history: Cold spraying is a solid-state powder deposition process with several unique characteristics, allowing
Received 1 April 2016 production of coatings or bulk components from a wide range of materials. The process has attracted
Received in revised form much attention from academia and industry over the past two decades. The technical interest in cold
13 June 2016
spraying is twofold: first as a coating process for applications in surface technology, and second as a
Accepted 14 June 2016
Available online 27 June 2016
solid-state additive manufacturing process, offering an alternative to selective laser melting or electron
beam melting methods. Moreover, cold spraying can be used to study materials behaviour under
extremely high strain rates, high pressures and high cooling rates. The cold spraying process is thus
Keywords:
Cold spraying
considered to be relevant for various industrial applications, as well as for fundamental studies in ma-
Coating terials science. This article aims to provide an overview of the cold spray process, the current under-
Additive manufacturing standing of the deposition mechanisms, and the related models and experiments, from a materials
Deposition mechanism science perspective.
Deformation © 2016 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

1. Introduction high temperature processing of materials, such as oxidation and


unfavourable structural changes. It is possibly the only continuous
Cold spraying (CS) is a solid-state material deposition technique, method to produce bulk components of metastable materials that
where micron-sized particles of a powder bond to a substrate as a are available only in the powder form, e.g. as obtained from me-
result of high-velocity impact and the associated severe plastic chanical attrition or gas atomisation.
deformation. Acceleration of particles to high velocities is obtained The invention of modern CS took place, somewhat serendipi-
via expansion of a pressurised and (ironically) ‘hot’ gas through a tously, in Russia about three decades ago. The initial patents are
diverging-converging nozzle. Despite heating the process gas e already more than 20 years old [1e3] but there have been a
which is to provide higher acceleration of the gas and also to growing number of related patents since the invention of CS [4].
facilitate particle deformation through thermal softening e the Early academic papers (in English) by the inventors and co-workers
feedstock remains in the solid state throughout the entire process; appeared about the same time [5e9]. These were shortly followed
hence the name ‘cold’ spraying. This is to underline the contrast to by others, e.g. Dykhuizen and Smith [10] Gilmore et al. [11] van
conventional thermal spraying where particles are completely or Steenkiste et al. [12] Kreye and Stoltenhoff [13] and Xiong et al. [14].
partially molten upon impact onto the substrate. Alternative ter- The interest in CS has been increasing at a relatively fast rate in the
minology includes cold gas spray, micro cold spray, cold gas dy- past two decades, as reflected by the number of publications per
namic spray, kinetic spray, supersonic particle deposition and metal year, Fig. 1. Scholarly publications on CS include several books and
powder application (MPA), all of which refer to the same principle review articles, covering various aspects of the process and its ap-
of solid-state powder deposition as described above. CS is used for plications [15e23]. The topic is broad and spans over various fields
coating and repair. It is also used for additive manufacturing (AM) of research, from fluid dynamics to solid mechanics. Even from
at relatively high deposition rates as compared to the methods materials perspective, there are various aspects regarding material
based on selective laser or electron beam melting. The main properties and relevant physical phenomena that deserve separate
advantage of CS is that it alleviates the problems associated with reviews.
The aim of the present paper is to provide an overview of CS
from a materials perspective, particularly for a broader audience in
materials research, e.g. with interest in dynamic phenomena,
* Corresponding author. Brunel University London, Brunel Centre for Advanced
Solidification Technology (BCAST), Uxbridge, Middlesex, United Kingdom. plasticity and phase transformations. In terms of application and
E-mail address: hamid.assadi@brunel.ac.uk (H. Assadi). properties, CS is perhaps best to be compared to thermal spraying

http://dx.doi.org/10.1016/j.actamat.2016.06.034
1359-6454/© 2016 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.
H. Assadi et al. / Acta Materialia 116 (2016) 382e407 383

Abbreviations and symbols Eth normalised thermal energy, 1


f powder feed rate, kg s1
F* force to compress particle to height of dp/2, N
Abbreviation l nozzle length (supersonic part), m
AD aerosol deposition lpre prechamber length, m
AM additive manufacturing m, n Johnson-Cook parameters, 1
ASI adiabatic shear instability pgas process gas stagnation pressure, Pa
BMG bulk metallic glass r nozzle expansion ratio, 1
CNT carbon nanotube s standoff distance, m
CS cold spraying t time, s
DE deposition efficiency T temperature, K
E-FEM Eulerian finite element method Tg glass transition temperature, K
FEM finite element method Tgas process gas stagnation temperature, K
FFT fast Fourier transform Tm melting temperature, K
FIB focused ion beam Tp particle temperature, K
GND geometrically necessary dislocation Tref reference temperature, K
HA hydroxyapatite Tsub substrate temperature, K
L-FEM Lagrangian finite element method Tx crystallization temperature, K
MD molecular dynamics vc velocity of contact point in explosive welding, m s1
MPA metal powder application vcr critical particle impact velocity for bonding, m s1
MPW magnetic pulse welding vgun traverse speed of gun, m s1
SEM scanning electron microscopy vp particle impact velocity, m s1
SPH smoothed-particle hydrodynamics x characteristic system dimension, m
TEM transmission electron microscopy a thermal diffusivity, m2 s1
b adiabaticity parameter (x2/a/t), 1
Symbol ε plastic strain, 1
a, b fitting constants in critical velocity formula, 11 ε_ plastic strain rate, s1
A, B Johnson-Cook parameters, Pa ε_ *
normalised plastic strain rate, 1
C Johnson-Cook parameter, 1 gcr critical shear strain at the start of localisation, 1
cp heat capacity of particle, J kg1 K1 gmax stress strain at max stress in adiabatic deformation, 1
d nozzle diameter, m h normalised particle impact velocity (vp/vcr), 1
dp particle diameter, m hE sum of normalised energy terms ðEk þ Eth Þ, 1
E total energy per unit mass of particle, J kg1 q normalised temperature, 1
Ek kinetic energy per unit mass of particle, J kg1 r density of particle, kg m3
Ek0 reference kinetic energy, J kg1 s flow stress, Pa
Ek normalised kinetic energy, 1 sUTS ultimate tensile strength of particle, Pa
Eth thermal energy per unit mass of particle, J kg1 u spraying angle, 1
0
Eth reference thermal energy, J kg1

processes, as in Refs. [24,25]. In view of relevant physical phe- reaching supersonic velocities [13,27]. Fig. 2 shows an overview of a
nomena, however, CS is most closely related to explosive powder high-pressure CS system. In this system, a fluidised powder feed-
shock compaction, magnetic pulse welding (MPW) [26] and stock is fed to the process gas upstream, i.e. at high pressure before
explosive welding. Likewise, some of the conventional thermal entering the nozzle. The particles of the powder are thus acceler-
spraying processes, such as arc or plasma spraying where there is ated by the process gas, reaching high velocities of up to 1200 m/s
extensive melting and resolidification, may be best compared to or more. The particle temperature upon impact can be 1000 K or
fusion welding from the viewpoint of solidification science. This more, depending on the process conditions. Powder preheating is
paper covers principles of CS, put into perspective by considering attained through a powder heater installed between the powder
common features of material deposition methods using high- feeder and the gun [28], or more conventionally, through an
velocity impact. A brief review of the key CS parameters and elongated prechamber just before the nozzle.
typical material systems are also presented. The focus of the present paper is on high-pressure CS, where the
gas pressure is typically above 1 MPa, the particles are micron-
2. Principle and setup sized, and the operation is conducted at open air or under near-
atmospheric pressure. Variations and modifications of CS are not
The main technical objective in CS is to ensure that the particles covered in this review. Some of these variations are briefly
of the feedstock powder impinge the substrate at or beyond a mentioned below.
critical velocity. This is achieved by means of a pressurised and
preheated process gas e typically nitrogen, in some cases helium e 2.1. Low-pressure cold spray
that expands through a (converging/diverging) de Laval nozzle,
Low-pressure cold spray system and high-pressure system were
developed in parallel. There is substantial difference between the
1
1 indicates dimensionless quantity. two processes in terms of the size and the cost of the available
384 H. Assadi et al. / Acta Materialia 116 (2016) 382e407

Fig. 1. Trend of interest in cold spray research as measured by the number of publi-
cations per year, as compared to that in plasma spraying and high-velocity oxy-fuel
spraying (data extracted from scopus.com).

commercial equipment, as well as the achievable impact velocities


and temperatures. The scope of application of low-pressure system
is hence limited. An example of comparison between high-pressure
and low-pressure CS is given by Koivuluoto et al. [29]. Fig. 2. An overview of a high-pressure cold spray system: (a) schematics of the main
components of the original system, (b) a commercialised cold spray gun/nozzle,
designed at Helmut Schmidt University, and (c) an original gun/nozzle developed by
2.2. Vacuum cold spray Anatolii Papyrin. The guns labelled (b) and (c) are in the same photo and have equal
distances from the camera.
Also known as aerosol deposition (AD) [30], vacuum cold spray
is a downscaled variation of CS, which uses submicron powders and
works under partial vacuum. It is commonly used to apply ceramic the deposition mechanism in CS requires an understanding of
coatings. The process is attracting a growing interest, particularly particle deformation.
for electronic device applications. A recent review on AD is pro-
vided by Hanft et al. [31].
3.1. Modelling of particle impact and deformation

2.3. Laser-assisted cold spray


Particles of cold spray powders are typically 10e50 mm in
diameter, and the process of particle impact e from the time of
In this technique CS is used in parallel with laser heating [32,33].
initial contact to total dissipation of kinetic energy e takes about
The laser is coupled to cold spray either as a means of surface
100 ns. At such small time and length scales, experimental moni-
preparation [34], or to enhance bonding by providing additional
toring with sufficient resolution of particle impact and the related
heat input mainly in low-pressure cold spray [35].
phenomena is conceivably difficult, if at all possible. As an alter-
native, numerical simulation of particle impact is used to provide
2.4. Electrostatic-force-assisted cold spray an insight to these phenomena. The simulations are normally based
on quantitative models of well-defined underlying physics, which
The use of an electrostatic field in parallel with CS has been are conceived to play a role in material deposition by CS. A major
considered conceptually in a few theoretical studies [36,37]. The fraction of the modelling and simulation work in this area has
idea has been to accelerate the particles further by means of elec- focused on an idealised case of perpendicular impact of a single
trostatic field and so to enhance particle impact velocity. This spherical particle on a flat substrate of the same material.
concept has apparently not yet been implemented.

3. Deposition mechanism 3.1.1. Underlying physics and material models


There are several physical phenomena that play a role in the
Material deposition by CS results from consecutive impacts of process of deposition in CS of deformable materials. The most
micron-sized solid particles on a substrate. The particle impact is important of these is viscoplastic deformation, which occurs under
associated with viscoplastic deformation of the interacting bodies, dynamic shear loading. In comparison, the elastic deformation of
leading to (1) sequential compaction of the feedstock powder into material plays a lesser role, and is therefore not central to model-
a deposit, and (2) metallurgical bonding over a significant fraction ling of particle impact. Modelling of viscoplastic deformation is
of the particle-particle interfaces. Both of these processes are usually pursued together with modelling of heat conduction. The
central to obtaining a dense and strong deposit; incomplete thermomechanical behaviour of material under high strain rate is
compaction results in high porosity, whereas poor bonding often described through constitutive equations that link the flow
results in low strength of the deposit. On the other hand, both stress (s) to strain (ε), strain rate (_ε) and temperature (T). An
compaction and bonding processes take place via particle impact example of such equations, among many, is given by Johnson and
and the resulting severe deformation. Therefore, understanding Cook [38] as follows
H. Assadi et al. / Acta Materialia 116 (2016) 382e407 385

  
s ¼ ð A þ Bεn Þ 1 þ C ln_ε* 1  qm (1)

where ε_ * is the equivalent plastic strain rate normalised with


respect to a reference strain rate, and q ¼ (T  Tref)/(Tm  Tref) is the
normalised temperature in which Tref is a reference temperature
(normally taken as room temperature), and Tm is the melting
temperature. A number of other material models can be used for
this purpose, e.g. as considered in Ref. [39]. While some models
appear to be more appropriate than others, selection of ‘the best’
material model remains a challenge. A main reason is that material
data hardly exist for the range of strain rates in CS e which can be as
high as 109 s1 [40]. Even for the low-strain-rate regime, there are
limited data available on the plastic constitutive properties of
materials in the form of micron-sized particles. Such data have
been evaluated for a few powder materials, such as copper and
MCrAlY using a single particle compression method [41] as
described in section 5. Moreover, accurate simulation of deforma-
tion during particle impact requires further refinements, in addi-
tion to that of plastic constitutive properties, which are yet to be Fig. 3. Schematics of stress-strain variation for a typical bulk metallic material during
implemented. These include the possible influence of strain gra- torsion experiment [47].
dients [42] on the deformation pattern during particle impact.
Under strain gradients, additional hardening may occur due to the
deformation, so that localisation and shear banding may occur at a
generation of geometrically necessary dislocations (GNDs) [43].
critical strain, gcr. Beyond this point, the local strain at the shear
There can also be additional softening due to microstructural
band increases sharply to high values, while the overall strain re-
changes, namely dynamic recrystallization [44], which has not
mains almost unchanged up to the point of rupture. As will be
been widely considered in the modelling of particle impact. It
discussed later in more detail, this phenomenon is considered to be
should be noted that the thermal softening as considered in
closely related to particle bonding in CS [40].
constitutive equations, as in Eq. (1), relates to thermally-activated
The above concepts are applicable mainly to crystalline metals
dislocation movement, and not to microstructural changes.
and alloys, with typical work hardening and thermally-activated
The work required for viscoplastic deformation is provided by
deformation behaviours. For other materials, different material
the kinetic energy of the impacting particle, a major fraction of
models and alternative analytical approaches may be needed. For
which is converted into heat during impact. The rest of the kinetic
instance, for the case of metallic glasses there is strain softening
energy, considered roughly to be less than 10%, is stored within the
instead of work hardening, so that the material is always suscep-
material as elastic energy associated with defects or residual stress.
tible to instability and shear banding. In contrast, for some inter-
The generated heat can dissipate through conduction, though this
metallic compounds instability may be difficult to achieve due to
is sometimes neglected, i.e. deformation is assumed to be adiabatic.
anomalous (non-monotonic) thermal softening e e.g. increase of
The assumption of adiabaticity is valid only when the system
strength with increasing temperature up to 800  C as in the case of
dimension (x)pffiffiffiffiffiis significantly greater than the thermal diffusion Ni3Al. For ceramic materials, plastic deformation is not conceivable,
distance ( at ). This is when:
so that alternative concepts and entirely different material models
x2 should be used to explain the deposition process. These additional
¼ b[1 (2a) considerations will be briefly discussed later.
at
in which x is a characteristic system dimension, a is the thermal
3.1.2. Numerical simulation
diffusivity, t is the process time, and b is a dimensionless parameter
Modelling of particle impact and deformation is mostly per-
representing the degree of adiabaticity. Considering that x scales
formed using mesh-based numerical methods such as finite-
with the particle diameter, dp, and assuming that t scales with the
element method (FEM). The main objective of numerical simula-
ratio of dp to impact velocity, vp, then it can be shown that:
tion is to work out temporal evolution of the deformation pattern
and the respective field variables (stress, strain and temperature)
b∝dp vp : (2b)
for various combinations of materials and impact conditions e i.e.
This means that for the same impact velocity, larger particles particle velocity, particle temperature and substrate temperature.
deform more adiabatically, and vice versa. In fact, only very large The results of numerical simulations are subsequently used for the
particles satisfy this condition [45]. Therefore, heat transfer should interpretation of particle bonding and coating formation. A recent
be considered for detailed quantitative analysis of particle defor- review of the work on modelling of particle impact is given by Li
mation e particularly when the role of particle size in deposition is et al. [48]. An overview of the employed numerical methods, their
to be considered. main outcome and possible future directions are given below.
The heat generation during plastic deformation, especially un- FEM has been most widely used in simulation of particle impact
der high-strain rates, may lead to strain localisation and shear and deformation in CS. Nevertheless, most work in the literature is
instability. This condition is normally manifested as shear banding about single particle impact and not complete build-up of a coating.
and failure in bulk materials [46]. As shown in Fig. 3, during shear A pioneering example is given by Dykhuizen et al. [49]. They
deformation of a bulk component thermal softening may dominate considered an axisymmetric model of particle impact under adia-
over strain and strain-rate hardening under adiabatic conditions, batic conditions, using CTH, an Eulerian-based FEM software
leading to overall strain softening. During strain softening, the package developed at Sandia National Laboratories [50]. Their
material becomes inherently unstable with respect to plastic simulations showed the development of shape and field variables,
386 H. Assadi et al. / Acta Materialia 116 (2016) 382e407

Fig. 4. Lagrangian-based FEM simulation of impact of a copper particle on copper substrate, assuming adiabatic conditions [40]. The contours and the arrows correspond to strain
and velocity fields, respectively.

such as temperature, during impact of a copper particle on a steel to be fully adiabatic or not; ASI may occur prematurely if heat
substrate for different impact velocities. These simulations indi- transfer is not considered. The role of heat transfer can be used as a
cated formation of material jets at the outer rim of the contact zone, basis to study the effect of particle size on the occurrence of ASI
similar to that observed in explosive welding, at impact velocities during impact [45]. According to Eq. (2), deformation in larger
higher than 400 m/s. Following this work, Assadi et al. simulated particles is more adiabatic, so that ASI becomes more prominent
particle impact using the Lagrangian formulation in Abaqus/ with increasing particle size. As shown in Fig. 6, ASI occurs only
Explicit [40], with a focus on the evolution of the field variables when both particle impact velocity and particle size are beyond a
near the interacting surfaces. Examples of the simulated deforma- certain value.
tion patterns are shown in Fig. 4. The prediction of ASI preceded experimental evidence, and
A main outcome of these simulations was to demonstrate that hence may be regarded as a significant outcome of simulations. An
adiabatic shear instability (ASI) could occur at the particle/substrate equally important prediction concerned the location and the extent
interface during particle impact. ASI was signified by sudden and of ASI. As shown in Fig. 7, ASI occurs initially at the outer rim of the
prominent rise in the plastic strain at the interacting surfaces, Fig. 5. particle/substrate contact zone. With the increase of particle
A number of factors e both physical and computational e were impact velocity, the ASI area extends from the circumference of the
found to influence the occurrence of ASI in the simulation of par- crater towards its centre. According to this prediction, the ASI never
ticle impact. Physical factors include density, strength, melting reaches the centre of crater regardless of the particle impact
temperature and specific heat of the interacting materials. These velocity.
factors will be discussed in more detail in the next section on Several other studies have used FEM to investigate particle
bonding criteria. A most important computational factor is the impact and deformation under various impact conditions and
mesh size, which strongly affects ASI; the coarser the mesh, the material combinations [45,52e58]. Most of these studies are based
higher the threshold impact velocity to induce shear instability. on Lagrangian formulation implemented in commercial software. A
Another computational factor is whether deformation is considered general feature in Lagrangian-based FEM (L-FEM) simulations is

Fig. 5. Calculated development of (a) maximum plastic strain and (b) maximum temperature on the surface of a copper particle impacting onto a copper substrate. Adiabatic shear
instability is indicated by a sudden rise in these variables, if the impact velocity exceeds 550 m/s by 30 m/s [40].
H. Assadi et al. / Acta Materialia 116 (2016) 382e407 387

Fig. 6. Calculated development of maximum temperature on the surface of (a) 5 mm, and (b) 15 mm copper particle impacting onto a copper substrate at different impact velocities.
Adiabatic shear instability occurs only in the larger particle, at or beyond 500 m/s [45].

that they can demonstrate ASI at the particle/substrate interface at Both Lagrangian and Eulerian finite-element methods are
high impact velocities. L-FEM is also considered to be computa- mesh-based. There have also been attempts to simulate particle
tionally more efficient. Nevertheless, L-FEM can be associated with impact using mesh-free methods, such as smoothed-particle hy-
excessive mesh distortion and hence unrealistic development of drodynamics (SPH) and molecular dynamics (MD). SPH is in
the deformation pattern. This problem does not exist in Eulerian- principle a Lagrangian method in the sense that the nodes move
based FEM (E-FEM), which offers a more realistic account of the with the material. Examples of application of SPH to cold spray
deformation geometry, Fig. 8. research include the work by Manap et al. [60] and Li et al. [61].
Despite achieving realistic deformation patterns, most of the SPH does not have the problems associated with mesh-based
existing examples of particle impact simulation using E-FEM are in methods, but at the price of a generally low spatial resolution.
quasi-2D domains and also lack sufficient resolution to show ASI Similar to mesh-based methods, SPH does not simulate directly
unequivocally. Moreover, definition of surface interaction is not the process of bonding and rebounding. MD in contrast resolves
straightforward in E-FEM; by default the respective interacting material deformation on atomistic length scales e and can hence
bodies ‘bond’ together as soon as they come into contact. Note that be used to simulate bonding e but it is limited to small systems,
continuum-level modelling methods as mentioned above do not namely to nanoparticles. The few existing examples of simulation
directly simulate the process of bonding and rebounding. Instead, of particle impacts using MD are limited to ceramic nanoparticles,
they simulate the physical phenomena that relate to bonding, which are relevant to aerosol deposition [62,63]. The main ob-
namely plastic deformation and heat transfer. The bonding in jectives in these simulations are: (1) to shed light on the bonding
standard E-FEM is therefore only artificial. With the implementa- mechanism, and (2) to understand mechanism of deformation in
tion of E-FEM in commercial software, nevertheless, this method is essentially non-deformable ceramic materials. Unlike continuum-
again being used more commonly in the study of particle impact in level methods, MD is particularly suitable to simulate the process
CS; see e.g. Ref. [56]. Contrary to the initial presumptions, ASI may of bonding. Preliminary results of MD simulations show that
be appropriately diagnosed in E-FEM [59]. An interesting example deformation of brittle materials may be possible through
of the application of E-FEM to particle impact in CS is the study of consecutive fracture and bonding, which is associated with grain
the deformation behaviour of oxide films on metallic particles by refinement, Fig. 10. Bonding can nevertheless happen only at
Yin et al. [56]. Interestingly, the concept of a surface oxide layer may certain combinations of particle size and particle impact velocity;
be employed as a basis to simulate bonding and rebounding. Fig. 9 high velocities or large particles may lead to fragmentation
shows an example of how this might be achieved. In this way, the instead of deformation and bonding. There are no indications of
bonding and rebounding process are simulated directly, without ASI or material jetting in the existing MD simulations of particle
the need to diagnose ASI. impact.

Fig. 7. Calculated development of (a) the temperature profiles along a meridian path on a copper particle impacting onto a copper substrate, and (b) the corresponding areas of
shear instability at the particle/substrate interface, for different impact velocities [51]. The dashed line in (a) indicates the threshold of the area shear instability, while the arrows
indicate the outer rim of the contact zone. The weighted distance in (a) is the square root of the respective inscribed contact area normalised with reference to the initial particle
surface area.
388 H. Assadi et al. / Acta Materialia 116 (2016) 382e407

Fig. 8. Comparison between large-scale impact experiment [45] and Eulerian-based FEM simulation, for 20-mm copper ball impacted onto a steel substrate at 750 m/s.

Overall, numerical simulation of particle impact in CS has shown through consecutive fracture and bonding. In this case, large par-
the following important features. For the case of metallic materials, ticles are more susceptible to fragmentation.
the deformation is highly non-uniform. It may be associated with
the occurrence of ASI, which forms initially at the outer rim of the 3.2. Bonding criteria
contact zone and extends e with increasing impact velocity e to-
wards the centre. The ASI region never fully covers the contact In cold spray deposition, there are two distinct types of bonding
zone. Moreover, when considering heat transfer, the threshold that should be distinguished from one another: (i) particle-to-
impact velocity to induce ASI is shown to increase with decreasing substrate bonding, which is necessary for the formation of the
particle size. That is, large particles are more prone to ASI. For the first layer of particles, hence relevant for the adhesion strength of
case of brittle materials and nanoparticles, deformation may occur the coating to the substrate, and (ii) particle-to-particle bonding,

Fig. 9. Simulation of bonding and rebounding of superficially oxidised copper particle impacting onto a copper substrate, using Eulerian-based FEM. The oxide layer is described by
a pressure-dependent yield model, allowing dilation under tensile stress [245]. The model parameters are set in such a way that rebounding occurs only when the impact velocity is
below 260 m/s, in agreement with the results of large-scale impact tests [45].
H. Assadi et al. / Acta Materialia 116 (2016) 382e407 389

Fig. 10. Results of MD simulations for a 35-nm ‘brittle’ particle impacting a rigid substrate at 375 m/s, showing (a) snapshot of the atomic positions, (b) Fast Fourier transform (FFT),
i.e. ‘diffraction pattern’ of this, and (c) inverse FFT of a filtered diffraction pattern to make the grain structure more visible [63]. The letters in (a) signify the grains formed as a result
of impact.

which concerns the build-up of the cold-sprayed deposit, hence compress and stretch the interfacial region via plastic deformation,
relevant for the cohesive strength and physical properties of the as shown schematically in Fig. 11. As will be discussed below,
deposited material. A major part of the published work on CS fo- bonding in CS is conceived to relate to this type of interfacial
cuses on the latter basic question, although there is equally much deformation, which leads to breakup of oxide layers and contact of
interest in the former from a practical viewpoint, particularly in metallic surfaces at atomic level. Thus, the oxide layer should be
applications where coating and substrate materials are dissimilar. considered as a geometrical constraint and not an ‘energy barrier’
Generally, bonding takes place as soon as atomically-flat clean to metallurgical bonding.
surfaces of two components are brought into contact. In principle,
this can occur without a need for heating, fusion or mechanical 3.2.1. The role of shear instability
interlocking. In practice, however, surfaces of metallic components The results of modelling of particle impact and deformation as
are rarely flat and are usually covered by an oxide layer. In order to discussed in section 3.1 do not directly explain the physics of
induce solid-state bonding, therefore, the surface oxide layers bonding in CS. Nevertheless, they shed light on the mechanism of
should be removed or broken up to allow exposure and tight interface deformation (Figs. 4e6) and oxide breakup (Fig. 9), which
contact of clean metallic surfaces. One way to achieve this is to are prerequisites for solid-state bonding. As suggested by various

Fig. 11. Schematics of interfacial plastic deformation resulting in cold welding. Under high strains, there is contact between fresh metallic surfaces and grain boundaries are formed
between the two metals.
390 H. Assadi et al. / Acta Materialia 116 (2016) 382e407

3.2.2. Bonding characteristics in large-scale impact processes


In view of the role of shear instability and interface deformation
in solid-state bonding, explosive welding and/or magnetic pulse
welding may be considered as closely-related large-scale physical
models of CS. Experimental study of bonding characteristics in
these large-scale processes is relatively straightforward; hence they
may be used as experimental bases to understand bonding mech-
anism in CS of deformable materials. Interestingly, high-velocity
impact processes share many common features, regardless of the
system dimensions. This is so because in dynamic deformation
problems e neglecting the effects of heat conduction and strain-
rate hardening e the field variables become unique functions of
the length-to-time ratio. In other words, the time-normalised
evolution of the deformation pattern becomes independent of the
system size for identical impact velocities. The reason for this size-
independence is explained below from a mathematical point of
view.
First, elastic-plastic deformation in dynamic problems may be
described by Newtonian equations of motion, where acceleration is
proportional to the difference between the applied and the internal
force vectors. Second, the force vectors as well as the lumped mass
scale with the system dimensions, if the flow stress is not affected
by strain-rate hardening and if there is no heat conduction (unlike
Fig. 12. Schematics of deformation localisation due to shear instability in (a) particle
in Fig 6). Consequently, for any given impact velocity, identical
impact in cold spraying, (b) explosive and/or magnetic pulse welding, and (c) shear velocity fields will be obtained for adiabatic systems of identical
banding under compression. The material flow in (a) and (b) takes place parallel to geometrical proportions. This leads in turn to identical displace-
the surface of the two interacting bodies in the same direction, resulting in jetting ment and stress fields, if time and spatial dimensions are scaled to
and bonding, whereas in (c) shear banding occurs within one component with
the same proportion, Fig. 13.
opposite material displacement, resulting in breakup and failure, as in high-speed
cutting. Overall, the relevant factor in impact processes is not the ab-
solute value of the kinetic energy; it is the energy density. The
particle size may indeed play a role, but only due to the fact that in
researchers, including the inventors of CS [64], the bonding smaller systems, strain-rate hardening and conductive heat trans-
mechanism in CS may be compared to that in explosive welding fer become more prominent; which has nothing to do with the
and magnetic pulse welding. In these processes, bonding takes kinetic energy being smaller for smaller particles. Strain-rate
place as a result of high-velocity impact of two solid objects. A hardening and conductive cooling lead in turn to restricted local-
common feature in both processes is the formation and ejection of a ised deformation at the interface and less extensive bonding for
material jet, as a result of ASI, at the interface of the interacting smaller particles. As mentioned in section 3.1, this type of size effect
bodies, Fig. 12. While shear instability is often associated with shear should be considered for quantitative analysis of ASI during particle
banding and failure e especially in materials under dynamic impact. Nevertheless, large-scale impact processes can still provide
loading e it is a key phenomenon that leads to bonding in both highly relevant albeit qualitative experimental information on the
explosive welding and cold spraying processes. This is also the case characteristics of the interfacial deformation and bond zones in
in magnetic pulse welding, which is a close variant of explosive cold spraying.
welding. Fig. 14 shows an example of bond zone microstructure in

Fig. 13. Lagrangian-based FEM simulation of impact at 400 m/s of copper on copper, for (a) diameter of 50 mm, 100 ns after initial contact, and (b) diameter of 50 mm, 100 ms after
initial contact. The mesh size in (b) is also 1000 times larger than that in (a). Heat conduction and strain-rate hardening are not considered.
H. Assadi et al. / Acta Materialia 116 (2016) 382e407 391

Fig. 14. Microstructure of Al-3 wt. % Cu plates, bonded together using an impact velocity (velocity of the flyer plate in the vertical direction) of 285 m/s and an impact angle of 8.6 .
The dark precipitates (q0 phase) are stretched up to more than 10 times their original length near the bond zone [65]. Equiaxed nanograins are formed at the interface.

explosive welding. Based on the elongation of the precipitates in Study of explosive welding thus reveals common characteristics
this example, the plastic strain in this region is accurately evaluated of solid-state bonding due to high-velocity impact, as follows:
to be in excess of 1000%, which is a manifestation of ASI at the Bonding (a) occurs above a certain impact velocity; (b) does not
contact surfaces. In this example, there is a grain-refined region at occur below a certain impact angle; (c) results from ASI; (d) is
the centre of the bond zone. The grain refinement may have associated with large plastic strains of above 1000%; (d) features
resulted from dynamic recrystallization, though it was originally recrystallization and grain-refinement. The analogy between
attributed to melting and re-solidification [65]. Similar micro- bonding characteristics in explosive welding and CS has also been
structural features have also been observed in magnetic pulse considered as a basis to devise laser shock experiments [67] as a
welding [66]. micron-scale physical model of particle impact in CS. These ex-
In explosive welding bonding takes place for certain combina- periments resemble explosive welding geometrically, but relate to
tions of impact angle and impact velocity, Fig. 15. For the case of particle impact in terms of system size.
Al-3 wt. % Cu plates, bonding occurs only if the impact velocity of An alternative large-scale physical model of CS was set up by
the flyer plate is between 200 and 500 m/s, and the impact angle is Schmidt et al. [45], where 20-mm spherical projectiles (balls) were
within a certain range. The impact angle and the impact velocity shot onto a substrate using a military gun. As compared to explosive
determine the velocity of the collision point e the point of contact welding, these experiments were geometrically more similar to CS,
between the flyer plate and substrate e which should be below the with the additional advantage that the impact velocity could be
speed of sound for successful bonding. accurately measured and controlled. Examples of ball impact tests
are shown in Fig. 16. In analogy with the observations in explosive
welding, the plastic strain near the interface can be in excess of
1000%, as evaluated considering elongation of grains in the steel
substrate. There is also grain refinement indicating dynamic
recrystallization in the copper ball near the interface. The magni-
tude of strain and the extent of recrystallization depend, never-
theless, on the location in the contact zone; deformation and grain
refinement is less prominent at the centre of the contact zone as
compared to its periphery. A most important feature is that
metallurgical bonding does not occur at the central region of the
contact zone e i.e. around the ‘south pole’ of the particle, the first
point of contact with the substrate. The prediction that there is no
bonding at the south pole of large-scale impact has been verified
experimentally e via tensile testing of miniature specimens cut out
from different locations over the interface e for copper on copper,
copper on steel, and steel on steel [68]. Lack of bonding at the
centre of contact zone of the interface is visually evident for the
case of copper on aluminium, Fig. 16d.

3.2.3. Bonding of single microparticles


Fig. 15. Window of bonding for explosive welding of AlCu3 plates. The dashed lines
The impact of single micron-sized particles can be studied by
correspond to the lower and upper limits of the collision-point velocity, vc, within using the so-called ‘wipe-test’ experiments, where a substrate is
which bonding is possible. Data are taken from Ref. [65]. rapidly moved in the traverse direction in front of a cold-spray jet.
392 H. Assadi et al. / Acta Materialia 116 (2016) 382e407

Fig. 16. Large impact tests with 20 mm copper ball shot onto (a)e(c) steel, and (d) aluminium substrates. The grains of the steel substrate (b) are elongated more than 10 times (c),
and the grains of the copper ball (a) are severely refined near the interface [45]. There is no indication of bonding at the centre of impact zone in (d) [140]. The average width of the
perspective shown in (d) is approximately 60 mm.

These experiments produce samples that are most closely related experimental evaluation of the plastic strain near the contact re-
to real cold-spray deposition, both in geometry and in length scale, gion is not feasible. Therefore, ASI and severe plastic deformation of
though with differences in surface temperature of the substrate. A the interfacial region may only be indirectly diagnosed, e.g. in view
major downside of wipe-test is that the impact conditions of the of features related to jet formation or bonding.
individual deposited particles are never exactly known. Moreover, Fig. 17 shows examples of such features; they clearly indicate jet
unlike in explosive welding (Fig. 14) or large impact tests (Fig. 16), formation and bonding during impact of single micron-sized

Fig. 17. SEM micrographs of single particle impacts in cold spraying, showing (a) jetting [40], (b) residues of a rebounded particle on the substrate at the outer rim of the contact
zone [51], (c) residues of substrate on a detached particle (by ultrasonic cavitation), and (d) imprint on the substrate, as a result of impact and detachment of a part of substrate
material, related to (c) [69].
H. Assadi et al. / Acta Materialia 116 (2016) 382e407 393

particles. A most prominent feature is that bonding seems to occur


initially at the periphery of the contact zone, and that it extends
towards the centre as the impact velocity is increased. In agreement
with the results of large-scale impact tests and simulations, there is
no bonding at the central area, e.g. as shown for the impact of
Ti6Al4V particles on titanium [69]. Note that the centre of impact is
exactly where the hydrostatic pressure reaches a maximum during
the impact. This clearly rules out the role of hydrostatic pressure as
a main cause of bonding.
The fact that bonding does not occur at the centre of contact
zone e both in the impact of 20 mm balls and in cold spraying of
microparticles e agrees well with the results of particle impact
simulations, Fig. 7. It is also consistent with the results of explosive
welding experiments. As shown in Fig. 15, to achieve successful
bonding in explosive welding, the flyer plate must collide onto the
substrate while it is tilted within a certain range of angles. For the
example shown in Fig. 15, there is no bonding below 4 . At parallel
or low-angle impacts, shear forces diminish e and also, the velocity Fig. 18. Window of deposition for cold spraying of titanium with nitrogen at 40 bar
of the contact point, vc, becomes greater than the speed of sound in and different initial gas temperatures as shown in the legend [59]. The extent of ASI is
increased by exceeding further away from the critical conditions for bonding (lower
the solid material e hence the viscoplastic deformation is sup-
bound of the grey region). Note that the particle temperature is always lower than the
pressed. This leads in turn to the suppression of ASI and jetting. In initial gas temperature (more prominently so for smaller particles).
the impact of spherical objects e be it a 20-mm ball or a 20-mm
particle e such a condition exists at the centre of the contact zone,
where the interacting surfaces are almost parallel. It should be impact velocities result in hydrodynamic erosion of the substrate
noted that bonding may also occur at the central zone due to fric- [45]. The region bound between these two characteristic velocities
tional dissipation in angular impact [70], but this is not a common defines the window of deposition in CS, as shown in Fig. 18.
case in CS.
3.2.5. Bonding of dissimilar materials
3.2.4. Window of deposition for metallic materials All of the above experiments and conjectures concern bonding
In view of the above findings, the following conclusions may be between two similar metallic materials of relatively high deform-
drawn on the mechanism of particle-particle bonding in CS of ability. The problem complicates further for the cases where
deformable (metallic) materials: bonding is between two different metals or between metal and
ceramic.
 Bonding is associated with ASI, leading to large plastic strains For the case of dissimilar coating/substrate materials, the
(~10) and jetting. bonding characteristics and hence the adhesion strength depend
 ASI and bonding occur beyond a certain impact velocity. on the differences in properties e mainly hardness and density e of
 ASI and bonding extend with increasing impact velocity, from the respective materials. As a rule of thumb, smaller differences in
the periphery towards the centre of the contact zone, but never properties results in better bonding. If the substrate is softer and
cover the entire interface. less dense than the impinging particles, the particles may penetrate
more deeply, resulting in the dissipation of the kinetic energy over
Based on these points, ASI may be considered as a prerequisite a longer time and over a larger volume of the substrate material.
for bonding in CS of metallic materials [71]. Subsequent studies by This leads to diminished particle deceleration and ‘impact power’
various other researchers confirm the role of ASI in bonding. Some (see section 5.3), hence suppressed ASI on the particle side. A large-
studies have underlined the role of interfacial melting in bonding scale example of this case is shown in Fig. 16d. Despite suppression
[64,72e74]. According to these studies, ASI may lead to partial of the ASI on the particle side, the coating may still exhibit good
melting for certain combinations of materials and impact condi- adhesion to the substrate, possibly because of ASI on the substrate
tions [75]. However, there is no consensus on whether or not side and arguably because of mechanical interlocking. This is not
melting is necessary for bonding; in view of the overall observa- the case when the substrate is harder than the particles, as me-
tions, melting does not seem to be a prerequisite for bonding. There chanical interlocking would be ruled out as a possible mechanism
has also been a discussion about the role of mechanical contribu- of bonding. On the other hand, metallic materials have been suc-
tion, i.e. interlocking [76]. This has nevertheless concerned the cessfully deposited on ceramics, and even on glass [80]. It is clear
bonding between certain combinations of coatings and substrates, that ceramics do not deform plastically, so that bonding cannot be
and not the general particle-particle bonding as in the build-up of explained based on the ASI conjecture. Also, mechanical inter-
CS deposits. The bonding mechanism has also been discussed in locking would not be a viable option. It is therefore not completely
view of the competition between the adhesion and the rebound clear how metal coatings bond to ceramic or glass substrates. There
energy [77e79]. Consideration of the rebound energy seems to be have been studies of the interface between aluminium coating and
relevant for some cases of materials e such as metallic glasses (see ceramic substrates [81,82]. It has been argued that mechanical
Fig. 20 below) e where the interface is liquefied and hence exhibits interlocking is not necessary for bonding for such cases, and that
little resistance to rebounding. For the case of solid interfaces, adhesion is influenced by structural match on the atomic scale [82].
however, postulation of ‘adhesion energy’ may be too simple an Dissimilar material bonding is also relevant for CS of powder
approach, as it disregards the role of fracture mechanics and plas- blends to produce composite deposits. In this case, also, similar
ticity in particle detachment. Regardless of the perceived mecha- properties of the powder materials would lead to better bonding.
nism of bonding, there is consensus that bonding occurs beyond a This is nevertheless rarely aimed for, as composites are intention-
critical velocity, and that this velocity generally decreases with ally produced from materials of often drastically different proper-
increasing particle temperature. It is also presumed that too high ties. Fortunately, successful deposition is possible, if at least one of
394 H. Assadi et al. / Acta Materialia 116 (2016) 382e407

the constituents has sufficient ductility to embed the other. In this particles cool down to temperatures below Tg. Impact temperatures
case, bonding between soft and hard particles would not be a have been estimated to be 130e340  C for CS of NiTiZrSiSn or
critical issue, as long as there is sufficient bonding between the soft Zr55Al10Ni5Cu30 with helium [16,84,85], well below the respective
embedding particles. Clearly, this works out only if the volume glass transition temperatures. It has been argued that the temper-
fraction of the hard constituent is meant to be below a certain limit. ature may rise locally at the particle-substrate interface [89]. The
This is so because the amount of hard phase in the deposit is always temperature may indeed rise locally to very high values in BMGs
lower than that in the feedstock; it reaches a saturation limit due to shear banding [83,87]. On the other hand, as a precursor to
depending on the size and properties of the hard and soft shear banding, disorder by deformation may be responsible for
constituents. local strain and strain-rate softening of metallic glasses [90,91]. It is
not yet clear which mechanism plays a more dominant role in cold-
spray deposition of BMGs.
3.2.6. Bonding of brittle materials Despite the very short exposure of particles to the gas stream
A broader question concerns the mechanism of bonding for (about maximum 0.003 s), a significant amount of the originally
intrinsically brittle materials, such as ceramics, metallic glasses and amorphous structure may crystallize at increased gas tempera-
WC-metal composites (cermets). Ceramic materials have been tures. The crystallization mostly occurs at the particle-particle in-
successfully deposited by aerosol deposition or vacuum cold terfaces, as shown by TEM investigations [89,92,93]. This is because
spraying. Also, high-pressure CS has been used for the deposition of of exceeding the corresponding crystallization temperature, Tx, and
metallic glasses. Despite the sporadically successful experiments, incubation time [87]. To verify this, one would need to work out the
CS of brittle materials is yet to be fully explored. For the case of thermal history of the interfacial region [89,92], which is deter-
ceramic materials, initial theoretical investigations using molecular mined by the interplay between the heat conduction and the heat
dynamics (discussed above) indicate that deposition is possible generation by deformation. Such an analysis is complicated
between two limiting impact velocities, Fig. 19. because the incubation time for crystallization decreases drastically
In addition, deposition of ceramics by AD seems to be viable with increasing strain rate [94,95].
only when the particle size is below a certain limit. The role of Interestingly, the deposition efficiency of BMGs passes through a
particle size is also evident in CS of ceramics and cermets. In both maximum by increasing the process gas temperature or pressure
cases, deposition has been successful only when agglomerates of [87,96]. Too high pressures or temperatures lead to particle
small particulates with certain levels of porosity were used. There is detachment and rebound [96,97]. An example of partial detach-
so far little information on the mechanism of deformation and ment of a particle is shown in Fig. 20. Detachment occurs whenever
bonding of such agglomerates. the interface is liquefied (T > Tg) and so, cannot withstand the
Bulk metallic glasses (BMGs) may also be classified as intrinsi- rebounding elastic forces. Successful deposition occurs only if the
cally brittle materials. Nevertheless, metallic glasses exhibit interface cools down fast enough, so that the viscosity increases to
different modes of deformation based on the temperature, loading sufficiently high values, or the temperature falls below Tg. This
conditions and geometry [83]. The deposition mechanism of conjecture has been experimentally verified, by demonstrating
metallic glasses is equally complex and not yet fully understood. decreased deposition efficiencies attained on preheated substrates
Some features in bonding and deposition of metallic glasses by CS [96].
are discussed below.
Study of individual splat morphologies indicates a mixture of
shear banding and localised viscous flow throughout the particle. 3.3. Quantification of the critical condition for bonding
There is also interface softening as a result of ASI, similar to that in
crystalline materials [84,85]. A presumption is that the particle The critical velocity for bonding in CS can be evaluated experi-
temperature should be above Tg, allowing for viscous flow of the mentally, by correlating the (measured or calculated) particle
material in the undercooled liquid state [86]. Recent investigations impact velocity with the measured deposition efficiency (DE) of the
on Fe73Cr2Si11B11C3 attribute successful deposition to viscous flow corresponding powder [27]. It can also be estimated through nu-
at high (>103) Reynolds numbers [87,88]. This interpretation merical simulation of particle impact, considering that bonding and
cannot explain experiments with helium as process gas, where ASI occur at the same particle impact velocity [40]. Moreover, the
following expression has been used to fit the evaluated critical
velocities for identical coating/substrate materials [45]:
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
 
vcr ¼ as=r þ bcp Tm  Tp (3)

where s is the temperature dependent flow stress, r is the density,


cp is the heat capacity, Tm is the melting temperature, Tp is the mean
temperature of particle upon impact, and a and b are fitting con-
stants. A good fit for powders of typical particle size distribution is
obtained when a ¼ 4 and b ¼ 0.25. The temperature-dependent
flow stress can be approximated as follows:

s ¼ sUTS ð1  qÞ (4)

where q is the normalised temperature as in Eq. (1). The equation


assumes a linear softening of the material up to the melting tem-
perature. The critical velocity is also found to be a function of
particle size, namely increasing with decreasing particle size. This
size dependence of the critical velocity has been attributed to the
Fig. 19. Window of deposition for simulation of impact of brittle nanoparticles [63]. role of heat conduction [45] and accordingly formulated as a
H. Assadi et al. / Acta Materialia 116 (2016) 382e407 395

Fig. 20. Example of particle rebounds and interface separation during cold spraying of amorphous Cu50Zr50 particle; (a) schematics, and (b) SEM of partial detachment [96].

function of adiabaticity [98]. The size-dependence of critical ve- thermal energy of particle required for bonding at zero kinetic
locity has been worked out and verified experimentally for copper. energy. Likewise, the reference kinetic energy Ek0 may be inter-
The size dependence for nickel has been estimated using FEM preted as the minimum kinetic energy of particle required for
simulations [99]. bonding at zero thermal energy. Thus, Eq. (6) is nothing more than a
Prediction of the critical velocity and its dependence on particle rearrangement of Eq. (3), in which Ek0 is simply 12v2cr at Tp ¼ Tref . In
size has also been estimated through alternative criteria for view of these considerations, a dimensionless energy-based
bonding. Examples of alternative criteria for bonding include those deposition parameter can be defined as follows:
based on the adhesion energy and rebound energy [100], and on
the evolution of maximum plastic strain, as worked out by E-FEM
. .
hE ¼ Ek þ Eth where Ek ¼ Ek Ek0 and Eth ¼ Eth Eth
0
(7)
simulations [101]. Moreover, the critical velocity has been shown to
be influenced by surface oxidation of the particles of the feedstock In fact, hE may be considered as a general characteristic
[102e104]. Higher surface oxidation normally results in increased parameter for both thermal and cold spraying, and the necessary
critical velocity. Clearly, smaller particles are more affected by condition for deposition can be expressed as:
surface oxidation, because of their larger surface to volume ratio.
Thus, surface oxidation should be considered as a factor in the
study of size-dependence of the critical velocity, in addition to that
hE  1 (8)
relating to adiabaticity. Further work is needed to refine the A property of the above criterion for deposition is that it does
analytical basis for the prediction of the critical velocity, for not concern the question of exceeding a critical velocity; instead, it
instance, to include the effect of oxide layer, non-linear thermal concerns exceeding a critical normalized energy content. Based on
softening behaviour, or the effect of substrate material properties. this, the window of deposition can be plotted on an energy dia-
Considering the role of particle temperature in bonding, Fig. 18, gram, where the critical conditions are signified by a straight line,
one may equally think of a ‘critical temperature’ for bonding. That 0 and E0 on the corresponding axes, Fig 21.
connecting Eth k
is, at a fixed impact velocity, increasing particle temperature
beyond a certain value results in bonding. Based on this viewpoint,
the critical condition for bonding may be formulated more gener-
ally in terms of the total energy of the particle upon impact, as
described below. It should be stressed nevertheless that the
following formulation does not include new theoretical analyses; it
is only a rewriting of the above criterion for bonding in terms of
newly defined parameters. The total energy per unit mass of a
particle can be considered to be the sum of two parts, the kinetic
energy and a nominal thermal energy, as follows

1  
E ¼ Ek þ Eth where Ek ¼ v2p and Eth ¼ cp Tp  Tref (5)
2
Note that the nominal thermal energy as defined here is
different from the real thermal energy, in that it disregards the
latent heat of fusion at temperatures higher than Tm. At the critical
impact condition for bonding, the following relations hold:

0
Eth jcr ¼ Eth q; where Eth
0
¼ cp ðTm  Tref Þ (6a)

 . Fig. 21. An energy-based representation of the window of deposition based on


Ek jcr ¼ Ek0 ð1  qÞ; where Ek0 ¼ asUTS =r þ bEth
0
2 (6b) equations (5)e(8). Deposition takes place when the normalised energy content, hE , is
above unity, i.e. it falls in the grey region. In this region, contours of constant hE would
be straight lines parallel to the lower bound, i.e. hE ¼ 1. The upper bound of the
where a and b are the same calibration factors as in Eq. (3). The deposition window is sketched schematically to mark the threshold of hydrodynamic
0 may be interpreted as the minimum
reference thermal energy Eth material removal or splashing away of droplets at too high impact velocities.
396 H. Assadi et al. / Acta Materialia 116 (2016) 382e407

Fig. 22. Tensile (cohesive) strength of cold-sprayed copper coatings as a function of (a) particle impact velocity, and (b) ratio of particle impact velocity to the critical velocity [98].
There is a unique correlation between the cohesive strength and the velocity ratio, regardless of the particle size and spraying conditions.

The postulation of the energy parameter is different from that of impact) combined with thermal misfit may result in complex re-
the velocity ratio h ¼ vp =vcr [98], in that the two components of sidual stress profiles. These unfavourable effects can be alleviated
energy in hE are normalised individually. This means that there is by appropriate post heat treatments, or by substrate preheating, to
no unique critical energy for bonding. It can be shown that the induce recovery and recrystallization. A brief discussion of the
following relationship exists between the normalised energy microstructures and properties of cold-sprayed deposits are given
parameter and the velocity ratio: below.

hE ¼ h2 ð1  qÞ þ q (9)
4.1. Microstructure
The above relationship implies that at higher particle temper-
atures, the energy parameter becomes independent of the velocity
There are three important microstructural features that are
ratio. Likewise, the energy parameter correlates well with the
commonly found in most cold-sprayed deposits: (1) severe plastic
velocity ratio at lower temperatures.
deformation as signified by a high density of dislocations and other
defects, (2) grain refinement due to dynamic recrystallization,
4. Deposit characteristics particularly near the interfaces, and (3) non-bonded interfaces,
namely at the centre of impact zones. These features appear to be
CS should be considered both as a coating method and as an common not only in cold spraying, but also in the closely-related
alternative to conventional AM for the production of bulk compo- process of warm spraying [105]. Another aspect of cold-sprayed
nents. Therefore, the relevant deposit properties concern both microstructures is that they are highly anisotropic [106]. Proper-
standard coating properties, such as adhesion strength, and the ties of cold-sprayed deposits are to a large extent influenced by
common bulk properties. Examination of various properties of these features. Fig. 23 shows an example of the microstructure of
cold-sprayed deposits shows that in many cases there is a unique cold-sprayed copper, showing various types of lattice defects near a
correlation between the property and the normalised impact particle-particle interface.
velocity, h ¼ v=vcr [98]. This is shown, as an example, for copper in Dislocations are among the most prominent defects in cold
Fig. 22. sprayed deposits. They can be homogeneously distributed or pile
It should be noted that the correlation between properties and up into dislocation walls. Moreover, dislocations can form persis-
velocity ratio is not universal, and that it is best applicable to low tent nano-sized loops that do not vanish by annealing, with im-
particle temperatures. For instance, near the melting temperature plications on the mechanical and electrical properties of the
the critical velocity approaches zero and hence the velocity ratio deposit [108]. The density of dislocations can be different from one
becomes infinitely large. At higher temperatures, therefore, the material to the other; e.g. it is normally higher for aluminium as
data might be better interpreted in view of the energy parameter, compared to copper and nickel [109], presumably due to differ-
Eq. (7). This needs to be explored in future work. In any case, ences in the stacking fault energy. On the other hand, the disloca-
interpretation of cold-sprayed deposit properties can be discussed tion density can be very low in nanograins of aluminium alloys; i.e.
conveniently in view of these normalised velocity or energy in feedstock materials that are already nanocrystalline before
parameters. spraying [110]. This is regarded as a common feature in the defor-
On a microscopic level, properties of cold-sprayed deposits are mation of nanocrystalline materials, where grain boundaries play a
inherited from their microstructure and state of deformation. major role in the mechanism of plastic deformation.
Generally, cold-sprayed deposits that are produced under optimum An example of electron backscatter diffraction (EBSD) maps of
conditions exhibit microstructures and properties similar to those cold-sprayed copper is shown in Fig. 24. The figure shows regions of
of highly cold-worked bulk material, e.g. high yield strength and highly deformed materials near the particle-particle interfaces, but
low elongation to failure. Moreover, high densities of lattice defects also a relatively high amount of misorientation within the particles.
restrict electron mobility and thus degrade conductivity. The same In the image quality (IQ) map the grey level correlates roughly with
applies to the central regions of the particle-particle interfaces, the severity of plastic deformation and lattice defects. Moreover,
which are not metallurgically bonded. The non-bonded interfaces the bright streaks (green in the colour version) in the kernel
also act as crack nuclei under tensile loading. Further to these average misorientation (KAM) map correspond to rotation gradi-
structural deficiencies, the shot-peening effect (due to particle ents, which correlate roughly with the density of GNDs. Note that
H. Assadi et al. / Acta Materialia 116 (2016) 382e407 397

Fig. 23. TEM micrograph of a cold-sprayed copper coating, showing various micro-
structural features, including (A) dislocation walls, (B) elongated nanograins, (C)
equiaxed nanograins, and (D) a micron-sized dislocation-free grain containing twins
[107].

the grains in the upper-left of the maps appear to be free of


misorientation, possibly indicating recovery or (static) recrystalli- Fig. 24. EBSD maps of the cross section of a cold-sprayed copper coating, showing (a)
image quality map and (b) kernel average misorientation (KAM) map (courtesy of Tom
zation. This is consistent with the dislocation-free region (D) as
J€apel, Max Planck Institute for Iron Research, Düsseldorf, Germany).
observed in Fig. 23. There are several examples of analysis using
EBSD in the literature, e.g. for cold-sprayed copper [111] and
aluminium [112]. Moreover, electron channelling contrast imaging
recrystallization in cold-sprayed deposits are reported in the liter-
(ECCI) has been used in the study of grain structure and defects in
ature [124,130e132].
cold-sprayed deposits; see e.g. Refs. [113,114]. Nanoscale twinning
It is a common notion that CS is particularly suitable for appli-
has also been observed in cold-sprayed nickel [115].
cations where a phase change is to be avoided. Incidentally, CS can
Detailed studies of the cross section of interfaces in cold-sprayed
bring about structural changes and phase transformations, some of
deposits indicate grain refinement due to dynamic recrystalliza-
which are hardly attainable by conventional thermal spraying
tion, e.g. as observed in zinc [116], titanium [117] and aluminium
methods. This is so because of the severe plastic deformation at
[118]. TEM studies have revealed no bonding at the central zones,
high strain rates of up to 109 s1, combined with high pressures of
e.g. as observed in cold-sprayed titanium [119]. Microstructural
up to more than 10 GPa and extremely high cooling rates. These
studies of cold-sprayed deposits also include extensive FIB/TEM
changes include formation of metastable microstructures, in addi-
analysis of the particle-particle and particle-substrate interfaces,
tion to nanocrystallisation and amorphization as mentioned above;
e.g. by Kang et al. [103]. These studies particularly focus on the role
see e.g. the review by Luo et al. [133]. In some cases, the reverse
of surface oxide layers in CS, but also provide concrete evidence for
may happen. That is, the impact can result in deformation-induced
dynamic recrystallization at the interfacial regions.
transformation of an initially metastable phase into a more stable
Recrystallization is often associated with the formation of
state, e.g. bcc to fcc transformation in steel 316L [134] and in
nanostructures [120e122]. However, the extent and distribution of
MCrAlY [135].
nanograins depend on the material; e.g. being less uniform in
nickel as compared to copper [123]. This may be attributed to dif-
ferences in activation energies for recrystallization and to the dif- 4.2. Mechanical properties
ferences in local temperatures due to different thermal
conductivities of copper and nickel. In the cases where the starting Mechanical properties of cold-sprayed deposits are much
material is already nanostructured, the nanocrystalline structure is influenced by the density of lattice defects, the non-bonded areas,
often retained after CS, e.g. as observed in a cold-sprayed AleMg and the often severe change in the grain size. For instance, nano-
alloy [110]. There have also been reports of dynamic amorphization indentation of cold-sprayed titanium has shown a correlation be-
in AleAl, NieAl [124], AleCu [125], Cu-Ni [126], Al-Mg [127] and tween grain size and hardness, following a Hall-Petch relationship
FeeAl [128,129] systems, over a layer of up to tens of nanometres [114]. Another important property of cold-spray deposits is the
thickness. For the case of metallic glasses, on the other hand, general lack of ductility due to severe deformation, though this can
CS may result in nanocrystallisation of the initially amorphous be improved by heat treatment [136e138]. Heat treatments can
phase [89]. More examples of amorphization and dynamic result in recovery and recrystallization of the material, and also
398 H. Assadi et al. / Acta Materialia 116 (2016) 382e407

increase the extension of bonded interfaces. In theory, partial re-


covery and recrystallization may occur during spraying of some
materials, resulting in some levels of ductility. This is consistent
with the observation of dislocation-free regions in as-sprayed
copper samples, Figs. 23 and 24. The role of non-bonded in-
terfaces in the mechanical behaviour of cold-sprayed deposits can
be explained as follows. For every bulk material there is a rela-
tionship between hardness and tensile strength. In cold-sprayed
deposits, however, there is often a deviation from this relation-
ship, indicating different plastic behaviour and strength under
tension and compression. Differences in tensile and compressive
strengths normally indicate the presence of discontinuities and
cracks in the material. Deviations from hardness-strength re-
lationships have been observed for copper [139] and for nickel
[140]. These deviations can be minimised by using optimum
spraying conditions, leading to close-to-bulk mechanical properties
of the cold-sprayed deposits [140].
Despite the lack of ductility, both the cohesive and adhesive
strengths of cold-sprayed coatings can be rather high. For instance,
Fig. 25. Residual stress measured using neutron scattering for copper coatings cold
adhesion strengths of up to 120 MPa have been reported for thick
sprayed on copper substrate [147]. The solid lines correspond to the progressively
copper coatings sprayed on steel substrates [141]. In a similar case, deposited coating model [148] fitted into the experimental data.
using extreme spraying conditions and helium as process gas, this
has been reported to be as high as 200 MPa [142]. In these studies,
special testing methods have been used, since the standard A major part of cold spray research has aimed to investigate e often
methods using glue is limited to adhesion strengths of up to about through fluid dynamics considerations e the influence of the pro-
80 MPa. cess parameters on the deposition efficiency and on the final
properties of the deposits. On the other hand, properties of the
4.3. Residual stress deposits are known to depend mainly on the particle impact con-
ditions e i.e. particle impact velocity, vp, particle impact tempera-
It is a common presumption that cold-sprayed coatings have ture, Tp, and surface temperature of the substrate (or of the already
compressive residual stresses. There have been numerical studies deposited layer) Tsub. Thus, the aim of most investigations has been
of residual stress through multiple particle impact simulations, to work out the relationship between the cold spray process pa-
which confirm this view [143e145]. Theoretical evaluation of the rameters and these three key parameters, i.e. vp, Tp and Tsub.
residual stress is nevertheless not straightforward. The sign, Numerous studies in this area are based on fluid dynamics con-
magnitude and profile of residual stresses in cold-sprayed coatings siderations, which are out of scope of the present paper. Never-
depend on various factors, such as cold-spray process parameters, theless, Table 1 offers a qualitative summary of the effect of various
e.g. as shown for aluminium coating [146]. An example of measured process parameters on the relevant impact parameters, including
residual stress profiles using neutron diffraction [147] is shown in those relating to fluid dynamics.
Fig. 25.
The results were shown to fit to Tsui and Clyne’s progressive 5.2. Powder
coating deposition model [148]. According to this model, the
coating deposition is considered as progressive formation of a new The particle size distribution is the main attribute of CS pow-
layer on the top of the previously deposited coating layers. This new ders. The particle size has opposing effects on the particle velocity
layer is formed with a characteristic deposition stress, which is and the critical velocity. As discussed above, the critical velocity
invariably compressive. Nevertheless, after spraying, the substrate generally decreases with increasing particle size. On the other
and coating system is normally cooled down to room temperature, hand, larger particles are more difficult to accelerate and also have a
so that a thermal misfit strain may arise. Thus, the residual stress larger heat capacity. In view of these complex influences, there is no
can be influenced by the heat input, spraying kinematics and the unique optimum particle size for CS. As will be discussed later, the
associated thermal history. Based on these factors, in addition to optimum particle size depends on various factors, including the
substrate and coating material properties and dimensions, the spraying conditions, nozzle specifications and the length of pre-
mean residual stress in the coating can be compressive or tensile chamber. Particle morphology also plays a role, as discussed in
[149,150]. Ref. [151]. At low spraying conditions, an irregular, e.g. dendritic
powder may lead to lower porosity, while at higher spraying con-
5. Cold spray parameters ditions spherical powders result in higher strengths. Moreover,
particle size and morphology together determine the free surface
Cold spray parameters may be divided into three groups as area per volume, which is yet another relevant factor for CS. A
follows: (1) process parameters, such as gas conditions, nozzle larger surface area could result in faster oxidation kinetics, and
design and spraying kinematics, (2) powder properties, and (3) hence in degraded coating quality. In the case of spraying reactive
substrate properties, as discussed below. Based on these parame- powder mixtures for subsequent heat treatments, e.g. for com-
ters, materials can then be sorted in terms of sprayability, that is, bustion synthesis of intermetallic compounds, particle size and
the capability of being successfully deposited by CS. morphology may additionally influence the diffusive reaction
kinetics.
5.1. Process parameters According to Eq. (3), the critical velocity is considered to depend
on the particle properties. These properties are often taken to be
There are several process parameters that play a role in CS [98]. the same as those of bulk material, both for the estimation of the
H. Assadi et al. / Acta Materialia 116 (2016) 382e407 399

Table 1
Summary of the effect of process parameters on the impact conditions.

Factor Parameter Effect on vp Effect on Tp Effect on Tsub

Gas Typea vp (He) > vp (N2) Tp (He) < Tp (N2)


Tgas Tgas[ 0 vp[ Tgas[ 0 Tp[ Tgas[ 0 Tsub[
pgas pgas[ 0 vp[b pgas[ 0 Tp[Y
Nozzle Length, l l[ 0 vp[ l[ 0 TpY l[ 0 TsubY
Exp. ratio, r r[ 0 vp[ r[ 0 TpY r[ 0 TsubY
Diameter, d d[ 0 vp[c
Prechamber Length, lpred lpre[ 0 Tp[
Procedure Standoff distance, s s[ 0 vp[Ye s[ 0 TpY s[ 0 TsubY
Spraying angle, uf u[ 0 vpY
Traverse speed, vgun vgun[ 0 TsubY
Powder feed-rate, f f[ 0 vpY f[ 0 Tp[ f[ 0 Tsub[
a
Main gas properties are molar weight, specific heat, specific heat ratio and viscosity.
b
Impact velocities of fine particles may drop at high gas pressures because of bow shock.
c
Larger nozzle diameters normally result in reduced boundary layer and frictional effects.
d
This is the distance between the powder injection point and the throat of the nozzle.
e
Increases initially up to a maximum (normally within a few centimetres) then decreases.
f
Effect on vertical component of the impact velocity.

critical velocity and for numerical simulations of particle impact. subsequent layers. This effect is nevertheless limited only to a few
While this is a valid assumption for density and heat capacity, the particle layers, i.e. to a few tens of microns. After about 50 mm,
strength and plastic properties of micron-sized particles as used for the already deposited material deforms in the same way as in a
CS may be substantially different from that of bulk material. For much thicker deposit.
some materials e such as MCrAlYs, which are produced in the For very soft substrates such as polymers cold-spray deposition
powder form e the bulk material data may not even exist. There- may require using relatively low particle velocities and tempera-
fore, it may be necessary to obtain mechanical strength of the single tures, to allow successful deposition without eroding the substrate
particles, e.g. through single particle compression tests as described away. Examples of metallic coatings on polymer substrates are
in Ref. [41]. In combination with numerical simulation of particle given in Refs. [153e157]. Low velocities may also be used for
compression, a simple correlation has been worked out between coating deposition of metals on ceramics, to avoid fracture of the
the ultimate tensile strength and the force, F*, at a displacement generally brittle substrate. Additionally, deposition of metals on
equal to half of the initial particle diameter, dp , as follows: ceramics may be facilitated by using an interlayer [158] e or by
substrate preheating, as discussed below.
F* Particle bonding is influenced not only by the particle temper-
sUTS zð0:9±0:1Þ (10)
d2p ature, but also by the surface temperature of the substrate or the
already deposited layers on which the particles impact [159]. Note
For the case of copper particles, the single-particle compression that the surface temperature is different from the initial substrate
tests have shown a significant difference between particle and bulk temperature. It is influenced by various factors including the initial
properties. substrate temperature, spraying conditions, spraying kinematics
and the thermal properties and dimensions of the substrate. The
effect of substrate preheating on bonding is demonstrated for CS of
5.3. Substrate
magnesium [160], aluminium [161], copper [162,163], nickel [164]
and steel 316L [165]. The trend and the intensity of this effect
Substrate hardness influences the adhesion between the coating
depend on the combination of coating-substrate materials and on
and the substrate. This can be explained in terms of particle
the surface oxidation of the substrate, which may occur due to
deceleration, or the ‘impact power’, which is the rate of kinetic
excessive preheating time or temperature [166]. For the case of
energy dissipation during impact of a single particle. In CS, the
copper sprayed on ceramic (alumina) substrates, increased sub-
impact power is typically around 50 W. For harder substrates, the
strate temperature resulted in significant increase of not only the
impact power is higher and the heat generation is restricted to
adhesive strength, but also the overall coating quality [167]. This
smaller volumes, leading to more pronounced temperature rise and
may be attributed to partial recovery of the deposit during
facilitated ASI on the particle side. Likewise, an impacting particle
spraying.
penetrates more deeply into a softer substrate, facilitating ASI on
Substrate roughness may also play a role; e.g. it can lead to
the substrate side, while particle deceleration and the corre-
higher deposition efficiency of the first layer [168]. However, it may
sponding impact power may be lower. The adhesion strength is
not necessarily result in increased bond strength [169]. An example
influenced by the combination of impact power and the distribu-
of the effect of various surface preparation methods on CS of an
tion of energy dissipation in particle and substrate. Therefore,
aluminium alloy is given in Ref. [170].
hardness of the substrate is an important consideration for the
adhesion strength.
Mechanical properties of the substrate influence not only the 5.4. Sprayability and parameter selection
adhesion strength, but also the intrinsic properties of the coating
such as the flattening ratio of the particles [152]. This is because Conventionally, the critical velocity has been taken as a measure
of the effect of substrate hardness on the deformation behaviour of sprayability of various materials. This can nevertheless be
of the already deposited layers. For instance, a hard substrate misleading. As mentioned above, many cold-spray properties
constrains the deformation of the first layer of deposited parti- depend on not only the critical velocity but also the particle impact
cles, thus influences the cohesion between the first and conditions. Therefore, sprayability should be more appropriately
400 H. Assadi et al. / Acta Materialia 116 (2016) 382e407

discussed in view of the velocity ratio, h, or of the energy parameter 6. Materials and applications
hE .
Table 2 shows a selection of materials, sorted based on the Cold spraying can be used for coating, repair and additive
respective critical velocities, and also sorted based on the calculated manufacturing (AM). CS deposition of almost all metallic materials
velocity ratios for different particle sizes and spraying conditions. and metal-ceramic composites to any thickness above 50 mm is
The dark grey (green in colour version) background indicates technically achievable. For a review of successfully cold-sprayed
sprayability, while the light grey (orange in colour version) back- and studied material systems and applications see e.g.
ground signifies melting and high risk of nozzle clogging. It should Refs. [20,171e174]. Fig. 26 shows a few examples of cold-sprayed
be noted that not only low-melting point materials, but also nickel coatings, as investigated by the authors and co-workers at Hel-
and nickel alloys are prone to nozzle clogging at high process gas mut Schmidt University.
temperatures, thus practically limiting sprayability. Countermea- Table 3 gives a summary of materials and applications in CS. The
sures include water cooling as well as using nozzle materials that list is non-exhaustive, partly because of the nondisclosure of the
are resistant to clogging. Moreover, nozzle clogging can be avoided work that is performed outside academia. The applications can be
by using helium instead of nitrogen as process gas e though this categorised based on the intended purpose, as follows: (i) to obtain
may be commercially justified only in special cases e so that the high purity deposits, (ii) to obtain superior properties through well-
same value of h is achieved at lower temperatures. bonded interfaces and (iii) to ensure low heat input to the substrate.
It is interesting to note that the order of materials changes from High purity and well-bonded interfaces result in high electrical and
one column to another in the table. For instance, based on the thermal conductivity, as well as in high mechanical strength and
critical velocity, tantalum appears as the second ‘easiest’ material high corrosion/oxidation resistance. High chemical purity and ho-
for CS, after zinc. With respect to the velocity ratio, however, it is mogeneity of the cold-sprayed deposits make them particularly
not sprayable at 30 bar and 400  C. This is because of high density of suitable for applications in electronics, chemical and biomedical
tantalum particles, making them difficult to accelerate. The oppo- industries. Besides offering unique deposit properties, CS may also
site is true for magnesium. It is also interesting to note that the be a suitable option for applications where the substrate is sensitive
order of materials changes in every case, by changing the spraying to thermal shock e e.g. in the case of coating metals on glass or
condition or the particle size. This means that there is no unique set ceramics, or in the repair of heat-sensitive components. It has also
of optimum particle size, and that a suitable particle size must be been considered as a method to improve fatigue strength of the
selected with respect to the intended spraying conditions. In this substrate [177].
table, copper is the only material that remains sprayable under all As discussed before, properties of cold-sprayed deposits may be
conditions considered. affected by defects and non-bonded interfaces. This problem can

Table 2
Sprayability of selected materials*.
H. Assadi et al. / Acta Materialia 116 (2016) 382e407 401

Fig. 26. Examples of cold-sprayed coatings [16,27,45,175,176].

nevertheless be minimised by using relatively high velocity ratios e ensure orthogonal spraying (h decreases with deviation from right
typically higher than 1.5. Moreover, post spraying heat treatments angle) and constant standoff distance. In practice, material depo-
may be used to increase the extent of bonded areas and to spher- sition is combined with machining, and the final part is often heat
odise the remaining discontinuities. In some cases e e.g. in the treated to improve ductility. The concept of using CS for AM e e.g.
production of Ni-based superalloy turbine blades e post heat to produce titanium parts e is almost as old as the process itself
treatments may not need extra efforts, as they may already be an [178]. Commercial implementation of CS for AM has nevertheless
integral part of the production procedure. been more recent; an example of this for the production of complex
Another technically interesting characteristic of the CS process steel components is shown in Fig. 27.
is the relatively small spray-line width of 6e8 mm, which makes it
suitable for AM applications. For applications in electronics, how-
6.1. Metallic materials
ever, the line width may need to be further reduced. This can be
achieved via masking techniques. In repair or AM applications, CS
There have been many studies on cold spraying of metallic
offers deposition rates that are by about an order of magnitude
materials. The literature on this topic alone already includes a few
higher than those in selective laser deposition techniques. For these
review articles. Table 4 gives only a selection of the published work
applications, elaborate robotics are normally required, mainly to
on CS of different metallic materials. The table excludes the work on

Table 3
An overview of current materials and applications in cold spraying.

Material Application Sector Merits

Cu, Ag Conductive layers; high power electronics Electronics High chemical homogeneity, no oxidation
Sensors Automotive Similar conductivity as bulk material
Heat transfer layers Energy
Corrosion protection Nuclear industries
Al, Mg alloys Repair Aerospace Low amounts of interface defects
Similar strength as bulk material
Al, Ti alloys, Additive manufacturing Aerospace Low amounts of interface defects
Cu, Steels Manufacturing of injection/die casting moulds Tooling Similar mechanical/fatigue strength as bulk material
Polymer industry High deposition rates
Complex combination of different materials in one component
Ta, Nb, Cu alloys Additive manufacturing Electronics High purity, high chemical homogeneity, no oxidation
Sputter targets Machinery
Ni, Ta, Cathodic corrosion protection Offshore Low amounts of interface defects
High-Cr steels Chemical Ind. High chemical homogeneity, no oxidation
Zn, Al alloys Anodic corrosion protection Offshore High chemical homogeneity, no oxidation
Chemical Dense coatings
Machinery
Cu alloys Tribological performance Maritime technologies Low amounts of interface defects
Cavitation resistance Machinery Similar mechanical strength as bulk material
Steels ferritic austenitic Repair Energy Low amounts of interface defects
Magnetic properties (induction heating) Consumer goods Similar strength as bulk material
Machinery Chemical homogeneity, no oxidation
Ni alloys Hot gas corrosion protection Energy Low amounts of interface defects
Repair Aerospace Chemical homogeneity, no oxidation
Ti alloys Biocompatible parts Medicine Chemical homogeneity, purity, no oxidation
402 H. Assadi et al. / Acta Materialia 116 (2016) 382e407

vacuum cold spraying, i.e. aerosol deposition (AD) [31]. High


pressure cold spraying of monolithic ceramic powder usually
results in fracture of particles and fragments embedded in the
surface of metallic substrates. There are nevertheless a few ex-
amples where open-air cold spray has been used to deposit ce-
ramics. Examples include CS of hydroxyapatite (HA) for
biomedical applications e pure HA [212,213], HA mixed with
titanium [214] or doped with silver [215] e and titania as pho-
toelectrodes [216e220]. In most cases, the thickness of the
deposited material is limited to a monolayer of particles. Also,
the success in cold spray deposition of ceramics appears to
depend strongly on the structure of particles, which can be ag-
glomerates of nanoparticles (see section 3.2.6). The work on CS of
ceramics is in its infancy.
Fig. 27. Example of an additively manufactured part with copper and tool steel with In contrast, there are numerous examples of CS of composite
cooling channel e drawing (left) and real part (right) e using MPA technology from
materials. In most cases, there is a metallic component that is to
Hermle Maschinenbau GmbH, Ottobrunn, Germany.
provide deformability and binding between the harder (often
ceramic) particulates. Almost any combination of metals and ce-
metallic glasses, which are discussed separately below. ramics is conceivable; hence many are explored, including binary
CS is a particularly attractive method for the deposition of CueW [221,222] and ternary Cu-CNT-SiC nanocomposites [223].
metastable materials such as metallic glasses, without dimensional Composite coatings can be made by spraying of powder mixtures
restriction. In this respect, CS may be considered as the only [224], coated particles [225], ball-milled mixtures [226] or sintered
practical method through which bulk metallic glasses of any size agglomerates. There are some 30 published journal papers on
may be produced. With the availability of experimental gas- cemented carbides alone; among them just to name a few are those
atomized amorphous feedstock powder in large quantities, first on WC-Ni [227] and WC-Co [228e231]. Moreover, many articles
reports on CS of metallic glass coatings were published about a report on CS of Al-based composites; examples of recent studies
decade ago [16,84e86,97]. By using helium as process gas, opti- include CS of Al5056-Inconel718 [232], AleAl2O3 [233] and
mised process conditions, and in some cases powder preheating, Al5056eSiC [234].
dense coating could be produced. Fig. 28 shows an example of a Polymers can also be used as the deformable component [235],
cold-sprayed Zr-based metallic glass. The obtained coatings though this is not common. The second phase is sometimes a
exhibited tribological and corrosive properties similar to that of temperature sensitive material such as CNT [223], nanodiamond
cast BMG [208e210]. More recent studies demonstrated that [236,237], metallic glass [238], or quasicrystal [239,240], for which
amorphous coatings of Fe-based BMGs could be deposited by using cold spray offers a unique merit of not inducing thermal degrada-
nitrogen as process gas [87,88,96,211]. tion. In some cases a cold sprayed composite is produced as a
precursor for subsequent reactive synthesis, e.g. to form interme-
tallic compounds [241e243] or to make use of the generated heat
6.2. Ceramics and composites as in thermite reaction [244]. There are yet many combinations of
materials for various purposes to be explored.
Generally, ceramics can be sprayed successfully only through

Table 4
A selection of published work on cold spraying of metallic materials.

Material Reference Remark

Ag [179] Electrical conductivity


Al alloys A most widely studied material system; nearly a hundred published papers Aerospace
Automotive
Bronze [180e182] Cavitation resistance
Cu [24,70,107,183e185] Early studies, reference material
Nanocrystals
Electrodes
CueNieFe [186] Aluminium production
MCrAlY [27,135,187] Phase transformation
Turbine blades
Mg alloys [188] Review
Ni [189,190] Hydrogen generation
Nanocrystals
NieCr [191e194] High-temperature applications
Ni-based superalloys [195,196] High-temperature applications
NiTi [197,198] Shape memory
Nanocrystals
Steel 316L, 304L [134,199e201] Heat treatment
Phase transformation
Stellite [202e204] Wear resistance
Ta [197,198] Oxygen evolution reaction
Corrosion protection
Ti [205,206] Review
Zn, Zn alloys [176,207] Printing rolls
Corrosion protection
H. Assadi et al. / Acta Materialia 116 (2016) 382e407 403

[4] E. Irissou, J. Legoux, A.N. Ryabinin, B. Jodoin, C. Moreau, Review on cold spray
process and technology: Part I e intellectual property, J. Therm. Spray
Technol. 17 (2008) 495e516.
[5] R.C. McCune, A.N. Papyrin, J.N. Hall, W.L. Riggs, P.H. Zajchowski, An explo-
ration of the cold gas-dynamic spray method for several material systems,
in: C.C. Berndt, S. Sampath (Eds.), Advances in Thermal Spray Science and
Technology, ASM International, Materials Park, 1995, pp. 1e5.
[6] A.O. Tokarev, Structure of aluminum powder coatings prepared by cold gas
dynamic spraying, Metal Sci. Heat Treat. 38 (1996) 136e139.
[7] A.P. Alkhimov, S.V. Klinkov, V.F. Kosarev, A.N. Papyrin, Gas-dynamic spraying
study of a plane supersonic two-phase jet, J. Appl. Mech. Tech. Phys. 38
(1997) 324e330.
[8] A.P. Alkhimov, V.F. Kosarev, A.N. Papyrin, Gas-dynamic spraying. An exper-
imental study of the spraying process, J. Appl. Mech. Tech. Phys. 39 (1998)
318e323.
[9] A. Papyrin, Cold spray technology, Adv. Mater. Process. 159 (2001) 49e51.
[10] R.C. Dykhuizen, M.F. Smith, Gas dynamic principles of cold spray, J. Therm.
Spray Technol. 7 (1998) 205e212.
[11] D.L. Gilmore, R.C. Dykhuizen, R.A. Neiser, M.F. Smith, T.J. Roemer, Particle
velocity and deposition efficiency in the cold spray process, J. Therm. Spray
Technol. 8 (1999) 576e582.
[12] T.H. van Steenkiste, J.R. Smith, R.E. Teets, J.J. Moleski, D.W. Gorkiewicz,
R.P. Tison, D.R. Marantz, K.A. Kowalsky, W.L. Riggs, P.H. Zajchowski,
B. Pilsner, R.C. McCune, K.J. Barnett, Kinetic spray coatings, Surf. Coat.
Technol. 111 (1999) 62e71.
Fig. 28. Optical micrograph of the cross section of a Zr55Al10Ni5Cu30 (at. %) bulk [13] H. Kreye, T. Stoltenhoff, Cold spraying e a study of process and coating
characteristics, in: Proceedings of the International Thermal Spray Confer-
metallic glass coating on an AlMg3 substrate, cold sprayed with helium at 30 bar and
ence, 2000.
600  C [16].
[14] T. Xiong, J. Wu, H.Z. Jin, M. Li, X. Liu, T.F. Li, Introduction to a new technology
e cold gas dynamic spray, Corros. Sci. Prot. Technol. 13 (2001) 267e269.
[15] V.F. Kosarev, S.V. Klinkov, A.P. Alkhimov, A.N. Papyrin, On some aspects of
7. Summary and outlook gas dynamics of the cold spray process, J. Therm. Spray Technol. 12 (2003)
265e281.
[16] F. G€ artner, T. Schmidt, T. Stoltenhoff, H. Kreye, Recent developments and
Cold spraying may be regarded as an established method for potential applications of cold spraying, Adv. Eng. Mater. 8 (2006) 611e618.
coating and additive manufacturing. In terms of industrial appli- [17] V.K. Champagne, The Cold Spray Materials Deposition Process: Fundamen-
cations, it offers unique merits such as low oxidation and preser- tals and Applications, Elsevier, 2007.
[18] A. Papyrin, V. Kosarev, S. Klinkov, A. Alkhimov, V.M. Fomin, Cold Spray
vation of the initial microstructure, not achievable by any other Technology, Elsevier, 2006.
powder deposition method. It can also be used as an experimental [19] Q. Wang, M. Zhang, Review on recent research and development of cold
method to study materials under ultra-high strain-rate deforma- spray technologies, Key Eng. Mater. 533 (2013) 1e52.
[20] A. Moridi, S.M. Hassani-Gangaraj, M. Guagliano, M. Dao, Cold spray coating:
tion or to explore shock-induced transformation. There has been review of material systems and future perspectives, Surf. Eng. 30 (2014)
much interest in research on cold spraying, and hence, much 369e395.
knowledge has been produced on various aspects of this process [21] S. Grigoriev, A. Okunkova, A. Sova, P. Bertrand, I. Smurov, Cold spraying: from
process fundamentals towards advanced applications, Surf. Coat. Technol.
within the past two decades. The mechanism of deposition, at least 268 (2015) 77e84.
for metallic materials, is well explored and extensively discussed. [22] J. Villafuerte, Modern Cold Spray: Materials, Process, and Applications, 2015.
The process has thus been optimised for various materials and [23] C.M. Kay, J. Karthikeyan (Eds.), High Pressure Cold Spray: Principles and
Applications, ASM International, 2016.
applications. Nevertheless, there is much room for further studies €rtner, H. Kreye, Microstructures and key
[24] T. Stoltenhoff, C. Borchers, F. Ga
in cold spraying of materials systems and substrate/coating com- properties of cold-sprayed and thermally sprayed copper coatings, Surf. Coat.
binations that have not been explored so far. Also, new criteria may Technol. 200 (2006) 4947e4960.
[25] M.F. Smith, Comparing cold spray with thermal spray coating technologies,
be needed to understand and optimise deposition of materials with
in: The Cold Spray Materials Deposition Process: Fundamentals and Appli-
special deformation behaviour such as intermetallics, metallic cations, 2007, pp. 43e61.
glasses, ceramics and composites. [26] M. Watanabe, S. Kumai, High-speed deformation and collision behavior of
pure aluminum plates in magnetic pulse welding, Mater. Trans. 50 (2009)
2035e2042.
Acknowledgements [27] T. Stoltenhoff, H. Kreye, H.J. Richter, An analysis of the cold spray process and
its coatings, J. Therm. Spray Technol. 11 (2002) 542e550.
The authors thankfully acknowledge the German Research [28] K. Lee, M. Lee, J. Yu, H. Kim, Effect of powder preheating temperature on the
properties of titanium coating layers manufactured by kinetic spraying,
Foundation (DFG), grant numbers KR 1109/ and KL1945/1; the Mater. Trans. 55 (2014) 622e628.
German Foundation of Industrial Research Associations (AiF), grant [29] H. Koivuluoto, A. Coleman, K. Murray, M. Kearns, P. Vuoristo, High pressure
numbers IGF 12.671 N, 15.503 N, 17135 N, 18449 BG; the German cold sprayed (HPCS) and low pressure cold sprayed (LPCS) coatings prepared
from OFHC Cu feedstock: overview from powder characteristics to coating
Federal Ministry of Economics and Technology (BMWi), grant properties, J. Therm. Spray Technol. 21 (2012) 1065e1075.
number 16IN0695; the German Federal Ministry of Education and [30] J. Akedo, Room temperature impact consolidation (RTIC) of fine ceramic
Research (BMBF), grant number 03SF0482; and the Volkswagen powder by aerosol deposition method and applications to microdevices,
J. Therm. Spray Technol. 17 (2008) 181e198.
Foundation, grant number Experiment 88 903 for financially sup- [31] D. Hanft, J. Exner, M. Schubert, T. Sto €cker, P. Fuierer, R. Moos, An overview of
porting the research on cold spraying at Helmut Schmidt University the aerosol deposition method: process fundamentals and new trends in
Hamburg. materials applications, J. Ceram. Sci. Technol. 6 (2015) 147e182.
[32] M. Bray, A. Cockburn, W. O’Neill, The laser-assisted cold spray process and
deposit characterisation, Surf. Coat. Technol. 203 (2009) 2851e2857.
References [33] E.O. Olakanmi, M. Doyoyo, Laser-assisted cold-sprayed corrosion- and wear-
resistant coatings: a review, J. Therm. Spray Technol. 23 (2014) 765e785.
[1] A.P.Alkhimov, V.F. Kosarev, N.I. Nesterovich, A.N. Papyrin. Method of [34] D.K. Christoulis, S. Guetta, E. Irissou, V. Guipont, M.H. Berger, M. Jeandin,
applying coatings (SU 1618778); prio. June 6th, 1986. J. Legoux, C. Moreau, S. Costil, M. Boustie, Y. Ichikawa, K. Ogawa, Cold-
[2] A.P. Alkhimov, V.F. Kosarev, A.N. Papyrin, A method of cold gas-dynamic spraying coupled to nano-pulsed Nd-YaG laser surface pre-treatment,
deposition, Sov. Phys. Doklady 35 (1990) 1047e1049. J. Therm. Spray Technol. 19 (2010) 1062e1073.
[3] A.P. Alkhimov, A.N. Papyrin, V.F. Kosarev, N.I. Nesterovich, M.M. Shushpanov. [35] M. Kulmala, P. Vuoristo, Influence of process conditions in laser-assisted low-
Gas-dynamic spraying method for applying a coating (US 5302414): Patent; pressure cold spraying, Surf. Coat. Technol. 202 (2008) 4503e4508.
1994. [36] T. Jen, L. Pan, L. Li, Q. Chen, W. Cui, The acceleration of charged nano-particles
404 H. Assadi et al. / Acta Materialia 116 (2016) 382e407

in gas stream of supersonic De-Lavel-type nozzle coupled with static electric Schneiden/Welding Cut. 37 (1985) 297e302.
field, Appl. Therm. Eng. 27 (2007) 2877e2885. [66] A. Berlin, T.C. Nguyen, M.J. Worswick, Y. Zhou, Metallurgical analysis of
[37] H. Takana, K. Ogawa, T. Shoji, H. Nishiyama, Computational simulation of magnetic pulse welds of AZ31 magnesium alloy, Sci. Technol. Weld. Join. 16
cold spray process assisted by electrostatic force, Powder Technol. 185 (2011) 728e734.
(2008) 116e123. [67] S. Barradas, V. Guipont, R. Molins, M. Jeandin, M. Arrigoni, M. Boustie,
[38] G.R. Johnson, W.H. Cook, A constitutive model and data for metals subjected C. Bolis, L. Berthe, M. Ducos, Laser shock flier impact simulation of particle-
to large strains, high strain rates and high temperatures, in: Proceedings of substrate interactions in cold spray, J. Therm. Spray Technol. 16 (2007)
the 7th International Symposium on Ballistics, The Netherlands, 1983, pp. 548e556.
541e547. [68] T. Schmidt, Kaltgasspritzen e Eine Analyse des Materialverhaltens beim
[39] S. Rahmati, A. Ghaei, The use of particle/substrate material models in Partikelaufprall und die daraus abgeleitete Prozessoptimierung: Thesis,
simulation of cold-gas dynamic-spray process, J. Therm. Spray Technol. 23 Helmut-Schmidt-Universita €t, Hamburg, Shaker Verlag, Aachen, 2007.
(2014) 530e540. €rtner, T. Klassen, S. Dosta, J.M. Guilemany, Single
[69] M. Villa Vidaller, A. List, F. Ga
[40] H. Assadi, F. G€ artner, T. Stoltenhoff, H. Kreye, Bonding mechanism in cold gas impact bonding of cold sprayed Ti-6Al-4V powders on different substrates,
spraying, Acta Mater. 51 (2003) 4379e4394. J. Therm. Spray Technol. 24 (2015) 644e658.
[41] H. Assadi, I. Irkhin, H. Gutzmann, F. Ga €rtner, M. Schulze, M. Villa Vidaller, [70] T. Schmidt, H. Assadi, F. Ga €rtner, H. Richter, T. Stoltenhoff, H. Kreye,
T. Klassen, Determination of plastic constitutive properties of microparticles T. Klassen, From particle acceleration to impact and bonding in cold spray-
through single particle compression, Adv. Powder Technol. 26 (2015) ing, J. Therm. Spray Technol. 18 (2009) 794e808.
1544e1554. [71] M. Grujicic, C.L. Zhao, W.S. DeRosset, D. Helfritch, Adiabatic shear instability
[42] N.A. Fleck, J.W. Hutchinson, A phenomenological theory for strain gradient based mechanism for particles/substrate bonding in the cold-gas dynamic-
effects in plasticity, J. Mech. Phys. Solids 41 (1993) 1825e1857. spray process, Mater. Des. 25 (2004) 681e688.
[43] N.A. Fleck, G.M. Muller, M.F. Ashby, J.W. Hutchinson, Strain gradient plas- [72] W. Li, C. Zhang, X. Guo, C. Li, H. Liao, C. Coddet, Study on impact fusion at
ticity: theory and experiment, Acta Metall. Mater. 42 (1994) 475e487. particle interfaces and its effect on coating microstructure in cold spraying,
[44] T. Sakai, J.J. Jonas, Dynamic recrystallization: mechanical and microstructural Appl. Surf. Sci. 254 (2007) 517e526.
considerations, Acta Metall. 32 (1984) 189e209. [73] G. Bae, S. Kumar, S. Yoon, K. Kang, H. Na, H. Kim, C. Lee, Bonding features and
[45] T. Schmidt, F. G€ artner, H. Assadi, H. Kreye, Development of a generalized associated mechanisms in kinetic sprayed titanium coatings, Acta Mater. 57
parameter window for cold spray deposition, Acta Mater. 54 (2006) (2009) 5654e5666.
729e742. [74] W. Li, C. Li, G. Yang, Effect of impact-induced melting on interface micro-
[46] T.W. Wright, The Physics and Mathematics of Adiabatic Shear Bands, Cam- structure and bonding of cold-sprayed zinc coating, Appl. Surf. Sci. 257
bridge University Press, 2002. (2010) 1516e1523.
[47] S.E. Schoenfeld, T.W. Wright, A failure criterion based on material instability, [75] P.C. King, G. Bae, S.H. Zahiri, M. Jahedi, C. Lee, An experimental and finite
Int. J. Solids Struct. 40 (2003) 3021e3037. element study of cold spray copper impact onto two aluminum substrates,
[48] W. Li, D.D. Zhang, C.J. Huang, S. Yin, M. Yu, F.F. Wang, H. Liao, Modelling of J. Therm. Spray Technol. 19 (2010) 620e634.
impact behaviour of cold spray particles: review, Surf. Eng. 30 (2014) [76] T. Hussain, D.G. McCartney, P.H. Shipway, D. Zhang, Bonding mechanisms in
299e308. cold spraying: the contributions of metallurgical and mechanical compo-
[49] R.C. Dykhuizen, M.F. Smith, D.L. Gilmore, R.A. Neiser, X. Jiang, S. Sampath, nents, J. Therm. Spray Technol. 18 (2009) 364e379.
Impact of high velocity cold spray particles, J. Therm. Spray Technol. 8 (1999) [77] J. Wu, H. Fang, S. Yoon, H. Kim, C. Lee, The rebound phenomenon in kinetic
559e564. spraying deposition, Scr. Mater. 54 (2006) 665e669.
[50] E.S. Hertel Jr., R.L. Bell, M.G. Elrick, A.V. Farnsworth, G.I. Kerley, J.M. McGlaun, [78] K. Wang, L. Kong, Y. Tao, T. Li, T. Xiong, Numerical simulation of minimal
S.V. Petney, S.A. Silling, P.A. Taylor, L.C.T.H. Yarrington, A software family for average bonding strength to suppress rebounding in cold spraying Cu/Cu: a
multi-dimensional shock physics analysis, in: Shock Waves@ Marseille I, preliminary study, J. Therm. Spray Technol. 24 (2014) 75e85.
Springer, 1995, pp. 377e382. [79] Y. Ji, G. Bae, K. Kang, C. Lee, Influence of the interface temperature and strain
[51] H. Assadi, F. G€ artner, T. Stoltenhoff, H. Kreye, Application of analytical gradients on the impact energy model of a soft particle on a hard substrate
methods for understanding and optimization of cold spray process, in: during kinetic spraying, Metals Mater. Int. 17 (2011) 335e340.
P. Heinrich (Ed.), High-velocity Oxy-fuel Flame Spraying (HVOF) Colloquium, [80] M. Song, H. Araki, S. Kuroda, K. Sakaki, Reaction layer at the interface be-
2003, pp. 49e59. tween aluminium particles and a glass substrate formed by cold spray,
[52] M. Grujicic, J.R. Saylor, D.E. Beasley, W.S. DeRosset, D. Helfritch, Computa- J. Phys. D Appl. Phys. 46 (2013) 195301.
tional analysis of the interfacial bonding between feed-powder particles and [81] P.C. King, S.H. Zahiri, M. Jahedi, Focused ion beam micro-dissection of cold-
the substrate in the cold-gas dynamic-spray process, Appl. Surf. Sci. 219 sprayed particles, Acta Mater. 56 (2008) 5617e5626.
(2003) 211e227. [82] R. Drehmann, T. Grund, T. Lampke, B. Wielage, K. Manygoats, T. Schucknecht,
[53] W. Li, H. Liao, C. Li, G. Li, C. Coddet, X. Wang, On high velocity impact of D. Rafaja, Interface characterization and bonding mechanisms of cold gas-
micro-sized metallic particles in cold spraying, Appl. Surf. Sci. 253 (2006) sprayed Al coatings on ceramic substrates, J. Therm. Spray Technol. 24
2852e2862. (2014) 92e99.
[54] G. Bae, Y. Xiong, S. Kumar, K. Kang, C. Lee, General aspects of interface [83] A.L. Greer, Y.Q. Cheng, E. Ma, Shear bands in metallic glasses, Mater. Sci. Eng.
bonding in kinetic sprayed coatings, Acta Mater. 56 (2008) 4858e4868. R Rep. 74 (2013) 71e132.
[55] S. Yin, X. Wang, W. Li, B. Xu, Numerical investigation on effects of in- [84] S. Yoon, H. Kim, C. Lee, Deposition behavior of bulk amorphous NiTiZrSiSn
teractions between particles on coating formation in cold spraying, J. Therm. according to the kinetic and thermal energy levels in the kinetic spraying
Spray Technol. 18 (2009) 686e693. process, Surf. Coat. Technol. 200 (2006) 6022e6029.
[56] S. Yin, X. Wang, W. Li, H. Liao, H. Jie, Deformation behavior of the oxide film [85] S. Yoon, C. Lee, H. Choi, H. Jo, Kinetic spraying deposition behavior of bulk
on the surface of cold sprayed powder particle, Appl. Surf. Sci. 259 (2012) amorphous NiTiZrSiSn feedstock, Mater. Sci. Eng. A 415 (2006) 45e52.
294e300. [86] H. Choi, S. Yoon, G. Kim, H. Jo, C. Lee, Phase evolutions of bulk amorphous
[57] S. Yin, X. Wang, X. Suo, H. Liao, Z. Guo, W. Li, C. Coddet, Deposition behavior NiTiZrSiSn feedstock during thermal and kinetic spraying processes, Scr.
of thermally softened copper particles in cold spraying, Acta Mater. 61 Mater. 53 (2005) 125e130.
(2013) 5105e5118. [87] A. Concustell, J. Henao, S. Dosta, N. Cinca, I.G. Cano, J.M. Guilemany, On the
[58] F. Meng, H. Aydin, S. Yue, J. Song, The effects of contact conditions on the formation of metallic glass coatings by means of cold gas spray technology,
onset of shear instability in cold-spray, J. Therm. Spray Technol. 24 (2015) J. Alloys Compd. 651 (2015) 764e772.
711e719. [88] J. Henao, A. Concustell, I.G. Cano, N. Cinca, S. Dosta, J.M. Guilemany, Influence
[59] H. Assadi, F. Ga €rtner, T. Klassen, Modeling and simulation of cold spray, in: of cold gas spray process conditions on the microstructure of Fe-based
C.M. Kay, J. Karthikeyan (Eds.), High Pressure Cold Spray: Principles and amorphous coatings, J. Alloys Compd. 622 (2015) 995e999.
Applications, ASM International, 2016. [89] S. Yoon, G. Bae, Y. Xiong, S. Kumar, K. Kang, J. Kim, C. Lee, Strain-enhanced
[60] A. Manap, O. Nooririnah, H. Misran, T. Okabe, K. Ogawa, Experimental and nanocrystallization of a CuNiTiZr bulk metallic glass coating by a kinetic
SPH study of cold spray impact between similar and dissimilar metals, Surf. spraying process, Acta Mater. 57 (2009) 6191e6199.
Eng. 30 (2014) 335e341. [90] K. Brinkmann, H. Teichler, Flow state in molecular-dynamics-simulated
[61] W. Li, S. Yin, X. Wang, Numerical investigations of the effect of oblique deformed amorphous Ni0.5Zr0.5, Phys. Rev. B 66 (2002) 1842051e1842059.
impact on particle deformation in cold spraying by the SPH method, Appl. [91] W.L. Johnson, M.D. Demetriou, J.S. Harmon, M.L. Lind, K. Samwer, Rheology
Surf. Sci. 256 (2010) 3725e3734. and ultrasonic properties of metallic glass forming liquids: a potential energy
[62] H. Ogawa, Atomistic simulation of the aerosol deposition method with zir- landscape perspective, MRS Bull. 32 (2007) 644e650.
conia nanoparticles, Mater. Trans. 47 (2006) 1945e1948. [92] S. Yoon, Y. Xiong, K. Kang, G. Bae, C. Lee, Phase separation in kinetic sprayed
[63] B. Daneshian, H. Assadi, Impact behavior of intrinsically brittle nanoparticles: bulk metallic glasses, J. Phys. D Appl. Phys. 42 (2009) 182007.
a molecular dynamics perspective, J. Therm. Spray Technol. 23 (2014) [93] S. Yoon, Y. Xiong, H. Kim, C. Lee, Dependence of initial powder temperature
541e550. on impact behaviour of bulk metallic glass in a kinetic spray process, J. Phys.
[64] A.N. Papyrin, V.F. Kosarev, S.V. Klinkov, A.P. Alkhimov, On the interaction of D Appl. Phys. 42 (2009) 082004.
high speed particles with a substrate under the cold spraying, in: Pro- [94] M.D. Demetriou, W.L. Johnson, Shear flow characteristics and crystallization
ceedings of the International Thermal Spray Conference, 2002. kinetics during steady non-isothermal flow of Vitreloy-1, Acta Mater. 52
[65] H. Kreye, M. Hammerschmidt, C. Woidneck, Discussion on the fusion char- (2004) 3403e3412.
acteristics during the process of explosion welding, Schweissen und [95] B. Lohwongwatana, J. Schroers, W.L. Johnson, Strain rate induced
H. Assadi et al. / Acta Materialia 116 (2016) 382e407 405

crystallization in bulk metallic glass-forming liquid, Phys. Rev. Lett. 96 (2006) [125] S. Guetta, M.H. Berger, F. Borit, V. Guipont, M. Jeandin, M. Boustie,
075503. Y. Ichikawa, K. Sakaguchi, K. Ogawa, Influence of particle velocity on adhe-
[96] A. List, F. Ga€rtner, T. Mori, M. Schulze, H. Assadi, S. Kuroda, T. Klassen, Cold sion of cold-sprayed splats, J. Therm. Spray Technol. 18 (2009) 331e342.
spraying of amorphous Cu50Zr50 alloys, J. Therm. Spray Technol. 24 (2014) [126] Y. Xiong, X. Xiong, S. Yoon, G. Bae, C. Lee, Dependence of bonding mecha-
108e118. nisms of cold sprayed coatings on strain-rate-induced non-equilibrium
[97] S. Yoon, C. Lee, H. Choi, H. Kim, J. Bae, Impacting behavior of bulk metallic phase transformation, J. Therm. Spray Technol. 20 (2011) 860e865.
glass powder at an abnormally high strain rate during kinetic spraying, [127] Q. Wang, D. Qiu, Y. Xiong, N. Birbilis, M. Zhang, High resolution micro-
Mater. Sci. Eng. A 448e451 (2007) 911e915. structure characterization of the interface between cold sprayed Al coating
[98] H. Assadi, T. Schmidt, H. Richter, J. Kliemann, K. Binder, F. G€ artner, T. Klassen, and Mg alloy substrate, Appl. Surf. Sci. 289 (2014) 366e369.
H. Kreye, On parameter selection in cold spraying, J. Therm. Spray Technol. [128] K.H. Ko, J.O. Choi, H. Lee, Pretreatment effect of Cu feedstock on cold-sprayed
20 (2011) 1161e1176. coatings, J. Mater. Process. Technol. 214 (2014) 1530e1535.
[99] G. Bae, K. Kang, H. Na, J. Kim, C. Lee, Effect of particle size on the micro- [129] K. Ko, J.O. Choi, H. Lee, Y. Seo, S. Jung, S. Yu, Cold spray induced amorph-
structure and properties of kinetic sprayed nickel coatings, Surf. Coat. ization at the interface between Fe coatings and Al substrate, Mater. Lett. 149
Technol. 204 (2010) 3326e3335. (2015) 40e42.
[100] J. Wu, H. Fang, S. Yoon, C. Lee, H. Kim, Critical velocities for high speed [130] S.H. Zahiri, D. Fraser, M. Jahedi, Recrystallization of cold spray-fabricated CP
particle deposition in kinetic spraying, Mater. Trans. 47 (2006) 1723e1727. titanium structures, J. Therm. Spray Technol. 18 (2009) 16e22.
[101] F.F. Wang, W. Li, M. Yu, H. Liao, Prediction of critical velocity during cold [131] Y. Zou, W. Qin, E. Irissou, J. Legoux, S. Yue, J.A. Szpunar, Dynamic recrystal-
spraying based on a coupled thermomechanical Eulerian model, J. Therm. lization in the particle/particle interfacial region of cold-sprayed nickel
Spray Technol. 23 (2014) 60e67. coating, Scr. Mater. 61 (2009) 899e902.
[102] W. Li, C. Zhang, H. Wang, X.P. Guo, H. Liao, C. Li, C. Coddet, Significant in- [132] C. Lee, J. Kim, Microstructure of kinetic spray coatings: a review, J. Therm.
fluences of metal reactivity and oxide films at particle surfaces on coating Spray Technol. 24 (2015) 592e610.
microstructure in cold spraying, Appl. Surf. Sci. 253 (2007) 3557e3562. [133] X. Luo, C. Li, F. Shang, G. Yang, Y. Wang, C. Li, High velocity impact induced
[103] K. Kang, S. Yoon, Y. Ji, C. Lee, Oxidation dependency of critical velocity for microstructure evolution during deposition of cold spray coatings: a review,
aluminum feedstock deposition in kinetic spraying process, Mater. Sci. Eng. Surf. Coat. Technol. 254 (2014) 11e20.
A 486 (2008) 300e307. [134] C. Borchers, T. Schmidt, F. G€ artner, H. Kreye, High strain rate deformation
[104] K.H. Ko, J.O. Choi, H. Lee, B.J. Lee, Influence of oxide chemistry of feedstock on microstructures of stainless steel 316L by cold spraying and explosive
cold sprayed Cu coatings, Powder Technol. 218 (2012) 119e123. powder compaction, Appl. Phys. A Mater. Sci. Process. 90 (2008) 517e526.
[105] K. Kim, M. Watanabe, S. Kuroda, Bonding mechanisms of thermally softened [135] C. Borchers, T. Stoltenhoff, M. Hahn, M. Schulze, H. Assadi, C. Suryanarayana,
metallic powder particles and substrates impacted at high velocity, Surf. F. G€artner, T. Klassen, Strain-induced phase transformation of MCrAlY, Adv.
Coat. Technol. 204 (2010) 2175e2180. Eng. Mater. 17 (2015) 723e731.
[106] W.B. Choi, L. Li, V. Luzin, R. Neiser, T. Gna €upel-Herold, H.J. Prask, S. Sampath, [136] A.C. Hall, D.J. Cook, R.A. Neiser, T.J. Roemer, D.A. Hirschfeld, The effect of a
A. Gouldstone, Integrated characterization of cold sprayed aluminum coat- simple annealing heat treatment on the mechanical properties of cold-
ings, Acta Mater. 55 (2007) 857e866. sprayed aluminum, J. Therm. Spray Technol. 15 (2006) 233e238.
[107] C. Borchers, F. G€ artner, T. Stoltenhoff, H. Assadi, H. Kreye, Microstructural [137] X. Meng, J. Zhang, W. Han, J. Zhao, Y. Liang, Influence of annealing treatment
and macroscopic properties of cold sprayed copper coatings, J. Appl. Phys. 93 on the microstructure and mechanical performance of cold sprayed 304
(2003) 10064e10070. stainless steel coating, Appl. Surf. Sci. 258 (2011) 700e704.
[108] C. Borchers, F. Ga €rtner, T. Stoltenhoff, H. Kreye, Formation of persistent [138] R. Huang, M. Sone, W. Ma, H. Fukanuma, The effects of heat treatment on the
dislocation loops by ultra-high strain-rate deformation during cold spraying, mechanical properties of cold-sprayed coatings, Surf. Coat. Technol. 261
Acta Mater. 53 (2005) 2991e3000. (2015) 278e288.
[109] C. Borchers, F. Ga €rtner, T. Stoltenhoff, H. Kreye, Microstructural bonding €rtner, T. Stoltenhoff, J. Voyer, H. Kreye, S. Riekehr, M. Koçak, Mechanical
[139] F. Ga
features of cold sprayed face centered cubic metals, J. Appl. Phys. 96 (2004) properties of cold-sprayed and thermally sprayed copper coatings, Surf.
4288e4292. Coat. Technol. 200 (2006) 6770e6782.
[110] M. Rokni, C.A. Widener, A. Nardi, V.K. Champagne, Nano crystalline high [140] T. Klassen, F. Ga €rtner, H. Assadi, Process science of cold spray, in: C.M. Kay,
energy milled 5083 Al powder deposited using cold spray, Appl. Surf. Sci. J. Karthikeyan (Eds.), High Pressure Cold Spray: Principles and Applications,
305 (2014) 797e804. ASM International, 2016.
[111] J. Cho, Y. Jin, D. Park, H. Kim, I. Oh, K. Lee, Manufacture and properties of cold [141] R.G. Maev, V. Leshchynsky, E. Strumban, D. Dzhurinskiy, J. Kocimski,
spray deposited large thickness Cu coating material for sputtering target, E. Maeva, Structure and mechanical properties of thick copper coating made
Metals Mater. Int. 17 (2011) 157e166. by cold spray, J. Therm. Spray Technol. 25 (2016) 113e122.
[112] K. Kang, H. Park, G. Bae, C. Lee, Microstructure and texture of Al coating [142] R. Huang, W. Ma, H. Fukanuma, Development of ultra-strong adhesive
during kinetic spraying and heat treatment, J. Mater. Sci. 47 (2012) strength coatings using cold spray, Surf. Coat. Technol. 258 (2014) 832e841.
4053e4061. [143] M. Saleh, V. Luzin, K. Spencer, Analysis of the residual stress and bonding
[113] Y. Zhang, N. Brodusch, S. Descartes, R.R. Chromik, R. Gauvin, Microstructure mechanism in the cold spray technique using experimental and numerical
refinement of cold-sprayed copper investigated by electron channeling methods, Surf. Coat. Technol. 252 (2014) 15e28.
contrast imaging, Microsc. Microanal. 20 (2014) 1499e1506. [144] R. Ghelichi, S. Bagherifard, D. MacDonald, I. Fernandez-Pariente, B. Jodoin,
[114] D. Goldbaum, R.R. Chromik, N. Brodusch, R. Gauvin, Microstructure and M. Guagliano, Experimental and numerical study of residual stress evolution
mechanical properties of Ti cold-spray splats determined by electron chan- in cold spray coating, Appl. Surf. Sci. 288 (2014) 26e33.
neling contrast imaging and nanoindentation mapping, Microsc. Microanal. [145] W. Li, K. Yang, D. Zhang, X. Zhou, Residual stress analysis of cold-sprayed
21 (2015) 570e581. copper coatings by numerical simulation, J. Therm. Spray Technol. 25
[115] G. Bae, K. Kang, C. Lee, Nanoscale deformation twinning at ultrahigh strain (2016) 131e142.
rates during kinetic spraying of nickel, Mater. Lett. 89 (2012) 320e323. [146] K. Spencer, V. Luzin, N. Matthews, M. Zhang, Residual stresses in cold spray
[116] C. Li, W. Li, Y. Wang, Formation of metastable phases in cold-sprayed soft Al coatings: the effect of alloying and of process parameters, Surf. Coat.
metallic deposit, Surf. Coat. Technol. 198 (2005) 469e473. Technol. 206 (2012) 4249e4255.
[117] G. Bae, K. Kang, J. Kim, C. Lee, Nanostructure formation and its effects on the [147] V. Luzin, K. Spencer, M. Zhang, Residual stress and thermo-mechanical
mechanical properties of kinetic sprayed titanium coating, Mater. Sci. Eng. A properties of cold spray metal coatings, Acta Mater. 59 (2011) 1259e1270.
527 (2010) 6313e6319. [148] Y.C. Tsui, T.W. Clyne, An analytical model for predicting residual stresses in
[118] K. Kang, J. Won, G. Bae, S. Ha, C. Lee, Interfacial bonding and microstructural progressively deposited coatings, Thin Solid Films 306 (1997) 23e33.
evolution of Al in kinetic spraying, J. Mater. Sci. 47 (2012) 4649e4659. [149] T. Suhonen, T. Varis, S. Dosta, M. Torrell, J.M. Guilemany, Residual stress
[119] K. Kim, M. Watanabe, K. Mitsuishi, K. Iakoubovskii, S. Kuroda, Impact development in cold sprayed Al, Cu and Ti coatings, Acta Mater. 61 (2013)
bonding and rebounding between kinetically sprayed titanium particle and 6329e6337.
steel substrate revealed by high-resolution electron microscopy, J. Phys. D [150] Z. Arabgol, H. Assadi, T. Schmidt, F. Ga €rtner, T. Klassen, Analysis of thermal
Appl. Phys. 42 (2009) 65304. history and residual stress in cold-sprayed coatings, J. Therm. Spray Technol.
[120] P.D. Eason, S.C. Kennett, T.J. Eden, I. Krull, B. Kowalski, J.L. Jones, In situ 23 (2014) 84e90.
observation of microstrain relief in cold-sprayed bulk copper during thermal [151] W. Wong, P. Vo, E. Irissou, A.N. Ryabinin, J. Legoux, S. Yue, Effect of particle
annealing, Scr. Mater. 67 (2012) 791e794. morphology and size distribution on cold-sprayed pure titanium coatings,
[121] G. Bae, J. Jang, C. Lee, Correlation of particle impact conditions with bonding, J. Therm. Spray Technol. 22 (2013) 1140e1153.
nanocrystal formation and mechanical properties in kinetic sprayed nickel, [152] S. Yin, X. Wang, W. Li, H. Jie, Effect of substrate hardness on the deformation
Acta Mater. 60 (2012) 3524e3535. behavior of subsequently incident particles in cold spraying, Appl. Surf. Sci.
[122] A.M. Birt, V.K. Champagne Jr., R.D. Sisson Jr., D. Apelian, Microstructural 257 (2011) 7560e7565.
analysis of cold-sprayed Ti-6Al-4V at the micro- and nano-scale, J. Therm. [153] R. Lupoi, W. O’Neill, Deposition of metallic coatings on polymer surfaces
Spray Technol. 24 (2015) 1277e1288. using cold spray, Surf. Coat. Technol. 205 (2010) 2167e2173.
[123] Y. Zou, D. Goldbaum, J.A. Szpunar, S. Yue, Microstructure and nanohardness [154] X. Zhou, A.F. Chen, J.C. Liu, X.K. Wu, J.S. Zhang, Preparation of metallic
of cold-sprayed coatings, Scr. Mater. 62 (2010) 395e398. coatings on polymer matrix composites by cold spray, Surf. Coat. Technol.
[124] Y. Xiong, K. Kang, G. Bae, S. Yoon, C. Lee, Dynamic amorphization and 206 (2011) 132e136.
recrystallization of metals in kinetic spray process, Appl. Phys. Lett. 92 [155] A. Ganesan, J. Affi, M. Yamada, M. Fukumoto, Bonding behavior studies of
(2008) 194101. cold sprayed copper coating on the PVC polymer substrate, Surf. Coat.
406 H. Assadi et al. / Acta Materialia 116 (2016) 382e407

Technol. 207 (2012) 262e269. S. Chandra, S.S. Yoon, Cold spray deposition of copper electrodes on silicon
[156] M. Gardon, A. Latorre, M. Torrell, S. Dosta, J. Fern andez, J.M. Guilemany, Cold and glass substrates, J. Therm. Spray Technol. 22 (2013) 1092e1102.
gas spray titanium coatings onto a biocompatible polymer, Mater. Lett. 106 [186] ,
G. Goupil, S. Jucken, D. Poirier, J. Legoux, E. Irissou, B. Davis, D. Guay, L. Roue
(2013) 97e99. Cold sprayed Cu-Ni-Fe anode for Al production, Corros. Sci. 90 (2015)
[157] P.C. King, A. Poole, S. Horne, R. Nys, S. Gulizia, M.Z. Jahedi, Embedment of 259e265.
copper particles into polymers by cold spray, Surf. Coat. Technol. 216 (2013) [187] O. Stier, Fundamental cost analysis of cold spray, J. Therm. Spray Technol. 23
60e67. (2014) 131e139.
[158] H. Nakano, M. Yamada, M. Fukumoto, E. Yamaguchi, Microstructure of [188] V.K. Champagne, Repair of Magnesium Components by Cold Spray Tech-
interfacial region between cold-sprayed copper coating and AlN substrate niques. The Cold Spray Materials Deposition Process: Fundamentals and
coated with sputtered titanium and copper, J. Therm. Spray Technol. 20 Applications, Elsevier, 2007.
(2011) 407e411. [189] L. Ajdelsztajn, B. Jodoin, J.M. Schoenung, Synthesis and mechanical proper-
[159] J. Legoux, E. Irissou, C. Moreau, Effect of substrate temperature on the for- ties of nanocrystalline Ni coatings produced by cold gas dynamic spraying,
mation mechanism of cold-sprayed aluminum, zinc and tin coatings, Surf. Coat. Technol. 201 (2006) 1166e1172.
J. Therm. Spray Technol. 16 (2007) 619e626. [190] M. Aghasibeig, H. Monajatizadeh, P. Bocher, A. Dolatabadi, R. Wuthrich,
[160] X. Suo, M. Yu, W. Li, M. Planche, H. Liao, Effect of substrate preheating on C. Moreau, Cold spray as a novel method for development of nickel electrode
bonding strength of cold-sprayed Mg coatings, J. Therm. Spray Technol. 21 coatings for hydrogen production, Int. J. Hydrogen Energy 41 (2015)
(2012) 1091e1098. 227e238.
[161] Y. Xiong, W. Zhuang, M. Zhang, Effect of the thickness of cold sprayed [191] N. Bala, H. Singh, S. Prakash, High-temperature oxidation studies of cold-
aluminium alloy coating on the adhesive bond strength with an aluminium sprayed Ni-20Cr and Ni-50Cr coatings on SAE 213-T22 boiler steel, Appl.
alloy substrate, Surf. Coat. Technol. 270 (2015) 259e265. Surf. Sci. 255 (2009) 6862e6869.
[162] M. Yu, W. Li, F.F. Wang, X. Suo, H. Liao, Effect of particle and substrate pre- [192] N. Bala, H. Singh, S. Prakash, J. Karthikeyan, Investigations on the behavior of
heating on particle deformation behavior in cold spraying, Surf. Coat. HVOF and cold sprayed Ni-20Cr coating on T22 boiler steel in actual boiler
Technol. 220 (2013) 174e178. environment, J. Therm. Spray Technol. 21 (2012) 144e158.
[163] S. Yin, X. Suo, Z. Guo, H. Liao, X. Wang, Deposition features of cold sprayed [193] M. Kumar, H. Singh, N. Singh, R. Joshi, Erosion-corrosion behavior of cold-
copper particles on preheated substrate, Surf. Coat. Technol. 268 (2015) spray nanostructured Ni-20Cr coatings in actual boiler environment, Wear
252e256. 332e333 (2015) 1035e1043.
[164] S. Yin, X. Suo, Y. Xie, W. Li, R. Lupoi, H. Liao, Effect of substrate temperature [194] H. Singh, T.S. Sidhu, J. Karthikeyan, S. Kalsi, Evaluation of characteristics and
on interfacial bonding for cold spray of Ni onto Cu, J. Mater. Sci. 50 (2015) behavior of cold sprayed Ni-20Cr coating at elevated temperature in waste
7448e7457. incinerator plant, Surf. Coat. Technol. 261 (2015) 375e384.
[165] Y. Xie, M. Planche, R. Raoelison, H. Liao, X. Suo, P. Herve , Effect of substrate [195] D. Levasseur, S. Yue, M. Brochu, Pressureless sintering of cold sprayed
preheating on adhesive strength of SS 316L cold spray coatings, J. Therm. Inconel 718 deposit, Mater. Sci. Eng. A 556 (2012) 343e350.
Spray Technol. 25 (2016) 123e130. [196] W. Wong, E. Irissou, P. Vo, M. Sone, F. Bernier, J. Legoux, H. Fukanuma, S. Yue,
[166] Y. Watanabe, C. Yoshida, K. Atsumi, M. Yamada, M. Fukumoto, Influence of Cold spray forming of Inconel 718, J. Therm. Spray Technol. 22 (2013)
substrate temperature on adhesion strength of cold-sprayed coatings, 413e421.
J. Therm. Spray Technol. 24 (2014) 86e91. [197] S. Tria, O. Elkedim, W. Li, H. Liao, Ball milled Ni-Ti powder deposited by cold
[167] K. Ernst, J. Braeutigam, F. Gaertner, T. Klassen, Effect of substrate tempera- spraying, J. Alloys Compd. 483 (2009) 334e336.
ture on cold-gas-sprayed coatings on ceramic substrates, J. Therm. Spray [198] S. Tria, O. Elkedim, R. Hamzaoui, X. Guo, F. Bernard, N. Millot, O. Rapaud,
Technol. 22 (2013) 422e432. Deposition and characterization of cold sprayed nanocrystalline NiTi, Pow-
[168] P. Richer, B. Jodoin, L. Ajdelsztajn, E.J. Lavernia, Substrate roughness and der Technol. 210 (2011) 181e188.
thickness effects on cold spray nanocrystalline Al-Mg coatings, J. Therm. [199] M. Villa, S. Dosta, J.M. Guilemany, Optimization of 316L stainless steel
Spray Technol. 15 (2006) 246e254. coatings on light alloys using cold gas spray, Surf. Coat. Technol. 235 (2013)
[169] T. Marrocco, D.G. McCartney, P.H. Shipway, A. Sturgeon, Production of tita- 220e225.
nium deposits by cold-gas dynamic spray: numerical modeling and experi- [200] B. Al-Mangour, P. Vo, R. Mongrain, E. Irissou, S. Yue, Effect of heat treatment
mental characterization, J. Therm. Spray Technol. 15 (2006) 263e272. on the microstructure and mechanical properties of stainless steel 316L
[170] M. Sharma, T.J. Eden, B. Golesich, Effect of surface preparation on the coatings produced by cold spray for biomedical applications, J. Therm. Spray
microstructure, adhesion, and tensile properties of cold-sprayed aluminum Technol. 23 (2014) 641e652.
coatings on AA2024 substrates, J. Therm. Spray Technol. 24 (2014) 410e422. [201] P. Coddet, C. Verdy, C. Coddet, F. Debray, F. Lecouturier, Mechanical prop-
[171] W. Li, H. Liao, H. Wang, Cold Spraying of Light Alloys. Surface Engineering of erties of thick 304L stainless steel deposits processed by He cold spray, Surf.
Light Alloys: Aluminium, Magnesium and Titanium Alloys, Elsevier, 2010. Coat. Technol. 277 (2015) 74e80.
[172] N. Bala, H. Singh, J. Karthikeyan, S. Prakash, Cold spray coating process for [202] N. Cinca, E. Lo pez, S. Dosta, J.M. Guilemany, Study of Stellite-6 deposition by
corrosion protection: a review, Surf. Eng. 30 (2014) 414e421. cold gas spraying, Surf. Coat. Technol. 232 (2013) 891e898.
[173] V.K. Champagne, D.J. Helfritch, Mainstreaming cold spray e push for appli- [203] N. Cinca, J.M. Guilemany, Structural and properties characterization of Stel-
cations, Surf. Eng. 30 (2014) 396e403. lite coatings obtained by cold gas spraying, Surf. Coat. Technol. 220 (2013)
[174] A.M. Vilardell, N. Cinca, A. Concustell, S. Dosta, I.G. Cano, J.M. Guilemany, 90e97.
Cold spray as an emerging technology for biocompatible and antibacterial [204] G. Lucchetta, R. Giusti, S. Vezzù, P. Bariani, Investigation and characterization
coatings: state of art, J. Mater. Sci. 50 (2015) 4441e4462. of Stellite-based wear-resistant coatings applied to steel moulds by cold-
[175] F. Ga€rtner, T. Stoltenhoff, T. Schmidt, H. Kreye, The cold spray process and its spray, CIRP Ann. Manuf. Technol. 64 (2015) 535e538.
potential for industrial applications, J. Therm. Spray Technol. 15 (2006) [205] T. Hussain, Cold spraying of titanium: a review of bonding mechanisms,
223e232. microstructure and properties, Key Eng. Mater. 533 (2013) 53e90.
[176] K. Onizawa, M. Schulze, F. G€ artner, T. Klassen, Cold gas spraying of zinc and [206] P. Vo, D. Goldbaum, W. Wong, E. Irissou, J. Legoux, R.R. Chromik, S. Yue, Cold-
zinc alloy coatings for printing applications, Weld. Cut. 11 (2012) 128e133. spray processing of titanium and titanium alloys, in: Titanium Powder
[177] C. Ziemian, M. Sharma, B. Bouffard, T. Nissley, T.J. Eden, Effect of substrate Metallurgy, Elsevier, 2015, pp. 405e423.
surface roughening and cold spray coating on the fatigue life of AA2024 [207] N.M. Chavan, B. Kiran, A. Jyothirmayi, P.S. Phani, G. Sundararajan, The
specimens, Mater. Des. 54 (2014) 212e221. corrosion behavior of cold sprayed zinc coatings on mild steel substrate,
[178] A.E. Segall, A.N. Papyrin, J.C. Conway Jr., D. Shapiro, A cold-gas spray coating J. Therm. Spray Technol. 22 (2013) 463e470.
process for enhancing titanium, JOM 50 (1998) 52e54. [208] H. Choi, H. Jo, K. An, S. Yoon, C. Lee, Tribological behavior of the kinetic
[179] N.M. Chavan, M. Ramakrishna, P.S. Phani, D.S. Rao, G. Sundararajan, The sprayed Ni59Ti16Zr20Si2Sn3 bulk metallic glass, J. Alloys Compd. 434e435
influence of process parameters and heat treatment on the properties of cold (2007) 64e67.
sprayed silver coatings, Surf. Coat. Technol. 205 (2011) 4798e4807. [209] S. Yoon, J. Kim, G. Bae, B. Kim, C. Lee, Formation of coating and tribological
[180] W. Li, C. Li, H. Liao, C. Coddet, Effect of heat treatment on the microstructure behavior of kinetic sprayed Fe-based bulk metallic glass, J. Alloys Compd.
and microhardness of cold-sprayed tin bronze coating, Appl. Surf. Sci. 253 509 (2011) 347e353.
(2007) 5967e5971. [210] S.J. Choi, H.S. Lee, J.W. Jang, S. Yi, Corrosion behavior in a 3.5 wt% NaCl so-
[181] X. Guo, G. Zhang, W. Li, L. Dembinski, Y. Gao, H. Liao, C. Coddet, Micro- lution of amorphous coatings prepared through plasma-spray and cold-
structure, microhardness and dry friction behavior of cold-sprayed tin spray coating processes, Metals Mater. Int. 20 (2014) 1053e1057.
bronze coatings, Appl. Surf. Sci. 254 (2007) 1482e1488. [211] A. List, F. G€
artner, T. Schmidt, T. Klassen, Impact conditions for cold spraying
[182] S. Krebs, F. Ga €rtner, T. Klassen, Cold spraying of Cu-Al-bronze for cavitation of hard metallic glasses, J. Therm. Spray Technol. 21 (2012) 531e540.
protection in marine environments, J. Therm. Spray Technol. 24 (2014) [212] A. Noorakma, H. Zuhailawati, V. Aishvarya, B. Dhindaw, Hydroxyapatite-
126e135. coated magnesium-based biodegradable alloy: cold spray deposition and
[183] R.C. McCune, W.T. Donlon, O.O. Popoola, E.L. Cartwright, Characterization of simulated body fluid studies, J. Mater. Eng. Perform. 22 (2013) 2997e3004.
copper layers produced by cold gas-dynamic spraying, J. Therm. Spray [213] M. Hasniyati, H. Zuhailawati, R. Sivakumar, B. Dhindaw, Optimization of
Technol. 9 (2000) 73e82. multiple responses using overlaid contour plot and steepest methods anal-
[184] J. Liu, X. Zhou, X. Zheng, H. Cui, J. Zhang, Tribological behavior of cold- ysis on hydroxyapatite coated magnesium via cold spray deposition, Surf.
sprayed nanocrystalline and conventional copper coatings, Appl. Surf. Sci. Coat. Technol. 280 (2015) 250e255.
258 (2012) 7490e7496. [214] X. Zhou, P. Mohanty, Electrochemical behavior of cold sprayed hydroxyap-
[185] D. Kim, J. Park, J. Lee, D. Kim, S.J. Tark, S. Ahn, J.H. Yun, J. Gwak, K.H. Yoon, atite/titanium composite in Hanks’ solution, Electrochim. Acta 65 (2012)
H. Assadi et al. / Acta Materialia 116 (2016) 382e407 407

134e140. Technol. 268 (2015) 180e189.


[215] N. Sanpo, M.L. Tan, P. Cheang, K.A. Khor, Antibacterial property of cold- [232] M. Yu, H. Chen, W. Li, X. Suo, H. Liao, Building-up process of cold-sprayed
sprayed HA-Ag/PEEK coating, J. Therm. Spray Technol. 18 (2009) 10e15. Al5056/In718 composite coating, J. Therm. Spray Technol. 24 (2014)
[216] M. Yamada, H. Isago, H. Nakano, M. Fukumoto, Cold spraying of TiO 2 pho- 579e586.
tocatalyst coating with nitrogen process gas, J. Therm. Spray Technol. 19 [233] J. Shockley, S. Descartes, P. Vo, E. Irissou, R.R. Chromik, The influence of
(2010) 1218e1223. Al2O3 particle morphology on the coating formation and dry sliding wear
[217] J. Lee, D. Kim, J. Park, Y. Cha, J. Yoon, H. Jeon, B. Min, M. Swihart, S. Jin, S. Al- behavior of cold sprayed Al-Al2O3 composites, Surf. Coat. Technol. 270
Deyab, S.S. Yoon, Graphene-titania hybrid photoanodes by supersonic ki- (2015) 324e333.
netic spraying for solar water splitting, J. Am. Ceram. Soc. 97 (2014) [234] M. Yu, X. Suo, W. Li, Y. Wang, H. Liao, Microstructure, mechanical property
3660e3668. and wear performance of cold sprayed Al5056/SiCp composite coatings:
[218] H. Gutzmann, Kaltgasspritzen von Titandioxid e Bindungsverhalten und effect of reinforcement content, Appl. Surf. Sci. 289 (2014) 188e196.
photokatalytische Eigenschaften: Thesis, Helmut-Schmidt-Universita €t, [235] K. Ravi, Y. Ichikawa, K. Ogawa, T. Deplancke, O. Lame, J. Cavaille, Mechanistic
Hamburg, 2014 (available online: http://edoc.sub.uni-hamburg.de/hsu/ study and characterization of cold-sprayed ultra-high molecular weight
volltexte/2014/3067/pdf/2014_Gutzmann.pdf). polyethylene-nano-ceramic composite coating, J. Therm. Spray Technol. 25
[219] M. Gardon, J.M. Guilemany, Milestones in functional titanium dioxide ther- (2016) 160e169.
mal spray coatings: a review, J. Therm. Spray Technol. 23 (2014) 577e595. [236] D. Woo, B. Sneed, F. Peerally, F. Heer, L. Brewer, J. Hooper, S. Osswald, Syn-
[220] I. Herrmann-Geppert, P. Bogdanoff, T. Emmler, T. Dittrich, J. Radnik, thesis of nanodiamond-reinforced aluminum metal composite powders and
T. Klassen, H. Gutzmann, M. Schieda, Cold gas spraying e a promising coatings using high-energy ball milling and cold spray, Carbon 63 (2013)
technique for photoelectrodes: the example TiO2, Catal. Today 260 (2016) 404e415.
140e147. [237] D. Woo, F. Heer, L. Brewer, J. Hooper, S. Osswald, Synthesis of nanodiamond-
[221] H. Kang, S.B. Kang, Tungsten/copper composite deposits produced by a cold reinforced aluminum metal matrix composites using cold-spray deposition,
spray, Scr. Mater. 49 (2003) 1169e1174. Carbon 86 (2015) 15e25.
[222] J. Voyer, T. Stoltenhoff, T. Schmidt, H. Kreye, Method and potential of the cold [238] N. Kang, P. Coddet, H. Liao, C. Coddet, The effect of heat treatment on
spray process, in: P. Heinrich (Ed.), High-velocity Oxy-fuel Flame Spraying microstructure and tensile properties of cold spray Zr base metal glass/Cu
(HVOF) Colloquium, 2003, pp. 39e47. composite, Surf. Coat. Technol. 280 (2015) 64e71.
[223] E. Pialago, O. Kwon, C.W. Park, Cold spray deposition of mechanically alloyed [239] X. Guo, J. Chen, H. Yu, H. Liao, C. Coddet, A study on the microstructure and
ternary Cu-CNT-SiC composite powders, Ceram. Int. 41 (2015) 6764e6775. tribological behavior of cold-sprayed metal matrix composites reinforced by
[224] R.G. Maev, V. Leshchynsky, Air gas dynamic spraying of powder mixtures: particulate quasicrystal, Surf. Coat. Technol. 268 (2015) 94e98.
theory and application, J. Therm. Spray Technol. 15 (2006) 198e205. [240] N. Khun, R. Li, K. Loke, K.A. Khor, Effects of Al-Cr-Fe quasicrystal content on
[225] H. Kim, D.J. Jung, C. Lee, The effect of Ni coating for diamond on the metal/ tribological properties of cold-sprayed titanium composite coatings, Tribol.
diamond composite coatings by cold spray deposition, in: Proceedings of the Trans. 58 (2015) 616e624.
International Thermal Spray Conference, 2009. [241] S. Dean, J. Potter, R. Yetter, T.J. Eden, V. Champagne, M. Trexler, Energetic
[226] E. Pialago, C.W. Park, Cold spray deposition characteristics of mechanically intermetallic materials formed by cold spray, Intermetallics 43 (2013)
alloyed Cu-CNT composite powders, Appl. Surf. Sci. 308 (2014) 63e74. 121e130.
[227] D. Lioma, N. Sacks, I. Botef, Cold gas dynamic spraying of WC-Ni cemented [242] A. Bacciochini, S. Bourdon-Lafleur, C. Poupart, M. Radulescu, B. Jodoin, Ni-Al
carbide coatings, Int. J. Refract. Metals Hard Mater. 49 (2015) 365e373. nanoscale energetic materials: phenomena involved during the
[228] R. Lima, J. Karthikeyan, C. Kay, J. Lindemann, C.C. Berndt, Microstructural manufacturing of bulk samples by cold spray, J. Therm. Spray Technol. 23
characteristics of cold-sprayed nanostructured WC-Co coatings, Thin Solid (2014) 1142e1148.
Films 416 (2002) 129e135. 
[243] V. Jan, J. Cupera, J. Cizek, On the search for producing intermetallics by
[229] P. Gao, C. Li, G. Yang, Y. Li, C. Li, Influence of substrate hardness on deposition diffusion reaction of cold spray bulk deposits, Surf. Coat. Technol. 268 (2015)
behavior of single porous WC-12Co particle in cold spraying, Surf. Coat. 216e223.
Technol. 203 (2008) 384e390. [244] A. Bacciochini, M.I. Radulescu, M. Yandouzi, G. Maines, J.J. Lee, B. Jodoin,
[230] G. Ji, X. Chen, H. Wang, X. Bai, Z. Dong, Deformation behaviors of cold- Reactive structural materials consolidated by cold spray: Al-CuO thermite,
sprayed WC-Co particles, J. Therm. Spray Technol. 24 (2015) 1100e1110. Surf. Coat. Technol. 226 (2013) 60e67.
[231] M. Couto, S. Dosta, J.M. Guilemany, Comparison of the mechanical and €rtner, T. Nanbu, Y. Kawashita, Y. Uehara, Modelling
[245] H. Assadi, T. Klassen, F. Ga
electrochemical properties of WC-17 and 12Co coatings onto Al7075-T6 of impact and bonding of inhomogeneous particles in cold spraying, in:
obtained by high velocity oxy-fuel and cold gas spraying, Surf. Coat. Proceedings of the International Thermal Spray Conference, 2014.