Experimental study of a MIG-MAG welding arc

F.Valensi(1), N.Pellerin(2), S.Pellerin(1), K.Musiol(3), Ch. de Izarra(1),
S.Zielińska(1,3) and F.Briand(4)
(1) LASEP – Centre Univ. de Bourges, BP 4043, 18028 Bourges cedex 2 – France
(2) CRMHT – 1D avenue de la Recherche Scientifique, 45071 Orléans cedex 2 – France
(3) Institut of Physics, Jagellonian University, ul. Reymonta 4, 30-459 Krakow – Poland
(4) CTAS – Air Liquide Welding, Saint Ouen l’Aumone, 95315 Cergy-Pontoise cedex – France
Abstract. Composition of the applied shielding gas has a strong influence on physical properties of the plasma and
parameters in the MIG-MAG welding process. In order to explain this phenomenon, the MIG/MAG welding arc plasma
was investigated by different methods at different mixtures of argon and carbon dioxide in the shielding gas: fast camera
to record distribution of spectral lines of the plasma components; SEM observations, EDX spectrometry and Castaing
microprobe analysis to study the microstructural modifications of the anode tip during the MIG-MAG welding process;
and optical emission spectroscopic to diagnose the welding arc. The obtained results and their implication on the MIGMAG
welding process will be presented during the conference.
Keywords: Welding arc, Argon, CO2, Visualisation
PACS: 52.50, 52.70, 52.77, 52.90
In the MIG-MAG (“Metal Inert Gas”-“Metal Active Gas”) welding process, the plasma arc burns between the
extremity of a fusible electrode (usually anode) and metal plate (usually cathode). The transfer mode of the metal
melted in the arc depends mainly on nature of the used gas, electrode dimensions and composition, and the density
of the welding current. Following the arc current and gas mixtures, the metal transfer can be made of three manners:
the regime of transfer by short-circuits (“short-arc”), the “globular transfer” and the “spray-arc”.
Composition of the applied shielding gas has a strong influence on physical properties of the plasma and
parameters of the welding process. In particular, increase of the percentage of carbon dioxide in argon, results in
increase of the transition current value while changing from the globular to spray mode of metal transfer during the
welding process. Beyond a given value it becomes impossible to reach spray mode, even for the highest currents.
In order to explain this phenomenon, the MIG/MAG welding arc plasma was investigated at different mixtures of
argon and carbon dioxide in the shielding gas. Applying a fast camera, fitted with an interferential filter, images of
the plasma showing the current distribution geometry has been recorded. We have noticed some phenomena not
described yet in the literature [1]. Particularly, results show in a very clear way that there is a limit of the percentage
of relative concentration CO2/Ar beyond which the shape of the arc is significantly modified.
The scheme of the experimental system is presented on Figure 1, and has been partially described in [1]. It is
organized around a welding set SAFMIG 480 TRS PLUS [2, 3]. The welding was performed under reverse-polarity
(wire-anode, workpiece-cathode) direct current in the constant current mode. The wire-anode was a mild steel
consumable electrode (AWS A5.17) with 1.2 mm diameter. The distance between the contact-tube and the metal
plate amounted 20 mm. The thickness of the workpiece was 8 mm for all measurements. The gas mixture based on
the argon and the carbon dioxide, was provided by two bottles of industrial gas. Two mass flow meters allow
measurement and control gas output. The measurements of voltage between

FIGURE 1 – Experimental set-up
[F1: 468.8nm interference filter; F2: Filter; M1..M7: Flat mirors; S1,S2: Spherical mirors; DP: Dove’s prism;
P: metallic plate; W: glass window; T: manually moving table; WEEQ step by step moving table; ST: Voltage probe; SC:
Current probe; S: Shunt]
The spectroscopic diagnostic of the welding arc is necessary to understand the observed changes in the mode of
droplet transfer. The used setup allows to study a chosen slice of the plasma column, and measures have been made
for various shielding gas compositions.
The distribution of spectral lines of the plasma components has been studied for visible and near ultra-violet
wavelengths. In this case the light intensity is simply integrated on the arc column width in order to get a single
With the assumption of plasma symmetry, the radial distribution of light intensity for the chosen slice of the arc
column is calculated, using the Abel inversion. An original method based on measurement of the stark broadening of
the ArI and FeI spectral lines has been used to estimate the temperature and the electronic density distributions in the
plasma, without hypothesis on its equilibrium state. Data are fitted to a Voigt profile from which the Stark
broadening is extracted. Under our experimental conditions, it is actually the dominant part of the lines broadening.

Radial distribution of electronic density and temperature profiles have been determined for the various studied
plasma slices, for conditions corresponding both to globular and spray mode. The main result is that the electronic
temperature in the central part of the arc never seems to exceed the value 15000 K, contrarily to the values given in
literature (T>20000 K on the axis).
Among other spectroscopic results, we can note especially the deficit in excited argon close to the axis of arc
when the amount of CO2 in the shielding gas is less than 5%, the relative weakness of iron lines on the edge of the
column, but also the increase of the FeII spectral lines in periphery of the arc.
On the basis of these results, it is clear that the modelisations available in the literature [4, 5] do not represent the
real situation: for pure argon, they are generally established in the local thermodynamical equilibrium (LTE)
hypothesis, and they do not take into account the metallic vapours influence and the induced temperature lowering.
Actually, to estimate the distribution of relative concentrations of argon and iron in the arc column, we are trying to
verify the LTE hypothesis by using different crossed diagnostic methods (Boltzmann plot, atomic to ionic lines
ratio…). But the zone of most intense iron line emission is limited to the center of the arc column, and possibilities

In order to further investigate the arc attachment at the consumable electrode, the microstructure of anode tip has
also been studied. The main hypothesis is that the cooling of the droplet constituting the end of the anode is fast
enough to preserve its structure. The tip of the wire is cut after the end of the welding sequence and then sliced in the
axial direction for SEM and EDX analysis.
Our results seem to show that the existence of an insulating oxide “gangue” around the tip of the anode is typical
of the globular transfer mode. Its thickness increases when the carbon dioxide concentration increases and the
current increases. Its bad electrical conductivity, as well as its high viscosity and melting temperature, can explain
the effect it has on the metal transfer mode. The high viscosity of the gangue prevents small droplets detachment;
this bad conductor hinders the current transfer and the arc needs a larger attach zone. That explains the bell arc shape
associated to globular transfer or short-arc. It has not been observed in spray transfer mode for pure argon shielding
gas. In those conditions, the plasma column has a well-defined conical shape and the curvature on the current lines is
clearly modified.
The comparison of metal transfer modes for experiments MIG in argon and MAG in argon + CO2 shielding gas
mixture shows an evolution from the short-arc to the globular transfer and then to the spray-arc when current values
are increased. The transition between globular and spray transfer modes is observed towards 240 A in pure argon,
and towards 410 A in gas mixture with 15% CO2.
SEM microstructural observations and EDX analysis, on quenched drops for MIG-MAG experiments have
allowed to characterize the physicochemical transformations of the anode wire, according to working conditions of
the process, such as the nature of the shielding gas and the arc current. CO2 in protective gas favours the formation
of the gangue by chemical oxidation-reduction reactions.
The plasma column diagnosis done by optical emission spectroscopy without hypothesis on the LTE state shows
that, in central part of the arc the temperature is different from what is calculated with numerical models. This work
must be continued to study the LTE state in the plasma, and try to estimate the distribution for relative
concentrations in argon and iron, and then the principal transport and thermodynamic characteristics (thermal and
electrical conductivity, viscosity…) in the arc column
50 μm
FIGURE 3 – Drop microstructure for MAG experiment at 330A in gas mixture – SEM micrograph in BSE mode (at 20%
CO2 vol.)
This work was supported in part by Air Liquide, Saint Ouen l’Aumone (France).
1 S.Zielinska, S.Pellerin, K.Musiol, Ch. de Izarra & F.Briand, ”Influence of the applied gas on the arc shape in MIG-MAG
welding”, 16th International Symposium on Plasma Chemistry, Taormina (Italie) – June 22/27, 2003 [Poster]
2 Safmig 330TRS Plus / Safmig 480TRS Plus – Instruction de sécurité d’emploi et d’entretien , documentation technique
SAF/Air Liquide Welding, Réf. 8695-0273, 1999.
3 Kit Safmig 480TR 16 – Instruction de sécurité d’emploi et d’entretien , documentation technique Air Liquide, Réf. 8695-044,
4 J.Haidar, J.Appl.Phys. 84 (1998) 3530
5 J.Haidar, J.Appl.Phys. 84 (1998) 3518

Опубликовано в рубрике Documents