Accurate quantification of surface diffusion is essential for controlling semiconductor morphology, yet reported values for silicon and germanium remain highly scattered. Conventional approaches based on ex situ snapshots or simplified models fail to capture dynamic evolution. Here, we present an in situ method to derive surface diffusion coefficients of silicon and germanium under Joule heating. Real-time reflectance monitoring combined with Raman thermometry enables simultaneous tracking of hole shrinkage and surface temperature, and the resulting trajectories are quantitatively matched to phase-field simulations with a sub-50 nm RMSD accuracy. From these fits, surface diffusion coefficients of 5.5 × 10-12-1.0 × 10-10 m2/s for silicon (1050-1200 °C) and 5.0 × 10-12-8.0 × 10-11 m2/s for germanium (750-900 °C) are obtained. Arrhenius analysis over these temperature windows yields activation energies of 3.25 eV (Si) and 1.96 eV (Ge). The germanium coefficients are up to 2 orders of magnitude higher than literature values measured under oxide-suppressed high-vacuum conditions. These results establish a quantitative framework linking surface diffusion kinetics to morphology evolution during annealing.
Kim et al. (Mon,) studied this question.