Low-porosity laser welding of 12.7mm thickness aerospace aluminium

Paper presented at INALCO 2007, October 24-26, Tokyo, Japan

Abstract

This paper summarises the results of a study aimed at investigating the suitability of a high-power Yb-fibre laser for producing low-porosity welds in a 12.7mm thickness Al-Zn-Mg-Cu aluminium alloy. Full penetration was demonstrated in the 12.7mm material using 7kW of Yb-fibre laser power in both an autogenous set-up and hybrid combination with a MIG arc. Both process configurations were capable of producing a porosity level below that of the stringent weld quality class defined in BS EN 13919-2 and AWS D17.1, by considering shielding gas supply and material surface preparation, and selecting an appropriate laser spot diameter and welding speed. To consistently eliminate keyhole-induced porosity, welding was performed at a speed lower than the maximum capable of full penetration. This welding speed varied with alloy composition, welding process, ie autogenous or hybrid, and laser spot diameter. The occurrence of porosity was also reduced by removing the surface oxide and selecting a larger laser spot diameter (0.6mm instead of 0.4mm). By dry machining top and bottomsurfaces of the samples and the edges to be welded, less than one hour prior to welding, and by the use of a low-moisture shielding gas delivery, the total level of weld metal porosity was reduced to levels acceptable to the weld quality comparators used in this study (BS EN 13919-2 and AWS D17.1).

Introduction

Since the inception of the aerospace industry, aluminium is the preferred material for civil aircraft fuselage structures, with riveting the manufacturing process of choice. An analysis carried out at the turn of the century, however, indicated that a move to welded airframe structures could lead to considerable weight reductions and manufacturing cost savings in the region of 30% over the riveted counterparts. ^[1] In recent years, this has led to a considerable amount of research into alternative joining processes to replace the conventional riveting. One such process under investigation is laser welding, because it offers low heat input, narrow welds with little thermal distortion, and flexibility in manufacture. In combination with a lightweight material such as aluminium, the benefits of laser welding for aerospace, as well as other industrial applications, are significant.

The potential of laser welding has already been demonstrated, in production, by Airbus, who now has several laser-welded stringer-to-skin assemblies incorporated in some of their most recent aircraft models, including the A318 and the A380. Based on their experience in thin-section fuselage structures, Airbus (UK) established a strategy for the development of fusion welding methods for wing structures. This resulted in a UK Department of Trade and Industry (DTI)supported collaborative project, with Airbus UK, Alcoa Europe, Cranfield University, Manchester University, QinetiQ and TWI, as Consortium partners, with as primary aim 'to investigate novel technologies for the welding of thick-section aluminium wing structures'. The work presented in this paper was part of this project, investigating the capability of using fibre-delivered laser technology to produce fully penetrating welds in a 12.7mm thickness aerospace aluminium alloy to be used for wing manufacture, with a resulting laser weld quality, in particular related to weld metal porosity, suitable for aerospace service.

Experimental Approach

Prior to this work, aluminium up to only 6-8mm in thickness could be welded in a single pass using the laser output power available at the 1µm (Nd:YAG) wavelength. For thicker samples, multi-pass or double-sided welding was required to obtain full penetration. This situation changed in 2003 with the introduction of Yb-fibre lasers, which are capable of producing laser light at a wavelength similar to that of Nd:YAG lasers and easily scaleable to deliver output powers significantly higher than those currently available with Nd:YAG laser technology. ^[2] Alongside this development in laser source technology, an alternative laser welding technique, hybrid laser-arc welding, has, in recent years, attracted wide-spread industrial interest. In this process, an electric arc isintroduced into the same molten pool as the laser, resulting in the capability to weld thicker materials or faster welding speeds to be achieved, as well as offering improved joint gap tolerance and weld quality. ^[3-4] Because of their reported advantages for welding thick-section materials, both the Yb-fibre laser and the hybrid laser-MIG process were investigated in this study, using 7kW of Yb-fibre laser power.

As is the case for other fusion processes when welding aluminium, specific material issues had to be considered in this investigation, to ensure a good weld quality. The presence of the tenacious surface oxide, the sensitivity (ofthe alloys used) to hot cracking, the high thermal conductivity and the high hydrogen-solubility in liquid aluminium, all increase the risk of weld imperfections, such as cracks, lack of fusion and porosity. In addition to these imperfections associated with fusion welding, keyhole instabilities can occur when welding with a power beam process such as the laser, resulting in porosity. Keyhole-induced porosity is particularly associated with partially penetrating welds, their occurrence reducing sharply when fully penetrating welds are made. ^[5] For the purpose of this work, a distinction was made between cavities, which were irregular in shape and observed on the weld centreline near the keyhole root, and coarse pores, which were spherical in shape with a diameter larger than 0.5mm and found throughout the weld depth. The bulk of the investigation was aimed at eliminating cavities and reducing the number and size of coarse porosity, typically found in thick-section aluminium laser welds, with some effort also spent in reducing hydrogen-induced porosity, also referred to as fine porosity, towards the end of the work. For the purpose of this paper, fine porosity is defined as those pores that are spherical in shape, smaller than 0.5mm in diameter and primarily caused by the entrapment of hydrogen.

Initial welding trials were carried out to establish autogenous laser welding conditions capable of producing full penetration in a 12.7mm thickness 7000-series Al-Zn-Mg-Cu aluminium alloy, with visually acceptable top and under-bead weld profiles. Trials were carried out under different conditions of welding position, laser orientation (in relation to the workpiece surface) and top-bead shielding. Based on these findings, autogenous laser and hybridlaser-MIG welding trials were performed to establish the minimum and maximum welding speeds for full penetration, with subsequent trials assessing the influence of alloy composition, surface oxide, laser spot size, welding speed and welding process, ie autogenous laser welding versus hybrid laser-MIG welding, on the occurrence of keyhole-induced cavities and coarse porosity in the laser-welded 12.7mm thickness aluminium alloy. In the final trials of this investigation, process conditions established in earlier TWI work for minimising the occurrence of fine porosity, ^[6] were applied to those conditions established in this work that resulted in cavity-free welds with a low level of coarse porosity. The level of porosity in all autogenous laser welds and hybrid laser-MIG welds produced in this study was assessed using radiographic examination.

Experimental set-up

All welding trials were performed on a 12.7mm thickness 7000-series Al-Zn-Mg-Cu aluminium alloy of proprietary composition and temper, further referred to as the 7xxx aluminium alloy. This alloy was selected, by Airbus UK, because of its potential (future) use in the manufacture of wing structures.

For both the autogenous laser and the hybrid laser-MIG process, welding was carried using a 7kW Yb-fibre laser, YLR-7000, manufactured by IPG, at a laser output power of 7kW, as measured at the workpiece using an OPHIR 10K-W powermeter. This laser power was focused into a 0.4mm or a 0.6mm diameter spot size, with all welds produced with the beam focus positioned on the material surface. For both processes, shielding gas was supplied to the top of the weld pool through the MIG torch, see Figure 1. Industrial grade argon and helium conforming to BS EN 439 were both evaluated for shielding of the top and underside of the weld bead, with a dewpoint for both gases of -70°C, containing less than 3ppm of moisture, as quoted by the supplier. The hybrid welding trials were performed using an ESAB AristoMIG 450 synergic MIG power source, with a commercially available 1.2mm diameter Al-Mg-Mn-Cr wire positioned 2mm behind the laser beam impingement point on the material surface and at a 30° angle with the laser axis.

All welds were produced by traversing the laser processing head (and MIG torch) over the samples, which were mounted in a stationary jigging arrangement. A sandwich-type jig was used, which comprised steel top clamping bars and a heavy-section steel backing plate with a copper insert, in which a 10mm wide and 12mm deep slot was machined to deliver shielding gas to the underside of the weld bead. The samples were clamped as close to the weld line as possible, to avoid problems with sample alignment and weld distortion, without compromising the joint access. The set-up as described above is shown in Figure 1.

 

Initial process optimisation

 

To obtain full penetration in the 12.7mm thickness 7xxx aluminium alloy, it was necessary to weld out-of-position to avoid burn-through and excessive sagging which was encountered when welding in the flat (PA) position. Both the horizontal-vertical (PC) and vertical-up (PF) welding position resulted in acceptable visual appearance of both top and underside of the weld, as shown in Figure 2 for a weld produced in the PF welding position. The use of helium as top-bead shielding gas created an enlarged weld pool, compared with argon, which impeded the closure of the top of the keyhole, thereby stabilising the process. This improved stability was also concluded from the CCTV footage during welding and the visual appearances of both top and underside of the weld.

Full penetration was achieved at a welding speed of 0.65m/min and 0.5m/min, when welding in the PF and PC welding position, respectively, using a 0.6mm diameter spot size. As a result of the higher welding speed, the PF position welds were made with a lower heat input, ie with less thermal damage to the heat-treated 7xxx aluminium alloy, compared with the PC position welds. When welded at the same welding speed of 0.5m/min, the weld area of a PF position weld was smaller than that of a PCposition weld, see Figure 3, suggesting that the welding heat remained constrained in a smaller area when welding in the PF position. It is thought that the position of the helium shielding flow with respect to the weld pool (supplied through the trailing MIG torch at an angle of 30° with the laser axis) contributed to this difference in weld shape, with the (cold) shielding gas affecting the melt pool temperature and cooling rate of the weld. Moreover, when welding in the PF position, the thinning of the leading edge of the keyhole wall (under the influence of gravity) may also improve the heat flow into the cold material in front of the keyhole.

When welding in the PF position, the occurrence of cavities was lower compared with welds produced in the PC welding position, for the same conditions of output power, spot size and welding speed. The volume of the thin sheath of molten material surrounding the front of the keyhole is small compared with the pool of molten material behind the keyhole. When welding in the PC position, this mass of molten material is subject to gravity, which may result in the keyhole collapsing. However, this is not the case when welding in the PF position, where, in contrast, gravity may smooth the weld pool motion, thereby contributing to keyhole stability. Moreover, the stability of the keyhole entrance is heavily affected by the behaviour of this large volume of liquid metal, and vice versa. When the keyhole momentarily becomes unstable, caused by a sudden expulsion of metal vapour from the keyhole, for example, a surface wave will be created, which travels away from the keyhole and bounces off the solidified weld material back towards the keyhole. This reflected surface wave will force the keyhole to close, but is counteracted by the vapour pressure inside the keyhole. Whether this will lead to the keyhole entrance closing, depends on the forces acting against this, most notably the vapour pressure, and on the intensity of the surface wave, which in turns depends on the viscosity of the molten material (determined by the alloy composition) and on the magnitude of the instability that causes it. ^[7-8]

Welds produced with the laser angled opposite to the direction of travel, ie pointing down, as shown in Figure 1, resulted in less sagging, particularly of the top bead. This configuration is further referred to as the arc pushing configuration, derived from the direction of the MIG nozzle in relation to that of travel. When keyhole laser welding, the bulk of the molten material is behind the keyhole and at the top of the weld bead. With the laser perpendicular to the material surface when welding in the PF position, this molten material sags and potentially drops out of the joint. Rotating the laser (axis) against this phenomenon, ie pointing the laser downward, countered this sagging, whilst rotation in the opposite direction exacerbated the problem. In this work, a 10° travel angle for the laser proved satisfactory, although different angles were not investigated.

Reference welding conditions

Autogenous laser welding was carried out in subsequent trials in the PF welding position and arc pushing configuration, using 7kW of Yb-fibre laser power focused on the material surface and helium shielding supplied through the MIGtorch. For a laser spot size of 0.4mm, the minimum and maximum welding speeds for full penetration were 0.5 and 0.83m/min respectively, whereas these reduced for a 0.6mm spot size to 0.5 and 0.65m/min, respectively.

Hybrid laser-MIG welding was carried out in the PF welding position and arc pushing configuration, the MIG arc trailing 2mm behind the laser beam, at a travel angle of 30° with the laser axis, Figure 1, using 7kW of Yb-fibre laser power focused on the material surface, 6m/min wire feed speed and helium shielding through the MIG torch. For a laser spot size of 0.4mm and 0.6mm, the maximum welding speed for full penetration was 0.92m/min and 0.94m/min, respectively. For both spot sizes, a welding speed as low as 0.5m/min, as achieved for the autogenous laser welding trials, still resulted in full penetration with an acceptable weld-bead appearance. The weld bead appearance of all hybrid laser-MIG welds displayed a regular surface ripple, with the cross-section and weld-bead appearances of a typical weld produced using a 0.6mm spot size at the maximum welding speed for full penetration shown in Figure 4.

Cavities and coarse porosity

Removing the high melting-temperature, tenacious oxide from the top and bottom surface of the 7xxx aluminium alloy, as well as from the edges to be welded, by dry-machining (as used here) just prior to welding, reduced the number of cavities and coarse pores. However, removing the oxide alone was not sufficient to produce cavity-free laser welds. When the oxide layer is thick and not removed prior to welding, the evaporating alloying elements can form high meltingpoint metal oxides during laser welding. These change the molten material viscosity, which impacts on the keyhole stability, and may also form a tenacious skin on the weld pool surface, which impedes the escape of formed pores. ^[9]

Keyhole instabilities in aluminium laser welds are often associated with the presence of low boiling point constituents, such as Mg and Zn, and the (violent) evaporation thereof during laser welding. To test this theory, fully penetrating welds were produced in 12.7mm thickness 2024 aluminium, another alloy frequently used in the aerospace industry, using identical welding conditions as those used for the 7xxx alloy. However, despite lower levels of both Mgand Zn in the 2024 alloy, the laser welds contained a larger level of cavities and coarse porosity compared with those produced in the 7xxx aluminium alloy. The other, most commonly accepted theory for keyhole instability is related to the viscosity/fluidity of the weld pool forming behind the keyhole, which can cause the top of the keyhole to collapse. Silicon (Si), for instance, makes the weld pool more fluid, ^[10-11] and the higher Si-content in the 2024 alloy compared with the 7xxx alloy, ie 0.5% compared with 0.1%, may explain why a higher number of keyhole-related pores were observed in the 2024 melt runs.

The hybrid laser-MIG process was also capable of reducing, but not eliminating, the occurrence of cavities and the number and size of coarse pores, compared with autogenous laser, when welding 12.7mm thickness 7xxx aluminium, as shown in Figure 5. A less vigorous weld pool motion was evident from the small surface ripple observed on both the top and under-bead of the hybrid laser-MIG welds, Figure 4 (produced at the maximum speed for full penetration) compared with the autogenous laser welds, Figure 2 (produced at the minimum speed for full penetration). This was the result of the MIG arc enlarging the weld pool at the top of the keyhole, dampening the weld pool motion, and thereby reducing the occurrence of keyhole instabilities and producing a more uniform weld bead profile, compared with autogenous laser.

A laser spot size of 0.6mm produced less cavities than a 0.4mm spot size when laser welding the 12.7mm thickness 7xxx aluminium. Equilibrium between pressure from vaporisation and from surface tension of the molten material is required for a stable keyhole, with the latter working to close the keyhole formed by the evaporation of the material. ^[12] A smaller spot size produces a higher surface tension, because of the smaller radius of the keyhole walls, ^[12] resulting in a keyhole that is more prone to collapse. ^[7] Moreover, the smaller the keyhole, the easier it is for the front, back and sidewalls of the keyhole to touch, and collapse.

Of those investigated, the factor that had the largest effect on the occurrence of cavities and coarse porosity in laser-welded 12.7mm thickness 7xxx aluminium, was welding speed. The influence of welding speed on the occurrence of cavities is shown in Figure 6. Using a 0.6mm diameter spot size, full penetration could be achieved at a welding speed of 0.65m/min and 0.94m/min for autogenous laser and hybrid laser-MIG welding, respectively. However, welds produced at these maximum speeds (for full penetration) contained cavities. When slowing down (whilst maintaining full penetration), less cavities were observed, due to the molten pool behind the laser keyhole getting larger, thereby dampening the weldpool motion and stabilising the keyhole. At welding speeds equal to or lower than 0.55m/min and 0.75m/min for autogenous and hybrid laser-MIG welding, respectively, no cavities were observed. At these cavity-free welding speeds, which depended on the laser spot size used, the number and size of the coarse and fine pores was also confirmed to be lower, with more time available at these lower speeds for any formed pores to escape the solidifying weld pool.

Fine porosity

A handful of autogenous laser and hybrid laser-MIG welds were subsequently produced using the conditions established earlier, at their respective cavity-free welding speeds of 0.55m/min and 0.75m/min. The welds were produced on samples of which top, bottom and edge surfaces to be welded were dry-machined immediately prior to welding and using a low-moisture shielding gas delivery system. The latter comprised a short, 1.5m long, polyamide gas line, which was stored at an elevated temperature for 24 hrs, acclimatised to laboratory conditions for at least 2 hours and purged with shielding gas at least 2 minutes immediately prior to welding.

Radiographs were take of all welds and pore counts carried out. For both the autogenous and hybrid laser-MIG welds, all pores observed were sub-surface and randomly distributed along the weld length, with clustered or linear porosity not observed in any of the welds. Based on the pore counts, as graphically displayed in Figure 7 (with the displayed values averages of all weld produced), the values for pore area and pore length were calculated in accordance with BS EN 13919-2 and AWS D17.1, respectively, and compared with the limits defined for the stringent weld quality class in each of these standards. This confirmed that both the autogenous laser and hybrid laser-MIG process were capable of consistently producing a level of porosity in accordance with the stringent weld quality classes defined in BS EN 13919-2 and AWS D17.1.

Conclusions

This study has demonstrated that full penetration can be achieved in a 12.7mm thickness Al-Zn-Mg-Cu aerospace aluminium alloy using 7kW of Yb-fibre laser power in either an autogenous set-up or in a hybrid combination with a MIGarc. Both the autogenous laser and the hybrid laser-MIG process are capable of producing a level of weld metal porosity in accordance with the most stringent weld quality classes defined in BS EN 13919-2 and AWS D17.1, by considering shielding gas supply and material surface preparation prior to welding, and selecting an appropriate laser spot size and welding speed. The progress made in the project duration, in terms of weld quality, in particular related to weldmetal porosity, is shown in Figure 8.

To consistently eliminate the occurrence of keyhole-induced cavities, welding had to be carried out at a welding speed lower than the maximum capable of achieving full penetration in the 12.7mm thickness aluminium. This cavity-freewelding speed varied with alloy composition, welding process ( ie autogenous laser or hybrid laser-MIG) and laser spot diameter used. The occurrence of cavities and coarse porosity was also reduced by removing the tenacious oxide prior to welding and by selecting a larger laser spot diameter, in this case, 0.6mm instead of 0.4mm.

By dry machining the top and bottom surface of the samples immediately next to the joint line, as well as the edges to be welded, less than one hour prior to welding, and by the use of a low-moisture shielding gas delivery, a total level of weld metal porosity in accordance with the stringent weld quality classes defined in BS EN 13919-2 and AWS D17.1 was achieved.

The PF welding position was preferred over the PC welding position, because faster welding speeds could be achieved for full penetration in the 12.7mm thickness Al-Zn-Mg-Cu aerospace aluminium alloy. The laser welds produced in the PF welding position contained less keyhole-induced porosity compared with those produced in the PC welding position.

Acknowledgments

This work was funded jointly by the Industrial Members of TWI, as part of the Core Research Programme, and by the UK Department of Trade and Industry (DTI) Aeronautics Research Programme. The author would like to give special thanksto Chris Allen and Paul Hilton of TWI for their technical contributions, and to all the Consortium partners for their input in the technical discussions, their guidance on the progress of the work and provision of the materials used. The assistance of Anthony Elliott and Peter Brown, who carried out the processing trials, and the staff at the Test House, for carrying out the radiography and sectioning, is gratefully acknowledged.

References

1. Dunkerton S B: 'Welding for the 21st century'. Materials World, Vol.10, No.3, March 2002, pp.18-19.
2. Verhaeghe G: 'Fibre lasers - A technology review'. LaserAM report, November 2006.
3. Bagger C, Olsen F O: 'Review of laser hybrid welding'. Journal of laser applications. Vol.17, No.1, February 2005, 2-14.
4. Dilthey U, Wieschemann A: 'Prospects by combining and coupling laser beam and arc welding processes'. Welding in the World. Vol.44, No.3, May-June 2000, pp.37-46.
5. Seto N, Katayama S, Matsunawa A: 'Porosity formation mechanism and suppression procedure in laser welding of aluminium alloys'. Welding International, Vol.15, No.3, March 2001, pp.191-202.
6. Verhaeghe G, Hilton P A, Barnes S: 'Achieving low-porosity laser welds in aerospace aluminium alloy'. Proceedings of Aerospace Manufacturing Technology Conference (AMTC), Montreal, Canada, 8-12 Sept 2003.
7. Hu B, Richardson I M: 'Autogenous laser keyhole welding of aluminium alloys'. Journal of Laser Applications, Vol.17, No.2, May 2005, pp.70-80.
8. Matsunawa A, Kim J-D, Seto N, Mizutani M, Katayama S: 'Porosity formation and solidification cracking in laser welding'. Proceedings of the Taiwan International Welding Conference TIWC'98 on Technology Advancements and New Industrial Applications in Welding, Taipei, Taiwan, 7-9 September 1998, pp.47-57.
9. Takahashi K, Sato S: 'Porosity reduction in CO ₂ laser welding of aluminium alloys - Influence of penetration, joint, oxygen gas and oxide films'. Welding International, Vol.14, No.6, June 2000, pp.439-446.
10. King F: 'Aluminium and its alloys'. Publ: Ellis Horwood Ltd, Chichester, UK, 1987. ISBN 0-7458-0013-0.
11. Zhao H, White D R, DebRoy T: 'Current isues and problems in laser welding of automotive aluminium alloys'. International Materials Review, Vol.44, No.6, June 1999, pp.238-265.
12. Duley W W: 'Laser welding'. Publ: John Wiley & Y Sons Inc, New York, 1999. ISBN 0-471-24679-4.