Illumination With Solid State Lighting

CHAPTER 1
INTRODUCTION
LIGHT-EMITTING DIODES (LEDs) have gained broad recognition as the ubiquitous
little lights that tell us our monitors are on, the phone is off the hook or the oven is hot. Recent
advances in AlInGaP Red and AlInGaN Blue and Green semiconductor growth technology
have enabled applications where in several single to several millions of these indicator style
LEDs can be packaged together to be used in full color signs, automotive interior and exterior
signaling applications including traffic signals. These more recent applications differ from the
“the ubiquitous little lights” of a decade ago in that the viewer is often not tens to hundreds of
centimeters from the LED source but may be tens to hundreds of meters away from the LED
source. Still the preponderance of applications require that the viewer look directly at the LED.

In this sense, even the “high brightness” or “high efficiency” LED applications are
dominated by indicator LEDs. This is NOT “Solid State Lighting”. Artificial “lighting” sources
are fluorescent tubes, 60-plus Watt incandescent light bulbs, high intensity discharge lamps etc.
which all share three key characteristics differentiating them on the evolutionary tree as a
species apart from the indicator lamp. First, they are rarely viewed directly. Light from a
lighting source is viewed in reflection off of the illuminated object. Second, the unit of measure
(flux) is the kilolumen (klm) or higher, not the mlm, lm or worse yet the Cd often used for
indicator LED lamps. Finally lighting sources are predominantly white with CIE color
coordinates very near the Planckian, producing good to excellent color rendering. Today there
really is no such thing as a commercial “solid state lamp” for use in illumination.

However, a branch in the evolutionary tree is forming and differences are beginning to
appear in the technologies used for low power LED indicators and the high power LED light
sources that will evolve into lighting sources. In this paper we will trace the common ancestors
for indicator and high power LEDs, look at the markets that are driving advancement of high
power LEDs, address technical challenges in moving toward true solid state lighting sources,
summarize recent advances in power flip chips, including lamp reliability, white LED
technology, and conclude with a look at what the future might hold for Illumination with SSL
Technology.

Solid-state lighting (SSL) is a type of lighting that uses semiconductor light-emitting

diodes (LEDs), organic light-emitting diodes (OLED), or polymer light-emitting diodes

Dept of EEE SICET 1

 Illumination With Solid State Lighting

(PLED) as sources of illumination rather than electrical filaments, plasma (used in arc lamps
such as fluorescent lamps), or gas. Solid state electroluminescence is used in SSL, as opposed
to incandescent bulbs (which use thermal radiation) or fluorescent tubes. Compared to
incandescent lighting, SSL creates visible light with reduced heat generation and less energy
dissipation. Most common "white LEDs” convert blue light from a solid-state device to an
(approximate) white light spectrum using photoluminescence, the same principle used in
conventional fluorescent tubes.

The typically small mass of a solid-state electronic lighting device provides for greater
resistance to shock and vibration compared to brittle glass tubes/bulbs and long, thin filament
wires. They also eliminate filament evaporation, potentially increasing the life span of the
illumination device. Solid-state lighting is often used in traffic lights and is also used in modern
vehicle lights, street and parking lot lights, train marker lights, building exteriors, remote
controls etc. Controlling the light emission of LEDs may be done most effectively by using the
principles of nonimaging optics. Solid-state lighting has made significant advances in industry.

Dept of EEE SICET 2

 Illumination With Solid State Lighting

CHAPTER 2

History and Evolution of Lighting Technologies

Fig 2. Hurricane lamp

The evolution of lighting is nearly 2 million years old. For most of human history, we
relied on the natural light of the sun and stars to work. This was essential when we lived in
agricultural communities but our cities and businesses now run 24/7. Without proper lighting,
our modern way of living would be impossible. Lights use over 58,000 terawatt hours of
electricity a year across the UK. This is the light we use in every business, home, and street in
Britain.

Let’s look at the history of lighting, from fire to LEDs, and see how lighting has
developed. Where lighting came from Manufactured light was one of the first inventions of
mankind. Our ancestors used primitive tools to create fire on demand instead of relying on
lightning to set fire to brush and ancient woodland. This became widespread by 125,000 BC
which allowed humans to occupy their first ever shelter – caves. The evolution of lighting and
shelter go hand in hand. It wasn’t until 4500 BC that humanity would be able to sustain a light
source without managing it. This was when the basic oil lamp was invented which was then
followed by the candle in 3500 BC.Before electricity, lamps were the focus of development by

Dept of EEE SICET 3

 Illumination With Solid State Lighting

inventors. The Argand lamp was invented in 1780 and it became a must have product in 18th
century Paris. This was because the lamp was much more efficient and it generated ten times
the light that a single candle could. In 1792, the first gas lamp was invented by William
Murdoch which revolutionised the use and evolution of lighting outside of the home.

Evolution of Lighting and the workplace Before the 19th century, factory work was
perilous. Early cotton mills relied on water or steam engines and any stray machinery could
harm a worker if they couldn’t properly repair it. It was Philips and Lee’s cotton mill in Salford
that began to change this by employing gas lamps. Similar developments pushed forward the
mining industry. Humphrey Davy’s safety mining lamp allowed coal miners to avoid setting
alight the flammable methane which existed naturally within the mine. Without the new lamp,
mining operations were deadly and many workers died. These early improvements saved
countless lives in dangerous industries and minimised injury even more. For modern offices
and businesses, the story was different. It wasn’t until the invention of electricity that the world
of work changed forever. The largest improvement in the evolution of lighting came from the
invention of alternating current.

The Great Barrington demonstration project showed this. Suddenly, 23 businesses

could be powered from one transformer which led the way to cheaper lighting for entire city
streets. New Developments in the Evolution of Lighting After electricity was made available
to homes and businesses on a city level, traditional methods of illumination slowly lost their
market share. They weren’t a viable option anymore after 1880 when Thomas Edison patented
a bulb that could last for 1500 hours. Slowly over the 20th century, the lightbulb developed to
become more efficient. With the creation of fluorescent bulbs, a single lightbulb could last for
over 10 years of regular use. Although more efficient bulbs were invented, it wasn’t until the
development of LEDs that lighting could be produced easily and effectively for all uses. LEDs,
or light-emitting diodes, are in almost every device now. At their smallest size, they can be
smaller than a single grain of salt.

For modern use, workplaces rely on LEDs to light an entire building. They last for up
to 25 years, making them more cost effective than traditional incandescent bulbs. The first
yellow LED was invented in 1927 and other colours were invented over time. Blue LED lights
weren’t invented until 1994 and an LED replacement for incandescent bulbs didn’t exist until
2011. Now that we have LED filaments, the original bulb was withdrawn from sale in the UK
and it has stopped being used in modern UK homes.

Dept of EEE SICET 4

 Illumination With Solid State Lighting

CHAPTER 3
LIGHTING TECHNOLOGY COMPARISON
1. Street lighting today

2. Comparing LED luminaries with traditional lighting fixtures

3. Street lighting technology comparison

3.1. STREET LIGHTING TODAY

Cyber Security in Electrical Systems Dept. of EEE SICET 12 Today, street lighting
commonly uses high-intensity discharge lamps, often HPS high pressure sodium lamps. Such
lamps provide the greatest amount of photopic illumination for the least consumption of
electricity. However when scotopic/photopic light calculations are used, it can been seen how
inappropriate HPS lamps are for night lighting. White light sources have been shown to double
driver peripheral vision and increase driver brake reaction time at least 25%. When S/P light
calculations are used, HPS lamp performance needs to be reduced by a minimum value of 75%.

A study comparing metal halide and high-pressure sodium lamps showed that at equal
photopic light levels, a street scene illuminated at night by a metal halide lighting system was
reliably seen as brighter and safer than the same scene illuminated by a high pressure sodium
system. New street lighting technologies, such as induction or LED lights, emit a white light
that provides high levels of scotopic lumens allowing street lights with lower wattages and
lower photopic lumens to replace existing street lights. LED lights are far superior to and can
easily replace traditional lamps with far higher light output. Two of the most important reasons
for this are directional light produced by LEDs and scotopic advantage of LED street lights.
Formal specifications around Photopic/Scotopic adjustments for different types of light sources
enables municipalities and street departments to test, implement and benefit from this new
technologies.

3.2. COMPARING LED LUMINARIES WITH TRADITIONAL

LIGHTING FIXTURE

Dept of EEE SICET 5

 Illumination With Solid State Lighting

1) Luminaire/ System efficiency

Source efficiency is based on the amount of light produced by a lamp at room

temperature. It is different from system efficiency that refers to the amount of usable light
delivered to the target area. There are several causes why raw lumens produced by a light
sourceare far higher than the actual light delivered on a surface. Losses due to trapped light –
Trapped light and reflection inefficiency are the first source of lower light output from
traditional lamps. As long as LEDs had not entered the scene all bulbs produced lights in a 360
degree sphere and the comparison was easy. LEDs changed all that with the directional nature
of their light. In a traditional bulb (incandescent, metal halide, HPS etc.) a considerable portion
of the light output is directed upwards. This light must then be reflected down. The efficacy of
the reflector in turn is determined by the quality of finish, operating conditions and ambient
temperature. The quality of the reflector and, therefore, the amount of light reflected degrades
over time. The actual amount of light coming out of a fixture is, thus, considerably lower.

An LED light by contrast ensures better light distribution with several sources of light.
All the light is produced and directed downwards. There are no problems of reflector effciency,
aging of reflector coating and consequent loss in light output. Losses due to Cover and lenses
– Lenses and glass covers are needed to direct light and to protect the bulb and reflectors from
dust and other damage. Both LED light and traditional fixtures suffer from these losses. Losses
due to operating temperature – The lumen output of LEDs is measured at 25 degrees but the
operating junction temperatures of LEDs can be high, resulting in a 10 % decrease in light
output. For the same reason LEDs are an excellent choice for outdoor lighting in cold climatic
conditions as the junction temperature is closer to the optimum operating temperature. Ballast
and LED driver losses – Ballasts reduce efficiency by 20%. LED drivers are more efficient and
reduce performance by 10% or less. Thus, traditional HPS bulbs are as efficient at producing
light as LEDs. When one considers the impact of systemic deficiencies, LEDs stand miles
ahead of the competition.

2) Scotopic superiority

Human eye responds better to shorter wavelengths of 510 and 550 nm. These are
conspicuous by their absence in HPS lights. Each LED photopic Lumen is equivalent to 2.4
Scotopic lumens. When it comes to Scotopic equivalent lumens HPS lamps are no match for
LED lights with richer light spectrum. Several advantages of LED lighting systems are

Dept of EEE SICET 6

 Illumination With Solid State Lighting

immediately visible. LED lights have significantly less glare, more uniform light distribution,
fewer shadows, and improved visibility. The best part is that this improvement is obtained with
fewer lumens and 40 % or more savings in electricity and maintenance costs.

3) Lumen Maintenance

Lumen maintenance is another area that decision makers need to consider. The human
eye can adapt well to up to 30% reduction in light levels. When light levels fall below the 30
% threshold reduction in light levels become evident and vision is compromised. Thus, lights
need to be designed based on average lumens over the useful life of a light.

Fig 3.2a Lumen Maintenance

A comparison of the lumen maintenance data of different types of lights is shown in the
graph above. A summary of the abbreviations, common name, and life till 70% of peak light
output levels is: PSMH / Pulse Start metal halide / 12000 hours CMH / Ceramic Metal Halide
/ 16000 hours HPS / High Power Sodium / 24000 hours LED / Light Emitting Diode / 50000
hours and more. A pulse start metal halide lamp with a peak light output of 10000 lumens will
require that the fixtures be designed for 7000 lumens. This would require compromises in
lamppost height, ground coverage and reduced spacing of light fixtures resulting in increased
costs. A long life LED light on the other hand, with a similar peak output would, need fixture
design for 9000 lumens.

4) Uniformity of light distribution and light hotspots under traditional lights

On the photo you can see the advantage of uniform lumen distribution achieved by LED
outdoor lights. The upward directed light in a traditional luminaire is reflected straight down,

Dept of EEE SICET 7

 Illumination With Solid State Lighting

which causes a ‘hotspot’ to form right below the luminaire while areas further away from the
bulb have poor light intensity.

Uniformity of light distribution and the presence of light hotspots are important factors
to consider when evaluating traditional lights. Traditional lights, such as incandescent bulbs or
fluorescent tubes, often exhibit uneven light distribution and create hotspots. This means that
certain areas may receive more light than others, leading to inconsistencies in illumination. In
contrast, LED lights have been shown to achieve more uniform lumen distribution, which helps
minimize hotspots and ensures a more even spread of light

Fig 3.2b LED Street Light vs HPS Street Lights

5) Color reproduction

Color discrimination is important in terms of safety to identify the colors of

environment and to instill a feeling of safety outdoors. Uniform lighting with LED parking area
lights, Light hot spots under HPS lights can be seen in the image above. The excellent color
reproduction with LED lights can be evaluated by observing the color of the grass. (Public
domain image from– The Office of Energy Efficiency and Renewable Energy

Fig 3.2c Gigavision Lighting

Dept of EEE SICET 8

 Illumination With Solid State Lighting

3.3 STREET LIGHTING TECHNOLOGY COMPARISON

We are living in a new and exciting era of changes regarding outdoor lighting. Major
changes are being experienced in most of the cities as a result of the technological advances
which are being promoted by the lighting industry. Nowadays, those responsible for lighting
in cities looking for a high quality light committed to the new values of the 21st century, such
as energy efficiency or eco- friendly designs. This means that a street light is no longer that
point of light that used to illuminate and had no further interaction than a simple ON/OFF.
Now, it has become a dynamic point of light with the possibility of adapting to the user’s
needs regardless of the date or place where it is installed.

1) Incandescent lamps

Incandescent Lamps are “standard” electric light bulbs that were introduced more than
125 years ago by Thomas Edison. They have the lowest initial cost, good color rendering and
are notoriously inefficient. They typically have short life spans and use significantly more watts
than CFLs and halogen lamps do to produce the same lumens, or light output. Incandescent
technology produces light by heating up a metal filament enclosed within the lamp’s glass.
More than ninety percent of the energy used by an incandescent light bulb escapes as heat, with
less than 10% producing light. Their use is most common in areas prone to frequent theft or
vandalism of light fixtures. In these locations a very high rate of replacement may make a case
for use of these cheap light bulbs. Anywhere else they are too wasteful to make sense. After
all, 5 % efficiency and a few hundred hours lifespan are difficult to consider when replacement
with LED systems use 7 times less energy.

Dept of EEE SICET 9

 Illumination With Solid State Lighting

Fig 3.3a Incandescent Lamp

2) High Intensity Discharge (HID) Lamps include:

• Mercury Vapor lamps (outdated and almost extinct)

Mercury Vapor Lamps were introduced in 1948. It was deemed a major improvement
over the incandescent light bulb, and shone much brighter than incandescent or fluorescent
lights. Initially people disliked them because their bluish-green light. Other disadvantages are
that a significant portion of their light output is ultraviolet, and they “depreciate”; that is, they
get steadily dimmer and dimmer with age while using the same amount of energy.

Mercury lamps developed in the mid 1960s were coated with a special material made
of phosphors inside the bulb to help correct the lack of orange/red light from mercury vapor
lamps (increasing the color rendering index(CRI)). The UV light excites the phosphor,
producing a more “white” light. These are known as “color corrected” lamps. Most go by the
“DX” designation on the lamp and have a white appearance to the bulb. As of 2008, the sale
of new mercury vapor streetlights and ballasts was banned in the United States by the Energy
Policy Act of 2005, although the sale of new bulbs for existing fixtures do continue.

Fig 3.3b Mercury Vapor Lamp

• Metal Halide lamps

In recent years, metal halide lamp (MH) streetlights have illuminated roadways, parking
lots and also warehouses, schools, hospitals and office buildings. Unlike the old mercury lights,

Dept of EEE SICET 10

 Illumination With Solid State Lighting

metal halide casts a true white light. It is not nearly as popular as its sodium counterparts, as it
is newer and less efficient than sodium. MH lamps operate at high temperatures and
pressures, emit UV light and need special fixtures to minimize risk of injury or accidental fire
in the event of a so called ‘non passive failure’ – or when the lamp bursts at the end of the
useful life. A small fire at the Harvard University greenhouse was started by one such lamp
that was not properly contained. These cannot start up at full brightness as the gases in the lamp
take time to heat up.

Additionally, every time the light is switched on a re-strike time of 5 to 10 minutes is needed
before the lamp can be switched on. These lamps are thus not suited to situations when
intelligent control systems are used to switch lights on and off. MH lamps suffer color shift as
they age though this has been improving. Actual life expectancy is about 10,000 to 12,000
hours on average.The mercury and lead content of these lamps is also a serious issue. A single
1500 watt lamp may contain as much as 1000 mg of mercury. High cost and low life hours has
kept them from becoming popular municipal lighting sources even though they have a much
improved CRI around 85.

Fig 3.3c Metal Halide Lamps

• High Pressure Sodium lamps (HPS)

HPS lamps were introduced around 1970 and are one of the more popular street lighting
options, the most efficient light source when compared to mercury vapor and metal halide

Dept of EEE SICET 11

 Illumination With Solid State Lighting

lamps (on a ‘lumen/ watt’ scale). The disadvantage is that they produce narrow spectrum light
mostly a sickly yellow in color. These lights have a very low color rendering Index and do not
reproduce colors faithfully.

These lights do not find favor with police departments as it is difficult to determine the
color of clothes and vehicles of suspects from eye witness accounts in the event of a crime.
Color-corrected sodium vapor lamps exist but are expensive. These “color corrected” HPS
lamps have lower life and are less efficient. There are two types of sodium vapor streetlights:
high-pressure (HPS) and low-pressure (LPS). Of the two, HPS is the more-commonly used
type.

Low Pressure Sodium lights are even more efficient than HPS, but produce only a
single wavelength of yellow light, resulting in a Color Rendering Index of zero, meaning colors
cannot be differentiated. LPS lamp tubes are also significantly longer with a less intense light
output than HPS tubes, so they are suited for low mounting height applications, such as under
bridge decks and inside tunnels, where the limited light control is less of a liability and the
glare of an intense HPS lamp could be objectionable.

Another issue of HPS lights is that they contain 1 to 22 mg of mercury for a 100 watt bulb with
an average of 16 mg per bulb.

They also contain lead. Unsafe disposal of these bulbs can lead to significant exposure
of human beings and wild life to mercury contaminated water and food. Issues with mercury
contamination and customer preference for full spectrum light has been fuelling the
replacement of these lights particularly in areas like self managed residential complexes where
people can directly pay for the quality of light.

3) Fluorescent lamps

The fluorescent lamp first became common in the late 1930s. These lamps are a form
of discharge lamp where a small current causes a gas in the tube to glow. The typical glow is
strong in ultraviolet but weak in visible light.

Dept of EEE SICET 12

 Illumination With Solid State Lighting

However, the glass envelope is coated in a mixture of phosphors that are excited by the
ultraviolet light and emit visible light. Fluorescent lamps are much more efficient than
incandescent lamps, but less efficient than High Pressure Sodium.

The major problems with standard fluorescent lamps for street lighting is that they are
large, and produce a diffused non-directional light. They are also susceptible to low voltage
failure, prone to breakage of glass parts and contain harmful mercury. Therefore the fixtures
needed to be large, and could not be mounted more than 20–30 feet above the pavement if they
were to produce an acceptable light level. Fluorescent lamps quickly fell out of favor for main
street lighting, but remained very popular for parking lot and outside building illumination for
roadside establishments.

Fig 3.3d Fluorescent lamps

4) Compact fluorescent lamp

Compact fluorescent lamps (CFL) have been used more frequently as time has
improved the quality of these lamps. These lamps have been used on municipal walkways and
street lighting though they are still rare at this time. Improvements in reliability still need to be
made. Some issues with them are limited lumen output, high heat build upin the self contained
ballast, low life/burnout due to frequent cycling (on/off) of the lamp, and the problem where
most fluorescent sources become dimmer in cold weather (or fail to start at all).

Dept of EEE SICET 13

 Illumination With Solid State Lighting

They also contain harmful mercury. CFL efficiency is high and CRI is excellent around
85. CFL produces a color temperature around 3000 K with its light being “soft white” around
that color temperature. Higher color temperatures are available.

Compared to general-service incandescent lamps giving the same amount of visible

light, CFLs use one-fifth to one-third the electric power, and last eight to fifteen times longer.
A CFL has a higher purchase price than an incandescent lamp, but can save over five times its
purchase price in electricity costs over the lamp's lifetime. Like all fluorescent lamps, CFLs
contain toxic mercury, which complicates their disposal. In many countries, governments have
banned the disposal of CFLs together with regular garbage. These countries have established
special collection systems for CFLs and other hazardous waste.

Fig 3.3e Compact fluorescent lamp

Dept of EEE SICET 14

 Illumination With Solid State Lighting

CHAPTER 4

APPLICATIONS AND POTENTIAL IMPACT

4.1 APPLICATIONS

SSL uses LEDs and OLEDs to convert electricity into light for illumination.2 Whereas
more than 90% of the power input into a traditional incandescent light bulb is wasted as heat,
SSL has potential energy efficiencies of 50% or greater. In addition, SSL can offer ultra-long
lamp lifetimes, rugged tolerance to mechanical stress, compactness, directionality, and
tunability in its light characteristics (e.g., spectral power distribution, color temperature,
temporal modulation). Applications for high-brightness LEDs and SSL broadly fall into two
categories, lighting for indication and for illumination . Indication is the use of a self luminous
source that is to be viewed directly, such as signage. Illumination is the use of light to make
other things visible by reflection from those objects, for example, the traditional incandescent
light bulb commonly in residential use and fluorescent lighting most often employed in
commercial and industrial settings.

Fig 4.1 Some current and near-term applications using high-brightness LEDs and SolidState
Lighting.

As indicators, high-brightness LEDs have many benefits. This is particularly so for

many monochromatic applications, in which LEDs are replacing extremely inefficient filtered
incandescent light sources. Some initial applications of monochromatic LEDs for indicators
have been (Haitz et al. 1999) traffic lights, safety/emergency lights, outdoor decorative
lighting, automobile tail lights, outdoor displays, and backlighting for widecolor-gamut graphic
displays. For example, LEDs have four distinct advantages over filtered incandescent sources
for traffic signals: higher brightness and directionality, improved daytime visibility; lower
operating costs saving up to $1,000 per intersection per year (derived from Navigant 2003);

Dept of EEE SICET 15

 Illumination With Solid State Lighting

and an expected seven-year lifetime versus one year for incandescent lamps. A red LED traffic
light quickly pays back the initial higher purchase cost (approximately $75 versus $3 for an
incandescent source) with lower energy usage; the seven-year cost of ownership of a red LED
traffic light is approximately one-third the cost of using an incandescent source (Navigant
2003).

As illumination sources, solid-state lights face much greater challenges, with an

accompanying larger payoff. Replacement of all incandescent and fluorescent light sources
with solid-state white lighting could have revolutionary global impact for energy savings and
accompanying environmental benefits, as discussed below. However, current white-light LEDs
are not yet up to the task. Today, white solid-state lights are mainly in specialized, low-flux
applications, such as shelf lighting, stair/aisle lights, accent lights, landscape path lights, and
flashlights.

4.2 IMPACT ON ENERGY CONSUMPTION

Because of its long-term goal of increasing the efficiency of energy consumption in the
U.S., the Department of Energy (DOE) is especially interested in SSL for generalpurpose
illumination. The Department of Energy-Energy Efficiency and Renewable Energy-Solid-
State-Lighting (DOE-EERE SSL) Roadmap (Navigant 2006) goal for white lighting is 50%
electrical-to-visible-optical power-conversion system efficiency by 2025. The present day
electrical-consumption mix for U.S. lighting is 42% incandescent, 41% fluorescent, and 17%
high-intensity discharges (Navigant 2002). This mix yields an aggregate electrical-to-visible-
optical-power-conversion efficiency of less than 15% for lighting.

The total energy consumption in the U.S. is about 10,000 Terawatt hours per year
(TWh/y). As illustrated in Figure 3, 38% of that energy is consumed as electricity, and of that,
22% is consumed by lighting. So, around 8% of the energy consumption is from lighting
(Navigant 2002). Assuming reasonable growth rates, by the year 2025, the U.S. may be
consuming as much as 1,000 TWh/y of electricity for lighting. At a price of $0.07/kWh, that is
about $70B/y. The worldwide trends are similar, with the consumption amounts increased by
factors of three to four.

Dept of EEE SICET 16

 Illumination With Solid State Lighting

If SSL, with a system efficiency of 50%, completely displaced current whitelighting

technologies, the impact would be enormous. The electricity used for lighting would then be
cut by 70% (or 1–15%/50%), and total electrical energy consumption would decrease by more
than 14% (or 22%x70%). Savings to the consumer in electricity cost would be around $50B/y

(or 70%x$70B/y) in 2025, and there would be a savings in electricity production of around 700
TWh/y (or 70% x 1,000TWh/y), or 80 GW average power.

Fig 4.2 Energy Savings Potential of SSL in General Lighting Applications

The savings in energy production by SSL would also be important for its impact on a
cleaner environment. Currently, the U.S. generates approximately 106 tons of carbon-
equivalent emissions to the atmosphere for every 6 TWh of electrical energy consumed. If the
U.S. continues to produce its electricity with roughly the same mix of technologies, the 700
TWh/y savings in energy through SSL would be a reduction of about 115 Megatons of carbon-
equivalent emissions per year.

Although the DOE-EERE SSL Roadmap goal of 50% system energy efficiency is
aggressive and will be challenging, there is no known fundamental physical (e.g.,
thermodynamic) constraint to cap the efficiency even at this value. If near 100% power
conversion efficiency could be achieved, the economic and environmental benefits above
would be even greater.

4.3 IMPACT ON HUMAN VISUAL EXPERIENCE

Dept of EEE SICET 17

 Illumination With Solid State Lighting

SSL could enhance in many fundamental ways the human visual “experience” with lighting,
and thereby enhance human functionality, safety, and comfort. Potential improvements in the
quality of white light include steady color output at all levels of illumination, the ability to
continuously vary output color temperature, simplified and flexible design for mounting
fixtures (including the elimination of lossy luminaries and diffusers), low voltage, and safe
power distribution.

A host of novel applications could also be enabled by SSL, for example (Schubert
2005):

• The ability to create large-area, diffuse lighting or, perhaps, weaving strands of light emitting
material into fabrics using OLED technology open up new architectural possibilities for room
designs and clothing.

• New interior design concepts are possible, such as integration with distributed color/ intensity
sensors for optimization of an entire lighting space

. • Lighting could be “personalized” in real time according to the preferences of an individual

entering a room; for example, tuning of color preferences might be important for an aging
population.

• Lighting could also be tailored to emphasize those features of the environment that are more
desirable to notice at a given time.

• Mood enhancement may be possible by programming interior lights to change room color
temperature throughout the day, corresponding to outdoor circadian cycles.

• Lighting could also be tailored for photopic versus scotopic (i.e., normal lighting vs. low-
lighting) conditions.

Dept of EEE SICET 18

 Illumination With Solid State Lighting

CHAPTER 6

TECHNOLOGY CHARACTERISTICS
The visible light produced from all sources, including solid-state, can be characterized
by various quantitative metrics. This section summarizes those most relevant to SSL .

The Spectral Power Distribution (SPD), P(λ), is the radiant power emitted by a source
at each wavelength or band of wavelengths over the entire spectrum (e.g., in units of W/nm).
The SPD provides a complete description of the spectral properties of a light source.

The color of a light source can be described by the Commission Internationale de

l’Éclairage (CIE) chromaticity coordinates, calculated as follows. Figure 4a displays the three
Color-Matching Functions, x(λ), y(λ) , and z(λ) , which define the “1931 CIE Standard
Observer.” The function y(λ) was defined to be identical to the photopic luminous efficiency
function, V λ , which describes the brightness perceived by the human eye of light of a
particular wavelength under day-brightness conditions (the scotopic luminous efficiency
function, which we do not discuss here, is the perceived brightness under night-brightness
conditions). The other two Color-Matching Functions, x(λ) and z(λ) , do not have such simple
physical interpretations (Fairman, Brill, and Hemmendinger 1997; Brill 1998), but, together
with the first, are necessary to determine the color of light, as described next.

Dept of EEE SICET 19

 Illumination With Solid State Lighting

Figure 6a. CIE 1931 Color Matching Functions.

Figure 6b. CIE 1931 (x,y) Chromaticity Diagram

The tristimulus values X, Y, and Z of a light source are defined as

X = k ∫∫ ∫ P(λ)x(λ)dλ; Y = k P(λ) y(λ)dλ; Z = k P(λ)z(λ)dλ , (1)

where k is a proportionality constant. The Y value will be equal to one of the

photometric quantities (such as luminous flux) of the light source if k=Km (Km=683 lm/W) is
used. The tristimulus values were derived such that X=Y=Z is equal-energy white.

The chromaticity coordinates x,y are calculated as x=X/(X+Y+Z) and y=Y/(X+Y+Z).

Any color of light can be represented on the CIE 1931 (x,y) chromaticity diagram, which is
shown in Figure 4b. The chromaticity diagram is a guitar-pick-shaped 2-D plot of all possible
hue and saturation values of a light source, independent of luminance. The perimeter of the
guitar pick is the locus of chromaticity coordinates for the spectral colors (i.e., monochromatic
light, wavelengths in nm).

The black arc plotted in the interior of the chromaticity diagram gives the chromaticity
coordinates of emission from a blackbody at temperatures between 1000 K and infinity. It is
called the Planckian locus and is a convenient way of representing a white light source.
Positions along the Planckian locus can be specified by the temperature of the blackbody, called
the color temperature, thereby reducing the hundreds of numbers in a SPD, or the two numbers
of chromaticity coordinate, to a single number.

A non-blackbody source is often described by the correlated color temperature (CCT),

which is the temperature of a blackbody radiator that most closely resembles the color from a

Dept of EEE SICET 20

 Illumination With Solid State Lighting

source of equal brightness. Lines of a specified (constant) CCT would cross the Planckian locus
at coordinates of that blackbody temperature, as indicated in Figure 6b. The terms “warm
white” and “cool white” are often used in lighting discussions. “Warm white” is similar to light
from an incandescent bulb and is somewhat yellow or red in appearance, whereas “cool white”
has more of a pure white (or even blue) tone. Interestingly, “warm white” actually has a lower
correlated color temperature (in the range of 2,700 K) than does “cool white” (with a CCT in
the range of 4,100 K). A powerful feature of the chromaticity diagram is that light produced by
mixing output from two sources having different chromaticity coordinates will fall on the line
connecting the coordinates of the sources alone; the position along the line is determined by
the weighted average of the brightness of each source.

Similarly, the chromaticity coordinate of light produced by mixing three light sources
will fall within the triangle formed by the chromaticity coordinates of the three sources as the
vertices. It is important to note that any color falling outside this triangle cannot be produced
by mixing the three sources. The complete subset of colors that can be produced by mixing n
light sources is referred to as the color gamut.

The definition of luminous efficacy depends upon whether one is referring to properties of (1)
radiation (in which it is often more precise to use the term “luminous efficacy of radiation,”
[LER]) or of (2) a source that emits radiation. Both usages are expressed in lm/W. In the first
case, Watts refers to the radiant power content of the radiation; in the second case, it refers to
the amount of electrical power input to the radiation source. The second usage is directly related
to the effectiveness of a device in converting input electrical power into visible light. Unless
stated otherwise, the second usage is meant when discussing SSL and efficacy targets.

Luminous Efficacy of Radiation (LER) is a measure of the amount of radiant power

(radiant flux in W) that can be observed by the eye as luminance (luminous flux in lm). The
LER is determined by the sensitivity of the eye to different wavelengths of light, (i.e., the
photopic luminous efficiency function, ) V (λ ) and the spectral power distribution (SPD, or
P(λ)) of the radiation. For monochromatic light of wavelength λ, the LER is simply K (λ) K V
(λ) rad m = ⋅ , where K m is a conversion factor relating radiant power (W) to luminous flux
(lm) at the single wavelength (555 nm) at which V (λ)=1; this factor is 683 lm/W. Thus, the
maximum possible luminous efficacy of radiation is, for monochromatic light at 555 nm
(green), at the peak of the eye’s responsiveness, 683 lm/W. For radiation with a distribution of

Dept of EEE SICET 21

 Illumination With Solid State Lighting

wavelengths, the LER Krad is found by integrating the product of the radiation’s SPD, P(λ),
and the photopic luminous efficiency function, ) V (λ , over all wavelengths, and normalizing
by the total amount of radiant power:

The Luminous Efficacy of a Source is a measure of the amount of radiant power emitted
by a source that can be observed by the eye as luminous flux divided by the electrical power
that is input into the source (Win), normalized by the same constant K m used above. Thus, the
luminous efficacy of a source, such as an LED or OLED, is given by .

Manipulating the two formulas above, it is easy to see that the luminous efficacy of a
source K is just the LER multiplied by the “wall-plug efficiency” (WPE), where

The Color Rendering Index (CRI) is a metric for color quality. It takes a subjective
measure and attempts to quantify it. The CRI gauges how colors of certain standard reference
objects appear when illuminated by a test light source, compared to the colors observed under
illumination by a reference light source of the same correlated color temperature; if there were
no difference between the colors rendered by the reference and test light sources, the test source
would have a CRI of 100. Light from a standard incandescent light bulb has a CRI near 100.
CRI is calculated from the SPD of the test light source. To evaluate the CRI of real light sources,
the SPD must be measured, although today mathematical models exist to simulate the SPDs of
various white LEDs (Ohno 2004).

Recently, there has been some discussion about the adequacy of the current CRI scale.
One problem is that it is based on an obsolete color space (the 1964 W*U*V* color space),
which is no longer recommended for use. Another problem is that it is based on eight reference
samples of low to moderate color saturation. On the one hand, this causes some Red Green
Blue (RGB) LEDs to have high CRI scores even though they render some poorly saturated
colors. On the other hand, RGB white LEDs tend to enhance color contrast, which is generally

Dept of EEE SICET 22

 Illumination With Solid State Lighting

preferred and can be an important feature of SSL, but is currently penalized by the present CRI
system (Ohno 2004). To solve these problems, new metrics are being proposed (e.g., see Davis
and Ohno 2005). Since color quality is very important for acceptance of SSL products in the
marketplace, these problems need to be studied further with visual experiments. The CIE
Technical Committee 1-62 "Colour Rendering of White LED Light Sources" is investigating
problems with the CRI scale, with a plan for developing a new metric.

Dept of EEE SICET 23

 Illumination With Solid State Lighting

CHAPTER 7

COST OF LIGHT
Illumination with Solid-state lighting relies on the conversion of electricity to visible
white light using solid materials. By taking advantage of direct electricity-to-light conversion
rather than processes ... cost-of-light comparable to, or lower than, that of existing lighting
technology. Where promising demonstrations of higher efficiency exist, they are typically
achieved in small devices

7.1 CAPITAL AND OPERATING COST BREAKOUT

The Cost of Light (COL) is the total cost of ownership for an SSL LED, i.e., the sum of
the purchase, maintenance, and operating costs over its entire lifetime. The commonly accepted
formula for calculating the COL is (IESNA 2000; Navigant 2006):

where CostofLight is the total cost of the light over its operational lifetime (in $/106
lmh); LampLumens is the light output of the lamp (in 1000 lumens); LampCost is the initial
cost of the lamp (in $); LaborCost is the labor cost to replace the lamp (in $); Lifetime is the
useful life of the lamp (in 1000 h); EnergyUse is the power consumption of the lamp (in W);
and EnergyCost is the cost of electricity (in $/kWh). For SSL, the useful lifetime of the lamp
is often measured at 70% lumen maintenance; the labor cost is typically taken to be $1 for a 1
klm lamp, which is used in our analysis.

In the formula above, the first term in the second set of parentheses determines the
Capital Cost, and the second term determines the Operating Cost; the sum of the two
determines the Cost of Light.

The COL for different white-lighting technologies is displayed in Figure 5; cost is

plotted in units of dollars per million lumen hours. The three solid curves represent contours of

Dept of EEE SICET 24

 Illumination With Solid State Lighting

constant COL. In Figure 5, the triangle, square, and diamond symbols represent the highest
market-value incandescent, fluorescent, and high-intensity discharge (HID) lamps,
respectively.

Figure 5. Factors contributing to SSL total cost of light

Of the three traditional lighting technologies, incandescent has the highest COL,
because of low luminous efficacy (high operating cost). As it develops, if SSL follows trends
of traditional lighting, one might expect the COL to fall along, or below, the trend (black line
in Figure 5) that capital cost represents about 1/5 of the operating cost. As the cost of energy
inevitably increases in the future, the ratio of capital cost to operating cost will decrease for all
technologies, including SSL.

The 100% efficient SSL point corresponds to the maximum possible luminous efficacy
of 386 lm/W (assuming a red, green, yellow, blue [RGYB] source with a CRI of 85), an
operating cost of $0.07/kWh/386 lm/W=$0.18/Mlmh, and a capital cost 1/5 this amount. Three
representative SSL white lights are represented in Figure 5. The point labeled “High CRI LED
2006” is a Lumileds LXHL-BW03 “Warm White” emitter. The two points labeled “Low CRI
LED 2006” represent Lumileds K2 series emitters driven at 0.35A and 1A (note that, as can be
seen in Table 1 at the end of this Appendix, these two LEDs have similar COL–driving the
LED “harder” produces more lumens, so the capital cost of the lamp decreases, but the
luminous efficacy decreases, so operating cost increases).4 The point labeled “WOLED” (for
“white OLED”) is taken from the 2002 OLED Roadmap (OIDA 2002), although it should be

Dept of EEE SICET 25

 Illumination With Solid State Lighting

noted that no WOLEDs are being manufactured for lighting, rendering this point very
speculative.

7.2 LUMINOUS EFFICACY AND POWER DELIVERY BREAKOUT

Another way of characterizing white lighting technologies is displayed in Figure 6, in

which the abscissa and ordinate correspond more directly to quantities the technologist can
impact. The first of these quantities is the “power delivery cost.” The power delivery cost is the
capital cost to make a lamp that can be driven at a certain power ($/kWin). It reflects the fact
that the more power one can supply to a lamp, the greater the light output, but at a higher
manufacturing cost (heat sinking the device being one important driver). The second of these
quantities is luminous efficacy (lm/W).

Note that these two quantities, together with the lamp lifetime, are sufficient to
determine the various components of the COL introduced earlier in Equation 5.
EnergyUse/LampLumens is just the reciprocal of luminous efficacy; LampCost/
(LampLumens x Lifetime) is just power delivery cost over luminous efficacy, divided by the
lamp lifetime.

Figure 7.2 Relationship between luminous efficacy and power delivery cost for different
lighting technologies.

The solid curves are contours of fixed COL. The meanings of the symbols are the same
as in Figure 7.1. Although incandescent lights have the lowest power delivery cost, their low
luminous efficacy gives them the highest overall COL. Thus, contours passing through the

Dept of EEE SICET 26

 Illumination With Solid State Lighting

incandescent point represent the highest COL. In proceeding from fluorescent to high-intensity
discharges, there is an increase in complexity (and, thus, power-delivery cost).

However, their luminous efficacies are greater, decreasing the COL. A hypothetical
“trajectory” for improving SSL is perpendicular to the contours of constant Cost of Light,
shown schematically as the dashed white curve. As shown in Figure 5, the 100% efficient SSL
point corresponds to the maximum possible luminous efficacy of 386 lm/W for an RYGB
source with CRI 85.

7.3 LUMINOUS EFFICACY

Luminous efficacy and wall-plug efficiency of LEDs are strong functions of emission
wavelength, as illustrated in the top and middle portions of Figure 7.3 , respectively. For several
of the lamps, the luminous efficacy and the electrical-to-optical power-conversion efficiency
are plotted for output at more than one drive current and voltage.

Figure 7.3a. Top – Luminous efficacy of commercially available LEDs as a function of

wavelength; Middle – External energy conversion efficiency as a function of wavelength; and
Bottom – LED wavelengths that give best luminous efficacy for a specified CRI.

The blue and beige lines correspond to the Lumileds Low CRI and High CRI white
LEDs, respectively. For the nitride-based LEDs (colored symbols in the range 455 to 530 nm),
the data correspond to: 0.35 A, 1.2 W (higher luminous efficacy points), and 1.0 A, 3.7 W.

Dept of EEE SICET 27

 Illumination With Solid State Lighting

Note that for LED devices, wall-plug efficiency drops significantly as emission
approaches the deep green and yellow regions. Efficiency in the nitride materials goes down
because of the difficulty of attaining high indium composition (which shifts emission to longer
wavelength) while still maintaining good material quality. (Generally, achieving high-In
content requires low growth-temperatures, resulting in poor quality material with greatly
increased defect densities.) Phosphide LED emission drops sharply as the Al content x
increases, where the alloy composition is (AlxGa1-x)0.5In0.5P; increasing Al fraction pushes
emission to shorter wavelengths. In this alloy system, the drop in emission with increasing x is
because the X-valley drops in energy relative to the Γ-valley, leading to an increasing number
of electrons involved in indirect transitions, which are not radiative. The band gap crosses over
from direct to indirect near x=0.5, leading to a nearly complete suppression of light emission.

Although the wall-plug efficiency of these visible-light LEDs has a minimum in the
region near 550 nm, the luminous efficacy of this series of LEDs is highest in this range (upper
portion of Figure 7). This is simply because the human eye response to light (also plotted in
the top part of Figure 7) is most sensitive at 555 nm.

High CRI is best achieved by a broad spectral distribution of light covering the entire
visible range. Mixing a small number of different wavelengths of light (e.g., three or four) to
approximate natural white light decreases the CRI , but improves the luminous efficacy by
concentrating the light output to subsets of the spectrum. Thus, there is a fundamental trade-
off between CRI and luminous efficacy for any light source.

For a given CRI, one can calculate the combination of RGB or RYGB wavelengths that
would give the maximum luminous efficacy. These wavelength combinations are given in the
bottom portion of Figure 7.3. This plot was generated using a white LED simulator (Ohno
2004). (The calculations assume that the source full-width at half maximum power is 20 nm
for the red, yellow, and blue sources and 30 nm for the green, and CCT=4000 K.) The black
dotted line in this plot represents the approximate upper limit of CRI that can be obtained using
the three-color (RGB) approach.

The CRI vs. luminous efficacy for current light sources is shown in Figure 8. The two
steep, solid (blue and beige) curves near the right of Figure 8 represent the maximum theoretical
luminous efficacy that could be obtained from either a RGB or RYGB color combination
scheme. For example, for a CRI of 50, the maximum luminous efficacy that can be achieved is

Dept of EEE SICET 28

 Illumination With Solid State Lighting

421 lm/W; for a CRI of 85, the maximum luminous efficacy is 386 lm/W (values at CCT=4000
K).

CRI and luminous efficacy are shown for several different incandescent (triangle
symbols), fluorescent (squares), and high-intensity discharge (diamonds) lamps; the areas of
these symbols are proportional to usage in Tlmh/yr (Teralumen hours per year). Of the three

traditional white lighting sources, incandescent lamps have the highest CRI, but the lowest
luminous efficacy. Fluorescent lamps and high-intensity discharge lamps provide lower CRI,
but much higher luminous efficacy. Also plotted are data from two current commercial LEDs
(the high-CRI and low-CRI devices from Lumileds discussed in previous plots). The dashed
white and blue curves in Figure 8 are hypothetical technology evolution curves that would
eventually achieve both a high CRI and the maximum possible luminous efficacy

Figure 7.3b. Map of Color Rendering Index vs. Luminous Efficacy for different white-lighting
technologies.

7.4 POWER DELIVERY COST

Dept of EEE SICET 29

 Illumination With Solid State Lighting

The power delivery cost can be further broken down into the ratio of the maximum input
power density to the device (Win/cm2 ) and the capital cost of making the LED normalized to
its chip area ($/cm2 ), illustrated in Figure 9. As the figure illustrates, different approaches
can achieve the same ratio.

Figure 7.4 Trade-off between LED lamp cost and maximum power input.
WOLEDs, for example, have very low lamp cost per chip area, but also very low chip
power density. High-power lasers have high lamp cost per chip area, but also high chip power
density. Conventional low-power LEDs have medium lamp cost per unit chip area, and medium
chip power density.
In 2006, the Luxeon K2 (Lumileds) device lies between the low-power LEDs and the
high-power lasers, and one can ask the question whether there is more "room" for the
technology to improve by lowering its lamp cost per chip area or by increasing its chip power
density. Of course, normally these quantities are strongly coupled (i.e., increasing chip power
density usually means increasing lamp cost per chip area). A technology "breakthrough" or
"shift" would be needed that allows one either to lower lamp cost per chip area without
decreasing chip power density, or to increase chip power density without raising lamp cost per
chip area.

Dept of EEE SICET 30

 Illumination With Solid State Lighting

CHAPTER 8
INORGANIC LIGHT-EMITTING DIODES (LEDS)
SSL technologies based on inorganic LEDs start with inorganic direct bandgap
semiconductors that emit in the visible spectrum. There are two inorganic LED-based
architectures in use for white lighting, as discussed below.

8.1 WHITE LIGHT CREATION ARCHITECTURES

The entire spectral range of visible light can be produced by III-V semiconductor LEDs.
AlGaInP LEDs can cover the spectral range from red to amber, and their commercial
production is relatively mature. AlGaInN LEDs are the prime candidates for producing
wavelengths from blue to yellow. Two main approaches for producing white light are being
pursued:

Color Mixing

This approach mixes output from three (or possibly four) different colors of LEDs.
First, it has the potential for higher luminous efficacy, provided that efficient LEDs of the
constituent colors can be developed; there is no Stokes loss from energy down-conversion, as
in the blue + phosphor or photon-recycling approaches. Color rendering index will be highest
by blending output from at least three different wavelengths. However, substantial problems
still exist with this approach. A major limitation is that efficient green and yellow LEDs are
still not currently available. A second problem is that LEDs of different colors tend to age and
degrade at different rates. Over time, the color balance and color rendering quality of a multi-
LED white light may decay significantly; electronic solutions exist, but at a cost. Third,
uniform color mixing can be a problem; color variation in the far-field pattern depends on
architecture. Fourth, even in the blue, overall efficiencies are far from 100%, and they decrease
significantly when driven at high currents.

Wavelength Down-Conversion

By using light from a blue LED to excite fluorescence from a yellow phosphor, the
combination of blue and yellow produces white light. Because of the early development of blue
LEDs and ready availability of yellow phosphors, the first white LED products, for example
flashlights, have used this scheme. The current state-of-the-art uses an InGaN LED emitting at
about 460 nm to excite a cerium-doped yttrium aluminum garnet (YAG) phosphor, which is
ground into a powder and dispersed in an epoxy cap on the device. There is an unavoidable

Dept of EEE SICET 31

 Illumination With Solid State Lighting

Stokes energy-loss in converting a blue photon to a lower-energy yellow photon. The color
quality in this approach can be enhanced, making it closer to that of incandescent lamps, by
additionally adding a red phosphor to improve CRI.

Ultraviolet (UV) LED pumping of RGB phosphors is another possible down conversion
approach. This scheme would potentially be simple to manufacture; it is similar to a television
or fluorescent lamp, with electron excitation replaced by a UV LED. Excellent color rendering
would be possible, but with fundamental limitations on efficiency because of phosphor
conversion efficiency and Stokes loss. In addition, LED packaging material would degrade
more quickly from exposure to high-energy near-UV photons.

8.2 WAVELENGTH CONVERSION MATERIALS

Rare-Earth Phosphors

Inorganic phosphors doped with rare-earth metals are used for a variety of applications.
A wide array of these phosphors has been developed for use with fluorescent lamps. The first
SSL white LEDs have used an yttrium aluminum garnet doped with trivalent cerium
(YAG:Ce+3) to convert output from a blue LED into very broad-band yellow light. The sum
of the two emissions appears white. For efficient white lighting with good color rendering,
quantum efficiencies >85% (for phosphors absorbing in the near-UV) at operating temperatures
>155o C (for phosphors in intimate contact with the LED) will be needed.

A red-emitting phosphor centered near 610 nm with narrow band emission (required
because of the steep drop-off in human eye sensitivity at longer wavelength) and absorption in
the near-UV or blue region has been difficult to achieve. Broader-band emission is acceptable
in the green region because of the eye’s sensitivity to a wider range of wavelengths in this
region. Phosphors based on divalent europium (Eu+2) are available in the green, and even into
the blue emission region

Photon Recycling Semiconductors

This approach is a photon down-conversion scheme similar to using yellow phosphors, in

which an AlGaInP photoluminescent quantum well (QW) or active layer is laminated to a GaN
LED. Some of the blue emission from the GaN chip is absorbed in the phosphide layer, which
emits complementary yellow light. The combined blue and yellow emission produces white
light.

Dept of EEE SICET 32

 Illumination With Solid State Lighting

Semiconductor Nanoparticles

White light can be produced using semiconductor nanoparticles. For example, the band-gap
of CdS nanoparticles can be tuned over the entire visible spectrum by the changing their size
(because of quantum confinement effects) and surface characteristics (e.g., coating the
nanoparticles with ZnS, or changing the nature of chemical groups bonded to the nanoparticle
surface).

8.3 LIGHT CREATION MATERIALS

Different families of inorganic semiconductor materials can contribute to solidstate

white lighting. The primary chemical systems used for LEDs are Group-III Nitrides and Group-
III Phosphides. It is, however, possible that breakthroughs in a different material system, for
example ZnO, will be important.

AlGaInN Materials

LEDs based on gallium nitride (GaN) and ternary alloys with indium (InGaN) and
aluminum (AlGaN), as well as quaternary alloys (AlGaInN) can span the entire visible
spectrum. The current applications for SSL utilize InGaN structures to produce high brightness
blue and green light; longer wavelength light can be efficiently generated by AlGaInP LEDs.
UV light from AlGaN LEDs could also be used to pump RGB phosphors, as mentioned above.
A schematic of a current nitride LED is shown in Figure 8.3.

Figure 8.3a. Schematic of a Nitride LED.

Nitride materials are usually grown by Metal Organic Vapor Phase Epitaxy
(MOVPE)—also referred to as Metal Organic Chemical Vapor Deposition (MOCVD)— from
organometallic sources (e.g., trimethyl-gallium, -indium, or -aluminum) and an excess of

Dept of EEE SICET 33

 Illumination With Solid State Lighting

ammonia. A major difficulty is the lack of low-cost, singlecrystal GaN to use as a growth
substrate. Group-III nitrides are normally grown on poorly matched sapphire (lattice mismatch
+16%, thermal expansion mismatch +39%) or more expensive silicon carbide (lattice mismatch
-3.5%, thermal expansion mismatch -3.2%) substrates. As a result, the films have a great
number (>108 /cm2 ) of dislocations and other structural defects, resulting in defect-mediated
nonradiative recombination of electronhole pairs and reduced mobility because of carrier
scattering from charged defect centers. An intermediate buffer (or nucleation) layer is normally
grown at reduced temperature between the substrate and the n-GaN layer (a GaN nucleation
layer on sapphire substrates, or AlN on SiC). This low temperature buffer reduces defect
densities from up to 1012 to 109 / cm2 . Further defect reduction (by roughly two orders of
magnitude) can be achieved by substrate patterning techniques such as Epitaxial Lateral
Overgrowth, Pendeo Epitaxy, or Cantilever Epitaxy; these approaches rely on spatial
“filtering,” terminating, and/or turning of threading dislocations, so they do not reach the
device active region.

GaN is readily n-doped with Si (usually using a silane source). However, p-type doping
with Mg (usually using the metal organic precursor bis-cyclopentadienyl magnesium, Cp2Mg)
is much more difficult, because of passivation by hydrogen during growth, and the magnitude
of the hole ionization potential associated with Mg. Depassivation of the Mg acceptors is
achieved by thermal annealing or low-energy electron beam irradiation.

Indium incorporation pushes emissions to longer wavelengths; indium fractions greater

than 20% are required for green LEDs. This represents a significant challenge in material
growth. Low temperatures are required for In incorporation because of lower thermal stability,
leading to poorer material. As In composition increases, latticemismatch strain also increases,
leading to a variety of strain-induced defects (e.g., pointdefects, V-defects, and carbon and
oxygen impurities) and lower optical efficiencies.

AlGaInP Materials. The (AlxGa1-x)0.5In0.5P alloys are nearly lattice matched to

GaAs, and production of LEDs emitting from 555 nm (yellow-green) to 650 nm (deep red) is
a relatively mature technology. The availability of single crystal GaAs substrates enables
growth of high-quality phosphide material by MOVPE. But the bandgap of GaAs is 1.42 eV
(870 nm) at room temperature, so this substrate absorbs emitted light below this wavelength,
greatly lowering the LED efficiency. Two solutions to this problem are illustrated in Figure
8.3b One way to prevent absorption of emitted light by the substrate is to insert a reflective

Dept of EEE SICET 34

 Illumination With Solid State Lighting

structure between the LED active region and substrate, illustrated in the left portion of Figure
11. The mirror structure is a Distributed Bragg Reflector (DBR),

Figure 8.3b. Left – AlGaInP LED with DBR mirror for improved light extraction; Right
AlGaInP LED wafer-bonded to transparent GaP substrate.

consisting of many (e.g., 5 to 50) alternating high-refractive index and lower-refractive index
layers. Because of the differences in refractive indices, a portion of the downwardly directed
light is reflected upward (and out of the device) at each layer interface. The mirror stack
thicknesses are adjusted so that all of the reflected waves are in constructive interference. The
DBR is highly effective for light incident normal to the DBR plane; glancing incident light,
however, is only partially reflected by the DBR.

Another approach is to remove the GaAs substrate after the epitaxial layers have been
grown, and then to bond the remaining structure to a transparent GaP substrate. The resulting
structure of a wafer-bonded LED is shown in the right half of Figure 11. Total light extraction
from the wafer-bonded AlGaInP LED can be more than a factor of two greater than the LED +
DBR design.

For low Al fraction, the internal quantum efficiency approaches 100%. (AlxGa1-
x)0.5In0.5P is a direct bandgap semiconductor for x < 0.5; above that composition, it is an
indirect-gap material. The crossover occurs at bandgap energy of about 2.23 eV (555 nm).
Thus, the AlGaInP LED quantum efficiency drops precipitously (Figure 7) at shorter
wavelengths because of the approach of the direct-indirect bandgap crossover. That is, as the
indirect-gap X-band becomes more populated, the radiative lifetime increases, allowing other
nonradiative processes to become more dominant. Efficiency also drops at higher drive currents

Dept of EEE SICET 35

 Illumination With Solid State Lighting

and operating temperatures because of poor carrier confinement in heterostructures as the

direct-indirect-gap crossover is approached.

ZnO Materials

ZnO-based alloys are another possibility for generation of light from the visible to the

near-UV. ZnO has a number of physical properties that make it a good potential candidate for
SSL. However, progress toward making it a practical material is still at an early stage.

The material is a wide-bandgap semiconductor (3.4 eV, comparable to the 3.5 eV of

GaN) with a wurtzite crystal structure. Single-crystal ZnO can now be produced, and
commercial 2 inch wafers are available, offering the possibility of homoepitaxy. The material
can be etched by wet chemical means, making it relatively easy to process. Because ZnO has
a high exciton binding energy (60 meV, compared to less than 30 meV for GaN), higher
operating temperatures are possible.

ZnO has a high intrinsic n-type conductivity, the source of which is not known. It has
been difficult to obtain p-doping; although there has been good recent progress, consistency is
still hard to achieve. Growth of high-quality ZnO films and heterostructures is still being
developed. It may be possible to tune the bandgap of ZnO by

Figure 8.3c. Some LED light extraction schemes

alloying with MgO (7.9 eV bandgap) or CdO (2.3 eV). However, these two oxides have cubic
crystal structures, so it may be difficult to add large fractions to ZnO without introducing
dislocations. Further, the use of the heavy metal Cd in commercial LED structures may not
prove acceptable because of long-term safety and environmental issues.

8.4 LIGHT EXTRACTION APPROACHES

In principle, 100% of the light could be extracted from an LED, but the current state-
of-the-art light extraction efficiency is about 50%. Limitations in light extraction include

Dept of EEE SICET 36

 Illumination With Solid State Lighting

internal reflection at interfaces and light absorption within the device or in the packaging.
Figure 8.3c shows some light-extraction methods.

Encapsulation can reduce the index step between the semiconductor and air, creating a
favorable geometry. High-index encapsulants are desirable. Limitations of this approach
include transparency and degradation of epoxy encapsulants from exposure to high
temperatures and intense radiation. For this reason, silicone encapsulants are replacing epoxy.

Chip shaping has been used to increase light extraction from LEDs. Lumileds used an
inverted pyramid shape to boost light output from a red (AlGaInP) LED to achieve wall-plug
efficiency as high as 50%. Cree has also used chip shaping to increase efficiencies in their
commercial blue and green LEDs.

Surface texturing produces random scattering at the surface, increasing the likelihood
that light will encounter a surface within its escape cone, thus increasing light extraction.

Photonic crystals can be used in multiple ways to increase light extraction. Two
dimensional photonic crystals can be used to scatter waveguided modes out of the active layer
region. Another approach is to use 2D photonic crystals to change the photonic density of states
in the active layer so that no in-plane modes are permitted. This would cause all emitted light
from the quantum wells to be normal to the LED surface, so that it would lie within the escape
cone and not be reflected. A third possibility is to increase the internal quantum efficiency by
enhancing the photonic density of states at the LED emission wavelength. Finally, 3D photonic
crystals could be utilized as highly reflective mirrors for resonant cavity LEDs and laser diodes,
which are described below.

Resonant cavity LEDs or Lasers pumping phosphors have the potential to be efficient
white-light sources. Resonant cavity LEDs and lasers have the highest energy conversion
efficiency of all optoelectronic devices (80% reported for a laser in the infrared). Despite
inherent energy loss because of photon down-conversion from exciting phosphors, the net
luminous efficacy of this scheme could be very high. However, capital costs to manufacture
these complex devices may make them too expensive to be practical.

8.5 THERMAL MANAGEMENT

Junction temperature of an LED affects luminous efficacy, color, and reliability. As high
brightness LEDs are driven harder, managing the heat from the semiconductor will be

Dept of EEE SICET 37

 Illumination With Solid State Lighting

increasingly important. Epoxy encapsulants quickly degrade if the temperature exceeds the
epoxy glass transition temperature. This can lead to device failure modes.

An increase of 75o C can reduce luminous flux in AlGaInP LEDs to one-half the room
temperature value; however, a similar increase in temperature reduces emission intensity of a
470 nm GaInN/GaN LED by only about 5% (Schubert 2003). The dominant emission
wavelength of a phosphide-based LED shifts to longer wavelength with increasing junction
temperature, about one nanometer for every 10o C (Lumileds 2002); the shift in emission
wavelength of a 400 nm nitride LED is about a factor of 3 smaller (Cho et al. 2005). (To put
these wavelength shifts into perspective, the human eye can detect wavelength shifts as small
as 2 nm to 4 nm, depending upon the color; this is the wavelength discrimination function.)

Removing excess heat from the lamp consists of two or more components of thermal
resistance in a serial configuration. As part of the packaging, the LED die is attached to a metal
slug (e.g., aluminum or copper) to conduct heat away from the lamp. Depending on the end-
use, the lamp might be connected to a printed circuit board or other electrical mounting system.
In any case, thermal management of heat flow from the lamp package through the mounting
and then to the surrounding ambient must be an engineering consideration for any application.

Dept of EEE SICET 38

 Illumination With Solid State Lighting

CHAPTER 9

ORGANIC LIGHT-EMITTING DIODES

Organic light-emitting diodes (OLEDs) are based on small molecules (Tang 1987),
dendrimers (Halim 1999), or polymers (Burroughes 1990). OLEDs have significant potential
for low-cost manufacturing and enable novel lighting architectures such as curved light
emitting surfaces. Issues of particular relevance to OLEDs include carrier injection and
transport, exciton formation, utilization, and conversion to light, operating lifetime,
encapsulation, and light extraction.

9.1 WHITE-LIGHT CREATION ARCHITECTURES

The two dominant material sets for white-light OLEDs are small molecule (discrete
molecular units with relatively weak inter-molecular bonding) and polymer (covalently bonded
repeat units where addition or subtraction of one repeat unit does not significantly affect the
properties of the molecule). While OLEDs typically have broad emission spectra, they are not
sufficiently broad for white light (the exception is emission from excimers or exciplexes, which
is usually relatively inefficient). General lighting, therefore, requires, as with inorganic LEDs,
mixing the light from at least two sources by either pumping a down-conversion phosphor,
combining two or three electroluminescent devices on a substrate using lateral patterning, co-
doping a single layer with spatially separated multiple chromophores, or stacking multiple
devices using transparent intervening electrodes to generate coaxial color mixing. Compared
to inorganic LEDs, organic devices offer more potential for stacked geometries since lattice
matching and strain are essentially irrelevant. However, patterning is more of a challenge since
organic semiconductors are typically damaged by photolithography chemicals such as resists,
etchants, and solvents.

OLEDs are fundamentally broad-area emitters requiring relatively large-area

substrates to generate the lumen output required for general lighting; however the lower
operating current density makes heat dissipation less of a limiting issue. The large area, low
brightness characteristic also potentially eliminates the need for a luminaire, since there is no
dazzle from such a distributed source. This not only saves capital cost, but also increases their
effective efficiency, since the reflector and diffuser enclosure required for a point source
reduces the lighting system efficiency by 30–50%.

Dept of EEE SICET 39

 Illumination With Solid State Lighting

Small-Molecule OLEDs (SM-OLEDs)

The most efficient OLEDs are currently based on “small molecule” materials (referred
to as SM-OLEDs); a schematic of a SM OLED is shown in Figure 9.1. SM-OLEDs consist of
many layers in a structure such as:

1. Substrawhich is usually transparent. However, it is possible to make either or both

electrodes of an OLED transparent (Burrows et al. 2000) so in some cases opaque
substrates such as metal foil can be used.
2. A transparent anode through which light is usually emitted; indium tin oxide (ITO) was originally
used but more complex compounds such as zinc indium tin oxide with more optimal work function
properties are now common in the displays industry.
3. A hole-injection layer (HIL) such as Poly(3,4-ethylenedioxythiophene)– Polystyrene (PEDOT-
PSS) or copper phthalocyanine
4. A hole-transport layer (HTL), such as N,N’-bis-(1-naphthyl)-N,N’-biphenyl-4 ,4’- diamine (NPD)
5. An emissive layer (EML) containing fluorescent and/or phosphorescent dyes usually doped into a
host matrix, discussed separately below
6. An electron transport layer (ETL), e.g., tris(8-hydroxyquinolinato) aluminum (Alq3); perhaps with
an additional hole-blocking molecule, e.g., bathocuproine (BCP), either added to this layer or as a
separate exciton and hole-blocking layer (HBL) between the EML and ETL layers to minimize
carrier leakage and exciton quenching at the cathode
7. A cathode consisting of a thin layer of LiF capped with aluminum. In top-emitting architectures,
the cathode is transparent, permitting the use of extremely low cost substrate materials, such as
metal foil.

Figure 9.1. Schematic diagram of a Small-Molecule OLED (SM-OLED) and a Polymer OLED
(PLED).

Dept of EEE SICET 40

 Illumination With Solid State Lighting

SM-OLED organic layers are usually grown by vacuum deposition (also known as
vacuum thermal sublimation) in which organic source molecules are heated to sublimation
within a vacuum chamber (base pressure around 10-6 Torr), for deposition on a substrate that
is usually close to room temperature. (Active cooling is typically not required due to the
relatively low temperature of the source materials.) The vacuum equipment required represents
a high capital cost but is capable of coating large areas and achieving high product throughput,
since lattice-matching constraints are absent for this materials system. Another growth
technique that has been applied recently is organic vapor phase deposition (OVPD), in which
a carrier gas transports organic molecules within a hot-walled, low-pressure (0.1–1 Torr)
growth chamber onto a cooled substrate (Baldo et al. 1998). This method yields higher
deposition efficiency than vacuum deposition, since only the substrate is coated while the
reactor walls remain clean; this may also reduce downtime for cleaning and allow for an extra
degree of control over thin film morphology. However, this, and the related technique (Sun,
Shtein, and Forrest 2005) of organic vapor jet printing (OVJP), are very new, and the limits on
deposition rate over large areas, throughput, and therefore cost are still unclear.

Solution-based processing for small molecule materials is now also being developed.
Branched molecules known as dendrimers have generated particular recent interest for this
processing method, and are claimed to combine the best features of small molecule and
polymeric materials.

The operating lifetime of OLEDs is inversely proportional to the operating current

density, which may be of concern for high brightness lighting applications. Recent results on
single-color OLEDs, however, show promise. For example, Universal Display Corporation
(UDC) has developed a sky blue (CIE 0.16,0.37) OLED with 100,000 hour lifetime at 200
cd/m2 and luminous efficacy of 20 cd/A, and a deeper blue (CIE 0.16,0.29) device with a
17,500 hour lifetime at 200 cd/m2 and 21 cd/A luminous efficacy. In early 2006, Thompson
(U. So. Cal.), Forrest (U. Michigan), and UDC announced white OLEDs using (a) side-by-side
RGB stripes or (b) mixed phosphors within the same EML, each with luminous efficacy of 20
lm/W at 800 cd/m2 .

Polymer OLEDs (PLEDs)

An alternate approach is to use a single polymercontaining emission layer between

two electrodes to produce a PLED (polymeric LED), shown schematically in Figure 13.

Dept of EEE SICET 41

 Illumination With Solid State Lighting

Polymers can be deposited over broad areas with relatively simple solution-based approaches
such as spin casting or doctor-blading, which are less capital-intensive than vacuum deposition.
The need to pattern more than one color of device on a single substrate has led to the further
development of ink-jet printing techniques.

The limits to the manufacturing cost of such techniques over the large areas required
for lighting is unknown. A further challenge of the solution process is the difficulty of making
multilayer heterostructure devices, which can give higher efficiency but require either
polymers with orthogonal solvent systems or polymers which can be cross-linked by thermal
treatment. Currently, PLEDs are less power efficient than smallmolecule devices. This is
primarily because of the lack of very high quantum efficiency eletrophosphorescence in
polymeric systems, which is not fully offset by the lower operating voltage of PLEDs compared
to SM-OLEDS. Conjugated polymers such as poly(phenylene)vinylene and polyfluorene
polymers have been used for OLEDs in the past for fluorescent devices. However, these
polymer systems appear to quench phosphorescence emission, and nonconjugated polymers
are now being used for higherefficiency phosphorescent devices. Nevertheless, 4 square-foot.
lighting panels have been demonstrated by General Electric (Duggal 2005).

White light can be produced by mixing polymer hosts with RGB-emitting

chromophores within a single EML, or in a multilayer structure in which different layers emit
different colors of light to produce white. In addition, polymers with different dopants can be
applied using commercial inkjet technology to apply RGB patterning (e.g., pixels or stripes of
three colors) to produce white light or for pixelation in full-color displays.

Some recent benchmark PLED results indicate the current state of the art in PLED
technology. As for single-color devices, in late 2005 Cambridge Display Technology
announced a blue PLED with 37,500 hour lifetime at 200 cd/m2 and 10 cd/A luminous efficacy,
and a red PLED with 125,000 hour lifetime at 200 cd/m2 and 7 cd/A luminous efficacy.

Concerning white OLEDs (WOLEDs), at the 2005 Fall MRS Meeting, Yang (UCLA)
reported a white PLED with 14 lm/W. In early 2006, Franky So’s group (U. Florida) also
reported a WOLED with 25 lm/W using a blue organic phosphorescent dye (FIrpic) and
wavelength down-conversion employing a nitridosilicate phosphor (Krummacher et al. 2006).

9.2 CARRIER INJECTION AND TRANSPORT

Dept of EEE SICET 42

 Illumination With Solid State Lighting

General features of carrier injection and transport are schematically shown in Figure
9.1 ; for simplicity, a single organic layer structure is used for illustration, with some details
near the interfaces such as band bending and defect states omitted. Many features of injection
and transport are common to both PLEDs and SM-OLEDs, although the details differ because
of stronger carrier localization in the latter class of materials. Indium tin oxide (ITO) is
illustrated here as the hole-injecting anode, and the medium for carrier transport is a conjugated
organic material, such as polyphenylene vinylene or tris (8-hydroxyquinoline) aluminum.
Delocalized π-bonding orbitals form the equivalent of the valance band, and anti-bonding π*
orbitals form the organic equivalent of the conduction band. The energy of the highest occupied
molecular orbital (HOMO) corresponds to the top of the valence band, and the lowest
unoccupied molecular orbital (LUMO) energy corresponds to the bottom of the conduction
band. (Although the LUMO and HOMO appear similar to conduction and valence bands in
semiconductors, charge carriers are much more localized in organics, and they are also more
strongly coupled to phonon modes. Where bands exist in any meaningful sense, they are very
narrow and background carrier densities are very low relative to conventional semiconductors.)

In the case where the interaction between adjacent materials is weak and the interface
dipole is negligible, the difference between the ionization potential of the organic HTL and the
work function of the anode (Φa) is the barrier for hole injection (ΔEh). The energy levels of
ITO and PPV are well aligned to give a relatively low barrier of 0.2 eV for hole injection
(Brown et al. 1992). In many cases, however, interfacial dipoles or surface chemistry act to
change these idealized barriers.

The cathode is a metal film deposited on top of the final organic layer via thermal evaporation.
The difference in energy between the cathode work function (Φm) and the electron affinity of
the LUMO (EA) is the barrier for electron injection (ΔEe). As drawn in Figure 14, a lower
cathode work function gives a lower electron injection barrier. For this reason, low work
function metals (e.g., barium or calcium) are typically used as OLED cathodes. Because of the
reactivity of these metals, the cathode is capped with a layer of aluminum. Again, however,
almost all effective OLED cathodes react with the ETL forming dipoles and defect states which
enhance electron injection. In essentially oxygenfree environments, covalent bonds can form
between the initially deposited metal atoms and the organic, modifying the electronic structure
of the system. The metal-organic interfacial region is on the order of 20–30 Ǻ, comparable to
the electron tunneling distance (Salaneck and Bredas 1996). The presence of even low amounts

Dept of EEE SICET 43

 Illumination With Solid State Lighting

of oxygen can cause formation of an insulating oxide layer which degrades performance.
Interfacemodification layers such as LiF are often used between the organic layer and the
cathode to improve this interface.

When a charged carrier is injected onto a conjugated molecule, it distorts the molecular
geometry which relaxes around the extra charge creating a polaron (Holstein 1959) which is
often referred to as a self-trapped state. The mechanism of carrier transport through organics is
often “polaron hopping” from one molecule to a neighbor. Disorder in the largely amorphous
organic films disrupts charge transport and plays a dominant role in determining the charge
carrier mobility.

9.3 EXCITON DYNAMICS AND LIGHT CREATION

Electrons and holes within the emission layer combine to form a neutral exciton, which
can then diffuse. If normal Langevin statistics prevail, singlet and triplet excitons are formed
in a 1:3 ratio. A crucial step towards making very efficient OLEDs is to convert all the singlet
and triplet excitons into light output. The first generations of OLEDs were based on fluorescent
dyes. In fluorescence, quantum mechanical spin is conserved when a singlet excited state
(exciton) emits a photon and drops to the singlet ground state. However, because a triplet-to-
singlet transition via light emission does not conserve spin, it is a “forbidden transition;” the
energy of the triplet excitons is wasted as dissipated heat. Thus, OLEDs utilizing fluorescence
alone were fundamentally limited to an internal quantum efficiency of 25%. There is ongoing
debate as to the exact singlet-triplet ratio in organics, and there is evidence for a material
dependence, with measurements suggesting well over 25% singlets in some PLEDs (Reufer et
al. 2005). There is some practical significance to this debate because the ability to engineer
materials for very high singlet exciton generation could potentially recover the exchange
energy that is currently lost in phosphorescent devices, yielding even higher efficiencies. Such
a device, however, has yet to be demonstrated.

Another approach is a phosphorescent OLED (PhOLED), in which the presence of a

heavy metal atom (e.g., Pt, Pd, Ir, Au) within a phosphor provides spin-orbit coupling and
mixing of the singlet and triplet states. The goal is conversion of all excitons to the triplet state,
followed by rapid phosphorescence; phosphorescent materials can relax from a mixed spin
metal ligand charge transfer excited state to the ground state by emission of a photon, and the
spin statistics of exciton creation are therefore irrelevant. Examples of such organometallics

Dept of EEE SICET 44

 Illumination With Solid State Lighting

phosphors are platinum octaethylporphine (PtOEP) and iridium tris(phenylpyridine)

(Ir(ppy)3), with triplet excited state lifetimes of ~100 μs for the former and <1 μs for the latter.
There is still an inherent energy loss (known as the exchange energy) when a singlet exciton is
converted by this process to a lower-energy triplet exciton. A recent approach is to combine
fluorescent and phosphorescent dyes to utilize the singlets for blue light and the triplets for
lower-energy green and red light (Sun et al. 2006). The relatively long excited-state lifetime of
triplet excitons also creates exciton-exciton and exciton-polaron quenching effects at high drive
currents, which result in efficiency loss at high brightness (Baldo, Adachi, and Forrest 2000).

9.4 ENCAPSULATION

Low-cost packaging and encapsulation approaches are needed to limit degradation of

OLEDs. Low work-function metals such as Ca are used as cathodes in OLEDs and must be
protected from reactions with oxygen and water. These ambient vapors will particularly
degrade the metal-organic interface, limiting device lifetime. Excited-state organic molecules
in the OLED EML are also very susceptible to oxidation (photo-oxidation), and must be
protected from contact with air. Ingress of water vapor likely leads to electrochemical reactions
near the electrodes and is particularly deleterious.

Currently, OLEDs are encapsulated by attaching a glass or metal lid above the cathode
using a bead of epoxy. Desiccant is incorporated in the package to absorb residual moisture
released from, and permeating through, the epoxy seal. This approach is limited to rigid and
small-area devices. For the envisioned flexible and large-area OLEDs required for lighting,
new encapsulation methods will be needed unless air-stable materials can be developed. High-
barrier coatings can be used to provide low-cost thinfilm encapsulation. These barriers must
be pin-hole free, robust, and tolerant to high temperatures that will occur when devices are
driven for high brightness. These barriers need to transmit < 10*6 g/m*2 day of water and <
10*5 cc/m*2 day of oxygen to ensure adequate OLED lifetimes (Burrows et al. 2001).

Dept of EEE SICET 45

 Illumination With Solid State Lighting

Furthermore, these targets will have to be met at low cost over large areas and in a
manufacturing environment. Achieving this may require new breakthroughs in both the science
and engineering of encapsulation

Figure 9.4. Inefficient light extraction in OLED device.

9.5 LIGHT EXTRACTION

As much as 80% of the light produced in an OLED can be lost to internal reflections
and waveguiding within the device or substrate layers (Figure 9.4). A number of different
approaches to improving the light extraction coefficient, Re, have been employed. Because of
the difference in refractive indices of air (1.0) and the glass substrate and organic layers (both
on the order of 1.5 to 1.8), only light incident on the interface within a certain escape cone will
pass through. Techniques for improving light extraction efficiency primarily try to increase the
effective light cone. Light extraction is even more of a problem for inorganic LEDs, where the
solid indices of refraction are much larger, on the order of 2.5. While the same fundamental
physics likely applies to both cases, the economics of light extraction from large area OLEDs,
rather than a small inorganic semiconductor die, likely creates new challenges.

Many approaches for improved light extraction are being pursued including:
roughening or texturing the outside surface of the glass substrate; corrugating the surface to
increase Bragg scattering in the forward direction; using a two-dimensional photonic crystal
structure to improve light output coupling; attaching an ordered array of micro lenses to the
glass substrate; shaping the device into a mesa structure, and including reflective surfaces
within the device.

Improvements in light extraction efficiency by factors of two to three have been

reported for most of the techniques listed above. Other issues that must be considered include
manufacturing cost and avoiding undesirable changes in the radiation pattern or an angular-
dependent emission spectrum. A single standard approach to light extraction for OLEDs has
yet to be established, and this is an important area for development.

Dept of EEE SICET 46

 Illumination With Solid State Lighting

CONCLUSION
As a result of the advances in semiconductor quality and of this device technology, the
LED efficiency exceeds the best available colored (filtered) incandescent, halogen, and fluores-
cent lighting in all colors, and the white LED efficiency exceeds the efficiency of both
incandescent and halogen sources.

The typically small mass of a solid-state electronic lighting device provides for greater
resistance to shock and vibration compared to brittle glass tubes/bulbs and long, thin filament
wires. They also eliminate filament evaporation, potentially increasing the life span of the
illumination device.

Dept of EEE SICET 47

 Illumination With Solid State Lighting

FUTURE SCOPE
There has been dramatic progress in the last few years on high power LEDs and there
is now clear difference in appearance and performance for LEDs designed as indicators and
LEDs that are the forefathers of solid state lamps. However, today even the most powerful
LEDs being made are still an order of magnitude too low in flux per LED, and two orders of
magnitude too high in cost per lumen to significantly penetrate the general illumination
market. Making the jump in flux/LED will require major improvements in packaging and
fixture integration that further reduce thermal resistance from LED junction to ambient into the
10-K/W range. Progress in reducing cost per lumen will come from process improvements and
the volume increases that result from market penetration of the pure color and high value niche
white applications. One hundred years after Edison’s discovery of a filament that made
incandescent bulbs practical, we cannot yet speak of “Solid State Illumination,” but we are near
enough to see the outlines of the future of solid state lighting.

Dept of EEE SICET 48

 Illumination With Solid State Lighting

REFERENCES
[1] N. Holonyak Jr. and S. F. Bevaqua, “Coherent (visible) light emission from GaAs P
junctions,” Appl. Phys. Lett., vol. 1, pp. 82–83, 1962. In 1995, Nick Holonyak received the
Japan Prize for his seminal work on LED’s and lasers.

[2] G. E. Moore, “Cramming more components onto integrated circuits,” Electron., vol. 38, no.
8, Apr. 1965

[3] R. Haitz, private communication.

[4] H. Rupprecht, J. M. Woodall, and G. D. Petit, “Efficient visible electroluminescence at

300 K from Ga Al As p–n junctions grown by liquid-phase epitaxy,” Appl. Phys. Lett., vol. 11,
pp. 81–83, 1967.

[5] H. Ishiguro, K. Sawa, S. Nagao, H. Yamanaka, and S. Koike, “High efficency GaAlAs light-
emitting diodes of 660 nm with a double heterostructure on a GaAlAs substrate,” Appl. Phys.
Lett., vol. 43, pp. 1034–1036, 1983

[6] C. P. Kuo, R. M. Fletcher, T. D. Osentowski, M. C. Lardizabal, M. G. Craford, and V. M.

Robbins, “High performance AlInGaP visible lightemitting diodes,” Appl. Phys. Lett., vol.
57, pp. 2937–2939, 1990

[7] S. A. Stockman and G. E. Stillman, “Hydrogen in III-V device structures,” Mater. Sci.
Forum, vol. 148–149, pp. 501–536, 1994

[8] M. R. Krames et al., “High-power truncated-inverted-pyramid (Al Ga ) In P/GaP light-

emitting diodes exhibting> 50% external quantum efficiency,” Appl. Phys. Lett., vol. 75, pp.
2365–2367, 1999

[9] S. Nakamura, “GaN Growth using GaN Buffer Layer,” Jpn. J. Appl. Phys., pt. 2, p. L1705,
1991.

[10] H. Amano, N. Sawaki, I. Akasaki, and Y. Toyoda, “Metalorganic vapor phase epitaxial
growth of a high quality GaN film using an AlN buffer Layer,” Appl. Phys. Lett., vol. 48, p.
353, 1986.

[11] W. Goetz et al., “Activation of acceptors in Mg doped GaN grown by metalorganic

chemical vapor deposition,” Appl. Phys. Lett., vol. 68, no. 5, pp. 667–669, 1996

Dept of EEE SICET 49

 Illumination With Solid State Lighting

[12] R. V. Steele, “High-brightness LED market overview,” in Proc. SPIE, vol. 4445, 2001, pp.
1–4.

[13] G. E. Höfler, D. A. Vanderwater, D. C. DeFevere, F. A. Kish, M. D. Camras, F. M.

Steranka, and I.-H. Tan, “Wafer bonding of 50-mm diameter GaP to AlGaInP–GaP light-
emitting diode wafers,” Appl. Phys. Lett., vol. 69, p. 803, 1996.

[14] J. J. Wierer, D. A. Steigerwald, and M. R. Krames, “High-power AlGaInN flip-chip

light-emitting diodes,” Appl. Phys. Lett., vol. 78, p. 3379, 2001.

[15] “LumiLeds lighting launches multi-format luxeon light sources,” Compound

Semiconductor Mag., vol. 7, no. 6, p. 11, 2001

[16] W. Goetz, F. Ahmed, J. Bhat, L. Cook, N. F. Gardner, E. Johnson, M. Misra, R. S. Kern,

A. Y. Kim, J. Kim, J. Kobayashi, M. R. Krames, M. Ludowise, P. S. Martin, T. Mihopoulos, A.
Munkholm, S. Rudaz, S. Salim, Y.-C. Chen, D. A. Steigerwald, S. A. Stockman, J. Sun, J. J.
Wierer, D. Vanderwater, F. M. Steranka, and M. G. Craford, “Power IIINitride LEDs,” in
International Conf. Nitride Semiconductors, Denver, CO, July 2001.

[17] Paul S. Martin, J. Bhat, C.-H. Chen, D. Collins, W. Goetz, R. Khare, A. Kim, M. Krames,
C. Lowery, M. Ludowise, G. Mueller–Mach, S. Subramanya, S.-C. Tan, J. Thompson, T.
Trottier, S.-C. , and R. Khare, “High Power White LED Technology for Solid State Lighting,”
in Proc. SPIE, San Diego, CA, July 2001.

[18] H. Chui, N. F. Gardner, P. N. Grillot, J. W. Huang, M. R. Krames, and S. A. Maranowski,

“High efficiency AlInGaP light-emitting diodes,” in Semiconductors and Semimetals, G. B.
Stringfellow and M. G. Craford, Eds. New York: Academic, 2000, vol. 64, p. 49.

[19] G. Wyszecki and W. S. Stiles, Color Science: Concepts and Methods, Quantitative Data
and Formulae, 2nd ed. New York: Wiley, 1982.

[20] S. Muthu, F. J. P. Schuurmans, and M. D. Pashley, “Red, Green, and Blue LEDs for white
light illumination,” IEEE J. Select. Topics Quantum Electron., vol. 8, pp. 333–338, Mar.–Apr.
2002.

[21] Lumileds Corporate Website [Online]. Available: www.lumileds.com/ solutions/LCD/

LCDindex.html

Dept of EEE SICET 50

 Illumination With Solid State Lighting

[22] R. Mueller-Mach, G. O. Mueller, M. R. Krames, and T. Trottier, “High-Power Phosphor-

Converted Light-Emitting Diodes Based on III–Nitrides,” IEEE J. Select. Topics Quantum
Electron., vol. 8, pp. 339–345, Mar.–Apr. 2002

Dept of EEE SICET 51