International Journal of Heat and Mass Transfer 50 (2007) 2217–2235

www.elsevier.com/locate/ijhmt

Effects of electromagnetic force on melt flow and porosity prevention

in pulsed laser keyhole welding
Jun Zhou, Hai-Lung Tsai *
Laser-Based Manufacturing Laboratory, Department of Mechanical and Aerospace Engineering, University of Missouri-Rolla,
1870 Miner Circle, Rolla, MO 65409-1350, United States

Received 13 March 2006; received in revised form 28 October 2006

Available online 29 December 2006

Abstract

Porosity formation in pulsed laser keyhole welding was found to be affected by two competing factors: (1) the solidification rate of
molten metal and (2) the back filling speed of molten metal during the keyhole collapse process. Porosity (pores/voids) was found in
welds when the solidification rate of molten metal exceeds the back filling speed of molten metal. In this study, the use of electromagnetic
force was proposed to control the back filling speed of molten metal, and a mathematical model was developed to investigate the effects
of electromagnetic force on the transient melt flow, keyhole dynamics, and porosity formation. The results demonstrate that porosity in
pulsed laser welding can be prevented by an applied electromagnetic force. Parametric studies to determine the desired strength of the
electromagnetic force and its duration were also conducted to achieve quality welds.
Ó 2006 Elsevier Ltd. All rights reserved.

Keywords: Laser welding; Electromagnetic force; Keyhole; Porosity prevention

1. Introduction close relationship with keyhole dynamics and the collapse

of the keyhole right after the termination of laser irradia-
Lasers have achieved increased significance for welding tion. In pulsed laser welding, during the keyhole formation
in recent years for their high power density and the result- process, molten metal is squeezed outward and upward,
ing small heat-affected zone in welds. Especially, with high surrounding the upper portion of the keyhole. Once the
depth-to-width aspect ratio and high welding speed, laser laser irradiation is terminated, the melt in the upper part
keyhole welding is more promising compared to conven- of the keyhole flows downward to fill the keyhole. At the
tional welding processes. However, in deep penetration same time, the upper part of the melt rapidly solidifies
laser welds, pores/voids (porosity) are frequently observed which prevents the melt from flowing to fill the keyhole,
[1–4] which deteriorate the strength of the weld. In order to leading to the formation of porosity (pores). Note porosity
optimize a laser welding process and to ensure high weld can also be caused by the entrainment of shielding gas and/
quality and strength, it is necessary to understand the or the dissolution of gas (for example, hydrogen) during
porosity formation mechanism in pulsed laser welding the solidification process which is normally called gas
and to find methods to reduce or eliminate porosity defects. porosity [8–10]. This study is limited to the porosity that
Over the years, a number of researchers have conducted is caused by the failure of back filling the keyhole due to
experiments to investigate porosity formation in pulsed premature solidification of molten metal which can occur
laser welding [5–7]. Based on these experimental observa- when the keyhole is relatively deep in pulsed laser keyhole
tions, it was found that the formation of porosity has a welding.
Zhou et al. [11,12] developed mathematical models to
*
Corresponding author. Tel.: +1 573 341 4945; fax: +1 573 341 4607. investigate the keyhole dynamics, heat transfer and fluid
E-mail address: tsai@umr.edu (H.-L. Tsai). flow, and the porosity formation process in pulsed laser

0017-9310/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.ijheatmasstransfer.2006.10.040
2218 J. Zhou, H.-L. Tsai / International Journal of Heat and Mass Transfer 50 (2007) 2217–2235

Nomenclature

A constant in Eq. (14) Pr surface tension

Av constant in Eq. (18) qconv heat loss by convection
B magnetic flux vector qevap heat loss by evaporation
B0 constant in Eq. (14) qlaser heat flux by laser irradiation
Bh magnetic flux density qrad heat loss by radiation
cp specific heat of metal qr radiation heat flux vector
cpl specific heat of plasma r–z cylindrical coordinate system
c the speed of light rf laser beam radius
C coefficient in Eqs. (2) and (3) rfo laser beam radius at the focal position
e charge of electron R gas constant
Ei ionization potential for neutral atom Rb radius of base metal
fa sulfur concentration in base metal ~
s vector tangential to local free surface
F volume of fluid function t time
g gravitational acceleration T temperature of metal
ge degeneracy factor of electron particle T0 reference temperature
gi degeneracy factor of ion particle TK plasma temperature outside of Knudsen layer
g0 degeneracy factor of neutral atom Tpl temperature of plasma
g quantum mechanical Gaunt factor T pl average plasma temperature
h enthalpy for metal Tw surface temperature of the liquid metal at the
hpl enthalpy for plasma keyhole wall
hconv convective heat transfer coefficient T1 ambient temperature
h Planck’s constant u velocity in r-direction
Hv latent heat for liquid–vapor U variable defined in Eq. (15)
Hb thickness of base metal v velocity in z-direction
I(r, s) total directional radiative intensity V velocity vector
Ib black body emission intensity Vr relative velocity vector (Vl  Vs)
Ic(r) collimate incident laser beam intensity distribu- W melt mass evaporation rate
tion on the focus plane Z charge of ion
I0(r, z) incident intensity from laser beam
Ir,m(r, z) incident intensity from mth reflection Greek symbols
J electric current flux vector / scalar electrical potential
jr electric current density in r-direction aFr Fresnel absorption coefficient
jz electric current density in z-direction bT thermal expansion coefficient
ka Planck mean absorption coefficient oc/oT surface tension temperature gradient
kb Boltzmann constant e surface radiation emissivity
kpl thermal conductivity of plasma e0 dielectric constant
k thermal conductivity of metal ef constant in Eq. (20)
K permeability function in Eqs. (2) and (3) c surface tension coefficient
Kpl inverse Bremsstrahlung (IB) absorption coeffi- cr specific heat ratio
cient x angular frequency of laser irradiation
ma atomic mass X solid angle
me electron mass j free surface curvature
mv variable defined in Eq. (42) ll dynamic viscosity
m constant in Eq. (24) r Stefan–Boltzmann constant
MK Mach number at the outer of the Knudsen layer re electrical conductivity
~
n vector normal to local free surface / angle of incident laser light
ne electron density in plasma q density of metal
ni ion density in plasma qpl density of plasma
n0 neutral particle density in plasma s~s Marangoni shear stress
Na Avogadro’s number
p pressure in liquid metal Subscripts
Plaser laser power 0 initial value
Pr recoil pressure c original incident laser light
 J. Zhou, H.-L. Tsai / International Journal of Heat and Mass Transfer 50 (2007) 2217–2235 2219

l liquid phase pl plasma

r relative to solid velocity s solid phase
(r, m) mth reflected laser beam

welding. Their studies indicate that porosity formation is effectiveness of the electromagnetic force on porosity pre-
affected by two competing factors. One is the solidification vention and to determine the desired strength and duration
rate of the molten metal and the other is the back filling of the electromagnetic force to achieve quality welds.
speed of the molten metal during the keyhole collapse pro-
cess. Porosity was found in the final weld when the solidi-
fication rate of the molten metal exceeds the back filling 2. Mathematical model
speed of liquid metal. Porosity formation has a close rela-
tionship with the depth-to-width aspect ratio of the keyhole Fig. 1 shows a schematic sketch of a pulsed laser keyhole
in pulsed laser keyhole welding. The larger the aspect ratio, welding process. A control volume method employing the
the easier the porosity and the larger the size of the pore/ volume of fluid (VOF) technique [22] and the continuum
void is formed. To reduce/eliminate the pore/void in the formulation [23] is used to calculate the momentum and
weld, a method was proposed to delay the solidification energy transport in the weld pool. The VOF technique
process by controlling the pulse shape of the laser irradia- can handle a transient deformed weld pool surface, while
tion [7,10,12]. This method prolongs the solidification rate the continuum formulation can handle fusion and solidifi-
and is effective in preventing/eliminating porosity for med- cation for the liquid region, the mush zone and the solid
ium depth-to-width aspect ratio laser keyhole welding. region. Plasma in the keyhole is treated as the vapor of
However, it failed for a keyhole with large depth-to-width weld material. Although the velocity and pressure change
aspect ratio [10,12]. dramatically across the Knudsen layer, the generic transla-
In this study, a new method is proposed to control the tion vapor flow along the keyhole is neglected [24] and, in
back filling melt flow during the keyhole collapse process. the present study, only the temperature distribution is
This can be achieved by increasing the back filling speed considered. Meanwhile, the pressure along the keyhole is
of the molten metal via the application of an electromag- considered to be approximately constant [25] and is compa-
netic force (Lorentz force). Electromagnetic force has been rable to the atmospheric pressure. Note in high power laser
found to be able to affect melting and solidification pro- welding (P8 kW), the plasma plume and its velocities in
cesses [13–15] and has been used in arc welding to change the keyhole can be very significant [4,5] and, hence, the
flow conditions [16–19]. In arc welding, when an electric assumption of no plasma flow in the present study is
current flows through the workpiece, it will interact with a limited to low power laser keyhole welding.
magnetic field to generate an electromagnetic force. This
electromagnetic force can change the momentum of the 2.1. Metal zone simulation
melt flow and to avoid weld slag or droplet detachment
under the effect of gravitation when welding in a ‘‘wall 2.1.1. Governing equations
position” [16,17]. In gas metal arc welding (GMAW), tem- The governing differential equations used to describe
perature and velocity fields, weld pool geometry and ther- the heat and mass transfer and fluid flow in a cylindrical
mal cycles were also observed to be strongly influenced by coordinate (r–z) system given by Chiang and Tsai [23] are
electromagnetic force [18]. Recently, experimental studies modified and used in the present study:
have been conducted to investigate the usage of electromag-
netic force in laser welding [20,21]. Electromagnetic force Continuity
was found to be able to affect the melt flow and weld pool
o
dynamics in laser welding to achieve good quality welds. ðqÞ þ r  ðqVÞ ¼ 0 ð1Þ
Although experimental observations can provide some ot
useful information, the underlying physics, such as temper- Momentum
ature and velocity evolutions during the keyhole formation
and collapse cannot be revealed. This is especially true for o
ðquÞ þ r  ðqVuÞ
the phenomena that occur inside the keyhole. However, the ot  
temperature and velocity evolutions during the keyhole q op ul q
collapse and the solidification processes play a significant ¼ r  ll ru   ðu  us Þ
ql or K ql
role on the formation of porosity. In this paper, mathemat- Cq2
ical models are developed to study the interplay between  0:5 ju  us jðu  us Þ  r  ðqfs fl V r ur Þ
the electromagnetic force and the keyhole dynamics, melt K ql
  
flow and heat transfer in pulsed laser keyhole welding. q
þ r  ls ur þ J  Bjr ð2Þ
Parametric studies are also conducted to investigate the ql
2220 J. Zhou, H.-L. Tsai / International Journal of Heat and Mass Transfer 50 (2007) 2217–2235

Shielding gas Laser light

F A B

Plasma zone

Hb
Solid – liquid
interface
E
Metal zone

r
D C

Rb

Fig. 1. Schematic sketch of the pulsed laser keyhole welding process.

o where jr and jz are the current density in r- and z-direction,

ðqvÞ þ r  ðqVvÞ
ot   respectively. In order to have the flexibility of controlling
q op ul q the magnitude and direction of the resulting electromag-
¼ qg þ r  ll rv   ðv  vs Þ
ql oz K ql netic force, an electrical potential and a magnetic field
Cq2 are applied independently. Based on our previous studies
 jv  vs jðv  vs Þ  r  ðqfs fl V r vr Þ [27], since the self-induced magnetic field intensity caused
K 0:5 ql
   by the external electrical current is much smaller than the
q externally applied magnetic field intensity, its effect is ne-
þ r  ls vr þ qgbT ðT  T 0 Þ þ J  Bjz ð3Þ
ql glected. Hence, Bh in Eq. (5) represents the externally ap-
plied magnetic flux density. The electrical current density
Energy
can be calculated in the following:
   
o k k
ðqhÞ þ r  ðqVhÞ ¼ r  rh  r  rðhs  hÞ Conservation of current
ot cp cp    
1 o o/ o o/
 r  ðqðV  V s Þðhl  hÞÞ ð4Þ re r þ re ¼0 ð6Þ
r or or oz oz
The physical meaning of each term appearing in the above
where re is electrical conductivity and / is electrical poten-
equations can be found in Ref. [23]. In Eqs. (1)–(4) the con-
tial. According to Ohm’s law, the electrical current density
tinuum density, specific heat, thermal conductivity, solid
in r- and z-direction is defined as
mass fraction, liquid mass fraction, velocity and enthalpy
are defined in Ref. [26]. The last term in Eqs. (2) and (3) o/ o/
jr ¼ re ; jz ¼ re ð7Þ
represents the electromagnetic force in the r- and z-direc- or oz
tion, respectively. Before electromagnetic force is applied,
these two items are zero and they are treated as body forces
after the electromagnetic force is applied. The calculation 2.1.2. Tracking of free surfaces
of electromagnetic forces is in the following: The algorithm of volume-of-fluid (VOF) is used to track
the dynamics of free surfaces [22]. The fluid configuration is
J  Bjr ¼ jr Bh ; J  Bjz ¼ jz Bh ð5Þ defined by a volume of fluid function, F(r, z, t), which tracks
 J. Zhou, H.-L. Tsai / International Journal of Heat and Mass Transfer 50 (2007) 2217–2235 2221

the location of free surface. The function F takes the value empty cell. Cells with F values between zero and one are
of one for the cell full of fluid and the value of zero for the partially filled with fluid and identified as surface cells.
The function F is governed by the following equation:
Table 1 dF oF
Boundary conditions (as shown in Fig. 3) for Eq. (6)
¼ þ ðV  rÞF ¼ 0 ð8Þ
dt ot
Boundaries AB BC CD DE EA
o/ I o/ o/ o/ 2.1.3. Boundary conditions
BCs re ¼ ¼0 ¼0 /=0 ¼0
oz pR22 oz or or The boundaries of the metal zone simulation are divided
Where R2 is the radius of the electrode as shown in Fig. 3. into five segments, as shown in Fig. 1.

Table 2
Thermophysical properties of 304 stainless steel and process parameters
Nomenclature Value
Constant in Eq. (14), A (Pa) 0.55
Constant in Eq. (18), Av 2.52
Vaporization constant in Eq. (14), B0 1.78  1010
Speed of light, c (m s1) 3  108
Specific heat of solid phase, cs (J kg1 K1) 700
Specific heat of liquid phase, cl (J kg1 K1) 780
Specific heat of plasma, cpl (J kg1 K1) 49.0
Charge of electron, e (C) 1.6022  1019
Ionization potential for neutral atoms, Ei (J) 1.265  1018
Sulfur concentration in base metal, f a (ppm) 100
Gravitational acceleration, g (m s2) 9.8
Degeneracy factors for electrons, ge 30
Degeneracy factors for ions, gi 30
Degeneracy factors for neutral atoms, g0 25
Quantum mechanical Gaunt factor, g 1.5
Convective heat transfer coefficient, hconv (W m2 K1) 80
Planck’s constant, h (J s) 6.625  1034
Latent heat of fusion, H (J kg1) 2.47  105
Thickness of substrate metal, Hb (mm) 3.0
Latent heat of vaporization, Hv (J kg1) 6.34  106
Boltzmann constant, kb (J K1) 1.38  1023
Thermal conductivity of liquid phase, kl (W m1 K1) 22
Thermal conductivity of plasma, kpl (W m1 K1) 3.74
Thermal conductivity of solid phase, ks (W m1 K1) 22
Atomic mass, ma (g) 9.3  1023
Electron mass, me (g) 9.1  1028
Mach number at the outer of the Knudsen layer, MK 1.2
Avogadro’s number, Na (mol1) 6.022  1023
Laser power, Plaser (W) 1700
Laser beam radius, rf (mm) 0.25
Laser beam radius at focus, rf 0 (mm) 0.25
Gas constant, R (J kg1 mol1) 8.3  103
Radius of substrate metal, Rb (mm) 20.0
Liquidus temperature, Tl (K) 1780
Reference temperature, T0 (K) 1700
Solidus temperature, Ts (K) 1670
Ambient temperature, T1 (K) 300
Average ionic charge in the plasma, Z 1
Thermal expansion coefficient, bT (K1) 4.95  105
Surface radiation emissivity, e 0.4
Dielectric constant, e0 14.2
Constant in Eq. (20), ef 0.2
Specific heat ratio, cr 1.67
Angular frequency of laser radiation, x (rad s1) 1.78  1014
Dynamic viscosity, ll (kg m1 s1) 0.006
Stefan–Boltzmann constant, r (W m2 K4) 5.67  108
Electrical conductivity, re (X1 m1) 7.14  105
Density of liquid phase, ql (kg m3) 6900
Density of plasma, qpl (kg m3) 0.06
Density of solid phase, qs (kg m3) 7200
2222 J. Zhou, H.-L. Tsai / International Journal of Heat and Mass Transfer 50 (2007) 2217–2235

2.1.3.1. Top surface inside the keyhole (AE in Fig. 1). For ficient c can be calculated as a function of temperature T
cells containing free surface, that is, cells that contain fluid and sulfur concentration f a [29]:
but have one or more empty neighbors, in the direction
c ¼ 1:943  4:3  104 ðT  1723Þ  RT  1:3  108
normal to the free surface, the following pressure condition   
must be satisfied [28]: 1:66  108
 ln 1 þ 0:00318f a exp ð12Þ
P ¼ Pr þ Pr ð9Þ RT

where P is the pressure at the free surface in a direction In this study, the sulfur concentration is assumed to be con-
normal to the local free surface. Pr is the surface tension stant, and the temperature-dependent Marangoni shear
and Pr is the recoil pressure. Pr is calculated by the follow- stress on the free surface in the direction tangential to the
ing formula: local surface is given by [30]:
oðV ~
sÞ oc oT
P r ¼ jc ð10Þ s~s ¼ ll ¼ ð13Þ
o~
n oT o~
s
where j is the free surface curvature, given by [26]:
Calculation of the evaporation-induced recoil pressure Pr is
     
~
n 1 ~
n complicated by the existence of a Knudsen layer over the
j¼ r ¼  r j~
nj  ðr  ~
nÞ ð11Þ vaporizing surface. Based on Knight’s model [31], the recoil
j~
nj j~
nj j~nj
pressure can be calculated by [32]:
where ~
n is the unit vector normal to the local free surface. pffiffiffiffiffiffi
For a pseudo-binary Fe-S system, the surface tension coef- P r ¼ AB0 = T w expðU =T w Þ ð14Þ

Fig. 2. A figure shows the porosity formation process in pulsed laser keyhole welding [12] (laser power is 1.7 kW and pulse duration is 15.0 ms).
 J. Zhou, H.-L. Tsai / International Journal of Heat and Mass Transfer 50 (2007) 2217–2235 2223

where A is the numerical coefficient and B0 is the vaporiza- The laser heat flux qlaser comes from the Fresnel absorption
tion constant. The coefficient A depends on the ambient of the incident intensity directly from the laser beam plus
pressure and its value varies from 0.55 for evaporation in the incident intensity from the multiple reflections:
the vacuum to 1 for the case of evaporation under a high X
n
ambient pressure. For atmospheric pressure, the coefficient qlaser ¼ I 0 ðr; zÞaFr ðu0 Þ þ I r;m ðr; zÞaFr ðum Þ ð19Þ
A is close to its minimal value of 0.55. B0 is at the value of m¼1
!
1.78  1010. Tw is the surface temperature of the liquid 2
1 1 þ ð1  ef cos uÞ e2f  2ef cos u þ 2 cos2 u
metal on the keyhole wall. The parameter U is defined as aFr ðuÞ ¼ 1  þ
2 1 þ ð1 þ ef cos uÞ2 e2f þ 2ef cos u þ 2 cos2 u
follows [32]:
ð20Þ
U ¼ ma H v =ðN a k b Þ ð15Þ
where ma is the atomic mass, Hv is the latent heat of evap- where / is the angle of the incident light with the normal of
oration, Na is the Avogadro’s number and kb is the Boltz- the keyhole surface, n is the total number of incident light
mann constant. from multiple reflections. ef is a material-dependent coeffi-
The energy on the top free surface is balanced between cient. In CO2 laser welding of mild steel, ef = 0.2 is used.
the laser irradiation, plasma-keyhole wall radiation, the I0(r, z) and Ir,m(r, z) are, respectively, the incident intensity
heat dissipation through convection, and metal vapori- from the laser beam and the mth multiple reflection at
zation. In general, since the velocity of the plume along the keyhole wall which are given as
 Z z0 
the surface is assumed to be zero [24], the heat loss due
to convection is omitted. The energy balance is given by I 0 ðr; zÞ ¼ I c ðrÞ exp  K pl dz ð21Þ
0
the following formula:  Z zm 
oT I r;m ðr; zÞ ¼ I r ðr; zÞ exp  K pl dz ð22Þ
k ¼ qlaser þ qrad  qevap ð16Þ 0
o~
n I r ðr; zÞ ¼ I 0 ðr; zÞð1  aFr Þ ð23Þ
In this study, the liquid/vapor evaporation model is used
due to the low intensity of laser irradiation. The heat loss where Ic(r) stands for the original collimated incident laser
due to surface evaporation can be written as [33] beam intensity, Ir,m(r, z) is the
R z reflected laser
R z beam intensity
at the m times reflections, 0 0 K pl dz and 0 m K pl dz are the
qevap ¼ WH v ð17Þ optical thickness of the laser transportation path, respec-
  tively, for the first incident and the multiple reflections,
18836 and Kpl is the plasma absorption coefficient due to the
logðW Þ ¼ Av þ 6:121   0:5 log T ð18Þ
T inverse Bremsstrahlung (IB) absorption [34]

Electrode
(Anode) A
•
C B
× × × • • •
jr
× × × • • • →
× × × • • •
B jz
× × × • • •

Current Fr
Fz

Fr = -jz Bθ
Fz = jr Bθ

zz
Workpiece
(Cathode)
r
D E
Fig. 3. A figure used to explain the direction and magnitude of electromagnetic force.
2224 J. Zhou, H.-L. Tsai / International Journal of Heat and Mass Transfer 50 (2007) 2217–2235

3.5 where n0 is neutral particle density which is 1026/cm3 for

iron [35], ge, gi and g0 are, respectively, the degeneracy fac-
3 tors for electrons, ions and neutral atoms, Ei is the ioniza-
tion potential for the neutral atoms in the gas. Assuming
2.5 the laser intensity distribution, Ic(r), is ideal Gaussian-like
and can be written as [36]
 2  
2 2P laser rf 2r2
z (mm)

I c ðrÞ ¼ exp  ð26Þ

pr2fo rfo r2f
1.5
where rf is the beam radius, rfo is the beam radius at the fo-
cal position, and Plaser is the laser power. In laser welding,
1
the keyhole surface temperature is much lower than that of
the plasma, so the radiation and emission of the surface can
0.5
102 (A/cm2) be omitted. Then qrad can be simplified as

0 qrad ¼ erðT pl 4  T 4 Þ ð27Þ

2 1 0 1 2
r (mm) where T pl is the average temperature of keyhole plasma.
3.5
2.1.3.2. Top surface outside the keyhole (AB in Fig. 1).
Boundary condition on the top surface outside the keyhole
3 is similar to that inside the keyhole. The differences lie in
the absence of plasma and multiple reflections. As shown
2.5 in Fig. 1, there is a shielding gas flow above the base metal,
which means that plasma outside the keyhole will be blown
away. So Eq. (19) can be written as
2
z (mm)

qlaser ¼ I 0 ðr; zÞaFr cos u ð28Þ

1.5 Since there is no plasma and the temperature of shielding
gas is much lower than that of the metal surface, the radi-
1 ation heat flux can be given as
qrad ¼ erðT 4  T 41 Þ ð29Þ
0.5 4 2
10 N/cm Here, T1 is the ambient temperature. Since there is a
shielding gas flow over the surface, the convection heat
2 1 0 1 2 loss cannot be omitted which is given by
r (mm) qconv ¼ hconv ðT  T 1 Þ ð30Þ
Fig. 4. The calculated (a) current density vector (b) electromagnetic force
vector distributions at t = 19.2 ms (I = 30 A and Bh = 300 mT).
2.1.3.3. Side surface (BC in Fig. 1).
 0:5   
ne ni Z 2 e6 2p me x oT
K pl ¼ pffiffiffi 3 1  exp  g k ¼ qconv ð31Þ

6 3me0 chx me 3 2 2pk b T pl k b T pl or
ð24Þ u ¼ 0; v ¼ 0 ð32Þ

where Z is the charge of ion in the plasma, e is the charge of

electron, x is the angular frequency of the laser irradiation, 2.1.3.4. Bottom surface (CD in Fig. 1).
e0 is the dielectric constant, ne and ni are the densities of oT
electrons and ions, respectively,  h is the Planck’s constant, k ¼ qconv ð33Þ
oz
m is a constant that is related to the specific laser being
u ¼ 0; v ¼ 0 ð34Þ
used and is 1 for CO2 laser, me is the electron mass, Tpl
is the plasma temperature, c is the speed of light, and g is
the quantum mechanical Gaunt factor. For weakly ionized 2.1.3.5. Symmetrical axis (DE in Fig. 1).
plasma in the keyhole, Saha equation [35] can be used to
calculate the densities of the plasma species: oT
¼0 ð35Þ
1:5   or
ne ni ge gi ð2pme k b T pl Þ Ei ov
¼ 3 exp  ð25Þ u ¼ 0; ¼0 ð36Þ
n0 g0 h k b T pl or
 J. Zhou, H.-L. Tsai / International Journal of Heat and Mass Transfer 50 (2007) 2217–2235 2225

The boundary conditions for the calculation of Eq. (6) are the thermal conductivity and specific heat of the plasma.
listed in Table 1. qr stands for the radiation heat flux vector. Note hpl =
cplTpl.
2.2. Plasma zone simulation The radiation source term $  (qr) is defined as
 Z 
2.2.1. Governing equations r  qr ¼ k a 4pI b  I dX ð38Þ
In current study, metal vapor in the keyhole is assumed 4p
to be a compressible, inviscid ideal gas. Since the heat pro-
duction by viscous dissipation is rather small in laser weld- where ka, Ib and X denote the Planck mean absorption
ing, the energy equation can be simplified as [37]: coefficient, blackbody emission intensity ðI b ¼ rT 4pl Þ and
  solid angle, respectively. When an intense laser pulse inter-
o k pl
ðqpl hpl Þ ¼ r  rhpl  qr acts with the vapor in the keyhole, a significant amount of
ot cpl
 Z z0  laser irradiation is absorbed by the ionized particles
þ K pl I c ðrÞ exp  K pl dz through the IB absorption. For simplicity, the plasma is
0 assumed to be an absorbing-emitting medium and the
X n  Z zm 
scattering effect is neglected. The radiation transport
þ K pl I r;m ðr; zÞ exp  K pl dz ð37Þ equation (RTE) has to be solved for the total directional
m¼1 0
radiative intensity I(r, s) [38]
where hpl and qpl represent, respectively, the enthalpy and
density of the plasma, kpl and cpl represent, respectively, ðs  rÞIðr; sÞ ¼ k a ðI b  Iðr; sÞÞ ð39Þ

Fig. 5. The liquid metal evolution for a medium depth-to-width ratio keyhole.
2226 J. Zhou, H.-L. Tsai / International Journal of Heat and Mass Transfer 50 (2007) 2217–2235

where s and r denote a unit vector along the direction of discontinuous, which can be calculated by the following
the radiation intensity and the local position vector. When formula [31]:
the plume within the keyhole is weakly ionized, the 2sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 32
absorption mechanism mainly depends on electron-neutral  2
TK 4 cr  1 m v pffiffiffi cr  1 mv
interaction and the plume behaves as an optically thin ¼ 1þp  p 5 ð41Þ
Tl cr þ 1 2 cr þ 1 2
medium. For the evaluation of the intensity and heat flux
divergence, the Planck mean absorption coefficient is given sffiffiffiffi
2
as [38] mv ¼ M k ð42Þ
cr
 0:5  1:5
128 p Z 2 e6 g ne ni
ka ¼ kb ð40Þ
27 me hrc3 T 3:5 where TK is the vapor temperature outside of the Knudsen
pl
layer, Tl is the liquid surface temperature adjacent to the
Knudsen layer, Mk is Mach number at the outer of the
2.2.2. Boundary conditions Knudsen layer and cr is the specific heat ratio. The value
2.2.2.1. Bottom surface inside the keyhole (EA in Fig. 1). of mv depends on the gas dynamics of the vapor flow away
Close to the liquid wall inside the keyhole, there is a so- from the surface. Mk = 1.2 is used in the present study [31].
called Knudsen layer where vaporization of material takes The vapor is assumed to be iron in the form of monatomic
place. The vapor temperature across the Knudsen layer is gas with molecular weight of 56 and cr = 1.67. The gas tem-

t = 15.0 (ms) t =19.8 (ms)

3 3

2 2

1 2.0 m/s 1 1.0 m/s

t = 18.6 (ms) t =20.4 (ms)

3 3

2 2

1 2.0 m/s 1 1.0 m/s

t = 19.2 (ms) t = 22.8 (ms)

3 3
z (mm)
z (mm)

2 2

1 2.0 m/s 1 1.0 m/s

1 0 1 1 0 1
r (mm) r (mm)

Fig. 6. The corresponding velocity distributions as shown in Fig. 5.

 J. Zhou, H.-L. Tsai / International Journal of Heat and Mass Transfer 50 (2007) 2217–2235 2227

500 1. Eqs. (1)–(4) are solved iteratively for the metal zone to
obtain velocity, pressure and temperature distributions
using the associated boundary conditions.
400
Case III 2. Eq. (37) is solved iteratively to obtain the plasma tem-
Magnetic flux density Bθ (mT)

perature distributions in the keyhole under the associ-

ated boundary conditions. The steps for solving Eq.
Case II (37) are listed below:
300
(a) Solve Eq. (39) using the associated boundary condi-
tions to get the total directional radiative intensity
Case I
distributions.
200 (b) Solve Eq. (38) to get radiation source term $  (qr).
(c) Solve Eqs. (25) and (24) in the order using the most
recent plasma temperature from the previous time
100 step to get the updated plasma absorption coefficient
Kpl.
(d) Solve Eq. (37) to get the updated plasma
0 temperature.
15 20 25 30 3. Solve VOF algorithm equation (8) to obtain the new
Time (ms)
domain for the metal and plasma zones.
Fig. 7. Magnetic flux used in porosity prevention for a large depth-to- 4. Update boundary conditions for the metal and the
width ratio keyhole.
plasma zones.
5. Eq. (6) and the associated boundary conditions are
perature outside the Knudsen layer is used as the boundary
solved to obtain current density in the r- and z-direc-
temperature. So the boundary condition is given by [38]:
tions. Then, Eq. (7) and Eq. (5) were solved for electro-
T pl ¼ T K ð43Þ magnetic force.
Z
1e 6. Advance to the next time step and back to Step 1 until
I ¼ eI b þ n  Xj dX0
Ij~ ð44Þ the desired time is reached.
p nX0 <0
~

In the calculation, before electromagnetic force is

applied, there is no electromagnetic force term in Eqs. (2)
2.2.2.2. Top surface outside the keyhole (FA in Fig. 1).
and (3) and, hence, the calculation of electromagnetic
T pl ¼ T 1 ð45Þ force, Step 5, is skipped.
The techniques for solving Eqs. (1)–(4) and (37) are
I ¼ I c ðrÞ ð46Þ
given by Wang and Tsai [26]. Following the MAC scheme,
the r- and z-velocity components are located at cell
face centers on lines of constants r and z, respectively; while
2.2.2.3. Symmetrical axis (EF in Fig. 1). the pressure, VOF function, temperature and absorbed
oT pl laser flux are located at cell centers. Since the temperature
¼0 ð47Þ and pressure field change more dramatically near the
or
oI keyhole, a non-uniform grid system with 202  252
¼0 ð48Þ points is used for the total computational domain of
or
5.0 mm  20.0 mm, in which smaller grids are concentrated
near the keyhole and larger grids for other parts. Due to
the axis-symmetry of the domain, only half of the grid
3. Numerical method points were used in the calculation. Calculations were exe-
cuted on the DELL OPTIPLEX GX270 workstations with
The solutions of transport equations in the metal zone LINUX-REDHAT 9.0 OS and it took about 6 h of CPU
and in the plasma zone are coupled; that is, the simulations time to simulate about 100 ms of real-time welding. The
of the metal and the plasma zone provide boundary condi- average time step is 104 s and the smallest time step is
tions for each other. However, there are large spatial and about 106 s.
physical differences between the metal and the plasma zone.
To enhance convergence rate and save calculation time,
different time and space resolutions are used for the metal 4. Results and discussion
and the plasma zone. The governing equations (Eqs. (1)–
(4), (6) and (37)) and all related supplemental equations The base metal is assumed to be 304 stainless steel. The
and boundary conditions are solved through the following process parameters and thermophysical properties used in
iterative scheme: the present study are summarized in Table 2. The laser
2228 J. Zhou, H.-L. Tsai / International Journal of Heat and Mass Transfer 50 (2007) 2217–2235

Fig. 8. The liquid metal evolution for a large depth-to-width ratio keyhole; Case I: small electromagnetic force.

energy is assumed to be in Gaussian distribution and the ing discussion. As shown in the figure at t = 28.0 ms, the
focus plane is on the top surface of the base metal. The back filling liquid metal from the top is blocked by the
laser irradiation duration varies in different study cases. solidification near the bottom of the keyhole. Hence, in
As the present study focuses on how the keyhole is back order to eliminate porosity, the back filling speed of the
filled during the keyhole collapse process, the keyhole for- liquid metal must be high enough to reach the bottom of
mation process is omitted and the temperature distribu- the keyhole before complete solidification. This can be
tions of the liquid metal and the plasma in the keyhole achieved by exerting an external force to accelerate the
are also omitted and they can be found in Ref. [11]. back filling speed of the liquid metal. It is well known that
when an electric field interacts with a magnetic field, an
4.1. Application of electromagnetic force in laser keyhole electromagnetic body force (Lorentz force) will be gener-
welding ated. Since steel is an electrically conductive material, elec-
tromagnetic body force J  B can be introduced in laser
Porosity formation mechanism and its prevention by welding to ‘‘stir” the weld pool to increase the back filling
pulse control have been discussed before [8,9,12]; hence, speed of the liquid metal during the keyhole collapse
the following discussion will be focused on the studies of process.
increasing the back filling speed of melt flow to prevent This idea is briefly shown in Fig. 3. In the figure, jr and jz
porosity formation in pulsed laser welding. Fig. 2 shows are, respectively, the electric current field distribution in the
a typical porosity formation process in pulsed laser welding r- and z-direction; Bh is the applied external magnetic flux;
which is reproduced from Ref. [12] to facilitate the follow- and Fr and Fz are, respectively, the electromagnetic forces
 J. Zhou, H.-L. Tsai / International Journal of Heat and Mass Transfer 50 (2007) 2217–2235 2229

in the r- and z-direction. In the model, an external electric in Fig. 4(b). This downward and inward electromagnetic
current J is applied via an electrode above the workpiece force can increase the downward and inward momentum
when the laser irradiation is shut off. The two components of the liquid metal and is helpful in eliminating porosity
of current, jr and jz are, calculated via Eq. (7). Since the discussed in the following.
self-induced magnetic field generated by this electric cur-
rent is relatively small (as compared to the externally 4.2. Porosity prevention by electromagnetic force
applied magnetic field), in order to intensify the electro-
magnetic body force in the weld pool, a DC electromagnet 4.2.1. Medium depth-to-width aspect ratio keyhole
is employed to generate a homogeneous magnetic field Bh For the same welding conditions used to produce the
in the direction as shown in Fig. 3. The magnitudes and results shown in Fig. 2, an external current 30 A and a con-
durations of the electric current and magnetic field vary stant magnetic flux Bh = 300 mT are applied as soon as the
in different study cases. Fig. 4 shows typical calculated elec- laser irradiation is shut off (at 15 ms) and they last for 5 ms.
tric current density distribution and electromagnetic force Fig. 5 shows the liquid metal evolution during the keyhole
distribution in a pulsed laser keyhole welding process at collapse process, and the corresponding velocity distribu-
t = 19.2 ms and I = 30 A and Bh = 300 mT. As shown in tion is shown in Fig. 6. Note in order to increase readabil-
Fig. 4(a), the electric current is flowing downward and out- ity, only one-half of the grid nodes are used for plotting
ward in the base metal which will result in an inward and the velocity distribution. At t = 15.0 ms, after the laser
downward electromagnetic force as shown in Fig. 3. This power is shut off, the recoil pressure is gone and the liquid
is verified by the electromagnetic force distribution shown metal near the top of the keyhole has a tendency to flow

t = 18.0 (ms) t = 25.8 (ms)

3 3

2 2

1 2.0 m/s 1 1.0 m/s

t = 19.2 (ms) t =28.8 (ms)

3 3

2 2

1 2.0 m/s 1 1.0 m/s

t = 20.4 (ms) t = 31.8 (ms)

3 3
z (mm)
z (mm)

2 2

1 2.0 m/s 1 1.0 m/s

1 0 1 1 0 1
r (mm) r (mm)

Fig. 9. The corresponding velocity distributions as shown in Fig. 8.

2230 J. Zhou, H.-L. Tsai / International Journal of Heat and Mass Transfer 50 (2007) 2217–2235

downward to fill the keyhole under the action of gravity bottom of the keyhole than that on the shoulder, the effect
and surface tension. As the electromagnetic force is down- of electromagnetic force on the liquid metal near the bot-
ward and inward, it will enhance the gravity and surface tom is not as remarkable as on that near the top of the key-
tension and shorten the time for the liquid metal to backfill hole. Hence, the application of electromagnetic force has
the keyhole. As shown in Figs. 5 and 6 from t = 15.0 ms to little influence on the solidification rate of the liquid metal
t = 18.6 ms, the downward and inward velocities of the near the bottom of the keyhole. By comparing Figs. 5 and
melt flow in the keyhole change more quickly than those 2, at t = 19.8 ms, it is seen the liquid metal still quickly
without electromagnetic force, shown in Fig. 2. This is solidifies near the bottom of the keyhole even when electro-
especially true for the liquid metal near the shoulder of magnetic force is being applied. However, the downward
the keyhole; the inward electromagnetic force ‘‘sweeps” velocity of liquid metal continues to be accelerated by
the molten metal surrounding the top of the keyhole into electromagnetic force and the bottom of the keyhole at
the keyhole. Comparing Figs. 5 and 2, it is seen that the t = 20.4 ms is filled before complete solidification. Hence,
keyhole is filled earlier for the case with the applied electro- at t = 22.8 ms, there is no pore/void formed in the final
magnetic force. weld.
As shown in Fig. 5, the liquid–solid interface (i.e., solid-
ification front) moves inward and upward from the keyhole 4.2.2. Large depth-to-width aspect ratio keyhole
in the metal after the laser irradiation is terminated. As the As discussed before, the formation of porosity in laser
electromagnetic force is greater near the top of the keyhole welding is strongly related to the depth-to-width aspect
(see Fig. 4(b)) and much less liquid metal exists near the ratio of the keyhole [12]. The larger the ratio, the easier

Fig. 10. The liquid metal evolution for a large depth-to-width ratio keyhole; Case II: medium electromagnetic force.
 J. Zhou, H.-L. Tsai / International Journal of Heat and Mass Transfer 50 (2007) 2217–2235 2231

the porosity can be formed. In order to prevent the poros- Figs. 8 and 9. In this case, the electromagnetic force is
ity at the root of the keyhole in the large depth-to-width applied at 18.0 ms right after the turn-off of the laser power
aspect ratio laser keyhole welding process, the back filling and is held for 7.0 ms. Similar to that shown in Fig. 4(b),
speed of the liquid metal must be accelerated faster than the direction of the electromagnetic force is inward and
that in the medium ratio pulsed laser keyhole welding pro- downward.
cess. In the following, the effects of the strength of electro- Under the action of electromagnetic force, the liquid
magnetic force and its duration on melt flow and porosity metal on the top flows inward and downward to quickly
formation will be discussed. The electric current used in refill the keyhole after the shut-off of the laser beam, as
these study cases are all at 30 A, while the magnitude and shown in Fig. 9. In this case, the keyhole is deeper, which
duration of the external magnetic flux vary as shown in means it requires a longer time for the liquid metal from
Fig. 7. Based on the relative magnitude of the magnetic the top to reach the bottom of the keyhole. Since the mag-
flux, in the following discussion we designate small, med- nitude of electromagnetic force is relatively small in this
ium, and large electromagnetic force, respectively, for Case case, the back filling speed of the liquid metal could not
I, Case II, and Case III. be accelerated high enough to reach the bottom of the key-
hole before complete solidification. This process is shown
Case I: Small electromagnetic force. The liquid metal evolu- in Fig. 8. As shown, the bottom of the back filling liquid
tion and corresponding velocity distribution are shown in metal solidifies at the depth of about 1.6 mm. After that,

t = 18.0 (ms) t = 25.8 (ms)

3 3

2 2

1 1.0 m/s 1 0.5 m/s

t = 19.2 (ms) t =34.8 (ms)

3 3

2 2

1 2.0 m/s 1 0.5 m/s

t = 20.4 (ms) t = 76.2 (ms)

3 3
z (mm)
z (mm)

2 2 Fully solidified

1 2.0 m/s 1

1 0 1 1 0 1
r (mm) r (mm)

Fig. 11. The corresponding velocity distributions as shown in Fig. 10.

2232 J. Zhou, H.-L. Tsai / International Journal of Heat and Mass Transfer 50 (2007) 2217–2235

although there is still some liquid metal on the top, it was of the keyhole before complete solidification. As shown
blocked by the solidified metal and could not flow farther in Fig. 10 at t = 25.8 ms, the keyhole was completely filled
to reach the bottom of the keyhole. Hence, a pore/void up by the liquid metal and the pore/void at the root of the
was found at the root of the keyhole, though the size of keyhole was eliminated. Also, as shown in Fig. 10 at
the pore/void is smaller than that shown in the case with- t = 25.8 ms, an extrusion was found in the center of the
out the use of an electromagnetic force. weld pool. This is due to the strong inward push of the
liquid metal by the electromagnetic force. However, under
the action of the hydrostatic force and surface tension, this
Case II: Medium electromagnetic force. In this case, the
squeezed liquid metal is able to flow back to the edge of
magnetic flux Bh is increased to 300 mT and the duration
the keyhole before complete solidification. The completely
of the magnetic field is 5.0 ms, as shown in Fig. 7. The
solidified weld pool is shown by a dashed curve at
liquid metal evolution and the corresponding velocity dis-
t = 76.2 ms.
tribution are shown in Figs. 10 and 11, respectively.
As shown in Fig. 11, since the magnitude of the applied
electromagnetic force is increased, the back filling speed of Case III: Large electromagnetic force. In this case, the mag-
the liquid metal from the top is faster than that shown in netic flux Bh is increased to 400 mT and its duration is
Fig. 9. Due to the strong downward momentum induced decreased to 3.0 ms, as shown in Fig. 7. The liquid metal
by the downward electromagnetic force, the liquid metal evolution and the corresponding velocity distribution are
on the top quickly flows downward and reaches the bottom illustrated in Figs. 12 and 13, respectively. Since the magni-

Fig. 12. The liquid metal evolution for a large depth-to-width ratio keyhole; Case III: large electromagnetic force.
 J. Zhou, H.-L. Tsai / International Journal of Heat and Mass Transfer 50 (2007) 2217–2235 2233

t = 18.0 (ms) t = 25.8 (ms)

3 3

2 2

1 1.0 m/s 1 0.5 m/s

t = 19.2 (ms) t =31.8 (ms)

3 3

2 2

1 2.0 m/s 1 0.5 m/s

t = 20.4 (ms) t = 76.2 (ms)

3 3
z (mm)
z (mm)

2 2 Fully solidified

1 2.0 m/s 1

1 0 1 1 0 1
r (mm) r (mm)

Fig. 13. The corresponding velocity distributions as shown in Fig. 12.

tude of electromagnetic force is greater, as shown in Fig. 13 metal is so strong, it takes a relatively longer time for the
at t = 19.2 ms, the downward velocity of the liquid metal is hydrostatic force and surface tension to even the liquid
greater than that in Fig. 11. This strong electromagnetic metal. At this time, since the solidification proceeds quite
force pushes the liquid metal from the top to flow down- fast, the liquid metal flowing from the center to the edge
ward to fill up the keyhole much faster than those in the quickly solidifies. The final weld shape is indicated by a
previous two cases. As shown in Fig. 12 at t = 20.4 ms, dashed-curve at t = 76.2 ms. It is clearly seen the undercuts
the liquid metal from the top almost completely fills up at the edge and an extrusion in the center of the final weld.
the keyhole. Hence, the pore/void at the bottom of the key- Hence, for the given welding conditions and the resulting
hole caused by rapid solidification after the shut-off of the keyhole as discussed above, the Case II medium electro-
laser power [12] is eliminated. magnetic force can be considered to be the optimum to
Although this greater electromagnetic force can help achieve quality welds.
prevent porosity formation at the root of the keyhole more
easily, it can cause some other welding defects. As shown in 5. Conclusions
Fig. 13, at t = 20.4 ms, the inward velocity of the liquid
metal becomes very strong due to the strong inward push Mathematical models have been developed to investi-
from this large electromagnetic force that squeezes the gate the effects of electromagnetic force on porosity preven-
liquid metal upward, as shown in Fig. 13 at t = 25.8 ms. tion in pulsed laser keyhole welding. The formation of
Even though electromagnetic force is turned off at porosity is strongly related to two competing factors: one
t = 21.0 ms, since the upward momentum of the liquid is the liquid metal solidification rate and the other is the
2234 J. Zhou, H.-L. Tsai / International Journal of Heat and Mass Transfer 50 (2007) 2217–2235

back filling speed of the liquid metal. If the back filling [10] A. Todate, Y. Ueno, M. Katsuki, S. Katayama, A. Matsunawa, YAG
speed of the liquid metal on the upper part of the keyhole laser weldability of carbon steel in CO2 shielding gas, in: Proc. of the
National Meeting of JWS 66, 2000, pp. 144–145.
is not large enough to be able to reach the bottom of the [11] J. Zhou, H.L. Tsai, P.C. Wang, Transport phenomena and keyhole
keyhole, porosity will be found in the final weld. dynamics during pulsed laser welding, ASME J. Heat Transfer 128 (7)
The use of an external electromagnetic force can (2006) 680–690.
increase the back filling speed of the liquid metal during [12] J. Zhou, H.L. Tsai, Porosity formation and prevention in pulsed laser
the keyhole collapse process. It allows the liquid metal welding, ASME J. Heat Transfer, in press.
[13] R. Sampath, N. Zabaras, Adjoint variable method for the thermal
from the upper part of the keyhole to reach the bottom design of eutectic directional solidification processes in an open-boat
of the keyhole in a short period of time. As shown in the configuration, Numer. Heat Transfer Part A: Appl. 32 (6) (2001) 655–
present study, for a medium depth-to-width aspect ratio 683.
laser keyhole welding process, applying an electromagnetic [14] E. Goncalves, M. Faghri, Y. Asako, M. Charmchi, Numerical
force is a very effective way to prevent porosity formation. solution of melting processes for unfixed phase-change material in the
presence of electromagnetically simulated low gravity, Numer. Heat
While for a large depth-to-width aspect ratio laser keyhole, Transfer Part A: Appl. 46 (4) (2004) 343–365.
the strength of the applied electromagnetic force and its [15] E. Goncalves, M. Faghri, Y. Asako, M. Charmchi, Numerical
duration are critical to prevent/eliminate porosity forma- solution of melting in side-heated rectangular enclosure under
tion and to prevent the formation of undercuts. When electromagnetically simulated low gravity, Numer. Heat Transfer
the applied electromagnetic force is too big, although Part A: Appl. 47 (4) (2005) 315–332.
[16] Y. Manabe, H. Wada, H. Kondou, Y. Hiromoto, Y. Kobayashi,
porosity at the root of the keyhole can be eliminated, other International Institute of Welding, IIW–XII, 1997, pp. 1484–1497.
welding defects, such as extrusion in the center and under- [17] G.M. Shelenkov, Weld. Production 3 (1997) 24–25.
cuts at the edge of the final weld, can be formed in final [18] S. Kou, Y.H. Wang, Computer simulation of convection in moving
welds. If an optimal electromagnetic force is applied, arc weld pools, Metall. Trans. 17A (1986) 2271–2277.
porosity can be eliminated and good quality welds can be [19] W.-H. Kim, H.G. Fang, S.-J. Na, Effect of various driving forces on
heat and mass transfer in arc welding, Numer. Heat Transfer Part A:
achieved for a high depth-to-width aspect ratio laser key- Appl. 32 (6) (1997) 633–652.
hole welding process. [20] M. Kern, P. Berger, H. Hugel, Magneto-fluid dynamic control of
seam quality in CO2 laser beam welding, Weld. J. 79 (3) (2000) 72s–
78s.
Acknowledgement [21] D. Lindenau, G. Ambrosy, P. Berger, H. Hugel, Effects of magnet-
ically supported laser beam welding of aluminum alloys, in: ICALEO
It is gratefully acknowledged that this work is partially Sec. A, 2001, pp. 308–318.
supported by the General Motors Corporation. [22] D.B. Kothe, R.C. Mjolsness, M.D. Torrey, Ripple: A Computer
Program for Incompressible Flows with Free Surfaces, LA-12007-
MS, Los Alamos National Laboratory, 1991.
References [23] K.C. Chiang, H.L. Tsai, Shrinkage-induced fluid flow and domain
change in two-dimensional alloy solidification, Int. J. Heat Mass
[1] S. Katayama, N. Seto, J. Kim, A. Matsunawa, Formation mechanism Transfer 35 (1992) 1763–1769.
and reduction method of porosity in laser welding of stainless steel, [24] I. Miyamoto, E. Ohmura, T. Maede, Dynamic behavior of plume and
in: Proc. of ICALEO Sec. G, 1997, pp. 83–92. keyhole in CO2 laser welding, in: Proc. ICALEO Sec. G, 1997, pp.
[2] S. Katayama, N. Seto, J. Kim, A. Matsunawa, Formation mecha- 210–218.
nism and suppression procedure of porosity in high power laser [25] J. Dowden, N. Postacioglu, M. Davis, P. Kapadia, A keyhole model
welding of aluminum alloys, in: Proc. of ICALEO Sec. C, 1998, pp. in penetration welding with a laser, J. Phys. D: Appl. Phys. 20 (1987)
24–33. 36–44.
[3] S. Katayama, A. Matsunawa, Formation mechanism and prevention [26] Y. Wang, H.L. Tsai, Impingement of filler droplets and weld pool
of defects in laser welding of aluminum alloys, in: Proc. of CISFFEL dynamics during gas metal arc welding process, Int. J. Heat Mass
6, vol. 1, 1998, pp. 215–222. Transfer 44 (2001) 2067–2080.
[4] T. Ishide, S. Tsubota, M. Nayama, Y. Shimokusu, T. Nagashima, K. [27] J. Hu, H.L. Tsai, Heat and mass transfer in gas metal arc welding.
Okimura, 10 kW class YAG laser application for heavy components, Part I: the arc, Int. J. Heat Mass Transfer, in press, doi:10.1016/
in: SPIE High-power Lasers in Manufacturing, vol. 3888, Osaka, j.ijheatmasstransfer.2006.08.025.
1999, pp. 543–550. [28] W. Duley, Laser Welding, John Wiley & Sons Inc, 1999.
[5] S. Katayama, N. Seto, M. Mizutani, A. Matsunawa, Formation [29] P. Sahoo, T. DeBroy, M.J. Mcnallan, Surface tension of binary
mechanism of porosity in high power YAG laser welding, in: Proc. of metal-surface active solute systems under conditions relevant to
ICALEO Sec. C, 2000, pp. 16–25. welding metallurgy, Metall. Trans. 19B (1988) 483–491.
[6] N. Seto, S. Katayama, A. Matsunawa, A high-speed simultaneous [30] R.T.C. Choo, J. Szekely, S.A. David, On the calculation of the free
observation of plasma and keyhole behavior during high power CO2 surface temperature of gas-tungsten-arc weld pools from first
laser welding, in: Proc. of ICALEO Sec. E, 1999, pp. 17–19. principles: part II. Modeling the weld pool and comparison with
[7] S. Katayama, S. Kohsaka, M. Mizutani, K. Nishizawa, A. Matsu- experiments, Metall. Trans. 23B (1992) 371–384.
nawa, Pulse shape optimization for defect prevention in pulsed laser [31] C.J. Knight, Theoretical modeling of rapid surface vaporization with
welding of stainless steels, in: Proc. of ICALEO, 1993, pp. 487–497. back pressure, AIAA J. 17 (1979) 519–523.
[8] S. Katayama, Y. Kobayashi, N. Seto, M. Mizutani, A. Matsunawa, [32] V. Semak, A. Matsunawa, The role of recoil pressure in energy
Effect of vacuum on penetration and defects in laser welding, in: Proc. balance during laser materials processing, J. Phys. D: Appl. Phys. 30
of ICALEO Sec. C, 2000, pp. 182–191. (1997) 2541–2552.
[9] S. Tsukamoto, I. Kawaguchi, G. Arakane, Suppression of welding [33] T. Zacharia, S.A. David, J.M. Vitek, Effects of evaporation and
defects in deep penetration CO2 laser welding, in: Proc. of ICALEO temperature-dependent material properties on weld pool develop-
Sec. C, 2000, pp. 7–15. ment, Metall. Trans. 22B (1991) 233–241.
 J. Zhou, H.-L. Tsai / International Journal of Heat and Mass Transfer 50 (2007) 2217–2235 2235

[34] Yu.P. Raizer, Laser-induced Discharge Phenomena, Consultants [37] R. Siegel, J.R. Howell, Thermal Radiation Heat Transfer, third ed.,
Bureau, New York, 1977. Hemisphere Publishing Corp, 1992 (Chapter 13).
[35] L. Spitzer, Physics of Fully Ionized Gases, Interscience Publishers, [38] R. Ho, C.P. Grigoropoulos, J.A.C. Humphrey, Gas dynamics and
New York, 1967. radiation heat transfer in the vapor plume produced by pulsed laser
[36] H. Kogelnik, T. Li, Laser beams and resonators, Appl. Opt. 5 (10) irradiation of aluminum, J. Appl. Phys. 79 (1996) 7205–7215.
(1966) 1550–1554.