Version of Record: https://www.sciencedirect.

com/science/article/pii/S0924013615301977
Manuscript_0aa149a41c87f810e01e50d98fccc690

Vaporizing Foil Actuator Welding as a Competing Technology

to Magnetic Pulse Welding
Marlon Hahn1,a, Christian Weddelinga, Geoffrey Taberb, Anupam Vivekb, Glenn S.
Daehnb, A. Erman Tekkayaa
a
Institute of Forming Technology and Lightweight Construction, TU Dortmund University, Baroper Str. 303,
44227 Dortmund, Germany
b
Department of Materials Science and Engineering, The Ohio State University, 477 Watts Hall, 2041 College
Rd., Columbus, OH 43210, USA
Keywords:
Collision welding, Vaporizing Foil Actuator Welding (VFAW), Magnetic Pulse Welding (MPW),
Analytical model
ABSTRACT

Photonic Doppler velocimetry was applied to compare magnetic pulse welding and vaporizing foil
actuator welding against each other in the form of lap joints made of 5000 series aluminum alloy
sheets under identical experimental conditions which are: charging energies of the pulse generator,
specimen geometry, initial distances between flyer and target plate. Impact velocities resulting from
rapidly vaporizing aluminum foils were up to three times higher than those of purely
electromagnetically accelerated flyer plates. No magnetic pulse welds were achieved, while every
vaporizing foil experiment yielded a strong weld in that failure always occurred in the joining
partners instead of in the weld seam during tensile tests. An analytical model to calculate the
transient flyer velocity is presented and compared to the measurements. The average deviation
between model and experiment is about 11 % with regard to the impact velocity. Hence, the model
may be used for the process design of collision welds generated by vaporizing foil actuators.

1. Introduction

Multi-material designs for complex lightweight applications become more and more important in the
context of reducing fossil fuel consumption and subsequent exhaust emissions. Within the scope of
such designs various materials are deployed with respect to the mechanical loads they are subjected
to. Conventional joining techniques are not always capable of meeting the challenging requirements
of those multi-material designs because of different thermal properties (e.g. melting point) of the
joining partners, for example. Screw and rivet connections are usually relatively heavy and
expensive, while adhesive bonding implies prolonged production times due to the curing process. As
a consequence, a rising demand for alternative joining methods can be noticed. In the case of firmly
bonded metals, solid-state joining of similar as well as conventionally almost unweldable dissimilar

1
Corresponding author. Tel.: +49 231 755 8415. E-mail: marlon.hahn@iul.tu-dortmund.de

© 2015 published by Elsevier. This manuscript is made available under the Elsevier user license
https://www.elsevier.com/open-access/userlicense/1.0/
metals (e.g. steel and aluminum alloys) through high velocity forming, also referred to as collision
welding, is believed to have great potential. As emphasized by Zhang (2010), another advantage of
collision welding is the elimination or at least minimization of problems associated with a heat-
affected zone (HAZ), such as the formation of brittle intermetallic phases or cracking in fusion
welding. Consequently, the strength of collision welds can reach or even exceed the one of the
weakest parent material.

2. Collision welding methods

Known collision welding methods are (Zhang, 2010): explosive welding (EXW; possible workpiece
dimensions are in the order of meters), laser impact welding (LIW; dimensions of the order of
millimeters), and magnetic pulse welding (MPW; dimensions of the order of centimeters). A more
recent method introduced by Vivek et al. (2013) is called vaporizing foil actuator welding (VFAW;
same dimensions as in MPW). The two latter methods constitute the core of the present work and are
treated in more detail in the ensuing paragraphs. All mentioned joining technologies basically
underlie the same physical mechanisms, which are depicted in Fig. 1.

Figure 1: Principle of collision welding (Mori et al., 2013).

At least one of the joining partners - the flyer - is accelerated rapidly to an appropriate velocity vim at
which it collides with the target plate under a certain impact angle, β, resulting in impact pressures of
the order of several gigapascals (Mori et al., 2013). As summarized by Shribman (2008), this comes
along with the formation of a jet that removes all oxides and surface contaminants in the weld area so
that an atomic bonding can be achieved between the two mating metal surfaces. The weld seam then
propagates with collision velocity vc, which is geometrically related to vim through β
(Mousavi and Sartangi, 2009). In the course of this, a wavy interface morphology as depicted in
Fig. 1 may evolve. Regarding the actual mechanism causing such a wavy pattern, a few theories exist
being still under discussion in the literature. A common explanation has been given by Ben-
Artzy et al. (2010). After reviewing earlier theories, the authors experimentally established that
interface waves were formed in a Kelvin-Helmholtz instability in the case of tubular MPW joints.
Reflected shock waves were found to be the reason for the liquid-like behavior of the metals across
their interface. Vivek et al. (2013) mention that successful welds generally occur at angles between

2
5 ° and 20 ° as well as at impact velocities ranging from 150 m/s to 1500 m/s. A weldability domain
of the crucial parameters vc and β required for a specific combination of materials may be represented
by a so-called welding window. It also reveals whether or not waves and interlayers are observable in
the weld seam. Kore et al. (2009) concluded from scanning electron microscopy (SEM) and x-ray
diffraction (XRD) analysis of Mg AZ31 to Al AA3003 magnetic pulse welds with wavy interface
morphology that the base materials do not undergo a melting and solidification stage. In contrast,
Göbel et al. (2010) used similar techniques to prove that both intermetallic phases in “meld pockets”
and virtually waveless intermetallic transition layers emerge during magnetic pulse welding of Al
tubes and Cu cylinders. Welding windows originate from EXW and are derived in detail in
Mousavi and Sartangi (2009) for the explosive welding of cp Ti and AISI 304 stainless steel.
Verstraete et al. (2011) yet point out that impact angle and velocity are not constant during MPW,
which makes the generation of such welding windows rather difficult compared to EXW. The same
applies to VFAW. Vivek et al. (2014c), however, lately applied the concept of welding windows
successfully to cp Ti-Cu 110 VFAW joints. Their work combined grooved target plates with
predetermined angles, photonic Doppler velocimetry to record the transient flyer velocities, and SEM
analyses.

2.1 Fundamentals of electromagnetic forming (EMF)

Since MPW makes use of the electromagnetic forming technology in order to join the flyer and
target plate, the functional principle of EMF is briefly presented in the following. The forming tool
consists of an electrically insulated coil, more generally referred to as actuator, which is connected to
a capacitor bank pulse power supply (EMF machine) and at the same time placed close to the
workpiece or flyer (compare Fig.2). This entire system may be approximated by a simple series RLC
(resistance-inductance-capacitance) circuit with constant elements (Winkler, 1973). The capacitor is
then discharged usually within a few tens of microseconds to accelerate the flyer
electromagnetically.

Figure 2: Schematic of MPW for lap joints as pictured in Weddeling et al. (2014).

3
Beerwald (2005) enumerates that typical charging energies differ from 1 to 100 kJ, charging voltages
are in the range of 3 to 25 kV. This results in peak currents of up to a few 100 kA. When the driving
current pulse passes through the pressure lead of the coil, opposing eddy currents are induced in the
workpiece under the terms of the electromagnetic laws of Lenz and Faraday. Two nearby currents
flowing in the opposite direction (primary current in the coil and secondary current in the workpiece)
repel each other due to the Lorentz force which acts as the forming force. High conductivity flyers
may be accelerated directly in EMF, otherwise, thin high-conductivity materials can be used as a
driver plate, as for example investigated by Li et al. (2013) for the electromagnetic launch of 1 mm
thick Ti-6Al-4V plates. Since the Lorentz forces take effect on both the workpiece and the coil, one
of the biggest problems in EMF with regard to mass production still is the relatively short lifetime of
a coil (Psyk et al., 2011). EMF coil designs can generally be classified into three categories
introduced by Harvey and Brower (1958): compression coils (usually solenoids), expansion coils
(also usually solenoids) as well as coils for sheet metal forming. For the latter coil type, several
conductor geometries have been developed. A recent one established by Kamal (2005) is named
uniform pressure electromagnetic actuator (UPEA) and was modified for magnetic pulse welding by
Weddeling et al. (2014). As studied in Zhang et al. (2010), MPW coils for direct lap joints can be
designed as a simple one-turn coil consisting of a pressure lead and a wider return path outside the
forming area (see Fig. 2). Aizawa (2003) showed that the return path can also act as a second
pressure lead if it is small enough and positioned below the target plate so that both joining partners
are accelerated against each other. The coil geometry greatly influences the circuit parameters,
especially the coil inductance. As explained in Daehn (2010), low capacitances and low inductances
favor high frequencies which are essential for inducing intense eddy current densities and thus
sufficiently high Lorentz forces in the workpiece. An upper bound for the maximum possible
frequency is given by the short circuit frequency of the EMF machine. Typical values range from 20
to 100 kHz (Henselek et al., 2004).

2.2 Fundamentals of vaporizing foil actuators (VFA)

The effects of electrically driven rapidly vaporizing foils (or wires), also referred to as electrical
explosion of conductors, have been the subject of several studies over the past decades
(Chace and Levin, 1960). Exemplary works include the production of nano-sized powders
(Zou et al., 2012a) or the shaping of high current pulses (Bealing and Carpenter, 1972). Other
common applications deal with shock wave studies (Weingart et al., 1976). However, vaporizing foil
actuators have not been used for welding until the recent work of Vivek et al. (2013). In the
following, the basic physical mechanisms of vaporizing conductors are outlined. VFAW basically

4
utilizes the same machinery as EMF, but in this case the discharge current of the capacitor rapidly
vaporizes a thin foil in order to launch the metal flyer plate, as indicated in Fig. 3.

Figure 3: Schematic of VFAW for lap joints.

When a large pulsed current passes through a thin metallic conductor (e.g. foil), rapid non-linear
changes in physical states take place. As described in Wang et al. (2011), these changes can be
divided into basically five stages: the heating stage, the melting stage, the heating stage of liquid
metal before the vaporizing stage, and the plasma forming stage. Once the actuator is in the gaseous
state its resistance is a couple of orders higher than in the initial stage. Chace and Levin (1960) claim
that the heating time of a fast electrical vaporization is long enough for current propagation over the
cross-section of the conductor, but shorter than the time required for the evolution of
magnetohydrodynamic and capillary instabilities. Chau et al. (1980) state that the heating time is also
shorter than the thermal diffusion time away from the conductor. Consequently, energy exceeding
the theoretical heat of vaporization can be deposited in the actuator before it vaporizes. The instant of
vaporization or burst time tB - and therefore the beginning of violent expansion – is indicated by the
peak in voltage and is initiated by the heat-induced destruction of the dynamic equilibrium between
the expansive vapor pressure and the counteracting Lorentz force around the conductor. Due to this
equilibrium and the restraining surrounding insulation, the actuator can maintain its shape and keep
overheating until the vaporization starts (Zou et al., 2012b). From tB on, the intense pressure of the
expanding vapor or plasma acts as forming pressure for the flyer plate. As investigated in
Vivek et al. (2014a), an elastomer plate (e.g. polyurethane) can be put between the foil and the
workpiece to serve as an elastic transfer medium for the pressure pulse, which is then spread over an
area larger than the actual geometry of the foil. A main factor of influence on achievable pressures,
and therefore flyer velocities, is constituted by the amount of electrical energy that can be
deposited into the foil until it bursts. Referring to Vivek et al. (2014a), it can be computed as
tB
ED = ∫UI dt (1)
0

5
where U represents the voltage drop across the length and I the current through the foil.
Grigoriev and Pavlenko (2009) experimentally found a linearly increasing correlation between the
energy supplied to various foil materials and the resultant pressure. The authors were able to inject
up to twice the energy required for boiling the foil material, yielding pressure amplitudes in the range
of 2.5 GPa. Sarkisov et al. (2004) proved that the amount of energy deposited into a metal wire
depends on the rate of energy deposition or heating rate and, therefore, on the current rate. Increasing
the current rate from 20 to 150 A/ns allowed the deposition of almost three times more energy. As
resumed in Sedoi et al. (1999), this conflicts with some earlier studies where authors state that the
storable energy per unit mass is a constant for a given metal. Chemical reactions can also play a
noteworthy role in vaporizing foil technology. Nelson et al. (1999) found that the exothermic
reaction between aluminum vapor and water leads to significantly higher peak pressures as compared
to similar underwater vaporizations of tungsten wires at the same input energy. By contrast,
Lee and Ford (1988) revealed that the chemical reaction, which is 2Al + 3H2O = Al2O3 + 3H2, only
contributes sparsely on the shock front pressure because it occurs approximately 200 μs after the
vaporization of the aluminum wire. The reaction kinetics of Al with air were studied by Sarathi et al.
(2004). They found that there is a time lag between the vaporization and the beginning of intense
oxidation that is at least twice as long as the vaporization period. In case of VFAW, it can therefore
be assumed that the travel of the flyer does not last long enough to allow chemical reactions to
develop before colliding with the target plate. Since the flyer velocity is a crucial parameter in
VFAW, there is a need to determine it. A common approach to estimate a final flyer plate velocity
vmax is to apply the so-called electrical Gurney model introduced by Tucker and Stanton (1975),
which originates from studies about explosives. If the vaporizing foil is assumed to be backed by
infinite mass, it can be written as follows:

2 Eeg 2ηPbτ
vmax = = . (2)
(m c + 1 3) (m c + 1 3)

Here, m and c represent the accelerated flyer mass and the mass of the vaporizing foil, respectively.
The product ηPbτ is defined as electrical Gurney energy Eeg, Pb is the peak burst power per unit mass,
η is the Gurney efficiency, and τ is a factor having units of time. Tucker and Stanton (1975)
concluded that η can be estimated at 0.5. They also developed and verified the equation

= (3)

where the only variable Jb is burst current density; K and n are empirical constants for a given
experimental setup. Another model for predicting flyer velocities for VFA, which comprises energy
balance considerations formulated by Osher et al. (1989), is derived in Vivek et al. (2014b). To

6
verify it, the authors experimentally varied the thickness of aluminum foil as well as the input energy
and recorded velocities ranging from 400 m/s to 1000 m/s for 0.508 mm thick flyers made of
2024 aluminum alloy. Yet, the models mentioned above do not allow adequate transient calculations
of velocity because they only incorporate the final kinetic energy of the flyer. Knowing the velocity
history of the flyer, however, is important for the process design in collision welding since the
impact with the target happens at a predetermined distance (standoff). A model to compute velocity-
time curves for vaporizing foils is introduced in Schmidt et al. (1977), but the governing differential
equations require a numerical solver to specify values for the velocity. A more complex simulation
model which takes heat conduction, magnetic pressure, and electrical power into account on the basis
of magnetohydrodynamic (MHD) equations is presented and experimentally verified in
Wang et al. (2011).

VFAW seems to be a promising alternative to MPW due to the following reasons: Higher flyer
velocities can be expected, there are no tool life problems since foils are disposable low-cost
actuators, workpieces do not need to be good electrical conductors. For collision welded lap joints
both methods have not yet been directly compared with each other under identical conditions of
charging energy, capacitor bank configuration, specimen material, and specimen geometry.
Therefore, such a comparison is carried out experimentally within the scope of this work. Since no
closed-form solutions for calculating velocity-time histories for VFA can be found in literature so
far, a simple model is also suggested and verified here to provide the possibility of estimating flyer
plate impact velocities without the need for extensive velocity measurements. It is an analytical
model and mainly based on continuity and mechanical equilibrium considerations as well as on a
modified version of the electrical Gurney model. The required input data for this model only consist
of current-time and voltage-time curves, geometric data, and some basic material parameters.

3. Analytical model for VFA

3.1 Velocity field of the flyer

The flyer plate deforms plastically during the violent vapor expansion. Thus, a velocity field that
does not violate the yield condition must be found for the flyer. Due to the fact that the flyer is only
fixed on two sites (via the spacers), a one-dimensional stress state may me assumed so that a cross-
section of the flyer can be treated as a beam. A further simplification is to assume rigid-perfectly
plastic beams where the yield condition can be expressed through the bending moment. The forming
pressure generated during vaporization acts in the region where the foil is placed. The flyer can then
be considered as a fully clamped beam under partial uniform loading p in the middle section. If p,
which is of the order of gigapascals in VFAW, is several times larger than the static pressure

7
necessary for a plastic collapse of the beam, a kinematically and statically admissible velocity field
for such a system may be taken from Martins and Symonds (1965). It is depicted in Fig. 4.

Figure 4: Velocity field of the flyer during vaporization.

There, the magnitude of the bending moment inside all horizontal parts of the beam equals the fully
plastic moment for the cross-section. Plasticized parts are connected to rigid beam parts by so-called
plastic hinges. These hinges remain stationary in case of temporally rectangular pressure pulses
(Jones, 1989). However, it is assumed that a middle fraction of the flyer also remains approximately
horizontal in the course of a time-varying pressure pulse until hitting the nearby target. Referring to
Jones (1989) and Fig. 4, the dynamic behavior of the beam is governed by the equation

∂2M ∂vb
= − p + sρ b (4)
∂y 2 ∂t

where M is the moment per unit length about the z-axis. The density of the flyer ρb as well as its
thickness s are assumed to stay constant. Since the moment does not vary with y inside the horizontal
beam parts (∂M/∂y = 0 =∂2M/∂y2), the velocity vb only depends on time t, so that Eq. (4) can be
rewritten as

dvb
p = sρ b (5)
dt

for the regarded middle part of the flyer. Eq (5) is nothing but Newton’s second law. It is noted that
this simplified approach is not suitable for free forming experiments without a target where larger
distances need to be covered.

8
3.2 Minimum gas particle velocity

The model proposed here does not incorporate an electromagnetic flyer acceleration arising from the
Lorentz force. Consequently, the flyer plate is assumed to be at rest before the foil bursts, although
this might not be the case for electrically conductive flyer materials. Until shortly before the burst the
foil is interpreted as an enclosed gas volume with constant initial density ρ0, as illustrated in Fig. 5.

Figure 5: Foil and flyer section before the burst.

According to the kinetic theory of gases (Loeb, 1927), the mean gas particle velocity v0 leads to a
homogenous pressure pmin for which the following applies:

ρ
pmin = 0 v02 . (6)
3
Keeping in mind that electromagnetic effects are neglected, the flyer will only be accelerated
sufficiently at the burst time if the particle velocity is high enough to cause a plastification of the
flyer. To ensure that v0 is not chosen too low, the flyer segment over the foil width w is seen as a
fully clamped beam here. The resulting elastic bending moment Mel in the flyer can then be derived
from Young and Budynas (2002) as

ρ v 2 w2 ∆z
M el = 0 0 (7)
36

where pressure has already been eliminated by inserting Eq. (6). In order to make the flyer move
plastically, the elastic moment has to reach the fully plastic moment Mpl, which is expressed as

1
M pl = σ Y s 2 ∆z . (8)
4

This equation, where σY denotes the flow stress of a perfectly plastic flyer material, can be found in
elementary bending theory (Reissner et al., 1990). After equating Mel and Mpl, rearranging yields

9
 σ Y 3s
v0 = . (9)
ρ0 w

This required minimum gas particle velocity will be used as a boundary condition.

3.3 Maximum flyer velocity

For estimating the maximum velocity vmax of the flyer plate the electrical Gurney model is used
(compare Eq. (2)), but in the simplified form of

−1 2
ED  m 1 
vmax =  +  . (10)
c  c 3

Here, ED is the energy deposition according to Eq. (1). The flyer mass that initially covers the
narrowest foil section and that is situated between the spacers is chosen as m, since almost this entire
segment is being pushed onto the target during the welding process. The energy deposition is
presumed to be concentrated in the narrowest foil section (active foil region). This implies that c can
be approximated by the mass of this active foil region. For the values used in this work a relatively
high m/c ratio of about 32 is obtained. Certainly, the point when vmax actually occurs needs to be
estimated, since Eq. (10) is just based on a final energy balance and does not provide information
about the time course. Once the vapor expansion starts, energy will still be injected into the
vaporizing foil and contribute to the flyer acceleration. Therefore, a conversion from resistive power
of the foil to mechanical power of the flyer is considered (beginning from the burst time tB):

dvb
ζUI = ζRI 2 = m vb . (11)
dt

ζ > 0 denotes an unknown conversion efficiency function, R the non-linear resistance of the foil.
Experimentally obtained (smoothed) voltage-time, current-time, as well as velocity-time traces are
given as illustration in Fig. 6.

10
Figure 6: Experimental voltage, velocity, and current traces for VFAW.

It is reasonable to expect a monotonically increasing flyer velocity (acceleration dvb/dt ≥ 0) until the
global maximum vmax is reached. Thus, with regard to Eq. (11) and Fig. 6, vmax may occur when the
current becomes (approximately) zero, at t = tE. Afterwards, in the course of the restrike, the vapor is
assumed to be already expanded to a state where it can no longer accelerate the flyer. Hence, the
second boundary condition is determined as

vb (tE ) = vmax . (12)

Yet it is noted that another approach for estimating tE must be developed if the current does not
exhibit a distinct minimum as in Fig. 6. The current flow during the formation of the restrike plasma
allows the capacitor to be discharged completely. More detailed information about restrike
phenomena can, for instance, be found in Sinton (2011).

3.4 Temporal evolution of flyer velocity

The kinetic gas theory, as incorporated in Eq. (6), only applies for a static flyer until right before the
foil bursts. Once the vapor or plasma expansion has started (t ≥ tB), particle velocity v, pressure p,
and density ρ vary temporally as well as spatially. This can be taken into consideration by utilizing
the continuity equation for compressible one-dimensional gas flow. A combination with the adiabatic
equation of state allows eliminating the density so that

∂p ∂p ∂v
+ v = −γp (13)
∂t ∂x ∂x

11
is obtained. This equation may be found in Pécseli (2012). For chemical elements in the gaseous
state (here: Al) or monatomic gases γ = 5/3 applies (Loeb, 1927). The left-hand side of Eq. (13) is
consistent with the substantial derivative of pressure, dp/dt. The total differential of v can be
expressed as

d x v ∂v
dv = dt v + d xv with ≡ . (14)
dx ∂x

It is reasonable to suppose dxv/dx > 0 and dtv/dt > 0 for the acceleration stage. To make it
mathematically manageable, a simplification stating that the velocity increases with time t in the
same way as it does spatially with x is made, but only for the particles situated on the flyer surface, at
x = xb. There, the total differential may then be written as

dv = dt v + d xv ≅ 2d xv . (15)

Taking this into account and keeping v = dx/dt in mind, Eq. (13), after some manipulations, can be
rewritten as

γ 2
dv = − dp . (16)
v p

The integration of this equation eventually leads to the following explicit pressure-velocity relation:

γ 2
v 
p = p0  0  at x = xb . (17)
 v 

The index 0 represents reference values at time tB. Since the Gurney model mentioned earlier implies
that the particle velocity at the surface of the flyer exactly conforms with the flyer velocity, the same
assumption is made here as well (see Fig. 7).

12
Figure 7: Schematic of one-dimensional flyer acceleration through VFA.

Equating Eq. (5) and Eq. (17) then results in a non-linear differential equation for the flyer velocity:

= K1v −γ 2
dv
at x = xb (18)
dt

with

p0 v 0 γ 2
K1 = . (19)
sρ b

A closed-form analytical solution of Eq. (18) can be found by using the method of separation of
variables.

2
2+γ  2 +γ
v(t ) =  ⋅ (K1t + K 2 ) (20)
 2 

The constants K1 and K2 are determined by the boundary value problem v(tB) = v0 and v(tE) = vmax,
which eventually provides the following:

2vmax(2 +γ ) 2 − 2v0(2 +γ ) 2
K1 = (21)
(2 + γ )(tE − tB )
,

( 2 +γ ) 2
2v
K2 = 0 − K1t B , (22)
2+γ
and by association

13
 K1sρ b
p0 = p(t B ) = . (23)
v0 γ 2

The pressure has its highest value at tB because the volume of the foil is still confined right before the
initiation of the burst. Afterwards, the pressure as well as the density of the expanding vapor
decrease. By contrast, the velocity increases until the maximum velocity is reached, naturally
followed by a deceleration stage which is not considered in the model presented here. For a given
energy deposition and, thus, maximum velocity, the peak pressure increases for lower values of
(tE - tB), meaning for shorter durations of vapor expansion. On the other hand, at elevated energy
depositions both the pressure and the velocity increase for a constant value of (tE – tB). Since
chemical reactions of the vapor are neglected in the model, the influence of the foil material on the
flyer velocity is mainly given by the simplified Gurney model in Eq (10). Attention should be paid to
the fact that the model is capable of providing the pressure-time, velocity-time, and displacement-
time history of the flyer just between the two points in time and . Since the displacement D(t) is
calculated by integrating velocity over time, the model can only be applied if the impact between
flyer and target (v = vim) occurs at or before D(tE). Potential errors between the model and
experimental data may also be attributed to the fact that adiabatic heating and rate effects causing
changes in the yield stress of the flyer material prior to the burst are not covered by the present
model. If all other values are kept constant in the model, it can be shown that a decreased yield stress
results in a slightly lower flyer velocity at a given distance, which is physically reasonable.

4. Experimental procedure

All experiments were performed using actuators connected to a Maxwell-Magneform pulse generator
having eight parallel capacitors. Each capacitor is discharged by two parallel ignitron switches. The
power supply has a total capacitance of C = 426 μF, an internal resistance of Ri = 10 mΩ, an internal
inductance of Li = 100 nH, and a maximum charging energy of 16 kJ at a charging voltage of
8.66 kV. The resultant short-circuit frequency is in the range of 25 kHz. The basic experimental
setups of the VFAW and the MPW system are quite similar, as can be seen in Fig. 8.

14
Figure 8: Schematic of the VFAW and MPW setup.

The actuator type, either a MPW coil or a vaporizing foil, is electrically insulated from and in close
proximity to the sheet metal flyer. Spacers, whose height defines the standoff distance, allow space
for the flyer to be accelerated plastically by the pressure pulse before hitting the target. The laser
beam of the photonic Doppler velocimetry (PDV) system, which is addressed later in this section, is
aligned orthogonally to the flyer surface through holes drilled in the target plate and fixture. This
diagnostic tool is used to acquire velocity-time histories over the course of each experiment. The
target hole is centered in relation to the upper surface of the actuator. For both the MPW and the
VFAW experiments the parameters varied are standoff distance and charging energy of the capacitor
bank, resulting in different impact angles and velocities. On that account, a full factorial matrix
containing three different energies (4.8 kJ, 6.4 kJ and 8 kJ) and three different standoff distances
(1.1 mm, 2.1 mm and 3.4 mm) - yielding nine combinations per process - was chosen. Partial
repetitions were done to ensure statistical certainty. Test matrices summarizing the experiments are
given in Table 1 a) and b). As the flyer velocity was recorded over time, experiments with higher
standoff distances could be used to also evaluate velocities at distances that correspond to the lower
standoff distances.

15
Table 1: Test matrix for a) VFAW experiments, b) MPW experiments.

a) VFAW experiments Charging energy Flyer velocity

also evaluated
X = 1 experiment 4.8 kJ 6.4 kJ 8 kJ at distance:

1.1 mm XX X X -
Standoff
2.1 mm X X XXX 1.1 mm
distance
3.4 mm X XXX X 1.1 mm; 2.1 mm

b) MPW experiments Charging energy Flyer velocity

also evaluated
X = 1 experiment 4.8 kJ 6.4 kJ 8 kJ at distance:

1.1 mm X X XX -
Standoff
2.1 mm X X X 1.1 mm
distance
3.4 mm X X XXX 1.1 mm; 2.1 mm

The horizontal distance between the spacers was always 32 mm. The flyer and target material
selected for all experiments is EN AW 5005A (AlMg1), a medium-strength aluminum alloy, which
is commonly applied in lightweight constructions. The aluminum sheets were not coated. In quasi-
static tensile tests the sheet metal was measured to have a yield strength of 130 MPa, an ultimate
tensile strength of 153 MPa, and Young’s modulus was found to be 70 GPa. These values refer to the
direction of rolling. All specimens (flyer and target plates in both VFAW and MPW) had a thickness
of 1 mm, a width of 75 mm, and a length of 100 mm. The density of all aluminum parts (specimens
and foils) is taken to be 2.70 g/cm3. Furthermore, tensile tests of the welded sheets were performed to
evaluate the weld quality. For reasons of safety to personnel, the setup for vaporizing foil actuators is
always deployed in a shrapnel containment enclosure which also provides noise abatement and
exhaust ventilation. The experimental assembly is shown in Fig. 9 where the setup is displayed
inversely to previous VFAW figures.

16
Figure 9: VFAW setup (upper fixture is not shown).

The foil is enclosed by polyester tape and additional Kapton tape (polyimide film) to guarantee good
insulation against the flyer and the lower steel fixture. In turn, two G10 (glass-fiber reinforced
epoxy) plates ensure the insulation of the steel fixture against the copper leads connected to the
capacitor bank. The foil ends are not taped so there is good electrical continuity when clamped to the
copper leads. Essentially uncoated pure aluminum foil (1000 series) with a thickness of 0.0762 mm
was used as actuator material. Each foil was cut out using a robust steel template. The active foil
region (see Fig. 9) is 52 mm long and 12.5 mm wide for every VFAW experiment. The bottom side
of the flyer is also taped with Kapton, with the objective to avoid arc formation on the flyer surface
and to ensure additional thermal and electrical insulation. Eventually, the setup is bolted together
using four threaded rods. The principal element of the MPW assembly, also displayed inversely to
the previous MPW figures, is pictured in Fig. 10.

17
Figure 10: MPW coil (upper fixture, target, and flyer plates as well as spacers are not shown).

It features a one-turn coil made of copper embedded in an insulating G10 plate. The cross-section of
the pressure lead has a width of 6.5 mm and a height of 13.3 mm. All copper-to-copper connections
of the coil are brazed and the only return path is positioned nearby and parallel to the pressure lead in
order to keep the inductance of the coil low. The return path is several times wider than the pressure
lead so as not to serve as a second pressure lead. Kapton tape is applied to the coil area for the
purpose of electrical insulation between the coil and the flyer plate.

Measurement of the transient discharge current was accomplished using a 100 kA : 1 V Rogowski
coil around one output lead of the of the capacitor bank and voltage histories were measured by
connecting a 1000 : 1 high-voltage probe to the terminals of the capacitor bank (in the connection
area of the actuators). For the purpose of recording the data collected by these probes, both of them
were connected to a LeCroy Waverunner 620Zi oscilloscope providing four channels, a bandwidth of
2 GHz, and a maximum sampling rate of 20 GS/s. Further data processing and analysis was done by
employing the commercial software Matlab. This extends to velocity-time histories acquired by the
PDV system. This method of measuring high velocities is described in Strand et al. (2004). A laser
beam with a wavelength of = 1550 nm is emitted by a focussing (or collimating) probe. It is then
reflected from the surface of the moving flyer causing a Doppler-shifted light that re-enters the lens
of the probe. Hence, the stationary probe serves as both source and receiver of the signal. A discrete
short-time Fourier transform (STFT) of the detected signal allows creating a spectrogram plot of the
beat frequency fb(t) that is equal to the difference between the Doppler-shifted frequency and the
frequency of the incident light signal. Flyer velocity histories v(t) can then be computed as

λ
v(t ) = 0 f b (t ) . (24)
2

18
A time window of 500 μs, a sampling rate of 5 GS/s, and a window size of 1000 with an overlap of
500 were chosen for the spectrograms of this study. Until the flyer impact, around 20 velocity points
were picked from each spectrogram plot to generate v(t) graphs. This method is similar to the one
shown in Vivek et al. (2014c). All key components of the PDV system used for the experiments
conform to those presented in Daehn et al. (2008).

5. Analysis of experiments

5.1 Energy deposition

As mentioned above, the amount of energy deposited into the foil prior to the instant when the foil
bursts (ED according to Eq. (1)) denotes an important value in the context of VFAW. Corresponding
experimental results are collected in Fig. 11, regardless of the standoff distance chosen for welding.

Figure 11: Influence of charging energy on energy deposition in VFAW experiments. The burst time
was consistently around 11 μs.

Neglecting energy losses in the copper parts of the VFAW setup, a consideration of the ratio of
average energy deposition to the latent heat of vaporization for the mass of the active foil region,
Hb = c· ΔH (ΔH = 10.78 MJ/kg for aluminum following Davis (1993)), shows an increase from the
value of 2 at a charging energy of 4.8 kJ to 2.5 at 8 kJ. The burst currents range from 98 kA at 4.8 kJ
to 110 kA at 6.4 kJ to 123 kA at 8 kJ of charging energy, while the associated peak voltages range
from 5.4 kV to 6.5 kV to 7.5 kV. These observations emphasize the assertion that the amount of
energy that can be deposited until tB is evidently not a constant for a given metal, as discussed in
Section 2.2. The accompanying increase in average heating rates ED/tB mapped in Fig. 11 also
supports the idea of a rate dependency. The burst times observed are about the same for all

19
experiments (average of 10.78 μs), so more energy is supplied to the foil during a time step when the
initial charging energy of the capacitor is raised. That is to say, the energy deposited increases with
charging energy if the circuit elements of the pulse generator (Ri, Li, C) are not changed over the
course of experiments. It is, however, conspicuous that there are relatively large ranges (in the area
of 15 %) concerning ED for a given charging energy and, thus, there are some overlaps with the next
value of input energy (compare Fig. 11). One reason for this could be the stochastic nature of
electrical vaporizations as they are based on the occurrence of instabilities. The observed increases
only show a relatively weak linear correlation on average, meaning that almost doubling the charging
energy just leads to a 25 % higher ratio of ED/Hb.

Regarding electromagnetic forming, the relevant energy partially converted into kinetic and
deformation energy of the flyer is not ohmic heating, but magnetic energy produced by the actuator.
It is therefore useful to look at the coil currents in this context. The peak currents obtained linearly
increase with the charging energy from 225 kA to 270 kA to 307 kA, which is more than twice as
high as the highest value of the VFAW experiments. The corresponding peak voltages in MPW
conform to the initial charging voltages (4.7 kV at 4.8 kJ, 5.5 kV at 6.4 kJ, 6.1 kV at 8 kJ) In general,
the workpiece deformation involves non-linear workpiece and mutual inductances and thus a time-
varying frequency behavior. Referring to Bühler and Bauer (1968), a significant frequency f can be
calculated on the basis of the point when the peak coil current is reached, Δtrise.

1 (25)
f =
4 ∆trise

Again, the remaining circuit elements were not varied, resulting in a frequency of f = 19.7 kHz for
every energy level.

5.2 Flyer velocities in VFAW and MPW

All other things being equal, higher energy depositions in case of VFAW and higher peak currents in
case of MPW lead to higher flyer plate velocities. The velocities recorded with the PDV system are
treated in this paragraph. Fig. 12 compiles flyer plate velocities at certain distances D and charging
energies for both processes, VFAW and MPW.

20
Figure 12: Comparison of flyer velocities in VFAW and MPW at certain distances and energy levels
using PDV.

The most notable fact is that vaporizing foils provide flyer velocities up to three times higher as
compared to MPW at the same charging energy, in numbers 930 m/s versus 300 m/s at the
maximum. Similar to the energy deposition, for an identical experimental configuration the VFAW
velocities shown in Fig. 12 exhibit deviations of up to 25 % with regard to the corresponding average
value. Resulting overlaps of velocity between different charging energies make it difficult to give
precise statements about correlations. Besides the mentioned stochastic nature of instabilities during
the energy deposition, the partly significant scatter for identical experimental configurations may be
attributed to edge effects of the foil. This means that imperfections or sharp edges of the foil due to
the cutting during the manufacture of the foil samples can cause local energy concentrations in
undesired regions during the vaporization, which eventually results in lower flyer velocities. This
assertion is supported by the work of Richardson et al. (1988) who underline that the foil corners
should be rounded carefully to avoid such a non-uniform heating. On average, however, the

21
following statements can be drawn for the trend lines displayed in Fig. 12. The flyer velocities at
distances of 1.1 mm do not increase as strongly with energy input as the velocities at larger standoff
distances. An explanation for this is given by the idea that the available energy is not yet converted
into kinetic energy of the flyer at such short distances, meaning that more travel is required if
substantially higher impact velocities are expected to be generated without changing the input
energy. The trend line for the standoff distance of the 3.4 mm increases, in turn, stronger from 4.8 kJ
to 6.4 kJ than from 6.4 kJ to 8kJ. As long as the maximum possible velocity is not yet reached at the
impact, a higher standoff distance would be needed to make even higher flyer velocities possible at
8 kJ. Hence, the VFAW velocity trend line for 3.4 mm rather increases in a non-linear manner with
charging energy, whereas almost linear correlations can be obtained for lower distances. Velocities
recorded during the MPW experiments do not show a variance like those of VFAW. None of the
MPW experiments yielded an actual weld (joining partners were removable by hand), while all
VFAW specimens were solidly welded. It must, however, be noted that neither of the setups was
optimized in a specific manner. In case of MPW, peak currents in the order of 300 kA at a charging
energy of 8 kJ are already in the maximum admissible regime of the pulse generator. That is why no
higher energy was tried to achieve higher velocities, and, thus actual magnetic pulse welds. It is
important to note that the impact velocity is not necessarily equal to the maximum velocity. The peak
value can also occur before hitting the target (see MPW experiments in Fig. 12) or the flyer may still
be accelerating at the point of impact. This can also be seen in Fig. 13, where selected velocity-time
curves of the VFAW and MPW experiments are plotted for the same initial configuration.

Figure 13: Selected velocity-time curves generated from PDV data for MPW and VFAW.

Further, the influence of the Lorentz force on the flyer velocity when employing VFA may be
derived from Fig. 13: Prior to the burst, the foil acts as a tool coil in EMF. In this particular case, the

22
electromagnetic launch provides velocities of about 50 m/s, followed by a rapid increase at the
instant of vaporization. As opposed to this, the velocity curve describing the MPW experiment
proceeds more continuously.

5.3 Testing of welded joints

One method of evaluating the quality of a welded lap joint is to compare its tensile shear strength
with an appropriate parameter of the parent material. To express strength in terms of stress, the actual
joining area as well as the stress distribution over this area must be known, which is usually hard to
measure non-destructively and is unknown in the present case. Besides, bending moments in the lap
joint affect the force measurements during a tensile test. Since the point when necking begins - the
ultimate tensile strength - is characterized by the maximum force, it is reasonable to compare the
maximum tensile force a lap joint can bear, Fmax, with the one of a reference strip having the same
cross-sectional geometry as the weakest joining partner, FUTS (strength ratio Fmax/FUTS). This
procedure could only be performed for the VFAW specimens because no magnetic pulse welds were
achieved. Resultant strength ratios are depicted in Fig. 14 a.

Figure 14: Evaluation of weld quality for VFAW experiments. a) Tensile test results for different
charging energies and standoff distances. b) Necking and thinning in the specimen. c) Etched
microsection of an exemplary weld seam.

23
Here, it can be seen that the strength ratio is close to one almost independent of energy input and
standoff distance. Even though the actual impact angle is unknown in this case, larger standoff
distances imply larger impact angles. Further, as shown before, greater charging energies enable
greater impact velocities. All investigated energies and standoff distances are apparently in a range
where the maximum strength of the weld is reached anyway, which indicates a large process window
for the given material combination. Necking always occurred in one of the joining partners farther
away from the clamping instead of in the weld zone. It can therefore be stated that the shear strength
of the VFAW joints is at least as high as the ultimate strength of the parent material. Owing to the
clamping force the flyer plate is subjected to in the region where the spacers are positioned (clamp
area), stress ratios lower than one can be explained through thinning effects during the forming stage
at that very place, resulting in a reduced maximum bearable test load. This statement is supported by
the fact that necking always started in the clamp area, which is pictured in Fig. 14 b. Due to elevated
forming pressures it is evident that higher charging energies facilitate thinning effects at larger
standoff distances (compare Fig. 14 a). The very first impact onto the target does not occur under an
oblique angle. Thus, a rebound effect is observable at the center, so that actual (symmetric) weld
seams are restricted to a small region near the center. Fig. 14 c shows a micrograph of the interface
cross-section of a weld zone for an 8 kJ experiment which indicates the typical wavy pattern
including regions where neither interlayers nor melt pockets could be found. Consequently, a solid
state weld can be expected there.

6. Verification of the analytical model for VFA

In this paragraph, the simplified analytical model presented within the scope of this work (Eq. (20)-
(23)) is applied to the current and voltage traces measured during the VFAW experiments (compare
Fig. 6). The results are compared with the corresponding results generated from PDV data for the
purpose of evaluating the accuracy and, therefore, the general applicability of the model. Since the
impact with the target happens at a predefined distance (standoff distance), the values at those
particular distances are of most interest when computing the flyer velocity and its time integral, the
displacement. To obtain a comprehensive overview concerning all vaporizing foil experiments, the
analytically calculated and recorded velocities are compiled in Fig. 15 for the varied charging
energies and distance levels.

24
Figure 15: Measured and analytical flyer plate velocities in VFAW for different charging energies
and distances of a) 1.1 mm, b) 2.1 mm, and c) 3.4 mm.

25
Naturally, the basic courses of the trend lines equal those of Fig. 12 as both figures are based on the
same experiments. Since the model involves energy depositions ED, their variances also affect the
analytical results. Like for the experimentally obtained flyer velocities, consequent overlaps of
calculated velocities between different charging energies complicate the establishment of clearly
defined relationships for any standoff distance in Fig. 15 a, b, and c. Nevertheless, the model
properly reflects the influence of energy deposition on the velocity at distinct distances. Single
deviations between experimental results and analytics range from perfect agreement to 34 %. The
overall average deviation is about 11 %, which is an acceptable value if the model is used as a first
approximation for the process design of a specific VFAW lap joint. The main reason for the
deviations may be that the analytical model postulates a constant mass of vaporized aluminum
(narrow active foil region) in Eq. (10), independent of energy input. The real energy density
distribution is most likely not restricted to exactly the same foil region at varied energy depositions,
though. Beside the velocities, the pressure acting on the flyer plate according to Eq. (5), which
complies with Newton´s second law, is also of interest. Amplitudes or peak values of pressure shall
be compared with the analytical model, too. On the basis of smoothed experimental velocity curves,
pressure-time histories can be approximated by applying a finite-difference method to Newton´s
second law. Central differences were used here for the necessary first time derivative of velocity.
Peak pressures arising from this approach are collected in Fig. 16 and range from about 1 GPa at a
charging energy of 4.8 kJ to more than 2.5 GPa at 8 kJ. The observable linear course in peak pressure
versus energy is qualitatively in accordance with the work of Grigoriev and Pavlenko (2009). It may
be noted that peak pressures are only of the order of 100 MPa or less when applying the same
approach to the MPW experiments.

26
Figure 16: Analytical and (indirectly) measured peak pressures in VFAW for various charging
energies regardless of standoff distance

With a maximum deviation of 18 % and an average deviation of just 8 % between analytical peak
pressures and those indirectly obtained from PDV data, the model for VFA also provides an adequate
pressure approximation. The peak or burst pressure is related to the instant of the burst when almost
no flyer displacement has happened yet. Accordingly, the standoff distance chosen for welding does
not really have an effect on the peak pressure. A complete and comparison of the temporal evolution
of pressure and flyer velocity between model and experiment is given in Fig. 17, where a standoff
distance of 3.4 mm is chosen in order to cover the largest possible displacement regarding the
experiments conducted.

27
Figure 17: Comparison of analytical and experimentally obtained results for a charging energy of
8 kJ and a standoff distance of 3.4 mm. a) Pressure and velocity plotted versus time. b) Velocity
plotted versus displacement.

Fig. 17 a accentuates the domain of definition of the analytical model starting with burst time tB.
According to PDV data, the impact on the target plate occurs at t = 14.9 μs, while t = 16.3 μs is
obtained for the analytically calculated impact, which is close to the end of the model area
(tE = 16.7 μs). Experimental and analytical curves show resembling shapes ending up at nearly the
same impact velocity of about 880 m/s. The virtually instantaneous rise to a peak pressure of circa
2.4 GPa is outside the range of the model and is therefore not part of it. Even so, the peak pressure in
itself and its subsequent decrease over time are sufficiently taken into account by the model. In
Fig. 17 b, the velocity is plotted over the associated flyer displacement. Good agreement between
analytics and experimental data in such a graph (v(D)) is of special importance for collision welding
when deciding on the standoff distance of a joint. Since Fig. 17 represents one of the best
comparisons, an example with a rather significant deviation between the experiment and the model
of approximately 25 % is given in Fig. 18 a and b.

28
Figure 18: Comparison of analytical and experimentally obtained results for a charging energy of
8 kJ and a standoff distance of 2.1 mm. a) Pressure and velocity plotted versus time. b) Velocity
plotted versus displacement.

The analytically computed impact velocity in Fig. 18 follows a course similar to the one shown in
Fig. 17 of up to 750 m/s while the corresponding measured flyer velocity shows a significant
decrease in acceleration so that only 600 m/s were reached at impact, although the same charging
energy of 8 kJ was chosen. This is the lower outlier also displayed in Fig. 15 b. Despite some distinct
outliers the curves belonging to the VFAW experiments still feature the same qualitative
progressions as in the previous example.

7. Conclusions

This work has shown that VFAW is a very suitable collision welding method and more efficient than
MPW as applied to solid state lap welding. Employing the example of a lap joint of 1 mm thick EN
AW 5005A sheets, both processes were directly compared under the same experimental conditions,
in concrete terms: charging energies, configuration of the capacitor bank, specimen geometry, and
standoff distances between flyer and target plate. Besides the advantage that vaporizing foils do not
have tool life issues as they are disposable low cost actuators, flyer impact velocities up to 900 m/s
were recorded, which is three times higher than the maximum of the corresponding MPW
experiments. Even at the lowest charging energy (4.8 kJ) impact velocities around 450 m/s were
reached using VFAW, with the result that a weld was formed in every VFAW experiment. By
contrast, not even the highest energy input (8 kJ) yielded an impact velocity high enough to produce
an actual magnetic pulse weld. Tensile tests showed that high joint strengths can be achieved with

29
VFAW, meaning failure always happened in one of the joining partners while the weld seam was
still intact. In addition, a simplified closed-form solution that allows plotting the time-dependent
flyer plate velocity with respect to displacement has been introduced for VFA. With an average
deviation of about 11 % in terms of impact velocity, this analytical model is in satisfying agreement
with PDV data, so that it can help predetermining the standoff distance needed for a specific VFAW
joint. It would be desirable to further verify the model by investigating different flyer thicknesses and
materials including those of lower electrical conductivity, since the irrelevancy of workpiece
conductivity constitutes another advantage of vaporizing foils against MPW.

Acknowledgements

This paper is based on investigations of the Collaborative Research Center SFB/TR 10, subproject
A10 “Joining by forming”, which is kindly supported by the German Research Foundation (DFG).
The authors would also like to acknowledge support of the U.S. Department of Energy through the
Clean Energy Research Centre for Clean Vehicle Collaboration (CERC-CVC), award number DE-
PI0000012. Also appreciated is the financial support of the German Academic Exchange Service
(DAAD-PROMOS grant). Many thanks to Steven Hansen from The Ohio State University for his
support with the experiments and for the fruitful discussions.

References

Aizawa, T., 2003. Magnetic pressure seam welding method for aluminium sheets. Welding International 17
(12), 929-933, doi: 10.1533/wint.2003.3199.

Bealing, R., Carpenter, P.G., 1972. Exploding foil devices for shaping meg-amp current pulses. J. Phys. E:
Sci. Instrum. 5, 889-892, doi: 10.1088/0022-3735/5/9/016.

Beerwald, C., 2005. Grundlagen der Prozessauslegung und -gestaltung bei der elektromagnetischen
Umformung. Dr.-Ing. Dissertation, Universität Dortmund, ISBN 3-8322-4421-2.

Ben-Artzy, A. Stern, A., Frage, N., Shribman, V., Sadot, O., 2010. Wave formation mechanism in magnetic
pulse welding. International Journal of Impact Engineering 37 (4), 397-404,
doi: 10.1016/j.ijimpeng.2009.07.008.

Bühler, H., Bauer, D., 1968. Ein Beitrag zur Magnetumformung rohrförmiger Werkstücke. Werkstatt und
Betrieb 110 (9), 513–516.

Chace, W.G., Levin, M.A., 1960. Classification of wire explosions. J. Appl. Phys. 31, 1298,
doi: 10.1063/1.1735823.

30
Chau, H.H., Dittbenner, G., Hofer, W.W., Honodel, C.A., Steinberg, D.J., Stroud, J.R., Weingart, R.C., Lee,
R.S., 1980. Electric gun: a versatile tool for high-pressure shock-wave research. Rev. Sci. Instrum. 51
(12), 1676-1681, doi: 10.1063/1.1136155.

Daehn, G.S., 2010. Energy field methods and electromagnetic sheet metal forming. In: Zhang, W. (Ed.),
Intelligent Energy Field Manufacturing: Interdisciplinary Process Innovations. CRC Press, pp. 471-
504, ISBN 978-1420071016.

Daehn, G.S., Zhang, Y., Golowin, S., Banik, K., Vivek, A., Johnson, J.R., Taber, G., Fenton, G.K., Henchi, I.,
L´Eplattenier, P., 2008. Coupling experiment and simulation in electromagnetic forming using Photon
Doppler Velocimetry. In: Proceedings of the 3rd International Conference on High Speed Forming -
ICHSF, Dortmund, Germany, pp. 35-44, online available at https://eldorado.tu-
dortmund.de/handle/2003/27105.

Davis, J.R., 1993. Aluminum and Aluminum Alloys. ASM International, p. 643, ISBN 978-0871704962

Göbel, G., Kaspar, J., Herrmannsdörfer, T., Brenner, B., Beyer, E., 2010. Insights into intermetallic phases on
pulse welded dissimilar metal joints. In: Proceedings of the 4th International Conference on High
Speed Forming – ICHSF, Columbus OH, USA, pp. 127-136, online available at https://eldorado.tu-
dortmund.de/handle/2003/27191.

Grigoriev, A.N., Pavlenko, A.V., 2009. Pressure generated by the electric explosion of metal foils. Technical
Physics Letters 35 (9), 865-868, doi: 10.1134/S1063785009090235.

Harvey, G.W., Brower, D.F., 1958. Metal Forming Device and Method. US-Patent Nr. 2976907.

Henselek, A., Beerwald, M., Beerwald, C., 2004. Design and adaptation of EMF equipment – from direct
acting multi-turn coils to separable tool coils for electromagnetic tube compression. In: Proceedings of
the 1st International Conference on High Speed Forming - ICHSF, Dortmund, Germany, pp. 275-284,
online available at https://eldorado.tu-dortmund.de/handle/2003/27029.

Jones, N., 1989. Structural impact, 1st ed.. Cambridge University Press, Cambridge, ISBN 978-0521628907.

Kamal, M., 2005. A uniform pressure electromagnetic actuator for forming flat sheets. Ph.D. Thesis, The Ohio
State University, online available at http://rave.ohiolink.edu/etdc/view?acc_num=osu1127230699.

Kore, S.D., Imbert, J., Worswick, M.J., Zhou, Y., 2009. Electromagnetic impact welding of Mg to Al sheets.
Science and Technology of Welding and Joining 14 (6), 549-553, doi: 10.1179/136217109X449201.

Lee, W.M., Ford, R.D., 1988. Pressure measurements correlated with electrical explosion of metals in water.
Journal of Applied Physics 64 (8), 3851-3854, doi: 10.1063/1.341365.

Li, F., Mo, J., Li, J., Zhou, H., Huang, L., 2013. Study on the driver plate for electromagnetic forming of
titanium alloy Ti-6Al-4V. Int. J. Adv. Manuf. Technol. 69 (1), 127-137, doi: 10.1007/s00170-013-
5002-1.

Loeb, L.B., 1927. Kinetic theory of gases, 1st ed.. McGraw-Hill Book Company, New York and London.

31
Martins, J.B., Symonds, P.S., 1965. Mode approximations for impulsively loaded rigid plastic structures.
Techn. Report from Brown University, Div. of Engineering, Providence RI, USA, online available at
http://www.dtic.mil/docs/citations/AD0621580.

Mori, K., Bay, N., Fratini, L., Micari, F., Tekkaya, A.E., 2013. Joining by plastic deformation. CIRP Annals –
Manufacturing Technology 62 (2), 673-694, doi: 10.1016/j.cirp.2013.05.004.

Mousavi, S.A.A. Akbari, Sartangi, P.F., 2009. Experimental investigation of explosive welding of cp-
titanium/AISI 304 stainless steel. Materials and Design 30 (3), 459-468,
doi: 10.1016/j.matdes.2008.06.016.

Nelson, L.S., Hogeland, S.R., Roth, T.C., 1999. Alumium-enhanced underwater electrical discharges for
steam explosion triggering. Sandia Labs. Report, Albuquerque NM, USA, 58 pages,
doi: 10.2172/12653.

Osher, J.E., Barnes, G., Chau, H.H., Lee, R.S., Lee, C., Speer, R., Weingart, C., 1989. Operating
characteristics and modeling of the LLNL 100-kV electric gun. IEEE Transactions on Plasma Science
17 (3), 392-402, doi: 10.1109/27.32247.

Pécseli, H.L., 2012. Waves and oscillations in plasmas, 1st ed.CRC Press, pp. 209-210, ISBN 978-
0415613033.

Psyk, V., Risch, D., Kinsey, B.L., Tekkaya, A.E., Kleiner, M., 2011. Electromagnetic forming – a review. J.
Mater. Process. Tech. 211 (5), 787-829, doi: 10.1016/j.jmatprotec.2010.12.012.

Reissner, J., Müller-Duysing, M., Dannenmann, E., Ladwig, J., 1990. Biegen. In: Lange, K., Umformtechnik-
Handbuch für Industrie und Wissenschaft, Band 3: Blechbearbeitung, 2. Aufl.. Springer Verlag, pp.
243-310, ISBN 3-540-50039-1.

Richardson, D.D., Northeast, E.D., Ryan, P.F.X., 1988. An exploding foil flying plate generator. Materials
Research Lab. Report, Melbourne, Australia, 37 pages.

Sarathi, R.,Chakravarthy, S.R., Venkataseshaiah, C., 2004. Studies on the generation and characterization of
nanoalumina powder using wire explosion technique. International Journal of Nanoscience 3 (6), 819
– 827, doi: 10.1142/S0219581X04002711
Sarkisov, G.S., Struve, K.W., McDaniel, D.H., 2004. Effect of current rate on energy deposition into
exploding metal wires in vacuum. Physics of Plasmas 11 (10), 4573-4581, doi: 10.1063/1.1784452.

Schmidt, S.C., Seitz, W.L., Wackerie, J., 1977. An empirical model to compute the velocity histories of flyers
driven by electrically exploding foils. Los Alamos Scientific Laboratory Report, Albuquerque NM,
USA, 14 pages.

Sedoi, V.S., Mesyats, G.A., Oreshkin, V.I., Valevich, V.V., Chemezova, L.I., 1999. The current density and
the specific energy input in fast electrical explosion. IEEE Transactions on Plasma Science 27 (4),
845-850, doi: 10.1109/27.782248.

Shribman, V., 2008. Magnetic pulse welding for dissimilar and similar materials. In: Proceedings of the 3rd
International Conference on High Speed Forming - ICHSF, Dortmund, Germany, pp. 13-22, online
available at https://eldorado.tu-dortmund.de/handle/2003/27107.

32
Sinton, R.P.W., 2011. Long distance exploding wires. Ph.D. Thesis, University of Canterbury, online
available at http://hdl.handle.net/10092/6586.

Strand, O.T., Berzins, L.V., Goosman, D.R., Kuhlow, W.W., Sargis, P.D., Whitworth, T.L., 2004.
Velocimetry using heterodyne techniques. In: Proc. SPIE 5580, 26th International Congress on
High-Speed Photography and Photonics, Alexandria, VA, USA, doi: 10.1117/12.567579.

Tucker, T.J., Stanton, P.L., 1975. Electrical Gurney energy: a new concept in modeling of energy transfer
from electrically exploded conductors. Sandia Labs. Report, Albuquerque NM, USA, 28 pages.

Verstraete, J., De Waele, W., Faes, K., 2011. Magnetic pulse welding: lessons to be learned from explosive
welding. Sustainable Construction & Design 2 (3), 458-464, ISSN 2032-7471.

Vivek, A., Brune, R.C., Hansen, S.R., Daehn, G.S., 2014a. Vaporizing foil actuator used for impulse forming
and embossing of titanium and aluminum alloys. J. Mater. Process. Tech. 214 (5), 865-875,
doi: 10.1016/j.jmatprotec.2013.12.003.

Vivek, A., Hansen, S.R., Daehn, G.S., 2014b. High strain rate metalworking with vaporizing foil actuator:
Control of flyer velocity by varying input energy and foil thickness. Review of Scientific Instruments
85 (7), 075101, doi: 10.1063/1.4884647.

Vivek, A., Hansen, S.R., Liu, B.C., Daehn, G.S., 2013. Vaporizing foil actuator: a tool for collision welding.
J. Mater. Process. Tech. 213 (12), 2304-2311, doi: 10.1016/j.jmatprotec.2013.07.006.

Vivek, A., Liu, B.C., Hansen, S.R., Daehn, G.S., 2014c. Accessing collision welding process window for
titanium/copperwelds with vaporizing foil actuators and grooved targets. J. Mater. Process. Tech. 214
(8), 1583-1589, doi: 10.1016/j.jmatprotec.2014.03.007.

Wang, G., Zhao, J., Luo, B., Jiang, J., 2011. Magnetohydrodynamics of metallic foil electrical explosion and
magnetically driven quasi-isentropic compression. In: Schulz, H. (Ed.), Hydrodynamics – Advanced
Topics. InTech, pp. 347-378, ISBN 978-953-307-596-9.

Weddeling, C., Hahn, M., Daehn, G.S., Tekkaya, A.E., 2014. Uniform pressure electromagnetic actuator – an
innovative tool for magnetic pulse welding. In: Proceedings of the International Conference on
Manufacture of Lightweight Components – ManuLight 2014, Dortmund, Germany, Procedia CIRP,
vol. 18, pp. 156–161, doi: 10.1016/j.procir.2014.06.124.

Weingart, R.C., Lee, R.S., Jackson, R.K., Parker, N.L., 1976. Acceleration of thin flyers by exploding foils:
application to initiation studies. In: Proceedings of the 6th International Symposium on Detonation,
Coronado CA, USA, pp. 653-663.

Winkler, R., 1973. Hochgeschwindigkeitsbearbeitung: Grundlagen und technische Anwendung elektrisch

erzeugter Schockwellen und Impulsmagnetfelder, 1. Auflage. VEB Verlag Technik, Berlin.

Young, W.C., Budynas, R.G., 2002. Roark´s formulas for stress and strain, 7th ed..McGraw-Hill, p. 193,
ISBN 0-07-072542-X.

33
Zhang, Y., 2010. Investigation of magnetic pulse welding on lap joint of similar and dissimilar materials.
Ph.D. Thesis, The Ohio State University, online available at
http://rave.ohiolink.edu/etdc/view?acc_num=osu1268135049.

Zhang, Y., Babu, S., Daehn, G.S., 2010. Impact welding in a variety of geometric configurations. In:
Proceedings of the 4th International Conference on High Speed Forming – ICHSF, Columbus OH,
USA, pp. 97-107, online available at https://eldorado.tu-dortmund.de/handle/2003/27194.

Zou, X.B., Mao, Z.G., Wang, X.X., Jiang, W.H., 2012a. Two different modes of wire explosion for nano-
powder production. A Letters Journal Exploring the Frontiers of Physics 97 (3), 35004,
doi: 10.1209/0295-5075/97/35004.

Zou, Y., He, B., Jiang, X., Wang, L., 2012b. Relationships between bridge foil parameters and input pulse
current. In: Proceedings of the 2012 International Symposium on Safety Science and Technology,
Procedia Engineering, vol. 45, 1020-1024, doi: 10.1016/j.proeng.2012.08.276

34