Thermomechanics, material flow and microstructure evolution during Friction Stir Processing of light cast alloys
Friction Stir Processing (FSP) is a solid-state processing technique which can be used to refine and modify as-cast microstructures for superior properties. The aim of the present research is to investigate the following fundamental aspects associated with FSP cast Al and Mg alloys: the quantitative relationships between processing speeds (rotation and linear speeds: ω and v) and the thermomechanical responses (tool torque–M, power–P, specific energy–Es and material flow volumes–Vflow); the details of material flow/deformation and microstructural evolution during FSP of a cast Al-Si alloy; the mechanism governing the removal of Beta-Mg17Al12 particles during FSP of cast Mg-Al alloys. Experimentally, FSP of A356 cast alloy were performed with wide ranges of ω and v. M was measured during each FSP experiment, and temperature (T) at various locations were monitored for selected experiments. Based on M, P and Es were calculated. Stir zone areas (Aflow) were measured using the metallographic samples to estimate Vflow values. FSP experiments, using tool-pin-breaking technique, were conducted on A356 plates under two representative conditions. Macro-scale flow, micro-scale deformation, and microstructural evolution were studied by means of Electron Backscatter Diffraction technique. Pin-breaking FSP experiments were also conducted on AZ91/AM60 cast alloys during which T was monitored, based on which the thermomechanical and metallurgical explanations for the removal of Beta-Mg17Al12 particles were investigated. The relationship between M and ω is found to be well described by an exponential decay function: M = Mo + Mfexp(–nω); while the influence of v on M can be described reasonably well by linearly relating Mo, Mf, and n to v. Together with the consideration of temperature data obtained, M is shown to intimately relate to material flow resistance to tool motion. Thus n and Mf can be adjusted for alloying effect in the low ω range, while such effect diminishes as ω increases. It is shown that tool shoulder flow volume generated per revolution (VS–rev = Ashouderv/ω) relates to M, which can be interpreted as energy input per revolution, in a form of M = Mo + Mf[1 – exp(–γVS–rev)]. The tool-pin flow volume does not require a proportional amount of energy input. The larger diameter of the shoulder compared to the pin, coupled with higher material flow stress near shoulder region are the fundamental causes of this. A new flow mechanism that explains the formation of the non-ring nugget during FSP A356 was identified. It is shown that regardless of the processing condition, the highly refined portion of the nugget zone clearly segregates from the less refined portion. How this macro-segregation relates to the difference in flow regime inside and outside thread spaces is demonstrated in detail. The deforming dendrites located ahead of the pin were traced and based on this the strain and strain rate during FSP were directly estimated. The mechanism governing the recrystallization of α-Al dendrites was identified primarily as Geometrical Dynamic Recrystallization. Recrystallized α-Al grains around the pin displayed a dominating “A” shear texture, although the local “A” texture must undergo a degree of rotation to obtain the ideal “A” texture due to the local texture frame misaligned with the ideal texture frame. It was found that this misalignment is closely related to the direction of material flow at the location under consideration. Finally, detailed evidences suggest that the major mechanism governing the removal of the eutectic beta-Mg17Al12 particles is through a sequence of incipient melting of beta-phase, Al rich liquid wetting recrystallized α-Mg grain boundaries thus leading to a significant increase in liquid/solid interface, transfer of Al solute from liquid to interiors of α-Mg and the growth of α-Mg into the liquid (resolidification).