No abstract

No abstract

Because joining dissimilar metals is often difficult by fusion joining, interest has been growing rapidly in using friction stir welding (FSW), which is considered a revolutionary solid-state welding process, as a new way to join dissimilar metals such as Al alloys to Mg alloys, Cu, and steels. Butt FSW of Al to Mg alloys has been studied frequently recently, but the basic issue of how the welding conditions affect the resultant joint strength still is not well understood. Using the widely used alloys 6061 Al and AZ31 Mg, the current study investigated the effect of the welding conditions, including the positions of Al and Mg with respect to the welding tool, the tool travel speed, and the tool rotation speed on the weld strength. Unlike previous studies, the current study (1) determined the heat input by both torque and temperature measurements during FSW, (2) used color metallography with Al, Mg, Al 3 Mg 2 , and Al 12 Mg 17 all shown in different colors to reveal clearly the formation of intermetallic compounds and material flow in the stir zone, which are known to affect the joint strength significantly, and (3) determined the windows for travel and rotation speeds to optimize the joint strength for various material positions. The current study demonstrated clearly that the welding conditions affect the heat input, which in turn affects (1) the formation of intermetallics and even liquid and (2) material flow. Thus, the effect of welding conditions in Al-to-Mg butt FSW on the joint strength now can be explained.

Cracking during solidification is a serious defect in welds and castings. A criterion for cracking that, unlike previous criteria/models, considers the grain boundary (where cracking occurs) was proposed recently. Based on this criterion, the maximum │dT/d(fS) 1/2 │ (T is temperature and fS fraction solid) of an alloy was proposed as an index for its crack susceptibility and the validity of the index was verified. The present study used the index to demonstrate the important effect of diffusion on the crack susceptibility and explained the surprisingly low crack susceptibility of Al-Mg alloys in welding. This was done by deriving closed-form equations to calculate the curves of crack susceptibility (the index) vs. solute content for binary Al-Si, Al-Cu and Al-Mg alloys. The curves were -shaped, consistent with those observed in the crack susceptibility tests of binary alloy systems. It was shown that diffusion from liquid into Al-rich dendrites reduces the peak of the -curve and shifts it to higher solute contents. The very high Mg solubility in solid Al, 17.5 wt% according to the Al-Mg phase diagram, makes Mg diffusion unusually significant, thus explaining the good weldability of Al-Mg alloys. The present study indicates the need to consider diffusion when comparing -curves of different alloy systems, where the phase diagram can affect diffusion, or different tests of the same alloy system, where the cooling rate during welding or casting can affect diffusion.

In welding, liquation cracking can occur in the partially melted zone, leaving open cracks along the edge of the weld bead. Likewise, solidification cracking can occur in the mushy zone, leaving open cracks inside the weld bead (which is called the weld metal or fusion zone). The present study aims at demonstrating that CALPHAD-based modeling can help predict the susceptibility of alloys to both types of cracking. The basic relationship between temperature T and the fraction of solid fS of an alloy can be calculated using thermodynamic software and a database based on the alloy composition. For liquation cracking the T-fS curve of the weld metal can be compared with that of the workpiece to assess the susceptibility. For solidification cracking, on the other hand, the T-(fS)1/2 curve of the weld metal can be used to calculate the susceptibility. The composition of the weld metal depends on the compositions of the workpiece and the filler metal, and the percentage of the workpiece in the weld metal (called dilution). The susceptibility predictions based on these curves and comparison with welding experiments will be demonstrated using Al alloys, Mg alloys, and carbon steels as examples.