Top transmission grating GaN LED simulations for light extraction improvement

Simeon Trieu, Xiaomin Jin, Ashton Ellaboudy, Bei Zhang, Xiangning Kang, Guoyi Zhang, Chang Xiong, Wei Wei, Sun Yong Jian, Fu Xing XingProceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE 2011 · DOI

TL;DR

This paper explores how top-surface transmission gratings can significantly boost the light extraction efficiency of Gallium Nitride (GaN) Light Emitting Diodes (LEDs). By simulating various grating designs using the FDTD method, the researchers found that optimized transmission gratings can improve light output by up to 165%, outperforming traditional reflection gratings.

What Problem Does This Paper Address?

Light Emitting Diodes (LEDs) are crucial for modern lighting and display technologies, but their efficiency is often limited by how well they can extract the light generated internally. For GaN-based LEDs, two main issues hinder this: 1) the high refractive index of GaN causes total internal reflection, trapping light within the device, and 2) generated photons can be absorbed by the multiple quantum well (MQW) region or defects before escaping. While various methods exist to improve light extraction, such as adding resin domes or using reflection gratings, this paper focuses on optimizing transmission gratings on the top surface of GaN LEDs. The goal is to find grating designs that effectively redirect internally trapped light out of the LED, thereby increasing overall efficiency.

Methods

The researchers employed the finite difference time domain (FDTD) method, a powerful computational technique for simulating electromagnetic wave propagation. This method allows for precise modeling of light-matter interactions, including reflection, refraction, and diffraction, and offers the flexibility to define custom material properties and complex geometries.

They simulated two primary grating shapes: conical and cylindrical pillars. These gratings were applied to the top surface of a GaN LED model. To represent the grating material, they considered three scenarios: non-lossy p-GaN, lossy indium tin oxide (ITO), and non-lossy ITO. The key parameters varied for the gratings were:

  • Grating cell period (Λ): The distance between repeating grating units.
  • Grating cell height (d): The depth of the grating features.
  • Grating cell width (w): The lateral dimension of the grating features.

The simulations explored a wide range of these parameters to identify optimal configurations. The results were then compared against a "flat plate" (no grating) reference case and also against previously studied bottom reflection gratings.

Datasets

Not discussed in the paper. The study relies on simulated data generated by the FDTD method, rather than experimental datasets.

Key Insights and Ideas

  • Top transmission gratings are superior: The study strongly suggests that top transmission gratings offer a more effective solution for light extraction improvement compared to bottom reflection gratings.
  • Close packing is key: Maximizing light extraction efficiency is achieved when grating cells are closely packed, meaning the grating period (Λ) is similar to or equal to the grating width (w). This suggests a move towards denser, potentially photonic crystal-like structures.
  • Material choice matters, but geometry is dominant: While the material properties (lossy vs. non-lossy ITO) influence the absolute efficiency, the geometric parameters (Λ, w, d) and the grating shape (conical vs. cylindrical) play a more significant role in determining the improvement in light extraction.
  • Nano-scale features are promising: The results indicate that grating structures with feature sizes approaching the wavelength of light (e.g., 460 nm) or its fractions (half-wavelength, quarter-wavelength) are likely to yield even better performance. This points towards future research in nano-scale photonic crystal designs.

Key Results

  • Significant Efficiency Gains: Optimized top transmission gratings achieved an impressive 165% improvement in light extraction efficiency compared to the no-grating case.
  • Outperforming Reflection Gratings: This 165% improvement significantly surpassed the gains from bottom reflection gratings (112%) and top-bottom combined gratings (42%).
  • Optimal Parameters: For non-lossy p-GaN, conical gratings with Λ=1 µm, w=1 µm, and d=200 nm showed the best results. For lossy ITO, conical gratings with Λ=4 µm, w=4 µm, and d=166 nm were optimal. For non-lossy ITO, conical gratings with Λ=4 µm, w=4 µm, and d=166 nm also showed high efficiency.
  • General Trend: In most cases, light extraction efficiency increased as the grating period (Λ) approached the grating width (w), indicating that denser grating patterns are more effective.

Limitations and Future Work

  • Simulation vs. Fabrication: The study is based on simulations. While FDTD is a powerful tool, real-world fabrication challenges and imperfections could affect the actual performance of these gratings.
  • Material Simplification: The use of simplified material models (e.g., non-lossy ITO) might not fully capture the complexities of real device materials.
  • Future Directions: The authors explicitly suggest exploring more closely packed structures, such as 3-fold symmetric photonic crystal structures with triangular symmetry, and investigating smaller feature sizes in the nano-scale, comparable to the wavelength of light (460 nm) and its fractions. This indicates a clear path for future research towards even higher efficiency LEDs.

Summary updated

Abstract

We study the top transmission grating’s improvement on GaN LED light extraction efficiency. We use the finite difference time domain (FDTD) method, a computational electromagnetic solution to Maxwell’s equations, to measure light extraction efficiency improvements of the various grating structures. Also, since FDTD can freely define materials for any layer or shape, we choose three particular materials to represent our transmission grating: 1) nonlossy p-GaN, 2) lossy indium tin oxide (ITO), and 3) non-lossy ITO (α=0). We define a regular spacing between unit cells in a crystal lattice arrangement by employing the following three parameters: grating cell period (Α), grating cell height (d), and grating cell width (w). The conical grating model and the cylindrical grating model are studied. We also presented in the paper directly comparison with reflection grating results. Both studies show that the top grating has better performance, improving light extraction efficiency by 165%, compared to that of the bottom reflection grating (112%), and top-bottom grating (42%). We also find that when grating cells closely pack together, a transmission grating maximizes light extraction efficiency. This points our research towards a more closely packed structure, such as a 3-fold symmetric photonic crystal structure with triangular symmetry and also smaller feature sizes in the nano-scale, such as the wavelength of light at 460 nm, half-wavelengths, quarter wavelengths, etc.
BibTeX
@inproceedings{trieu2011top,
  author = {Simeon Trieu and Xiaomin Jin and Ashton Ellaboudy and Bei Zhang and Xiangning Kang and Guoyi Zhang and Chang Xiong and Wei Wei and Sun Yong Jian and Fu Xing Xing},
  title = {{Top transmission grating GaN LED simulations for light extraction improvement}},
  booktitle = {Proceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE},
  year = {2011},
  doi = {10.1117/12.871595},
}
Figure