Metal Oxidation Composites from Al-Mg Alloys

2 Directed Metal Oxidation Composites from Al-Mg Alloys 

The mean intercept lengths of the ceramic and alloy channels, measured perpendicular to the growth direction, are about 5 pm and 2 pm, respectively. These intercept lengths are not related to the grain sizes, and both phases are monocrystalline over distances hundreds of times larger than the measured intercept lengths. In effect, the liquid metal in the composite develops a grain size upon solidification which is similar to that seen in a slowly cooled, cast aluminum alloy. 

The above mechanism was consistent with several experimental observations. MgO was known to be preferentially formed on the surface of Al-Mg alloys under many circumstances, as described earlier, either because of cation demixing in spinel, to form periclase, or due to rapid diffusion of Mg in the solid state and subsequent surface segregation. The single crystal oxidation product clearly precluded any mechanism that involved repeated nucleation of corundum (alumina) grains, which in any event is known to he difficult at temperatures below 1200°C. Finally, the presence of a film of molten alloy at a distance of about 1 pm from the surface accounted for the relative absence of a growth rate dependence on composite thickness (a rather slow decrease is actually observed). 

The steady-state microstructure described above is a simpUfication. The MgO layer is far from uniform, both with respect to position and time [50,56,66], and thicknesses 0.1 pm has occasionally been observed [66]. In some samples the interface between aluminum and MgO was found to contain spinel [47], while the solidified metallic phase frequently contained nanocrystals of MgO [66]. Even less consistent with the steady state hypothesis was the apparent equilibrium between aluminum and MgO in the presence of a magnesium concentration that was necessarily, and experimentally, close to the three-phase equilibrium with spinel and alumina [50]. Cation de-mixing of spinel, as a source of MgO (periclase), seems unlikely when spinel is the minor phase or apparently absent. An alternative hypothesis assumes that oxidation of magnesium in the vapor phase is responsible for the presence of MgO, an assumption that is supported by the demonstration of 

DM0 composite growth from aluminum alloys that contained no Mg but other, alternative volatile elements (Na, Li, Zn). In each case DMO growth occurs at temperatures for which the vapor pressure of the volatile solute is an appreciable fraction of the partial pressure of oxygen in the gas phase. 

---