Analysis of light extraction efficiency of GaN-based light-emitting diodes

Abstract

GaN-based light emitting diodes (LEDs) as one of the most important light source in next-generation solid-state lighting have been extensively studied and remarkable progress has been obtained. However, the light extraction efficiency (LEE) is not sufficient to satisfy application requirements. Most of the photons generated in multiple quantum well (MQW) always are trapped inside the semiconductor because both the reflection index of GaN and InGaN are higher than that of air, which results in total internal reflection (TIR). Great efforts were made in enhancing light extraction of LEDs experimentally in previous investigation. However, detailed theoretical studies in predicting the LEE of different types of LEDs are not available. In this paper the light extraction efficiency(LEE) of conventional chip (CC), flip chip (FC) and vertical chip (VC) is investigated using Monte Carlo ray tracing method is presented in conventional chip (CC). Monte Carlo ray tracing simulation based on statistics is known to be one of the most suitable methods to analyze the dependence of the light extraction efficiency of LEDs on the variety of the device parameters. Diffused bottom surface was found to be better for improvement of LEE than the perfect mirror surface. When there is a material with higher refraction index on the top surface the VC structure has the highest LEE no matter the bottom surface is perfect mirror or diffuse surface, which is above 80%. It is indicated that VC is potential in high-power GaN-based LEDs from the results.