Mechanics of Materials 40 (2008) 982–1000

Contents lists available at ScienceDirect

Mechanics of Materials
journal homepage: www.elsevier.com/locate/mechmat

Thermomechanical analyses of ultrasonic welding process using

thermal and acoustic softening effects
A. Siddiq *, E. Ghassemieh
Mechanical Engineering Department, The University of Sheffield, Mappin Street, Sheffield, Yorkshire S1 3JD, UK

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

Article history: Ultrasonic welding process is a rapid manufacturing process used to weld thin layers of
Received 1 August 2007 metal at low temperatures and low energy consumption. Experimental results have shown
Received in revised form 13 June 2008 that ultrasonic welding is a combination of both surface (friction) and volume (plasticity)
softening effects. In the presented work, a very first attempt has been made to simulate the
ultrasonic welding of metals by taking into account both of these effects (surface and vol-
ume). A phenomenological material model has been proposed which incorporates these
Keywords:
two effects (i.e. surface and volume). The thermal softening due to friction and ultrasonic
Ultrasonic welding
Thermomechanical analysis
(acoustic) softening has been included in the proposed material model. For surface effects a
Friction laws friction law with variable coefficient of friction dependent upon contact pressure, slip, tem-
Ultrasonic softening perature and number of cycles has been derived from experimental friction tests. Thermo-
mechanical analyses of ultrasonic welding of aluminium alloy have been performed. The
effects of ultrasonic welding process parameters, such as applied load, amplitude of ultra-
sonic vibration, and velocity of welding sonotrode on the friction work at the weld inter-
face are being analyzed. The change in the friction work at the weld interface has been
explained on the basis of softening (thermal and acoustic) of the specimen during the ultra-
sonic welding process. In the end, a comparison between experimental and simulated
results has been presented showing a good agreement.
Ó 2008 Elsevier Ltd. All rights reserved.

1. Introduction Hu et al., 1991; Mayer and Schwizer, 2002; Sheaffer and

Ultrasonic welding is a process in which ultrasonic en-
ergy is used to create a solid-state bond between two
pieces of metal. Ultrasonic welding is a versatile and pow-
erful joining technique in the microelectronic packaging
industry because of the low temperature, high yield rate
and flexibility of the process (Harman, 1997). The main
advantages of ultrasonic welding include, absence of
liquid–solid transformations, low energy consumption, no
atmosphere control required, works for dissimilar metals,
low temperature allows embedding of electronics, such
as sensors and actuators and most importantly, it is
environmental friendly and very fast (Dushkes, 1973;
* Corresponding author. Tel.: +44 7726467348.
E-mail address: a.siddiq@shef.ac.uk (A. Siddiq).
Levine, 1991).
The mechanism of creation of an ultrasonic weld has
been under study since many decades but still not fully
understood (Tucker, 2002). A very simple definition of
the creation of a weld was proposed by Tucker (2002),
i.e. creation of the weld is a process in which ultrasonic
interfacial motion between the two mating surfaces breaks
and disperse the surface oxides, dirt and other contami-
nants leaving clean intimate surfaces which then create
bonds. Joshi (1971) has performed studies on ultrasonic
welding of aluminium, copper and gold wires. A series of
tests were performed to quantify temperature rise to
determine if there exists localized melting during the weld
formation. It was found that interfacial temperature read-
ings were less than 70 °C suggesting that localized melting
does not occur during ultrasonic welding of wire bonds. It
was also found that inter-diffusion at dissimilar bond

0167-6636/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.mechmat.2008.06.004
 A. Siddiq, E. Ghassemieh / Mechanics of Materials 40 (2008) 982–1000
interfaces do not occur. Transmission electron microscopy
studies of ultrasonic wire bonding were performed by
Krzanowski (1990). Transmission electron microscopy
(TEM) was performed to examine ultrasonic bonding of
aluminium wire to various metal substrates. There was
no evidence found of inter-diffusion of atoms from sub-
strate to the foil or vice versa. Both Joshi (1971) and
Krzanowski (1990) found substantial amount of plastic
flow around the hard members. Harman (1997) found that
the plastic deformation is not only seen with in the micro-
structure of the weld zones but also at the interface of the
mating materials. Transmission electron microscopy stud-
ies by James (1990) and Kazumasa et al. (2003) showed
that vacancies or voids, debris or surface contaminants or
oxides, and dislocations were frequently observed near
the bonding interface. Harman and Albers (1977) per-
formed studies on aluminium and gold wire bonding in
microelectronics. Their experimental results showed that
ultrasonic bonding takes place primarily by means of a
deformation mechanism (ultrasonic softening) rather than
heating or sliding mechanism. Schwizer et al. (1999) stud-
ied the ultrasonic ball bonding process. It was shown that
stick–slip motion is necessary for high quality bonding.
Prieb (1999) performed experimental studies on the ultra-
sonic joining of copper–copper, copper–aluminium, cop-
per–silver and copper–steel metal pairs. It was found
that severe plastic deformation of the metal plays an
essential role during joining. It was also found that region
of maximum deformation is located below the top surface
at the foil/sonotrode interface. It was also found that larg-
est microstructural changes occur near the upper surface
at the foil/sonotrode interface. It was proposed that main
mechanisms which lead to the formation of a metallic
bond are removal of oxide coatings, metal plastification
and even flow in the boundary region. No evidences were
found of volume melting of the metals, melting of the met-
als along the contact interface, diffusion of the metals or al-
loy components into each other at the weld interface.
Ultrasonic welding of two different alloys of aluminium,
i.e. AA 6061 and AA 3003 was performed by Kong et al.
(2003) and Kong et al. (2004a,b). Kong et al. (2003) per-
formed ultrasonic welding of aluminium alloy 6061. It
was found that a thick oxide film (magnesium oxide) exists
along the weld interface. This oxide film gets compacted
due to the dynamic interfacial stresses, generated during
ultrasonic vibration, causing the oxide layer to form brittle
ceramic bonds at the weld interface. It was also shown that
the density of metallurgical bond can be increased by a
simple cleaning procedure. The cleaning process consisted
of simple cleaning with the help of a degreaser (petroleum
distillate) and wiped with a clean cotton cloth to remove
oxides and other contaminants. Micro-hardness tests on
welded and un-welded AA 6061 specimens were per-
formed in Kong et al. (2004a,b). It was found that hardness
near the weld interface is larger than the hardness away
from the interface. This difference in hardness values de-
picted that both, surface effect (friction) at the weld inter-
face and ultrasonic softening (acoustic softening) in the
material, were present during the welding. Similar studies
were performed on aluminium alloy 3003 using ultrasonic
welding (Kong et al., 2004a,b). It was found that unlike AA
983
6061, AA 3003 does not require cleaning prior to the weld-
ing. Gunduz et al. (2005) performed studies on the ultra-
sonic welding of zinc and aluminium at elevated
temperatures (513 K). It was found that at such high tem-
peratures, weld interface exhibits structures indicative of
enhanced inter-diffusion and local melting of aluminium
and zinc solid solution.
Cheng and Li (2007) investigated the heat generation
and temperature profile during ultrasonic metal welding
using micro sensor arrays. The materials used for welding
were copper alloy and nickel substrate. It was found that
temperature ranges from 100 to 250 °C for various loading
cases.
Daud et al. (2006, in press) performed studies (both
experimental and simulation) on ultrasonic assisted ten-
sion and compression behaviour of aluminium alloy
1050. All simulations were done using implicit mechanical
analysis without temperature effects. A phenomenological
approach was presented to simulate the deformation
behaviour of aluminium alloy 1050 by reducing the friction
forces when ultrasonic vibration is present. The friction
force was reduced by assigning a very small friction coeffi-
cient. It was concluded that the effect of ultrasonic vibra-
tion on bulk properties of metals cannot be explained in
terms of stress superposition and surface effects. It was
also proposed that a clear understanding can be developed
by studying, how is the ultrasonic energy absorbed by alu-
minium microstructure. Doumanidis and Gao (2004) and
Gao and Doumanidis (2002) performed the mechanical
analysis of an ultrasonic spot welding process of a metal
foil on a substrate. The mechanical analysis was based on
the definition of frictional boundary conditions at the
foil/substrate interface. The friction boundary condition
was defined by using the experimentally measured strain
on the substrate surface and adjacent to the ultrasonic
probe. A good agreement of time dependent strain was
found at the foil/substrate.
Although, a number of researchers have performed
studies on ultrasonic metal working and have reported
both surface and volume effects in ultrasonic processing
of materials. But the mechanisms describing these effects
are not fully explained in quantitative manner for the case
of ultrasonic welding. Also, a very few attempts have been
made to simulate the ultrasonic welding process
(Doumanidis and Gao, 2004; Gao and Doumanidis, 2002).
In most of the theoretical and simulated works, the effect
of ultrasonic vibration is attributed in the friction
coefficient rather than taking into account both surface
and volume effects.
In the present paper, a material model based on cyclic
plasticity theory has been proposed to take into account
the volume effects while a kinematic friction model has
been proposed to include the contribution of surface ef-
fects during the ultrasonic welding. The proposed material
and friction model are discussed in detail in the following.
2. Deformation behaviour of aluminium alloy 6061
Aluminium alloy (AA) 6060/6061 is one of the most
commonly used aluminium alloy due to its versatile
984 A. Siddiq, E. Ghassemieh / Mechanics of Materials 40 (2008) 982–1000
thermal and mechanical properties. The typical composi-
tion of aluminium alloys 6060/6061 in weight percent is:
magnesium (0.8–1.2), silicon (0.4–0.8), iron (0.18–0.7),
copper (0.15–0.4), zinc (0.2–0.25), titanium (0.001–0.15),
manganese (0.1–0.15), chromium (0.04–0.35) and balance
aluminium. Due to its good toughness, surface finish, wel-
dability and easy workability, AA 6060/6061 has found
many applications in the industry, such as aerospace com-
ponents, marine fittings, automotives, and other electrical
and electronics equipments. One of the widely used appli-
cations of AA 6060/6061 is in metal–matrix composites. In
all such applications, material has to bear all kinds of
mechanical and thermal loads. In this section a thermome-
chanically motivated cyclic plasticity model has been pro-
posed to model the deformation behaviour of AA 6060/
6061, which takes into account Bauschinger effect, plastic
ratchetting and shake down along with thermal softening
with increasing temperature.
Experimental investigations to study the deformation
behaviour of AA 6060/6061 have been performed in the
past (Swearengen, 1972; Sih and Chou, 1989; Yang and
Wang, 1993; Davis, 1993; Hopperstad et al.,1995a,b; Srivat-
san et al., 1997; Lee et al., 2000). Swearengen (1972) has
performed experimental studies on the deformation
behaviour of AA 6061. Swearengen (1972) showed that
thermally activated flow stress in AA 6061 arises from
the simultaneous contributions of two parallel processes,
which are, dislocation intersection and shear of the glide
planes. It was also found that at high strain rates the room
temperature behaviour may begin to incorporate one or
both of the thermally activated processes. Sih and Chou
(1989) studied the uniaxial tension behaviour of AA 6061
at elevated temperature. They found out the non-equilib-
rium thermal/mechanical response of AA 6061 at elevated
temperatures. It was shown that at a specific surrounding
temperature, during the uniaxial tension loading, temper-
ature of the specimen decreases locally due to expansion
and then after some duration of time the temperature be-
comes the same as surrounding with reversal in heat flux.
Yang and Wang (1993) studied the cyclic fracture behav-
iour of AA 6061 over and extensive range of positive mean
with dk being the plastic multiplier which satisfy the fol-
stress. It was shown that secondary-stage creep rate and
lowing consistency conditions (Kuhn–Tucker type)
fracture process are slowed down by decreasing mean
stress values for low cycle stress amplitude.
Hopperstad et al. (1995a,b) studied the deformation
behaviour of AA 6060 at room temperature. The main focus
of the work was to investigate the factors, such as the shape
of the hysteresis loops, cyclic hardening and softening along
with the memory of prior strain histories, during deforma-
tion. It was found that AA 6060 exhibited cyclic hardening
with stress range increasing in an exponential manner
during the first cycles and eventually saturates to a stabi-
lized stress amplitude range after a certain number of
cycles. Fig. 1 shows 1st cycle and 40th cycle during defor-
mation of AA 6060 at room temperature (Hopperstad
et al., 1995a,b). Srivatsan et al. (1997) studied the effect of
temperature on the cyclic response and fracture character-
istics of AA 6061. It was found that AA 6061 shows softening
at all test temperatures. Lee et al. (2000) studied the dy-
namic response of the AA 6061 material at high strain rate
impact loading. It was found that the impact response of
AA 6061, strongly depends upon applied strain rate, which
results in the variation of work hardening rate and strain
rate sensitivity. It was also shown that rate of work harden-
ing decreases with decreasing strain and strain rates.
For the presented work, the data presented in Davis
(1993) about the monotonic deformation behaviour of AA
6061 is used along with the experimental uniaxial cyclic
stress–strain curves of Hopperstad et al. (1995a,b).
3. Material model
3.1. Isothermal cyclic plasticity model
The thermomechanical cyclic plasticity model proposed
in this work is based on the combined nonlinear isotropic/
kinematic hardening model for time independent cyclic
plasticity presented by Chaboche and coworkers (Chaboche,
1977, 1979, 1986, 1989; Chaboche and Rousselier, 1981,
1983; Lemaitre and Chaboche, 1990; Chun et al., 2002;
Freed and Chaboche, 1989).
The basic constitutive equations of the model for uniax-
ial loading case are summarized below:
The total strain tensor during deformation is the sum of
elastic strain tensor and plastic strain tensor, given by
e ¼ eel þ epl ð1Þ
Stresses are computed using the elastic stress strain
relation
r ¼ Ceel ¼ Cðe  epl Þ ð2Þ
Yield function (criterion) is given by
F ¼ jr  aj  ðr0 þ RÞ ¼ 0 ð3Þ
with a being the back stress tensor due to kinematic hard-
ening, R is the isotropic hardening term and r0 being the
initial yield stress.
Plastic strain during deformation is given by
oF
depl ¼ dk ð4Þ
or
F 6 0; dk P 0 dk  F ffi 0 ð5Þ
The nonlinear isotropic hardening rule adopted here was
presented in Lemaitre and Chaboche (1990), Huber and
Tsakmakis (1999), and Chun et al. (2002). The isotropic
hardening (R) which describes the expansion of the yield
surface is defined as an exponential function of accumu-
lated plastic strain (i.e. isotropic hardening is dependent
on the plastic strain history), which is given by
 
pl
R ¼ Q 1  ebe ð6Þ
where epl is the equivalent plastic strain, while Q and b are
material parameters to be identified by inverse modelling.
Q is the maximum change in the size of the yield surface,
and b is the rate at which the size of the yield surface
changes with changing plastic strains.
A nonlinear kinematic hardening proposed by
Armstrong and Frederick (1966) has been used to capture
 A. Siddiq, E. Ghassemieh / Mechanics of Materials 40 (2008) 982–1000 985

250

1st cycle (Hopperstad et al., 1995a, b)

200

150

100

50
Stress (MPa)

0
-1.6 -1.2 -0.8 -0.4 0 0.4 0.8 1.2 1.6
-50

-100

-150

-200 40th cycle (Hopperstad et al., 1995a, b)

-250
Strain (%)

Fig. 1. Experimental cyclic stress–strain curves of AA 6060 for 1st and 40th cycle (Hopperstad et al., 1995a,b).

nonlinear hardening behaviour and smooth transition from
elastic to plastic deformation (Chaboche, 1977, 1979, 1986,
1989; Chaboche and Rousselier, 1981, 1983; Ohno and
Wang, 1993a,b; Lemaitre and Chaboche, 1990; Jiang
and Kurath, 1996; Wang and Barkley, 1998, 1999; Abdel
and Ohno, 2000; Geng and Wagoner, 2000, 2002). The
evolution of back stress (a) is given by
1
a_ ¼ C ðr  aÞe_ pl  cae_ pl
r0
where r0 = ry + R with ry being the yield stress at zero
plastic strain, while C and c are the material parameters
which can be identified from cyclic testing. The c term
determines the rate at which the saturation value of kine-
matic hardening decreases with increasing plastic defor-
mation. C is the kinematic shift of the yield surface.
The calibration of C and c is performed using the stress
strain data of stabilized cycle. The evolution law of back
stress in Eq. (7) when integrated for a uniaxial case is given
by (ABAQUS, 2006)
C pl
 pl
a¼ 1  ece þ a1  ece
c
where a1 is obtained from stabilized cycle and is given by
s Rth ¼ Q 1  ebe  ð1  ^hm Þ
a1 ¼ r1  r
with r1 is the stress at the start of the stabilized cycle and
rs is the stabilized size of the yield surface, which is given
by
ðr1 þ rn Þ
rs ¼
2 htransition is the transition temperature, at or below which
with r1 and rn are the stress at the start and end of the sta-
bilized cycle.
3.2. Cyclic plasticity model with thermal softening
Many metals that exhibit aging processes have shown
strong temperature path history dependence (Swearengen,
1972; Sih and Chou, 1989; Srivatsan et al., 1997). The ther-
momechanical plasticity theories have been under study
since many years. Many researchers have proposed differ-
ent models to describe inelastic behaviour for metals un-
ð7Þ der different thermal and mechanical loading conditions.
The thermomechanical coupling term used in the present
study is adopted from Johnson and Cook (1985). A similar
approach was proposed by McDowell (1992), in which the
temperature terms were only included in the description
of isotropic and kinematic hardening rules. This procedure
implicitly takes into account any nonisothermal response
as temperature path history-dependent. In the present
study, the thermomechanical term is adopted from
Johnson and Cook (1985) hardening model. The inclusion
of such thermomechanical coupling term is quite straight
forward. This was done by embedding the temperature
term (1  hm) in the nonlinear isotropic and kinematic
ð8Þ hardening model. The modified nonlinear isotropic hard-
ening law is given by
 
pl
ð10Þ
where m is the material parameter and ^
h is the nondimen-
sional temperature given as
^h ¼ h  htransition ð11Þ
hmelt  htransition
ð9Þ
there is no temperature dependence on yield stress, and
hmelt is the melting temperature.
986 A. Siddiq, E. Ghassemieh / Mechanics of Materials 40 (2008) 982–1000
Similarly, the modified nonlinear kinematic hardening
law is given by
 
C pl pl
ath ¼ ð1  ece Þ þ a1  ece ð1  hm Þ
c
3.3. Cyclic plasticity model with thermal and acoustic
(ultrasonic) softening
Acoustic (ultrasonic) softening is defined as the de-
crease in the plastic (yield) limit of a material under in-
tense ultrasonic vibration. Acoustic softening has been
under investigation since many decades (Dawson et al.,
1970; Green, 1975; Hansson and Tholen, 1978; Langenecker,
1966; Mordyuk, 1975; Severdenko et al., 1973; Winsper
and Sansome, 1968).
Langenecker (1966) performed the studies on the effect
of ultrasonic vibration on the deformation behaviour of
aluminium and zinc metals. It was found that yield stress
of metals reduces significantly when intense ultrasonic
stress waves are applied during deformation. It was also
found that amount of reduction of yield stress is directly
proportional to the acoustic energy input to the specimen.
It was also found that as long as ultrasound energy applied
is below a critical value, typical for the different materials,
no permanent changes appear in the properties of metals.
These results also revealed that the ultrasonic energy den-
sity required to produce deformation in aluminium was
roughly 107 times less than the thermal energy.
Langenecker (1966) proposed that this difference in ener-
gies is due to the fact that the absorption of ultrasonic en-
ergy is localized, i.e. in the vacancies, dislocations and
grain boundaries while thermal energy is absorbed uni-
formly in the material. Winsper and Sansome (1968) per-
formed studies on the deformation behaviour of different
materials, such as mild steel, stainless steel, hard copper,
and hard and soft aluminium, in the presence of ultrasonic
vibration. It was found that the reduction in stress can be
explained by a mechanism of superposition of static and
oscillatory stresses. It was also reported that ultrasonic en-
ergy in the elastic region does not affect the stress/strain
curves of the material. Dawson et al. (1970) presented an
overview of the effect of ultrasonic vibration in the defor-
mation of different metals. Green (1975) proposed a theo-
retical model which takes the behaviour of the material as
nonlinear elastic. Using this nonlinear elastic behaviour, a
reduction in Young’s modulus has been calculated.
Hansson and Tholen (1978) studied the deformation
behaviour of commercial aluminium in the presence of
ultrasonic energy. It was proposed that the local heating
due to the absorption of ultrasonic energy at internal de-
fects (e.g. dislocations) dissipates fast due to the relatively
high thermal conductivity of aluminium giving only a
slight overall increase in the specimen temperatures (al-
most negligible). It was also proposed that the acoustic
softening explanations presented by previous researchers
do not take into account the effect of static stress in lower-
ing the necessary ultrasonic energy transferred to the
material for dislocation generation, interaction and mobil-
ity. Green (1975), Mordyuk (1975), and Severdenko et al.
(1973) carried out experiments on many metals (zinc, cad-
mium, aluminium, copper, steel, and tungsten). It was
found that a decrease in plastic (yield) limit is proportional
to the intensity of oscillations and independent of their fre-
ð12Þ quency in the range of 15–80 kHz. Mordyuk (1975) also
found that the value of this decrease in yield limit does
not depend on the degree of preliminary deformation in
the rang of residual elongations below 16% and on the tem-
perature within the range 30–500 °C. Gilman (2001) pro-
posed a model to describe the contraction of an extended
dislocation. It was proposed that a critical speed exists at
which an extended dislocation will have a higher total en-
ergy than a unit dislocation minus the energy of its stack-
ing fault. At this critical velocity, the extended dislocations
in the material will contract into unit dislocations. As the
extended dislocations have contracted, screw dislocations
will cross-glide freely without the aid of the thermal acti-
vation. It was also shown that critical velocity for contrac-
tion is about one-third the velocity of shear waves in the
material. Rusynko (2001) proposed a generalized theory
of plasticity for the case of plastic deformation of metals
in the presence of ultrasonic vibrations.
In the present work a simple phenomenological ap-
proach has been used to include the acoustic (ultrasonic)
softening during deformation of the material. The main
assumption for this approach comes from the above dis-
cussion (Green, 1975; Langenecker, 1966; Mordyuk,
1975; Severdenko et al., 1973). The main assumptions
are the decrease in plastic (yield) limit is proportional to
the ultrasonic intensity. Also, the decrease in plastic (yield)
limit is independent of the frequency (in the range of
15–80 kHz), temperature (in the range of 30–500 °C) and
preliminary deformation.
A phenomenological softening term dependent upon
the ultrasonic energy density per unit time has been intro-
duced in the relations of isotropic and kinematic hardening
terms (Eqs. (10) and (12)). The modified equations of iso-
tropic and kinematic hardening are given by
Rultrasonic ¼ Rth  ð1  d  Eultrasonic Þ2 ð13Þ
2
aultrasonic ¼ ath  ð1  d  Eultrasonic Þ ð14Þ
where Rth is the isotropic hardening rule with thermal soft-
ening, defined in Eq. (10), d is ultrasonic softening param-
eter which has to be identified from experiments of
deformation behaviour of the material in the presence of
ultrasonic energy. Eultrasonic is the ultrasonic energy density
per unit time transferred from the sonotrode to the mate-
rial. ath is the kinematic back stress term with thermal
softening, defined in Eq. (12).
The above discussed hardening models, Eqs. (10)–(14),
are implemented in ABAQUS user subroutine UHARD
(user-defined hardening subroutine). This user subroutine
is called at all material integration points of elements for
which the material definition includes user-defined isotro-
pic or cyclic hardening for metal plasticity (ABAQUS, 2006).
This user subroutine can be used to define a material’s iso-
tropic yield behaviour, the size of the yield surface in a
combined hardening model or material behaviour depen-
dent on field or state variables. The implementation is very
straight forward. The UHARD user subroutine requires
 A. Siddiq, E. Ghassemieh / Mechanics of Materials 40 (2008) 982–1000
three quantities to be defined, i.e. Hardening model (Eqs.
(10)–(14)), rate of change of hardening rule with respect
to the accumulated plastic strain epl , which is given by
dRultrasonic pl
¼ Q  b  ebe ð1  ^hm Þ  ð1  d  Eultrasonic Þ2 ð15Þ
depl
 
daultrasonic pl pl
¼ Cece  a1  c  ece ð1  hm Þ  ð1  d  Eultrasonic Þ2
depl
ð16Þ
And rate of change of hardening rule with respect to the
temperature, which is given by
h It should also be noted that term phenomenological has
dRultrasonic pl
¼  Q  m  ð1  ebe Þ
dh^
#
ð^hm1 Þ
  ð1  d  Eultrasonic Þ2 ð17Þ
ðhmelt  htransition Þ
"  account activation of slip systems or dislocations due to
daultrasonic C pl pl
¼ ð1  ece Þ þ a1  ece
dh^ c
#
ð^hm1 Þ
  ð1  d  Eultrasonic Þ2 ð18Þ
ðhmelt  htransition Þ
The above discussed nonlinear isotropic/kinematic harden-
ing law requires the identification of six parameters
(Q, b, C, c, m, d) from cyclic stress–strain data, thermal soft-
ening data and acoustic (ultrasonic) softening data.
It should be noted that the presented plasticity model
has been used in the framework of finite deformation the-
ory based on spatial configuration (ABAQUS, 2006). Selec-
tion of the infinitesimal strain or finite strain
formulation, which requires the selection of appropriate
work conjugate stress and strain measures (Green and
Naghdi, 1965; Hill, 1970; Simo and Miehe, 1992), is based
on the keyword NLGEOM available in ABAQUS (2006).
Detailed discussion about the formulation is available in
ABAQUS (2006), which is not repeated here for brevity.
Also, the theory presented in this work is a particular
type of thermomechanical Mises plasticity theory with
analytical forms of the hardening law. Thermodynamical
consistency of such theories for infinitesimal and finite
deformation cases has been discussed in detail in Simo
and Miehe (1992), Abed and Voyiadjis (2007), Ristinmaa
and Vecchi (1996), and Chaboche (1989). It should be
noted that current plasticity theory has an additional inde-
pendent state variable (ultrasonic energy, Eultrasonic). This
means that the Helmholtz free energy function is a func-
tion of elastic strain ðeel
ij Þ, internal state variables (harden-
ing rule bn; n = 1,2, . . . n), temperature (h), gradient of
temperature (rih) and additionally ultrasonic energy
(Eultrasonic). The modified Helmholtz free energy function
is given by
w ¼ wðeel
ij ; bn ; h; ri h; Eultrasonic Þ
This potential when substituted in thermodynamic
inequality obtained from the conservation of energy and
the Clausius–Duhem inequality (refer to Chaboche, 1989)
ends up with an additional term which involves partial
derivative of the Helmholtz free energy function with re-
spect to the ultrasonic energy (Eultrasonic) times the time
987
derivative of ultrasonic energy (Eultrasonic). This additional
term is given by oE ow  E_ ultrasonic . Similar to e_ el _
ij and h
ultrasonic
(Chaboche, 1989), Ėultrasonic is also an observable variable
and for the validity of the inequality oE ow ¼ 0. Therefore,
ultrasonic
the rest of the dissipation inequality preserves the form of
Chaboche (1989).
Using the presented approach in Chaboche (1989) and
Simo and Miehe (1992), one can obtain a similar free en-
ergy function which is a function of thermal potential,
themomechanical coupling potential, elastic potential and
hardening potential (Simo and Miehe, 1992).
been used in the sense that hardening and softening rules
are defined as analytical functions and does not take into
account basic physical mechanisms involved which could
be the outlook of this research area. For example, using
micromechanics based material models which take into
thermal and acoustic energies. These models could be
based on crystal plasticity framework (Siddiq et al., 2007)
or discrete dislocation models (Hartmaier et al., 2005).
3.4. Material parameter identification
The material model discussed above with nonlinear iso-
tropic and kinematic hardening rules along with thermo-
mechanical coupling and ultrasonic softening, involves
six material parameters (Q, b, C, c, m, d). The role of these
material parameters has been discussed in the previous
section. The identification of these material parameters is
performed using inverse modelling method. In this method
experimental stress–strain curves are compared with the
simulated stress–strain curves for each set of parameters
and the difference between the experimental and simu-
lated stress–strain curves is minimized by the variation
of these parameters. The process is repeated until a good
agreement is achieved between experimental and simu-
lated stress–strain curves. To identify large number of
parameters, such as for the case of crystal plasticity theory,
when number of unknown parameters ranges between 7
and 30, an automatic identification procedure has been
proposed in Siddiq and Schmauder (2005) and Siddiq
(2006). In the present study there were only six parame-
ters involved, therefore, the identification of these param-
eters has been performed manually.
The thermal and mechanical properties, taken from
Davis (1993), Kong (2005), are given in Table 1.
The identified set of isothermal material parameters
(Q, b, C, c) are given in Table 2 and the comparison of the
experimental and simulated stress–strain curves is shown
in Fig. 2.
Table 1
Thermal and mechanical properties of AA 6060/6061
Thermal properties Elastic properties
Thermal conductivity 235 W/m K Young’s modulus 66.24 GPa
Thermal expansion 23.4e6/°C Poisson’s ratio 0.33
coefficient
Specific heat 896 J/kg K Yield stress 50 MPa
Density 2700 kg/m3
988 A. Siddiq, E. Ghassemieh / Mechanics of Materials 40 (2008) 982–1000

Table 2 60
Identified nonlinear isotropic/kinematic hardening parameters

Initial Yielding (MPa)

50
Q (MPa) b C (GPa) c
100 20.0 15.0 60.0 40

30

The experimental fit for the temperature dependent ini- 20

tial yielding has been performed using the experimental 10 Experimental (J.G. Kaufman,1999)
data obtained from Kaufman (1999) and Davis (1993) for Simulated
aluminium alloy 6061. The best fit to the experimental 0
0 50 100 150 200 250 300 350
temperature dependent initial yielding is shown in Fig. 3, Temperature (deg. C)
with m = 1.642081.
The ultrasonic softening parameter (d) has been identi- Fig. 3. Comparison between experimental and simulated initial yielding
fied by comparing the experimental results of ultrasonic as a function of temperature.
softening (Langenecker, 1966) of aluminium. The value of
ultrasonic softening parameter (d) is found to be
60
1.3  106 m2/W. The experimental and simulated initial Experimental (Langenecker, 1966)
yield stress as a function of ultrasonic energy density per 50 Simulated

Yield Stress (MPa)

unit time is plotted in Fig. 4. The results in figure show a
good agreement. 40
The final set of six material parameters for aluminium
alloy 6061 is given in Table 3. 30

20
4. Friction model
10
Friction and wear plays an important role during the
sliding of two surfaces under cyclic loading, such as high 0
0.0E+00 1.0E+05 2.0E+05 3.0E+05 4.0E+05 5.0E+05 6.0E+05
cycle fatigue in turbine blades and ultrasonic processes, Ultrasonic Energy (W/m2)
etc. In this work a friction law has been proposed which
comprises of static and kinematic friction components. Fig. 4. Effect of ultrasonic energy per unit time on initial yielding.
A time dependent friction coefficient was proposed in
Gao and Doumanidis (2002), ultrasonic welding process
was analyzed using finite element methods. It was as- of the unbonded foil–substrate interface surface. A
sumed that friction coefficient is uniform at all locations time dependent friction coefficient was defined using the

250

200

150

100

50
Stress (MPa)

0
-1.6 -1.2 -0.8 -0.4 0 0.4 0.8 1.2 1.6

-50

-100

-150
Initial Cycle (Experiment)
Initial Cycle (Simulation)
-200 40th Cycle (Experiment)
40th Cycle (Simulation)
-250
Strain (%)

Fig. 2. Comparison between experimental and simulated stress–strain curves of initial and 40th cycle.
 A. Siddiq, E. Ghassemieh / Mechanics of Materials 40 (2008) 982–1000 989

Table 3 4.1. Description of the friction model

Final set of parameters for aluminium alloy 6061

Q (MPa) b C (GPa) c m d The friction model proposed in this work is based on the
100 20.0 15.0 60.0 1.642081 1.3  106 m2/W
experimentally measured strain on the substrate surface
and adjacent to the ultrasonic probe. It was shown that
friction coefficient increases almost linearly until a specific
time and then remains the same and then starts decreasing
as time progresses. Fretting fatigue behaviour of AA 6061
was studied experimentally in Naidu and Raman (2005).
The effect of different parameters, such as contact pres-
sure, stress amplitude and number of cycles, on coefficient
of friction was studied. It was found that coefficient of fric-
tion decreases with the increasing contact pressure. It was
also found that, as the number of cycles increases, the coef-
ficient of friction also increase and after a certain number
of cycles it saturates to a steady value. Effect of ultrasonic
on upsetting of a model paste was analyzed in Huang et al.
(2002). A constant coefficient of friction was used based on
Coulombs friction model. As soon as the ultrasonic vibra-
tion is applied on the specimen, the friction coefficient
was reduced to simulate the experimental behaviour. The
limitations of a constant coefficient of friction in Coulomb’s
friction model have been studied in Naboulsi and Nicholas
(2003). A non-classical Coulomb’s friction model was
developed which allow the coefficient of friction to be a
function of local contact pressure and local slip magnitude.
It was shown that with different combinations of the pro-
posed model parameters provide better comparison with
experimental stress states. Effect sliding velocity and tem-
perature on coefficient of friction of aluminium alloy was
studied in Zhang et al. (2006). It was found that velocity
has no effect on the friction coefficient while friction coef-
ficient increases as the temperature increases until 150 °C
and then starts decreasing. A phenomenological model of
variable friction coefficient, for the 2D cases, in fretting fa-
tigue process has been proposed in Cheikh et al. (2006).
The proposed model is based on kinematic and isotropic
hardening of the coefficient of friction as a function of
accumulated slip between the contacting surfaces. The
friction model consisted of three parts static friction coef-
ficient, isotropic friction behaviour and kinematic friction
behaviour. In other studies, such as (Cheikh et al., 2006;
Dick and Cailletaud, 2006; Petiot et al., 1995; Stromberg,
1999), coefficient of friction has always been assumed to
dependence of coefficient of friction l, on number of cycles
N, temperature T, and parameters a and b which depend
upon magnitude of slip and contact pressure.
The friction model proposed in this work is based on the
friction experiments performed in Cheikh et al. (2006), Gao
and Doumanidis (2002), Huang et al. (2002), Naboulsi and
Nicholas (2003), Naidu and Raman (2005), Zhang et al.
(2006). As discussed before, Naidu and Raman (2005)
found that coefficient of friction increases with increasing
number of cycles and after a certain number of cycles, coef-
ficient of friction saturates to a steady value (Fig. 5).
A simple logarithmic correlation has been used to de-
fine the experimental friction behaviour. The correlation
is given by
liso ¼ ls þ ls  ða  logðNÞ þ bÞ ð19Þ
where a and b are friction parameters depend upon the
magnitude of the slip amplitude and contact pressure,
while ls is the initial static coefficient of friction, and N is
number of cycles. The values of a and b are determined
from the experimental results of AA 6061 presented in
Naidu and Raman (2005). For the case of contact pressure
of 50 MPa and stress amplitude of 193 MPa (equivalent
to a slip magnitude of 8.4 mm for the case of AA 6061)
are found to be a = 0.323 and b = 0.1. The comparison of
the model with the experimental result for this case is
shown in Fig. 6.
The values of a and b are identified for other contact
pressure and displacement amplitudes. The values identi-
fied are given in Table 4. In order to implement the values
of a and b as a function of contact pressure and displace-
ment amplitude, a correlation has been deduced. The cor-
relation is of the form:
a ¼ a1  a2  P ð20Þ
and
b ¼ b1  b2  P ð21Þ
The values of a1, a2, b1, and b2 are identified from the
already given values of a and b for different contact
1.4
Contact Pressure = 50 MPa
Coefficient of Friction

1.2
be constant. On the other hand, as discussed above, exper-
imental studies performed in Cheikh et al. (2006), Gao and 1
Doumanidis (2002), Huang et al. (2002), Naboulsi and
Nicholas (2003), Naidu and Raman (2005), and Zhang et 0.8

al. (2006), found that the coefficient of friction is not a con- Max. Cyc. Stress 169 MPa
0.6
stant. It was also found (Cheikh et al., 2006; Gao and Max. Cyc. Stress 193 MPa
Doumanidis, 2002; Huang et al., 2002; Naboulsi and 0.4 Max. Cyc. Stress 217 MPa
Nicholas, 2003; Naidu and Raman, 2005; Zhang et al., Max. Cyc. Stress 243 MPa
0.2 Max. Cyc. Stress 265 MPa
2006) that the coefficient of friction depends upon slip mag-
nitude, contact pressure, number of cycles and temperature. 0
10 210 410 610 810
In the following, a friction model based on Coulomb’s stick/
Number of Cycles
slip formulation has been proposed by taking into account
physical quantities, such as, slip magnitude, contact pres- Fig. 5. Coefficient of friction as a function of number of cycles (Naidu and
sure, number of cycles and temperature. Raman, 2005).
990 A. Siddiq, E. Ghassemieh / Mechanics of Materials 40 (2008) 982–1000

1.4 Table 5
Identified values of a1, a2, b1, and b2 for different displacement amplitudes
Coefficient of Friction
Experiment (Naiduand Raman, 2005 )
1.2
Displacement a1 a2 b1 b2
1 amplitude (lm)
8.4 0.362 0.000827 0.07303 0.000592
0.8 10.4 0.25 0.00033 0.1447 0.000313
Friction Model 12.4 0.2924 0.00054 0.18218 0.00015
0.6 14.4 0.2192 0.0000854 0.2157 0.000015

0.4 Contact Pressure = 50 MPa

0.2 Displacement amplitude = 8.4 μm
a = 0.323; b = -0.1 Table 6
0 Friction parameters for temperature dependent friction coefficient
10 110 210 310 410 510 610 710 810 910
Number of Cycles p q r S t
8.485e10 8.842e7 1.969e4 9.762e3 1.12
Fig. 6. Comparison between friction model and experiments (Naidu and
Raman, 2005) for a contact pressure = 50 MPa and stress
amplitude = 193 MPa.

pressures. The values are given in Table 5 for different dis-

placement amplitude values.
The temperature dependence of coefficient of friction is
introduced using the experimental observation in Zhang et
al. (2006). Zhang et al. (2006) studied the friction behav-
iour of aluminium allow at different temperatures. It was
found that coefficient of friction increases with the increas-
ing temperature until a specific temperature and then de-
creases. This temperature dependence has been taken into
account by including an additional fourth order polyno-
mial as a function of temperature. The modified coefficient
is given by

l ¼ liso  ðp  T 4 þ q  T 3 þ r  T 2 þ s  T þ tÞ
The additional friction parameters p, q, r, s, t are identified
using the experimental results (Zhang et al., 2006) for alu-
minium alloy and are given in Table 6 while experimental
and friction model results are plotted in Fig. 7.
4.2. Implementation of the friction model
The proposed friction model has been implemented in
ABAQUS user subroutine (FRIC) to define user friction
model. The implementation is based on Coulomb’s fric-
tion law with stick–slip algorithm (ABAQUS, 2001) with
jsfricj 6 l  P, i.e. if jsfric < l  P then stick and if jsfricj = l  P
ð22Þ
Fig. 7. Coefficient of friction as a function of temperature (experiments
and friction model).
then slip. The numerical implementation of simple
Coulomb’s friction law with constant friction coefficient
has been discussed in ABAQUS (2001). In this work, the
already available friction subroutine (ABAQUS, 2001)
has been modified by including Eqs. (19)–(22)). This
had been done by replacing the constant coefficient of
friction with a coefficient of friction that is dependent
upon temperature, contact pressure, amplitude of vibra-
tion and number of cycles.

Table 4
Values of a and b identified from experimental results (Naidu and Raman, 2005)

Contact pressure 25 MPa 30 MPa 50 MPa 100 MPa 125 MPa 155 MPa 175 MPa
Amplitude of vibration = 8.4 lm
a 0.328 0.327 0.323 0.3095 0.2787 0.2267 0.1945
b 0.098 0.0985 0.1 0.1098 0.1321 0.17 0.1933
Amplitude of vibration = 10.4 lm
a 0.2625 0.26319 0.26595 0.275 0.2825 0.2975 0.32
b 0.1575 0.1576 0.1582 0.1675 0.1775 0.191 0.21
Amplitude of vibration = 12.4 lm
a 0.2725 0.2712 0.2659 0.25116 0.2388 0.2091 0.1824
b 0.187 0.188 0.192 0.1920 0.196 0.2054 0.214
Amplitude of vibration = 14.4 lm
a 0.212 0.2114 0.2089 0.2035 0.199 0.188 0.1782
b 0.2162 0.2163 0.2164 0.217 0.2172 0.2176 0.219
 A. Siddiq, E. Ghassemieh / Mechanics of Materials 40 (2008) 982–1000
5. Ultrasonic welding model and boundary conditions
Typical ultrasonic metal welding setup is shown in
Fig. 8. The welding setup consists of three main compo-
nents, a foil, a substrate and a sonotrode (attached to the
ultrasonic welding unit). Substrate is fixed to an anvil at
the bottom surface, a foil is placed at the top surface of
the substrate and with the help of the sonotrode load is ap-
plied in the vertical direction (as shown in Fig. 8). The
sonotrode vibrates at a frequency of 20 kHz in the direction
perpendicular to the rolling (welding) direction.
The process parameters during ultrasonic welding are
applied load (Papplied), velocity of sonotrode (V), frequency
of the ultrasonic vibration (f) and amplitude of vibration
which are normally varied in order to check the effect of
each of these parameters on weld quality.
The geometric parameters for the present study are
width of the specimen (w) 20 mm, thickness of the sub-
strate (ts) 1 mm, thickness of the foil (tf) 100 lm. The ra-
dius of the sonotrode used in the present study is 25 mm
(Kong, 2005). In the present work, effect of the ultrasonic
welding process parameters (applied load, velocity of
sonotrode, and amplitude of ultrasonic vibration) is stud-
ied. Frequency of the ultrasonic vibration (f = 20 kHz) is al-
ways kept constant during the present study. The applied
load is varied from 25 to 175 MPa. The amplitude of vibra-
tion is varied from 8.4 to 14.4 lm while velocity of sono-
trode is varied from 27.8 to 38.8 mm/s.
Finite element analyses of the ultrasonic welding spec-
imen have been performed using coupled temperature–
displacement analysis. The finite element model has been
constructed for a small region of approximately 2 mm, as
shown in Fig. 9. The selected width of the model (2 mm)
has been chosen based on the width of the contact area
during the specific weld cycle (which ranges from 0.133
to 0.324 mm depending upon applied loading). This size
of the model is chosen so that the distance of the weld area
is far enough from the boundary. The geometry of the
model from the side view is shown in Fig. 10. As discussed
above, foil has the dimensions of 2 mm  20 mm  0.1 mm
P a p p lie d
f
V
S o n o tr o d e
ts
il
Fo
te
s tra
tf
S ub
w
Fig. 8. Ultrasonic metal welding setup.
991
while substrate is of dimensions 2 mm  20 mm  1 mm
(20 mm is in the direction orthogonal to the side view
shown in Fig. 10). In reality, sonotrode is a solid cylindrical
shaped bar (Kong, 2005), however to save computation
time the geometry of the sonotrode during this work has
been modelled as hollow with an inner diameter of
50 mm and outer diameter of 50.4 mm (Fig. 10). The
choice of thickness of the hollow sonotrode walls
(0.2 mm) is based on heat conducted through sonotrode.
The thickness chosen in this work ensures that the heat
generated at the foil/sonotrode interface does not reach
the inner surface of the hollow sonotrode modelled in this
work. Process parameters used during the simulations, i.e.
velocity of sonotrode, applied load and vibration ampli-
tudes, are applied through a reference point associated
with the sonotrode as shown in Fig. 10. Normal load is ap-
plied in the vertical direction; rolling velocity of the sono-
trode is applied using the velocity boundary condition in
the rolling direction (as shown in Fig. 10) while ultrasonic
vibration is applied orthogonal to the side view in Fig. 10
using displacement boundary conditions with a periodic
amplitude curve (A = Amaxsinxt). Where A is the ampli-
tude of the vibration dependent on time, Amax is the max-
imum amplitude of the vibration, x is the angular velocity
and t is the total time. The values of the process parame-
ters used are already given in the previous paragraph. The
bottom surface of the substrate is fixed in all directions as
was during the real experiments (Kong, 2005). Initial tem-
perature of 20 °C has been prescribed to the complete
geometry. For the current analyses all the free edges
(other than the contact surfaces and left and right edges
in Fig. 10) in the model are under free convection with
convection coefficient of 30 W/m K and ambient tempera-
ture of 20 °C. The left and right edge boundaries are as-
sumed to be at 20 °C, by prescribing the temperature
boundary condition. These temperature boundary condi-
tions have been used, based on the experimental findings
of Cheng and Li (2007), which showed that for different
weld times the temperature away from the weld regions
decrease as one moves away from the weld area. Also,
heat flux has been tracked during all the simulations
and it has also been found that heat flux never reached
these boundaries as the weld time is very small, i.e. 4–
9 ms, which also makes the assumption of temperature
(20 °C) valid for left and right edges. The bottom surface
of substrate and top surface of the sonotrode are also kept
at 20 °C. Throughout this work foil and substrate have
been modelled using material properties of aluminium al-
loy 6061 (see Section 3.4) while sonotrode is modelled as
steel (elastic) with properties given in Table 7.
Finite element mesh consisted of 31,842 coupled tem-
perature-displacement elements (C3D8RT). Thermome-
chanical interaction properties between foil/sonotrode
interface and foil/substrate interface are discussed in the
following. The normal contact properties between foil/
sonotrode interface and foil/substrate interface are defined
using hard contact formulation available in ABAQUS
(2006). In this contact formulation contact constraint is en-
forced with a Lagrange multiplier representing the contact
pressure in a mixed formulation (ABAQUS, 2006). In this
case
992 A. Siddiq, E. Ghassemieh / Mechanics of Materials 40 (2008) 982–1000

Papplied

f
V
Path, parallel to the direction of vibration Sonotrode

ts
2 mm
2 mm

il
Fo

e
at
bstr
tf
Su

1 mm w

Fig. 9. Finite element model of ultrasonic welding specimen.

Applied Load
All dimensions are in millimetres
Al
R25
Rolling
Velocity of
sonotrode
Reference Point
Associated with
Sonotrode
Foil
Foil Sonotrode

0.2
R25 0.1

Substrate 1

*Ultrasonic vibration is orthogonal to this plane

Fig. 10. Geometry of ultrasonic welding specimen (side view).

p ¼ 0 for h < 0 ðcontact is openÞ; The friction properties between foil and substrate con-
p > 0 for h ¼ 0 ðcontact is closedÞ; tact surface is defined using the thermomechanical friction
model proposed in this work (Section 4). Friction proper-
where h is the contact closure. ties used are already discussed in Section 4. The friction
 A. Siddiq, E. Ghassemieh / Mechanics of Materials 40 (2008) 982–1000 993

Table 7 0.035
Thermal and mechanical properties of sonotrode (steel) 8.4 micron
10.4 micron
Thermal properties Elastic properties 0.03 12.4 micron
14.4 micron

Friction work (mJ/mm2)

Thermal conductivity 80 W/m K Young’s modulus 200 GPa
0.025
Thermal expansion 11e6/°C Poisson’s ratio 0.27
coefficient
Specific heat 440 J/kg K 0.02
Density 7800 kg/m3
0.015

coefficient for sonotrode/foil interface (steel/aluminium) is
defined using a pressure dependent isothermal coefficient
friction in Coulomb’s friction model. The pressure depen-
dent coefficient of friction between sonotrode and foil is
plotted in Fig. 11. Through out this work it is assumed that
all of the friction energy generated between different con-
tacting surfaces is converted into heat energy. This has
been achieved by using ‘‘gap heat generation” option al-
ready available in ABAQUS (2006). With the help of this
option, user can define what fraction of the friction energy
is converted into heat. For the present study this fraction is
set to 1, i.e. all of the friction energy is converted into heat
energy. It is also assumed that the heat energy is equally
transferred to both the contacting surfaces in every inter-
face. Results of the thermomechanical ultrasonic welding
analyses have been discussed in the following.
6. Results and discussion
6.1. Effect of applied load and displacement amplitude on
friction work between foil/substrate interface
The friction work at the interface of foil and substrate is
plotted in Fig. 12. These values are obtained when friction
work saturates (approximately after 22–28 cycles for the
applied loading between 25 and 175 MPa). The results
plotted in Fig. 12 are for the case when velocity of sono-
trode was 34.5 mm/s. It is found that as the applied load
increase the friction work at the interface of foil and sub-
strate decrease. It has been found that this decrease in fric-
tion work has two main reasons, friction dissipation at the
foil/sonotrode interface (surface effect) and ultrasonic soft-
ening of the foil material (volume effect).
0.01
0.005
0
0 20 40 60 80 100 120 140 160 180 200
Applied Pressure (MPa)
Fig. 12. Friction work at foil/substrate interface, velocity = 34.5 mm/s.
Both of these factors are found to decrease the transfer
of ultrasonic vibration due to softening of the foil material.
The softening comprises of two parts, i.e. thermal softening
(surface effect) which is due to the friction dissipation at
the interface and ultrasonic softening which is due to the
ultrasonic energy being transferred to the material. The
first factor, i.e. the friction dissipation at foil/sonotrode
interface causes higher plastic dissipation at the foil/sono-
trode interface as the applied load increases. The increase
in plastic dissipation in foil at foil/sonotrode interface is
due to the higher thermal softening, i.e. increasing temper-
ature because of increasing heat generation. The equiva-
lent plastic strain at a point in the middle of the foil
surface which is in contact with the sonotrode is plotted
in Fig. 13. The results plotted in Fig. 13 are for the case
when velocity of sonotrode was 34.5 mm/s and amplitude
of ultrasonic vibration was 8.4 lm. The plot shows that as
the applied loading increases the amount of plastic strain
increases which is due to the increase in friction dissipa-
tion at the foil/sonotrode interface. The second factor, i.e.
ultrasonic softening of foil material (volume effect) also re-
sults in higher plastic deformation as the loading is in-
creased. It is found that ultrasonic energy transferred to

80

70
Equivalent Plastic Strain

0.9
Coefficient of Friction

0.85 60

0.8 50
0.75
40
0.7
30
0.65
20
0.6
0.55 10
0.5 0
0 20 40 60 80 100 120 140 160 180 200 0 20 40 60 80 100 120 140 160 180 200
Contact Pressure (MPa) Applied Pressure (MPa)

Fig. 11. Coefficient of friction (measured) between sonotrode (steel) and Fig. 13. Equivalent plastic strain in foil at foil/sonotrode interface
foil (aluminium alloy). (amplitude = 8.4 rmum; velocity = 34.5 mm/s).
994 A. Siddiq, E. Ghassemieh / Mechanics of Materials 40 (2008) 982–1000

the foil increases with the increasing load, i.e. making it
softer. The plot of the ultrasonic energy per second trans-
ferred to the foil material at different points in the foil
along the thickness direction is plotted in Fig. 14. It can
also be seen that ultrasonic energy transferred in the foil
along the thickness direction decreases as one moves from
sonotrode surface towards substrate surface.
It can also be seen in Fig. 12 that as the amplitude of
ultrasonic vibration increase, the friction work at the inter-
face of foil and substrate decrease. This shows that the
increasing amplitude of ultrasonic vibration has the similar
effect as that of the increasing load. The reasons are the
ultrasonic softening (volume effects) and thermal soften-
ing due to friction (surface effects). As the amplitude of
ultrasonic vibration increase the ultrasonic energy trans-
ferred per second also increase resulting into higher soft-
ening. Ultrasonic energy per second transferred to the
foil at foil/sonotrode interface is plotted in Fig. 15. The plot
in Fig. 15 shows that as the amplitude of ultrasonic vibra-
tion increases the amount of ultrasonic energy transferred
also increases.
Also, as the amplitude of ultrasonic vibration increase
friction dissipation increase causing higher equivalent
plastic strains at foil/sonotrode interface.
6.2. Effect of sonotrode velocity on friction work between
foil/substrate interface
Friction work as a function of velocity of sonotrode for
different applied loads and constant vibration amplitude
(8.4 lm) has been plotted in Fig. 16. It can be inferred from
the plot that as the velocity of the sonotrode increase, the
friction work also increases. This can be explained on the
basis of higher loading rate of the sonotrode which allows
less number of cycles to reach up to the specific applied
load when compared with smaller velocities of sonotrode,
resulting in less thermal softening due to friction dissipa-
tion at foil/sonotrode interface. This lesser thermal soften-
ing at the foil/sonotrode interface increase the ultrasonic
energy transferred to the foil/substrate interface resulting
in higher friction dissipation.
Similar trend of friction work has been found for differ-
ent amplitudes of the ultrasonic vibration. Fig. 17 shows
the plot of friction work between foil/substrate interface
as a function of velocity for different displacement ampli-
tudes and constant applied load (155 MPa). It can be seen
that as the amplitude of the ultrasonic vibration increases
the friction work also increase. The reason is the same as

0.06
50000 25 MPa Vibration amplitude = 8.4 micron
Friction Work (mJ/mm2)

100 MPa
45000 0.05 125 MPa
Ultrasonic Power (W/m2)

25 MPa 155 MPa

40000 40 MPa 175 MPa
125 MPa 0.04
35000 155 MPa
175 MPa
30000 0.03
25000
0.02
20000
15000
0.01
10000
5000 0
25 30 35 40 45
0 Velocity (mm/sec)
0 0.02 0.04 0.06 0.08 0.1 0.12
Distance along the thickness direction of the Foil (mm)
Fig. 16. Friction work between foil/substrate interface as a function of
Fig. 14. Ultrasonic energy per second transferred to the foil along the velocity of sonotrode (vibration amplitude = 8.4 lm).
thickness direction.

0.016
Friction work (mJ/mm2)

80000 8.4 micron

0.014 10.4 micron
Ultrasonic Power (W/m2)

70000 12.4 micron

0.012 14.4 micron
60000
0.01
50000
0.008
40000
0.006
30000
0.004
20000
0.002
10000 Applied load = 155 MPa
0
0 25 30 35 40 45
8 9 10 11 12 13 14 15
Amplitude of ultrasonic vibration (μm) Velocity of Sonotrode (mm/sec)

Fig. 15. Ultrasonic energy per second transferred to the foil at foil/ Fig. 17. Friction work between foil/substrate interface as a function of
sonotrode interface (velocity = 34.5 mm/s; applied load = 25 MPa). velocity of sonotrode (applied load = 155 MPa).
 A. Siddiq, E. Ghassemieh / Mechanics of Materials 40 (2008) 982–1000 995

explained above, i.e. higher the velocity, higher is the load-
ing rate which allow less number of cycles to reach up to
the specific applied load, resulting in less thermal soften-
ing due to friction dissipation at foil/sonotrode interface.
6.3. Plastic deformation
As discussed in the previous section, plastic deforma-
tion in the specimen is caused by two different phenom-
ena, namely surface and volume effects. It was also
shown in the previous section that as the applied load in-
creases the plastic deformation in the foil at foil/sonotrode
interface increases (Fig. 18). This is due to the increase in
friction dissipation (surface effect) at foil/sonotrode inter-
face along with the increasing ultrasonic softening (volume
effect) due to ultrasonic energy transferred from sonotrode
to the foil. Fig. 19 shows the plot of equivalent plastic
strain in the foil at foil/substrate interface. It is found that
as the applied load increases the equivalent plastic strain
also increases until a loading value which is enough to
cause plastic deformation in the material without the
application of ultrasonic vibration and after this loading
equivalent plastic strains saturate.
Similar behaviour has been shown by the substrate, as
shown in Fig. 19. The plot in Fig. 19 shows that as the
amount of applied load increases the plastic deformation
also increases until a loading value which causes plastic
deformation before the ultrasonic vibration has been
started.
Comparison of Figs. 13, 18, and 19 shows that the re-
gions of the foil near foil/sonotrode interface undergo se-
vere plastic deformation. This high plastic deformation is
due to high friction dissipation and ultrasonic energy
transferred to the foil near foil/sonotrode interface (Figs.
14 and 15). It can also be inferred from Figs. 18 and 19 that
amount of plastic deformation in foil near the foil/sub-
strate interface is almost twice as high as the plastic defor-
mation in the substrate. As explained before the reason for
having larger plastic deformation in foil is due to the dual
effect, i.e. surface (friction dissipation at foil/substrate
interface) and volume (ultrasonic softening) effects. On
the other hand the substrate has only the surface effects,
i.e. friction dissipation at the foil/substrate interface, to
cause plastic deformation.
6.4. Temperature at the interfaces

0.07 As discussed in Section 1, temperature of the specimen

due to the friction dissipation at the interface is found to be
0.06
Equivalent Plastic Strain

only 15–55% of the melting temperature (650 °C) of AA

0.05 6061 for various applied loads. Figs. 20–22 show the plots
of the temperature profile at foil and substrate surfaces at
0.04 foil/sonotrode and foil/substrate interface. The tempera-
ture profile is plotted along the path which is parallel to
0.03 the direction of sonotrode vibration and orthogonal to
the sonotrode rolling direction (see Fig. 9).
0.02
The plot in Fig. 20 shows that as the amount of applied
load increases the temperature of the foil surface at foil/
0.01
sonotrode also increases and this causes larger thermal
0 softening as discussed in the previous sections. It should
0 20 40 60 80 100 120 140 160 180 200 also be noted that temperatures are well below the meting
Applied Pressure (MPa) temperature of AA 6061, i.e. 650 °C.
Fig. 21 shows the plot of temperature profile at foil sur-
Fig. 18. Equivalent plastic strain in the foil at foil/substrate interface
(amplitude = 8.4 rmum; velocity = 34.5 mm/s). face at foil/substrate interface along the path parallel to the
direction of the vibration of sonotrode. The plot shows the
similar trend as discussed above, i.e. the increase of tem-
0.04 perature with increasing applied load. The temperature is
again found to be well below the melting temperature of
0.035 AA 6061.
Equivalent PlasticStrain

0.03 The plot of temperature profile at substrate surface at

foil/substrate interface is shown in Fig. 22. The tempera-
0.025 ture is again found to be increasing with increasing applied
load. Also, the temperature is found to be 10–37% of the
0.02
melting temperature of the AA 6061. The results are in
0.015 agreement with the experimental results of Cheng and Li
(2007), it was found that temperature for the case of cop-
0.01 per and nickel were found to be in the range of 100–250 °C,
i.e. copper and nickel never reach to the melting tempera-
0.005
ture during ultrasonic welding.
0 One of the most important conclusion which can be in-
0 50 100 150 200 ferred from Figs. 20–22 is, there exists a temperature gra-
Applied Load (MPa)
dient from sonotrode to the foil and then to the substrate.
Fig. 19. Equivalent plastic strain in the substrate at foil/substrate This can be identified from Figs. 20–22 that highest tem-
interface (amplitude = 8.4 rmum; velocity = 34.5 mm/s). perature is at foil/sonotrode interface then it starts
996 A. Siddiq, E. Ghassemieh / Mechanics of Materials 40 (2008) 982–1000

400

175 MPa
350
155 MPa
300 125 MPa
100 MPa
Temperature (deg. C)
250

200

150

25 MPa
100

50

0
0 2 4 6 8 10 12 14 16 18 20
Distance along the path in the direction of sonotrode vibration (mm)

Fig. 20. Temperature of foil surface at foil/sonotrode interface (amplitude = 8.4 rmum; velocity = 34.5 mm/s).

350

175 MPa
300

155 MPa
250
125 MPa
Temperature (deg. C)

100 MPa
200

150

100 25 MPa

50

0
0 2 4 6 8 10 12 14 16 18 20
Distance along the path in the direction of sonotrode vibration (mm)

Fig. 21. Temperature of foil surface at foil/substrate interface (amplitude = 8.4 rmum; velocity = 34.5 mm/s).

decreasing as the one move towards the foil/substrate
interface. This important conclusion justifies the argument
given in many experimentalists, that friction only plays a
secondary role during ultrasonic welding, i.e. it is only re-
quired to break and disperse the surface impurities and
oxide layers. The main bonding mechanism is mechanical
interlocking of atoms when two surfaces are in contact at
atomic scale, i.e. the clearance between the two surfaces
is in atomic distances. This small atomic distance is then
enough for the surface atoms to form a bond.
 A. Siddiq, E. Ghassemieh / Mechanics of Materials 40 (2008) 982–1000 997

300

250 175 MPa

155 MPa

Temperature (deg.C)
200 125 MPa

100 MPa

150

100

25 MPa

50

0
0 2 4 6 8 10 12 14 16 18 20
Distance along the path in the direction of sonotrode vibration (mm)

Fig. 22. Temperature of substrate surface at foil/substrate interface (amplitude = 8.4 rmum; velocity = 34.5 mm/s).

6.5. Friction stress at the foil/substrate interface studied were aluminium, brass, copper, and stainless
steel. They found that the as the amplitude of the ultra-
Kumar and Hutchings (2004) performed experimental sonic vibration increases the friction stresses decrease.
studies on the friction behaviour of metals under longi- The reduction in friction observed was in the range of
tudinal and transverse ultrasonic vibration. The materials 60–80%. Fig. 23 shows the plot of friction stress at foil/

180

160
Friction Stress atFoil/Substrate Interface (MPa)

140
175 MPa

120

100

80

60

40
25 MPa
20

0
8 9 10 11 12 13 14 15
Amplitude of Ultrasonic Vibration (μm)

Fig. 23. Friction stress at foil/substrate interface (velocity = 34.5 mm/s).

998 A. Siddiq, E. Ghassemieh / Mechanics of Materials 40 (2008) 982–1000
substrate interface as a function of amplitude of ultra-
sonic vibration of sonotrode. The results are plotted for
two different applied loads, i.e. 25 and 175 MPa. It can
be inferred from the plot in Fig. 23 that as the amplitude
of ultrasonic vibration increases the amount of friction
stress decreases and the decrease for these two load
cases is found to be 48% and 57.5%, respectively. The
reduction of friction stresses under ultrasonic vibration
show similar behaviour as was observed during the
experiments (Kumar and Hutchings, 2004). As explained
in the previous sections, this reduction in friction stres-
ses is due to the ultrasonic and thermal softening of
the material.
6.6. Comparison with experimental results
In order to compare the simulated results with the
experiments, peel tests are performed on the AA 6061
specimen welded at an applied load of 155 MPa and a
velocity of sonotrode equals to 34.5 mm/s. Using the peel
test force–displacement curves, fracture energies have
been computed using the procedure described in Korn et
al. (2002), Siddiq and Schmauder (2005, 2006), and Siddiq
et al. (2007).
Friction work between foil/substrate interface and
fracture energies computed from peel test curves are
plotted in Fig. 24 as a function of amplitude of ultrasonic
vibration for an applied load of 155 MPa. It can be in-
ferred from Fig. 24 that as the amplitude of ultrasonic
vibration increase the experimental fracture energy ob-
tained from peel tests decrease. Similar trend can be seen
in Fig. 24 for friction work at the foil/substrate interface,
i.e. as the amplitude of ultrasonic vibration increase fric-
tion work at the foil/substrate interface decrease. The rea-
son for the decrease in experimental fracture energy can
0.0085
Experimental Fracture Energy
0.008
0.0075
Friction work (mJ/mm2)
0.007
Simulated Friction work
0.0065
0.006
0.0055
0.005
8.4
Amplitude of ultrasonic vibration (m)
Fig. 24. Friction work and fracture energy as a function of amplitude of ultrasonic vibration (velocity = 34.5 mm/s; load = 155 MPa).
be explained on the basis of decreasing friction work. As
explained in the previous sections, the friction energy
serves a secondary purpose for welding, i.e. it is required
to break and disperse the oxide layers and other impuri-
ties to bring the two contacting surfaces closer to each
other (approx. atomistic distances). Therefore, as the fric-
tion work at the foil/substrate interface decreases the dis-
persion of the oxide layer and other impurities also
reduce, causing less number of bonds at the Foil/substrate
interface. This lower density of bonds ultimately results
into a lower fracture energy as exhibited by the welded
specimen of AA 6061 specimen during experiments (Fig.
24).
Similar trends can be seen for other applied loadings,
see Fig. 25. It is found that as the applied loading or ampli-
tude of ultrasonic vibration increase the experimental frac-
ture energy decrease, similar trends have been shown by
the friction work at the weld interface during thermome-
chanical analyses. The reason for this decrease in friction
work has been explained before.
7. Conclusion
A thermomechanical analyses of ultrasonic welding of
aluminium alloy has been presented in this work. Both vol-
ume and surface effects have been incorporated in the pro-
posed cyclic plasticity model with thermal and acoustic
softening along with a friction law with variable coefficient
of friction depending upon the contact pressure, amount of
slip, temperature and number of cycles. Results show that:
– Friction plays a secondary role during ultrasonic weld-
ing. Results suggest that, in real experiments, friction
work only breaks up the oxide layer at the weld inter-
face and disperse it along and near the interface.
1400
1300
Fracture Energy (J/m2)
1200
1100
1000
900
800
10.4 12.4 14.4
 A. Siddiq, E. Ghassemieh / Mechanics of Materials 40 (2008) 982–1000
1700
135 MPa
1600 155 MPa
175 MPa
Fracture Energy (J/m )
2
1500
1400
1300
1200
1100
1000
900
800
8 9 10 11 12
Amplitude of ultrasonic vibration (ηm)
Fig. 25. Experimental fracture energy (left) and simulated friction work (right) as a function of amplitude of ultrasonic vibration (velocity = 34.5 mm/s).
– The maximum temperature reached during the ultra-
sonic welding is well below melting temperatures of
the joining materials which is in agreement with the
experimental results.
– Our three-dimensional simulations showed that mate-
rial immediately next to the sonotrode receives the
ultrasonic energy which introduces the pre-softening
of the material before the sonotrode has moved on it.
This pre-softening of material keeps on moving ahead
of the sonotrode in the rolling direction of sonotrode.
This could be one of the reasons why it has been
observed during the ultrasonic consolidation of fibre
embedding experiments that material immediately
flows around the fibre as soon as the load is applied.
– It is also found that friction stresses at the weld inter-
face reduce due to thermal and acoustic softening
which is in conjunction with the experimental results.
– Comparison with the experimental results shows a
good agreement between simulated and experimental
results.
– The microstructural and finite element analyses at sub-
micron level is necessary to further understand the
deformation phenomena involved during ultrasonic
welding process which is planned in near future using
crystal plasticity theory with ultrasonic softening
effects.
Acknowledgements
The authors thankfully acknowledge the financial sup-
port of EPSRC (Engineering and Physical Sciences Research
Council) and MOD (Ministry of Defence through the grant
(GR/T19988) and the collaborative support of the Solidica
Ltd.
References
ABAQUS, 2006. ABAQUS, version 6.5, Hibbit and Karlsson (online
documentation).
ABAQUS, H., Karlsson & Sorensen, Inc., 2001. Coulomb Friction in User
Subroutine FRIC. Rhode Island.
Abdel, K.M., Ohno, N., 2000. Kinematic hardening model suitable for
ratcheting with steady-state. Int. J. Plasticity 16, 225–240.
999
0.014
135 MPa
155 MPa
175 MPa
0.012
Friction work (mJ/mm )
2
0.01
0.008
0.006
0.004
0.002
13 8 9 10 11 12 13
Amplitude of ultrasonic vibration (ηm)
Abed, F.H., Voyiadjis, G.Z., 2007. Thermodynamic consistent formulations
of viscoplastic deformations in FCC metals. J. Eng. Mech. ASCE 133,
76–86.
Armstrong, P.J., Frederick, C.O., 1966. A Mathematical Representation of
the Multiaxial Bauschinger Effect. CEGB Report No. RD/B/N 731.
Chaboche, J.L., 1977. Viscoplastic constitutive equations for the
description of cyclic and anisotropic behaviour of metals. Acad.
Polo. Sci., Sev. Sc. Et. Techn. 25 (33).
Chaboche, J.L., 1979. Modelization of the Strain Memory Effect on the
Cyclic Hardening of 316 Stainless Steel, SMIRT-5, Division L, Berlin.
Chaboche, J.L., 1986. Time independent constitutive theories for cyclic
plasticity. Int. J. Plasticity 5, 247–302.
Chaboche, J.L., 1989. Constitutive equations for cyclic plasticity and cyclic
viscoplasticity. Int. J. Plasticity 5, 247–302.
Chaboche, J.L., Rousselier, G., 1981. On the plastic and viscoplastic
constitutive equations based on the internal variables concept. In:
SMIRT-6 Post. Conf., Paris, T.P. ONERA, pp. 8–11.
Chaboche, J.L., Rousselier, G., 1983. On the plastic and viscoplastic
constitutive equations. J. Press. Vess. Technol. 105, 153–164.
Cheikh, M., Quilici, S., Cailletaud, G., 2006. Presentation of KI-COF, a
phenomenological model of variable friction in fretting contact. Wear
262, 914–924.
Cheng, X., Li, X., 2007. Investigation of heat generation in ultrasonic metal
welding using micro sensor arrays. J. Micromech. Microeng. 17 (2),
273–282.
Chun, B.K., Kim, H.Y., Lee, J.K., 2002. Modelling the Bauschingers effect for
sheet metals. Part II: Applications. Int. J. Plasticity 18, 597–616.
Daud, Y., Lucas, M., Huang, Z., 2006. Superimposed ultrasonic oscillations
in compression tests of aluminium. Ultrasonics 44, 511–515.
Daud, Y., Lucas, M., Huang, Z., 2007. Modelling the effects of
superimposed ultrasonic vibrations on tension and compression
tests of aluminium. J. Mater. Process. Technol. 186 (1–3), 179–190.
Davis, J.R., 1993. ASM Specialty Handbook: Aluminium and Aluminium
Alloys. ASM International, Ohio.
Dawson, G.R., Winsper, C.E., Sansome, D.H., 1970. Application of high- and
low-frequency oscillations to the plastic deformation of metals. Metal
Form. 9, 254–261.
Dick, T., Cailletaud, G., 2006. Fretting modelling with a crystal plasticity
model of Ti6Al4V. Comput. Mater. Sci. 38, 113–125.
Doumanidis, C., Gao, Y., 2004. Mechanical modelling of ultrasonic
welding. Weld. J. 4, 140–146.
Dushkes, S.Z., 1973. A design study of ultrasonic bonding tip. IBM J. Res.
Dev., 230–235.
Freed, A.D., Chaboche, J.L., 1989. Viscoplasticity: A Thermodynamic
Formulation. NASA Technical Memorandum 102388.
Gao, Y., Doumanidis, C., 2002. Mechanical analysis of ultrasonic bonding
for rapid prototyping. J. Manuf. Sci. Technol. 124, 426–434.
Geng, L., Wagoner, R.H., 2000. Springback Analysis with a Modified
Nonlinear Hardening Model. SAE 2000-01-0410.
Geng, L., Wagoner, R.H., 2002. The effects of plastic anisotropy and
hardening on sheet springback. Int. J. Mech. Sci. 44, 123–148.
Gilman, J.J., 2001. Contraction of extended dislocations at high speeds.
Mater. Sci. Eng. A, 84–86.
Green, R.E., 1975. Non-linear effects of high-power ultrasonics in
crystalline solids. Ultrasonics 13, 117–127.
1000 A. Siddiq, E. Ghassemieh / Mechanics of Materials 40 (2008) 982–1000
Green, A.E., Naghdi, P.M., 1965. A general theory of an elastic plastic
continuum. Arch. Rat. Mech. Anal. 18, 251–281.
Gunduz, I.E., Ando, T., Shattuck, E., Wong, P.Y., Doumanidis, C.C., 2005.
Enhanced diffusion and phase transformations during ultrasonic
welding of zinc and aluminum. Scripta Mater. 52 (9), 939–943.
Hansson, I., Tholen, A., 1978. Plasticity due to superimposed macrosonic
and static strains. Ultrasonics 16, 57–64.
Harman, G., Albers, J., 1977. The ultrasonic welding mechanism as applied
to aluminium and gold wire bonding in microelectronics. IEEE Trans.
PHP 13, 406–412.
Harman, G.G., 1997. Wire Bonding in Microelectronics – Materials,
Process, Reliability and Yield. McGraw-Hill, New York.
Hartmaier, A., Buehler, M.J., Gao, H., 2005. Two-dimensional discrete
dislocation models of deformation in polycrystalline thin metal films
on substrates. Mater. Sci. Eng. A 400, 260–263.
Hill, R., 1970. Constitutive inequalities for isotropic elastic solids under
finite strain. Proc. R. Soc. Lond. A314, 457.
Hopperstad, O.S., Langseth, M., Remseth, S., 1995a. Cyclic stress–strain
behaviour of alloy AA6060. Part I: Uniaxial experiments and
modelling. Int. J. Plasticity 11, 725–739.
Hopperstad, O.S., Langseth, M., Remseth, S., 1995b. Cyclic stress–strain
behaviour of alloy AA6060. Part II: Biaxial experiments and
modelling. Int. J. Plasticity 11, 741–762.
Hu, S.J., Lim, G.E., Foong, K.P., 1991. Study of temperature parameter on
the thermosonic gold wire bonding of high-speed CMOS. IEEE Trans.
Compon. Hybr. Manuf. Technol. 14, 855–858.
Huang, Z., Lucas, M., Adams, M.J., 2002. Influence of ultrasonics on
upsetting of a model paste. Ultrasonics 40, 43–48.
Huber, N., Tsakmakis, C., 1999. Determination of constitutive properties
from spherical indentation data using neural networks. Part II:
plasticity with nonlinear isotropic and kinematic hardening. J.
Mech. Phys. Solids 47, 1589–1607.
James, E.K., 1990. Transmission electron microscopy of ultrasonic
wire bonding. IEEE Trans. Compon. Hybr. Manuf. Technol. 13,
176–180.
Jiang, Y., Kurath, P., 1996. Characteristics of the Armstrong–Frederick type
plasticity models. Int. J. Plasticity 12, 387–415.
Johnson, G.R., Cook, W.H., 1985. Fracture characteristics of three metals
subjected to various strains, strain rates, temperatures and pressures.
Int. J. Eng. Fract. Mech. 21, 31–48.
Joshi, K.C., 1971. The formation of ultrasonic bonds between metals.
Weld. J. 50, 840–848.
Kaufman, J.G., 1999. Properties of Aluminum Alloys: Tensile, Creep, and
Fatigue Data at High and Low Temperatures, ASM, Metals Park, OH
44073-0002, USA.
Kazumasa, T., Mitsuo, U., Yoshihiro, T., 2003. Au bump interconnection
with ultrasonic flip–chip bonding in 20 micron pitch. Jpn. J. Appl.
Phys. 42, 2198–2203.
Kong, C.Y., 2005. Investigation of Ultrasonic Consolidation for Embedding
Active/Passive Fibres in Aluminium Matrices. Loughborough
University, Loughborough, UK.
Kong, C.Y., Soar, R.C., Dickens, P.M., 2003. Characterisation of aluminium
alloy 6061 for the ultrasonic consolidation process. Mater. Sci. Eng. A
360, 99–106.
Kong, C.Y., Soar, R.C., Dickens, P.M., 2004a. A model for weld strength in
ultrasonically consolidated components. Prof. IMechE Part C: J. Mech.
Eng. Sci., 83–91.
Kong, C.Y., Soar, R.C., Dickens, P.M., 2004b. Optimum process parameters
for ultrasonic consolidation of 3003 aluminium. J. Mater. Process.
Technol. 146, 181–187.
Korn, D., Elssner, G., Cannon, R.M., Ruhle, M., 2002. Fracture properties of
interfacially doped Nb/Alumina bicrystals. I: Fracture characteristics.
Acta Mater. 50 (15), 3881–3901.
Krzanowski, J.E., 1990. A transmission electron microscopy study of
ultrasonic wire bonding. IEEE Trans. CHMT 13, 176–181.
Kumar, V.C., Hutchings, V., 2004. Reduction of the sliding friction of
metals by the application of longitudinal or transverse ultrasonic
vibration. Tribol. Int. 37, 833–840.
Langenecker, B., 1966. Effects of ultrasounds on deformation
characteristics of metals. IEEE Trans. Son. Ultrason. SU-13, 1–8.
Lee, W., Shyu, J., Chiou, S., 2000. Effect of strain rate on impact response
and dislocation substructure of 6061-T6 aluminium alloy. Scripta
Mater. 42, 51–56.
Lemaitre, J., Chaboche, J.L., 1990. Mechanics of Solid Materials. Cambridge
University Press.
Mayer, M., Schwizer, J., 2002. Ultrasonic bonding: understanding how
process parameters determine the strength of Au–Al bonds. In: Intl.
Symposium on Microelectronics, IMAPS, Denver, CO, USA, 2002, pp.
626–631.
McDowell, D.L., 1992. A nonlinear kinematic hardening theory for cyclic
thermoplasticity and thermoviscoplasticity. Int. J. Plasticity 8, 695–
728.
Mordyuk, N.S., 1975. Influence of ultrasonic oscillations on the physical
properties of metals and alloys (in Russian). Naukova Dumka, Kiev.
Naboulsi, S., Nicholas, T., 2003. Limitations of the Coulomb friction
assumption in fretting fatigue analysis. Int. J. Solids Struct. 40, 6497–
6512.
Naidu, N.K.R., Raman, S.G.S., 2005. Effect of contact pressure on fretting
fatigue behaviour of Al–Mg–Si alloy AA-6061. Int. J. Fatigue 27, 283–291.
Ohno, N., Wang, J.D., 1993a. Kinematic hardening rules with critical state
of dynamic recovery. Part I: Formulation and features for ratcheting
behaviour. Int. J. Plasticity 9, 373–390.
Ohno, N., Wang, J.D., 1993b. Kinematic hardening rules with critical state
of dynamic recovery. Part II: Application to experiments of ratcheting
behaviour. Int. J. Plasticity 9, 391–403.
Petiot, C., Vincent, L., Van, K.D., Maouche, N., Foulquier, J., Journet, B.,
1995. An analysis of fretting-fatigue failure combined with numerical
calculations to predict crack nucleation. Wear 181, 101–111.
Prieb, V., 1999. Structural Aspects of Metal Joining, Stretching and
Stamping using Ultrasonics. 1st Memory Alloys GmbH.
Ristinmaa, M., Vecchi, M., 1996. Use of couple-stress theory in elasto-
plasticity. Comput. Meth. Appl. Mech. Eng. 136, 205–226.
Rusynko, A.K., 2001. Mathematical description of ultrasonic softening of
metals within the framework of the synthetic theory of plasticity.
Mater. Sci. 37 (4), 671–676.
Schwizer, J., Mayer, M., Bolliger, D., 1999. Thermosonic ball bonding:
friction model based on integrated microsensor measurements. In:
IEEE International Electronics Manufacturing Technology Symposium
(IEMT), Austin Texas, 1999, p. 108.
Severdenko, V.P., Klubovich, V.V., Stepanenko, A.V., 1973. Metal working
under pressure with ultrasound (in Russian). Nauka i Tekhnika,
Minsk.
Sheaffer, M., Levine, L., 1991. How to optimize and control the wire
bonding process. Solid State Technol., 67–70.
Siddiq, A., 2006. Modelling of Crystal Plasticity Effects in the Fracture of a
Metal/Ceramic Interface – Bridging the Length Scales. University of
Stuttgart, Germany, Stuttgart.
Siddiq, A., Schmauder, S., 2005. Simulation of hardening in high purity
niobium single crystals during deformation. Steel Grips: J. Steel Relat.
Mater. 3, 281–286.
Siddiq, A., Schmauder, S., 2006. Interface fracture analyses of a bicrystal
specimen using cohesive modelling approach. Modell. Simul. Mater.
Sci. Eng. 14 (6), 1015–1030.
Siddiq, A., Schmauder, S., Huang, Y., 2007. Fracture of bicrystal metal/
ceramic interfaces: a study via the mechanism-based strain gradient
crystal plasticity theory. Int. J. Plasticity 23 (4), 665–689.
Sih, G.C., Chou, D.M., 1989. Nonequilibrium thermal/mechanical response
of 6061 aluminium alloy at elevated temperature. J. Theor. Appl.
Fract. Mech. 12, 19–31.
Simo, J.C., Miehe, C., 1992. Associative coupled thermoplasticity at finite
strains: formulation, numerical analysis and implementation.
Comput. Meth. Appl. Mech. Eng. 98, 41–104.
Srivatsan, T.S., Sriram, M.K., Daniels, C., 1997. Influence of temperature on
cyclic stress response and fracture behaviour of aluminium alloy
6061. Eng. Fract. Mech. 56, 94–99.
Stromberg, N., 1999. A Newton method for three-dimensional fretting
problems. Int. J. Solids Struct. 36, 2075–2090.
Swearengen, J.C., 1972. The thermo-mechanical behaviour of 6061
aluminium–magnesium–silicon alloy. Mater. Sci. Eng. A 10, 103–117.
Tucker, J.C., 2002. Ultrasonic Welding of Copper to Laminate Circuit
Board. Worcester Polytechnic Institute.
Wang, H., Barkley, M.E., 1998. Strain space formulation of the Armstrong–
Frederick gamily of plasticity models. ASME J. Eng. Mater. Technol.
120, 230–235.
Wang, H., Barkley, M.E., 1999. A strain space nonlinear kinematic
hardening/softening plasticity models. Int. J. Plasticity 15, 755–777.
Winsper, C.E., Sansome, D.H., 1968. The influence of oscillatory energy on
the stresses during plastic deformation. J. Inst. Metals 96, 353–357.
Yang, A., Wang, Z., 1993. Cyclic deformation and fracture behaviour of
aluminium alloy 6061 under the action of positive mean stress.
Metall. Trans. A 24, 2083–2093.
Zhang, C.B., Zhu, X.J., Li, L.J., 2006. A study of friction behaviour in ultrasonic
welding (consolidation) of aluminium. In: AWS Conference: Session 7:
Friction and Resistance Welding/Materials Bonding Processes.