Increasing the Strength of GMA Welds in 6000 Series Aluminum Alloys

By W.H.S. Lawson and H.W. Kerr

Background

Among aluminum alloys used in structural fabrication, alloy 6061-T6 is the most popular of precipitation-strengthened alloys that derive their high strength from heat treatment. When alloy 6061 is welded, the welding thermal cycle dissolves the precipitates in the part of the heat affected zone closest to the weld metal, giving a softer and weaker region than the unaffected base metal. Filler metals are different in composition from the base metal, which further complicates the weldment properties. The tensile strength of a sound, defect-free weldment in alloy 6061 in the as-welded condition can be as little as 50% of the specification strength of un-welded 6061-T6 material. Allowable working stresses in relevant design codes such as the ASME codes reflect this low strength expectation. The end result is a significant limitation on the economic and technical competitiveness of weld-fabricated products using 6061 and similar alloys.

As part of the undergraduate welding specialization program of the University of Waterloo’s Mechanical Engineering Department, students undertake a final year design project on a welding-related subject. Because of a perceived lack of information, the CWB Group’s Gooderham Centre proposed that a study be undertaken to optimize the strength of 6061 alloy weldments, including the effects of post weld heat treaments. To date, three groups of students have investigated several aspects of the strength-optimization problem. This article will summarize what has been learned so far about various aspects of the properties of such welds, using gas metal arc welding.

All test welds have been made to a common basic specification as listed in Table I. Welding and testing conditions developed by the students have explored several variables, principally:

• base-metal cleaning, edge preparation and fitup;
• welding-process parameters (voltage, current, travel speed, electrical stickout);
• filler-metal selection/weld-bead composition;
• post-weld heat treatment.

Table I: Common Material and Welding Conditions
Base Metal:	AA 6061-T6, 6.3 mm (1/4") thick plate
Process:	gas metal arc welding, single wire 1.2 mm (.045"), argon shield spray transfer
Joint Design:	full penetration butt weld on non-consumable backing, welding from one side in one or two passes; vee-groove preparations with various groove angles, lands and root gaps

To help clarify the effects of welding and heat-treatment variables, tensile testing has been supplemented by microhardness testing . Since hardness is nearly proportional to local tensile strength in these alloys, hardness profiles can show local effects of welding and heat-treatment variables and explain variations of strength and failure location.

Geography of a Weld Zone

The varying thermal history around a weld causes the microstructure and properties to vary greatly from the weld-bead centre to the outer edge of the heat-affected zone (haz). However, for simplicity in this discussion, we have separated the weld zone into only three different regions:

1. an outer part of the haz on each side of the weld, in which the peak temperature achieved is well below that required for melting or for dissolving the precipitates. In this area, the strengthening precipitates experience coarsening ("overaging"), which reduces their effectiveness.
    
2. the inner part of the haz where the peak temperature is high enough to dissolve precipitates, but the material composition is still that of 6061 alloy. In the as-welded condition, the hardness and strength depend on any re-precipitation that may occur during cooling or over longer times after welding.
    
3. the weld-fusion zone, which is fully melted by the welding process and has a composition different from that of the base metal, i.e, it is a mixture of melted base metal and added filler metal of a different composition.

Depending on the weld-bead composition and the thermal history (weld thermal cycle(s) and post-weld heat treatment if any), any one of these three zones can be the "weakest link" that limits the weld strength.

The composition of alloy 6061 is extremely susceptible to solidification cracking. A filler metal of different composition is used to shift the final weld-bead composition to a less crack-sensitive range. One of the questions is what effect this composition change has on the strength of the weld bead, especially after post-weld heat treatment. The filler metals traditionally used in welding alloy 6061 are 5356 (5% Mg) or 4043 (5.5% Si). The nominal compositions of alloy 6061 and of the welding filler metals examined in this work are summarized in Table II.

Table II: Nominal Compositions of Base Metals and Filler Metals (wt. %)
Alloy	Si	Mg	Cu
6061	0.4-0.8	0.8-1.2	0.15-0.40
5356	**	4.5-5.5	0.10*
4043	4.5-6.0	0.05*	0.30*
4643	3.6-4.6	0.1-0.3	0.10*
4047	11.0-13.0	0.10*	0.30*
* Maximum, ** Si + Fe <0.50

The 6061 plate used in the projects was 6.3 mm thick. Butt GMA welds in this thickness normally employ two weld passes. The student projects have attempted a single-pass procedure, for maximum productivity. To get reliable penetration, a high welding current is required, which results in significant melting of base metal. Most of the single-pass welds have shown dilution by the base metal of 50 to 70%; that is, the weld-bead composition comprised 30% to 50% filler metal, balance base metal.

As-Welded Properties

Figure 1 shows the results of microhardness testing of as-welded samples for single-pass procedures for three of the filler metals. As discussed above, in the as-welded condition the dissolution of the precipitates close to the fusion zone results in a soft weak zone. At a greater distance from the fusion zone, where precipitates are coarsened, the hardness and strength gradually increase towards the level of the unaffected base metal. Since the hardness of a properly hardened 6061-T6 plate should be about 100 Vickers, the welding operation has evidently reduced the material tensile strength by 40% to 45%. Different hardness values in the fusion zone were obtained for different filler metals, but the hardness values indicate that generally the strength was lowest in the haz (as long as fusion defects were not present).

Figure 1: Vickers hardness traverses (300g test load) from weld centreline to outer haz of GMA welds in alloy 6061 in the as-welded condition using three different filler metals. Coupons were refrigerated after welding to minimize natural aging prior to test.

Full-Heat-Treatment Properties

If a weldment fabricated in alloy 6061 is completely re-heat-treated (that is, including solution treatment, quenching and aging), the haz should be completely eliminated and the final weldment tensile strength could be up to 100% of the specification strength of 6061-T6, if the weld fusion zone can respond fully to precipitation strengthening. In Figure 2, hardness traverses are shown for welds similar to those in Figure 1 but given a full three-step T6 heat treatment after welding. It can be seen that the 4043 and 4643 weldments reached essentially 100% of baseplate specification strength , but that in the 5356 weldment the weld metal showed very little response to heat treatment.

Figure 2: Vickers hardness traverses from weld centreline to outer haz of GMA welds subjected after welding to full T6 precipitation heat treatment (solution treatment, quench and artifical aging).

The solution heat-treatment step for alloy 6061 requires the material to be soaked at 545° C (which is less than 30° C below the melting temperature) and then rapidly water quenched. This often is not a feasible heat treatment for larger welded fabrications, and also risks part warpage and/or partial melting.

Post Weld Aging Heat Treatment

A third alternative is to process as-welded fabrications using only the aging portion of the T6 heat-treatment sequence. This requires a heat treatment of 190° C for 4 hours, which is much more feasible for large weldments. In Figure 3, the results of post-weld aging are shown for weldments similar to those of Figure 1. As expected, a considerable hardness gain was obtained in the inner haz and in the weld metal of all but the 5356 sample. The outer part of the haz showed very little strengthening, as expected since this zone was already in an over-aged condition.

As implied by Figure 3, tensile testing of defect-free weld samples made with all three 4000 series filler metals and post-weld heat treated (aged only), showed tensile strengths at or above 70% of the specification strength of 6061-T6. Failure locations were in the haz about 12 mm (0.5") outboard of the weld-fusion boundary, which is where hardness testing had showed the softest part of the weld region, caused by the original overaging. Typical broken tensile test specimens shown in Figure 4 illustrate the shift of failure location from outer haz to weld fusion zone, depending on which is the softest part of the weldment.

Figure 3. Vickers hardness traverses across GMA welds given
an aging heat treatment of 190° C for 4 hrs after welding.

Figure 4: Typical tensile test specimens from the student projects,
showing failure in the outer haz as typical of aged welds,
and weld fusion zone failure as typical of some as-welded conditions
and of the welds made with 5356 filler after aging.

The Role of Weld-Metal Composition

Previous research has demonstrated that a net weld-metal analysis of greater than 2% Mg or Si is required to reliably suppress solidification cracking. All filler metals employed in these projects can achieve this result with the dilutions used in this work.

We have found that the strength of the weld-fusion zone depends strongly on the ability of the resultant composition to respond to post-weld precipitation strengthening. The precipitation reaction in turn results from coordinated action of the Mg and Si alloy additions. This work has clarified that some important differences in weld-metal precipitation depend on filler-metal composition. When filler metal 5356 was used, the weld bead responded much less to post-weld strengthening heat treatments (Figures 2 and 3), leaving the weld metal relatively soft. (and frequently comprising the weakest part of the weldment). Conversely, when Si was present in excess concentration (i.e., when a 4000 series filler alloy was used), the efficiency of the precipitation-strengthening process depended mainly on the amount of Mg present, and could closely match the base metal’s response if sufficient Mg (~0.3%) were present.

As shown in Table II, three different 4000 series filler-metal compositions are normally commercially available. Based on the alloying element effects discussed above, we have been able to confirm manufacturers’ recommendations as to what application regime is best for each, as follows:

• alloy 4043, with intermediate Si content is a suitable filler-metal choice for high-current GMA welds with 50% to 70% dilution, yielding sufficient Si to suppress solidification defects, while the necessary Mg is supplied by the melted base metal.
   
• alloy 4643 is designed for use in multi-pass, multi-layer arc welds in which base-metal dilution is much lower and weld beads consist mainly of added filler metal. Mg is added to the filler metal to achieve adequate net Mg analysis in the weld bead and ensure a good post-weld precipitation response.
   
• alloy 4047 adds Si to the weld bead at a much higher rate and is therefore most useful for processes and procedures that have inherently high base-metal dilution, above 70%. For instance, it is a popular choice for laser welding. One of the projects in this study employed 4047 filler metal at 50% to 70% dilution and obtained good weld-bead quality and tensile properties. Although it is beyond the scope of this study, we suspect that there may be some disadvantages to having excess silicon content in the weld bead (i.e., in ductility or corrosion behaviour), and therefore 4043 may be preferred over 4047 for procedures having intermediate levels of base-metal dilution.

Conclusions and Implications

Through careful choice of welding procedure, filler metal and post-weld heat treatment, significant and reliable increases in weldment strength can be obtained compared to traditional expectations for GMA structural welds in 6061-T6 plate.

The final strength of each zone in a weldment is dependent on the efficiency of the precipitation-strengthening mechanism in that zone. To ensure that the weld-fusion zone will not be the softest and weakest part of the weldment, a suitable 4000 series filler metal should be used, and the procedure should be designed so that dilution will produce an optimum weld-bead chemistry (~2-3% Si and ~0.3% Mg). Weld beads made with Al-Mg based 5000 series filler metals such as alloy 5356 have a much lower potential strength.

In the post-weld aged condition, the weakest part of a weldment made with optimal fusion zone composition will be the outer, overaged, haz. Therefore, the next step in upgrading the strength potential of alloy 6061 weldments is to explore methods of reducing the depth and width of this hardness "valley" in the haz. Based on information currently available, at least two promising approaches are evident. One is to carry out welding on material that is solution heat treated but not yet artificially aged (i.e., in T4 condition rather than T6). The other approach is to reduce the welding heat input as much as possible (i.e., to use larger numbers of smaller weld beads). In future student projects, we hope to clarify the feasibility and benefits of such procedural changes.

Acknowledgments

The projects summarized here were carried out by final year students Michael Armstrong, Kevin Johnston, Lindsay Derrah, Bradley Ellis, Katarina Slavik and Thomas Koeppe. The authors also want to express their appreciation to Bruce James of the CWB Gooderham Centre who assisted in supervision of the projects. Financial and/or material support was provided by Materials and Manufacturing Ontario, Alcotec Wire Inc., and the CWB Group.

W.H.S Lawson is Adjunct Professor and H.W. Kerr is Distinguished Professor Emeritus in the Department of Mechanical Engineering, University of Waterloo.