D S Howse and W Lucas
Published in Science and Technology of Welding and Joining, 2000, Vol. 5, No. 3, pp 189-193, by IOM Communications Ltd for The Institute of Materials.
The authors work in the Arc, Laser and Sheet Processes Department of TWI, Granta Park, Great Abington, Cambridge, CB1 6AL, United Kingdom.
The mechanisms by which active fluxes increase the penetration of conventional TIG welds (A-TIG) are reviewed. The most dominant mechanism for increased penetration is considered to be arc constriction rather than a change in the surface tension of the molten pool. An experimental programme of work was carried out using the A-TIG flux in combination with a number of welding processes. The plasma process was investigated as it gives greater penetration than conventional TIG welding by increasing current density. The CO2 laser and electron beam processes which do not rely on a current carrying arc as the heat source for welding were also investigated. Macrosections taken from the welds made by these processes showed that the A-TIG flux was only effective when the weld pool is produced by an arc or plasma. Where there was no arc or plasma present, the flux had little effect.
Activated fluxes that increase the penetration of TIG welds (A-TIG fluxes) offer a means of significantly increasing the productivity of the welding process. A-TIG fluxes were first utilised in the late 1950s by the EO Paton Institute of Electric Welding in the former Soviet Union. The first published papers [1,2] described the use of the fluxes for welding titanium alloys and the first reference to the use of the technology for steels was in 1968  . The use of these fluxes reduces both the need for edge preparations and increases productivity due to the reduction in the number of weld passes required to make the joint. It is possible using this approach to achieve full penetration TIG welds in 12mm material with a single pass  . These fluxes have been introduced recently to the West and have generated a great deal of interest in how the fluxes can increase penetration in TIG welds. A number of mechanisms have been proposed to account for the observation that the arc visibly constricts in the presence of an activating flux. At the onset of the research reported here, it was considered that the concentrated arc energy might increase weld penetration through either an arc or weld pool flow mechanism.
2. Review of mechanisms
The first published reference found which relates to activated fluxes for TIG welding dates from 1965 
and refers to the improvement in the penetration of titanium alloys using an oxygen free activated flux. The effect and physical characteristics of the process, such as a drop in the welding current and an increase in the welding voltage associated with the flux are noted in at least one more paper 
. The first paper which refers to a similar effect for steels was published in 1968 
A number of observations have been made about the effect of the fluxes on the TIG arc for titanium alloys and stainless steels. It is claimed that the ability of the flux to wet the surface of the molten pool has an effect on the suitability and that the composition can be altered to modify the surface tension and produce better wetting  . However, it is not entirely clear as to how this wetting of the surface contributes to the effectiveness of the flux other than it must, when liquid, cover the surface of the molten pool  . This wetting of the surface of the weld and altered surface tension of the flux itself is not believed to be related to the Marangoni fluid flow effect in the weld pool. A change in Marangoni flow, has been used to explain variable penetration in welds  . This change in fluid flow is related the Thermal Coefficient of Surface Tension (TCST) of the molten pool. If the TCST is negative, the cooler peripheral regions of the pool will have a higher surface tension than the centre of the weld pool and the flow will be outwards creating a wide shallow weld pool. In materials with a positive gradient, this flow is reversed to the centre of the weld pool and in the centre the molten material flows down  . This creates a narrower deeper weld pool for exactly the same welding conditions ( Fig.1).
Fig.1. Marangoni flow a) negative coefficient giving shallow penetration and b) positive coefficient giving deep penetration.
There are other mechanisms proposed for altered weld bead penetration and the use of oxide coatings to improve weldability in steels is also a recognised practice  . This technique is thought to stabilise the arc roots in TIG welding but could be a similar effect to that produced by the A-TIG fluxes.
In terms of the A-TIG fluxes themselves, AG Simonik published a paper on the welding of titanium with fluxes made up of CaF2 and AlF3  . In this paper, he proposed a theory for the effectiveness of the fluxes based upon an arc constriction mechanism.
Simonik also demonstrated that the effectiveness of a flux constituent in constricting the arc was linked to a higher temperature of molecule formation and this was proved by experimental work.
In reviewing the above literature, although Simonik's theory appears plausible and does concur with his experimental observations, the model proposed for the TIG arc does not agree with current theory in which the arc is comprised of a central ionised (plasma) column rather than neutral atoms.
Lucas and Howse  have applied Simonik's principle of electron absorption to account for the observed constriction of the arc and the increased weld pool penetration. The mechanism is based on the concept that the TIG arc is comprised of the following four regions ( Fig.2):
Fig.2. Mechanism proposed by Lucas and Howse for the A-TIG process
Plasma column - Current carried by the electrons and ions produced by the thermal ionisation of the shielding gas. Anode/cathode - High potential drop to maintain the current flow as the gas is cooled by the electrode (the plasma temperature is much greater than the electrode). Cathode - Under the bombardment by positive ions, the high temperature creates the conditions for the thermionic emission of electrons. Anode - Under the influence of the anode potential drop, the electrons accelerate but then the kinetic energy is transferred to the anode.
Thus, the heat required to form the weld pool is principally derived from the transfer of the kinetic energy of the electrons as they are absorbed into the surface. The amount of heat produced at the surface is determined by the energy acquired on accelerating across the anode drop and their heat of condensation. Constriction of the arc will increase the temperature at the anode because of the increase in current density and the higher arc voltage.
It is considered that the vaporised flux will constrict the arc by capturing electrons in the outer regions of the arc in a similar manner to that proposed by Simonik. Electron absorption is effected by the attachment of electrons to vaporised molecules and dissociated atoms to form negatively charged particles. Electron attachment can only take place in the cooler peripheral regions where the electrons have low energy in a weak electric field. Towards the centre of the arc where there is a strong electric field, high temperatures and very high energy electrons, ionisation will dominate. Thus, restricting current flow to the central region of the arc will increase the current density in the plasma and at the anode resulting in a narrower arc and a deeper weld pool.
The proposed mechanism is supported by the observed relative effectiveness of the flux constituents by Simonik. For example, arc constriction will be promoted by flux constituents whose molecules or atoms have a large electron attachment cross section. Thus, halogen compounds which have a large electron attachment cross section when dissociated, will have a strong affinity for electrons. Other compounds, such as metal oxides, which have a lower electron attachment diameter but a higher dissociation temperature, are equally effective in constricting the arc as they can provide a greater number of vaporised molecules and atoms in the outer regions of the arc.
Based upon the above, a practical experimental procedure was devised to confirm the proposed mechanism for the A-TIG flux. This programme of work provided a comparison of the A-TIG effect with other arc/plasma techniques which achieve deep penetration by increasing the ionisation potential of the shielding gas (argon/helium mixtures) and using a higher current density heating source (plasma welds). Laser and electron beam welds were included to investigate how the A-TIG flux would affect the weld bead shape in a weld made with no current carrying arc or plasma. It was the intention with the latter welds to separate out the effect of any arc or plasma and any possible effect of change in thermal coefficient of surface tension.
3. Experimental procedure
The material used during these trials was a single cast of austenitic stainless steel AISI 316L material of 6mm thickness. The chemical composition of this material is shown in Table 1. The flux composition used throughout these trials was that produced by the Paton Institute for stainless steels identified as AFP SS1. This flux is made up of a mixture of metal oxides.
Table 1 Chemical analysis of 6mm thickness austenitic AISI 316L steel used for weldability trials
3.2 Welding procedures
In order to investigate the effect of the flux and to provide a comparison with the other processes used in this work, the flux was first used to produce a melt run weld with a conventional TIG torch. The fluxes are supplied in powder form and are mixed with acetone and painted onto the surface to be welded. The acetone evaporates, leaving a layer of the flux adhering to the surface of the material to be welded. In the first instance a weld was made on a single plate half coated with the flux and the arc moved from an uncoated region to a coated region. The weld parameters were monitored with a QA Weldcheck arc monitoring system. The shielding gas used was industrially pure argon (>99.9%).
Another weld was then made using a similar procedure. This weld was made with a 75% helium - 25% argon mixture but with the same welding current as the weld made with pure argon shielding mixture. Again, this weld was made by moving the welding head from an uncoated region of the plate into a region coated by the flux.
The effect of the flux was also investigated for plasma welding. Two keyhole welds were made, again both with and without the flux, using both pulsed and continuous welding current. Again, melt run welds were made.
A number of laser welds were made with varying power and travel speed on the stainless steel material. The intention was to produce deep penetration laser welds with the keyhole mechanism and also de-focus the beam to give a similar weld pool size to a TIG weld and investigate whether the fluxes affected the shape of the weld bead.
Welds were also made using electron beam welding. The flux used for this exercise, although the same composition of AFP SS1, was applied by an aerosol applicator which had some lacquer binder to attach the flux to the surface. It was thought that if the powder form was used for these welds, the flux could lift off the surface without being melted.
Macrosections were taken of all the welds, made both with and without the A-TIG flux.
4. Results and discussion
4.1 TIG welds
The welds with argon shielding gas were made with a fixed stand-off distance of 0.8mm from the tip of the tungsten to the workpiece. The first published reference which discusses the phenomenon of increased penetration of these fluxes in some detail identifies an increase in arc voltage associated with the fluxes and also a reduction in the welding current 2. The initial TIG trials carried out at TWI also show an increase in the arc voltage of typically 0.5 V but not a reduction in the welding current.
It can be clearly seen from Fig.3 that the A-TIG flux causes increased penetration and a reduced width of bead. For these TIG welds, the weld cross sectional area is also increased by 30 to 80%. Although some of this can be attributed to the increase in voltage and change in cooling rate due to the surrounding material, it is also possible that the flux has increased the efficiency factor of the process, i.e. more of the arc energy is transmitted as heat input to the parent material.
a) without A-TIG flux at 187A, 9.5V, 100mm/min and 1.07 kJ/mm arc energy (AH1898)
b) with A-TIG flux 187A, 10V, 100 mm/min and 1.12 kJ/mm arc energy. 6mm thickness stainless steel (AH1899)
Fig.3. Macrosections of TIG welds made with argon shielding
The welds made with the helium 25% argon shielding gas were also made with a fixed stand off of 0.8mm ( Fig.4). It is well known that the use of helium as a shielding gas will give a greater arc voltage than for pure argon at a given arc length  . However, it can be seen from these welds that the increase in arc voltage of 1.0V due to using the helium in the shielding gas does not have the same effect as the increase in arc voltage noted when the A-TIG flux is used. If Fig.3 is compared with Fig.4 it can be seen that the two effects, increased voltage due to the higher ionisation potential of the helium causing a wider weld bead, and increased voltage to the A-TIG flux causing a deeper weld bead, give very different results. The welds made with the helium based gas mixture, although showing an increased heat input, do not affect the penetration in the same way as the A-TIG fluxes. It can clearly be seen that the increase in arc voltage for the helium rich shielding gas also leads to a spreading of the arc and an increase in the width of the bead, rather than a reduction in the width of the bead which is characteristic of the A-TIG flux welds. When the helium rich shielding gas and the fluxes are used together, it can clearly be seen that the combined effects are cumulative and appear to be independent of one another.
a) without A-TIG flux at 187A, 11V, 100mm/min and 1.23 kJ/mm arc energy (AH1916)
b) with A-TIG flux at 187A, 11.8V, 100 mm/min and 1.32 kJ/mm arc energy. 6mm thickness stainless steel (AH1915)
Fig.4. Macrosections of TIG welds made with 75% helium 25% argon shielding
4.2 Plasma welds
The macrosections of the continuous current plasma welds are shown in Fig.5
. It can be seen that the flux has increased the width of the penetration bead and increased the cross sectional area producing a weld bead shape very similar to the A-TIG flux TIG weld. The plasma welds show that the effect is not restricted entirely to TIG welds. Clearly, the mechanisms acting to cause increased penetration in the TIG welds with the fluxes have some effect on plasma welds also. The effect of the flux on the plasma weld is not as pronounced as for the TIG welds but this would be expected. The plasma welds were made in the keyhole mode. Any constriction of the plasma arc will not increase the penetration of the weld as it is producing a fully penetrating weld already. For TIG welding, the mechanism is very different. The arc is 'softer' than for plasma and does not penetrate the workpiece. The weld is made in a conduction limited mode and any constriction of the arc will increase the penetrating power of the arc. The results seen here indicate that the flux has had an effect on the plasma process and the reduced cap width indicates that the plasma arc was constricted by the flux.
a) without A-TIG flux at 177A, 26.5V, 210mm/min and 1.34 kJ/mm arc energy (AH1900)
b) with A-TIG flux at 177A, 26.5V, 210 mm/min and 1.34 kJ/mm arc energy. 6mm thickness stainless steel (AH1903)
Fig.5. Macrosections of continuous current plasma welds
4.3 CO2 Laser welds
Macrosections of two of the CO2
laser welds are shown in Fig.6
. Again, the weld bead shape is changed when it moves into the A-TIG flux AFP SS1.
a) without A-TIG flux (AH1963)
b) with A-TIG flux. 6mm thickness stainless steel (AH1964)
Fig.6. Macrosections of CO2 laser welds 2.4 kW, 1.0m/min
These welds were made to investigate the effect of the A-TIG fluxes on the molten pool without the influence of a current carrying welding arc. It can clearly be seen that the A-TIG fluxes do also have an effect on the penetration of the weld in the conduction limited mode and some effect in the deep penetration mode. There is, however, a plasma associated with CO 2 laser welds and it is thought that the wider top portion of the laser weld profile is caused by the heating effect of the plasma rather than the effect of the laser beam itself. It was noted whilst welding that the plasma visibly constricted when the weld moved into the flux coated region. It is possible that the fluxes could cause a constriction of the non current carrying plasma associated with the laser in the manner originally proposed by Simonik for the TIG arc. This constriction would lead to a lesser effect of the laser beam being scattered and hence produce a greater power density beam. It may then also be assumed that the higher density beam could produce a depressed molten pool which would cause increased activity and enhanced flow in the weld pool. It has been noted that for the TIG welds made with A-TIG fluxes in this programme of work, in some instances, the arc does appear to submerge slightly also giving a depressed molten pool.
4.4 Electron beam welds
The electron beam melt runs were made to investigate welding without the usual arc or plasma in order to attempt to separate any effects of arc or plasma constriction and surface tension caused by the flux. The welds were made as melt runs, using three heat inputs successively on the same plate. In order to assess the effect of any heat build up in the plate, the last melt runs repeated the heat input of the first. By using a combination of beam de-focus and beam deflection, the beam's power density was modified to simulate that of a typical TIG arc. Again, two of the weld sections are shown both with and without the A-TIG flux in Fig.7
. It can be seen that the electron beam melt runs do not show major increases in penetration due to the A-TIG fluxes. There is some difference between the larger of the weld beads compared with and without flux but this was not observed in the other melt runs made.
a) without A-TIG flux (AJ1372)
b) with A-TIG flux. 6mm thickness stainless steel (AJ1377)
Fig.7. Macrosections of electron beam welds 16.0 mA, 6.5 kHz
It was concluded that the A-TIG flux AFP SS1 had a minor effect on the melt run bead shape in some cases. However, overall there was no consistent major effect on the penetration of any of the electron beam melt runs when compared to the other welding processes investigated in this work.
5. Concluding remarks
The proposed fundamental mechanisms for weld pool penetration in TIG welding range from arc effects, electromagnetic effects and weld pool flow and it is likely that most of these effects exist in an equilibrium condition in TIG welding. However, in comparing the TIG, plasma, CO 2 laser welds with the electron beam welds, it is possible to say that the effect of the A-TIG fluxes is clearly a very separate effect to changes in weld pool flow due to an altered surface tension effect proposed by Heiple and Roper. The theoretical explanation proposed by Lucas and Howse suggests an arc constriction effect which is given further credence in that this constriction is clearly visible whilst welding. It is thought that this constriction of the arc leads to an increase in current density which in turn leads to greater arc forces acting on the molten pool which produce the increase in penetration depth. Other work carried out  has also shown that the fluxes reduce the effect of poor weld bead penetration caused by cast to cast variation in stainless steels. This in itself indicates that the proposed arc effect is a different and dominant mechanism to the Marangoni/surface tension flow effect and shows that the problem of cast to cast variation could be overcome by using these A-TIG fluxes.
The most plausible mechanism at present is that the arc or plasma is constricted by the action of the A-TIG fluxes and that the associated increase in current density results in increased forces which alter the molten pool flow to give increased penetration. Although there are a number of effects resulting from the use of activating fluxes it is thought that the increased penetration associated with the fluxes is not caused by a change in the thermal coefficient of surface tension (Marangoni flow).
This work was carried out within the Core Research Programme of TWI and was funded by Industrial Members of TWI. The authors would like to thank C Hardy for carrying out the TIG and plasma welds, M Hardy for carrying out the laser welds and B Dance for his advice and assistance in producing the electron beam welds. The authors would also like to thank colleagues at TWI for their advice during the work.
- S.M. Gurevich et al: 'Improving the penetration of titanium alloys when they are welded by argon tungsten arc process' Automatic Welding, 1965, 9, 1-4.
- S.M. Gurevich and Zamkov: 'Welding titanium with a non-consumable electrode using fluxes' Automatic Welding, 1966, 12, 13-16
- A.M Makara et al: 'High-tensile martensitic steels welded by argon tungsten arc process using flux' Automatic Welding, 7, 1968, 78-79.
- K. Yushchenko et al: 'ATIG [activating flux TIG] welding of carbon-manganese and stainless steels', Paper 2, 'Welding technology Paton Institute', Proc. Conf. Abington, Cambridge, Oct. 1993.
- N.M. Voropai and Lebedeva: 'Physical properties of welding fluxes based on TiO 2 formed in melting activated wires' Automatic Welding, 3, 1989, 19-23.
- A.M. Makara et al: 'High-tensile martensitic steels welded by argon tungsten arc process using flux' Automatic Welding, 7, 1968, 78-79.
- Rogers et al: 'Minor effects on superalloy weld pools' Proc., 3rd International Conference on Trends in Welding Research, Gatlinburg, TN, USA, June, 1993, 709-712
- C.R. Heiple and J.R. Roper : 'Mechanism for minor element effect on GTA fusion zone geometry' Welding Journal, 1982, 61 (4).
- J. Tapp and P. Simpson : 'Effects of calcium content in austenitic stainless steel on autogenous GTA [TIG] welding' Welding Review International, 3, 1993, 149-150.
- US Patent 3,584,187 'Method of welding stainless steel' Filed April 3 1969, Patented June 8 1971.
- A.G. Simonik: 'The effect of contraction of the arc discharge upon the introduction of electro-negative elements' Welding Production, 1976, 3, pp 49-51.
- O.E. Ostrovskii et al: 'The effect of activating fluxes on the penetration capability of the welding arc and the energy concentration in the anode spot' Welding Production, 3, 1977, 3-4.
- W. Lucas, D.S. Howse: 'Activating flux - increasing the performance and productivity of the TIG and plasma processes' Welding and Metal Fabrication, 1996, January, pp 11 - 17.
- W. Lucas: 'Shielding gases for arc welding - part 1' Welding and Metal Fabrication, 1992, June, 218-225.
- B.E. Paton et al: 'The weldability of structural steels after refining by remelting' Automatic Welding, 6, 1974, 1-4.