High Performance Green LEDs for Solid State Lighting DOE SSL Workshop January 29, 2019 PIs Shuji Nakamura James S. Speck Steve DenBaars Claude Weisbuch Materials Department University of California Santa Barbara, CA 93106 Core team Cheyenne Lynsky Ryan White Guillaume Lheureux Bastien Bonef Abdullah Alhassan Additional support Yuh-Renn Wu (NTU) Prime recipient: UCSB Agreement # DE-EE0008204 SSL Project Manager: Dr. Joel Chaddock 1
30
Embed
High Performance Green LEDs for Solid State Lighting
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
High Performance Green LEDs for Solid State Lighting
DOE SSL WorkshopJanuary 29, 2019
PIsShuji NakamuraJames S. SpeckSteve DenBaarsClaude Weisbuch
Materials DepartmentUniversity of CaliforniaSanta Barbara, CA 93106
Prime recipient: UCSBAgreement # DE-EE0008204SSL Project Manager: Dr. Joel Chaddock
1
Motivation
Fundamental efficiency of RYGB diode direct vs. pc-LED
Goal: WPE (PCE) > 44% to exceed pc-G
DOE 2018 SSL R&D Opportunities
Motivation
WPE = EE x IQE x LEE
EE = Vph/VF (where hν = qVph)
λgreen = 525 – 540 nm … Vph = ~2.25 V
Commercial LEDs c. 2016-2018 … VF (20 or 35 A/cm2) > 3 VEE < 0.75
DOE 2018 SSL R&D Opportunities
Motivation
Estimates for green LED efficiencies …
WPE = EE x IQE x LEE0.27 = 0.68 x 0.47 x 0.85
Three areas of focus:
EE: Identify and engineer all barriers to electron and hole transport
IQE:SRH reduction – material qualityAuger – reduced carrier density in active QW
Engineering Against SRH
Engineering Against SRH
Engineering Against SRH
Engineering against SRH
*High T QW growth via high TMI flow*AlGaN cap at same T as QW*Multistep GaN barrier (higher T, switch to H2 carrier gas)
Assumption: reduced SRH via higher T and high TMI flow
Voltage Reduction
Voltage Reduction
Timeline of VF reduction
Project Kickoff09/01/2017VF = 4.6 V at 20 A/cm2 for 5 QW
Remove EBLY1Q1
Role of QW #Y1Q2
Reduced Al content in capY1Q2
Ohmic p-contactsY1Q4
Role of SLY1Q4
End of Year 110/31/18VF = 3.1 V at 20 A/cm2 for 5 QW
Voltage Reduction
Patterned Sapphire Substrate (PSS)
2.5 μm UID GaN
1.8 nm GaN UID3 nm InGaN QW
5 nm n-GaN
2 nm Al0.10Ga0.90N cap layer
8.3 nm p+-GaN
X QW
27 nm n-GaN
6 nm GaN barrier
1.5 μm n-GaN, [Si] = 4x1018 cm-3
130 nm p-GaN, [Mg] = 5x1019 cm-3
10 nm p+-GaN, [Mg] = 2.5x1020 cm-3
45x SL 2.65 n-In0.04GaN0.96
Achieved low VF green LEDsLow Al content AlGaN capOhmic p-contactsIncreased SL period from 10 to 45p+ layer after last QBReduced GaN QB thickness from 9 to 6 nm
Experimental I-V curve for 1, 3, 5 QW green LEDs
Reduced VF from 4.6 V to 3.1 V for a 5 QW green LED
Voltage Reduction
Patterned Sapphire Substrate (PSS)
2.5 μm UID GaN
1.8 nm GaN UID3 nm InGaN QW
5 nm n-GaN
2 nm Al0.10Ga0.90N cap layer
8.3 nm p+-GaN
X QW
27 nm n-GaN
6 nm GaN barrier
1.5 μm n-GaN, [Si] = 4x1018 cm-3
130 nm p-GaN, [Mg] = 5x1019 cm-3
10 nm p+-GaN, [Mg] = 2.5x1020 cm-3
45x SL 2.65 n-In0.04GaN0.96
Achieved low VF green LEDsLow Al content AlGaN capOhmic p-contactsIncreased SL period from 10 to 45p+ layer after last QBReduced GaN QB thickness from 9 to 6 nm
Experimental I-V curve for 1, 3, 5 QW green LEDs
Reduced VF from 4.6 V to 3.1 V for a 5 QW green LED
Voltage of Green LEDs
*VF >> Vph (best reports ΔV ~0.4 V)*Why?
*Need to identify barriers to carrier transport
Advanced Characterization and Simulations
Numerical Tools
APT In concentration map of an In0.29Ga0.71N QW (courtesy of B.Bonef)
10 nm
Alloy fluctuations play a major role in nitride devices
Need to be taken into account
Major computation issue in semiconductor physics
Requires solving Schrodinger equation for electrons and holes in a random, disordered potential
[0001]
13
𝑯𝑯𝑢𝑢 =ħ2
2𝑚𝑚∆ 𝑢𝑢 + 𝑉𝑉𝑢𝑢 = 1 (2)
𝑯𝑯ψ =ħ2
2𝑚𝑚∆ψ + 𝑉𝑉ψ = 𝐸𝐸ψ (1)
Idea: Replace Schrodinger equation with the landscape equation
1/u acts as an effective confining potential
Replacing ψ by (u 𝜙𝜙) in (1) leads to
−ħ2
2𝑚𝑚1𝑢𝑢2𝑑𝑑𝑑𝑑𝑑𝑑(𝑢𝑢2𝛻𝛻𝜙𝜙) +
1𝑢𝑢 𝜙𝜙 = 𝐸𝐸𝜙𝜙
M. Filoche and S. Mayboroda, PNAS 109, 14761 (2012)D. Arnold et al. PRL 116, 056602 (2016)
1/u describes the localization energies for localized state
Can be used to predict local DOS
Landscape Theory
Poisson-landscape-drift-diffusion solverSelf-consistent algorithmFast convergence Enables 3D simulation of nitride devices
Blue LED simulationsExperimental parameters are usedExcellent agreement with commercial blue LEDs
C.K. Li et al., PRB 95, 144206 (2017)M. Filoche et al., PRB 95, 144204 (2017) 15
Landscape Theory
Poisson-landscape-drift-diffusion solverSelf-consistent algorithmFast convergence Enables 3D simulation of nitride devices
Blue LED simulationsExperimental parameters are usedExcellent agreement with commercial blue LEDs
C.K. Li et al., PRB 95, 144206 (2017)M. Filoche et al., PRB 95, 144204 (2017) 16
Landscape Theory
3D Landscape-Poisson Solver
*100X – 1000X faster than 3D Schrodinger-Poisson*Facilitates 3D simulations including natural alloy disorder
*Marked improvement in device I-V prediction
*Alloy fluctuations (experiment and theory)Percolative paths for carrier transportPockets for locally high carrier density and enhanced Auger