Physics of Welding

Download as pdf or txt
Download as pdf or txt
You are on page 1of 8

in Advanced Joining Technologies,

Technologies,
T.H. North, ed., Chapman and Hall, London, U.K.,
U.K., 61, 1990.
1990.

THE PHYSICS OF ARC WELDING PROCESSES


T.W.EAGAR
Department of Materials Science and Engineering,
Massachusetts Institute of Technology

Abstract
welding is an extremely complex process; however, due to
its commercial importance, it is essential that a more
thorough study of the various processes be undertaken.
Three examples of areas requiring greater understanding of
arc physics are presented. The first discusses heating or
cooling of the metal at the cathode; the second describes
variations in heat transfer using various shielding gases
and the last describes droplet formation in gas metal arc
welding.
Key Words: Arc Welding, Arc Physics, Shielding Gases, Gas
Metal Arc Welding.
1. Introduction

Langmuir, ~omptonand Cobine are but a few of a number of


outstanding physicists who contributed greatly to our
understanding of atmospheric pressure arcs during the first
half of this century. However, during the past forty
years, few physicists have studied such problems. I have
been told by some physicists that arc physics is too
mundane and too simple a problem to challenge a modern day
physicist. This, of course, is merely an excuse.
Atmospheric pressure arcs are too complex and represent too
difficult a problem for these people to study. Stated
simply, atmospheric pressure arcs and welding present
problems which are too difficult and too complicated to be
approached by the type of physicists that are educated in
today's world.
The complexity of welding is readily apparent when one
considers that fusion welding involves temperature
gradients of thousands of degrees, over distances of less
than a centimeter, occuring on a time scale of seconds,
involving multiple phases of solids, liquids, gases and
plasma. This complexity causes most scientists to shun
welding as unscientific and. unworthy of thoughtful
analysis. It is only the true engineer who is willing to
deal with a process of such exceeding complexity in order

I ;~

to achieve the end result of a fabricated product of


commercial and societal usefulness. One of the challenges
facing those who engineer welded structures and components,
is communication of the importance and significance of
welding to other scientists and to the general public. One
of the purposes of this paper is to describe some of the
complexity of several welding processes, and to demonstrate
how study of arc physics can help provide a better
understanding of arc welding processes.
2. Anode and Cathode Spot Heating

Cobine [I9551 and others have shown that the heat transferred
to the anode, QA, is given by

where: I is the welding current in amperes


(o,
is the work function or heat of condensation of the
electrons expressed in volts
I T A is the anode voltage drop
k is the Boltzmann's constant
T is the absolute temperature of electrons in the
plasma, and
e is the electron charge
By symmetry, the heat at the cathode should be given by

where the sign of the current is reversed due to reversal


of electron flow. Such a simplistic formulation of cathode
heating may not be valid due to the much more complex (and
essentially unknown) physics which occur in the cathode
spot [Ecker, 19801. Nonetheless, the sign reversal of the
current clearly indicates that the exiting electrons cool
the cathode. The removal of electrons to the plasma causes
a cooling of the cathode even as a vaporizing gas removes
heat from a liquid.
Experimental confirmation of cathodic cooling was
obtained by Ishikawa et al. 119651 who used an infrared
pyrometer to show that the hottest portion of a gas
tungsten welding arc is not located at the tip, but several
millimeters up the electrode. When the cathode is liquid,
as in gas metal arc welding, convection within the weld
pool essentially negates the local cooling effect of
electron vaporization. Thus, cathodic cooling is an
important factor when the cathode is solid; but when the
I
cathode is liquid, cathodic cooling has little influence on 1
I

temperature gradients, although it may alter the overall


heat balance.
Nonetheless, cathodic cooling may be important in
influencing cathode spot motion on the liquid weld pool.
Electron emission from the weld pool is influenced by the
temperature of the cathode; a hotter cathode produces
emission more readily. When gas metal arc welding of most
metals, the cathode spot wanders rapidly across the liquid
surface. Does it do this because cathodic cooling produces
easier electron emission in the hotter areas adjacent to
the cathode spot? Or does cathodic cleaning of surface
active oxygen at the cathode spot create a region of easier
electron emission on adjacent surface regions where a thin
oxygen film remains? Does the addition of oxygen to argon
when gas metal arc welding of steel produce an oxygen
surface film that influences arc stability? What is the
effect of metal vaporization on the properties of the
plasma immediately above the cathode spot? Does this high
concentration of easily ionized metal vapors create
stability for the cathode spot or does it cause the spot to
wander about the cathode surface? These and other
questions remain unanswered, yet our process modelling
tools have progressed to a sufficient level that some could
be answered - at least qualitatively.
Clearly, there is much to be learned about the science
of atmospheric pressure arcs as they relate to welding.
Some of these questions could lead to answers of commercial
importance.
3. Heat Transfer UsingVarious Shielding Gases

It is well known that one will melt more metal with a 100
ampere helium arc than with a 100 ampere argon arc. The
conventional wisdom says that this is due to the higher
ionization potential of helium (24.6 eV) as compared to
argon (15.8 eV). It is assumed based upon the Saha
equation that this will result in a higher arc temperature.
Unfortunately, based on experimental measurements [Key,
19831, there is not a higher temperature in the helium
welding arc; and even if there were, it could not explain
the much greater melting efficiency when using helium. The
correct explanation lies in the variation of the thermal
conductivity of the gases.
Although eighty percent of the heat in the argon arc is
carried by the electron current, twenty percent is due to
heat conductance across a gas boundary layer at the anode
surface. IQuiqley et al., 1973; Dinesclu and Pfender, 19801
Heat flow across this boundary layer is controlled by the
thermal conductivity of the gas. For monatomic gases such
as helium and argon, the thermal conductivity is controlled
by the mass diffusivity of the atoms, which, according to
the kinetic theory of gases, is proportional to the inverse

square root of the mass of the atom [Dunn and Eagar, 19861.
Thus, with an atomic weight ten times as great as helium,
argon has a thermal conductivity only 30 percent as large
as helium. All other factors being equal, helium will
conduct three times as much heat across the boundary layer
as will argon. Table 1 illustrates this point. Although
only 20 percent of the heat entering the anode from the
argon arc is conducted across the boundary layer, 43% of
the heat in a helium arc will be conducted'across the
boundary layer. (These figures are only "back of the
envelope" calculations. Many other parameters change when
switching from argon to helium gas. Other factors such as
anode voltage drop, electric current distribution, arc
temperature or radiation will change, but by insignificant
amounts as compared to the thermal conductivity.)
As pointed out by Giedt et al. [1989], the actual
melting efficiency of the arc welding process is relatively
low (i.e. on the order of 20 percentor less). Thus an
increase of 40% in the heat supplied could double the
volume of metal melted. This is why helium shielded arcs
melt more metal than argon shielded arcs.

Shielding
Gas

Electric
Current

Conduction
across gas
boundary
layer

Argon

80%

20%

20%

Helium

80%

60%

43%

Percent by
conduction

Table 1. Relative Heat Transfer from the Welding Plasma in


Argon and Helium Shielding Gases (All heat values
normalized to 100% for argon).
In addition to these pure gases, this hypothesis can
explain why gas mixtures such as 75 Ar - 25 He or 50 Ar 50 He will increase the rate of heat transfer over that of
pure argon. Such mixtures have a thermal conductivity
intermediate to the two pure gases. The ionization
potential theory completely fails to explain why argon rich
mixtures increase the rate of heat transfer. Since only 5
to 30% of the arc need be ionized, gas mixtures of less
than 70 percent He contain no He ions. [Kim, 19861 If the
He does not ionize, its ionization cannot be used to
explain the more efficient heat transfer.
One may also ask why Argon-5% Hz shielding gases also
transfer heat more efficiently than pure argon. Using the
simple mass diffusivity, thermal conductivity relationships
presented above with a simple rule of mixtures, one would

conclude that Ar-5% H2 would have approximately a 25%


greater thermal conductivity than pure argon. This cannot
explain the greater rate of melting observed as this
increase in thermal conductivity would result in only a 5%
increase in the amount of heat transformed to the anode.
The explanation in this case is due to a third mechanism of
heat conduction. In addition to the flow of the electric
current and the thermal conductivity across the gas
boundary layer, diatomic gases such as hydrogen also have
an enhanced thermal conductivity due to dissociation of the
gas in the plasma and recombination in the gas boundary
layer. This enhancement of thermal conductivity in
polyatomic gas plasmas is known as the reactive thermal
conductivity. [Dunn, 19841 In fact, the thermal
conductivity of Ar-5% H2 is enhanced even more than 25%
over that of pure argon, due to this contribution of the
reactive thermal conductivity.

4. Droplet Formation in GMAW


Photography of the gas metal arc process shows that the
anode spot corona varies markedly with different shielding
gases. With helium or carbon dioxide shielding, the anode
spot is attached to the bottom of the pendant drop on the
end of the electrode. The size of the spot does not change
markedly as the current increases. With argon, the anode
spot is more diffuse and grows noticably in size as the
current increases. At sufficiently high currents (for
example, above 23.0 amperes with 1.6 mm diameter steel
electrodes) the anode spot fills the molten drop and begins
to climb the cylindrical side walls of the electrode.
Since most of the heat carried to the anode is carried by
the welding current, the current in the anode spot causes
the cylindrical side walls of the electrode to melt [Y.S.
Kim, 19891. The end of the electrode forms a tapered
geometry as shown in Figure 1. Such a tapered electrode
'tip geometry does not develop in pure helium or pure carbon
dioxide shielding gases.
The tapered tip creates a smaller diameter for
attachment of the droplet to the electrode by surface
tension forces [Y.S. Kim, 19891. This produces smaller
droplets than would be present in the absence of taper
formation. The smaller droplets formed with argon
shielding from a tapered electrode are seen clearly in
Figure 2, as compared with helium and carbon dioxide
shielding.
The formation of the taper occurs during gas metal arc
welding of steel, aluminum and titanium, when using argon
shielding gas [Y.S. Kim, 19891. There are a number of
consequences of this taper formation. Firstly, Y.S. Kim
[I9891 has shown that the melting rate of the electrode is
controlled by heat transfer across the liquid-solid

Fig. 1. Development of a taper as welding current


increases when using argon shielding gas. (a) 205
amperes, (b) 237 amperes, (c) 253 amperes, (d) 281
amperes, and (e) 310 amperes.

150

200

250

300

WELDING CURRENT, amperes

Fig. 2.

Droplet size measured in three different shielding


gases as a function of welding current for 1.6 mm
diameter steel electrodes.

boundary between the drop and the solid electrode. Since


the taper reduces the area of this boundary, taper
formation reduces the rate of melting of the electrode. At
high currents, the melting rate in helium and carbon
dioxide is greater than in argon due to absence of the

taper. ~ h u sthe greater melting rate in helium is not due


to the higher ionization potential of this gas as is
commonly assumed (see Section 3 of this paper), but to a
larger liquid-solid contact area in which heat is
transferred to the solid electrode.
Secondly, Y.S. Kim has shown that the formation of a
taper can reduce the range of one drop per pulse transfer
in pulsed current GMAW. As the drops become smaller with
argon shielding gas due to taper formation, the pulsing
frequency must increase accordingly. By suppressing the
formation of the taper, both large and small drops can be
released from the end of the electrode during pulsed
current welding. The largest drops are equal to the
droplet size created at the low current portion of the
pulse in the absence of the taper, while the smallest drops
are equivalent to those produced during the peak of the
current pulse. This greatly increases the range of
frequencies which are permissible for one drop per pulse
operation [Y.S. Kim and Eagar]. By addition of helium or
carbon dioxide to the argon shielding gas, the anode spot
is constricted and does not climb the cylindrical side
walls of the electrode as readily. Thus taper formation is
suppressed and the acceptable range of frequencies for
pulsed current GMAW is broadened and the process becomes
more easily controlled. This example illustrates how an
understanding of the physics of droplet formation can aid
the welding engineer in selection of the best shielding gas
composition for pulsed current gas metal arc welding.
5. Conclusion

It is seen that an understanding of the arc physics can


make a very complex process such as welding, appear much
clearer. By understanding the process, it becomes easier
to control and easier to modify our procedures to ensure
'that high quality, reliable joints are produced. With so
many arc process variations and with an ever increasing
number of materials to be joined, there is ample
opportunity for many people to contribute to a better
understanding of the physics of welding processes.
6. Acknowledgements

The author expresses his appreciation to the U.S.


Department of Energy, the Office of Naval Research and the
National Science ~oundationfor support of the research
presented herein.

References
Dinulescu, A. and Pfender, E. (1980) ~nalysisof the Anode
Boundary Layer ofHigh Intensity Arcs. J. Acpl. Phvs.,
5J(6) 3149-3157.
Dunn, G.J. (1984) Metal Vapors in Gas Tungsten welding
Arcs. S.M. Thesis, Dept. of Materials Science and
~ngineering,MIT, Cambridge, MA.
Dunn, G.J. and..Eagar, T.W. (1986) Metal Vapors in Gas
Tungsten Arcs: Part 11. Theoretical Calculations of
Transport Properties. Metall. Trans., ~ ( l O ) ,1865-1871.
Ecker, G. (1980) Theoretical Aspects of the Vacuum Arc, in
Vacuum Arcs, Theorv and Aoplication (ed. J.M. Lafferty),
John Wiley, NY, pp. 228-320.
~ i e d t ,W.H., ~allerico,L.N. and Fuerschbach, P.W. (1989)
GTA Welding Efficiency: Calorimetric and Temperature Field
Measurements. Weldina J,, &(l) 28s-32s.
Key, J.F., Chan, J.W. and McIlwain, M.E. (1983) Process
Variable Influence on Arc Temperature Distribution.
Welding J., 62(7), 179s-184s.
Kim, E.W. (1986) Visible Light missions During Gas
Tungsten Arc Welding. S.M. Thesis, Dept. of Materials
Science and Engineering, MIT, cambridge, MA.
Kim, Y.S. (1989) Metal Transfer in Gas Metal Arc Welding.
Ph.D. Thesis, Dept. of Materials Science and Engineering,
MIT, Cambridge, MA.
Kim, Y.S. and Eagar, T.W. Metal Transfer in Pulsed Current
GMAW. Submitted to Weldinu J..
Quigley, M.B.C., Richards, P.H., Swift-Hook, D.T. and Gick,
A.E.F. (1973) Heat Flow to the Workpiece From a TIG Welding
Arc. J. P h ~ s .D. ~ l .
P~vs.,6, 2250-2258.
Savage, W.F., Strunck, S.S. and Ishikawa, Y. (1965) The
Effect of Electrode Geometry in Gas Tungsten-Arc Welding.
Weldinq J., U(ll), 489s-496s.

You might also like