Comment on pyramidal structure formation at the interface between III/V semiconductors and silicon

TL;DR

This paper discusses the atomic structure of GaP/Si interfaces, focusing on the formation of pyramidal structures, and highlights discrepancies between recent DFT studies and experimental observations regarding interface energetics.

Contribution

It critically examines previous DFT studies on pyramidal structure formation at GaP/Si interfaces, emphasizing the importance of chemical potentials in interfacial energy calculations.

Findings

DFT studies neglected chemical potential dependence

Experimental data contradict some DFT predictions

Interface structure critically influences defect densities

Abstract

GaP/Si(100) is considered as pseudomporphic virtual substrate for III/V-on-Si integration in order to reduce defects related to polar-on-nonpolar heteroepitaxy. The atomic structure of the GaP/Si(100) heterointerface is decisive to yield low defect densities and its dependence on nucleation conditions is still under debate. Recently, Beyer et al. suggested the formation of a 'pyramidal' structure as a general mechanism at polar-on-nonpolar interfaces [A. Beyer et al., Chem. Mat. 28, 3265 (2016)]. However, their DFT studies neglected the dependence of the calculated interfacial energies on appropriate chemical potentials and their findings are contradictory to recent and past experimental data.