The solidification of the beta-Sn phase in Sn-based solder alloys is often unpredictable and difficult to induce, commencing heterogeneously from a highly-undercooled liquid. In addition, the crystal structure of beta-Sn is highly anisotropic and influences properties in the bulk material, such as elasticity, thermal expansion, resistivity, and diffusivity. Couple this intrinsic anisotropy with the difficulty in nucleation, and the outcome is a high level of variability in solder joint mechanical and thermal properties. Microstructural results, such as primary phase formation and beta-Sn grain size, as well as measurable experimental values related to the efficacy of beta-Sn nucleation, such as beta-Sn undercooling, range widely in solder alloys due to experimental variation across the literature in regard to a number of different variables, including, but not limited to: solder sample volume, solidification cooling rate, substrate materials used, fluxes used during experimentation, atmospheric conditions, overall solder composition, etc. Despite the wide variation in the existing literature, it is clear that the examined approaches have yet to result in a reliable method for enhanced beta-Sn nucleation, particularly for the small sample volumes that are relevant to microelectronics applications. Thus, my research is aimed to address this need for enhanced beta-Sn nucleation, while actively seeking to promote grain refinement though the application of solidification theory, common to large-scale castings, to the small-scale of microelectronics.

My work has examined the nucleation, growth, and structure of beta-Sn via the implementation of the following thermodynamic and solidification approaches: (1) Al micro-alloying additions, (2) introduction of inoculating phases, and (3) introduction of growth-restricting solutes in the small solder sample volumes (0.16--0.30 cm3 -- solder sphere diameters of 250--500 microm) relevant to the microelectronics industry. The results were an enhanced understanding of beta-Sn nucleation and structure through industrial, governmental, and academic collaborations, implementing both experimental and computational methods.

First, the effects of Al micro-alloying on common industrial solder alloys was examined. In our initial studies, Al additions resulted in a range of microstructure effects, including the formation of Cu-Al IMCs. Through collaboration with colleagues at Iowa State University (ISU), Ames Laboratory, and Nihon-Superior Co., the nucleation and thermodynamic stability of CuxAly IMCs was investigated through the compositional variations of Sn-Cu-Al and Sn-Ag-Cu-Al alloys. The alloys produced CuxAly particles of varying morphologies and stoichiometric phases, and a trend of increasing CuxAly volume fraction with increasing Al alloying content was established.

Through continued collaborations with ISU, Ames Lab, and Nihon-Superior, the potential nucleant relationship between CuxAly and Cu6Sn5 was explored as a means to manipulate the number and size of Cu6Sn5 nucleation sites through the particle size refinement of CuxAly via rapid solidification. Cooling rates spanning eight orders of magnitude were used to refine the average CuxAly and Cu6Sn5 particle sizes down to submicron ranges. Deep etching of the samples revealed the three-dimensional microstructures and illuminated the epitaxial and morphological relationships between the CuxAly and Cu6Sn5 phases. Initial solidification cooling rates within the range of 10 3 to 104 °C/s were found to be optimal for realizing particle size refinement and maintaining the CuxAly/Cu 6Sn5 nucleant relationship. Additionally, the coarsening behavior of CuxAly and Cu6Sn5 was characterized after multiple (1--5) re-melting (reflow) cycles via DSC between 20 °C and 250 °C. Little-to-no coarsening of the Cu xAly particles was observed for the two compositions studied.

For Cu6Sn5, a bimodal size distribution was observed for the alloy with Cu6Sn5 phase temperature stability of 240 °C, with large, faceted growth of Cu6Sn5, while the alloy with Cu6Sn5 phase temperature stability of 267 °C displayed no significant increase in the average particle size after reflow. The link between original alloy composition, reflow undercooling, and subsequent IMC coarsening behavior was discussed by using calculated solidification paths. The reflowed microstructures suggested that the heteroepitaxial relationship between the CuxAly and the Cu6Sn5 was maintained after reflow cycling. These studies highlighted the utility of manipulating solder IMC spatial relationships, during both initial solidification and reflow cycling, as an important tool in improving solder joint mechanical properties and reliability through IMC structural and morphological coordination. (Abstract shortened by ProQuest.).