UNIT-III

Fiber Optical sources and coupling

Basic semiconductor properties:
Introduction to sources:
The principal light sources used for optical communications applications are
heterojunction-structured semiconductor LASER diodes or injected laser diodes (ILDs) and
LEDs. A heterojunction consists of two adjoining semiconductor materials with different
band gap energies. These are suitable devices for optical transmission systems because
a) They have adequate output power for a wide range of applications.
b) Their optical power output can be directly modulated by varying the input current to
the device
c) They have high efficiency and
d) Their dimensional characteristics are compatible with those of optical fiber.
The light emitting region of both LEDs and LASER diodes consists of a pn junction
constructed of direct band gap III-V semiconductor materials. When this junction is
forward biased, electrons and holes are injected into the p and n regions respectively.
These injected majority carriers can recombine either radiatively (by emitting photon of
energy h) or non-radiatively (where upon, the recombination energy is dissipated in the
form of heat). This pn junction is known as the active or recombination region.
A major difference between LEDs and laser diodes is that the optical output from an
LED is incoherent, where as that from laser diode is coherent. In coherent sources, the
optical energy is produced in an optical resonant cavity. The optical power released from
the cavity has spatial and temporal coherence, meaning that the released energy is highly
monochromatic and the output beam is very directional. The output radiation has a broad
spectral width, since the mitted photon energies range over the energy distribution of the
recombining electrons and holes. An incoherent energy is emitted into a hemisphere
according to a cosine power distribution and thus has a large beam divergence.
To a large extent these two sources fulfill the major requirements for an optical fiber
emitter which are listed below:
a) A size and configuration compatible with launching light into an optical fiber. Ideally,
the output should be highly directional.
b) Must accurately track the electrical input signal to minimize distortion and noise.
Ideally, the source should be linear.
c) Should emit light at wavelengths where the fiber has low losses and low dispersion
ad where the detectors are efficient.
d) Preferably capable of simple signal modulation over a wide bandwidth extending
from audio frequencies to beyond the GHz range.
e) Must couple sufficient optical power to overcome attenuation in the fiber plus
additional connector losses and leave adequate power to drive the detector.
f) Should have a very narrow spectral width in order to minimize dispersion in the
fiber.
g) Must be capable of maintaining a stable optical output which is largely unaffected
by changes in ambient conditions (e.g.: temperature).
h) It is essential that the source is comparatively cheap and highly reliable in order to
compete with conventional transmission techniques.

The spatially directed coherent optical output from laser diode can be coupled into
both single mode and multimode fibers. In general LEDs are used with multimode fibers
because its incoherent optical power from an LED can be coupled in sufficient quantities
to be useful. LEDs have been employed in high speed local area applications in which
several wavelengths can be transmitted on the same fiber. Here, the broad spectral
emission of LED is splitted into narrow spectral slices by using waveguide grating array.
This technique is called spectral slicing. These slices are each centered at a different
wavelength, they can be individually modulated extremely with independent data streams
and simultaneously sent on the same fiber.
Some of the semiconductor properties are energy gaps, intrinsic and extrinsic
materials, pn junctions, and direct and indirect band gaps.
Semiconductor is a material which has the conductivity lieing between that of metals
and insulators. The basic semiconducting materials are SI and Ge. A silicion (Si) atom has
four electrons I its outer shell by which it makes covalent bonds with its neighbouring
atoms in a crystal. Si is located in IV group of periodic table.
Energy bands: The conductance properties are well explained with the aid of energy
band diagrams as shown in figure below.

In a pure crystal at low temperatures, the conduction band is completely empty of

electrons and the valence band is completely full. These two bands are separated by an
energy gap, or bad gap, where no energy levels are present. As the temperature is
increased some electrons are thermally excited across the band gap. For Si, this excitation
energy must be >1.1eV, which is bad gap energy. This increases the concentration of
electrons in conduction band and leaves equal concentration p of holes in valence band
as shown in figure (b). Both electrons and holes are mobile within the material, so that
both contribute to electrical conductivity, i.e., an electron in the valence band can move
into a vacant hole. This action makes the hole move in the opposite direction to that of
electron flow.
The concentration of electrons and holes is known as intrinsic concentration (i). For
perfect material with no impurities it is given by
n = p = i = K .

where constant K = 2

2 k BT
2
h

Eg
2Tk B

3 /2

. (memh)3/4 is characteristic of material.

T is the temperature in 0K, kB is Boltzmanns constant, h is Planks constant, me and

mh are effective masses of electron and hole respectively.
The conduction can be increased by adding impurities from group V elements (P, As,
Sb). This process is called doping and the doped semiconductor is called extrinsic
material. P, As, Sb have five electrons in outer shell. When they replace a Si atom, four are
used for covalent bonding and the fifth is loosely bound electron and available for
conduction.

As shown in above figure, this gives rise to an occupied level, just below the
conduction band called donor level. The impurities are called donors. This increases the
free electron concentration in conduction band as shown in above figure (b). I this
material current is caused by electrons, so it is called n-type material.
The conduction can also be increased by adding group III elements (B, I, Ga) which
have three electrons in outer shell. The three electrons make covalent bonds and a hole
with same properties of donor electron is created. This gives rise to an unoccupied level
just above the valence band as shown in figure below. Conduction occurs when electrons
are excited from the valence bad to this acceptor level (so called because the impurity
atoms leave accepted electros from valence band). Correspondingly, the free hole
concentration increases in the valence band as shown in figure (b) below. This is called ptype material because the conduction is a result of flow of holes.

p-n junction potential:

It is necessary to use both p and n materials to make semiconductor devices. The

junction between the two material regions is known as the pn junction. When a junction is
created, the majority carriers that is electrons on n-side and holes on p-side diffuse across
the junction. This causes holes to appear on the n-side and electrons on the p-side. As a
result, a electric field appears across the junction as shown in figure. This field prevents
further diffusion of majority carriers across the junction and the junction is depleted of
mobile charge carriers once equilibrium has been established. This region is called either
the depletion region or space charge region.
Forward bias of pn junction:
When the positive terminal of an external battery is connected to the p-type
material and the negative terminal to the n-type material, the junction is said to be
forward biased. Once the pn junction is forward-biased the magnitude of the barrier
potential is reduced. The electrons on n side and holes on p side are allowed to diffuse
across the junction.

Reverse bias of pn junction:

When the positive terminal of an external battery is connected to the n-type
material and the negative terminal to the n-type material, the junction is said to be
reverse biased. As a result of the reverse bias, the width of the depletion region will
increase on both p and n sides. This effectively increases the barrier potential and
prevents the flow of majority carriers across the junction and minority carriers can move
with the field across the junction.

Direct and indirect band gaps:

Semiconductors are classified as either direct or indirect band gaps based on the
h
shape of band gap as a function of momentum K (K = C ). For direct semiconductor
materials, both electrons and holes have the same momentum and so the photon radiated
is of h energy. In the indirect band gap the momentums are different and so the energy
radiated is increased or decreased by indirect band gap energy (Eind).

Advantages of LEDs over LASER:

a) Simpler fabrication: there are no mirror facets and in some structures no striped
geometry.
b) Cost: the simpler construction of LED leads to reduced cost which is always likely to
be maintained.
c) Reliability: the LED does not exhibit catastrophic degradation and has proved far
less sensitive to gradual degradation than the junction laser. It is also minimum to
self pulsation and modal noise problems.
d) Generally less temperature dependence: the light output against current
characteristic is less affected by temperature than the corresponding characteristic
for the injection laser. And also the LED is not a threshold device and therefore
raising the temperature does not increase the threshold current above the operating
point and hence halt operation.
e) Simpler drive circuitry: this is due to generally lower drive currents and reduced
temperature dependence which makes temperature compensation circuitry
unnecessary.
f) Linearity: ideally, the LED has a linear light output against current characteristic,
unlike the injection laser. This can prove advantageous where analog modulation is
concerned.

These advantages combined with the development of high radiance output

relatively high radiance bandwidth devices have ensured that the LED remains an
extensively used source for optical fiber communications.
The light source materials for shorter wavelength region are GaAs/GaAlAs and for
longer wavelength region are InGaAsP/InP.
LED Structures: LEDs are the best light source choice for optical communication
systems that require bit rates less than approximately 100-200 MBPS, with multimode
fiber-coupled optical power in terms of W. LEDs requires less complex drive circuitry than
LASER diodes since no thermal or optical stabilization circuits are needed and their
fabrication cost is less with higher yields.
Requirements of LED source:
a) It must have high radiance output: Radiance is a measure of the optical power
radiated into a unit solid angle per unit area of emitting surface in watts.
b) Fast emission response time: It is the time delay between the application of a
current pulse and the onset of optical emission.
c) High quantum efficiency: The quantum efficiency is related to the fraction of
injected electron-hole pairs that recombine radiatively.

Quantum efficiency =

due
No. of photonsemitted / sec recombination

No . of injected carriers

To achieve high radiance output 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 pn junction. Optical confinement is for preventing absorption of the
emitted radiation by the material surrounding pn junction.
To achieve carrier and optical confinement, there are two types of LED structures,
homojunctional and single and double heterojunctional configuration. Generally, double
heterojunctional configuration is most preferable, because it has two different alloy layers
o each side of the active region. By means of this sandwich structure of differently
composed alloy layers, both the carriers and the optical field are confined in the central
active layer. The band gap differences of adjacent layers confine the charge carriers, and
the differences in the indices of refraction of adjoining layers confine the optical field to
the central active layer. This dual configuration leads to both high efficiency and radiance.
The figure below shows the compositions of LED configuration and injected carriers and
refractive index differences.

The device performance can also be influenced by optical absorption I the active
region (self-absorption), carrier recombination at the hetero junctions, doping
concentrations of the active layer, injection carrier density and the thickness of active
layer.
LED Configurations:
Two basic configurations that can be used for fiber optics are:
a) Surface emitted LED (SLED)
b) Edge emitted LED (ELED)
Surface emitted LED (SLED):
In SLED, the plane of active light-emitting region is oriented perpendicularly to the
axis of the fiber as shown in figure below:

Circular etched well: It is etched through the substrate of the device, to accommodate a
fiber core in order to accept the emitted light. It is normally 50m in diameter and upto
2.5m thick.

Confinement layers: To increase coupling efficiency and the radiance output to be

maximum.
Circular active region: It provides low thermal impedance, and so increases the current
density to maximum.
Epoxy resin: It decreases the refractive mismatch between confinement layers.
Active layer is used to increase the radiance output.
Radiation/emission pattern: It is an isotropic pattern with a 120 0 half-power beam
width. This isotropic pattern from surface emitter is called a Lambertian pattern. In this
pattern, the source is equally bright from all directions, but the power diminishes as
cos where is the angle between the viewing direction and the normal to the surface.
This decrease is caused, because the projectional area decreases as cos. Thus the power
is down to 50% of its peak at =600, so that the total half-power beam width is 1200. Thus
the power launched into the fiber by heterojunction is about 60% to 80%. This is shown in
figure below:

The radiation is maximum at core axis, and some of the power is lost and some of
the power is reflected. Therefore internally generated power is different from externally
generated power.
Edge emitted LED (ELED):
It consists of an active junction region and two guiding layers. The refractive index
of the two guiding layers is lower than that of active region, but higher than the index of
the surrounding material. This structure forms a waveguide channel that directs the
optical radiation towards the fiber core. To match the fiber core diameters (50-100m),
the contact stripes for the edge emitter LED are 50-70m wide. The length of the active
region varies from 100 to 150m. the radiation patter of the edge emitter is more
directional than that of the surface emitter.

In the plane parallel to the junction, where there is no effect of waveguide, the
emitted beam is lambertian with half power beam width of II = 1200. In the plane
perpendicular to the junction, the half power beam width has been made as small as 25300 by proper choice of waveguide thickness.
Advantages of ELED over SLED:
a) Coupling efficiency (c) is increased.
b) Its modulation band width (i.e., -band and spectral line width) is larger; hence it is
a highly divergent source.
c) The ELEDs can be fabricated using InGaAsP/InP in order to be operated at a
wavelength of 1.3m, where losses are minimum.
d) The stripe geometry of ELED reduced internal absorption of carriers and increases
external power efficiency.
Light source materials:
In direct-band-gap material electrons and holes can recombine directly across the
band gap without needing a third particle to conserve momentum. Only in direct-bandgap material radiative recombination is sufficiently high to produce an adequate level of
optical emission. Many binary compounds made from compounds of a group III element
and that of a group V element are direct band gap materials. Various ternary and
quaternary combinations of binary compounds of these elements are also direct-gap
materials.
For operation in the 800-to-900 m spectrum, the principal material used is the
ternary alloy Ga1-xAlxAs. The ratio x of aluminum arsenide to gallium arsenide determines
the band gap of the alloy and correspondingly, the wavelength of the peak emitted
radiation. An example of the emission spectrum of a Ga 1-xAlxAs LED with x = 0.08 is shown
in figure (b).

The width of the spectral pattern at its half-power point is known as the full-width
half-minimum (FWHM) spectral width.
At longer wavelengths the quaternary alloy In 1-xGaxAsyP1-y is one of the prinicipal
material. By varying the mole fractions x and y in the active area, LEDs with peak output
powers at any wavelength between 1.0 and 1.7m can be constructed.
The alloys GaAlAs and InGaAsP are chosen to make semiconductor light sources
because it is possible to match the lattice parameters of the heterostructure interfaces by
using a proper combination of binary, ternary, and quaternary materials.

The matching of the crystal lattice parameters of the two adjoining heterojunctions
is required to reduce interfacial defects and to minimize strains in the device as the
termperature varies. These factors directly affect the radiative efficiency and lifetime of a
light source.
The quantum-mechanical relationship between energy E and frequency is
E = h = hc/
the peak emission wavelength in m can be expressed as a function of the bad-gap
energy Eg in electron volts by the equation
(m) =

1.240
E g ( eV )

The below table lists the bandgap energies of some device materials
Semiconductor
materials
Silicon (Si)
GaAs
Germanium (Ge)
InP
Ga0.95Al0.03As

Bandgap
energy (eV)
1.12
1.43
0.67
1.35
1.51

A heterojunction with matching lattice parameters is created by choosing two

material compositions that have the same lattice constant but different band-gap
energies.
In ternary alloys the energy gap in electron volts for values of x between 0 and 0.37
(the direct-band-gap region) can be found from the empirical equation
Eg = 1.424 + 1.266x + 0.266x2
The band-gap energy and lattice-constant range for the quarternary alloy InGaAsP
are much larger as shown in figure below.

These materials are generally grown on an InP substrate, so that lattice-matched

configurations are obtained by selecting a compositional point along the top dashed line I
the above figure, which passes through the InP point. Along this line, the compositional

parameters x and y follow the relationship y=2.20x with 0 x 0.47. For In 1-xGaxAsyP1-y
compositions that are lattice matched to InP, the band gap in eV varies as
Eg = 1.35 0.72y + 0.12y2
The typical spectral patterns for edge-emitting and surface-emitting LEDs at
1310nm are shown in figure below:

The table below lists the typical characteristics of SLED and ELED.
LED type

Material

Waveleng
th (nm)

Operating
current
(mA)

SLED
SLED
ELED

GaAlAs
InGaAsP
InGaAsP

850
1310
1310

110
100
110

Fibercoupled
power
(W)
40
15
30

Nominal
FWHM
(nm)
35
80
150

Quantum efficiency and LED power:

Equal densities of electrons and holes are injected for charge neutrality in the
crystal. When carrier injection stops the carrier density returns to equilibrium value. The
excess carrier density decays exponentially with time according to the relation
n = n0 e-t/

(1)

where n0 is initially injected excess electron density and time constant is carrier
lifetime. The value of ranges between ms to fractions of ns depending on the material
composition and device defects.
The excess carriers can recombine either radiatively or non-radiatively. Nonradiative recombination effects include optical absorption in the active region, carrier
recombination at the heterojunction interfaces and the energy released during an
electron-hole recombination is transferred to another carrier in the form of kinetic energy.
When there is a constant current flow into an LED, an equilibrium condition is
established. The total rate at which carriers are generated is the sum of the externally
supplied and the thermally generated rates.
The externally generated rate =

J
qd

(2)
The thermal generation rate =

(3)
where J is current density in A/cm 2 and d is the thickness of recombination region.
The rate equation for recombination in LED is written as

dn
dt

J
qd

(4)

The equilibrium condition is found by setting (4) equal to zero.

dn
dt

n=

=0=

J
qd

J
qd

(5)

Equation (5) gives the steady state electron density in the active region when a
constant current is flowing through it.
Internal quantum efficiency (int):
The internal quantum efficiency in the active region is the fraction of the electron
hole pairs that recombine radiatively. If the radiative recombination rate is R r and nonradiative recombination is Rnr, then the internal quantum efficiency int is the ratio of the
radiative recombination rate to the total recombination rate
int

Rr
R r + Rnr

(6)
Therefore the radiative recombination life time, for exponential decay of excess
carriers is
r

n
Rr

(7)
and the non-radiative recombination life time is
nr =

n
R nr

(8)
1
int =

1+

r
nr

(9)

where is the bulk recombination lifetime, given by

1

(10)

1
r

1
nr

In the direct bandgap semiconductors R r and Rnr are equal so that the int is about
50% for simple heterojunction LED. For double heterojunction LEDs it ranges between 60
and 80%.
If the current injected into the LED is I, then the number of recombinations per
second is equal to Rr + Rnr = I/q
(11)
Substitute (11) in (6)
I
Rr =

(12)

As Rr is the total number of photons generated per second and that each photon has
an energy h, the optical power generated internal to the LED is

Pint

int

hcI

I
q

(13)
External quantum efficiency (int):
As all the internally generated photons will not exit the device, to find emitted
power, ext is needed which is defined as the ratio of the photons emitted from the LED to
the number of internally generated photons. To calculate ext number of reflections at the
surface of LED is to be considered. At the interface of a material boundary only that
fraction of light falling within a core defined by the critical angle =/2 - will cross
c

the interface. We have

=sin-1(n2/n1). Here n1 is the refractive index of semiconductor

material and n2 is the refractive index of outside material, normally air.

The quantum efficiency is calculated by

c

ext

1
T ( ) ( 2 sin ) .d
4
0

(14)
where T() is the Fresnel transmission coefficient or Fresnel transmittivity and it depends
on incidence angle .

For

simplicity

for

normal

incidence,

T(0)

4 n1 n 2

(n1 +n2 )2

T(0)

4n
2
(n+1)

(15)
Assuming

outside

medium

is

air

and

letting

n1

n,

(16)
Then

ext

1
n(n+1)2

(17)
Therefore, the optical power emitted from LED is
P
P = ext . Pint =

n(n+1)

(18)

Modulation of an LED:
The frequency response of an LED is determined by three factors:
a) The doping level in the active region
b) The injected carrier lifetime i in recombination region and
c) Parasitic capacitance of LED.
If the drive current is modulated at a frequency , the optical output power of device
varies as
P() = P0 [1+(i)2]-1/2
where P0 is the power emitted at zero modulation frequency.
The parasitic capacitance causes a delay of carrier injection into the active junction
and so the optical output is delayed. This delay is negligible if a small constant forward
bias is applied to the diode. Under this condition the above equation is valid and the
modulation response is limited only by the carrier recombination time.
The modulation bandwidth of an LED can be defined in terms of electrical or optical.
Electrical bandwidth is actually found via the associated electrical circuitry. Thus, the
modulation bandwidth is defined as the point where the electrical signal power designated
P(), has dropped to half of its constant value resulting from the modulated portion of
optical signal. This is the electrical 3-dB point i.e., the frequency at which output power is
reduced by 3-dB with respect to input electrical power as shown in figure below.

As optical source exhibits a linear relation between light power and current we have
P() = I2()/R.
The ratio of the output electrical power at the frequency to the power at zero
modulation is given as
Ratioelec

P()
= 10 log ( P(0) ) = 10 log

I 2 ()
I 2 (0)

( )

where I() is the electrical current in the detection circuitry

2

The electrical 3-dB occurs at frequency where P() = P(0)/2. It happens when,

=1/2 which implies

I ()
I (0)

I ()
I 2(0)

= 0.707.

Optical bandwidth: It is the frequency where optical power P() = P0/2. The 3-dB
frequency is determined from the ratio of the optical power at frequency to the
unmodulated value of the optical power. Since, the detected current is directly
proportional to the optical power, this ratio is
Ratioopt = 10 log

( II()
(0) )

The optical 3-dB point occurs at frequency where the ratio of currents is 1/2.
LED transient response:
It is known from the device rise time. The rise time is the time required for the
device to make transition from 10% of its final light output value to 90% of its final value.
Rise time deter mines the maximum modulation rate of the device. If the fall time of the
LED is negligible compared to its rise time, the maximum modulation rate is reciprocal of
the rise time. Rise time is limited by several factors. One main factor which limits rise time
is space charge capacitance Cs in the active region. The rise time for the LED is given by
tr = 2.2

2 kT C s
+
qIp

where k Boltazmanns constant, T- temperature of the device, C s space charge

capacitance Ip size of drive current step function, q electron charge.

Therefore the rise time is decreased by decreasing Cs and increasing Ip.

Power band width product:
For LED it is given by
P(f) =

hcJ
2 q r

where P is the peak optical power, h is the Planks constant, q is the electron charge, is
the peak emission wavelength, f is bandwidth, c is the velocity of light, J is the injected
carrier current density, r is the radiative recombination life time.
Disadvantages of LED:
a) Quantum efficiency is low.
b) Damages due to high voltage and over current.
Applications:
a)
b)
c)
d)
e)

In various displays.
As indicator lamps
As indicators in measuring equipment
Sources in opto-couplers and
In IR remote controls.

LASER diodes:
For high performance (bit rates) applications LED is not applicable due to its broad
spectrum of emitted light. This problem can be overcome by using LASER diodes, though
it is inadequate for many applications. The LASER diodes have sharp emission line with a
line width of about two orders of magnitude narrower than that of LEDs. Its modulation
bandwidth is above 50GHz. It has superior spatial coherence. So a LASER beam does not
spread much as the beams of other sources. Thus they can be focused to give a very high
intensity.
LASER operation:
LASER is an acronym for light amplification by stimulated emission of radiation. Its
light output is highly directional, coherent and monochromatic. The first LASER was by
Maiman in 1961. Later on several improvements are made to it to have various types of
LASERs like Ruby laser, semiconductor lasers, CO2 laser, power lasers etc.
The principle of operation/ Basic steps in production of LASER beam:
Laser action is a result of three processes. They are
a) Absorption
b) Spontaneous emission and
c) Stimulated emission

Absorption: It can be explained by considering two-energy levels: ground state energy

level (E1) and excited state (E2). Usually, under normal conditions the electrons stay in the
lower energy level (say E 1). When a photon of certain energy is incident on this electron, it
absorbs that energy and its energy is increased and is shifted to the higher energy level
(say E2). This transfer of electron is purely due to the increase of its energy. The incident
photon should have at least the energy (E2-E1).
Spontaneous emission: The excited state of an electron is unstable, it shortly returns to
the ground state by emitting a photo of energy h. It occurs without any external
stimulation ad so it is called spontaneous emission. These are isotropic emissions and of
random phase, thus appear as a narrow band Gaussian output.

Stimulated emission: It is the key process for laser beam generation as it involves in
the amplification of light. An excited electron falls back to the ground state due to
spontaneous emission. If the incident photon is allowed to strike this excited electron
before the spontaneous emission takes place, there is no chance for this electron to go to
the higher energy level. Then it emits another photon. So, at the output two photons are
obtained, one which strikes the excited electron and the other coming out because of
excited electron leaving the higher level. The emitted photon is in phase with the incident
photon and travels in the same direction. This is known as stimulated emission. To cause
this amplification, the necessary conditions are:
a) The incident photons should strike the excited electrons before the occurrence of
spontaneous emission and
b) The number of excited electrons should be more than that of electrons in the ground
level in the total structure.

Population inversion: The process in which the number of excited electrons is made
larger than that of the electrons at ground level is called population inversion.

To (increase the speed of) get easier stimulated emission the time period for the
electrons in the excited state should be made longer. Thus the number of excited
electrons is more in stimulated emission which causes amplification. This amplified light is
called LASER beam which is coherent and monochromatic in nature.
Advantages of LASERs:
a)
b)
c)
d)
e)
f)

Laser light is coherent and monochromatic.

It has narrow spectral bandwidth (2nm or less), good directionality and stability.
It has high quantum efficiency, high modulation rate.
Divergence of its light output is small.
Its light can be modulated externally or it has ability of direct modulation and
Response time < 1ns.

LASER diode types:

They are capable of coupling several tens of mWs of useful optical power into fibers
with small core diameters. Usually, LASER diodes are multilayered heterojunction devices.
Same as in LED, double heterojunction confines optical output and carrier injection. The
current confinement in a LASER diode is in a small lasing cavity. It requires complicated
construction in LASER structure.
Stimulated emission in LASER takes place from optical transitions between
distributions of energy states in the valence and conduction bands. The radiation in the
laser diode is generated within a Fabry-Perot resonator cavity. This cavity is much smaller,
approximately 250-500m long, 5-15m wide and 0.1-0.2m thick. Generally these
dimensions are referred as longitudinal, lateral ad transverse dimensions of cavity.
The Fabry-Perot resonator is a pair of flat reflecting mirrors directed toward each
other to enclose the cavity. The mirror facets are constructed by making two parallel clefts
along natural cleavage planes of the semiconductor crystal. The purpose of these mirrors
is to provide strong optical feedback in the longitudinal direction. With this feedback the
device is converted into an oscillator with a gain mechanism that compensates optical
losses in the cavity. The device oscillates at the resonant frequencies of the laser cavity,
at which the gain is sufficient to overcome losses. The sides of cavity are formed simply,
by roughening the edges of the device to reduce unwanted emissions in these directions.

Distributed feedback laser (DFB laser):

In DFB laser, the cleaved facets are not required for optical feedback. The
fabrication of this device is similar to that of Fabry-Perot type, except that the lasing
action s obtained from Bragg reflectors (gratings) or periodic variations of the refractive
index called distributed feedback configurations. They are incorporated into the
multiplayer structure along the length of the diode. The full optical output is needed at the
front facet of the laser i.e., the one to be aligned with an optical fiber. To reduce optical
losses, a dielectric reflector can be deposited on the rear laser facet and also with this
dielectric reflector the threshold current density (the point at which lasing starts) is
reduced and the external quantum efficiency is increased. With a 6 layer reflector, the
reflectivities greater than 98% have been achieved.

Cavity/LASER modes and threshold conditions:

The optical radiation within the resonance cavity of a laser diode sets up a pattern
of electric and magnetic field lines called modes of cavity. These are divided basically into
TE and TM modes. Each set of modes are described interms of the longitudinal, lateral ad
transverse half-sinusoidal variations of EM fields along major axes of cavity.
Longitudinal modes: They are related to the length L of the cavity. They determine the
main structure of the frequency spectrum of the emitted optical radiation. L is much larger
than the lasing wavelength. So many longitudinal modes exist.
Lateral modes: They lie in the plane of pn junction. These depend on the width of the
cavity ad determine the shape of the lateral profile and beam.
Transverse modes: They are associated with EM field ad beam profile. They lie in the
plane perpendicular to the pn junction. They determine radiation pattern and the
threshold current density of laser.

Lasing conditions and resonant frequencies can be determined by electromagnetic

wave propagation in the longitudinal direction (along axis normal to the mirrors) in terms
of electric field phasor.
E(z,t) =I(z) ej(t-z)
where I(z) is optical field intensity, is optical radian frequency and is propagation
constant.
Lasing is a condition at which light amplification becomes possible in laser diode.
The requirement for lasing is population inversion be achieved. This conditions is clear by
the relation below
I(z) =I(0) exp[{g(h) (h)}z]

(1)

where I(0) is optical intensity at z=0, is effective absorption coefficient of material, g is

gain coefficient in Fabry-Perot cavity, is optical field confinement factor.
The equation (1) also shows that the stimulated emission rate into a given mode is
proportional to the intensity of radiation in that mode. The radiation intensity at a photon
energy h varies exponentially with distance z, that it traverses along the lasing cavity
according to the relation in equation (1).
Lasing occurs when the gain of one or several guided modes is sufficient to exceed
the optical loss during one round trip through the cavity i.e., z=2L. During this round trip,
only the fractions R1 and R2 of optical radiation are reflected from the two ends 1 and 2
respectively where R1 ad R2 are the mirror reflectivities or Fresnel reflection coefficients
given by
R=

n1n 2
n1 +n2

(2)

where n1 and n2 are the refractive indices of two materials.

For this lasing condition, equation (1) becomes
I(2L) = I(0)R1R2exp[2L{ g(h) (h)}]

(3)

At the lasing threshold, a steady state oscillation takes place and the magnitude and
phase of the returned wave must be equal to those of original wave. This gives amplitude
of
I(2L) = I(0)

(4)

e-j2L = 1

(5)

The modes which have sufficient gain for sustained oscillation are determined by
equation (4). The condition to reach the lasing threshold is the point at which optical gain
is equal to the total loss t, in the cavity. From equation (4) this condition is

1
1
ln (
)
2L
R 1 R2

gth = t = +

= +end

(6)

where end is the mirror loss in the lasing cavity;

Thus for lasing to occur gain g gth.
For laser structures that have strong carrier confinement, the threshold current
density (Jth) for stimulated emission can be related to the lasing optical gain by
gth = Jth

(7)

where is a constant that depends on the specific device construction.

Resonant frequencies:
The phase of the returned wave is given by
e-j2L = 1
The above condition is hold when 2L = 2m

(1)

where m is an integer
As = 2n/ which implies m = 2Ln/ = 2Ln/C

(2)

The above equation states that the cavity resonates when an integer m of halfwavelengths spans the region between the mirrors. Since in all lasers, the gain is a
function of frequency there will be a range of frequencies or wavelengths for which
equation (2) holds. Each of these frequencies corresponds to a mode of oscillation of the
laser. Depending on the laser structure, any number of frequencies can satisfy magnitude
and phase equations. Thus some are single mode lasers and some are multimode. The
relation between gain and frequency is assumed to have Gaussian form given by
g() = g(0) exp

( 0)2
2 2

(3)

where 0 is central wavelength of the spectrum, is the spectral width of the gain ad g(0)
is proportional to the population inversion.
Frequency/wavelength spacing between modes of multimode laser:
Consider longitudinal modes, in each mode there may be several transverse modes
that arises from one or more reflections of the propagating wave at the sides of the
resonator cavity. Now consider two successive modes of frequencies m-1 and m
represented by the integers m-1 and m.
From equation of integer m-1 =
(1)

2 ln
c

m-1

and m =

2 ln
c

(2)
Subtracting (1) from (2) we have

2 ln
c

= 1

(3)
where = m-1 - m
Therefore frequency spacing = c/2Ln

(4)

This can be related to the wave length spacing by

=

which implies

(5)
2
2 ln

(6)

From equations (5) and (6) the output spectrum of a multimode laser follows the
gain vs frequency plot as shown in figure below.

The number of modes, their heights and their spacings depends on the laser
construction.
LASER diode rate equations:
The relationship between optical output power and the diode drive current can be
determined from the rate equations that govern the interaction of photons and electrons
in the active region. For a p junction with a carrier-confinement region of depth d, the rate
equations are given by

d
dt

= Cn + Rsp - ph

(1)
= stimulated emission + spontaneous emission +photon loss
, and

which governs the number of photons

dn
dt

J
qd

n
sp

- Cn

(2)
= injection + spontaneous recombination + stimulated
emission
which governs the number of electrons.
C is a coefficient describing the strength of the optical absorption and emission
interactions, Rsp is the rate of spontaneous emission into the lasing mode, ph is the
photon lifetime, sp is the spontaneous recombination lifetime, and J is the injectioncurrent density.
The first term in eqn (1) is a source of photos resulting from stimulated emission.
The second term describes the number of photos produced by spontaneous emission
which is smaller than the first term. The third term indicates the decay in the number of
photons caused by loss mechanisms in the lasing cavity.
The first term in eqn (2) represents the increase in the electron concentration in the
conduction band as current flows into the device. The second and third terms give the
number of electrons lost from the conduction band due to spontaneous and stimulated
emissions respectively.
The steady state is characterized by the LHS of eqns (1) and (2) being equal to zero.
Assuming Rsp is negligible and d /dt must be positive when is small, we have
Cn -

1
ph

(3)
This shows that n must exceed a threshold value n th in order for

to increase.

Using this equation this threshold value can be expressed in terms of the threshold
current Jth needed to maintain an inversion level n=n th in the steady state when the
number of photons =0

nth
sp

J th
qd

(4)
This equation defines the current required to sustain an excess electron density in
the laser when spontaneous emission is the only decay mechanism.
Consider the photon ad electron rate equations in the steady-state condition at the
lasing threshold. Then eqns (1) and (2) become
0

= Cnth

0=

J
qd

+ Rsp -

s
ph

(5)

nth
sp

- Cnth s

(6)
where

is the steady-state photon density. Adding equations (5) and (6), and by using

(4) solving for

yields the number of photons per unit volume

s

ph
qd (J-Jth) + ph Rsp

(7)
The first term in above equation is the number of photons resulting from stimulated
emission and second term gives the spontaneously generated photons.
External quantum efficiency: The external differential quantum efficiency ( ext) is
defined as the number of photons emitted per radiative electron-hole pair recombination
above threshold. Assuming that the gain coefficient remains fixed at g th above threshold
ext is given by
i (g th )
gth

ext =

(1)

i is the internal quantum efficiency whose value lies between 0.6-0.7 at room
temperature. If ext is calculated from the straight-line portion of the curve for the emitted
optical power P versus drive current I, then
ext
(2)

q
Eg

dp
dI

0.8065

(m)

dp(mW )
dI (mA )

dp is the incremental change in the emitted optical power in milliwatts for a

incremental change dI in the drive current in milliamperes. For standard semiconductor
lasers, external differential quantum efficiency ranges between 15-20 percent per facet.
High-quality devices have differential quantum efficiencies of 30-40 percent.
Fiber alignment and joints:
The optical loss encountered at the interface is a major consideration with all types
of fiber-fiber connections. Even when the two joined fiber ends are smooth and
perpendicular to the fiber axes, and the two fiber axes are perfectly aligned, a small
proportion of light may be reflected back into the transmitting fiber causing attenuation at
the join. This phenomenon is known as Fresnel Reflection. It is associated with the step
changes in refractive index at the jointed interface i.e., glass-air-glass. The magnitude of
this partial reflection of the light transmitted through the interface may be estimated
using the classical Fresnel formula for light of normal incidence given by
r=

n1n
n1 +n

( )

(1)

where r is the fraction of light reflected at a single interface, n 1 is the refractive index of
the fiber core and n is the refractive medium between the two jointed fibers (i.e., for air
n=1).
Fresnel reflection at both fiber interfaces must be taken into account to determine
the amount of light reflected at a fiber joint. This loss in dB at a single interface is given
by
LossFres = -10log10(1-r)

(2)

Hence using the relationships in equations (1) and (2) it is possible to determine the
optical attenuation due to Fresnel reflection at a fiber-fiber joint.
a) Fresnel reflection may give a significant loss at a fiber joint even when all other
aspects of the connection are ideal.
b) The effect of Fresnel reflection at a fiber-fiber connection can be reduced to a very
low level through the use of an index matching fluid in the gap between the jointed
fibers, which has the same refractive index as the fiber core.
c) Fresnel reflection is only one possible source of optical loss at a fiber joint.
d) A potentially greater source of loss at a fiber-fiber connection is caused by
misalignment of the two jointed fibers.

Fiber alignment:
Any deviations in the geometrical and optical parameters of the two optical fibers
which are jointed will affect the optical attenuation or insertion loss through the
connection. It is not possible within any particular connection technique to allow for all
these variations. Hence, there are inherent connection problems when jointing fibers with.
For example

a)
b)
c)
d)

Different core and/or cladding diameters;

Different numerical apertures and/or relative refractive index differences;
Different refractive index profiles;
Fiber faults (core ellipticity, core concentricity, etc);

The losses caused by the above factors together with those of Fresnel reflection are
usually referred to as intrinsic joint losses.
Therefore the best results are achieved with compatible fibers which are manufactured
to the lowest tolerances. In this case still there is a problem of quality of fiber alignment
provided by the jointing mechanism.
The below figure illustrates the possible misalignments between coupled compatible
optical fibers at joints.

Misalignment may occur in three dimensions:

a) The separation between the fibers called longitudinal misalignment.
b) The offset perpendicular to the fiber core axes called lateral/radial/axial
misalignment.
c) The angle between the core axes called angular misalignment.
Optical losses resulting from these three types of misalignment depend upon the fiber
type, core diameter and the distribution of optical power between the propagating modes.
Examples of the measured optical losses due to various types of misalignment are
shown in figures below

Fig (a) shows the attenuation characteristics for both longitudinal and lateral
misalignment of a 50m core diameter graded index fiber. It may be observed that the
lateral misalignment gives significantly greater loss per unit displacement than the

longitudinal displacement. In this case a lateral displacement of 10m gives about 1dB
insertion loss where as similar longitudinal displacement gives loss of around 0.1dB.

Fig (b) shows the attenuation characteristics for angular misalignment of two
multimode step index fibers with numerical apertures of 0.22 and 0.3. A loss of around
1dB is obtained with angular misalignment of 4 0 and 50 for numerical aperture = 0.22 and
0.3 fibers respectively. It may also be observed from figure that the effect of an index
matching fluid in the fiber gap causes increased losses with angular misalignment.
Therefore, it is clear that relatively small levels of lateral and/or angular misalignment can
cause significant attenuation at a fiber joint. This is especially the case for small core
diameter, less than 50m, fibers which are currently employed for most
telecommunication purposes.
Multimode fiber joints:
Theoretical and experimental studies of fiber misalignment in optical fiber
connections allow approximate determination of the losses encountered with the various
misalignments of different fiber types. Some expressions are determined to calculate
losses due to lateral and angular misalignment of optical fiber joints and longitudinal
misalignment is the least important effect and may be largely avoided in fiber connection.
All groups predict that larger the numerical aperture, the greater the spread of the output
light and the higher the optical losses at a longitudinally misaligned joint.
Theoretical expressions for the determination of lateral and angular misalignment
losses are by no means definitive. The experimental results from different sources tend to
vary due to difficulties of measurement.
Lateral misalignment loss:
Lateral misalignment reduces the overlap region between the two fiber cores.
Assuming uniform excitation of all the optical modes in a multimode step index fiber the
overlapped area between both fiber cores approximately gives the lateral coupling
efficiency lat. Hence, the lateral coupling efficiency for two similar step index fibers may
be written as
n
16 1
n

lat =

( )
n
1+
(
[ n )]
1

2 cos

1 /2

( ) ( ) [ ( )]
y
y
y

1
2a
a
2a

(1)

where n1 is the core refractive index, n is the refractive index of medium between fibers,
y is the lateral offset of the fiber core axes and a is the fiber core radius
The lateral misalignment loss in dB may be determined using
Losslat = -10 log10lat dB

(2)

The predicted losses obtained using the formulate in (1) and (2) are generally
slightly higher than the measured values due to the assumption that all modes are equally
excited. This assumption is only correct for certain cases of optical fiber transmission.
Also, it may be assumed that the index is matching and hence no Fresnel reflection which
makes the first term in eqn (1) equal to unity as

n1
n

= 1. This is valid if the two fibers

ends are assumed to be in close contact i.e. no air gap in between and gives lower
predicted losses.
Lateral misalignment loss in multimode graded index fibers assuming uniform
distribution of optical power throughout all guided modes was calculated by Gloge. He
estimated that the lateral misalignment loss was dependent o the refractive index
gradient proportional for small lateral offset ad may be obtained from
Lt =

2 y
a

( )( +2
+1 )

for 0 y 0.2a

(3)

where the lateral coupling efficiency was given by

lat = 1-Lt in dB

(4)

with a parabolic refractive index profile = 2

Lt =

8 y
3 a

()

= 0.85

( ay )

(5)

A further estimate including leaky modes, gave a revised expression for the lateral
misalignment loss given in eqn (4) of 0.75

index fibers where =

( ay )

( ay )

. This analysis was also extended to step

and gave lateral misalignment losses of 0.64

( ay )

and 0.5

for the cases of guided modes only ad both guided plus leaky modes respectively.

From the plot between insertion loss and lateral misalignment measured for a 50m
diameter GI fiber it may be observed that the losses predicted are very pessimistic. A
model which is found to predict insertion loss due to lateral misalignment in GI fiber with
greater accuracy was proposed by Miller and Mettler. In this model they assumed that the

power distribution at the fiber output to be a Gaussian form. Unfortunately the analysis
involves integration using numerical techniques. Therefore the estimation of insertion
losses due to lateral misalignment in multimode GI fiber is limited to the use of Gloges
formulae.
Angular misalignment:
Angular misalignment losses at joints in multimode step index fibers may be
predicted with reasonable accuracy using an expression for the angular coupling
efficiency ang is given by
n
16 1
n

ang =

( ) 1 n
[ n 2 ]
n
1+ ( )
[ n]
4

(5)
where is the angular displacement in radians and is the relative refractive index
difference for the fiber.
Then the insertion loss due to angular misalignment is obtained by
Lossang = -10log10ang

(6)

a) From eqns (5) and (6), for smaller values of , the larger the insertion loss due to
angular misalignment. This appears intuitively correct as small values of imply
small numerical aperture fibers, which will be more affected by angular
misalignment. It is confirmed by the measurements shown in graph between
insertion loss vs angular misalignment.
b) Insertion loss due to angular misalignment can be reduced by increasing the
numerical aperture of fiber. But this is the factor for increase in the loss due to
longitudinal misalignment at a joint.
c) Major factors
which cause fiber-fiber intrinsic losses were
i)
Mismatch in fiber core diameters
ii)
Mismatch in numerical apertures
iii)
Difference between the refractive index profiles.
Longitudinal misalignment loss:
Assuming all the modes are equally excited in a multimode step index or multimode
graded index fiber, and the numerical apertures and index profiles are the same, then the
losses resulting from a mismatch of core diameters is given by

LossesCD =

where a1 and a2 are core radii of the transmitting and receiving fibers.

a) It may be noted form eqn (a) that o loss is incurred if the receiving fiber has a larger
diameter than the transmitting one. In addition, only a relatively small loss around
0.09dB is obtained when the receiving fiber core diameter is 1% smaller than that of
the transmitting fiber.
b) When the transmitting fiber has a higher numerical aperture than receiving fiber,
some of the emitted light will fall outside the acceptance angle of receiving fiber
and therefore they will not be coupled to the receiving fiber through the joint.
Assuming a uniform modal power distribution, and fibers with equivalent refractive
index profiles and core diameters, then the loss caused by a mismatch of numerical
apertures is given
by
LossNA =
where NA1 and NA2 are the numerical apertures of transmitting and receiving fibers
respectively.
Equation (b) is valid for both step index and graded index fibers and in
common with equation (a) it demonstrates that no loss occur when the receiving
parameter i.e., numerical aperture is larger than transmitting one.
c) Finally, a mismatch in refractive index profiles results a loss given as
LossRI =

where 1 and 2 are the profile parameters of transmitting and receiving fibers.
d) When connecting from a step index fiber with 1 =

to a parabolic profile graded

index fiber with 2 = 2 both having the same core diameter ad axial numerical
aperture, then a loss of 3 dB is produced. The reverse connection does not incur a
loss due to refractive index profile mismatch.

The intrinsic loss obtained at multimode fiber-fiber joints due to core diameter
mismatch, numerical aperture mismatch and refractive index profile mismatches can be
combined into a single expression as
Lossint =

The above equation assumes that the three mismatches occur together.
Single mode fiber joints:
Misalignment losses at connections in single mode fibers have been theoretically
considered by Marcuse ad Gambling et al. The analysis which was instigated by Marcuse
is based upon the Gaussian or near Gaussian shape of the modes propagating in single
mode fibers for both step index and graded index. Further development of this theory by
Gambling et al. gave simplified formulae for both the lateral ad angular misalignment
losses at joints in single mode fibers.
a) In the absence of angular misalignment, Gambling et al. calculated that the loss T l
due to lateral offset y was given by
y 2
Tl = 2.17
dB
(1)
w

( )

where w is the normalized spot size of the fundamental mode. The normalized spot
size for the LP01 mode is obtained from the empirical formula
w=a

0.65+1.62 v 3 /2 +2.88 v 6
21/ 2

in m

(2)

where a is the core radius and v is the normalized frequency.

b) The insertion loss Ta caused by an angular misalignment in radian at a joint in a
single mode fiber may be given by
2
w n1 v
Ta = 2.17
dB
(3)
aN.A

where n1 is the fiber core refractive index and N.A is the numerical aperture of fiber.
Note: i) In the above two equations it is assumed that the spot sizes of the modes in the
two coupled fibers are the same.
ii) Gambling et al. also derived expressions for combined losses due to both lateral and
angular misalignments. They indicate that for small total losses [<0.75dB] a reasonable
approximation is obtained by simply combining (1) and (3).
iii) Aside from actual magnitudes of the respective misalignments, the insertion losses
incurred are also strongly dependent on normalized frequency of the fiber. This is
especially the case where angular misalignment loss is < 0.3 dB obtained when =10 with
fibers of appropriate v-value. For low loss single mode fiber joints it is important that
angular misalignment is better than 10.
The theoretical model developed by Marcuse has been utilized by Nemota and
Makimoto in deriving a general equation for determining the coupling losses between
single mode fibers. Their full expression takes account of all the extrinsic factors (lateral,
angular, longitudinal misalignments and Fresnel reflection), as well as intrinsic factors
associated with the connection of fibers with unequal mode field diameters.

Lossint = -10log10

[(

w
w
4 02 + 01
w01 w 02

(4)

where w01 and w02 are the spot sizes of transmitting and receiving fibers respectively.
Equation (4) enables the additional coupling loss and resulting from mode field diameter
mismatch between two single mode fibers to be calculated.
Note: the intrinsic loss through a single mode fiber joint is independent of the direction of
propagation i.e., any fiber can be transmitting/receiving.
Fiber splicing:
A fiber splice is a permanent or semi-permanent joint between the two fibers. These
are typically used to create log optical links or in situations where frequent connection and
disconnection are not needed. In making and evaluating such splices, one must consider
a) the geometrical differences I the two fibers, b) fiber misalignments at the joints and c)
the mechanical strength of fiber.
Splicing techniques:
Splicing techniques include
a) The fusion splice
b) V-groove mechanical splice and
c) Elastic-tube splice
The first technique yields a permanent joint, whereas the other two types of splices can
be disassembled if necessary.
a) Fusion splice: Fusion splices are made by thermally bonding together the prepared
fiber ends as pictured in the figure.

In this method the fiber ends are first prealigned ad butted together. This is
done either in a grooved fiber holder or under a microscope with micro
manipulators. The butt-joint is then heated with an electric arc or laser pulse so that
the fiber ends are momentarily melted and hence bonded together. This technique
can produce very low splice losses <0.06dB. However, care must be exercised in
this technique, since surface damage due to handling, surface defect growth
created during heating and residual stresses induced near the joint as a result of
changes in chemical composition arising from the material melting can produce a
weak splice.
b) V-groove splicing: In the V-groove splice technique, the prepared fiber ends are
first butted together in a V-shaped groove as shown. Then they are bonded together
with a adhesive or are held in place by means of a cover plate. The V-shaped
channel can be either a grooved-silicon, plastic, ceramic or metal substrate. The
splice loss in this method depends on the fiber size (outside dimensions and core-

diameter variations) and eccentricity (the position of the core relative to the center
of fiber).

c) Elastic tube splice: Elastic tube splice is shown in figure cross-sectionally. It is a

unique device that automatically performs lateral, longitudinal and angular
alignment. It splices multimode fibers to give losses in the same range as
commercial fusion splices, but loss equipment and skill are needed. The central hold
diameter is slightly smaller than that of the fiber to be spliced and it is tapered at
both ends for easy fiber insertion. When a fiber is inserted, it expands the hole
diameter so that the elastic material exerts a symmetrical force on the fiber. This
symmetry feature allows an accurate and automatic alignment of the axes of the
two fibers to be joined. A wide range of fiber diameters can be inserted into the
tube. Thus the fibers to be spliced do not have to be equal in diameters, since each
fiber moves into position independently relative to the tube axes.