Light-Emitting Diodes for Solid-State Lighting and Beyond

E. Fred Schubert and Jong Kyu Kim

The Future Chips Constellation
Department of Electrical, Computer, and Systems Engineering
Department of Physics, Applied Physics, and Astronomy
Rensselaer Polytechnic Institute
Troy NY 12180

1 of 40

Let’s keep the lights on – and strongly reduce power

It is feasible to reduce energy consumption for lighting by 50%

– and keep the lights on!

2 of 40

1
 Traditional applications for LEDs and OLEDs

Texas Inst. Pulsar AT&T

Kodak

3 of 40

Recent applications

Japan
Taiwan

Japan

China

Germany

Germany

United States Taiwan China

4 of 40

2
 Solid-state lighting

ƒ Inorganic devices:
• Semiconductor plus phosphor illumination devices
• All-semiconductor-based illumination devices
ƒ Organic devices:
• Remarkable successes in low-power devices
(Active matrix OLED monitors, thin-film transistors, etc.) Comp. Semiconductors, 2006

• Substantial effort is underway to demonstrate high-power devices

• Anticipated manufacturing cost and luminance of organic devices are
orders of magnitude different from inorganic devices

Predicted growth of LED market

5 of 40

Lighting sources one has never seen before

ƒ Light-emitting large-area panels, wallpaper, and even curtains

ƒ OLEDs are light sources do not blind

Siemens, 2005

adopted from Edward Hopper

OLED-panel lighting in movie theater Siemens, 2005

6 of 40

3
 OLED versus LED

Opto Tech Corp. Osram Corp.

OLEDs are area sources LEDs are point sources

They do do not blind They are blindingly bright
Suitable for large-area sources Suitable for imaging-optics applications

ƒ Luminance of OLEDs: 102 – 104 cd/m2

ƒ Luminance of LEDs: 106 – 107 cd/m2
ƒ Luminance of OLEDs is about 4 orders of magnitude lower
ƒ OLED manufacturing cost per unit area must be 104 × lower

OLEDs LEDs
Low-cost reel-to-reel manufacturing Expensive epitaxial growth
7 of 40

Energy conservation – A singular opportunity

Nobel Laureate Richard Smalley: “Energy is the single most

important problem facing humanity today” and “conservation
efforts will help the worldwide energy situation”.
Testimony to US Senate Committee on Energy and Natural Resources, April 27, 2004 1943 – 2005

ƒ Solid-state lighting is singular opportunity for conservation of energy

Multiple light-emitting diodes LED with wavelength (λ) converter

8 of 40

4
 Quantification of solid-state lighting benefits

ƒ Energy benefits
• 22 % of electricity used for lighting
• LED-based lighting can be 20 × more efficient than
incandescent and 5 × more efficient than fluorescent lighting
ƒ Financial benefits
• Electrical energy cost reduction, but also savings resulting from
less pollution, global warming
ƒ Environmental and economic benefits
• Reduction of CO2 emissions, a global warming gas
• Reduction of SO2 emissions, acid rain
• Reduction of Hg emissions by coal-burning power plants
• Reduction of hazardous Hg in homes

Hg
Cause: CO2 Cause: SO2 CO2 ,SO2, NOx, Hg, U Cause: Waste heat
and acid rain

Antarctica Czech Republic Switzerland United States 9 of 40

Quantification of benefits
Global benefits enabled by solid-state lighting technology over period of
10 years. First numeric value in each box represents annual US value.
US uses about ¼ of world’s energy.
Savings under Savings under
“5.5% scenario” “11% scenario”

Reduction in total energy 43.01 × 1018 J × 5.5% × 4 × 10 = 43.01 × 1018 J × 11% × 4 × 10 =

consumption = 94.62 × 1018 J = 189.2 × 1018 J

Reduction in electrical energy 228.9 TWh × 4 × 10 = 457.8 TWh × 4 × 10 =

consumption = 9,156 TWh = 32.96 × 1018 J = 18,310 TWh = 65.92 × 1018 J

Financial savings 22.89 × 109 $ × 4 × 10 = 45.78 × 109 $ × 4 × 10 =

= 915.6 × 109 $ = 1,831 × 109 $

Reduction in CO2 emission 133.5 Mt × 4 × 10 = 5.340 Gt 267.0 Mt × 4 × 10 = 10.68 Gt

Reduction of crude-oil 12.03 × 106 barrels × 4 × 10 = 24.07 × 106 barrels × 4 × 10 =

consumption (1 barrel = 159 liters) = 481.2 × 106 barrels = 962.4 × 106 barrels

Number of power plants not 35 × 4 = 140 70 × 4 = 280

needed

(*) 1.0 PWh = 1000 TWh = 11.05 PBtu = 11.05 quadrillion Btu “=” 0.1731 Pg of C = 173.1 Mtons of C
1 kg of C “=” [(12 amu + 2 × 16 amu) / 12 amu] kg of CO2 = 3.667 kg of CO2 Schubert et al., Reports on
Quantitative data based on Schubert et al., Reports on Progress in Physics, invited review, to be published (2006)
see also R. Haitz et al. Adv. in Solid State Physics, Physics Today 2001; see also US DOE (2006)
Progress in Physics (2006)
10 of 40

5
 Fundamental innovations

ƒ Innovation in materials
• Omnidirectional reflectors
• New materials with unprecedented low refractive index
• New materials with very high refractive index

ƒ Innovation in devices
• White LEDs with remote phosphors
• Solid-state lighting

ƒ Innovation in systems
• Figures of merit
• Future smart lighting systems

11 of 40

Light-emitting diodes with reflectors

To avoid optical losses, ideal device structures possess either:

Perfect Transparency or Perfect Reflectivity

Example of reflective structure: (after Osram Corporation)

Example of transparent structure (after Lumileds Corporation)

12 of 40

6
 Omni-directional reflector (ODR)
Search for the “perfect reflector”: R = 100% for all Θ, all λ, and TE and TM

High reflectivity is important!

• Omni-directional reflection characteristics

• High reflectivity (> 99 %)
• Electrical conductivity
• Broad spectral width

13 of 40

AlGaInP and GaInN LEDs with ODR

AlGaInP LED GaInN LED

λ = 650 nm, MQW active region λ = 460 nm, MQW active region
AlGaAs window layer Sapphire substrate
GaAs substrate removed, Si submount

14 of 40

7
Employment of specular and diffuse reflectors in GaInN LEDs
ƒ Specular versus diffuse reflectors

Specular reflector Diffuse reflector

ƒ Waveguided modes are trapped modes in specular reflector

ƒ Light escape enabled by roughening reflector surface: diffuse reflector

AFM: Diffuse reflector roughened by dry Reflectivity measurement

etching with polystyrene nano mask

(a) Polystyrene 400 nm (b) Polystyrene 700 nm

15 of 40

Figure of merit for DBR: Index contrast ∆n

nh − n1 ∆n
ƒ Fresnel reflectance of interface r = =
nh + n1 nh + n1
2
2 ⎡ 1 − (n n )2m ⎤
ƒ DBR reflectance RDBR = rDBR = ⎢ l h
⎥
⎢⎣ 1 + (n l nh )
2m
⎥⎦
2 λ Bragg ∆n
ƒ Spectral width of stop band ∆λ stop =
neff

L1 + L2 L + L2 n1 + n2
ƒ Penetration depth Lpen ≈ = 1
4r 4 ∆n

n1 2 2 ∆n
ƒ Critical angle (max. angle for θc ≈
high reflectivity) n0 n0 n1 + n2

Î By increasing index contrast ∆n, figures of merit improve

Î New materials are required
16 of 40

8
 From total internal reflection … to waveguiding

17 of 40

New class of materials: Low-n materials

ƒ Dense materials n ≈ 1.4: SiO2 (n = 1.45); MgF2 (n = 1.39)

ƒ Low-n: refractive index n < 1.25
ƒ Xerogels (porous SiO2)
• Gill, Plawsky, et al. 2001, 2005
ƒ Oblique-angle evaporation
• Xi et al., 2005, 2006
• Technique was developed in the 1950s
ƒ Both techniques suitable for low-loss LEDs

Low-n xerogels, after Gill and Plawsky, 2005 Low-n SiO2

18 of 40

9
 New class of materials: Low-n materials

ƒ Pore sizes << λ (Rayleigh scattering)

ƒ Pore sizes 2 – 8 nm achieved
ƒ Maxwell’s equations: n2 = εr (= k)
ƒ Low-k material in Si technology (field dielectric)
ƒ Low-n films are new class of materials with distinct properties

Thin-film interference Optical micrograph Atomic force micrograph

n = 1.08 Î world record!

19 of 40

Low-index layer and reflector data

ƒ Reflector has 100 × lower mirror losses than metal reflectors

ƒ Reflector has > 100 × lower mirror losses than DBRs
ƒ Suitable for low-loss LEDs
20 of 40

10
Innovation in high-refractive index encapsulation materials
ƒ Fundamental problem of light extraction
• Index mismatch between semiconductor and surrounding air
• Fresnel reflection and total internal reflection

ƒ Encapsulation materials with high

refractive index would solve light-
extraction problem
ƒ Titania nanoparticles in
• Silicone
• Epoxy
• PMMA
ƒ Titania, TiO2, n = 2.68
ƒ Polymer: n ≈ 1.6
ƒ Î Mixture n > 2.0

Lester et al. US Patent 5,777,433 (1998)

Layered approach reduces scattering 21 of 40

High-refractive index encapsulation materials

Without surfactant With surfactant

AFM of TiO2 nano-particles in epoxy

ƒ Optical scattering in film with

poissonian distribution of nano-
particles?
ƒ Ordered distributions feasible?

Shiang and Duggal J. Appl.

Phys. 95, 2880 (2004)

22 of 40

11
 Solid-state lighting technologies and figures of merit

Lighting technologies

Figures of merit of new lighting technologies

Quantity Desirable value
ƒ Luminous source efficiency 150 lm/W
ƒ Luminous flux 1000 lm
ƒ Color rendition capability (CRI) 75 or greater
ƒ Color temperature 2500 – 6000 K
ƒ Lifetime 100 000 hrs
ƒ Cost < $ 10 per lamp
23 of 40

White solid-state lighting sources

ƒ Different technical approaches

• Blue LED plus yellow phosphor
• UV LED plus RGB phosphor
• Phosphor has excellent color stability
• Multiple LEDs
• Which one is best?
ƒ Efficiencies
• Incandescent light bulb: 17 lm/W
• Monochromatic green: 680 lm/W
• Di-chromatic white source: 420 lm/W
• Trichromatic white source: 300 lm/W with
excellent color rendering (CRI > 90)
• Demonstrated with SSL sources: 100 lm/W

ƒ What is the optimum spatial distribution of phosphors?

• Proximate and remote distributions

24 of 40

12
 Innovation in white LEDs – Phosphor distribution

Proximate distribution
(after Goetz et al., 2003)

Proximate distribution
(after Goetz et al., 2003)

Remote phosphor distributions reduce absorption

of phosphorescence by semiconductor chip

Kim et al., Jpn. J. Appl. Phys. Exp. Lett. 44, L649 (2005)

Luo et al., Appl. Phys. Lett. 86, 243505 (2005)

Remote distribution
Narendran et al., Phys. Stat. Sol. (a) 202, R60 (2005)
(after Kim et al., 2005)
25 of 40

Ray tracing simulations

ƒ Ray tracing simulations prove

improvement of phosphorescence
efficiency for
• Remote phosphor
• Diffusive reflector cup

26 of 40

13
 Experimental results

ƒ Improvement of phosphorescence efficiency:

• 15.4 % for blue-pumped yellow phosphor
• 27.0 % for UV pumped blue phosphor

27 of 40

Novel loss mechanisms in white lamps with remote phosphor

Lord Rayleigh
(1842–1919)
ƒ Whispering Gallery Mode
ƒ Trapped optical mode
ƒ Solution:
• Non-deterministic element that breaks symmetry
• Suppression of trapped whispering-gallery modes
28 of 40

14
Remote phosphors with diffuse and specular reflector cups

Specular reflector cup Diffuse reflector cup

ƒ Reflectance versus angle

ƒ Surface texture by bead
blasting
ƒ Diffuse reflectance increased
by two orders of magnitude

29 of 40

Color rendition

ƒ A light source has color rendering capability

ƒ This is the capability to render the true colors of an object

Example: False color rendering

ƒ What is the color of a yellow banana when illuminated with a red
LED?
ƒ What is the color of a green banana when illuminated with a yellow
LED?

Î Î

30 of 40

15
 Examples of different color renditions

Solid-state lighting is
crosscutting technology
that will enable brilliant
displays with the most
vivid colors ever seen

High CRI Low CRI

illumination source illumination source
Franz Marc “Blue Horse” (1911)

CRI = 90 CRI = 62
Siemens, EU Olla project (2005) 31 of 40

Color Temperature

As temperature
increases, hot objects
sequentially glow in the
red, orange, yellow,
white, and bluish white

red, 1000 K ≈ 730 °C

Example: Red-hot
horseshoe orange, 1300 K

bluish white, 10 000 K

white, 6000 K yellow, 2100 K

ƒ Hot physical objects exhibit heat glow (incandescence) and a color

ƒ Planckian radiator = Black, physical object with temperature T
ƒ Color temperature = Temperature of planckian radiator with same
location in chromaticity diagram 32 of 40

16
 Demonstration of trichromatic source

ƒ Color rendering index (CRI) depends strongly on alloy broadening

ƒ 64 lm/W demonstrated at this time (CRI = 84) for trichromatic sources
ƒ For some applications CRI is irrelevant
• For such applications, 680 lm/W would be possible with perfect SSL device
33 of 40

Luminous efficacy and CRI for tri-chromatic source

34 of 40

17
 The Future: Smart Sources
Smart light sources can be controlled and tuned to adapt to different
requirements and environments

35 of 40

The Future: Smart Sources

Smart light sources will enable a wealth benefits and new functionalities
ƒ Example: Communicating automotive lights and room lights

ƒ Example: Circadian lights

36 of 40

18
 The Future: Smart Sources

ƒ Agriculture

ƒ Bio-imaging

ƒ Displays

NASA 37 of 40

Future transportation systems

ƒ … seat belts … air bags … anti-lock brake … electronic stability control …

and

… communicative traffic lights …

38 of 40

19
 Conclusions

ƒ Solid-state lighting is revolutionary technology that enables

• Huge energy savings
• Less global-warming gas and acid-rain-causing gas emissions
• Reduced dependency on foreign oil

ƒ Fundamental innovation required to satisfy needs. Examples:

• New low-n materials n = 1 .08
• Omnidirectional reflectors with 100 × lower mirror losses than metal reflectors
• High-refractive index encapsulants would be very beneficial
• Remote phosphor distributions with higher luminous performance

ƒ Novel applications enabled by “smart” light sources

• Circadian lighting systems
• Communications systems
• Intelligent transportation systems

39 of 40

Acknowledgements

ƒ Dr. Jong Kyu Kim, Profs. Shawn Lin, Christian Wetzel, Joel Plawsky, William Gill,
Partha Dutta, Richard Siegel, and Thomas Gessmann, Dr. Alex Tran (RPI), Drs.
Jaehee Cho, Cheolsoo Sone, Yongjo Park, (Samsung SAIT) Drs. Art Fischer,
Andy Allerman, Dan Koleske, and Mary Crawford (Sandia) Students: Sameer
Chhajed, Charles Li, Pak Leung, Hong Luo, Frank Mont, Alyssa Pasquale,
Chinten Shah, Jay Shah, JQ Xi, Yangang Andrew Xi (RPI)

ƒ Non-commercial agencies:
• National Science Foundation
• Army Research Office
• Department of Energy
• Sandia National Laboratories
• New York State, NYSTAR

ƒ Companies:
• Crystal IS
• Samsung Advanced Institute of Technology
• Applied Materials
• Troy Research Corporation

40 of 40

20