Laser welding of steel sheet - the carbon monoxide alternative

TWI Bulletin, January - February 1997

Ian Jones, Senior Research Engineer, helped develop techniques for high power laser welding at TWI. Specifically welding of steel plate with and without filler wire, leading to the production of guidelines for application of laser welding in ship construction.

CO lasers have potential advantages over CO ₂ lasers due to the shorter wavelength of light produced (5-6µm), which results in the possibility of transmission of the energy through fibre-optic cables at low power levels and reduced plasma effects for welding applications. Paul Hilton and Ian Jones describe the results obtained on a fast axial flow CO laser developed within Eureka Project EU113 capable of operating at continuous output powers of 2.5kW for welding of mild steel.

The results reported here are part of the Eureka EU113 carbon monoxide laser project. This project, which began in 1992 and is due for completion in 1996, consists of a group of collaborators made up from commercial and research organisations from France, Russia and the UK. The principal objective of the EU113 CO project is to demonstrate that CO lasers have inherent advantages over the well-established CO ₂ and Nd:YAG lasers in both industrial and medical applications, and also that they can compete successfully in the market place. In order to exploit these advantages fully, the project aimed to develop not only lasers suitable for a range of applications but also optical fibres and optical components for the wavelength range of 5-6µm in which the CO laser operates. The preliminary results, described here are obtained using the fast axial flow CO laser (located at TWI) on welding of mild steel.

The fast axial flow CO laser

Four organisations from the project were involved in developing the fast axial flow laser source: Laser Ecosse now Howden Lasers (an industrial laser manufacturer taking on the role of laser head development); Farnell Hivolt (a power supply manufacturer who developed the electrical power source); Loughborough University (who undertook investigations into DC electrode design and gas flow behaviour) and TWI (who commissioned this laser and will be the evaluator of the laser's performance for materials processing applications).

The laser development involved conversion of a well-established design for a 5kW CO ₂ laser, using 20 fast axial flow tubes arranged within a recirculating gas flow loop. A discharge tube pressure of 120mbar is used with a gas composition of CO/He/N ₂ in the ratio of 1.2/60/4. Considerable attention must be paid to prevention of air leaks in the system as CO laser performance is very sensitive to O ₂ partial pressure and to the presence of water vapour. The latter problem is minimised by using bakeable water traps containing molecular sieve material. Introduction of a controlled amount of O2 however, as well as that of Xe, to the gas mix, does produce a significant power increase. Gas temperatures are mainta maintained at -10°C at the discharge tube inlets using several water-to-gas heat-exchangers within the gas flow loop. Power output drops if higher temperatures are used, especially above +20°C.

During experimental trials reported below, the beam from the laser was directed via two flat copper mirrors to a 150mm focal length 90° off-axis parabolic form copper mirror. The total enclosed beam path for this system had a length of 2m, and was purged continuously with the laboratory 'dry' compressed air system. A maximum power of 2.5kW at the laser output could be achieved for periods of several hours. However, with the beam line still closed and purged, the Joule stick measurement of power at the focusing parabola was about 500W less than that near the output window. It would also appear that the drop in power along the beam path was greater when closer to the laser.

Mirror surface and condition were eliminated as a cause for the above effect. The most likely source of this power drop is believed to be variable atmospheric absorption of the beam through the beam tubes. Schellhorn and von Bülow [1] using a gas dynamically cooled CO laser operated in a semi-closed gas cycle allowing an operation cycle of eight seconds, have reported significant thermal blooming effects in humid CO laser beam lines and resultant loss in beam quality. Gebhardt [2] has also reported that absorption due to water vapour is much higher for the spectral range of the fast axial flow CO laser when compared to the spectral range of the gas dynamic cryogenically cooled CO laser. It is clear that careful control of atmosphere in the beam path may be necessary for certain CO lasers.

Experimental details

Experimental welding work using the CO laser consisted of melt runs carried out in 6.2mm thick C-Mn steel. The work had two main objectives:

• To compare the welding performance when using different plasma disrupting gases at a range of welding speeds.
• To assess the effect of the gases on transition from deep penetration to conduction limited welding as power is reduced

All the melt runs were carried out using the CO laser beam focused at the top surface of the testpieces via a 150mm focal length paraboloidal mirror. The focused spot diameter at this position was estimated at 0.4mm, by measuring the diameter of the hole punched in a very thin aluminium foil when using the laser in pulsed mode. The gas for plasma disruption was applied at a flow rate of 30 litre/min via a 4mm diameter bore copper tube at 45° to the work surface. The gas was directed at the weld pool centre blowing in the welding direction. Helium, argon and carbon dioxide gases were used. In addition a selected number of trials were carried out with no gas applied.

To investigate the first objective, melt runs were carried out using a set workpiece power of 1.8kW, with a range of welding speeds from 0.10 to 3.0 m/min and with each of the gases. For the second objective, to examine the transition from deep penetration to conduction limited welding, the melt runs were carried out at a set speed of 0.6m/min and with applied beam power ranging from 1.8kW down to 0.7kW. The weld penetration, weld width and shape were measured.

Results

Weld penetration obtained with different plasma disrupting gases using a CO laser power of 1.8kW at the workpiece is shown in Fig.1. At high welding speeds above 1.0 m/min and up to 3.0 m/min, the weld penetration tended to be higher when using argon and no (in air) applied gas. Helium and CO ₂ gas gave up to 30% lower penetration at 3.0 m/min. Figure 2 shows the effect of the plasma disrupting gases when a 0.6 m/min welding speed was used. At low speeds, below 1.0 m/min, CO ₂ gas gave the deepest welds and resulted in full penetration (6.2mm) at speeds as high as 0.3 m/min. The other gases tended to restrict weld penetration, again by as much as 30% in the case of helium. However, welding was still possible even at very low speeds ( eg 0.1 m/min) with all the gases. At this speed all the welds were wide and fully penetrating. There is, however, an indication of gross porosity when welds are made in air under these conditions.

At a constant welding speed of 0.6 m/min, weld penetration reduced with decreased applied beam power from approximately 4mm at 1.8kW to 1mm at 0.7kW. Again the gas used was important in defining the penetration achieved, as shown by Fig.3. Use of helium resulted in the lowest penetration throughout the power range, with argon and CO ₂ giving deeper penetration. The graph (Fig.3) indicates step changes in weld penetration. At a power of 1.8kW, deep penetration high aspect ratio welds are produced in all the gases. In CO ₂ and air the top bead is especially narrow. With reducing power the penetration and aspect ratio reduces. The deep penetration weld spike is lost at approximately 1.2kW with helium and at 1.0kW with CO ₂ and argon. Ultimately beam coupling is almost completely lost at approximately 0.85kW in helium and at 0.7kW with CO ₂. Welds made in argon appear to retain reasonable coupling even at 0.7kW.

Discussion

Effect of applied gases

The type of gas used in high energy density welding, for either plasma disruption or shielding, can give rise to a number of effects. Depending on the gas and on the degree of ionisation brought about by the laser beam there may be one or a combination of the following effects:

• exothermic heating reaction
• cooling by heat transfer to the gas
• weld pool surface absorption and surface tension effects by the gas or gas products

The high penetration and high aspect ratio welding seen with CO ₂ gas may be a result of some exothermic heating and changes in the surface tension and flow of the welds. Free oxygen can have these effects, following release from separated CO ₂ mole molecules. This may also explain similar effects of welding in air, where in some cases the penetration is enhanced.

Helium tends to have a cooling effect due to low mass of the atom and tends to resist ionisation. For these reasons it is commonly used in high power CO ₂ laser welding to control detrimental plasma formation above the weld. However, in relatively low power CO laser welding, there seems to be little need for plasma control and helium merely serves as a cooling medium. The cooling effect may explain reduced weld penetration over the full range of trials when helium was used and the transition to conduction limited low aspect ratio welds at the relatively high powers of 0.9-1.2kW, compared to 0.7-0.9kW with argon and CO ₂.

Based on these results it would appear that cheaper gases than helium could be used in CO laser welding. This is a similar conclusion to that formed by Schellhorn and von Bülow [1] when they compared CO and CO ₂ laser welding.

Comparison with CO ₂ laser welding

Direct performance comparisons with a CO ₂ laser are difficult to make unless the beam profiles and power applied at the workpieces are fully understood. However, it is possible to make some general observations based on previous experience of CO ₂ laser processing.

The plasma formed above the CO laser weld appears visibly different to that seen in the CO ₂ laser welding. The colour is white with CO rather than blue. This may be related to welding performance at low speeds. The CO laser appears to give improved results at very slow welding speeds. At speeds less than 0.4 m/min, even relatively low power CO ₂ beams can be distorted by the plasma formed above the weld. Helium plasma control is used to overcome the plasma restrictions at low speed. Even with this plasma control there is usually no further increase in weld penetration as speed is reduced below 0.4 m/min. The CO laser beam appears to be less affected by plasma and hence allows a deeper maximum weld penetration at low weld speeds. In this respect, the CO laser appears to be similar to the Nd:YAG laser, which can also be used to make welds at very low speeds with little plasma disruption.

Conclusions

CO laser welding has been carried out under helium, argon and CO ₂ plasma disrupting gases, and in air at a range of powers and welding speeds. The following results have been found:

• In contrast with CO ₂ lasers, benefit is gained in terms of weld penetration by working at low speeds (<0.4 m/min) with a low power laser.
• Welds made in helium have the lowest penetration of the gases tested and it leads to loss of coupling with the workpiece at low applied powers (less than 1.2kW).
• Welds made in argon and CO ₂ indicate enhanced beam coupling at low applied powers and welds made in CO ₂ gas have a deeper, narrower aspect when high powers (eg 1.8kW) are used.
• Welds made in air are also generally quite deeply penetrating, but can be prone to porosity.

References