C A Olivier, P A Hilton, J D Russell
Paper presented at ICALEO Conference, 15-18 November 1999 in San Diego, CA, USA.
The beams from three separate 4kW Nd:YAG lasers have been optically combined to produce a workpiece power of 9kW CW, delivered in a single spot, via a 1mm diameter step index fibre. This high power beam has been used to investigate the welding of thick section C-Mn steel. In the down hand (1G) position, it was possible to produce fully penetrating melt runs, of good quality, at a speed of 0.7m/min in 12mm thick material. In the horizontal/vertical (2G) position, good quality, fully penetrating melt runs were obtained on 15mm thick material at a speed of 0.4m/min.
Industries which make use of steel plate (above 10mm thick) have been slow to adopt laser welding when compared to the rapid take up and productive use of laser welding seen in the industry sectors which utilise thin steel sheet. There are several reasons for this, one of which is the fact that until recently, little work on the development of laser welding procedures and validation of laser weld properties has been undertaken. This position is now changing in, for example, the shipbuilding industry, where classification society documents on 'guidelines for CO 2 laser welding in ship construction'  , can be found. Another important factor, is that at thicknesses of 10mm and above, only CO 2 lasers, with their relatively complex mirror beam delivery systems, have been capable of single pass welding. An interesting alternative to the CO 2 laser for thick section welding would be the Nd:YAG solid state laser, whose output at 1064nm, can be easily passed through a small diameter optical fibre. This type of beam delivery system lends itself well to the welding of both simple and complex structures. Another benefit of the ten times difference in wavelength when using the Nd:YAG laser over the CO 2 laser, is that the intense plasma cloud which eventually limits the weld penetration at low speeds, should not be as evident. Only recently, however, have CW Nd:YAG lasers become commercially available at output powers in the region 4-5 kW, and at best these lasers can only address the welding of 10mm thick plate at very low speeds. In order to influence further the development of high power solid state lasers, work is needed to demonstrate the effectiveness of Nd:YAG laser radiation for welding steel in the thickness range of up to about 20mm. In the absence of single source CW Nd:YAG lasers with sufficiently high powers for this thickness range, it is necessary to find a means of combining the beams from two or more lower power lasers, in order to achieve sufficient power for, say, a fully penetrating weld in 15mm thick steel. Combining the beams from Nd:YAG lasers can be performed in various ways [2,3] . In the work reported here, the beams from three separate Lumonics Multiwave Auto Nd:YAG lasers have been optically combined into a single fibre beam delivery system, to produce a workpiece power of up to 10kW CW. This is believed to be the highest CW Nd:YAG power reported. The beam combining unit, manufactured by HIGHYAG Lasertechnology, is equipped with three input receptacles, which admit the standard fibre delivered output from the Multiwave Auto lasers. Fibres of 0.6mm diameter are used between the three lasers and the beam combining unit. The beams emerging from these fibres have a beam parameter product of about 30mm.mrads. These beams are separately expanded, collimated and focused within the beam combiner. Provision is made to focus each beam onto the end of a 1mm diameter fibre, which is used to pass the combined beam to a processing head. In the process head, the expanding beam emerging from the fibre is collimated and focussed to a small spot. A schematic representation of the system can be seen in Figure 1.
Fig.1. Schematic representation of the beam combining unit
The design of the processing head allows for two focus spot sizes to be utilised. The first, which provides 1 to 1 imaging of the 1mm diameter fibre, was designed to produce a reasonably long stand-off distance (150mm) between the last lens element and the workpiece, without compromising the overall diameter of the process head, which does not exceed 100mm. The second optical system provides a greater stand-off distance of 220mm, while still maintaining the 100mm diameter for the process head. The compromise is the spot size, which in this latter case is increased to 1.4mm diameter. The beam combining unit itself was provided with several sensors at various points to assist alignment of the laser beams and fibre and to monitor the performance of the device during long term operation. Laser workpiece power has been estimated by measuring the power available from each laser separately and comparing these values with the measured optical power loss through the beam combiner and process head. The maximum process power available has been assumed to be the sum of the power available from the three separate lasers. Using this method, a maximum workpiece power of 9kW CW has been estimated. It is expected, however, that if all three lasers were equipped simultaneously with a new set of flashlamps, then 10kW would be available at the workpiece.
2 Welding experiments on thick section steels
Whilst it is clear that 9kW, delivered in a single focussed spot, would be very effective for the high speed welding of thin materials, the first series of work with the beam combining device has investigated what welding performance is possible in steels up to 20mm thick. Further work on both thick and thin section material is ongoing and will be reported in due course. For the experiments reported here, the beam combining process head was mounted on the arm of a Kawasaki JS30 robot, the latter being mounted on a 6m linear traverse.
It is generally expected that the 'plume' produced during keyhole laser welding is a partially ionised plasma producing strong atomic line emission in the UV and visible portions of the spectrum. Mueller et al  have spectroscopically investigated the plume produced during CW Nd:YAG laser welding of steel with a single Lumonics Multiwave Auto laser, and report little UV emission and significantly reduced atomic line emissions. They conclude that the Nd:YAG 'plume' is best described as a 'hot gas' plume. Work with 10kW CO 2 lasers has shown that the CO 2 induced plasma can be effectively controlled when welding 12mm thick section steel for example, with the use of a side gas jet at speeds down to 0.5m/min  . Below these speeds the CO 2 plasma becomes violent and difficult to control, the weld profile widening considerably and penetration reducing as more laser power is absorbed by the plasma. Initial welding results on C-Mn steels, with the beam combiner system at 7.5kW CW power and low speeds, showed that the Nd:YAG process was particularly energetic, in the orange and yellow parts of the visible spectrum, with an irregular and 'fiery' plume, which could extend several cms up and out from the weldpool region. The first series of experiments conducted at 9kW power at the workpiece therefore concentrated on suppression of this plume using a series of different systems, amongst which a co-axially supplied assist gas and a side jet applied gas were investigated. The benefits of the flexibility of fibre optic delivered beams are reduced somewhat if uni as opposed to omni directional plume suppression devices are required, and this is why the work with a co-axial gas system was first undertaken. The larger of the two available stand-off distances was chosen in the process head and a range of air knives were mounted horizontally below the lens and coverslip to provide protection against fume and spatter. Variables for the co-axial nozzle experiments included the nozzle exit diameter, nozzle exit to workpiece distance and gas type and flow rate. Variables for the side jet experiments included nozzle exit diameter, gas/laser beam interaction impingement point, gas type and flow, as well as nozzle attitude. The majority of results were obtained at welding speeds in the range 0.2 to 0.7m/min.
(a) Condition optimisation
When operating the co-axial 'plume' suppression system, a series of melt runs were made at 9kW workpiece power. The general observations from this work were:
- The deepest penetration observed (16.5mm) was produced with the smallest nozzle exit diameter and stand-off distance investigated. However, such welding trials also resulted in irregular and disturbed top beads.
- Smooth top beads could be obtained by either increasing the nozzle diameter or increasing the stand-off distance but at the expense of a significant loss in penetration.
- Sectioning of selected melt runs showed that, whatever the welding parameters, when using the co-axial gas delivery system, substantial widening of the top section of the melt profile was evident.
When operating with the side jet suppression system, gas flow rate, impingement position, nozzle diameter and nozzle attitude, all had a notable effect on the weld beam geometry. The general observations arising from this work were:
- Under optimised conditions, weld penetrations of up to 18mm were observed, with smooth and regular top beads. Radiographic examination of the melt runs showed low porosity content (B quality as defined by EN-ISO-13919-1). Typicalmacro sections and weld top beads can be seen in Figure 2.
- At these same conditions the part of the plume in the visible region was reduced, albeit with less visible control than is generally observed when trying to control plasma formation when welding with an equivalent CO 2 laser. All optimised sections at 0.2m/min exhibited some enlargement of the melt profile at the top of the weld.
Fig.2. Typical top bead and melt run profile obtained in 20mm C-Mn steel at 9kW workpiece power and with a side gas jet
The remainder of the results reported were made with a side jet suppression system.
(b) Fully penetrating melt runs
When working in the down hand position (1G) it was only possible to produce reasonable fully penetrating melt runs, in steel of thicknesses up to 12mm. Figure 3
shows a macrosection of such a melt run made with 9kW workpiece power at a speed of 0.7m/min. Any attempt to produce fully penetrating melt runs in 15mm thick C-Mn steel resulted in inconsistent penetration, accompanied by severe weld sagging as the molten metal was influenced by gravity. Figure 4
shows a typical macrosection. In order to suppress any detrimental effect of gravity, the working position was changed to horizontal-vertical (2G) position. In this case, it was possible to obtain B quality (EN-ISO-13919-1), fully penetrating melt runs in 15mm C-Mn steel, as can be seen from Figure 5
, which was made with 9kW workpiece power and a welding speed of 0.35m/min.
Fig.3. Fully penetrating melt run in 12mm C-Mn steel at 0.7m/min and 9kW workpiece power
Fig.4. Macro section showing undercut top bead in 14mm thick C-Mn steel
Fig.5. Fully penetrating melt run in 15mm thick C-Mn steel at a speed of 0.35m/min in the horizontal/vertical position (2G)
After producing approximately 100m of weld during the first series of experiments with the beam combiner, it was decided to change the optical system to weld with the 1mm diameter focus spot, (at the reduced stand off distance of 150mm). For these experiments, little deviation from the previously optimised conditions was allowed in order to protect the lens systems from damage. With the 1mm spot, fully penetrating melt runs in 15mm thick material could be produced, but at no gain in welding speed.
Process capability and characteristics
The system used to combine the power from three separate 4kW CW Nd:YAG lasers proved to be successful in delivering 9kW at the workpiece which, after careful optimisation of all process parameters, lead to the production of melt runs in steel plates up to 15mm thick, at speeds up to 0.4m/min.
When comparing the efficiency of the Nd:YAG laser beam with that of a CO 2 laser beam for welding of thick section steels, two aspects ought to be considered. It has previously been shown  , that welding of 15mm thick structural steel can be achieved with 9kW of CO 2 laser workpiece power at a speed of 0.5-0.6m/min, a speed which is slightly faster than that reported in the present work for the equivalent Nd:YAG laser power. One of the reasons for this difference is that the amount of molten metal in the Nd:YAG laser welding process is significantly higher than in the CO 2 process, as the diameter of the focused spot from the 9kW Nd:YAG laser is much larger than that generally produced by the CO 2 laser beam. However, welding of sections thicker than 15mm would certainly require a CO 2 workpiece power higher than 9kW, as decreasing the welding speed below 0.5m/min is not an option due to the aggressive plasma. In the case of the high power Nd:YAG laser system, the present work showed that control of the plume could be achieved at speeds down to 0.2m/min, thus resulting in weld penetrations up to 18mm with 9kW workpiece power. Therefore, for similar power levels, the maximum weld penetration achievable with a high power Nd:YAG laser beam is about 20% higher than with a CO 2 laser beam. Although the plume observed in Nd:YAG laser welding is visibly larger than with the CO 2 process, it appears to be less detrimental in limiting the weld depth due to speed limitations, in comparison to CO 2 laser welding, for reasons yet to be fully investigated but probably linked to the degree of ionisation of the metal vapour.
The recommended maximum thickness to be welded in the flat position is estimated at 12mm. For thicknesses up to 14mm, top bead undercut and heavy underbeads should be expected. For steel plate thicknesses above 15mm, the welding position should be changed to horizontal-vertical, as the combination of low welding speeds, large molten pool and gravity, results in process instabilities and severe weld sagging. If welding in the flat position is absolutely required in thicknesses higher than 12mm, two welding passes might be considered in order to reduce the influence of gravity on the weld geometry.
It has been shown  when CO 2 laser welding, a generally linear relationship holds between weld depth and the product of the laser beam intensity and the focal spot radius at a given welding speed. Figure 6 shows the same plots for Nd:YAG laser welding using workpiece powers in the range from 3.5kW (with a single laser) to 9kW (with the beam combining unit). The lower set of data (triangles) corresponds to a weld speed of 1.5m/min. The other three sets of data all correspond to a travel speed of 0.5m/min, but with focal spot sizes of 0.6, 1.0 and 1.4mm respectively. Although the data for 0.6 and 1.0mm diameter spots lie on a similar line, clearly the data taken with the 1.4mm diameter spot (all at 9kW power) does not fit the same relationship. This behaviour is not yet understood. It would appear, however, that the beam with the larger spot performs better, in terms of weld depth, with thicker materials at slower speeds, and that there are plume effects when welding with Nd:YAG lasers, which are power rather than intensity dependent.
Fig.6. Graph of weld penetration against laser beam intensity times spot radius for Nd:YAG workpiece powers between 3.5kW and 9kW
Nd:YAG laser manufacturers are currently striving to increase the beam parameter product of their equipment. While these steps may not be necessary for the welding of thicker materials, if flexible, fibre delivered beam delivery systems are required, the better beam quality of the laser will result in more compact process heads with the benefit of maintaining small physical diameters and combining this with large stand off distances.
When optimised welding conditions were used, both full and partial penetration welds in C-Mn steels exhibited very low levels of porosity and satisfied all the criteria required for the stringent (B) quality level, according to EN-ISO-13919-1. Again, this might constitute an advantage for the high power Nd:YAG laser welding process when compared with CO 2
laser systems, as porosity can sometimes be a problem when welding thick section steels with CO 2
lasers, especially when partial penetration welds are considered.
As mentioned above, the design of the processing head allowed for two focus spot sizes to be utilised: 1mm or 1.4mm. Such diameters are equal to two to three times the diameter of a typical high power CO 2
laser welding beam (about 0.5mm). The main consequence of the large spot diameters used in this work, is that the welds produced were larger than typical CO 2
laser welds. This could provide a significant advantage for high power Nd:YAG laser systems, as it is expected that a larger amount of molten material should extend gap tolerance when compared to what is currently acceptable for high power CO 2
The work described in this paper is part of a large research programme conducted at TWI and currently supported by a group of 22 industrial sponsors including companies in the following areas:
- - Nuclear
- - Shipbuilding
- - Structural steelwork
- - Line Pipe
- - Aerospace
- - Automotive
From these early results, the envisaged main interests for the heavier industries in high power CW Nd:YAG lasers, include the possibility of a flexible system capable of welding up to 18mm thick structural steel in one pass (possibly up to 20mm with a V preparation) or more in several passes, with an excellent weld quality and extended process tolerance, whilst maintaining the added advantage of low distortion when compared to arc welding.
The beams from three 4kW Nd:YAG lasers have been combined by using a 3-in-1 type beam combining unit which enabled a CW workpiece power of 9kW to be produced.
The type of and parameters for a plume control system were optimised for workpiece powers up to 9kW and speeds between 0.2m/min and 0.7m/min. The optimum system used a side jet oriented in line with the welding direction and performed satisfactorily after careful optimisation of its position, the type of gas being used and the gas flow rate.
A full parameter optimisation allowed a consistent, fully reproducible welding operation to be carried out and single pass, full penetration melt runs were successfully produced in 12mm and 15mm thick structural steel at welding speeds up to 0.7m/min and 0.4m/min respectively. All welds satisfied the stringent quality level imposed by the EN-ISO-1319-1 standard.
- 'Guidelines for the Approval of CO 2 Laser Welding', Lloyds Register of Shipping, March 1997.
- Bostanjoglo, G, Beck Th. and K Richter. '6-kW Nd:YAG Laser System for Welding Applications'. Lasers in Material Processing - SPIE Vol. 3097: pp. 129-136.
- Narikiyo, T., Miura H. Fujinaga S. Ohmori A. and K. Inoue. 'Combinations of Two Nd:YAG Laser Beams and their Welding Characteristics'. Journal of Laser Applications 11, 2, April, 99; pp. 91-95.
- Mueller, R.E., Gu H. Ferguson N. Ogmen M. and W.W, Daley. 'Real Time Optical Spectra Monitoring of Laser Welding Plumes'. ICALEO 98. C132.
- British Patent Application No.44269/76 'Plasma Control Gas Jet'.
- Dawes, C., 'Laser Welding - A Practical Guide' Abington Publishing. ISBN 1855730340.
- Rosen, H.G., 'Influence of System Parameters on Laser Material Processing Performance'. Proc. ICALEO 86. pp. 201-206.