Joining aluminium alloy MMCS

TWI Bulletin, September/October 1992

Mike Gittos is a Principal Metallurgist in the Materials Department. He graduated from Manchester University, joining the then Metallurgy Department at TWI in 1974. Although most of his work has been concerned with the welding of non-ferrous metals, his published work also relates to ferritic and stainless steels as well as plastics materials.

Philip Threadgill of TWI's Materials Department is Section Leader of Non-Ferrous and Advanced Materials and is responsible for business development in advanced metallic materials. He graduated from the Department of Metallurgy at University College, Swansea with a bachelor's degree in Physical Metallurgy in 1968 and completed his doctorate in 1973. During the mid-seventies he was a research fellow at Liverpool University where his main interests were in ferrous powder metallurgy and the creep strength of complex magnesium alloys. During the late eighties he was seconded briefly to the Edison Welding Institute in Columbus, Ohio.

Wide ranging developments in improved MMC (metal matrix composite) alloys have taken place recently and there are now several industrial applications in place or under detailed consideration. As Mike Gittos and Philip Threadgill report, staff at TWI have investigated the feasibility of joining a number of particulate reinforced aluminium alloys both to themselves and to other unreinforced aluminium alloys.

With all new materials, a need to join them to the same or to other materials is inevitable, and it can be argued that the industrial potential of the materials will be severely limited if joining procedures cannot be developed. Ideally, joining technology should provide a variety of processes suitable for manufacturing high integrity joints with optimum mechanical properties, and ease of inspection. It is unlikely that all these criteria can be fully satisfied for MMCs in the short term.

Solid phase processes

Friction welding has proved very successful in making sound joints. Two components are moved against each other under pressure, developing frictional heat. The relative motions can be of several types, e.g. rotary, orbital or linear. In the present work rotary motions have been most widely used. A soft layer is formed at the interface, and this undergoes intense plastic deformation, much of the materials being extruded from the joint. Once this layer has formed, a solid state bond exists between the two components, and it is normal to allow the bond to cool under pressure.

From a metallurgical point of view, examination of the interface (which is often hard to find exactly) shows a continuous microstructure with no evidence of microstructural discontinuities across the interface. What is sometimes found is evidence of fragmentation of the reinforcement, caused by repeated impacts during the friction process. This phenomenon is not always observed, see Fig.1.

So far, only limited mechanical property data have been obtained, but the results are encouraging. Yield values in cross weld tensiles of up to 90% of parent material value have been obtained. It is probable that suitable PWHT would give even better results, as the age hardening matrix would have been subjected to solution treatment or overageing, depending on location and thermal cycle.

Joining trials using diffusion bonding have demonstrated that joints can be made by this process. Preliminary trials on pure Al/20%SiC, and 6061/20%SiC produced successful bonds, using Cu, Ag or no interlayers. Metallographic examination confirms the need for precise control of the process variables, to ensure that particulate rich zones, or particulate free zones are not formed at the interface. Some typical bond line microstructures are shown in Fig.2.

Fig. 2. Microstructures in diffusion bonds showing:

a) Insufficient

(a) Al/20%SiC, Cu interlayer

b) Correct

b) 6061/20%SiC, Ag interlayer

Fig.2c) Excessive mass transport

c) Al/20%SiC, Ag interlayer)

Mechanical data on these processes are few, but it is known that a bond which appears sound when examined microstructurally does not necessarily give good mechanical properties. Diffusion bonding of aluminium is not easy, because of the tenacious and stable oxide, so it is difficult to achieve reproducibly satisfactory bonds in the matrix alloys; addition of the reinforcing particulate imposes further restrictions on process variables.

Fusion processes

Fusion processes are economical, flexible, and generally require only modest capital investment. They are well established for aluminium alloys, and are attractive for joining metal matrix composites. However, experience has highlighted a number of problems which must be considered. The first is that molten aluminium MMC materials are very viscous, even at temperatures considerably in excess of the melting point.

This means that it is difficult to use conventional filler alloys. The two liquids do not mix satisfactorily in the few seconds that the pool remains molten, and care must be taken to minimise the melting of the MMC material. This effect may not be a problem if low strength welds are required, but it has significant implications for design of welded structures which incorporate metal matrix composites. A section through a TIG weld in an 8090/20%SiC alloy, using a 4047A filler ( Fig.3) shows an extreme case of the lack of mixing. However, Duralcan have demonstrated that aluminium MMCs can be TIG or MIG welded with procedures chosen to minimise dilution, thus overcoming the poor miscibility of the molten MMC. A second feature is that of rejection of the particulate by the solidifying interface. Most MMCs are reasonably homogeneous, but following fusion welding, ceramic particles agglomerate in the interdendritic regions, reducing their efficiency as a reinforcement.

The magnitude of this effect and its significance are clearly influenced by the particle size and the cooling rate, which determines the dendritic cell size. In large weld pools, gravitational effects may also lead to enhanced macro segregation, although there are no known investigations of this.

With certain particulate/matrix combinations, there is a significant risk of particle/ matrix interactions. In particular, SiC will react with aluminium at high temperatures to produce an aluminium carbide, Al ₄C ₃ ( Fig.4).

This phase is extremely brittle, but is also unstable, and reacts with any moisture to form acetylene. Experience has shown that the life expectancy of welds where this reaction has occurred can be very short, perhaps only days. However, the reaction can be suppressed by use of high Si alloy matrices, or by avoidance of high energy density processes which raise the peak weld pool temperature.

Alternatively, the use of other reinforcements such as Al ₂O ₃, which is stable in molten Al, will overcome the problem.

Another problem may occur with some MMCs made by powder metallurgical techniques. If the occluded gas content is too high, gas will come out of solution at high temperatures, giving extensive cracking in the HAZ. However, this does not appear to be a problem with current powder route MMCs.

Finally, a phenomenon which has so far defied rational explanation is a mass segregation effect found in some resistance welds, in which the SiC particles segregate preferentially to the periphery of the molten nugget, see Fig.5. It is not clear whether this is caused by a hydrodynamic effect generating a centrifugal force, or is perhaps a surface tension effect. In practice, the explanation is of academic interest only, as the effect can be suppressed by careful control of the welding parameters.

Conclusions

Industrial implementation of MMCs is to a large degree dependent on development of various manufacturing processes, of which joining is only one. Work to date has proved encouraging, but it is evident that much more is required to develop joining processes further, and to understand the responses of the materials to these processes.

Acknowledgement

This work was funded by Industrial Members of TWI and the Minerals and Metals Division of the UK Department of Trade and Industry.