Regression based computer vision analysis of volume-fraction effect on Pb-Sn solid-liquid coarsening in microgravity

Outcomes of coarsening experiments performed in solid-liquid mixture of various Pb-Sn systems, under microgravity, are investigated to explicate the influence of volume fraction on coarsening kinetics. A computer vision technique capable of efficiently ascertaining the number of Sn-rich solid precipitate is developed. After sufficient training and validation, this object detection model is employed to micrographs of Pb-Sn systems captured at different stages of coarsening. The change in the number density of the precipitate with time is monitored for different Pb-Sn systems of varying phase fractions, ranging from 5 to 80% precipitates, and the corresponding kinetics coefficient is determined using regression. Examining the kinetic coefficients in relation to phase fractions unravels that the volume of Sn-rich solid has negligible effect on the steady-state coarsening rate of the Pb-Sn solid-liquid mixture under microgravity. The indefinite and marginal effect of precipitate volume-fraction on kinetic coefficient apparently substantiates the trans-interface diffusion governed coarsening in Pb-Sn system, as opposed to generally held dominance of matrix diffusion.