Introduction
At present, the traditional fiber grating demodulation system uses a large volume and high price, which limits its popularization and application fields. It is particularly important to produce an on-chip integrated light source for the fiber grating demodulation system. Although there have been some reports of monolithic integrated silicon-based light sources [1, 2], it is still challenging to achieve efficient monolithic integrated silicon-based light sources in the short term. At present, the integrated light source adopts two methods: an external light source and a hybrid integrated on-chip light source [3]. As one of the on-chip integrated light sources, light-emitting diodes (LEDs) have their unique advantages, including low cost, low power consumption, small size, long life and high reliability. However, the main disadvantage of LED as an on-chip integrated light source is its low light extraction efficiency. This is because the semiconductor material of the LED active layer has a higher refractive index than air, and the light is totally reflected at the interface between the LED medium and the air, so that most of the light cannot be emitted from the LED. It is emitted and absorbed by metal contacts, substrates or active layers, resulting in low light-emitting efficiency of LEDs [4-6]. How to improve the light-emitting efficiency of LEDs has become a hot issue.
There are many ways to improve the efficiency of LED light extraction, such as growth distribution Bragg reflection layer structure [7], surface roughening [8], fabrication of transparent substrate [9] and flip chip bonding [10]. The Bragg reflection layer is composed of multiple layers of high refractive index and low refractive index materials. The larger the refractive index difference, the higher the DBR reflectivity, thereby reducing the light absorption of the substrate and improving the light extraction efficiency of the LED; the surface roughening is by increasing the transmittance. The direction of light that satisfies the law of total reflection is changed, and then the other surface is reflected back without being totally reflected and transmitted through the interface to achieve the purpose of improving light emission; the transparent substrate is formed by reducing the absorption of light by the substrate. The light exiting is improved; the flip-chip bonding structure can reduce the absorption of light by the active layer and the free carriers caused by the reflection of light inside the LED, and increase the light emission. These methods have different improvements in LED light extraction efficiency, but the effects are not ideal.
Since Yablonovitch proposed photonic crystal (PC) in 1987 [11], photonic crystals have attracted the interest of researchers. Since the two-dimensional photonic crystal is periodically arranged by a medium having a different dielectric constant, it has
There is a photonic band gap similar to a semiconductor forbidden band. The photonic band gap limits the light field propagation in the horizontal direction, improves the light emission in the vertical direction, and achieves the purpose of improving the light output efficiency of the LED. Research on the use of photonic crystals to improve the light-emitting efficiency of LEDs has been reported both at home and abroad. The foreign Taesung Kim research team studied infrared LEDs [12], which increased the light-emitting efficiency of LEDs by 75% after adding photonic crystals on their surfaces. The domestic semiconductor research group used electron beams. Exposure and inductively coupled plasma technology produces photonic crystals on InP-based LED surfaces [13], which maximizes LED light extraction efficiency by 93%. However, the photonic crystal band structure is not systematically analyzed, and the photonic crystal structure with the best LED light extraction efficiency is sought.
In this paper, the basic structure of the on-chip C-band LED in the integrated micro-system researched by the research group is designed, and the structure of the C-band LED is simulated. The spectral characteristic curve and output optical power curve of the C-band LED are obtained. The finite-difference time-domain (FDTD) method is used to calculate the different duty ratios of different arrays (Rp=r/a, r is the photonic crystal cell radius, a is the photonic crystal lattice constant). The photonic crystal band structure is used to analyze the relationship between the forbidden band width of the photonic crystal and the duty cycle Rp. The forbidden band theory is used to find the optimal Rp and the normalized center frequency f0(a/λ), so that the photonic crystal structure parameters for improving the light-emitting efficiency of the C-band LED are obtained.
1 C-band LED basic structural parameters
The basic structure of the designed C-band LED is shown in Figure 1. It is a continuous growth of 4 layers on the n-InP substrate, which are: n-InP buffer layer, InGaAsP active layer, p-InP confinement layer and p-InGaAsP. Top level.
