Your SlideShare is downloading. ×
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Modeling of Devices with SimuLED
Upcoming SlideShare
Loading in...5
×

Thanks for flagging this SlideShare!

Oops! An error has occurred.

×
Saving this for later? Get the SlideShare app to save on your phone or tablet. Read anywhere, anytime – even offline.
Text the download link to your phone
Standard text messaging rates apply

Modeling of Devices with SimuLED

7,988

Published on

LED and laser diode design and optimization …

LED and laser diode design and optimization
http://www.str-soft.com/products/SimuLED/

Published in: Technology, Business
1 Comment
7 Likes
Statistics
Notes
  • This presentation was published at LinkedIn.com by Alex Galyukov
    http://www.linkedin.com/in/alexgalyukov
    http://www.str-soft.com/products/SimuLED/
       Reply 
    Are you sure you want to  Yes  No
    Your message goes here
No Downloads
Views
Total Views
7,988
On Slideshare
0
From Embeds
0
Number of Embeds
10
Actions
Shares
0
Downloads
0
Comments
1
Likes
7
Embeds 0
No embeds

Report content
Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
No notes for slide

Transcript

  • 1. STR Group, Ltd. engineering tool for LED and laser diode design and optimization October 2008 1
  • 2. GaN total market and important optoelectronic sectors 10 HB LED market ($ billion) OIDA Total GaN market ($ billion) 10 1 1 High-brightness LED market history 0.1 1995 1998 2001 2004 2007 2010 Compilation of Year forecasts from various sources 0.1 Laser diode market ($ billion) 1995 1998 2001 2004 2007 2010 10 Strategies Unlimited Year M. Leszczynski Strategy Analytics 1 up to now, the total GaN market nearly Forecasts from corresponds to the various sources 0.1 market of III-N LEDs 1995 1998 2001 2004 2007 2010 Year 2
  • 3. Main application areas of III-nitride LEDs and LDs Green Blue Violet UV UV 550-520 nm 470-420 nm 405 nm 365-340 nm 280-250 nm Solid-state Water & air Solid-state Curing and lighting with disinfection lighting by drying of inks, phosphors coatings, and RGB mixing Food steri- adhesives Signs lization Signs Medicine Medicine Traffic lights Traffic lights False banknote Chemical Display Entertainment detection back-lighting catalysis HR optical Automotive lithography Entertainment DVD writing New generation of and playing Projection TV DVD systems Optical data (~300 nm) storage 3
  • 4. Challenges in modeling advanced light-emitting devices Complex multi-scale 3D geometry of state-of-the-art LEDs and LDs Coupled current-spreading and heat-transfer problems; very non-linear equations Non-ordinary properties of novel III-nitride and II-oxide materials → Huge simulation time and computer resources demanded ! 4
  • 5. Hybrid approach to modeling LED dice I-V characteristic, 3D ray tracing emission spectrum, series resistance, temperature distribution, external 3D model of light extraction efficiency current spreading efficiency & heat transfer U p −n = Fn − Fp p-n Fn difference neutral neutral junction between the region region Fp region quasi-Fermi levels at the x boundaries of p-n junction current density, IQE, and region z emission spectra versus 1D model of y p-n junction bias carrier transport & light emission 5
  • 6. Hybrid approach to modeling LED dice semitransparent electrode: p-n junction unipolar 2D current spreading region: conductivity distributed non- J = −σ∇U p-pad linear resistor neutral p-region J z = J z (U p −n ) n-pad z ηint = ηint (J z ) neutral y n-region x Calculated with SiLENSe substrate or specified manually 6
  • 7. SimuLED™ – coupled software tools for LED modeling SiLENSe™ – 1D simulator of carrier injection and light emission in III-N and II-O LED structures SpeCLED™ – 3D simulator of current spreading and heat transfer in LED dice 90 120 60 RATRO™ – 3D ray-tracing analyzer of 30 150 light propagation and extraction 180 0 in LED dice 330 210 240 300 270 7
  • 8. SiLENSe SiLENSe – software for development and optimization of LED/laser diode heterostructures 8
  • 9. SiLENSe – 1D drift-diffusion simulation of LED heterostructure Band diagrams Carrier concentrations Electric filed Radiative and non- radiative recombination Internal emission efficiency Carrier fluxes Energy levels in QWs Emission and gain spectra Editable database of materials properties 9
  • 10. SiLENSe – 1D simulator for LED heterostructures specification of individual layer parameters: thickness, doping, and composition layer-by-layer Input data LED structure visualization specification 10
  • 11. SiLENSe – 1D simulator for LED heterostructures editable database of materials properties that are automatically identified from the input layer parameters 11
  • 12. SiLENSe – basic equations Poisson equation account of distributed ( )( ) polarization dPz0 ∇ ε 33∇ϕ = q N A − N D + n − p − ∗ − + doping in graded- dz composition materials Continuity equations ∇ ⋅ J n − qR = 0 ∇ ⋅ J p + qR = 0 , account of Equations for carrier fluxes Fermi statistics fo J n = µ n n ∇Fn , J p = µ pp ∇Fp degenerate carriers 12
  • 13. SiLENSe – basic equations Recombination model R=R +R +R +R SR dis A rad Radiative Shockley-Read recombination of Auger non- non-radiative non-equilibrium radiative recombination at carriers recombination point defects non-radiative recombination at threading IQE = R rad /R dislocations 13
  • 14. SiLENSe – basic equations Quantum-mechanical analysis of emission spectrum complex valence band structure 14
  • 15. SiLENSe – Competitive advantages Advanced physical models: new Polar/nonpolar/semipolar heterostructures Distributed polarization doping in graded-composition AlInGaN alloys Original model of non-radiative recombination at dislocations Original model of the effect of localized new states in InGaN QWs on IQE of LED structure Easy to learn: it requires ~1-2 days to start simulations after installing the package Fast operation: the simulator allows full analysis of ~5-10 heterostructures a day SiLENSe is helpful not only for device engineers but also by people doing epitaxial growth of LED and LD heterostructures 15
  • 16. SiLENSe Application examples 16
  • 17. Blue MQW LED heterostructure n-GaN barrier NDB = 5×1017 – 5×1018 cm-3 UID-In0.13Ga0.87N N Characterization of 12 nm 3 nm (MQW) the structure: structure: S.S. Mamakin et al., p-GaN n-GaN p-Al0.2Ga0.8N Semiconductors 37 NA = 7×1019 cm-3 ND = 2×1018 cm-3 NA = 7×1019 cm-3 (2003) 1107 0.5 µm 4-5 µm 0.1 µm 60 nm Recombination rates Band diagram Carrier concentrations 1021 6 30 10 2 j = 38 A/cm non-radiative Recombination rate (cm s ) Carrier concentration (cm-3) 28 -1 10 1020 5 reconbination -3 26 10 radiative 1019 4 recombination 24 Energy (eV) 10 electrons 18 3 10 22 10 Fn 20 holes 10 1017 2 Fp 18 10 16 1 10 16 10 15 0 10 14 j = 38 A/cm2 10 j = 38 A/cm2 12 14 10 -1 10 600 650 700 750 800 850 900 600 650 700 750 800 850 900 600 650 700 750 800 850 900 Distance (nm) Distance (nm) Distance (nm) 17
  • 18. Factors affecting the internal quantum efficiency Internal emission efficiency 0 10 8 -2 -1 Nd= 10 cm 10 operation External efficiency temperature experiment (b026) -1 10 -2 10 with ηext= 13 % 17 -3 5x10 cm dislocation 18 -3 1x10 cm density 18 -3 3x10 cm -2 10 -3 18 -3 10 5x10 cm Light emission efficiency 1.0 n-GaN -4 -3 -2 -1 0 1 2 3 PL intensity (arb.units) 10 10 10 10 10 10 10 10 17 -3 n0= 1×10 cm 2 Current density (A/cm ) 0.8 {} 0.6 0.4 16 -3 ∆n = 5×10 cm 17 -3 ∆n = 5×10 cm 0.2 structure 18 -3 ∆n = 5×10 cm MQW barrier design 0.0 4 5 6 7 8 9 10 11 doping 10 10 10 10 10 10 10 10 -2 Dislocation density (cm ) 18
  • 19. Effect of barrier doping on operation of MQW LEDs 21 21 10 10 only the QW 17 -3 19 -3 1 1 ND= 5x10 cm ND= 2x10 cm 20 20 adjacent to Concentration (cm ) Concentration (cm ) 10 10 3 3 0 0 AlGaN EBL 19 19 10 10 Energy (eV) Energy (eV) -1 electrons -1 electrons gives a major 18 18 10 10 rise to -2 -2 holes holes 17 17 10 10 recombination -3 -3 16 16 10 10 at a high barrier -4 -4 15 15 doping 10 10 460 480 500 520 540 560 460 480 500 520 540 560 Distance (nm) Distance (nm) 19 -3 28 ND= 2x10 cm 28 10 10 Recombination rate Recombination rate 25 25 10 10 nonradiative 22 22 10 10 19 19 10 10 nonradiative radiative 16 16 10 10 17 -3 ND= 5x10 cm radiative 13 13 10 10 460 480 500 520 540 560 460 480 500 520 540 560 Distance (nm) Distance (nm) 19
  • 20. Experimental identification of most efficient QW A. David et al, Appl. Phys. Lett. 92 Appl. Lett. (2008) 053502 (Phillips Limuleds) Limuleds) On the basis of wavelength-resolved far-field pattern measurements the conclusion was made that only the QW adjacent to p-AlGaN blocking layer emits light effectively 20
  • 21. LED heterostructure with a wide GaN active region 5 1.8 0.8 Internal quantum efficiency 2 2 j = 26 A/cm 200 nm p-GaN j = 20 A/cm 0.7 1.5 4 EL intensity (a.u.) 200 nm p-AlGaN 0.6 1.2 200 nm n-GaN active region 0.5 3 Energy (eV) Fn 0.9 200 nm n-AlGaN 0.4 GaN 2 0.3 active 0.6 region 0.2 1 4-6 µm n-GaN 0.3 Fp 0.1 d 0 0.0 0.0 8 9 10 10 10 10 -2 Dislocation density (cm ) -1 sapphire substrate 400 600 800 1000 1200 Distance (nm) 0.8 3.5 Internal emission efficiency A.S.Usikov et al, Phys.Stat.Solidi 0.7 External efficiency (%) 3.0 8 -2 (c) 0 (2003) 2265 (TDI, Inc.) Nd = 2x10 cm 0.6 2.5 0.5 2.0 0.4 1.5 The LED structure is grown by HVPE 0.3 9 -2 Nd = 10 cm 1.0 that is much cheaper compared to 0.2 0.5 0.1 conventional MOCVD technique 0.0 0.0 0 5 10 15 20 25 30 2 Current density (A/cm ) 21
  • 22. Threading dislocation effect on IQE of deep-UV LEDs 10 -2 -1 10 Nd = 10 cm At Nd < 107-108 cm-2, IQE is no -3 External efficiency 10 Internal efficiency longer limited by non-radiative with ηext= 2 % carrier recombination at -2 10 threading dislocation cores -4 10 -3 2 200 x 200 µm device 10 2 -5 10 1 x 1 mm device 0 10 -2 -1 0 1 2 3 4 10 10 10 10 10 10 10 Internal efficiency 2 Current density (A/cm ) -1 10 A.J.Fisher et al, Appl.Phys.Lett. 84 Appl.Phys.Lett. -2 10 (2004) 3394 8 -2 Nd = 10 cm (Sandia Labs) 9 -2 Nd = 10 cm -3 10 10 -2 Nd = 10 cm Wavelength ~ 290 nm -3 -2 -1 0 1 2 3 4 10 10 10 10 10 10 10 10 Output power up to 1.35 mW 2 Current density (A/cm ) Current 300 mA Forward voltage 9.4 V 22
  • 23. Distributed polarization doping in graded-composition alloys EBL design XAlN graded- nominal: 0.5 composition zero PC constant AlGaN 0.3 composition 0.1 polarization [0001] charge (PC) XAlN GaN dP negative graded-down: ρ=− z 0.5 PC descending dz 0.3 composition 0.1 [0001] Distributed polarization XAlN doping has been proposed, graded-up: positive for the first time, for HEMTs 0.5 ascending PC 0.3 composition D. Jena et al., Phys.Stat.Solidi 0.1 (c) 0 (2003) 2339 [0001] 23
  • 24. Band diagrams of LDs with different EBL designs Partial current density (A/cm ) improvement of band line-up in 2 1 2 j = 5.3 kA/cm the LD structure and suppression 4 10 0 of electron leakage Energy (eV) electron -1 current -2 3 10 Partial current density (A/cm ) hole current 2 1 2 -3 j = 5.3 kA/cm 4 10 0 -4 2 10 T = 300 K Energy (eV) electron -1 graded-down -5 current 1650 1700 1750 1800 -2 3 10 Partial current density (A/cm ) Distance (nm) hole current 2 1 -3 2 j = 5.3 kA/cm nominal 4 10 0 -4 2 10 T = 300 K Energy (eV) M. Kneissl, et al., -1 -5 electron current APL 82 (2003) 2386 1650 1700 1750 1800 -2 3 10 Distance (nm) hole current -3 graded-up minor effect of -4 2 10 composition T = 300 K -5 grading on band 1650 1700 1750 1800 diagram: a lower Distance (nm) barrier to holes 24
  • 25. Use of distributed polarization to improve EBL design 0.5 Internal quantum efficiency constant composition 1.0 graded-up Injection efficiency 0.4 graded-down 0.8 9 -2 Nd= 10 cm 0.3 9 -2 Nd= 10 cm 0.6 T = 300 K threshold T = 300 K threshold 0.2 0.4 constant composition 0.1 0.2 graded-up graded-down 0.0 0.0 0 1 2 3 4 5 0 1 2 3 4 5 10 10 10 10 10 10 10 10 10 10 10 10 2 2 Current density (A/cm ) Current density (A/cm ) Graded-down EBL provides a dramatic suppression of the electron leakage in the LD and a high IQE at room temperature. Graded-up EBL affects weakly the injection efficiency and IQE 25
  • 26. IQE rollover caused by Auger recombination -23 rollover of IQE is predicted 10 InSb Coefficient Ctot (cm s ) , direct for blue SQW LED like that 6 -1 -25 10 CdHgTe indirect fabricated at Nichia GaSb -27 10 InAs 1.2 -29 GaP InGaN InGaAs Internal quantum efficiency 10 GaAs n-Auger p-Auger 1.0 Ge Si InP AlGaAs -31 6H-SiC 10 both n/p 4H-SiC without 4x10-31 0.8 -33 Auger 10 0.6 -35 10 0.0 0.6 1.2 1.8 2.4 3.0 3.6 0.4 Bandgap (eV) 2x10-31 4x10-31 0.2 empirical estimation of 0.0 -3 -1 1 3 5 10 10 10 10 10 the Auger recombination 2 Curent density (A/cm ) coefficient 26
  • 27. IQE rollover caused by Auger recombination External quantum efficiency 0.6 0.6 Internal quantum efficiency 0.5 0.5 400 nm 0.4 0.4 470 nm 0.3 0.3 428 nm pulsed blue SQW LED 460 nm 0.2 0.2 blue MQW LED 530 nm pulsed 0.1 -31 6 -1 0.1 C = 5x10 cm s a τnr = 50 ns b experiment modeling 0.0 0.0 0 1 2 0 1 2 10 10 10 10 10 10 2 2 Current density (A/cm ) Current density (A/cm ) Close correlation between the measures EQE and IQE predicted for SQW and MQW LED structures with account of Auger recombination in the active region 27
  • 28. New structure design accounting for Auger recombination 0.4 Internal quantum efficiency 6 QWs 13 nm 0.3 active 2 QWs layer 0.2 8 -2 Nd = 5x10 cm λ = 430 nm six 3 nm QWs 0.1 13 nm active layer λ = 430 nm 0.0 1 10 100 1000 2 Current density (A/cm ) Measurements reported in: N. F. Gardner, in: Simulation results, accounting for et al., APL 91 (2007) 243506 Auger recombination (Phillips Lumileds) Lumileds) 28
  • 29. Hybrid ZnO/AlGaN LED single heterostructure -1 n-ZnO p-Al0.12Ga0.88N e1 EC -2 e2 7×1017 cm-3 5×1017 cm-3 n= Tunnel p= -3 Energy (eV) 0.8 µm 1.0 µm interface -4 hh1 emission Ya.I.Alivov et al, -5 hh2 Appl.Phys.Lett 83 (2003) 4719 -6 EV -7 I = 1 mA -8 -1 990 995 1000 1005 1010 I = 19 mA T = 315 K Distance (nm) -2 Wavelength (nm) Energy (eV) -3 440 420 400 380 360 Fn -4 experiment Intensity (arb.units) p-AlGaN n-ZnO tunnel -5 emission (350 K) bulk Fp -6 emission -7 975 1000 1025 1050 Distance (nm) 2.7 2.8 2.9 3.0 3.1 3.2 3.3 3.4 3.5 3.6 Type-II band alignment Energy (eV) 29
  • 30. Hybrid II-O/III-N LED double heterostructures 29 4 2 10 10 Recombination rate (cm s ) Partial current density (A/cm ) -3 -1 2 2 2 j = 475 A/cm j = 475 A/cm radiative 1 n-CdZnO @ 400 K @ 400 K 3 EC 10 26 10 n-CdZnO 0 non-radiative Energy (eV) 2 n-MgZnO 10 -1 p-AlGaN n-ZnO n-MgZnO p-GaN 23 10 -2 1 10 p-GaN p-AlGaN -3 20 10 n-ZnO 0 10 EV -4 [0001] 17 -1 -5 10 10 480 510 540 570 600 480 510 540 570 600 Distance (nm) Distance (nm) hybrid II-O/III-N DHS LEDs provide a high IQE even at elevated operation temperatures. Excellent carrier confinement can be obtained in a CdZnO active region. 30
  • 31. Comparison of performances of various ZnO-based LEDs 1 10 Internal quantum efficiency CdZnO DHS 2 j = 30 A/cm @ 400 K 0 10 ZnO DHS @ 400 K -1 10 -2 10 80% ZnO/AlGaN SHS 6% @ 350 K -3 p-i-n ZnO 10 @ 300 K -4 10 6 -2 9 -2 8 -2 8 -2 1x10 cm 1x10 cm 2x10 cm 2x10 cm comparative analysis of various ZnO-based LED structures has shown advantage of double-heterostructures to get a high IQE 31
  • 32. SiLENSe SiLENSe – Laser Edition 32
  • 33. Special options for laser characteristics Combined with the main functionality of the SiLENSe simulator, new options provide analysis and optimization of III-nitride laser diodes • Computation of the waveguide TE- and TM-modes • Advanced approximation of the refractive index dispersion in nitride materials • Birefringence is taken into account • Computation of the optical gain and losses • Computation of the gain spectrum and optical confinement factor for each quantum well • Optical loss because of the free carriers • Laser characteristics • Threshold current density, differential quantum efficiency 33
  • 34. Specific features of III-nitride lasers metallic electrode z [0001] has remarkable L metallic contact effect on the waveguide modes III-nitride LDs suffer from insufficient waveguiding heterostructure caused by low refractive index variation with 0 x AlInGaN composition substrate frequently mode leakage in the y substrate occurs in III-nitride lasers 34
  • 35. 375 nm UV LD on sapphire substrate (waveguiding) 10 3.0 Electric Field Intensity (a.u.) M. Kneissl et al., 9 GaN Glads 2.8 8 2.6 Appl. Phys. Lett. 7 Refractive Index 2.4 TE modes 82 (2003) 2386 6 WG 2.2 m=1 5 2.0 m=8 4 1.8 m=9 3 1.6 2 1.4 1 1.2 0 -1 1.0 -5 -4 -3 -2 -1 0 1 2 µ Thickness (µm) lasing of high-order modes is typical of LDs fabricated on sapphire substrate; a careful optimization of free-carrier losses is therefore required 35
  • 36. 375 nm UV LD on sapphire substrate (oscillation threshold) 0.30 Threshold current density (A/cm ) 2 24 0.25 Injection efficiency L = 1.5 mm 21 0.20 Rout= 0.5 18 Rback= 0.9 0.15 15 0.10 0.05 12 IQE = 71-74% m=7 m=9 m=8 0.00 9 360 365 370 375 380 Oscillation wavelength (nm) 6 2.5 3 Optical confinement factor (%) 2.0 0 360 365 370 375 380 TE-polarization 1.5 Oscillation wavelength (nm) m=7 m=8 1.0 m=9 huge threshold current variation with the 0.5 oscillation wavelength is caused by (i) electron 0.0 leakage in p-region of the LD and (ii) change in 360 365 370 375 380 Oscillation wavelength (nm) the optical confinement factor 36
  • 37. SpeCLED SpeCLED – package for development and optimization of LED dice 37
  • 38. SpeCLED – main options and effects considered 3D coupled computation of the current spreading and heat transfer provides the following information: 3D distributions of the electric potential, current density, and temperature in the whole die 2D distributions of the p-n junction bias, current density, internal quantum efficiency, and temperature in the active region plane I-V characteristic, series resistance, emission spectrum, external quantum efficiency and wall-plug efficiency Internal visualization of the simulation results 38
  • 39. SpeCLED – easy geometry specification Support of both planar and vertical LED dice Import of some CAD-created die configurations ... and get 3D geometry ! Input 2D layout ... 39
  • 40. SpeCLED – automatic grid generation square-LED: unstructured 3D grid with 85 000 cells multi-pixel LED: 3D grid with 1 100 000 cells combining structured and unstructured meshes 40
  • 41. Import of SiLENSe results into SpeCLED 5000 Current density (A/cm ) 2 300 K compatibility of SiLENSe 4000 400 K 500 K and SpeCLED 3000 2000 1000 0 2.6 2.8 3.0 3.2 3.4 3.6 3.8 Active layer bias (V) Internal quantun efficiency 0.5 300 K 0.4 400 K 500 K 0.3 0.2 0.1 manual approximation of 0.0 input data is also possible 2.6 2.8 3.0 3.2 3.4 3.6 3.8 Active layer bias (V) 41
  • 42. Assigning materials properties in SpeCLED anisotropic conductivity can be used to account for a superlattice spreading layer smart setting of the materials properties, including piecewise, spatially non- polynomial, and uniform functional properties approximations may be described via “function” option 42
  • 43. SpeCLED – advantages of using the package Coupled consideration of the current spreading and heat transfer in the dice including conductive new substrates and pads Analysis of very complex LED geometry, including multi-pixel LED arrays Advanced hybrid 3D/1D approach provides fast operation on desktop PC: coupled problem requires ~3-5 h of computing; current spreading problem alone requires ~0.5-3 h 43
  • 44. SpeCLED Application examples 44
  • 45. Current spreading in blue LED dice n-GaN barrier UID-In0.13Ga0.87N NB = 2×1018 cm-3 , 12 nm 2× 3 nm (MQW) [0001] p-Al0.15Ga0.85N p-GaN n-GaN ND = 2×1018 cm-3 NA = 2×1019 cm-3 NA = 1.5×1019 cm-3 × × × 4 µm 200 nm 60 nm two LED die designs Conventional LED Power LED A = 0.0466 mm2 ; A = 1 mm2 ; substrate-down flip-chip mounting mounting S.S. Mamakin, et al., Semiconductors 37 (2003) 1107; W. Goetz et al, presentation at ICNS-4, Denver (2001) ICNS- 45
  • 46. Schematic of a planar LED die design Planar chip design is typical of LEDs semitransparent fabricated on sapphire etched p-electrode substrates mesa p-pad n-pad light is extracted sapp through the hire subs n-contact semitransparent trate layer p-electrode heat sink 46
  • 47. Current crowding and in-plane temperature non-uniformity current crowding occurring at the electrode edge (I = 80 mA) produces a very non-uniform A/cm2 2000 in-plane EL intensity 1000 distribution 500 100 10 n-electrode 320 K 420 pad 390 340 350 a local overheating depends 370 n-electrode on the current density and pad die configuration 47
  • 48. I-V characteristic and output power of blue LED 120 800 Output optical power (mW) ηext= 9% 100 1 700 10 Temperature (K) Current (mA) 80 600 experiment 500 0 60 10 with thermal ak effects pe 400 40 without thermal 300 -1 average 10 20 effects 200 0 0 1 2 3 10 10 10 10 2.5 3.0 3.5 4.0 4.5 5.0 Input electric power (mW) Forward voltage (V) Data: W. Goetz et al., presentation Data: S.S. Mamakin et al., at ICNS-4, Denver (2001) ICNS- Semiconductors 37 (2003) 1107 account of thermal effects is necessary to provide adequate predictions 48
  • 49. Simulation of high-power LED operation Jmax = 1300 A/cm2 , Jave = 72 A/cm2 0 10 ηext= 30% Output power (W) -1 10 -2 10 Experiment Computation -3 10 -2 -1 0 1 10 10 10 10 Input power (W) Data: W. Goetz et al., Tmax= 470 K presentation at ICNS-4, ICNS- Tave= 360 K Denver (2001) 49
  • 50. Schematic of a vertical LED die design Vertical chip design p-electrode on LED structure totally covers is typical of LEDs the surface fabricated on conductive SiC substrates 3 cross-sections SiC te bstra su flip-chip mounting is 2 assumed in simulations for this chip 1 light is extracted n-electrode through the SiC substrate 50
  • 51. Current density distributions in horizontal substrate cross-sections 1 in-plane current density distribution becomes quite uniform near the active region 2 Jmax = 105 A/cm2 Jmin = 103 A/cm2 Jmax = 394 A/cm2 Jmin = 42 A/cm2 3 n-pad Jmax = 114 A/cm2 Jmin = 99 A/cm2 I = 50 mA 51
  • 52. RATRO RATRO – module for optical design and optimization of LED dice 52
  • 53. RATRO – 3D ray-tracing analysis of optical properties of LED dice Prediction of the light extraction efficiency from an LED die Prediction of the far- and near-field emission patterns Analysis of light polarization new Original model of light scattering on surfaces patterned with hexagonal or rectangular pyramids or holes new Original model of light transmission through semitransparent multilayer metallic electrodes new Consideration of non-uniform electroluminescence intensity distribution in the active region plane (if used in combination with SpeCLED) Various die configurations, including shaped substrate Internal visualization of the simulation results 53
  • 54. RATRO – 3D ray-tracing analyzer for optical properties of LED dice 2.0 Relative extraction efficiency Ag 1.8 p-electrode Ni-Au 1.6 p-electrode 1.4 Experiment 1.2 RATRO modeling 1.0 440 460 480 500 520 540 Dominant wavelength (nm) Data: D.A. Steigerwald, et al. IEEE J. on Selected Topic in Quantum Electronics 8 (2002) 54
  • 55. RATRO – patterned surfaces Original model is developed to describe light interaction with surfaces patterned with regular array of hexagonal or rectangular pyramids or holes 55
  • 56. RATRO Application examples 56
  • 57. Output light intensity at planar LED die surfaces semitransparent p-electrode p-pad n-pad LED structure sap waveguiding phir e sub leading to light stra te extraction through the side walls of sapphire heat sink substrate 57
  • 58. Radiation far-field pattern of planar LED die Extraction efficiency 0 through top surface is 8.8% 330 30 correlates well with fitting 300 60 results obtained for Lumileds low-power chip 270 90 bottom emission occurs largely through 240 120 the side walls of sapphire substrate, 210 150 providing two-peak 180 angular dependence 58
  • 59. SimuLED Application of the full package 59
  • 60. Modeling of multi-pixel LED array operation semitransparent p-pad p-electrode n-pad mesa 300×300 µm2 square LED A. Chakraborty et al. Interdigitated multi-pixel array (IMPA) (UCSB), Appl.Phys.Lett 88 containing a hundred of 30×30 µm2 pixels (2006) 181120 60
  • 61. 3D-visualization of the IMPA die geometry n-bus n-contact layer placed in the mesa semitransparent p-electrodes p-contact p-bus is not layer included 61
  • 62. Operation of LED heterostructure 1 3 1 3 2 10 2 10 j = 415 A/cm Current density (A/cm ) j = 160 A/cm Current density (A/cm ) 2 2 0 0 Energy (eV) Energy (eV) 2 2 -1 -1 10 electrons 10 electrons p-AlGaN n-GaN n-GaN p-GaN p-AlGaN p-GaN -2 -2 1 1 -3 10 -3 10 -4 -4 InGaN/GaN MQW InGaN/GaN MQW holes 0 holes 0 10 10 -5 -5 650 700 750 800 850 900 950 650 700 750 800 850 900 950 Distance (nm) Distance (nm) electron leakage ~16% electron leakage ~7% Holes are mainly injected in the quantum well adjacent to the p-AlGaN blocking layer. As a result, more than ~95% of all photons are emitted just from this well. The other wells operate under non-optimal conditions. 62
  • 63. IQE as a function of current density and temperature Auger Internal quantum efficiency recombination IQE 0 10 and electron 0.8000 leakage are the main factors 0.6000 -1 controlling IQE 10 at high current 0.4000 densities 0.2000 -2 10 0.05000 300 400 obtained by Te 500 4 10 0.01000 m 2 10 2 600 1D modeling pe 0 -2 10 cm ) 700 ra 10 y (A/ -4 of LED -6 10 nsit tu 800 10 nt de re re structure Cu r (K ) 63
  • 64. Current crowding and active region overheating in square LED I = 50 mA Lsp~ 80 µm (A/cm2) j (A/cm2) j ∆T = 12 K I = 500 mA Lsp~ 70 µm T(K) T(K) ∆T = 350 K 64
  • 65. Current crowding and active region overheating in IMPA LED Extremely high uniformity I = 1000 mA is predicted for the Lpix= 30 µm current density (~1%) and ∆T = 45 K temperature (~6%) j (A/cm2) Much lower active region overheating of IMPA LED is due to a 1000 larger area which Maximum temperature (K) actually emits light 900 800 square 700 600 T(K) 500 IMPA 400 300 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 Forward current (A) 65
  • 66. Current –voltage characteristics 2.1 , experiment Series resistance ( ) 1.8 , theory 1.5 Current (A) Theory Experiment 1.2 IMPA Square 7.2-9.5 8 0.9 IMPA 1.2 1 0.6 square 0.3 0.0 Excellent agreement 0 1 2 3 4 5 6 7 8 between the predicted and Forward bias (V) measured series resistance Discrepancy between the theoretical and R square ≈ (L sp / σ d c p) −1 = 7 measured turn-on voltage is attributed to non-ohmic behavior of p-contact. 66
  • 67. Ray-tracing simulation of light extraction from the dice 20 million rays is Top – 3% used to generate a cross section Bottom – 7% smooth far-field Side walls – 5% top view radiation pattern 0 330 30 300 60 270 90 bottom view 240 120 210 150 180 far-field emission pattern 67
  • 68. Light extraction from square LED die Non-uniform Wave-guiding resulted in distribution of the strong light extraction optical power over through the side walls of the die surfaces sapphire substrate IF = 100 mA bottom view top view 68
  • 69. Light extraction from IMPA LED die Very uniform experiment Weak variation of distribution of the emission intensity emission power over the back among the pixels sapphire substrate IF = 1000 mA 10 W/cm2 bottom view top view 69
  • 70. Output optical power as a function of current 40 700 η back= 6% 120 experiment 35 Output power (mW) theory Output power (mW) IMPA Temperature (K) 30 100 600 25 80 IMPA 20 500 60 square 15 active 40 10 layer 400 , experiment 20 5 , theory η back= 6% 0 0 300 0.0 0.2 0.4 0.6 0.8 0 1 2 3 4 Forward current (A) Forward current (A) deviation of the theoretical curve from experimental points may be caused by insufficiently accurate approximation of temperature- dependent materials parameters 70
  • 71. Main advantages of using SimuLED™ package Advanced physical models Adopted to use by device engineers and PhD students: it is easy to learn and to use Optimized by efficiency: operates very fast, as compared to other packages, and requires minimum computer resources (can be run in personal computers) Is rapidly developing to include new important physical mechanisms 71
  • 72. Other device-related simulators offered by STR BESST™ – 1D simulator for short-period superlattice (SPSL) properties and LEDs consisting of the SPSL regions SELES™ – simulator of In surface segre- gation in III-nitride heterostructures with prediction of their optical properties FETIS™ – simulator of III-nitride FETs and HEMTs 72
  • 73. Contact information Consulting service & software support: str-contact@str-soft.com Information on commercial software www.str-soft.com Detailed info is available on request: • Demo version • Physical summary • Code description • GUI manual • Trial license 73
  • 74. Our customers STR currently provides software and consulting services to over 40 companies and Academic Institutions in USA, Europe, and Asia. We are grateful to those of our SimuLED™ customers who permitted us to refer their names: • Technische Universitat Berlin, Institut fur Festkorperphysik, Germany • Fraunhofer Institut Angewandte Festkörperphysik, Freiburg, Germany • Tokyo Institute of Technology, Japan • Industrial Technology Research Institute of Taiwan, Taiwan • School of Electrical and Computer Engineering, Georgia Institute of Technology, GA, USA • Electrical & Engineering Department, University of Delaware, DE, USA • College of Optics and Photonics, University of Central Florida, FL, USA • Institute of High Pressure Physics), Polish Academy of Sciences, Warsaw, Poland • Department of Materials Science and Engineering, Meijo University, Nagoya, Japan • Department of Electrical Engineering, The National Central University, Jhongli, Taiwan • Advanced Optoelectronic Devices Laboratory, National Taiwan University • Department of Applied Mathematics and Physics, State University, Vladimir, Russia • Interdisplinary Graduate School of Science and Engineering, Tokyo Institute of Technology • Semiconductor Device Laboratory, Yamaguchi University, Ube, Japan • Korea Polytechnic University, Siheung City, Korea 74 • Gwangju Institute of Science and Technology, Gwangju, Korea
  • 75. Testimonials I'm happy to be reference for the SiLENSe program. We have been very happy with the ease of the user interface, and the underlying physical models are comprehensive. We have been particularly impressed by the ability to include parameters for materials other than for the GaN system into the database, which has helped our development work on ZnO-based LEDs. We look forward to a long association with this excellent software package. Dr. Steve Pearton, Department of Materials Science and Engineering, University of Florida, USA I think SiLENSe is very useful not only as an educational tool for students, but also as a designing tool to optimize blue-LED structures. We are planning to apply this very effective tool to optimize UV LED structures. Prof. Hiroshi Amano, Department of Materials Science and Engineering, Meijo University, Nagoya, Japan We are using it and we like it. Prof. Stanislaw Krukowski, Institute of High Pressure Physics, Polish Academy of Sciences, Warsaw, Poland 75
  • 76. Testimonials Your code is useful and the simulations have good agreement with experimental results. Last year the results of simulations have been reported on two conferences. Sergey Nikishin, Associate Professor Texas Tech University, Electrical & Computer Engineering, Lubbock, TX, USA SiLENSe (Simulator of Light Emitters based on Nitride Semiconductor) is an excellent software tool for user, especially for the beginner of the simulation and modeling. It supports the user-friendly interface that even if a person who has no idea about programming can run this simulator. This software based on the simulation of nitride (especially GaN) light emitting diode (LED). But it can also be applied to other materials system such as ZnO-based LED. Sang Youn Han, Department of Materials Science and Engineering, University of Florida, USA 76

×