Compositional degradation of the electron blocking layers through solid-solution in GaN-based laser diodes

Electron leakage currents seriously hinder GaN-based blue laser diodes (LDs) from high wall-plug efficiencies. Inserting an ultra-thin AlGaN electron blocking layer (EBL) in the epitaxy structure is a major technique to suppress the leakage currents for which a high Al composition in the EBL is necessary. Despite many studies on the optimization of the compositions of EBLs, it is questionable whether they reach the designed value in real growths by metal-organic vapor phase epitaxy. We investigate the influence of the growth conditions of upper cladding layers (CLs) on the underlying EBLs. A strong composition degradation of the EBL is observed when the growth rate of the CL is low, which drastically reduces the output performance of both LEDs and LDs. A 30-nm fast-growing protecting layer can efficiently prevent the EBL from such degradation. The phenomenon cannot be explained by a composition pulling effect nor an etch effect by hydrogen, but by a mutual solid solution between the EBL and the adjacent CL. The solution process is found thermally favored by calculating the Gibbs energy where strain and entropies are considered. It is inferred that the chemically active Ga adatoms at the surface play an important role in accelerating the solution process. Based on these considerations, we introduce a random walk model to clarify the kinetic influence of CL growth rates on EBL degradation semi-quantitatively. The results help to understand the subtle process in the growth of heterostructures and the transport process of GaN-based LDs.

Automated summary

Inserting an ultra-thin AlGaN electron blocking layer is a major technique to suppress electron leakage currents in GaN-based blue laser diodes. It has been found that the growth conditions of the upper cladding layers affect the composition of the underlying electron blocking layer and significantly reduce the output performance of LEDs and LDs. By introducing a random walk model, the kinetic influence of cladding layer growth rates on electron blocking layer degradation can be quantitatively clarified.