Wired for sound results... pulsed MIG welding of titanium

TWI Bulletin, January/February 2006

A novel consumable has been developed which offers a promising future for the joining of titanium.

Tasos Kostrivas graduated with an MSc and PhD from two of the world's leading academic centres for welding metallurgy, Cranfield University and Ohio State University. Since joining TWI's CRA, Surfacing and Analysis section he has worked on various aspects of non-ferrous metallurgy, leading failure investigations and research programmes. His recent work on titanium includes high temperature creep-fatigue interactions, sustained load cracking and MIG welding.

Lee Smith manages TWI's CRA, Surfacing and Analysis section, covering all aspects of stainless steel and non-ferrous metal, metallurgy and corrosion, surfacing and material analysis. Lee has ten years experience at TWI, following graduation as a metallurgist from Birmingham University and post doctorate work on non-ferrous alloys. In addition to managing the section, Lee leads some of TWI's more high profile consulting projects.

Geoff Melton has over twenty years of experience in research, development and technical support in arc welding equipment and consumables. Eleven of those years were spent with a leading global manufacturer of welding equipment and consumables during which time Geoff was active in the British and European manufacturers' associations for welding products. He is also the Chairman of the British and European committees for arc welding equipment standardisation.

Alex Plewka recently completed the second year of his PhD studentship, sponsored by TWI, to research the mechanisms and incidence of porosity in titanium welds. He is studying at Birmingham University and his academic supervision is provided by Martin Strangwood and Claire Davis.

The production of high-quality, welded titanium components can be expensive due to the inherent incompatibility of many welding processes with the material. This has resulted in TIG and electron beam welding being the most applied joining processes. It might be assumed that MIG welding would be more common, but historical difficulties have marginalised this process so that it is only typically used for the lowest quality applications (appliqué armour, for example). Recently, Daido Steel has produced titanium wire that is claimed to create greater arc stability and reduced porosity and spatter. As Tasos Kostrivas et al report, this has been achieved by the adoption of a novel wire production process that modifies the wire surface, and the use of optimised pulse parameters.

As with any titanium fusion welding process, porosity can arise from entrapment of shielding gases, intrinsic to the process, or from gas bubble formation from absorbed surface contaminants 'extrinsic' to the process. Whilst a more stable arc will reduce spatter and mitigate intrinsic porosity, the cleanliness and preparation method of the mating surfaces and the time lapsed between surface cleaning and welding will have the greatest influence on the extrinsic weld metal porosity of titanium welds.

In order to assess the quality, specifically weld spatter and weld metal porosity, of MIG welded titanium, the present programme considers conventional and novel wire for fabricating commercial purity titanium. Various joint preparation methods were used in an attempt to differentiate between the intrinsic and extrinsic weld metal porosity and so make a more informed assessment of the 'quality' of the welds. In addition, the fume and ozone generation for the process was measured, to determine compliance with modern health and safety legislation. This is particularly important since prior to this work there was no published information on the fume or ozone generation for titanium welding processes.

Objectives

• To determine if a novel titanium wire addresses the quality concerns previously associated with MIG welding of titanium, namely weld spatter and porosity, by comparing welding with the novel wire directly with a conventional wire, using different Ar/He shielding gas mixtures.
  
  
• To determine the influence of the joint face preparation method on porosity and provide greater confidence in distinguishing between intrinsic and extrinsic porosity.
  
  
• To produce data for fume and ozone generation for MIG welding of titanium using the novel and conventional titanium wires.

Experimental approach

Bead-on-plate deposits were made on 12mm thick ASTM B265 Grade 2 titanium plates using both a novel wire, yet matching the composition of ASTM Grade 2, and a conventional AWS ERTi-2 wire with an OTC Daihen Inverter 350 pulsed MIG welding power source. Chemical analysis was carried out to establish the composition of the parent material and novel wire and the results are given in Table 1. Deposits were made using a range of welding parameters and different Ar/He shielding gas mixtures, ranging from pure Ar to pure He. Argon is the most widely used inert gas while helium provides a hotter arc and potentially improved weld metal transfer. Argon was used as purging gas for the full penetration butt welds.

Table 1 Chemical analysis data of parent material, weld metal and welding wires (TWI analysis references S/05/87 and S/05/129).

Fig.1. The joint face preparation

Optimised welding conditions, supplied by the consumable manufacturer, with argon shielding gas were used to produce full penetration butt welds in 6mm thick ASTM B265 Grade 2 titanium plates ( Fig.1) with the novel wire. The joints were prepared using different preparation methods and time delays between surface preparation and welding. The surface preparation methods included milling with and without lubrication, surface cleaning using ScotchbriteTM and a lint-free cloth, and acid pickling. Details of the plate preparation techniques are shown in Table 2.

Table 2 Details of surface preparation techniques

The completed welds were examined using X-ray radiography and the film was analysed using macro-image analysis software. Samples prepared with different surface preparation techniques were examined using surface profilometry, scanning electron microscopy (SEM) and also energy dispersive X-ray (EDX) spectroscopy to link the observed surface features with the weld metal porosity. Particulate fume and ozone sampling was carried out for the novel and conventional wires according to BS EN ISO 15011 Part 1 and relevant TWI procedures, respectively.

Results and discussion

The initial bead-on-plate welding parameters and gas optimisation trials demonstrated that small changes to parameters such as voltage, current and peak-current-time had marked effects on arc stability, spatter generation and bead profile. The optimised welding conditions for both wires and shielding gases are shown in Table 3. The novel wire was found to perform best with pure argon gas shielding, whilst conventional wire performed best with pure He shielding.

Table 3 Optimised welding parameters for the conventional and novel welding wires using different shielding gases. Peak current was 400A and background current was 60A.

The novel wire demonstrated a smooth wire feed and produced welds which were straight and consistent with low spatter compared to the conventional wire. Examination under an optical microscope showed the novel wire to be smoother and more consistent in diameter than the conventional wire ( Fig.2-4). These characteristics are expected to reduce wear of the contact tip and thus, increase consistency of current pick-up which promotes arc stability over conventional wire. Energy-dispersive X-ray (EDX) analysis showed presence of Ca on the surface of the novel wire ( Fig.5). This may act to minimise the wandering of the arc root and enhance arc stability. As a result, the novel wire performed much better than conventional wire with respect to arc stability and minimised spatter.

Optical micrographs showing the microstructure of the welding wires. Note the difference in grain size and surface roughness
a) Novel wire
b) Conventional wire

Fig.3. The surface of conventional wire, secondary electron image

Fig.4. The surface of the novel wire, secondary electron image

Fig.5. EDX spectrum from the surface of the novel wire showing calcium peak

Variation in the diameter of the conventional wire is anticipated to increase mechanical damage at the contact and thus, produce variation in the wire feed rate and reduced arc stability. Changes in wire feed rate, such as those observed for the conventional wire, are expected to alter the arc length. Despite having a voltage control system in the power source which compensates for changes of the arc length by altering the pulsing parameters, frequent adjustments of pulsing parameters also compromised arc stability.

The values of surface roughness for the surface preparation techniques applied to the titanium plates were found to be between 0.6µm and 1.6µm. There was little variation in surface roughness between the different joint face surface preparations although SEM and EDX analysis demonstrated consistent differences in both surface topography and cleanliness.

Examination of the finished welds showed effective gas shielding with no signs of surface oxidation ( Fig.6). Radiography showed both a degree of porosity ( Fig.7) and pore sizes comparable with those reported in relevant literature, although overall the pore density was low. The as-wet machined and pickled sample with a one-day delay prior to welding gave the lowest overall porosity level. Table 4 shows the results of the porosity investigation. The butt welds were characterised class A with regard to both maximum pore size (1.5mm) and maximum accumulated length (6mm) according to AWS D17.1: 2001. They also met the acceptance limits for the same criteria in SAE AMS 2689A - Aerospace Material Specification ( Table 4). However, the porosity levels were still greater than can be achieved using TIG or EB welding, although further reductions may be possible given additional optimisation.

Fig.6. Photograph showing the cap of a full penetration butt weld made with the novel titanium wire and argon shielding gas

Fig.7. Photomicrograph showing the microstructure of a pulsed MIG weld produced using G-Coat (ASTM Grade 2) wire and ASTM Grade w parent material. Pore is highlighted circle

(micrograph supplied by Birmingham University)

Table 4 Summary of porosity measurements for MIG butt welds produced with different surface preparation methods

The particulate fume and ozone emission rates for the novel wire, when using Ar shielding gas, were 0.32mg/s and 1.02ml/min, respectively. The corresponding rates for the conventional titanium wire, using helium shielding gas, were 4.04mg/s and 2.45ml/min and the emission rate for ozone using argon shielding gas was 3.01ml/min. Table 5 summarises the measurement results for particulate fume and ozone emissions for the novel and conventional titanium wire. These are compared with the ozone emission rates from MIG and TIG welding of other materials such as steel and aluminium.

Table 5 Summary of particulate fume and ozone emission rate measurements

It can be seen that pulsed MIG welding with the novel wire and argon shielding gas provided the lowest ozone and particulate fume emission rates.

Conclusions

The novel titanium wire with Ar shielding gas gave a stable arc with lower weld metal spatter and intrinsic weld metal porosity than that achievable using conventional titanium wire with Ar, Ar/He or He shielding and successfully addressed the quality concerns previously associated with MIG welding of titanium.

Low extrinsic weld metal porosity can be achieved when the surfaces of the mating plates, in the vicinity of the weld area, are acid pickled followed by acetone rinsing and welding within one day. However, porosity levels of the welds generated in the present programme were greater than can be achieved with either TIG or EB welding, although further refinement may be possible.

Very low particulate fume and ozone emissions were obtained with the novel wire and argon shielding (lower than those in TIG welding of steel and aluminium). Ozone was also low for welds made using conventional wire and He shielding gas, but fume generation was sufficiently high to warrant the use of local extraction.

Recommendations

This work demonstrated that pulsed MIG welding can be used for high productivity manufacture of titanium structures. Stable arc and smooth weld metal transfer characteristics with reduced spatter were achieved using a novel titanium wire and Ar shielding gas. In addition, particulate fume and ozone production was very low for MIG welding with the novel wire when it was compared with the emission rates for conventional titanium wire (with He shielding) or steel. Low weld metal porosity was achieved when the surfaces of the mating plates, in the vicinity of the weld area, were acid pickled followed by acetone rinsing and immediate welding. This cleaning method is compatible with industrial experience and it can be used as a standard cleaning practice of titanium prior to welding. However, further refinement of the welding parameters should be attempted to reduce intrinsic porosity levels.