fiber materials, fiber fabrication techniques,

In selecting materials for optical fibers, following requirements must be satisfied

1. It must be possible to make long, flexible fibers from the material.
2. The material must be transparent at a particular optical wavelength in order for the fiber
to guide light efficiently.
3. Physically compatible materials having slightly different refractive indices for the core
and cladding must be available.
Commonly available materials that satisfy the above requirements are the glass and plastics.
The majority of fibers are made of glass consisting of silica (SiO2) or silicate. Plastic fivers are
less popular because of its substantially higher attenuation than glass fibers. Plastic fibers are
used in short distance applications and in abusive environments, where the greater mechanical
strength of plastic fibers offers an advantage over the use of glass fibers.

Glass Fibers
Glass is made by fusing mixtures of metal oxide, sulphides or selenides. The resulting material is
a randomly connected molecular network. When glass is heated up from room temperature, it
remains a hard solid upto several hundred degrees centigrade. As the temperature increases
further, the glass gradually begins to soften until at very high temperatures it becomes a viscous
liquid.
Melting temperature of glass refers to an extended temperature range in which the glass becomes
fluid enough to free itself fairly quickly of gas bubbles.
The most commonly used oxide glass material is silica (SiO2) which has a refractive index of
1.458 at 850 nm.Doping silica with GeO2 , P2O5 increases the refractive index of silica, where as
doping with fluorine or B2O3 decreases the refractive index.Since cladding must have a lower
refractive index than the core, the possible materila used are,
1. GeO2- SiO2 core; SiO2 cladding
2. P2O5-SiO2 cladding
3. SiO2 Core; B2O3- SiO2 cladding
4. GeO2- B2O3- SiO2 core; B2O3- SiO2
The raw material for silica is sand.Some of the desirable properties of silica glass, are a
resistance to deformation at temperatures as high as 1000o C, a high resistance breakage from
thermal shoke because of its thermal expansion, good chemical durability and high transparency
in both visible and infrared regions of interest to fiber optic communication systems.
Its high melting temperature is a disadvantage if the glass is prepared from a molten state.
However, this problem is partially avoided when using vapour deposition techniques.

Halide Glass Fibers:

Fluride glasses (discovered in 1975) have exteremly low transmission losses at mid-infrared
wavelength (0.2 – 0.8 μm with lowest loss being around 2.55μm).Fluride glasses belong to a
general family of halide glasses.
Fluride glass uses ZrF4 as the major component and glass network former. Fluride glass forms
the core of a glass fiber. To make a lower refractive index glass for the cladding , ZrF4 is
replaced by HaF4.
These glasses offer intrinsic minimum losses of 0.01 to 0.001 dB/km, but fabrication for long
lengths of these fibers is difficult.

Active Glass Fibers

Rare – earth elements (atomic number 57 to 71) when incorporated into normal passive glass
materials (like Silica and halide glasses) gives the resulting material new optical and magnetic
properties like amplification, attenuation , phase retardation etc.

Plastic –Clad Glass Fibers

For short distance applications, where higher losses are tolerable, the less expensive plastic-clad
silica fibers can be used. These fibers are composed of silica cores with lower refractive –index
cladding being a polymer (plastic) material. These fibers are referred to as PCS fibers.
Commonlr used material for the silica core is high purity natural quartz. A coomon cladding
material is a silicone resin having a refractive index of 1.405 at 850 nm. Silicone resin is also
frequently used as a protective coating for other types of fibers. Another popular plastic cladding
material is perfluoronated ethylene propylene (Teflon) . The low refractive index 1.338 , of this
material results in fibers with potentially large numerical aperures.
Plastic claddings are only used for step – index fibers. The core diameters are larger (150 to 600
μm) than the standrd 50 μm diamter core of all glass gradded-index fibers and the larger
differencfe in the core and the cladding indices results in a high numeriacal aperture. This allows
llow cost large area light sources to be used for coupling optical power into these fibers.
Plastic Fibers
All plastic multimode step index fibers are good transmission media for fairly short and low cost
links. They exhibit greater signal attenuation than glass fibers, but also offer toughness and
durability and greater mechanical flexibility. The high refractive index differenced that can be
achieved between the core and the cladding materials yields numerical apertures as high as 0.6
and large acceptance angles of up to 700 Example of lastic fiber constructions are:
A polysterene core (n1 = 1.6 ) and a methyl methacrylate cladding (n2 = 1.49) to give an NA of
0.6.
A polymethyle methacrylate core (n1 = 1.49) and a cladding made of its copolymer (n2= 1.4) to
give an NA of 0.5.

The direct –melt methods:

In this technique the optical fibers are made directly from the molten state of purified
components of silicate glasses.
Outside Vapor Phase Oxidation
In this process, fist a layer of SiO2 particle called soot is deposited from a burner onto a rotating
graphite or ceramic mandrel. The glass soot adheres to his bait rod and layer by layer , a
cylindrical , porous glass perform is built up.
By properly controlling the constituents of the metal halide vapour stream during the deposition
process, the glass compositions and dimensions desired for the core and cladding can be
incorporated into the preform. Either step or graded index perform can thus be made.
When the deposition process is complete ed , the mandrel is removed and the porous tube is then
vetrified in a dry atmosphere at a high temperature ( above 14000C) to a clear glass perform.
This clear prefrom is subsequently mounted ina fiber – drawing tower and made into a fiber.The
central hole in the tube prefrom collapses during this drawing process.

Vapor Phase Axial Deposition

In this methos the SiO2 particles are formed emerge from the toches, they are deposited onto the
end surface of a silica glass rod which acts as a seed.A porous prefrom is grown in the axial
direction by moving the rod upwars.
The rod is also continuously roteated to maintain cylindrical symmetry of the particle deposition
. As the porous perform mves upward, it is transformed into a solid , transparent rod prefprm by
zone melting with the carbon ring heater.
The resultant perform can then be drawn into a fiber by heating it in another furnace.
The advantage of this method is that
The perform has no central hole as ocured incase of OVPO process
The perform can be fabricated in continuous lengths which can affect process costs and product
yields
The fact that the deposition chamber and the zone-melting ring heater are tightly connected to
each other in the same enclosure allows the achievemen of a clean environment.

Modified chemical vapour deposition

In this method, the glass vapour particles arising from the reaction of the constituent metal
halide gases and oxygen flow through the inside of a revolving silica tube. As the SiO2 particles
are deposited , they are sintered to a clear glass layer by an oxyhydrogen torch which travels
back and forth along the tube.
When the desired thickness of glass has been deposited the vapour flow is shut off and the tube is
heated strongly to cause it to collapse into a solid rod perform.
The fiber that is subsequently drawn from this prefrom rod will have a core that consists of the
vapour –deposited material and a cladding consisting of the original silica tube.
Plasma –Activated Chemical Vapor Deposition
In this method deposition occurs within a silica tube.
However a nonisothermal microwave plasma operating at low pressure initiates the chemical
reaction.
The silica tube is held at a temperature in the range of 1000 to 12000C to reduce the mechanical
stresses in the growing glass films, a moving microwave resonator operating at 2.45 GHz
generates a plasma inside the tube to activate the chemical reaction.
This process deposits clear glass material directly on the tube wall. When the desired glass
thickness is deposited, the tube is collapsed into a preform.
Double-Crucible Method
In this method, glass rods for the core and the cladding materials are first made separately by
melting mixures of purified powders to make the appropriate glass composition. /these rods are
then used as feedback for each oft wo concentric crucibles.The inner crucibles contains molten
core glass and the outer one consists of the cladding glass. The fibers are drawn from the ,molten
state through orifices in the bottom of the two concentric crucible in a continuous production
process.
The process is a continuous one , but have to be carefull to avoid contamination during the
melting.

Mechanical Properties of fibers

Mechanical properties of fibers play a very important role when they are used as the transmission
medium.
The fibers must be able to withstand the stresses and strains that occur during the cabling process
and the loads induced during the installation and service of the cable. During cable manufacture
and installation the loads applied to the fiber can be either impulsive or gradually varying. Once
the cable is in place, the service loads are usually slowly varying ones, which can arise from
temperature variations or a general settling of the cable following installation.
Strength and static fatigue are the two basic mechanical characteristic on glass optical fibers. The
longitudinal breaking stress of pristine glass fibers is comparable to that of metal wires. The
cohesive bond strength of the constituent atoms of a glass fibers governs its theoretical intrinsic
strength. Maximum tensile strength of 14 GPa (2 x 106 lb/in2) have been observed in short –
gauge –length glass fibers. In practice the existence of stress concentrations at surface flaws on
microcracks limits the medium strength of long glass fibers.
Since an optical fiber generally contains many flaws having a random distribution of size, the
fracture strength of a fiber must be viewed statistically. If F(σ,L) is defined as the cumulative
probability that a fiber of length L will fai below a stress level σ, then under the assumption that
the laws are independent and random distributed in the fiber and that the fracture will occur at
the most severe flaw , we have
F(σ, L) = 1 – e LN(σ)
Where N(σ) is the cumulative number of flaws per unit length with a strength less than σ.
In contrast to strength , which relates to instantaneous failure under an applied load , static
fatigue relates to the slow growth of pre-existing flaws in the glass fiber under humid conditions
and tensile stress. This gradual flaw growth causes the fiber to fail at a lower stress level than
that which could be reached under a strength test. A flaw propagates through the fiber because of
chemical erosion of the fiber material at the flaw tip. The primary cause of this erosion is the
presence of water in the environment, which reduces the strength of the SiO2 bonds in the glass.
The speed of the growth reaction is increased when the fiber is put under stress.
Certain fiber materials are more resistance to static fatigue than others, with fused silica being
the most resistance of the glasses in water. In general coating which are applied to the fiber
immediately during the manufacturing process affords a good degree of protection against
environmental corrosion.
Another important factor to consider is dynamic fatigue. Whne an optical cable is being installed
in a duct, it experience repeated stress owing to surging effects. The surging is caused by varying
degrees of friction between the optical cable and the duct or guiding tool in a man hole on a
curved route. Varying stresses also arise in aerial cables that are set into transverse vibration by
the wind.

FIBER OPTIC CABLES.

For practical applications , The fibres need to be incorporated in some type of cable structures.
The cable structure will vary depending on whether the cable is to pulled into underground or
intrabuilding ducts, buried directly in the ground, installed on outdoor poles, or submerged under
water. In each application certain fundamental cable design principle applies.
One important mechanical property is the maximum allowable axial load on the cable, which
determines the maximum length of the cable without fracture that can be reliably installed. Since
static fatigue occurs very quickly at stress level above 40 percent of the permissible elongation
and very slowly below 20 percent , fiber elongation during cable manufacture and installation
should be limited to 0.1 to 0.2 percent.
Plastic strength members and high tensile strength organic materials are used as cable to avoid
the effects of elctromagnetic induction and to reduce the cable weight. With good fabrication
techniques the optical fibers are isolated from other cable components, they are kept close to the
neural axis of the cable and room is provided for the fibers to move when the cable is flexed or
stretched.
Another factor o consider is fiber brittleness Since glass fibers donot deform plastically , they
have a low tolerance for absorbing energy from impact loads. Hence the outer sheath of an
optical cable must be designed to protect the glass fibers inside room impact forces. In addition ,
the outer sheath should not crush when subjected to side forces, and it should provide protection
from corrosive environmental elements. In underground installation , a heavy gauge –metal
outer sleeve may also b required to protect against potential damage from burrowing rodents.
In design optical fibre cables , several types of fiber arrangements are possible and a large
variety of components could be included in the construction. In a hypothetical two fiber design, a
fibre is first coated with a buffer material and placed loosely in a tough, oriented polymer tube,
such s polythene. For strength purposes this tube is surrounded by strands of array mid yarn
which in turn is encapsulated in a polyurethane jacket. A final outer jacket of polyurethane ,
polyethylene or nylon binds the two encapsulated fiber units together.

Larger cables can be created by stranding several basic fiber building blocks around a central
strength members. The fiber units are bound onto the strength member with paper or plastic
binding tape, and then surrounded by an outer jacket.
Energy Bands
Semiconductor materials have conduction properties that lie between those of metals and
insulators. The conduction properties can be interpreted with the aid of theenergy band
diagram. In a pure crystal at low temperature the conduction band is comoletely empty of
electrons and the valence band completely full. These two bands are separated by an energy
gap, or band gap in whuch no energy leels exista.As the temperature is raised , some of the
electrons ate thermally ecxited across the band gap. For Si this excitation energy must be
greater than 1.1 eV which is the band gap energy.This gives ridse to concentration of n free
electrons in the conduction band, which leaves behind an equal concentration p of vacancies or
holes in the valenve band.Both the free electron and the holes are mobile with in the material,
so that both can contribute to electrical conductivity, that is an electron in the valence band can
move to vacant hole.This action makes the hole move in the opposite direction to the electron
flow.

The concentration of electrons and holes is known as intrinsic carriers concentration n i and is
given by
N = p = ni = K exp ( - Eg / 2 kB T)
Where K is a constant and is a characteristic of the material.
The conduction can be increased by adding traces of impurities from the group V elements. This
process is called doping. When group five elements replaces a Si atom four electron are used
for covalent bonding and the fifth loosely bound electron is available for conduction, giving rise
to an occupied level just below the conduction band called donor level. The impurities are
called donors because they can give rise to free electron to the conduction band. As the current
is carried by electrons , it is called n-type material.
The conduction can also be increased by adding group III materials which have three electrons
in the outer shell, which makes three covalent bond and hole is created .This give rise to an
unoccupied level just above the valence band. Correspondingly the free holes concentration
increases in the valence band and called as the p-type materil.
A material containing no impurity is called as intrinsic material. Because of thermal vibration of
the crystal atoms , some electrons in the valence band gain enough energy to be excited to the
conduction band. This thermal generation produces free electron – hole pairs. In the opposite
recombination process,a free electrion reeases its energy and drops into free hole in the
valence band. The generation and recombination rates are equal in equilibrium. If n the number
of electron concentration and p is the hole concentration then for an intrinsic material
Pn = p0 no = ni2
The inroductioof small qualntities of chemical imputities into crystal produces an extrinsic
semiconductor.
The pn junction
Doped n- and p – semiconductor material junction is responsible for electrical characteristics of
semiconductor devices. When pn junction is created, the majority of carriers diffuse across it.
This causes electrons an to fill holes in the p side of the junction and causes holes to appear
across the junction. The field prevents further net movement of charges once equilibrium has
been established. The junction now has no mobile carriers and is called as depletion region.
When the pn junction is reverse biased , the width of the depletion region will increase on both
the side of the junction.This effectively increases the barrier potential and prevents any
majority carriers from flowing across the junction. However minority carriers can move with the
field acoss the junction.The minority current flow is small at normal temperature and oprting
voltages., but it can be significant when excess carriers are created as in case of photodiode.

When the pn junction is forward biased the magnitude of the barrier potential is reduced.
Conduction band electrons on the n side and valence band holes on the p side are thereby
allowed to diffuse across the junction. Once across, they significantly increase the minority
carriers concentration, and the excess carriers then recombine with he oppositely charged
majority carriers. The recombination of excess minority carriers is the mechanism by which
optical radiation is generated.
3.1.2 p–n Junctions
Extrinsic semiconductor is made n-type or p-type by doping it with impurities whose atoms have
an excess valence electron or one less electron compared with the semiconductor atoms. In the
case of ntype semiconductor, the excess electrons occupy the conduction-band states, normally
empty in undoped (intrinsic) semiconductors. The Fermi level, lying in the middle of the
bandgap for intrinsic semiconductors, moves toward the conduction band as the
dopant concentration increases. In a heavily doped n-type semiconductor, the Fermi level Ef c
lies inside the conduction band; such semiconductors are said to be degenerate.
Similarly, the Fermi level E f v moves toward the valence band for p-type semiconductors and
lies inside it under heavy doping.
In thermal equilibrium, the Fermi level must be continuous across the p–n junction. This is
achieved through diffusion of electrons and holes across the junction. The charged impurities left
behind set up an electric field strong enough to prevent further diffusion of electrons and holds
under equilibrium conditions. This field is referred to as the built-in electric field.

Fig (a) shows the energy-band diagram of a p–n junction in thermal equilibrium andunder
forward bias.

When a p–n junction is forward biased by applying an external voltage, the built in electric field
is reduced. This reduction results in diffusion of electrons and holes across the junction. An
electric current begins to flow as a result of carrier diffusion The current I increases
exponentially with the applied voltage V according to
I = Is[exp(qV/kBT)−1], (3.1.15)
where Is is the saturation current and depends on the diffusion coefficients associated with
electrons and holes.

In a region surrounding the junction (known as the depletion width), electrons and holes are
present simultaneously
when the p–n junction is forward biased. These electrons and holes can recombine through
spontaneous or stimulated emission and generate light in a semiconductor optical source.The p–n
junction shown in Fig. (a) is called the homojunction, since the same semiconductor material is
used on both sides of the junction. A problem with the homojunction is that electron–hole
recombination occurs over a relatively wide region (∼ 1–10 μm) determined by the diffusion
length of electrons and holes. Since the carriers are not confined to the immediate vicinity of the
junction, it is difficult to realize high carrier densities. This carrier-confinement problem can be
solved by sandwiching a thin layer between the p-type and n-type layers such that the bandgap of
the sandwiched layer is smaller than the layers surrounding it. The middle layer may or may not
be doped, depending on the device design; its role is to confine the carriers injected inside it
under forward bias. The carrier confinement occurs as a result of bandgap discontinuity at the
junction between two semiconductors which have the same crystalline structure (the same lattice
constant) but different bandgaps. Such junctions are called heterojunctions, and such devices are
called double heterostructures. Since the thickness of the sandwiched layer can be controlled
externally (typically, 0.1μm),
high carrier densities can be realized at a given injection current. Figure b) shows the energy-
band diagram of a double heterostructure with and without forward bias.

Optical transmitter
Under normal conditions all materials absorbs light rather than emit it. If the photon energy hν of
the incident light of frequency ν is about the same as the energy difference Eg = E2−E1, the
photon is absorbed by the atom, which ends up in the excited state. Incident light is attenuated as
a result of many such absorption events occurring inside the medium. The excited atoms
eventually return to their normal “ground” state and emit light in the process. Light emission can
occur through two fundamental processes known asspontaneous emission and stimulated
emission.
In the case of spontaneous emission, photons are emitted in random directions with no phase
relationship among them.
Stimulated emission, by contrast, is initiated by an existing photon. The remarkable feature of
stimulated emission is that the emitted photon matches the original photon not only in energy (or
in frequency), but also in its other characteristics, such as the direction of propagation. All lasers,
including semiconductor lasers, emit light through the process of stimulated emission and are
said to emit coherent light. In contrast, LEDs emit light through the incoherent process of
spontaneous emission.
The use of a heterostructure geometry for semiconductor optical sources is doubly beneficial. As
already mentioned, the bandgap difference between the two semiconductors helps to confine
electrons and holes to the middle layer, also called the active layer since light is generated inside
it as a result of electron–hole recombination. How ever, the active layer also has a slightly larger
refractive index than the surrounding p-type and n-type cladding layers simply because its
bandgap is smaller. As a result of the refractive-index difference, the active layer acts as a
dielectric waveguide and supports optical modes whose number can be controlled by changing
the active-layer thickness (similar to the modes supported by a fiber core). The main point is that
a heterostructure confines the generated light to the active layer because of its higher
refractive index.

It is this feature that has made semiconductor lasers practical for a wide variety
of applications.

Nonradiative Recombination
When a p–n junction is forward-biased, electrons and holes are injected into the active
region, where they recombine to produce light. In any semiconductor, electrons and holes can
also recombine nonradiatively. Nonradiative recombination mechanisms include recombination
at traps or defects, surface recombination, and the Auger recombination. In the Auger
recombination process, the energy released during electron–hole recombination is given to
another electron or hole as kinetic energy rather than producing light.
All nonradiative processes reduce the number of electron–hole pairs that emit light. Their effect
is quantified through the internal quantum efficiency, defined as Rrr

Where Rrr is the radiative recombination rate, Rnr is the nonradiative recombination rate, and
Rtot = Rrr + Rnr is the total recombination rate. It is customary to introduce Rrr = N / τrr and ,
Rnr = N / τnr , where N is the total carrier density.

The internal quantum efficiency is then given by

.
The radiative and nonradiative recombination times vary from semiconductor to semiconductor.
A semiconductor is said to have a direct bandgap if the conduction band minimum and the
valence-band maximum occur for the same value of the electron wave vector. The probability of
radiative recombination is large in such semiconductors, since it is easy to conserve both energy
and momentum during electron–hole recombination.
By contrast, indirect-bandgap semiconductors require the assistance of a phonon for conserving
momentum during electron–hole recombination. This feature reduces the probability of radiative
recombination. Typically, for Si and Ge, the two semiconductors commonly used for electronic
devices are not suitable for optical sources because of their indirect bandgap. For direct-bandgap
semiconductors such as GaAs and InP ,stimulated emission dominates.
It is useful to define a quantity known as the carrier lifetime such that it represents
the total recombination time of charged carriers in the absence of stimulated
recombination. It is defined by the relation
Rspon + Rnr = N/τ

Semiconductor Materials

Almost any semiconductor with a direct bandgap can be used to make a p–n homojunction
capable of emitting light through spontaneous emission. In the case of heterostructure devices ,
their performance depends on the quality of the heterojunction interface between two
semiconductors of different bandgaps. To reduce the formation of lattice defects, the lattice
constant of the two materials should match to better than 0.1%. Nature does not provide
semiconductors whose lattice constants match to such precision. However, they can be fabricated
artificially by forming ternary and quaternary compounds in which a fraction of the lattice sites
in a naturally occurring binary semiconductor (e.g., GaAs) is replaced by other elements. In the
case of GaAs, a ternary compound Al xGa1-xAs can be made by replacing a fraction x of Ga
atoms by Al atoms. The resulting semiconductor has nearly the same lattice constant, but its
bandgap increases.
Direct and Indirect Band Gaps
In order for electron transition to take place to or from the conduction band with the absorption
or emission of a photon, respectively, both energy and momentum must be conserved.
Semiconductors are classified as direct band gap or indirect band gap materials depending on the
shape of the band gap. In direct band gap materials electron and hole have the same momentum
and hence recombination may take place releasing a photon.

In case of indirect –band-gap materials , the conduction band minimum and the valence band
maximum energy levels occurs at different values of momentum.In that case band to band
recombination must involve a third particle to conserve momentum , since photon momentum is
very small. Phonons also released in this case.
In fabricating semiconductor devices , the crystal structure of the various material regions have
to be considered. In any crystal structure , single atoms or group of atoms are arranged in a
repeated pattern in space. This periodic arrangement defines a lattice, and the spacing between
the atoms or group of atoms is called lattice spacing or the lattice constant which is typically a
few angstroms.

LIGHT EMITTING DIODES

3.2 Light-Emitting Diodes

A forward-biased p–n junction emits light through spontaneous emission, a phenomenon referred
to as electroluminescence. In its simplest form, an LED is a forward biased p–n homojunction.
Radiative recombination of electron–hole pairs in the depletion region generates light; some of it
escapes from the device and can be coupled into an optical fiber. The emitted light is incoherent
with a relatively wide spectral width (30–60 nm) and a relatively large angular spread.
Semiconductor light emitting diodes are the best light source for optical communication systems
requiring bit rate less than approximately 100-200 Mb/s together with multimode fiber coupled
optical power in the tens of microwatts.

LED Structure:
To be useful in fiber transmission applications an LED must have a high radiance output, a fast
emission response time and high quantum efficiency.
Its radiance is a measure in watts of the optical powr radiated into a unit solid angles per unit
area of the emitting surface. High radiance is necessary to couple sufficiently high power levels
into fiber.
The emission response time is the time delay between the application of a current pulse and the
onset of optical emission. This time delay is the factor limiting the bandwidth with which the
source can be modulated directly by varying the injected current.
The quantum efficiency is related to the fraction of injected electron-hole pairs that recombine
radiatively.
To achieve a high radiance and high quantum efficiency the LED structure must provide a means
of confining the charge carriers and the stimulated optical emission to the active region of the pn
junction where radiative recombination takes place. Carrier confinement is used to achieve a
high level of radiative recombination in the active region of the device which yields a high
quantum efficiency. Optical confinement is of importance for preventing absorption of the
emitted radiation by the material surrounding the pn junction.

To achieve carrier and optical confinement , LED with double-heterostructure is commonly

used. It has two different alloy layers on each side of the active region.
/
The double-heterojunction LED

The DH LED is consists of a p-type GaAs layer sandwiched between a p-type AlGaAs and an n-
type AlGaAs layer. When a forward bias is applied electrons from the n-type layer are injected
through the p–n junction into the p-type GaAs layer where they become minority carriers. These
minority carriers diffuse away from the junction , recombining with majority carriers (holes) as
they do so. Photons are therefore produced with energy corresponding to the bandgap energy of
the p-type GaAs layer.
The injected electrons are inhibited from diffusing into the p-type AlGaAs layer because of the
potential barrier presented by the p–p heterojunction . Hence, electroluminescence only occurs in
the GaAs junction layer, providing both good internal quantum efficiency and high-radiance
emission. Furthermore, light is emitted from the device without reabsorption because the
bandgap energy in the AlGaAs layer is large in comparison with that in GaAs. The DH structure
is therefore used to provide the most efficient incoherent sources for application within optical
fiber communications.

LED structures

There are six major types of LED structure and although only two have found extensive use in
optical fiber communications, two others have become increasingly applied. These are

the surface emitter,

the edge emitter,

the superluminescent and

the resonant cavity LED respectively.

The other two structures, the planar

and dome LEDs, find more application

as cheap plastic-encapsulated visible devices for use in such areas as intruder alarms, TV channel
changers and industrial counting. However, infrared versions of these devices have been used in
optical communications mainly with fiber bundles.

Planar LED
The planar LED is the simplest of the structures that are available and is fabricated by either
liquid- or vapor-phase epitaxial processes over the whole surface of a GaAs substrate. This
involves a p-type diffusion into the n-type substrate in order to create the junction. Forward
current flow through the junction gives Lambertian spontaneous emission and the device emits
light from all surfaces. However, only a limited amount of light escapes the structure due to total
internal reflection, and therefore the radiance is low.

Dome LED
A hemisphere of n-type GaAs is formed around a diffused p-type region. The diameter of the
dome is chosen to maximize the amount of internal emission reaching the surface within the
critical angle of the GaAs–air interface. Hence this device has a higher external power efficiency
than the planar LED. However, the geometry of the structure is such that the dome must be far
larger than the active recombination area, which gives a greater effective emission area and thus
reduces the radiance.

Surface emitter LEDs

A method for obtaining high radiance is to restrict the emission to a small active region within
the device. The technique involve homostructure devices to use an etched well in a GaAs
substrate in order to prevent
heavy absorption of the emitted radiation, and physically to accommodate the fiber. These
structures have a low thermal impedance in the active region allowing high current densities and
giving high-radiance emission into the optical fiber. Furthermore, considerable advantage may
be obtained by employing DH structures giving increased efficiency from electrical and optical
confinement as well as less absorption of the emitted radiation. This type of surface emitter LED
(SLED) has been widely employed within optical fiber communications.

The structure of a high-radiance etched well DH surface emitter* for the 0.8 to 0.9 µm
wavelength band

The internal absorption in this device is very low due to the larger bandgap-confining layers, and
the reflection coefficient at the back crystal face is high giving good forward radiance. The
emission from the active layer
is essentially isotropic, although the external emission distribution may be considered
Lambertian with a beam width of 120° due to refraction from a high to a low refractive index at
the GaAs–fiber interface. The power coupled Pc into a multimode step index fiber may be
estimated from the relationship [Ref. 12]:

Edge emitter LEDs

DH edge emitter LED (ELED). takes advantage of transparent guiding layers with a very thin
active layer (50 to 100 μm) in order that the light produced in the active layer spreads into the
transparent guiding layers, reducing
self-absorption in the active layer. The consequent waveguiding narrows the beam divergence to
a half-power width of around 30° in the plane perpendicular to the junction. However, the lack of
waveguiding in the plane of the junction gives a spontaneous (Lambertian) output with a half-
power width of around 120°. Most of the propagating light is emitted at one end face only due to
a reflector on the other end face and an antireflection coating on the emitting end face. The
effective radiance at the emitting end face can be very high giving an increased coupling
efficiency into small-NA fiber compared with the surface emitter. However, surface emitters
generally radiate more power into air (2.5 to 3 times) than edge emitters since the emitted light is
less affected by reabsorption and interfacial recombination.

The enhanced waveguiding of the edge emitter enables it to couple 7.5 times more power into
low-NA fiber than a comparable surface emitter. It has been found that lens coupling with edge
emitters may increase the coupling efficiencies by comparable factors (around five times).

The stripe geometry of the edge emitter allows very high carrier injection densities for given
drive currents. Thus it is possible to couple approaching a milliwatt of optical power into low-NA
(0.14) multimode step index fiber with edge-emitting LEDs operating at high drive currents (500
mA).
Edge emitters have also been found to have a substantially better modulation bandwidth of the
order of hundreds of megahertz than comparable surface-emitting structures with the same drive
level. In general it is possible to construct edge-emitting LEDs with a narrower linewidth than
surface emitters, but there are manufacturing problems
with the more complicated structure (including difficult heat-sinking geometry) which moderate
the benefits of these devices.
The ELED comprises a mesa structure with a width of 8 m and a length of 150 m for current
confinement. The tilted back facet of the
device was formed by chemical etching in order to suppress laser oscillation. The ELED active
layer was heavily doped with Zn to reduce the minority carrier lifetime and thus improve the
device modulation bandwidth. In this way a 3 dB modulation bandwidth of 600 MHz can be
obtained.

Superluminescent LEDs

Another device geometry which is providing significant benefits over both SLEDs and ELEDs
for communication applications is the superluminescent diode or SLD. This device type offers
advantages of:
(a) a high output power;
(b) a directional output beam; and
(c) a narrow spectral line width – all of which prove useful for coupling significant optical power
levels into optical fiber. Furthermore the super radiant emission process within the SLD tends to
increase the device modulation bandwidth over that of more conventional LEDs. The SLD
structure requires a p–n junction in the form of a long rectangular stripe. However, one end of
the device is made optically lossy to prevent reflections and thus suppress lasing, the output
being from the opposite end. For operation the injected current is increased until stimulated
emission, and hence amplification, occurs (i.e. the initial step towards laser action), but because
there is high loss at one end of the device, no optical feedback takes place. Therefore, although
there is amplification of the spontaneous emission, no laser oscillation builds up. However,
operation in the current region for stimulated emission provides gain causing the device output to
increase rapidly with increases in drive current due to what is effectively single-pass
amplification. High optical output power can therefore be obtained, together with a narrowing of
the spectral width which also results from the stimulated emission.

The GaAsP/InP SLD device which emits at 1.3 μm comprises a buried active layer within a V-
shaped groove on the p-type InP substrate. This technique provides an appropriate structure for
high-power operation because of its low leakage current. Unlike the aforementioned SLD
structures which incorporate AR coatings on both end facets to prevent feedback, a light
diffusion surface is placed within this device. The surface, which is applied diagonally on the
active layer of length 350 μm, serves to scatter the backward light emitted from the active layer
and thus decreases feedback into this layer. In addition, an AR coating is provided on the output
facet. As it is not possible to achieve a perfect AR coating, the above structure is therefore not
left totally dependent on this feedback suppression mechanism.

Resonant cavity and quantum-dot LEDs

The resonant cavity light-emitting diode (RC-LED) is based on planar technology containing a
Fabry–Pérot active resonant cavity between distributed Bragg reflector (DBR) mirrors. A
quantum well is then embedded in this active cavity. Since the cavity is confined to a micrometer
size, the RC-LED is therefore also referred to as a microcavity light-emitting diode. The basic
structure for an RC-LED consist of an active region consisting of InGaAsP multiquantum wells
is positioned in the optical resonant cavity which is located between two DBR mirrors, one each
at the bottom and the top of the active cavity.
Current confinement is obtained through the ion implantation technique in the top mirror while
the RC-LED structure constitutes a Fabry–Pérot resonator where the optical cavity mode is in
resonance amplifying the spontaneous emission from the active layer. The reflectivity of the
bottom DBR mirror is kept to a maximum (i.e.
higher than 90%) by incorporating a large number of gratings (i.e. more than 40) whereas the
surface DBR mirror is made semitransparent by introducing fewer gratings (i.e. about 15)
creating low facet reflectivity (i.e. 40 to 60%) to allow the optical signal to exit through this
mirror. Since these devices incorporate DBR mirrors they may also be referred to as grating-
assisted RC-LEDs.
In this context the structure is similar to that of a vertical cavity surface-emitting laser (VCSEL)
excepting that the emitting side of the DBR mirror of the resonant cavity is emitransparent.
Therefore light is emitted as a result of resonantly amplified spontaneous emission and
stimulated emission does not occur. The operation of the device is, however, similar to the
VCSEL exhibiting low facet reflectivity on the top mirror and without threshold limitations.
Based on the cavity design, RC-LEDs can be constructed to emit from either the bottom or the
surface of the device structure. Although they can be fabricated for longer wavelength operation
at both 1.3 μm and 1.55 μm, RC-LEDs are generally fabricated for operation over a range of
wavelengths between 0.85 and 0.88 μm and also at 0.65 μm for use with plastic optical fiber.
Although the growth process for
RC-LEDs is more complex than for conventional devices, their enhanced features, such as the
highly directional circular output beam and improved fiber coupling efficiency, make
overcoming the fabrication problems worthwhile. External quantum efficiency of the RCLED is,
however, reduced to around 6 to 10% when operating at a wavelength of 1.3 μm or 1.55 μm due
to the increased linewidth broadening at these longer wavelengths.
Nevertheless, even this value of external quantum efficiency proves sufficient to provide
comprising a layer of InAs quantum dots covered by InGaAs is positioned at a distance from a
gold-coated mirror on the device surface. The active region comprises a single layer of quantum
dots while a AlGaAs layer is grown between the GaAs substrate and the active region in order to
confine the injected carriers. To enhance output signal power, the quantum-dot layer is
positioned at half the emission wavelength distance from the surface mirror. The optical signal
reflected by the mirror therefore constructively interferes with the radiation emitted downwards
from the active layer resulting in a fourfold increase in optical signal power being collected from
the substrate side. QD-LEDs with 10 mW output power when operating at wavelengths of 1.30
μm and 1.55 μm have also
been successfully fabricated. QD-LEDs based on a resonant cavity with external quantum
efficiency of greater than
20% have also been demonstrated. For nonresonant cavity QD-LEDs increased quantum
efficiency can be obtained by introducing thin active layers at the surface of the LED. Such
devices are referred to as surface-textured thin-film LEDs.
In this structure an optical signal which suffers total internal reflection is scattered internally by
the textured top surface and hence changes its angle of propagation. After reflection from the
back reflector the optical signal can be coupled to the output of the LED. An external quantum
efficiency of 29% at higher transmission rates of 1 Gbit s1 has been demonstrated with this
structure and it was further improved to 40% when incorporating an optical lens on the top of the
device.

LED characteristics

Optical output power

The ideal light output power against current characteristic for an LED (depicted for an isotropic
device) is shown in the figure. It is linear corresponding to the linear part of the injection laser
optical power output characteristic before lasing occurs.Intrinsically the LED is a very linear
device and hence it tends to be more suitable for analog transmission where severe constraints
are put on the linearity of the optical source. However, in practice LEDs do exhibit significant
nonlinearities which depend upon the configuration utilized. It is therefore often necessary to use
some form of linearizing circuit technique (e.g. predistortion linearization or negative feedback)
in order to ensure the linear performance of the device to allow its use in high-quality analog
transmission systems. It may be noted that the surface emitter radiates significantly more optical
power into air than the edge emitter, and that both devices are reasonably linear at moderate
drive currents.
In a similar manner to the injection laser, the internal quantum efficiency of LEDs
decreases exponentially with increasing temperature. Hence the light emitted from these devices
decreases as the p–n junction temperature increases. It may be observed that the edge-emitting
device exhibits a greater temperature dependence than the surface emitter and that the output of
the SLD with its stimulated emission is strongly dependent on the junction temperature. This
latter factor is further emphasized in the light output against current characteristics for a
superluminescent LED displayed. These characteristics show the variation in output power at a
specific drive current over the temperature range 0 to 40 °C for a ridge waveguide device
providing lateral current confinement. The nonlinear nature of the output characteristic typical of
SLDs can also be observed with a knee becoming apparent at an operating
temperature around 20 °C. Hence to utilize the high-power potential of such devices at
elevated temperatures, the use of thermoelectric coolers may be necessary.
It should also be noted that resonant cavity LEDs have shown a similar reduction in
output power when operated at higher temperatures.

Output spectrum

The spectral linewidth of an LED operating at room temperature in the 0.8 to 0.9 μm
wavelength band is usually between 25 and 40 nm at the half maximum intensity points
(full width at half power (FWHP) points). For materials with smaller bandgap energies
operating in the 1.1 to 1.7 μm wavelength region the linewidth tends to increase to around
50 to 160 nm. This becomes apparent in the differences in the output spectra between surface-
and edge-emitting LEDs where the devices have generally heavily doped and lightly doped (or
undoped) active layers respectively. It may also be noted that there is a shift to lower peak
emission wavelength
(i.e. higher energy) through reduction in doping in and hence the active
layer composition must be adjusted if the same center wavelength is to be maintained
The differences in the output spectra between InGaAsP SLEDs and ELEDs caused
by self-absorption along the active layer of the devices are displayed in Figure. It
may be observed that the FWHP points are around 1.6 times smaller for the ELED than
the SLED. In addition, the spectra of the ELED may be further narrowed by
the superluminescent operation due to the onset of stimulated gain and in this case the
linewidth can be far smaller (e.g. 30 nm) than that obtained with the SLED.

The output spectra also tend to broaden at a rate of between 0.1 and 0.3 nm °C-1 with
increase in temperature due to the greater energy spread in carrier distributions at higher
temperatures. Increases in temperature of the junction affect the peak emission wavelength as
well, and it is shifted by +0.3 to 0.4 nm °C-1 for AlGaAs devices and
by +0.6 nm°C-1 for InGaAsP devices.

Modulation bandwidth

The modulation bandwidth in optical communications may be defined as the electrical 3 dB

point or the frequency at which the output electric power is reduced by 3 dB with respect to the
input electric power. As optical sources operate down to d.c. level the high-frequency
3 dB point have to be considered the modulation bandwidth being the frequency range between
zero and this high-frequency 3 dB point. The optical bandwidth is significantly greater than the
electrical bandwidth. The difference between them (in frequency terms) depends on the shape of
the frequency
response for the system. However, if the system response is assumed to be Gaussian, then
the optical bandwidth is a factor of √2 greater than the electrical bandwidth.
The modulation bandwidth of LEDs is generally determined by three mechanisms.
These are:

(a) the doping level in the active layer;

(b) the reduction in radiative lifetime due to the injected carriers;
(c) the parasitic capacitance of the device. Assuming negligible parasitic capacitance, the
speed at which an LED can be directly current modulated is fundamentally limited by the
recombination lifetime of the carriers, where the optical output power Pe() of the device
(with constant peak current) and angular
modulation frequency is given by
where  is the injected (minority) carrier lifetime in the recombination region and Pdc is
the d.c. optical output power for the same drive current from spontaneous recombination rather
than stimulated emission, coupled with the increased numbers of nonradiative centers at higher
doping levels. Thus at high modulation bandwidths the optical output power from conventional
LED structures decreases
The reciprocal relationship between modulation bandwidth and output power may be

Reliability

LEDs are not generally affected by the catastrophic degradation mechanisms which can
severely affect injection lasers. Early or infant failures do, however,
occur as a result of random and not always preventable fabricational defects. Such failures
can usually be removed from the LED batch population over an initial burn-in operational
period. In addition, LEDs do exhibit gradual degradation which may take the
form of a rapid degradation mode* or a slow degradation mode.
Rapid degradation in LEDs is similar to that in injection lasers, and is due to both the
growth of dislocations and precipitate-type defects in the active region giving rise to
dark line defects (DLDs) and dark spot defects (DSDs), respectively, under device aging
[Ref. 69]. DLDs tend to be the dominant cause of rapid degradation in GaAs-based LEDs.
The growth of these defects does not depend upon substrate orientation but on the injection
current density, the temperature and the impurity concentration in the active layer.
Good GaAs substrates have dislocation densities around 5 × 10-4 cm-2. Hence, there is
less probability of dislocations in devices with small active regions. DSDs, and the glide
of existing misfit dislocations, however, predominate as the cause of rapid degradation in
InP-based LEDs