Plasma/Plume Control for High Power Nd:YAG Laser Welds

C H J Gerritsen
TWI Ltd, Cambridge, United Kingdom

C A Olivier
TWI Ltd, Cambridge, United Kingdom (currently with British Energy, Gloucester, United Kingdom)

Paper presented at 6th International Conference on Trends in Welding Research, 15 - 19 April 2002, Callaway Gardens Resort, Pine Mountain, Georgia, USA

Abstract

Nd:YAG lasers are now commercially available with power levels up to 10kW, and therefore of increasing interest for welding of thicker sections (e.g. up to 15mm in steel), for example for off-road vehicles and shipbuilding applications. However, because travel speeds for deep penetration laser welding are generally relatively low (<1m/min), plasma/plume formation above the keyhole can cause a reduction in laser power at the workpiece, limiting penetration depth and weld quality. Although this issue has already been extensively investigated for CO ₂ laser welding, the results may not be directly transferable to Nd:YAG laser welding since the wavelength of the laser light as well as the plume characteristics are different. This paper reports on an extensive investigation of different nozzle designs, process gases and set-up parameters, specifically for Nd:YAG laser welding of thick section C-Mn steel. Characterisation was performed in terms of weld penetration and quality, for melt runs made at 7.5 and 9kW of laser power. Using the optimised set-up, penetration levels of 15mm could then comfortably be achieved.

1. Introduction

The technical and economic benefits of high power laser welding are already exploited in many industrial sectors, mainly for welding of relatively thin sections (typically up to 6mm for steel). The lasers used for these applications are almost exclusively CO ₂ and Nd:YAG lasers, which are now commercially available at power levels up to 45kW and 10kW, respectively.

The availability of such high power laser sources generates an increasing interest in single-pass welding of thicker sections (i.e. greater than 10mm thickness in steel), for example in the areas of shipbuilding, structural steel work, off-road vehicles, power generation, containment plant, offshore structures and line pipe. The advantages of laser welding for these applications result from the deep penetration, reducing the number of weld passes and the thermal distortion and increasing the joint completion rate.

Deep penetration welds are made via the so-called keyhole welding mechanism. A keyhole is generated if the power density in the focused laser spot is high enough (~10 ⁴W/mm ^{2 [1]} ) to cause melting and vaporisation of the metal before significant quantities of heat are removed from the processing zone via thermal conduction. The keyhole essentially is a cylindrical hole with molten walls that are kept from collapsing mainly by the vapour pressure inside the keyhole. Via inverse Bremsstrahlung ^[2] , the vapour inside the keyhole is ionised and forms a plasma of ions and free electrons, which dramatically improves the energy coupling between the laser beam and the workpiece.

To achieve deep penetration laser welds, however, it is usually necessary to weld at travels speeds of less than 1m/min, at which speeds the process can be adversely affected by vapour and plasma periodically escaping from the keyhole and forming a cloud above it. In the cloud, the vapour and plasma absorb and diffusely re-radiate some of the laser beam's energy. This can generally be seen from the weld by a decreased penetration depth and widened top bead (which may lead to the so-called 'nail-head' or 'wine-glass' weld profile).

Whereas with CO ₂ laser welding the cloud is thought to be a partly-ionised gas plasma consisting of ions, electrons and neutral atoms ^[2-4] , in Nd:YAG laser welding it is thought to be merely a 'hot gas' or vapour plume consisting of neutral atoms only ^[4-6] . This arises because the energy absorption coefficient for inverse Bremsstrahlung is proportional to the square of the wavelength, which makes the cloud more transparent to Nd:YAG (wavelength λ = 1.064µm) than CO ₂ laser light (wavelength λ = 10.6µm). For that reason, it may be expected that there are differences in the most effective plume control measures for CO ₂ and Nd:YAG laser welding as well.

Considerable research effort has been spent over the years to investigate this plasma/plume formation and to develop techniques for suppressing or at least reducing it. ^{[e.g. 2,3,7-10]} So far, this research has mostly concentrated on high power CO ₂ laser welding, which has been more commonly used because of the higher powers available and the lower cost per kilowatt of laser power. Nonetheless, high power Nd:YAG lasers (typically 4kW) are now available, and of greater application due to the flexibility that the optical fibre beam delivery of Nd:YAG lasers can give. Furthermore, with new techniques such as diode pumping and disc lasers, the available Nd:YAG laser power is set to rise.

This paper describes research specifically into plume control for deep penetration Nd:YAG laser welding.

2. Objectives

The objectives of the research reported here were:-

• To establish the effectivity of different process gases and gas mixtures for plume control during high power Nd:YAG laser welding;
• To establish an effective gas delivery system for plume control during high power Nd:YAG laser welding.

3. Experimental approach

3.1 Materials

Two C-Mn steels were used in samples of 150x300mm, with a thickness of 14mm and 15mm respectively. The chemical compositions of the steels are listed in Table 1.

Table 1: Chemical compositions (in weight%) of the steels used.

Thickness	C	Mn	Si	Ni	Cr	Cu	Al	P	S
14mm	0.12	1.28	0.009	0.019	0.015	0.014	0.036	0.015	0.005
15mm	0.17	1.08	0.25	0.021	0.019	0.022	0.029	0.014	0.014

3.2 Equipment

To achieve laser powers of 7.5 and 9kW (laser powers quoted in this paper are at the workpiece), three Nd:YAG laser sources were used simultaneously. Each laser was capable of an average continuous wave (CW) workpiece power of 3.5kW. The output from each laser was guided through a 0.6mm core diameter fibre optic cable to a bespoke beam combining unit. Inside this device, the output from all three fibres was focused onto the end of a single fibre optic cable of core diameter 1mm and this fibre optic cable transmitted the combined laser output to the focusing head, giving a laser power at the workpiece of up to 10kW. The focusing head comprised a recollimating lens and focusing lens producing a focused spot of nominally 1.4mm in diameter at a stand-off distance of 220mm. The optics were protected from contamination and damage by weld spatter and fume by three air knives, aligned below an anti-reflection-coated glass cover slide.

3.3 Experimental approach

Rust was removed from both the top and bottom surface of the samples by milling and the samples were degreased with acetone prior to welding. Partial penetration melt runs were made in the different types of steel at laser powers of 7.5 and 9kW.

Firstly, to establish the most suitable plume control set-up, different gas delivery systems were investigated. A schematic of each delivery system tested and the parameters under investigation can be found in Figure 1. Tested were the angled jet, which is the preferred system for CO ₂ laser welding, plus modifications to attempt and improve the performance and tolerance to set-up. As helium is generally considered to be the best plasma control gas when CO ₂ laser welding ^{[e.g. 11]} , it was used for these experiments. The laser power used was 7.5kW at a travel speed of 0.5m/min; the steel used was 14mm thickness C-Mn steel. The commonly used co-axial shielding nozzle was also tested (at 9kW), but this was already reported earlier ^[12] .

Secondly, using the best performing delivery system, the influence of the process gas itself was tested. The common process gases helium, nitrogen, argon and carbon dioxide as well as mixtures (helium-nitrogen, helium-carbon dioxide, helium-oxygen, argon-oxygen) were studied. An indication of the different properties of the unmixed gases can be found in Table 2. The laser power used was again 7.5kW and the travel speed 0.5m/min; the steel used was 14mm thickness C-Mn steel.

Table 2: Properties of the used process gases (as indication only).

Gas	Atomic mass	Thermal conductivity at 1200K (W/m/K)	Ionisation potential (eV)
He	2	0.4	24.5
N 2	14	0.08	15.7
O 2	16	0.08	12.5
Ar	18	0.05	15.7
CO 2	22	0.08	13.8

Lastly, the performance of the optimised set-up was further evaluated when attempting full penetration melt runs in 14mm thickness at 7.5kW and in 15mm thickness steel at 9kW of laser power.

3.4 Weld quality assessment

All melt runs were firstly evaluated on their external appearance. If satisfactory, they were also radiographed in order to detect the presence of porosity, solidification cracks and other internal defects. The radiographic weld quality was assessed against the criteria set in the laser welding standard BS-EN-ISO 13919-1:1997 ^[13] and assigned a quality Class (stringent (B), intermediate (C), moderate (D)).

Transverse cross-sections were prepared at a location thought to be indicative of that melt run. The sections were ground, polished, and etched in a 2% solution of nital to reveal the weld bead and heat affected zone. The cross-sections were assessed in terms of penetration, bead shape, and incidence of solidification cracking and porosity.

4. Results

4.1 Evaluation of gas delivery systems

Circular cross-section angled jet

The conditions that were used for these experiments can be found schematically in Figure 1. Changes in the nozzle diameter, gas flow rate or impingement position for a circular nozzle all had a significant effect on the weld bead geometry and penetration depth. A series of observations could be noted from these experiments:

• The deepest penetration depth observed was 12mm. This was achieved with as plume control parameters:-

  - Nozzle diameter 2mm, impingement position +1mm, helium flow rate 30 l/min;
  - Nozzle diameter 2mm, impingement position +2mm, helium flow rate 30 or 40 l/min;

   

• Deep penetration melt runs were associated with a small plume and a smooth top bead. Radiographic examinations of these melt runs showed low levels of porosity. However, the occurrence of short solidification cracks, although rare,was noted.

   

• When optimum parameters were not used, the penetration depth decreased to 8mm and the plume was seen to significantly increase in size. The top bead of all low penetration melt runs showed a 'pulsing effect', which was also linkedto extensive cracking. An example of a top bead showing this pulsing effect can be seen in Figure 2.

   

• No melt runs of satisfactory penetration depth and weld quality were produced with the 1.2mm or 4mm diameter nozzles.

Rectangular cross-section angled jet
A nozzle with a rectangular cross-section did not provide any improvement or increased tolerance to set-up, even though it was designed to spread the gas jet more widely and evenly across the interaction zone and thereby improve the tolerance to positioning. In fact, within the range of parameters investigated, most melt runs exhibited a medium to large plume, a 'pulsing' or disturbed top bead and shallow penetration.

Rectangular cross-section horizontal jet
A rectangular cross-section nozzle with its axis parallel to the workpiece did not enable the production of a melt run of acceptable quality over the range of parameters investigated. Most melt runs exhibited a 'pulsing' top bead and relatively low penetration depths (8-10mm). In these experiments, the position of the nozzle with respect to the workpiece and the laser beam-material interaction point proved to be of little influence on the penetration depth. In addition, it was noticed that the weld shape and appearance were not directly linked to the visible plume size, as seen previously.

Circular cross-section angled jet with shaped nozzle
Only two experiments were performed with an angled circular cross-section jet with shaped exit nozzle as indicated in Figure 1c. Neither of the experiments produced a melt run of acceptable quality. Even the use of very high helium flow rates and gas bottle pressures did not allow control of the plume. Both melt runs exhibited pulsing top beads and small penetration depths.

Summary
From the experiments, it was concluded that of the gas delivery systems tested with helium, the optimum conditions, enabling 12mm penetration in C-Mn steel with a Nd:YAG workpiece power of 7.5kW and at a travel speed of 0.5m/min, were as follows:-

• Circular gas jet of diameter 2mm;
• Jet oriented in line with the welding direction and trailing the beam at a 35° angle to the workpiece;
• 10mm nozzle-to-workpiece stand-off distance;
• Impingement point 2mm ahead of the beam;
• Helium flow rate of 40 l/min.

4.2 Evaluation of different process gases

Several different process gases and gas mixtures were tested using the best performing gas delivery set-up (circular cross-section angled jet). The experiments showed that, for every gas or mixture, the flow rate had a significant effect on the weld bead geometry and penetration depth. A re-optimisation was therefore performed for each in terms of flow rate and nozzle position. The laser power used was again 7.5kW at a travel speed of 0.5m/min.

The maximum penetration depth achieved was 12.5mm with argon, nitrogen, carbon dioxide or a 90% helium-10% oxygen mixture. The melt runs produced with these process gases, apart from those made with carbon dioxide which were not radiographed, exhibited low or acceptable porosity levels (to Class B (stringent) of BS-EN-ISO 13919-1:1997 ^[13] ).

Very short cracks (1mm) were occasionally detected in some of the melt runs made with nitrogen. However, although not further investigated, this was not thought to be related to the process gas. It is interesting to note, that the use of nitrogen did allow the plume to be controlled, whereas with high power CO ₂ laser welding, it tends to give a very hot and fiery plume.

Argon was found to give the best overall results because it was most tolerant to set-up variations, is inert and does not have an alloying effect. It was therefore concluded that the optimum set of welding conditions, enabling a 12.5mm penetration depth in C-Mn steel with a Nd:YAG laser power of 7.5kW at a travel speed of 0.5m/min, was as follows:-

• Circular gas jet of diameter 2mm;
• Jet oriented in line with the welding direction and trailing the beam at a 35° angle to the workpiece;
• 10mm nozzle-to-workpiece stand-off distance;
• Impingement point 2mm ahead of the beam;
• Argon flow rate of 20 l/min.

4.3 Verification of optimised set-up

Experiments at 7.5kW

Full penetration melt runs were attempted in 14mm thick C-Mn steel with a workpiece power of 7.5kW. The procedure followed was to use the optimum plume control parameters established earlier and reduce the travel speed until full penetration was achieved. However, it was noticed that as the travel speed changed, the optimum position of the angled jet and argon gas flow rate changed as well, and the set-up had to be amended accordingly.

Full penetration was achieved at travel speeds of 0.30 and 0.35m/min ( Figure 3) using the following plume control set-up:-

• Circular gas jet of diameter 2mm at a 35° angle;
• 10mm nozzle-to-workpiece stand-off distance;
• Impingement point 3mm ahead of the laser beam at the workpiece surface;
• Argon gas flow 30 l/min.

The melt run made at 0.30m/min showed little porosity, acceptable under Class B according to BS-EN-ISO 13919-1:1997 [13] . The melt run made at 0.35m/min was not radiographed. As was to be expected for full penetration in the PA position in this plate thickness, significant weld metal sagging occurred.

Experiments at 9kW
As full penetration melt runs were attempted in the 15mm thickness steel as well, the welding position was changed to the PC position, to prevent weld metal sagging.

Full penetration melt runs could indeed be achieved using the optimised angled jet plasma control at a laser power of 9kW and a travel speed of 0.30m/min at zero focus ( Figure 4) and at 0.35m/min at -3mm focus ( Figure 5). No significant change in plume control set-up was required to cope with the change of welding orientation, although plume control was still necessary. The plume control set-up used was:-

• Circular gas jet of diameter 2mm at a 35° angle;
• 10mm nozzle-to-workpiece stand-off distance;
• Impingement point 2mm ahead of the laser beam at the workpiece surface;
• Argon gas flow 20 l/min.

The melt runs exhibited smooth top beads and consistent under-beads. All melt runs contained very low porosity levels and in that sense qualified as Class B according to BS-EN-ISO 13919-1:1997. ^[13]

5. Discussion

5.1 Introduction

The extent of experimental trials undertaken in this study illustrates the large number of process variables involved, and the importance of controlling the plume effectively to achieve deep penetration and good quality melt runs - or welds - using high power Nd:YAG laser beams. It should also be noted that, due to the large number of variables investigated (materials, process gases, gas delivery systems, etc), the trials were not performed to a statistically designed set of experiments. Nonetheless, it is felt that the results do indicate general trends.

5.2 Observations

When reviewing all the experiments, the following general observations were made:-

• As was to be expected, plume control becomes increasingly difficult as the laser power increases and/or the travel speed decreases, because the increased input of laser energy tends to promote plume formation.
• For a constant laser power and travel speed, the diameter of the angled jet, its position (i.e. impingement point) and the process gas and flow rate all have a significant effect on the weld profile and penetration.
• The ionisation potential of the process gas has no obvious importance in the control of the Nd:YAG 'hot gas' plume. In fact, all gases and gas mixtures could be used to make deep penetration melt runs, providing the set-up of the gas jet was correct for that specific gas. This observation differs from what is commonly known for high power CO ₂ laser welding (at travel speeds of less than 1m/min) where helium is, by far, the best process gas for plasma control due to its high ionisation potential and thermal conductivity. With Nd:YAG, even nitrogen could be used, which generally gives a very fierce plasma when used in high power CO ₂ laser welding.
• The optimum gas flow rate is gas specific. Lighter gases (He) and gas mixtures (He-CO ₂, He-O ₂) require high flow rates (40-60 l/min). All heavier gases (Ar, N ₂, CO ₂) and gas mixtures (Ar-O ₂), despite significantly lower ionisation potentials and thermal conductivity, require lower flow rates (10-30 l/min) to maintain melt runs of similar or even higher penetration.
• When attempting to achieve full penetration melt runs, the position of the gas jet and consequently the gas flow has to be re-optimised if either or both the laser power and the travel speed vary significantly.

The parameters which critically affect the success of the welding operation therefore include:-

• Nozzle size and shape;
• Positioning of the plume control jet;
• Gas type for plume control and shielding of the weld;
• Gas flow rates;
• Material composition and thickness;
• Laser power;
• Travel speed.

Most of these parameters mutually influence one another.

5.2 Mechanism

When reviewing all the observations made, more than preventing ionisation of the plume and cooling it, the role of the gas jet appears to be to assist in keeping the keyhole open by hitting a precise location, at the front of the keyhole, with a specific momentum. Should the size or the orientation of the keyhole vary, due, for example, to a change in laser power, travel speed or penetration depth, the optimum position of the gas jet and the optimum gas flow rate will change.

The actual mechanism by which the gas jet influences the formation or displacement of the plume is not readily understood. It could be that the gas jet prevents its escape and forces it into the keyhole and/or pushes the plume that has escaped from the keyhole away from the interaction zone.

Whatever its exact role, accurate positioning of the gas jet was found to be critical for plume control. Furthermore, because the angled jet is non-axisymmetric with regard to the laser beam, it will be very difficult to apply to non-linear joints, unless a robotic device is used which can keep the angled jet aligned with the joint line and keyhole. But even then, the influence of the change of direction on the keyhole may affect the efficiency of the plume control device.

6. Conclusion

No one recipe for plume control suitable for all Nd:YAG laser welding applications can be given, due to the complexity and interdependence of many factors. However, this study has shown that an angled jet is a good starting point for most deep penetration applications in steel when using the following set-up and conditions:-

• Circular cross-section, 2mm in diameter;
• Jet following the laser beam in line with the seam and at an angle of 35° to the workpiece surface;
• 10mm nozzle-to-workpiece stand-off distance;
• Gas flow impingement point about 2mm ahead of laser beam-material interaction point for argon;
• Argon at a flow rate of 10-40 l/min.

Using this set-up as a basis, it may need to be further optimised for each particular application, material and set of welding parameters.

Acknowledgements

This work was funded partly by the members of TWI Group Sponsored Project 'Exploitation of High Power Nd:YAG Laser Processing' and partly by the Industrial Members of TWI through the Core Research Programme. The support of the Department of Trade and Industry of the United Kingdom for the former project is also gratefully acknowledged.

References

1. C. J. Dawes, Laser welding, p16, Abington Publishing, Abington, Cambridge, UK (1992).
2. S. Tsukamoto, K. Hiraoka, Y. Asai, H. Irie and M. Oguma, Proc. conf. Trends in welding research, 431-436 (1998).
3. A. Poueyo, G. Deshors, R. Fabbro, A.M. de Frutos and J.M. Orza, Proc. conf. LAMP 1992, 323-328 (1992).
4. R. E. Müller, H. Gu, N. Ferguson, M. Ogmen and W. Daley, Proc. conf. ICALEO 1998, C132-137 (1998).
5. G. K. Lewis, D.A. Cremers and R.D. Dixon, Proc. 4th Int. Colloquium on welding and melting by electron and laser beams, 211-218 (1988).
6. J. Greses, P.A. Hilton, C.Y Barlow and W.M. Steen, Proc. conf. ICALEO 2001, Paper 808 (2001).
7. J. T. Pauley and J.D. Russell, U.K. Patent 1,591,793.
8. Z. Sun, A.S. Salminen and T.J.I. Moisio, Proc. conf. Power Beam Technology 1990, 60-68 (1990).
9. I. M. Norris, TWI Members Report 333 (1987).
10. M. Ono, K. Nakada and S. Kosuge, Quarterly Journal of the Japan Welding Society, 10 (2), 9-14 (1992).
11. C. J. Dawes, Laser welding, p100, Abington Publishing, Abington, Cambridge, UK (1992).
12. C. A. Olivier, P.A. Hilton and J.D. Russell, Proc. conf. ICALEO 1999, D233-D241, 1999.
13. BS-EN-ISO 13919-1: 1997, Welding - Electron and Laser Beam Welded Joints - Guidance on Quality Levels for Imperfections Part 1: Steel, British Standards Institution, London, U.K. (1997).