Material bond, formation, and growth, in Al/Mg compound castings

Abstract

Al/Mg compound castings possess a number of benefits compared to conventional single material castings. High specific strength, low weight, economic viability and the possibility for a “tailored-to-need” design have been cited. In compound castings, the material bond is of great importance to the mechanical stability and soundness of the compound casting and thus ultimately their economic and practical viability. Whilst aluminium and magnesium have been successfully joined by a number of processes, little is known about the formation of the material bond in the compound casting process and how to influence it. Consequently, in this study a methodology was devised to assess the factors influencing the formation of this material bond. The results of the experimentation were characterised using SEM, EDS, XRD, DSC measurements and mechanical testing. As similar to the solid state joining of the two metals, the primary constituents of the material bond were found to be the intermetallic phases β(Al3Mg2) and γ(Al12Mg17). Aluminium’s intrinsic oxide layer did not inhibit the formation of a bond at the interface between the two metals. In fact, the oxide layer was found to be reduced by the magnesium melt upon contact, forming MgO. The resultant MgO is then dispersed in the surrounding melt. Dissolution of the solid aluminium into the magnesium melt and the subsequent precipitation of the material bond from the liquid phase was identified to be the main mechanism behind the formation of the material bond. A direct correlation between the amount of aluminium and thickness of the material bond exists. Based on the results from experimentation, a step-by-step model was developed to explain and understand the underlying mechanism and processes that contribute to the formation and growth of the material bond in Al/Mg compound castings. The different steps were identified as: 1. Initial contact of melt and solid metal 2. Dissolution of the solid metal and solidification of the melt 3. Growth of the material bond due to solid state diffusion The addition of silicon and zinc resulted in the formation of phases according to the corresponding ternary phase diagrams. The addition of silicon resulted in the formation of an Mg2Si phase. The formation of this phase was found to be an exception, resulting from a diffusion reaction process rather than precipitation from the liquid phase. Unlike silicon, zinc was dissolved by the magnesium and crystallised during solidification on the crystal lattice of the intermetallic phases β and γ. The zinc- rich phases τ1 and φ were only formed in the presence of an abundance of zinc and displayed a very low solidification range of 367-433°C. The low melting temperature and rapid and imbalanced diffusion of zinc caused the formation of evenly shaped voids (similar to the Kirkendall effect). This was only observed with zinc concentrations greater than 15wt%. Based on the observations made during experimentation and assessment of the compound castings, a model was developed to explain the formation of these aforementioned voids. The temperature distributions during casting, solidification and cooling at the solid/liquid interface were measured and simulated with the CAE software package Magmasoft 5.31. As the formation of the interface is a complex interaction of temperature, dissolution, diffusion and solidification it cannot be precisely simulated. Despite a number of discrepancies between simulation and reality, a simulated and measured distribution was determined, that was in relatively good accordance, especially at higher temperatures. The melt composition near the solid/liquid interface was changed by the dissolution of aluminium into the magnesium melt. As a result, the melt near the solid/liquid interface solidified last, thus causing shrinkage, which negatively affected the soundness of the bond. Furthermore, soundness was impaired by fractures and cracks, thought to originate during cooling or handling of the castings. Mechanical strength of the material bond was approximated with Vickers micro-hardness measurements and push-out testing. Regardless of the used alloys and parameters, the bond displayed high brittleness with the push-out tests revealing that the material bond always fails in the aluminiumrich region. The magnesium- rich phases γ and φ (in the presence of zinc) showed signs of increased ductility (compared to the aluminium- rich phases β and τ1). Overall push-out resistance of the bond varied between 5-25MPa. Although no correlation between the thickness of the material bond and pushout resistance was evident, push-out resistance was found to be greatly affected by the soundness of the castings. Micro-hardness of the intermetallic phases in the material bond was up to 300Hv higher than that of the used aluminium and magnesium alloys. The magnesium- rich phases β and φ were found to exhibit a slightly lower hardness than the aluminium- rich phases β and τ1, whilst retaining some ductility. Several problems/drawbacks of compound castings have been identified in this study. Especially the low mechanical strength of less than 25MPa, which is considerably below than that of other joining techniques, is to be seen as problematic. Moreover, the material bond was found to be highly susceptible to galvanic corrosion, with signs of corrosion already being present on the samples after preparation. Lastly, fractures and pores were commonly found within in the bond of the compound castings, reducing the soundness of it. These problems/drawbacks will need to be overcome if Al/Mg compound castings are ever to become a viable alternative to already established and proven joining techniques.