Findings
Nanophotonic Quantum Well LEDs: Analysis of Light Extraction Efficiency in GaN-Based Coaxial Microwall Light Emitting Diodes.
REU Student: Michael Wabaunsee
Graduate Student Mentor: Mohsen Nami, Ashwin .K. Rishinaramangalam
Faculty Mentor: Dr. Daniel Feezell
A. Results
Bulk gallium nitride:
We started with a simple simulation, where theoretical results are easily achievable. We put a point source in the gallium nitride (GaN) bulk material and tried to find extraction efficiency into air and epoxy. The index of refraction of epoxy in the simulation is 1.5.
Point Source at the center of Bulk GaN (n=2.52) / Theory / Simulation
58.93 / 58.94
Model / Extraction Efficiency in air (%)
Point Source at the center of Bulk GaN (n=2.52) / Theory / Simulation
24.63 / 24.66
Flip chip LED:
We moved to a more complicated structure, which is called Flip Chip LED. The geometry and structure parameters was shown in the following picture:
Material / Color / W (µm)×L(µm)×T(µm) / Index of refraction / a(1/cm)n-Gan / Green / 300×300×(50-300) / 2.52 / Variable
p-GaN / Purple / 300×300×0.15 / 2.52 / Variable same as n-GaN
Source (InGaN) / Yellow / 300×300×0.02 / 2.77 / 0
Ag Mirror / Gray / 310×310×0.1 / --- / Mirror: R=95% A=5%
Coaxial microwall structure:
Material and device parameters used in the simulation:
Material / Color / Thickness(µm) / Index of refraction / a(1/cm)Indium tin oxide (ITO) / Yellow / 0.2 / 2.035 / 1000
P-GaN / Light Blue / 0.15 / 2.52 / 30
Source (InGaN) / Yellow / 0.02 / 2.77 / 0
n-GaN Microwall / Green / Height(2µm)×Width(2µm) / 2.52 / 3
SiNx / Dark Blue / 0.15 / 2.055 / 0
Sapphire / Purple / 400 / 1.78 / 0
n-GaN Substrate / Green / Height(2µm)×Width(4µm) / 2.52 / 3
Ag / Not Shown / 0.1 / --- / Mirror: R=95% A=5%
We found the relation between the extraction efficiency and microwall length, as shown in the graph. By increasing the length of the microwall, more rays would be trapped inside the structure due to total internal reflection (TIR) and thus the extraction efficiency decreases.
Adding Texture:
Three different types of textures have been simulated.
· Hemispherical texture.
· Prism type texture.
· Equilateral triangular base pyramid.
1) Hemispherical Texture:
Dx (nm),Dy (nm) / R (nm) / Extraction Efficiency (%) / K500 / 24 / 68.58 / 7.23×10-3
150 / 20 / 70.68 / 5.58 ×10-2
600 / 200 / 75.98 / 3.48 ×10-1
150 / 50 / 76.79 / 3.48 ×10-1
150 / 54 / 77.26 / 4.06 ×10-1
100 / 45 / 78.03 / 6.35×10-1
24 / 11.5 / 78.21 / 7.20 ×10-1
24.5 / 12 / 78.28 / 7.53×10-1
2) Prism type texture:
Dx (nm),Dy (nm) / L (nm) / W(nm) / H(nm) / Extraction Efficiency (%) / K
100 / 20 / 20 / 10 / 69.67 / 0.04
50 / 20 / 20 / 10 / 73.6 / 0.16
50 / 25 / 20 / 10 / 74.25 / 0.2
50 / 35 / 20 / 10 / 75.16 / 0.28
50 / 40 / 20 / 10 / 75.26 / 0.32
50 / 45 / 20 / 10 / 74.75 / 0.36
50 / 47 / 20 / 10 / 73.9 / 0.376
50 / 49 / 20 / 10 / 72.72 / 0.392
50 / 35 / 40 / 20 / 77.69 / 0.56
50 / 40 / 47 / 23.5 / 77.65 / 0.752
3) Equilateral triangle based pyramid texture:
Dx (nm),Dy (nm) / W(nm) / H(nm) / Extraction Efficiency (%) / K100 / 46.5 / 16.5 / 71.2 / 0.093
93 / 88 / 30 / 73.88 / 0.387
Results of adding texture:
· Adding texture increases the light extraction efficiency (LEE) by ≈ 12%.
· Hemispherical texture increases the LEE the most.
· Increasing the texture filling factor for hemispherical textures can increase the LEE.
Multi-structure coaxial microwalls
Length (µm) / Number of Micro-walls / Texture Filling Factor / Extraction (%)950 / 1 / 0 / 67.02
950 / 1 / 0.5 / 77.63
950 / 2 / 0 / 66.01
950 / 2 / 0.5 / 75.63
950 / 4 / 0 / 65.07
950 / 4 / 0.5 / 73.89
Aspect Ratio:
Length(µm)) / Width(µm) / Height(µm) / Texture Filling factor / Aspect ratio / Extraction Efficiency950 / 16 / 2 / 0 / 1/8 / 67.6
950 / 8 / 2 / 0 / 1/4 / 67.88
950 / 4 / 2 / 0 / 1/2 / 68.46
950 / 2 / 2 / 0 / 1 / 67.02
950 / 2 / 4 / 0 / 2 / 65.9
950 / 2 / 8 / 0 / 4 / 62.74
950 / 2 / 16 / 0 / 8 / 58.38
Length (µm) / Width (µm) / Height (µm) / Texture Filling Factor / Aspect Ratio / Extraction Efficiency (%)
950 / 16 / 1 / 0 / 1/16 / 68.03
950 / 8 / 1 / 0 / 1/8 / 68.87
950 / 4 / 1 / 0 / 1/4 / 69.55
950 / 2 / 1 / 0 / 1/2 / 69.83
950 / 1 / 1 / 0 / 1 / 69.05
950 / 1 / 2 / 0 / 2 / 67.48
950 / 1 / 4 / 0 / 4 / 64.48
950 / 1 / 8 / 0 / 8 / 62.42
Conclusions for Aspect Ratio:
• For structures with same aspect ratios, the smaller structure has a better LEE.
• Both structures have a maximum LEE at A=0.5.
• Smaller structure is more sensitive to aspect ratio lesser than 1, while the bigger structure is more sensitive to aspect ratio greater than 1.
· Changing absorption coefficient for ITO:
Conclusions:
• We found that the TIR of the microwall sides result in a strong dependence of LEE on microwall length.
• Shorter walls exhibit higher LEE.
• The LEE is dependent on the ITO absorption coefficient.
• The LEE increases greatly with ITO texturing due to the decreased TIR from sides.
• Increasing the aspect ratio beyond 1 decreases the LEE. The best aspect ratio seems to be 0.5 for coaxial microwall structure.
Future work:
• Simulation of an array of coaxial microwall LEDs.
• Exploring nano scale structures for better LEE using finite-difference time-domain (FDTD) based simulations.
5