Geert Verhaeghe and Paul Hilton, TWI Ltd
Paper presented at ICALEO 2005, 24th International Congress on Applications of Lasers & Electro-Optics, October 31-November 4, 2005, Miami, Florida, USA. Paper #511.
For welding, cutting and surface engineering applications, the use of high-power fibre-delivered beams from solid-state lasers offers many advantages. Only 10 years ago, the only available laser source of this type was a lamp-pumped Nd:YAG laser, a laser with modest beam quality. Notwithstanding this modest beam quality, these laser sources have been widely adopted for welding applications, particularly in the automotive industry. Today however, the range of fibre-delivered continuous wave (CW) laser beams has increased significantly, as has the available laser beam quality. The choice of laser source for a particular application is no longer obvious. This paper addresses the issue of how the laser beam quality of CW fibre-delivered lasers affects welding performance, by comparing a series of welds made in thin and thick section aluminium and steel, using a constant laser power from laser sources with different beam qualities.
Only a few years ago, anyone wishing to utilise the benefits of fibre-optic beam delivery from a laser source, would need to consider the capital outlay, the running costs, the reliability and the capability of the laser to perform the process required. What would not need to be considered, at that time, was the type of laser source, as the only CW fibre-delivered laser source available would have been the lamp-pumped Nd:YAG rod laser. Today's user of fibre-delivered laser power, must now add to the above list, the type of technology used in generating the laser beam. This is due to the rapid advances made over the last few years in the generation of CW infrared radiation laser sources with wavelengths suitable for transmission down an optical fibre. Diode-pumped Nd:YAG rod lasers, Yb-fibre lasers and Yb:YAG thin-disc lasers, are all now commercially available at powers up to at least 4kW.
Laser users are now also faced with the additional questions of laser beam quality and brightness, as the original lamp-pumped CW Nd:YAG rod lasers could only achieve a beam parameter product of about 20-25mm.mrad at 4kW, whereas some of the advanced laser technology mentioned above can operate at 4kW with a beam parameter product as low as 2mm.mrad.
A consequence of a high laser beam quality, i.e. a low value of beam parameter product, is that the beam can be focussed into a small diameter optical delivery fibre. This translates, on the processing side, for a focusing optic of a given diameter, into a better focusability of the laser beam into a minimum beam waist diameter (spot size). Or, for a given spot size and focusing optics of a given diameter, a higher beam quality will produce a higher brightness (defined as the ratio of the power density in the beam waist and the solid angle formed by the focusing beam cone). For a given spot size, a higher beam quality will produce a greater depth of focus at the beam waist. A high beam quality will also allow greater distances between the focusing lens and the workpiece, providing the practical advantage of minimising spatter damage to the optical system.
Each of the advanced laser technologies mentioned above has its advantages and, in some cases, its disadvantages. It is not the purpose of this paper to comment in this area, but rather to investigate the welding process capability of a range of CW fibre-delivered laser sources and beam focusing systems, with beam parameter products between 23 and 4mm.mrad, in a controlled series of experiments on aluminium and steel, to determine depth of penetration as a function of welding speed for a constant laser power of 4kW.
The general welding performance of CW Nd:YAG rod lasers, up to 10kW workpiece power, has been described by Russell and Hilton.  Verhaeghe and Hilton  have reported on the welding performance of a 7kW Yb-fibre laser and Weberpals, Russ, Dausinger and Hügel  have reported on the welding performance of a 4kW Yb:YAG thin-disc laser. Hügel  has reported on the welding performance of the diode-pumped Nd:YAG rod laser. Some work has also compared the performance of different laser sources, for example Bartel, Pathe, Roatzsch, and Weick  , who discuss the influence of beam quality when welding with Nd:YAG and CO 2 lasers and Ream  , who has compared the welding efficiencies of the Nd:YAG, disc and fibre lasers. Verhaeghe and Hilton  also compared fibre laser results with CO 2 and Nd:YAG data. A common problem with the comparisons in [2 , 5] and  , however, is that data is obtained over different (sometimes lengthy) time periods, on slightly different samples, with different spot sizes and at different powers, altogether making it difficult to draw concise conclusions. In the work reported here, all the experiments were performed on the same materials, all laser powers were measured with the same power meter, and the optical systems were chosen to produce, as well as a 'smallest' spot diameter, a spot diameter as close to 0.4mm, for each of the different laser sources used. The objective of this work was to investigate the performance of CW fibre-delivered laser sources for the welding of aluminium and steel, for a range of laser beam parameter products and focused spot sizes.
Experimental work programme
The characteristics of the lasers used in these experiments and their beam delivery systems are given in Table 1. Four different lasers, with beam parameter products from 23 to 4mm.mrad were used.
Table 1 Characteristics of the lasers and beam delivery systems used in experiments
|Laser||Laser 1||Laser 2||Laser 3||Laser 4|
|Nd:YAG||Yb:Fibre||Yb:Fib re||Yb:YAG disc|
|Delivery fibre diameter
|Collimating lens focal length
|Focusing lens focal length
|Nominal beam waist
|Measured beam waist(86% pts)
|Beam parameter product *
|Rayleigh length *
|Laser power at the workpiece
|Power density at beam waist for 4000W
(kW/mm 2 )
|Brightness (at 4000W)
(10 5 W/mm 2 .steradian)
|* Derived from beam caustic data
With these four lasers, seven combinations of delivery fibre, collimating lens and focusing lens, produced beam waists in the range from 0.61 to 0.14mm in diameter. In addition, four of the beam focusing systems were configured to produce a beam waist very close to 0.4mm in diameter. In the work reported here, all welding trials were made with the beam waist positioned on the surface of the workpiece, however the optical systems used produced a range of Rayleigh lengths from 9.9 to 1.1mm.
Promotec and Primes laser beam analysers were used to measure the beam caustic in the region of the beam focus. In one series of measurements, both the Promotec and the Primes units were used for the same laser with the same processing optics, revealing only small differences in measured values, within ±3%. The minimum beam waist diameters, beam parameter products, and Rayleigh lengths, given in Table 1 are those calculated by the beam analyser software using 86% intensity values.
In all cases a flat cover slide was used to protect the focusing optic. In all the experiments an Ophir 8000W power meter, with a claimed accuracy of ±5%, was used to measure the laser power in the focused beam.
In all cases the laser power was adjusted so that the welding trials reported here were carried out with a laser power of 4000W at the workpiece, although some of the lasers used for the trials, listed in Table 1, could operate well above this power.
Trials were carried out on 5mm and 10mm thickness S275 grade C-Mn steel and 5083-O aluminium alloy. The samples, 300mm by 75mm, were machined to give a tapered profile, such that the thickness along the 300mm length varied continuously from 1 to 5mm and from 4 to 10mm for the 5mm and the 10mm thickness samples respectively. The samples were clamped with the machined side facing down. To eliminate differences in heat-sink, the same clamping arrangement was used for all four lasers.
An airknife was used with each beam focusing system to reduce the risk of smoke, fume and/or spatter damaging the cover slide and focussing optic. The aluminium weld pool was shielded using 8 and 5 litre/min of argon applied to the top and bottom of the weld, respectively. No gas shielding was used for the steel welds.
Melt runs were completed in the flat (PA) position at various welding speeds, using a fixed laser power, with the laser beam perpendicular to and the laser spot positioned on the material surface. The points at which full penetration was lost, were noted and the thickness of the tapered sample measured at those points. The average of at least two of these thickness values per welding speed condition was used to construct graphs typical of those shown in Figure 1, relating depth of penetration to welding speed, for each of the seven systems listed in Table 1.
Results and discussion
Although the majority of the results discussed in this paper are related to the welds in aluminium, similar trends were observed for the melt-runs made in C-Mn steel. As stated above, in practical welding terms, a lower beam parameter product, i.e. better beam quality, translates into either a better focusability or, for a given focusing lens diameter and spot size, a higher brightness, larger stand-off distance and depth of focus.
Effect of spot size
Figure 1 shows the dependence of spot size on the depth of penetration achieved in aluminium, for different welding speeds, carried out using a laser with a beam parameter product of 4mm.mrad. The depth of welding is significantly affected by the spot size, as power density will increase for smaller spot sizes  .
Fig. 1. Depth of penetration as a function of welding speed for the welding of aluminium using a laser with a beam parameter product of 4mm.mrad focused to spot sizes of 0.14 and 0.40mm
Figure 1 shows that to obtain full penetration in 4 and 6mm thick 5083-O aluminium alloy for instance, approximately 60% and 21% higher welding speed is possible with a 0.14mm spot size, compared with a 0.4mm spot size. This graph also shows, on the other hand, that welding at fixed speeds of 5 and 2m/min, means that 5.0 and 6.8mm of aluminium can be penetrated with the smaller spot, compared with only 4.3 and 6.4mm, using the larger spot size. This is an increase in depth of penetration of 16 and 6% respectively.
The same effect was observed for the systems with beam parameter products of 23 and 7mm.mrad, where the smaller spot size enhanced the welding performance, giving improvement in speed or depth of penetration. The smaller the decrease in spot size, however, the smaller the gain in welding performance. This is demonstrated in Figure 2 for a laser with a beam parameter product of 23mm.mrad, where the spot size was reduced from 0.61 to 0.44mm, i.e. a 39% reduction, compared with the data given in Figure 1, where the spot size reduction was nearly three-fold for a laser with a beam parameter product of 4mm.mrad. This resulted in a 19 and 17% increase in welding speed for obtaining full penetration in 4 and 6mm thickness aluminium, compared with 60 and 21% for the results shown in Figure 1. For all lasers, this enhanced welding performance was speed dependent, with less of an increase in welding performance noticeable for the slow speeds/thick sections.
Fig. 2. Depth of penetration as a function of welding speed for welding aluminium using a laser with a beam parameter product of 23mm.mrad focused to a 0.44 and 0.61mm spot size
The weld cross-sections in Figure 3, taken from bead-on-plate runs, show the influence of the spot size on weld geometry. Both welds were carried out using a system with a beam parameter product of 7mm.mrad and at a welding speed of 15m/min, the laser focus on the material surface and using 4000W of power at the workpiece. The cross-section on the left was made with a spot size of 0.34mm, whereas the cross-section on the right was made with a 0.20mm spot size. An area analysis of both cross-sections, using AutoCAD software, revealed no significant difference in molten area, i.e. less than 2.5%, between the two spot sizes. This would indicate that the melting efficiency is the same in both cases.
Fig. 3. Cross-sections of melt runs completed in 5083-O aluminium alloy using a laser with a beam parameter product of 7mm.mrad at a welding speed of 15m/min, using 4000W of laser power at the workpiece and a spot size of0.34mm (left) and 0.20mm (right) positioned on the material surface
Figure 4 shows the gain in depth of penetration for a spot size of 0.14mm relative to a spot size of 0.4mm, as a function of welding speed, for the system with a beam parameter product of 4mm.mrad. The graph shows that the advantage of moving to a small spot size is limited to less than 10% for welding speeds lower than around 7.5m/min, but increases sharply, and apparently linearly, for higher speeds. This indicates that the mechanism that determines depth of penetration, changes significantly at a welding speed of around 7.5m/min.
Fig. 4. Percentage gain in depth of penetration in aluminium as a function of welding speed, for a spot size of 0.14mm, relative to a spot size of 0.40mm, using a laser with a beam parameter product of 4mm.mrad
A significant speed dependence can also be observed in Figure 5, which shows the depth of penetration obtained for each of the seven systems detailed in Table 1, plotted against the inverse of the spot size, for the three different welding speeds of 1, 5 and 15m/min.
Fig. 5. Depth of penetration in aluminium plotted as a function of the inverse spot size at welding speeds of 1, 5 and 15m/min
Up to a value of 3 mm -1, i.e. corresponding to spot sizes between 0.3 and 0.61mm used in the trials here, the data points show an approximate linear behaviour, with different slopes, corresponding to the different welding speeds, as expected. What is interesting, is that above l/spot size values of 3 mm-1, changes in the slope of the data become obvious. At the slowest welding speed of 1m/min, no additional gain in penetration whatsoever can be seen, for any spot size below 0.3mm in diameter.
This behaviour is similar for the welding speeds of 5 and 15m/min, however, at these speeds some increase in depth of penetration is still observed above the value of 3 mm -1 , although the inflection point is still clear. Weberpals  has observed similar effects when welding steel and aluminium using a thin-disc solid-state laser. In his study, the proportionality was maintained for spot diameters as low as 0.20mm, but below this spot size, the weld penetration actually fell. Weberpals indicated that the divergence angle of the focused beam could play a part in this. However, in the work reported here, when depth of penetration was plotted against beam divergence angle, no dependence was evident.
It has been shown by Greses  that when laser welding steel with an Nd:YAG laser, no plasma is generated above the weld pool. Instead, an energetic plume of thermally excited vapour with a typical black body spectrum can be seen. Olivier  noted that, when welding with a high power Nd:YAG laser at speeds around and below 1m/min, additional depth of penetration could be gained when this energetic plume was displaced using a 'heavy' gas, such as argon. Helium shielding, as used in the work reported here, would have no effect on this plume and it is therefore possible that the trends observed in Figure 5 could be due to attenuation of the incident laser beam by the plume generated. An alternative explanation for the behaviour shown in Figure 5, could be due to the shape and stability of the laser-induced keyhole. If this was the case, however, it might be expected that the melting efficiency would change and differences would be seen in the shape of the weld penetration, particularly in the region of the inflection points in Figure 5. Although this is not obvious from the cross-sections shown in Figure 3, a further evaluation in this area will be undertaken. A closer study of plume and possibly plasma formation, at the extremes of power density, and the dynamics of molten metal flow is also recommended to help understand this behaviour.
Effect of beam quality
Another useful comparison is to investigate the depth of penetration achieved using the same spot size, as a function of beam parameter product. As can be seen from Table 1, certain combinations of laser and available optics produced spot sizes close to 0.4mm. Figure 6 compares the welding performance, for a spot size close to 0.4mm, at the two extremes of beam parameter product used in this work, i.e. 4 and 23mm.mrad. Figure 6 shows gains in both depth of penetration and welding speed, when using the system with a beam parameter product of 4mm.mrad, for all speeds above 1m/min.
Based on the results shown in Figures 1 and 2, which demonstrate that welding performance in aluminium only marginally improves for small reductions in spot size, the gains in performance shown in Figure 6 can realistically be attributed to the change in beam parameter product (with only a 10% difference in spot size between the two curves displayed). This enhanced process performance for the system with a lower beam parameter product, translates as a larger depth of penetration for a constant welding speed, or a faster welding speed for a given thickness of aluminium plate. The gain in depth of penetration as a result of improving the beam quality, appears similar for both low and high welding speeds, e.g. 1 and 15m/min.
Fig. 6. Depth of penetration as a function of welding speed for welding aluminium using beam parameter products of 4 and 23mm.mrad (focused into 0.40mm and 0.44mm spots, respectively)
This beam parameter product comparison, however, may be too simple. When discussing fibre lasers in particular, many people refer to the term brightness (in some cases incorrectly). Laser beam brightness is defined as the ratio between the power density in the beam focus and the solid angle determined by the beam cone emerging from the focusing lens. As mentioned earlier, for measurements made in this work, plotting the depth of penetration for a given welding speed against beam cone angle showed no obvious trend. However, when the depth of penetration is plotted against the brightness, the results, as shown in Figure 7, are more interesting.
Fig. 7. Depth of penetration in aluminium plotted against the brightness in the focused laser spot for welding speeds of 1, 5 and 15m/min. The lines are a guide to the eye.
The depth of penetration increases with an increasing laser beam brightness up to around 33x10 5 W/mm 2 .steradian, which appears to be the optimum brightness for maximising the depth of penetration when welding aluminium, regardless of travel speed. Beyond this brightness, the depth of penetration apparently reduces, although it is noted that this behaviour is currently based only on one data point. This means that, using an 'optimal' brightness of around 33x10 5 W/mm 2 .steradian in equation (1), the beam transmission parameters can be determined to give a maximum depth of penetration for the welding of aluminium.
||P = power measured at the workpiece, W
F = focal length of focussing lens, mm
ω0 = beam waist radius, mm
D = laser beam aperture, mm
From the data points in Figure 8, it can also be seen, for instance, that for a welding speed of 1m/min, to achieve a depth of penetration of 8mm, it does not appear necessary to use a laser system with a brightness of more than about 10 6 W/mm 2 .steradian. The same is true for a depth of penetration just under 4mm for a welding speed of 5m/min and also for 2mm depth of penetration for a welding speed of 15m/min.
Figure 8 shows the same plot as that in Figure 7, but for the welding of steel instead of aluminium. In the case of steel, it would appear that the optimum brightness for achieving the largest depth of penetration differs slightly depending on the welding speed, i.e. between 32 and 38 x10 5 W/mm 2 .steradian for welding speeds of 1 and 15m/min, respectively.
The reasons for this optimum brightness and its apparent independence of welding speed and material, are not fully understood at the time of writing this paper.
Fig. 8. Depth of penetration in steel plotted against the brightness in the focused laser spot for welding speeds of 1, 5 and 15m/min. The lines are a guide to the eye.
Figures 9 and 10 demonstrate the performance improvement, for aluminium and steel respectively, achieved in this experiment, using the system with the smallest beam parameter product and spot size, i.e. 4mm.mrad and 0.14mm, compared with the system with the largest beam parameter product and spot size, i.e. 23mm.mrad and 0.61mm. This shows that choosing a laser source with a low beam parameter product and a small spot size can improve the welding performance for both steel and aluminium.
Fig. 9. Depth of penetration in aluminium plotted as a function of welding speed for the two extremes of focused spot size and beam parameter product used in the experiments.
Fig. 10. Depth of penetration in steel plotted as a function of welding speed for the two extremes of focused spot size and beam parameter product used in the experiments.
Choosing a laser welding system
The results shown in Figure 5, indicate that in choosing an optimum welding system with the capability to process over a range of welding speeds and material thicknesses, there would appear to be no real benefit in using a focused spot smaller than 0.3mm in diameter. Combining this figure with the 'optimum' brightness figure of 33x10 5 W/mm 2 .steradian, indicates that this should be achievable using a lens with a focal length of around 350mm, for a focusing system with an aperture of 50mm, for instance. If the numerical aperture of the beam delivery fibre is of the order of 0.2, then, in order to achieve the 0.3mm spot size with a 175mm focal length collimating lens, a delivery fibre with a diameter of about 0.15mm would be necessary. In order to use such a fibre, the beam parameter product of the laser required, would then have to be between 5 and 7mm.mrad.
The performance of a series of CW fibre-delivered laser systems has been assessed, at a laser power of 4000W measured at the workpiece, for the welding of aluminium and steel. The work investigated spot sizes from 0.61 to 0.14mm and beam parameter products from 23 to 4mm.mrad. The work has allowed the following conclusions to be drawn:
- For a given laser beam quality, a smaller spot will generally produce either a faster welding speed for a given depth of penetration, or an increase in depth of penetration for a given welding speed. The smaller the reduction in spot size, the smaller the gain.
- For a laser with a fixed beam parameter product of 4mm.mrad, the gain in depth of penetration when moving from a spot diameter of 0.4mm to one of 0.14mm, is small for welding speeds less than 7.5m/min, but then increases linearly above this speed.
- For a given spot diameter of 0.4mm, a laser with a better beam quality generally produces either a faster welding speed for a given depth of penetration, or an increase in depth of penetration for a given welding speed.
- Little increase in depth of penetration can be seen, for any beam quality, for spot diameters smaller than 0.3mm.
- The graph of depth of penetration versus laser beam brightness, shows, that for the welding speeds between 15 and 1m/min, there is an 'optimum' brightness for maximising the depth of penetration in both aluminium and steel. This optimum brightness is between 32 and 38 x10 5 W/mm 2 .steradian, regardless of welding speed or material.
The authors would like to thank Anthony Elliott, Paul Fenwick and Harvey Whitmore, from TWI, for their assistance in carrying out the welding trials. The authors are also grateful for the assistance of colleagues at Trumpf and IPG in realising some of the work described in this paper. This study has been made possible with contributions from the Yorkshire and Humber Regional Development Agency.
- Russell, J.D., Hilton, P.A. (2001) The development of a 10kW Nd:YAG laser facility, in Proceedings of the 7 th International Welding Conference: High Productivity Joining Processes: fundamentals, Applications Equipment, Aachen, Germany, vol.1, 299-310.
- Verhaeghe, G., Hilton, P.A., (2005) Battle of the sources - using a high power Yb-fibre laser welding steel and aluminium, in Proceedings of the 3 rd International WLT Conference of Lasers in Manufacturing, Munich, Germany, 33-38.
- Weberpals, J., Russ, A., Dausinger, F. Hügel, H. (2005) Influence of the focus diameter in laser welding with thin disk laser, in Proceedings of the 3 rd International WLT Conference of Lasers in Manufacturing, Munich, Germany, 39-42.
- Hügel, H. (2000) New solid-state lasers and their application potentials, in Optics and Lasers in Engineering, vol34, 213-229.
- Bartel, W., Pathe, D., Roatzsch, R., Weick, J.M. (1997) Influence of beam quality when welding with Nd:YAG and CO 2 lasers, in Proceedings of the Lasers in Materials Processing Conference, Munich, Germany, vol.3097, 147-156.
- Ream, S.L. (2005) Disc and fibre gain ground, Industrial Laser Solutions, Feb 2005, 15-19.
- Greses J, Hilton P A, Barlow C Y, Steen W M (2002) Plume attenuation under high power Nd:YAG laser welding, in Proceedings of ICALEO 2002, Scottsdale, AZ, US, vol.94, 10pp.
- Olivier C A, Hilton P A, Russell J D (1999) Materials processing with a 10kW Nd:YAG laser facility, in Proceedings of ICALEO 1999, San Diego, CA, USA, section D, 233-241.
Geert Verhaeghe is a Mechanical (MSc) and European Welding Engineer (EWE), who started his career in 1994 at OCAS, the Belgian R&Dmp;D centre of Arcelor, working on tailor-welded blanks for the ULSAB (Ultra Light Steel Auto Body) project. Since 1996, he has been a Senior Project Leader at TWI Ltd, where he has managed a variety of projects on a range of arc and laser processes and applications. His particular experience is in the welding of aluminium, hybrid laser-arc welding and, more recently, fibre laser processing.
Paul Hilton is Technology Manager at TWI where he has the responsibility for the strategic development of laser material processing. He is a past president of the UK's Association of Industrial Laser Users and is a current board member of the European Laser Institute.