Crack propagation in low dislocation density quantum dot lasers epitaxially grown on Si

Abstract

Low threading dislocation density in epitaxial lasers on Si is required for high performance and robust devices for silicon photonic integrated circuits. However, as the threading dislocation density is further reduced, a point is reached where it is energetically favorable for cracking to occur in the layers after cooldown to room temperature due to the thermal expansion coefficient mismatch between the film and the substrate. This can be solved in most cases by increasing the optical confinement and reducing the total layer thickness. We combine models of dislocation motion (controlling plastic relaxation) and thin film channel cracking to describe the impact of dislocation density and cooling rate, which addresses a well-known and previously unsolved problem in heteroepitaxial growth for optoelectronic and electronic devices. Agreement between predictions and experiments illustrates that the model is effective in identifying critical dislocation densities, film thickness, and cooling rates that avoid cracking. We believe that this work is the first attempt to merge the fracture mechanism and dislocation-mediated plastic relaxation in semiconductor films to solve a practical problem in optoelectronic materials.