Laser and Hybrid Laser-MIG Welding of Aluminium Alloy

C M Allen ^1, , G Verhaeghe ^1, , P A Hilton ^1, , C P Heason ^2, , P B Prangnell ^3,

¹ TWI Ltd., Granta Park, Gt. Abington, Cambridge CB1 6AL, United Kingdom.
² Now at Corus Research, Development & Technology, Swinden Technology Centre, Moorgate, Rotherham, South Yorkshire S60 3AR, United Kingdom. Formerly of ^3
3 Materials Science Centre, Manchester University, Grosvenor St., Manchester M1 7HS, United Kingdom.

Paper presented at International Conference on Aluminium Alloys (ICAA 10), Vancouver, 9-13 July 2006.

Abstract

Fusion welding of 7xxx aluminium alloy plates has been investigated for aerospace applications using autogenous laser welding and hybrid laser-MIG welding. Nd:YAG and Yb-fibre lasers have been used, with two different focussed spot sizes in each case. Autogenous and hybrid welding of 12.7mm thick plate using the Yb-fibre laser with a 0.6mm diameter spot was selected for further development, on the basis of penetration and weld quality achieved. These welds were acceptable to the highest quality class B (stringent) of BS EN ISO 13919-2:2001, with a porosity of only 0.3% of the cross-sectional area of the weld, and close to class A of AWS D17.1. Transverse proof strengths of ~60% of parent material were achieved. Development of hybrid welding is ongoing with novel fillers to refine weld metal grain structure and improve weld properties.

Introduction

Traditionally, riveting has dominated the joining of thin sheet aluminium structures in aircraft. However, to maintain competitiveness in aircraft production novel manufacturing routes and joining processes are being developed. Laser welding is one such process, already demonstrated by Airbus to deliver manufacturing and weight saving advantages in fuselage construction using thin (2-3mm) 6xxx aluminium alloy sheet. ^[1]

Challenges remain for thicker (>6mm) structures made from high strength, limited weld ability 7xxx and 2xxx alloys, which cannot easily be substituted by more weldable variants. Welding procedure development is necessary to overcome strength loss, and improve weld quality, of 7xxx fusion welds. This paper focuses on the development of laser welding processes to reduce heat input, increase weld quality; by reduction of porosity and elimination of cracking, and improve weld properties by novel filler additions.

Experimental

A 7xxx aluminium alloy in proprietary temper has been chosen. Single pass single sided full penetration butt welding of 12.7mm plate has been performed with a fibre optic delivered IPG Yb-fibre laser capable of high power output. ^[2] Focussed spot diameters used were either 0.4mm or 0.6mm. As a start point for potential future development of a simultaneous double sided welding procedure using the Yb-fibre laser beam split in to two equal lower power beams, single sided butt welding of 6.35mm plate has also been performed using a Trumpf Nd:YAG laser operated at half the output power of the fibre laser, with either a 0.45mm or 0.6mm diameter spot. Both autogenous laser welding and hybrid laser-MIG welding processes have been investigated, the latter using a 5556A (Al-5Mg) filler wire. For hybrid Yb-fibre laser-MIG welding, an ESAB AristoMIG 450 power source was used, and for hybrid Nd:YAG laser-MIG welding, a FroniusTPS450 power source. When hybrid welding, standard pulsed synergic programs on both arc power sources were used, appropriate to the wire consumable type and diameter used. In both hybrid welding processes, the laser led the MIG arc. Weld top bead shielding was provided by a flow of 25l/min of He through the MIG torch, with an underbead shielding of 5l/min of Ar.

Initial trials were performed to select a reduced number of processes from the large number of possible options, in terms of choice of laser source and spot size, and choice between autogenous laser welding and hybrid laser-MIGwelding. In all cases processes were considered with respect to complexity/ease of operation, the depth of consistent penetration achievable, and typical weld quality achieved. Weld cap and root profiles were less critical, as in application these are commonly machined off to avoid potential stress concentrations. Following selection, further development included minimisation of welding heat input, and addressed improvement of weld quality through factors such as material and filler preparation and cleanliness, and shielding gas quality and delivery. Parallel work targeted future filler development for hybrid welding, addressing the effect of novel filler compositions in laser welds using pre-placed inserts, as these were unavailable at the time in wire form. Inserts between 0.5 and 2mm thick were cast and rolled by the University of Manchester.

Results and discussion

Process Selection. The results of the initial selection trials are summarised in Table 1.

Table 1. Results of initial selection trials

Laser source and thickness penetrated	Spot diameter [mm]	Laser welding process
		Autogenous	Hybrid
Nd:YAG on 6.35mm thickness	0.45	Poor weld quality	Suitable for further development
	0.6	Inconsistent penetration	Inconsistent penetration
Yb-fibre on 12.7mm thickness	0.4	Poor weld quality	Suitable for further development
	0.6	Suitable for further development	Suitable for further development

Autogenous Nd:YAG laser welding. Full penetration in the 6.35mm thickness plate was achieved using either a laser spot diameter of 0.45 or 0.6mm. Due to a two-fold reduction in power density between the 0.45mm and0.6mm spot sizes, a three-fold reduction in welding speed was required to obtain full penetration in the latter case. Moreover, that penetration was not consistent over the entire length of the 300mm long test weld. Therefore,autogenous welding with the 0.6mm spot was not pursued further. With the 0.45mm spot, consistent penetration was achieved, but with poor internal weld quality. Welds contained a number of irregularly shaped cavities, in addition to spherical gas porosity. Cavities, as seen in Fig.1, are associated with momentary keyhole collapse. Collapse occurs when a fluctuation in the keyhole leads to a transient imbalance between the metal vapour pressure holding the keyhole open and the metallostatic pressure of the surrounding weld pool trying to close the keyhole. ^[3] It is thought by the authors that for any given fluctuation in the keyhole, the narrow diameter keyhole in the case of the smaller spot size could be more likely to close than in the case of the larger diameter keyhole with the larger spot size. The slower welding speed and longer solidification time in the latter case may also allow more time for any associated porosity to escape.

Fig.1. Radiograph of autogenous Nd:YAG butt weld in 6.35mm thick 7xxx plate, made using a 0.45mm diameter focussed spot, showing presence of coarse cavities (arrowed) associated with keyhole collapse duringwelding

Hybrid Nd:YAG-MIG welding. Lack of consistent penetration and slow speed again precluded a 0.6mm laser spot size as a viable process option. The 0.45mm spot size in the hybrid configuration did, however, achieve an internal weld quality that was free of the cavities observed in the corresponding autogenous welds, as shown in Fig.2. This could be due to the larger weld pool size in the hybrid technique, and hence slower solidification time and easier passage of porosity out of the weld pool prior to freezing. Some finer porosity was still present in these welds, however.

Fig.2. Top bead photograph (left), under bead photograph (centre) and radiograph (right) of hybrid Nd:YAG laser-MIG butt weld in 6.35mm thick 7xxx plate made using a 0.45mm diameter spot

Autogenous Yb-fibre laser welding. Full penetration in the 12.7mm thickness plate was achieved using either a laser spot diameter of 0.4 or 0.6mm. Welding speeds were ~25% faster in the case of the smaller spotdiameter than in the larger. Welds made with the smaller spot size contained cavities, associated with keyhole collapse, as was the case for Nd:YAG laser welds made with the smaller spot size in 6.35mm plate.

Hybrid Yb-fibre laser-MIG welding. Both laser spot sizes produced weld qualities that were free from cavities. One such example is shown in Fig.3. Similar welding speeds were achieved for both spot sizes. The fact that a good weld quality without cavities was achieved with a 0.4mm laser spot size in a hybrid configuration, but not achieved in the autogenous configuration, as was also seen in the Nd:YAG laser welds in 6.35mm plate, could again indicate that this is due to an increase in weld pool size. The presence of fine porosity in the hybrid welds was not considered further at this stage in the initial trials.

Based on the ability of a 0.6mm Yb-fibre laser spot size to achieve full penetration in 12.7mm plate, consistently and in a single pass, in the autogenous and the hybrid configuration, both these process variations were selected for further process optimisation.

Fig.3. Cross-section (top) and radiograph (below) of a hybrid Yb-fibre laser-MIG weld in 12.7mm thick 7xxx plate made using a 0.6mm diameter spot. Dashed lines on the cross-section indicate a typical final machined thickness used in application, confirming that weld cap and root profiles would be acceptable

Process Optimisation. The welding speed ranges over which the selected welding processes produced welds without cavities were established. In terms of minimising heat input and optimising mechanical properties, the faster speeds were then selected for both processes. In the hybrid process, decreasing wire feed speed, which reduced arc voltage and current, further reduced the heat input, although this was only of the order of 7% or less, and not anticipated to be significant. Weld quality was further improved in terms of fine porosity by concentrating on the effects of cleaning parent material, paying special attention to the use of low moisture shielding gas and its delivery system, and (for hybrid welding) filler wire prior to welding. These modifications resulted in weld qualities within the most stringent class, class B, of BS EN ISO 13919-2:2001, with a porosity of only 0.3% of the cross-sectional area of the weld, and close to the most stringent class, class A, of AWS D17.1, deemed to be a more rigorous standard related to quality of fusion welds in aluminium. At time of writing, further work is ongoing to consistently produce weldsto class A of AWS D17.1.

Weld Characterisation. Hybrid welds can possess an advantage over autogenous welds as the associated filler addition can be used to improve the mechanical properties of the weld metal.

In order to ensure that this advantage can be achieved through thickness, the uniformity of the mixing of the filler addition in a hybrid weld was investigated. Fig.4 shows an electron probe microanalysis (EPMA) map, showing the distribution of Zn in longitudinal section in the weld centre plane in a hybrid weld. Both the parent alloy and filler contain Mg, but only the parent contains Zn. Therefore, mapping of Zn can be used to show, by inference, the distribution of the filler. Fig.4 shows that the distribution of Zn, and hence Mg, is not uniform, with more Mg being near the top of the weld.

Fig.4. EPMA map of Zn distribution along welding direction in a hybrid Yb-fibre laser-MIG weld

Before considering the implications of non-uniform weld metal composition through thickness, Fig.5 shows the variation with heat input of minimum hardness values after proprietary post weld heat treatment (PWHT) for both Yb-fibre laser and hybrid Yb-fibre laser-MIG welds, from mid-thickness positions, in both the weld metal and the HAZ. The softest material was weld metal in both cases. Heat inputs were generally higher for the hybrid welds, from the additional heat input introduced by the MIG arc, in spite of faster welding speeds used. Fig.6 shows hardness profiles for both types of welds, using the optimum heat input consequently determined, measured transverse to welding direction. As Fig.6 shows, the weld metal hardness values towards the weld top and in the mid-thickness of the hybrid weld were very similar, indicating that the non-uniform weld metal composition, as shown previously in Fig.4, was not likely to significantly affect mechanical properties measured at different thickness positions. The ratio of the minimum hardness in the weld metal to the parent material hardness as a percentage was between 65and 70%.

Fig.5. Variation of minimum hardness values with available heat energy for Yb-fibre laser and hybrid Yb-fibre laser-MIG welds, after PWHT, from mid-thickness positions, in both the weld metal and the HAZ

 

Longitudinal and transverse tensile tests were carried out by QinetiQ on autogenous and hybrid welds. Tensile test blanks, 2mm thick, were cut transverse to the welding direction, and parallel to welding direction centred on the weld centreline and on the position of the minimum recorded hardness in the HAZ. Table 2 summarises the results after PWHT, compared to specified minimum properties for the parent material. All welds were made parallel to plate rolling direction (L).

Fig.6. Transverse hardness profiles for Yb-fibre laser and hybrid Yb-fibre laser-MIG welds, after PWHT, showing mid-thickness profiles for both weld types, and also a profile towards the top of the weld for the hybridweld

 

The following conclusions can be drawn from Table 2:

• The transverse weld proof strengths of both weld types were similar, with the transverse tensile strength and elongation of the hybrid weld being slightly lower.

• The ratio of the transverse weld proof strength to the specified parent material proof strength as a percentage was of the order of ~60%.
• The properties of the two weld metals (autogenous and hybrid) were significantly reduced with respect to parent properties, in line with the weld metal solidification microstructure, which included severe microsegregation and interdendritic eutectic films.
• The tensile strength and elongation of the hybrid weld metal were slightly lower than for the autogenous weld metal.
• The properties of the HAZ of the two weld types were very similar, which contrasts to the measurement of a softer HAZ in the hardness profile from the hybrid weld. In both welds the HAZ had higher strengths and larger elongations than the corresponding weld metal.

Table 2. Tensile properties of Yb-fibre laser and hybrid Yb-fibre laser-MIG welds. Number in brackets indicates number of tensile tests from which mean values were calculated. All welds were made parallel to plate rollingdirection.

These results, in terms of static mechanical properties, suggest that is no advantage in hybrid welding over autogenous welding, at least using the conventional filler wire composition selected in this work. However, there is meritin pursuing non-commercially available filler compositions, to determine if further improvements can be achieved in weld properties, in particular weld metal hardness, proof strength and elongation. As an example of an initial resultof work in progress, Fig.7

Material condition	Test orientation	Sample position	Mean 0.2% proof stress, MPa	Mean tensile strength, MPa	Mean elongation, %
Minimum specified properties for parent plate	Parallel to rolling (L)	n/a	580	605	7
	Transverse to rolling (LT)	n/a	565	595	7
Autogenous weld	Parallel to welding (L)	Weld metal	311 (1)	380 (1)	3 (1)
	Parallel to welding (L)	HAZ minimum	462 (3)	543 (3)	12 (3)
	Transverse to welding (LT)	Across weld	349 (3)	408 (3)	2 (3)
Hybrid weld	Parallel to welding (L)	Weld metal	314 (2)	362 (2)	1 (2)
	Parallel to welding (L)	HAZ minimum	468 (3)	545 (3)	12 (3)
	Transverse to welding (LT)	Across weld	348 (3)	388 (3)	1 (3)

shows the grain structure achieved in the weld metal with one such filler composition, containing a grain refiner addition, compared to a filler of conventional composition. In this work both were produced by laser welding over a pre-placed insert, the novel filler composition being unavailable at the time as a drawn wire. Ongoing work with higher levels of additions in novel filler compositions, is anticipated to lead to associated improvements in weld metal strength and elongation, as well as the grain refinement presented here.

Fig.7. Refined weld metal grain structure achieved in a Yb-fibre laser weld using a pre-placed insert of novel filler composition (top) compared to that achieved using the conventional filler (below)

Conclusions

The main conclusions of this work are:

• Single pass, single sided full penetration butt welds in 12.7mm thick 7xxx aluminium alloy plate can be produced using a high power Yb-fibre laser, with either an autogenous or hybrid laser-MIG welding process.
• Both of these types of weld were acceptable to class B of BS EN ISO 13919-2:2001, with a porosity of only 0.3% of the cross-sectional area of the weld, and close to class A of AWS D17.1.
• The ratio of the transverse weld proof strength to the specified parent material proof strength as a percentage was of the order of ~60%.
• The weakest and least ductile part of the weld in both cases was in the weld metal, due to the solidification microstructure, which included severe micro segregation and interdendritic eutectic films.
• Use of novel filler additions can achieve grain refinement, and ongoing work is anticipated to lead to a corresponding improvement in mechanical properties.

Acknowledgements

The authors would like to thank QinetiQ, one of the partners in the DEFUSE project in which this work has been carried out, for tensile testing of welded samples, and the U.K. government DTI Aeronautics Research Programme, forfinancial support.

References

1. Palm F: Plenary session given at Workshop on Laser Applications in Europe, Fraunhofer IWS, Dresden, Germany, 23-24 November 2005.
2. Verhaeghe G: Welding Journal, Vol.84, No.8, Aug 05, pp. 56-60.
3. Matsunawa A, Mizutani M, Katayama S and Seto N: Welding International, Vol. 17, No. 6, pp. 431-437.