Photonic-crystal surface-emitting lasers (PCSELs) are characterized by single longitudinal mode, high beam quality, small divergence angle, and narrow linewidth. Compared with traditional edge-emitting lasers (EELs) or vertical cavity surface-emitting lasers (VCSELs), PCSELs offer advantages in beam quality, high-speed modulation, and power performance. Their unique photon confinement properties allow for a high-quality, single-mode beam at the watt scale by increasing the gain area of the active region. This makes them become ideal future laser sources, combining the benefits of GaN materials and blue lasers to achieve high power output with excellent beam quality under single-device, single-mode conditions. Consequently, they have attracted significant attention. In traditional GaN-PCSEL designs, the holes of the photonic crystal (PhC) are typically buried near the active layer to enhance the optical field confinement factor. However, even with this approach, the confinement factor remains limited, and the complex fabrication process involved in burying PhCs can lead to structural disorder, weakening the photonic crystal resonance effect and coupling strength. Furthermore, the refractive index limitations of AlGaN and the high-quality epitaxial growth requirements make GaN material systems less amenable to efficient distributed Bragg reflectors (DBRs) and effective optical field confinement layers compared to GaAs and InP-based systems. As a result, high-performance GaN-based surface-emitting lasers have faced challenges in advancement. Therefore, we present a theoretical simulation of blue light PCSELs, which is expected to provide valuable insights for practical applications.

The GaN-PCSEL laser designed in this study utilizes a unique combination of the photonic crystal layer and the DBR/AlGaN composite confinement layer, as shown in Fig. 1. The design of the PCSEL laser gain cavity is based on layer parameters of traditional GaN active systems. Initially, we use the plane wave expansion (PWE) method to obtain the TE-mode photonic crystal cavity energy band diagram and determine the lattice constants. Since resonance in PCSELs occurs in the PhC layer and the adjacent gain medium, the etched holes resonate at specific wavelengths by forming a periodic photonic crystal structure. We systematically investigate the relationship between the thickness of the photonic crystal layer and etching depth and use rigorous coupled wave analysis (RCWA) and finite difference time domain (FDTD) methods to simulate the device. We obtain the resonance map and the electric field distribution of the fundamental mode across the device. In addition, we simulate the transmission and reflection spectra of GaN/porous-GaN DBRs using the finite element method (FEM) and compare the effects of DBRs on slope efficiency.

We achieve a porous-GaN/GaN DBR structure with a reflectivity of up to 99.9% and a wide reflective bandwidth of over 100 nm (Fig. 3). The slope efficiency of the GaN-PCSEL is doubled (Fig. 7), and the PhC confinement factor reaches up to 8% for the appropriate photonic crystal thickness and etching depth [Figs. 4(g) and 4(h)], with a gain threshold of 217 cm^-1.

In this study, we propose using TiO₂ as a photonic crystal layer material on the GaN surface to leverage its high refractive index for modulating the resonant optical field distribution. Combining it with porous-GaN/GaN DBRs, which have excellent reflective properties, results in a PCSEL structure with a high confinement factor, low gain threshold, and high power. By optimizing the photonic crystal and active region parameters, we achieve a PhC confinement factor exceeding 8% and a gain threshold of 217 cm^-1 at the appropriate photonic crystal thickness and etching depth. In addition, the porous-GaN/GaN DBR composition achieves over 99% reflectivity and a broad reflective bandwidth of more than 100 nm in the blue wavelength range, doubling the slope efficiency of the GaN-PCSEL. However, increasing slope efficiency may lead to higher gain thresholds. Through careful parameter selection and trade-off optimization, relatively high slope efficiencies can be achieved while maintaining a low threshold.