Optical Fiber Communications, 3e - Gerd Keiser

E&CE 477
Photonic Communications
Systems & Devices
Winter 2006
Instructor: Hamed Majedi
Content
1- Overview of Photonic Communications
2- Optical Fiber: Waveguiding, Propagation Modes
- Single Mode Fiber
- Fiber Materials & Fabrication Procedures
3- Signal Degradation in Optical Fibers
4- Photonic Sources & Transmitters: LED & Laser Diodes
- Single Mode Lasers, Modulation & Noise
5- Laser-Fiber Connections (Power Launching & Coupling)
6- Photodetectors
7- Digital Photonic Receivers & Digital Transmission systems
8- WDM & Photonic Networks
Lab & Computer Simulations
• Lab sessions
- Fiber Attenuation Measurement
- Dispersion Measurement
- Spectral Attenuation Measurements
• Computer Simulations using Photonic Transmission Design

Suite 1.1 Lite
1- Bit error rate estimation of digital single channel fiber-optic link
2- Influence of fiber dispersion on the bit error rate
3- Fiber dispersion compensation by three different methods
4- Four channel WDM transmission by four wave mixing
5- Comparison of external vs. direct laser modulation for various bit rate
6- Two channel WDM add/drop multiplexer using fiber Bragg gratings
& circulators.
Chapter 1
Overview of Photonic
Communications
Optics
• Optics is an old subject involving the generation, propagation
& detection of light.
• Three major developments are responsible for rejuvenation of
optics & its application in modern technology:
1- Invention of Laser
2- Fabrication of low-loss optical Fiber
3- Development of Semiconductor Optical Device
As a result, new disciplines have emerged & new terms describing them
have come into use, such as:
- Electro-Optics: is generally reserved for optical devices in
which electrical effects play a role, such as lasers, electro-optic
modulators & switches.
Photonics
• Optoelectronics: refers to devices & systems that are
essentially electronics but involve lights, such as LED, liquid
crystal displays & array photodetectors.
• Quantum Electronics: is used in connection with devices &
systems that rely on the interaction of light with matter, such
as lasers & nonlinear optical devices.
• Quantum Optics: Studies quantum & coherence properties of
light.
• Lightwave Technology: describes systems & devices that are
used in optical communication & signal processing.
• Photonics: in analogy with electronics, involves the control of
photons in free space and matter.
Photonic Communications
• Photonics reflects the importance of the photon nature of light. Photonics
& electronics clearly overlap since electrons often control the flow of
photons & conversely, photons control the flow of electrons.
• The scope of Photonics:
1- Generation of Light (coherent & incoherent)
2- Transmission of Light (through free space, fibers, imaging systems,
waveguides, … )
3- Processing of Light Signals (modulation, switching, amplification,
frequency conversion, …)
4- Detection of Light (coherent & incoherent)
• Photonic Communications: describes the applications of
photonic technology in communication devices & systems,
such as transmitters, transmission media, receivers & signal
processors.
Why Photonic Communications?
• Extremely wide bandwidth: high carrier frequency ( a wavelength of
1552.5 nm corresponds to a center frequency of 193.1 THz!) &
consequently orders of magnitude increase in available transmission
bandwidth & larger information capacity.
• Optical Fibers have small size & light weight.
• Optical Fibers are immune to electromagnetic interference (high voltage
transmission lines, radar systems, power electronic systems, airborne
systems, …)
• Lack of EMI cross talk between channels
• Availability of very low loss Fibers (0.25 to 0.3 dB/km), high
performance active & passive photonic components such as
tunable lasers, very sensitive photodetectors, couplers, filters,
• Low cost systems for data rates in excess of Gbit/s.
BW demands in communication systems
Type & Format Uncompressed Compressed
applications
Voice, digital 4 kHz voice 64 kbps 16-32 kbps
telegraphy
Audio 16-24 kHz 512-748 kbps 32-384 kbps
(MPEG, MP3)
Video conferencing 176 144 or 352  2-35.6 Mbps 64 kbps-1.544
288 frames @ 10- Mbps (H.261
30 frames/s coding)
Data transfer, E- 1-10 Mbps
commerce,Video
entertainment
Full-motion 720480frames @ 249 Mbps 2-6Mbps (MPEG-2)
broadcast video 30 frames/s
HDTV 1920 1080 1.6 Gbps 19-38 Mbps
frames@ 30 frames (MPEG-2)
/s
Early application of fiber optic communication
• Digital link consisting of time-division-multiplexing (TDM) of 64 kbps
voice channels (early 1980).
Optical Fiber communications, 3rd ed.,G.Keiser,McGrawHill, 2000

SONET & SDH Standards
• SONET (Synchronous Optical NETwork) is the network standard used in
north America & SDH (Synchronous Digital Hierarchy) is used in other
parts of the world. These define a synchronous frame structure for sending
multiplexed digital traffic over fiber optic trunk lines.
• The basic building block of SONET is called STS-1 (Synchronous

Transport Signal) with 51.84 Mbps data rate. Higher-rate SONET signals
are obtained by byte-interleaving N STS-1 frames, which are scramble &
converted to an Optical Carrier Level N (OC-N) signal.
• The basic building block of SDH is called STM-1 (Synchronous Transport

Module) with 155.52 Mbps data rate. Higher-rate SDH signals are achieved
by synchronously multiplexing N different STM-1 to form STM-N signal.
SONET & SDH transmission rates
SONET level Electrical level Line rate (Mb/s) SDH equivalent
OC-1 STS-1 51.84 -
OC-3 STS-3 155.52 STM-1
OC-12 STS-12 622.08 STM-4
OC-24 STS-24 1244.16 STM-8
OC-48 STS-48 2488.32 STM-16
OC-96 STS-96 4976.64 STM-32
OC-192 STS-192 9953.28 STM-64

Evolution of fiber optic systems
• 1950s:Imaging applications in
medicine & non-destructive testing,
lighting
• 1960s:Research on lowering the fiber
loss for telecom. applications.
• 1970s:Development of low loss
fibers, semiconductor light sources &
photodetectors
• 1980s:single mode fibers (OC-3 to
OC-48) over repeater sapcings of 40
km.
• 1990s:Optical amplifiers (e.g.
EDFA), WDM (wavelength division
multiplexing) toward dense-WDM.
Optical Fiber communications, 3rd ed.,G.Keiser,McGrawHill,

2000
Operating range of 4 key components in the 3 different
optical windows

Major elements Of typical photonic comm link

Installation of Fiber optics

WDM Concept

Chapter 2
Optical Fibers: Structures,

Waveguiding & Fabrication
Theories of Optics
• Light is an electromagentic phenomenon described by the same theoretical
principles that govern all forms of electromagnetic radiation. Maxwell’s
equations are in the hurt of electromagnetic theory & is fully successful in
providing treatment of light propagation. Electromagnetic optics provides
the most complete treatment of light phenomena in the context of classical
optics.
• Turning to phenomena involving the interaction of light & matter, such as
emission & absorption of light, quantum theory provides the successful
explanation for light-matter interaction. These phenomena are described by
quantum electrodynamics which is the marriage of electromagnetic theory
with quantum theory. For optical phenomena, this theory also referred to as
quantum optics. This theory provides an explanation of virtually all
optical phenomena.
• In the context of classical optics, electromagentic radiation propagates in
the form of two mutually coupled vector waves, an electric field-wave &
magnetic field wave. It is possible to describe many optical phenomena
such as diffraction, by scalar wave theory in which light is described by a
single scalar wavefunction. This approximate theory is called scalar wave
optics or simply wave optics. When light propagates through & around
objects whose dimensions are much greater than the optical wavelength,
the wave nature of light is not readily discerned, so that its behavior can be
adequately described by rays obeying a set of geometrical rules. This
theory is called ray optics. Ray optics is the limit of wave optics when the
wavelength is very short.
Quantum Optics
Electromagnetic Optics
Wave Optics
Ray Optics
Engineering Model
• In engineering discipline, we should choose the appropriate & easiest

physical theory that can handle our problems. Therefore, specially in this
course we will use different optical theories to describe & analyze our
problems. In this chapter we deal with optical transmission through fibers,
and other optical waveguiding structures. Depending on the structure, we
may use ray optics or electromagnetic optics, so we begin our discussion
with a brief introduction to electromagnetic optics, ray optics & their
fundamental connection, then having equipped with basic theories, we
analyze the propagation of light in the optical fiber structures.
Electromagnetic Optics
• Electromagnetic radiation propagates in the form of two mutually coupled
vector waves, an electric field wave & a magnetic field wave. Both are
vector functions of position & time.
• In a source-free, linear, homogeneous, isotropic & non-dispersive media,
such as free space, these electric & magnetic fields satisfy the following
partial differential equations, known as Maxwell’ equations:

 E
 H   [2-1]
t

 H
  E   [2-2]
t

E  0 [2-3]

H  0 [2-4]
• In Maxwell’s equations, E is the electric field expressed in [V/m], H is the
magnetic field expressed in [A/m].
 [F/m] : Electric permittivi ty

 [H/m] : Magnetic permeabili ty
  : is divergence operation
 : is curl operation
• The solution of Maxwell’s equations in free space, through the wave

equation, can be easily obtained for monochromatic electromagnetic
wave. All electric & magnetic fields are harmonic functions of time of the
same frequency. Electric & magnetic fields are perpendicular to each other
& both perpendicular to the direction of propagation, k, known as
transverse wave (TEM). E, H & k form a set of orthogonal vectors.
Electromagnetic Plane wave in Free space
Ex
Direction of Propagation k
x
z z
y
By
An electromagnetic wave is a travelling wave which has time

varying electric and magnetic fields which are perpendicular to each
other and the direction of propagation, z.
S.O.Kasap, optoelectronics and Photonics Principles and Practices, prentice hall, 2001
Linearly Polarized Electromagnetic Plane wave

E  e x E 0 x cos(ωt - kz) [2-5]

H  e y H 0 y cos(ωt  kz) [2-6]
where :
ω  2f : Angular frequency [rad/m] [2-7]
k : Wavenumber or wave propagation constant [1/m]

2
 :Wavelength [m]
k
E0 x 
  [] : intrinsic (wave) impedance [2-8]
H0y 
1
v [m/s] : velocity of wave propagation [2-9]

E and B have constant phase
in this xy plane; a wavefront
z E
E
k
Propagation
B
Ex
Ex = Eo sin(wt–kz)
A plane EM wave travelling alongz, has the same Ex (or By) at any point in a
given xy plane. All electric field vectors in a givenxy plane are therefore in phase.
The xy planes are of infinite extent in thex and y directions.
Wavelength & free space
• Wavelength is the distance over which the phase changes by 2 .
v
 [2-10]
f
• In vacuum (free space):
10 9
0  [F/m]  0  4  10 7 [H/m]
36 [2-11]
v  c  3  10 8 m/s  0  120 []
EM wave in Media
• Refractive index of a medium is defined as:
c velocity of light (EM wave) in vacuum 

n    r r
v velocity of light (EM wave) in medium  0 0
[2-12]
 r : Relative magnetic permeability
 r : Relative electric permittivity
• For non-magnetic media ( r  1) :
n  r [2-13]
Intensity & power flow of TEM wave
1  
• The poynting vector S  E  H for TEM wave is parallel to the
2
wavevector k so that the power flows along in a direction normal to the
wavefront or parallel to k. The magnitude of the poynting vector is the
intensity of TEM wave as follows:
2
E0
I [W/m 2 ] [2-14]
2
Connection between EM wave optics & Ray
optics
According to wave or physical optics viewpoint, the EM waves radiated by
a small optical source can be represented by a train of spherical wavefronts
with the source at the center. A wavefront is defined a s the locus of all
points in the wave train which exhibit the same phase. Far from source
wavefronts tend to be in a plane form. Next page you will see different
possible phase fronts for EM waves.
When the wavelength of light is much smaller than the object, the
wavefronts appear as straight lines to this object. In this case the light wave
can be indicated by a light ray, which is drawn perpendicular to the phase
front and parallel to the Poynting vector, which indicates the flow of
energy. Thus, large scale optical effects such as reflection & refraction can
be analyzed by simple geometrical process called ray tracing. This view of
optics is referred to as ray optics or geometrical optics.
Wave fronts
(constant phase surfaces) Wave fronts
Wave fronts
k
  P E
k r
rays 
P
O
z
A perfect plane wave A perfect spherical wave A divergent beam
(a) (b) (c)
Examples of possible EM waves
General form of linearly polarized plane waves
Any two orthogonal plane waves

Can be combined into a linearly
Polarized wave. Conversely, any
arbitrary linearly polarized wave
can be resolved into two
independent Orthogonal plane
waves that are in phase.

E  e x E0 x cos(ωt  kz)  e y E0 y cos(ωt  kz)

E  E  E0 x  E0 y
2 2
[2-15]
E0 y
  tan (
1
)
E0 x
Elliptically Polarized plane waves

E  e x Ex  e y E y
 e x E 0 x cos(ωt  kz)  e y cos(ωt  kz   )
2
 Ex   E y   E x  E y 
2
      2   cos   sin 2 

   
 E0 x   E0 y   E 0 x  E 0 y  [2-16]
2 E 0 x E 0 y cos 
tan( 2 ) 
E0 x  E0 y
2 2

Circularly polarized waves

Circular polarizati on : E0 x  E0 y  E0 &    [2-17]
2
 : right circularly polarized, - : left circularly polarized

Laws of Reflection & Refraction
Reflection law: angle of incidence=angle of reflection
Snell’s law of refraction:
n1 sin 1  n2 sin 2 [2-18]

Total internal reflection, Critical angle
Transmitted
(refracted) light
kt  2  90
2 n2 Evanescent wave
n 1 > n2
1
ki
1 kr c
Critical angle
1   c TIR
Incident Reflected
light light
n2 (c)
(a) sin  c  (b)
n1
Light wave travelling in a more dense medium strikes a less dense medium. Depending on
the incidence angle with respect to c , which is determined by the ratio of the refractive
indices, the wave may be transmitted (refracted) or reflected. (a)1   c (b) 1   c (c)
1   c and total internal reflection (TIR).
n2
sin  c  [2-19]
n1
Phase shift due to TIR
• The totally reflected wave experiences a phase shift however
which is given by:
N n 2 cos 2 1  1  p n n 2 cos 2 1  1
tan  ; tan  [2-20]
2 n sin 1 2 sin  1
n1
n
n2
• Where (p,N) refer to the electric field components parallel or

normal to the plane of incidence respectively.
Optical waveguiding by TIR:
Dielectric Slab Waveguide
Propagation mechanism in an ideal step-index optical waveguide.

Launching optical rays to slab waveguide
n2
sin  min  ; minimum angle that supports TIR
n1
[2-21]
Maximum entrance angle,  0 max is found from

the Snell’s relation written at the fiber end face.
n sin  0 max  n1 sin  c  n1  n2

2 2
[2-22]
Numerical aperture:
NA  n sin  0 max  n1  n2  n1 2
2 2
[2-23]
n1  n2
 [2-24]
n1
Optical rays transmission through dielectric slab
waveguide

n1  n 2 ;    c   c O
2
For TE-case, when electric waves are normal to the plane of incidence
 must be satisfied with following relationship:

 n1 d sin  m   n1 cos   n2
2 2 2 
tan     [2-25]
  2   n1 sin  
 
Note
• Home work 2-1) Find an expression for  ,considering that the electric
field component of optical wave is parallel to the plane of incidence (TM-
case).
• As you have seen, the polarization of light wave down the slab waveguide
changes the condition of light transmission. Hence we should also consider
the EM wave analysis of EM wave propagation through the dielectric slab
waveguide. In the next slides, we will introduce the fundamental concepts
of such a treatment, without going into mathematical detail. Basically we
will show the result of solution to the Maxwell’s equations in different
regions of slab waveguide & applying the boundary conditions for electric
& magnetic fields at the surface of each slab. We will try to show the
connection between EM wave and ray optics analyses.
EM analysis of Slab waveguide
• For each particular angle, in which light ray can be faithfully transmitted
along slab waveguide, we can obtain one possible propagating wave
solution from a Maxwell’s equations or mode.
• The modes with electric field perpendicular to the plane of incidence (page)
are called TE (Transverse Electric) and numbered as: TE 0 , TE 1 , TE 2 ,...
Electric field distribution of these modes for 2D slab waveguide can be
expressed as:

Em ( x, y, z, t )  e x f m ( y) cos(ωt   m z ) [2-26]
m  0,1,2,3 (mode number)
wave transmission along slab waveguides, fibers & other type of optical
waveguides can be fully described by time & z dependency of the mode:
cos(ωt   m z ) or e j (wt   m z )
TE modes in slab waveguide

Em ( x, y, z, t )  e x f m ( y) cos(ωt   m z )
m  0,1,2,3 (mode number)
Modes in slab waveguide
• The order of the mode is equal to the # of field zeros across the guide. The
order of the mode is also related to the angle in which the ray congruence
corresponding to this mode makes with the plane of the waveguide (or axis
of the fiber). The steeper the angle, the higher the order of the mode.
• For higher order modes the fields are distributed more toward the edges of
the guide and penetrate further into the cladding region.
• Radiation modes in fibers are not trapped in the core & guided by the fiber
but they are still solutions of the Maxwell’ eqs. with the same boundary
conditions. These infinite continuum of the modes results from the optical
power that is outside the fiber acceptance angle being refracted out of the
core.
• In addition to bound & refracted (radiation) modes, there are leaky modes
in optical fiber. They are partially confined to the core & attenuated by
continuously radiating this power out of the core as they traverse along the
fiber (results from Tunneling effect which is quantum mechanical
phenomenon.) A mode remains guided as long as n2 k    n1k
Optical Fibers: Modal Theory (Guided or
Propagating modes) & Ray Optics Theory
n1 n2
n1  n2
Step Index Fiber
Modal Theory of Step Index fiber
• General expression of EM-wave in the circular fiber can be written as:
  
E (r ,  , z, t )   Am E m (r ,  , z, t )  AmU m (r ,  )e j ( ωt   m z )
m m
  
H (r ,  , z, t )   Am H m (r ,  , z, t )   AmVm (r ,  )e j ( ωt   m z )
m m
[2-27]
 
• Each of the characteristic solutions Em (r,  , z, t ) & H m (r ,  , z, t ) is
called mth mode of the optical fiber.
• It is often sufficient to give the E-field of the mode.

U m (r,  )e j (ωt  m z ) m  1,2,3...

• The modal field distribution, U m (r ,  ) , and the mode
propagation constant,  m are obtained from solving the
Maxwell’s equations subject to the boundary conditions given
by the cross sectional dimensions and the dielectric constants
of the fiber.
• Most important characteristics of the EM transmission along the fiber are

determined by the mode propagation constant,  m (ω) , which depends on
the mode & in general varies with frequency or wavelength. This quantity
is always between the plane propagation constant (wave number) of the
core & the cladding media .
n2 k   m (ω)  n1k [2-28]

• At each frequency or wavelength, there exists only a finite number of
guided or propagating modes that can carry light energy over a long
distance along the fiber. Each of these modes can propagate in the fiber
only if the frequency is above the cut-off frequency, ω c , (or the source
wavelength is smaller than the cut-off wavelength) obtained from cut-off
condition that is:
 m (ωc )  n2 k [2-29]
• To minimize the signal distortion, the fiber is often operated in a single

mode regime. In this regime only the lowest order mode (fundamental
mode) can propagate in the fiber and all higher order modes are under cut-
off condition (non-propagating).
• Multi-mode fibers are also extensively used for many applications. In
these fibers many modes carry the optical signal collectively &
simultaneously.
Fundamental Mode Field Distribution
Mode field diameter

Polarizations of fundamental mode

Ray Optics Theory (Step-Index Fiber)
Skew rays
Each particular guided mode in a fiber can be represented by a group of rays which
Make the same angle with the axis of the fiber.

Different Structures of Optical Fiber

Mode designation in circular cylindrical
waveguide (Optical Fiber)
TE lm modes : The electric field vector lies in transverse plane.
TM lm modes : The magnetic field vector lies in transverse plane.
Hybrid HE lm modes :TE component is larger than TM component.
Hybrid EH lm modes : TM component is larger than TE component.
y
l= # of variation cycles or zeros in direction. r
m= # of variation cycles or zeros in r direction. 
x
z
Linearly Polarized (LP) modes in weakly-guided fibers ( n1  n2  1 )
LP0m (HE 1m ), LP1m (TE 0m  TM 0m  HE 0m )
Fundamental Mode: LP01 (HE 11 )
Two degenerate fundamental modes in Fibers
(Horizontal & Vertical HE 11 Modes)

Mode propagation constant as a function of frequency
• Mode propagation constant,  lm (ω), is the most important transmission

characteristic of an optical fiber, because the field distribution can be easily
written in the form of eq. [2-27].
• In order to find a mode propagation constant and cut-off frequencies of
various modes of the optical fiber, first we have to calculate the
normalized frequency, V, defined by:
2a 2a
V n1  n2 
2 2
NA [2-30]
 
a: radius of the core,  is the optical free space wavelength,
n1 & n2 are the refractive indices of the core & cladding.
Plots of the propagation constant as a function of normalized
frequency for a few of the lowest-order modes
Single mode Operation
• The cut-off wavelength or frequency for each mode is obtained from:
2n2 w c n2
 lm (ω c )  n2 k   [2-31]
c c
• Single mode operation is possible (Single mode fiber) when:
V  2.405 [2-32]
Only HE11 can propagate faithfully along optical fiber

Single-Mode Fibers
  0.1% to 1% ; a  6 to 12 m ;
V  2.3 to 2.4 @ max frequency or min 
• Example: A fiber with a radius of 4 micrometer and n1  1.500 & n2  1.498

has a normalized frequency of V=2.38 at a wavelength 1 micrometer. The
fiber is single-mode for all wavelengths greater and equal to 1 micrometer.
MFD (Mode Field Diameter): The electric field of the first fundamental
mode can be written as:
r2
E (r )  E 0 exp(  2
); MFD  2W0 [2-33]
W0
Birefringence in single-mode fibers
• Because of asymmetries the refractive indices for the two degenerate modes
(vertical & horizontal polarizations) are different. This difference is referred to as
birefringence, B f :
B f  n y  nx [2-34]

Fiber Beat Length
• In general, a linearly polarized mode is a combination of both of the

degenerate modes. As the modal wave travels along the fiber, the
difference in the refractive indices would change the phase difference
between these two components & thereby the state of the polarization of
the mode. However after certain length referred to as fiber beat length, the
modal wave will produce its original state of polarization. This length is
simply given by:
2
Lp  [2-35]
kB f
Multi-Mode Operation
• Total number of modes, M, supported by a multi-mode fiber is
approximately (When V is large) given by:
V2
M  [2-36]
2
• Power distribution in the core & the cladding: Another quantity of
interest is the ratio of the mode power in the cladding, Pclad to the total
optical power in the fiber, P, which at the wavelengths (or frequencies) far
from the cut-off is given by:
Pclad 4
 [2-37]
P 3 M
Chapter 3
Signal Degradation in
Optical Fibers
Signal Attenuation & Distortion in
Optical Fibers
• What are the loss or signal attenuation mechanism in a fiber?
• Why & to what degree do optical signals get distorted as they
propagate down a fiber?
• Signal attenuation (fiber loss) largely determines the maximum

repeaterless separation between optical transmitter & receiver.
• Signal distortion cause that optical pulses to broaden as they
travel along a fiber, the overlap between neighboring pulses,
creating errors in the receiver output, resulting in the limitation
of information-carrying capacity of a fiber.
Attenuation (fiber loss)
• Power loss along a fiber:
Z=0 Z= l
 p l
P(0) mW P (l )  P (0)e mw
 p z
P( z )  P(0)e [3-1]
• The parameter  p is called fiber attenuation coefficient in a units of for

example [1/km] or [nepers/km]. A more common unit is [dB/km] that is
defined by:
10  P(0) 
 [dB/km ]  log    4.343 p [1 / km] [3-2]
l  P(l ) 
Fiber loss in dB/km
z=0 Z=l
P(0)[dBm ]
P(l )[dBm ]  P(0)[dBm ]   [dB/km ]  l[km] [3-3]
• Where [dBm] or dB milliwat is 10log(P [mW]).

Optical fiber attenuation vs. wavelength

Absorption
• Absorption is caused by three different mechanisms:

1- Impurities in fiber material: from transition metal ions (must
be in order of ppb) & particularly from OH ions with
absorption peaks at wavelengths 2700 nm, 400 nm, 950 nm &
725nm.
2- Intrinsic absorption (fundamental lower limit): electronic
absorption band (UV region) & atomic bond vibration band
(IR region) in basic SiO2.
3- Radiation defects
Scattering Loss
• Small (compared to wavelength) variation in material density, chemical
composition, and structural inhomogeneity scatter light in other directions
and absorb energy from guided optical wave.
• The essential mechanism is the Rayleigh scattering. Since the black body
radiation classically is proportional to  4 (this is true for wavelength
typically greater than 5 micrometer), the attenuation coefficient due to
Rayleigh scattering is approximately proportional to  . This seems to me
4
not precise, where the attenuation of fibers at 1.3 & 1.55 micrometer can be
exactly predicted with Planck’s formula & can not be described with
Rayleigh-Jeans law. Therefore I believe that the more accurate formula for
scattering loss is
1
 hc 
 scat  5 exp( )
 k B T 
h  6.626  10 34 Js, k B  1.3806  10 23 JK -1 , T : Temperatur e
Absorption & scattering losses in fibers

Typical spectral absorption & scattering
attenuations for a single mode-fiber

Bending Loss (Macrobending & Microbending)
• Macrobending Loss: The

curvature of the bend is much
larger than fiber diameter.
Lightwave suffers sever loss due
to radiation of the evanescent
field in the cladding region. As
the radius of the curvature
decreases, the loss increases
exponentially until it reaches at a
certain critical radius. For any
radius a bit smaller than this
point, the losses suddenly
becomes extremely large. Higher
order modes radiate away faster
than lower order modes.
Microbending Loss
• Microbending Loss:
microscopic bends of the fiber
axis that can arise when the
fibers are incorporated into
cables. The power is dissipated
through the microbended fiber,
because of the repetitive
coupling of energy between
guided modes & the leaky or
radiation modes in the fiber.

Dispersion in Optical Fibers
• Dispersion: Any phenomenon in which the velocity of propagation of any
electromagnetic wave is wavelength dependent.
• In communication, dispersion is used to describe any process by which any

electromagnetic signal propagating in a physical medium is degraded
because the various wave characteristics (i.e., frequencies) of the signal
have different propagation velocities within the physical medium.
• There are 3 dispersion types in the optical fibers, in general:
1- Material Dispersion
2- Waveguide Dispersion
3- Polarization-Mode Dispersion
Material & waveguide dispersions are main causes of Intramodal

Dispersion.
Group Velocity
• Wave Velocities:
• 1- Plane wave velocity: For a plane wave propagating along z-axis in an
unbounded homogeneous region of refractive index n1 , which is
represented by exp( jωt  jk1 z) , the velocity of constant phase plane is:
w c
v  [3-4]
k1 n1
• 2- Modal wave phase velocity: For a modal wave propagating along z-axis
represented byexp( jωt  jz ) , the velocity of constant phase plane is:
ω
vp  [3-5]

3- For transmission system operation the most important & useful type of
velocity is the group velocity, V g . This is the actual velocity which the
signal information & energy is traveling down the fiber. It is always less
than the speed of light in the medium. The observable delay experiences by
the optical signal waveform & energy, when traveling a length of l along
the fiber is commonly referred to as group delay.
Group Velocity & Group Delay
• The group velocity is given by:
dω
Vg  [3-6]
d
• The group delay is given by:
l d
g   l [3-7]
Vg dω
• It is important to note that all above quantities depend both on frequency
& the propagation mode. In order to see the effect of these parameters on
group velocity and delay, the following analysis would be helpful.
Input/Output signals in Fiber Transmission
System
• The optical signal (complex) waveform at the input of fiber of length l is
f(t). The propagation constant of a particular modal wave carrying the
signal is  (ω). Let us find the output signal waveform g(t).
w is the optical signal bandwidth.
z-=0 Z=l
w c  w
~
f (t ) 
w
 w
f (w )e jwt dw [3-8]
c 
w c  w
~
g (t ) 
w
 w
f (w )e jwt  j (w ) l dw [3-9]
c 
If w  w c
d 1 d 2
 (w )   (w c )  (w  w c )  (w  w c ) 2  ... [3-10]
dw w w c 2 dw 2
w w c
w c  w / 2 w c  w / 2 d
j wt  j [  ( w c )  (w w c )]l
~ ~ dw
  f (w )e
jw t  j  (w ) l
g (t )  f (w )e dw  w w c
dw
w c  w / 2 w c  w / 2
w c  w / 2 d
jw ( t  l )
~ dw
 e  j  (w c ) l  f (w )e dw
w w c
w c  w / 2
 j  (w c ) l d
e f (t  l )  e  j (w c )l f (t   g )
dw
[3-11]
w w c
d l
g  l  [3-14]
dw w w c Vg
Intramodal Dispersion
• As we have seen from Input/output signal relationship in optical fiber, the
output is proportional to the delayed version of the input signal, and the
delay is inversely proportional to the group velocity of the wave. Since the
propagation constant,  (ω) , is frequency dependent over band width ω
sitting at the center frequency ω c , at each frequency, we have one
propagation constant resulting in a specific delay time. As the output signal
is collectively represented by group velocity & group delay this
phenomenon is called intramodal dispersion or Group Velocity
Dispersion (GVD). This phenomenon arises due to a finite bandwidth
of the optical source, dependency of refractive index on the
wavelength and the modal dependency of the group velocity.
• In the case of optical pulse propagation down the fiber, GVD causes pulse
broadening, leading to Inter Symbol Interference (ISI).
Dispersion & ISI
A measure of information
capacity of an optical fiber for
digital transmission is usually
specified by the bandwidth
distance product BW  L
in GHz.km.
For multi-mode step index fiber
this quantity is about 20
MHz.km, for graded index fiber
is about 2.5 GHz.km & for single
mode fibers are higher than 10
GHz.km.

How to characterize dispersion?
• Group delay per unit length can be defined as:
g d 1 d 2 d
  
2c d
[3-15]
L dω c dk
• If the spectral width of the optical source is not too wide, then the delay
d
difference per unit wavelength along the propagation path is approximately g
For spectral components which are apart, symmetrical around center d
wavelength, the total delay difference  over a distance L is:
d g L  d 2 d  
2
      2  
d 2c  d d2 
d d  L   d 2 
 w  w  L w [3-16]
dw dw  V g   dw
2

d 2
• 2  is called GVD parameter, and shows how much a light pulse
dw 2
broadens as it travels along an optical fiber. The more common parameter
is called Dispersion, and can be defined as the delay difference per unit
length per unit wavelength as follows:
1 d g d  1 
   2c  2
D  [3-17]
L d d  V g 
  2
• In the case of optical pulse, if the spectral width of the optical source is
characterized by its rms value of the Gaussian pulse   , the pulse
spreading over the length of L,  g can be well approximated by:
d g
g     DL  [3-18]
d
• D has a typical unit of [ps/(nm.km)].
Material Dispersion
Input Cladding
v g ( 1 )
Core Output
Emitter v g ( 2 )
Very short
light pulse
Intensity Intensity Intensity

Spectrum, ² 
Spread, ² 
 t t
1 o 2 0 
All excitation sources are inherently non-monochromatic and emit within a

spectrum, ² , of wavelengths. Waves in the guide with different free space
wavelengths travel at different group velocities due to the wavelength dependence
of n1. The waves arrive at the end of the fiber at different times and hence result in
a broadened output pulse.
© 1999 S.O. Kasap, Optoelectronics (Prentice Hall)
Material Dispersion
• The refractive index of the material varies as a function of wavelength, n ( )
• Material-induced dispersion for a plane wave propagation in homogeneous
medium of refractive index n:
d 2 d 2 d  2 
 mat L  L  L n (  )
dω 2c d 2c d   
L dn 
 n    [3-19]
c d 
• The pulse spread due to material dispersion is therefore:
d mat L  d 2 n
g     2  L  Dmat ( ) [3-20]
d c d
Dmat ( ) is material dispersion

Material Dispersion Diagrams

Waveguide Dispersion
• Waveguide dispersion is due to the dependency of the group velocity of the
fundamental mode as well as other modes on the V number, (see Fig 2-18
of the textbook). In order to calculate waveguide dispersion, we consider
that n is not dependent on wavelength. Defining the normalized
propagation constant b as:
 / k  n2
2 2 2
 / k  n2
b  [3-21]
n1  n2
2 2
n1  n2
• solving for propagation constant:
  n2k (1  b) [3-22]
• Using V number:
V  ka(n1  n2 )1/ 2  kan2 2
2 2 [3-23]
Waveguide Dispersion
• Delay time due to waveguide dispersion can then be expressed as:
L d (Vb) 
 wg  n2  n2 
dV 
[3-24]
c

Waveguide dispersion in single mode fibers
• For single mode fibers, waveguide dispersion is in the same order of
material dispersion. The pulse spread can be well approximated as:
d wg n2 L  d 2 (Vb)
 wg     L  Dwg ( )  V [3-25]
d c dV 2
Dwg ( )

Polarization Mode dispersion
Intensity
t
Output light pulse
z 
n1 y // y Core Ex
 = Pulse spread
Ex Ey
n1 x // x Ey
t
E
Input light pulse
Suppose that the core refractive index has different values along two orthogonal
directions corresponding to electric field oscillation direction (polarizations). We can
take x and y axes along these directions. An input light will travel along the fiber with Ex
and Ey polarizations having different group velocities and hence arrive at the output at
different times
© 1999 S.O. Kasap,Optoelectronics (Prentice Hall)

Polarization Mode dispersion
• The effects of fiber-birefringence on the polarization states of an optical are
another source of pulse broadening. Polarization mode dispersion (PMD)
is due to slightly different velocity for each polarization mode because of
the lack of perfectly symmetric & anisotropicity of the fiber. If the group
velocities of two orthogonal polarization modes are vgx and vgy then the
differential time delay  pol between these two polarization over a
distance L is
L L
 pol   [3-26]
v gx v gy
• The rms value of the differential group delay can be approximated as:
 pol  DPMD L [3-27]

Chromatic & Total Dispersion
• Chromatic dispersion includes the material & waveguide dispersions.
Dch ( )  Dmat  Dwg

[3-28]
 ch  Dch ( ) L 
• Total dispersion is the sum of chromatic , polarization dispersion and other

dispersion types and the total rms pulse spreading can be approximately
written as:
Dtotal  Dch  D pol  ...

[3-29]
 total  DtotalL 
Total Dispersion, zero Dispersion
Fact 1) Minimum distortion at wavelength about 1300 nm for single mode silica fiber.
Fact 2) Minimum attenuation is at 1550 nm for sinlge mode silica fiber.
Strategy: shifting the zero-dispersion to longer wavelength for minimum attenuation and dispersion.

Optimum single mode fiber &
distortion/attenuation characteristics
Fact 1) Minimum distortion at wavelength about 1300 nm for single mode
silica fiber.
Fact 2) Minimum attenuation is at 1550 nm for sinlge mode silica fiber.
Strategy: shifting the zero-dispersion to longer wavelength for minimum
attenuation and dispersion by Modifying waveguide dispersion by
changing from a simple step-index core profile to more complicated
profiles. There are four major categories to do that:
1- 1300 nm optimized single mode step-fibers: matched cladding (mode
diameter 9.6 micrometer) and depressed-cladding (mode diameter about 9
micrometer)
2- Dispersion shifted fibers.
3- Dispersion-flattened fibers.
4- Large-effective area (LEA) fibers (less nonlinearities for fiber optical
amplifier applications, effective cross section areas are typically greater
than 100 m2 ).
Single mode fiber dispersion

Single mode fiber dispersion

Single mode Cut-off wavelength & Dispersion
2a
• Fundamental mode is HE11 or LP01 with V=2.405 and c  n1  n2
2 2
V
• Dispersion: [3-30]
d
D ( )   Dmat ( )  Dwg ( ) [3-31]
d
  D( ) L  [3-32]
• For non-dispersion-shifted fibers (1270 nm – 1340 nm)

• For dispersion shifted fibers (1500 nm- 1600 nm)
Dispersion for non-dispersion-shifted fibers
(1270 nm – 1340 nm)
S0 0 2 2
 ( )   0  (  ) [3-33]
8 
•  0 is relative delay minimum at the zero-dispersion wavelength 0 , and S0
2
is the value of the dispersion slope in ps/(nm .km).
dD
S 0  S (0 )  [3-34]
d  0
S0 
0 4 
D ( )  1 ( )  [3-35]
4   
Dispersion for dispersion shifted fibers (1500
nm- 1600 nm)
S0
 ( )   0  (  0 ) 2 [3-36]
2
D( )  (  0 )S0 [3-37]

Example of dispersion
Performance curve for
Set of SM-fiber

Example of BW vs wavelength for various optical sources for
SM-fiber.

MFD

Bending Loss

Bending effects on loss vs MFD

Bend loss versus bend radius
a  3.6 m; b  60m

n n
  3.56 10 3 ; 3 2  0.07
n2

Kerr effect
Temporal changes in a narrow optical pulse that is subjected to Kerr nonlinearity in
A dispersive medium with positive GVD.
n  n0  n2 I Kerr nonlinearity in fiber, where I is the intensity of

Optical wave.
First-order Soliton
Temporal changes in a medium with Kerr nonlinearity and negative GVD. Since dispersion tends to broaden the pulse, Kerr
Nonlinearity tends to squeeze the pulse, resulting in a formation of optical soliton.
Chapter 4
Photonic Sources
Contents
• Review of Semiconductor Physics

• Light Emitting Diode (LED)
- Structure, Material,Quantum efficiency, LED Power,
Modulation
• Laser Diodes
- structure, Modes, Rate Equation,Quantum efficiency,
Resonant frequencies, Radiation pattern
• Single-Mode Lasers
- DFB (Distributed-FeedBack) laser, Distributed-Bragg
Reflector, Modulation
• Light-source Linearity
• Noise in Lasers
Review of Semiconductor Physics
k B  1.38 1023 JK -1
a) Energy level diagrams showing the excitation of an electron from the valence band to the conduction band.
The resultant free electron can freely move under the application of electric field.
b) Equal electron & hole concentrations in an intrinsic semiconductor created by the thermal excitation of
electrons across the band gap
n-Type Semiconductor
a) Donor level in an n-type semiconductor.

b) The ionization of donor impurities creates an increased electron concentration distribution.

p-Type Semiconductor
a) Acceptor level in an p-type semiconductor.

b) The ionization of acceptor impurities creates an increased hole concentration distribution

Intrinsic & Extrinsic Materials
• Intrinsic material: A perfect material with no impurities.
Eg
n  p  ni  exp(  ) [4-1]
2k BT
n & p & ni are the electron, hole & intrinsic concentrat ions respective ly.
E g is the gap energy, T is Temperatur e.
• Extrinsic material: donor or acceptor type semiconductors.
pn  ni
2 [4-2]
• Majority carriers: electrons in n-type or holes in p-type.

• Minority carriers: holes in n-type or electrons in p-type.
• The operation of semiconductor devices is essentially based on
the injection and extraction of minority carriers.
The pn Junction
Electron diffusion across a pn junction

creates a barrier potential (electric field)
in the depletion region.

Reverse-biased pn Junction
A reverse bias widens the depletion region, but allows minority carriers to move freely with the applied field.

Forward-biased pn Junction
Lowering the barrier potential with a forward bias allows majority carriers to diffuse across the junction.

Direct Band Gap Semiconductors
Indirect Band Gap Semiconductors
E E E
CB
Indirect Bandgap, Eg
Ec CB
Direct Bandgap Eg Photon CB Ec Er Ec
Ev kcb Phonon
Ev Ev
VB
VB kvb VB
–k k –k k –k k
(a) GaAs (b) Si (c) Si with a recombination center
(a) In GaAs the minimum of the CB is directly above the maximum of the VB. GaAs is
therefore a direct bandgap semiconductor. (b) In Si, the minimum of the CB is displaced from
the maximum of the VB and Si is an indirect bandgap semiconductor. (c) Recombination of
an electron and a hole in Si involves a recombination center .
Light-Emitting Diodes (LEDs)
• For photonic communications requiring data rate 100-200 Mb/s

with multimode fiber with tens of microwatts, LEDs are usually
the best choice.
• LED configurations being used in photonic communications:
1- Surface Emitters (Front Emitters)
2- Edge Emitters
Cross-section drawing of a typical
GaAlAs double heterostructure light
emitter. In this structure, x>y to provide
for both carrier confinement and optical
guiding.
b) Energy-band diagram showing the
active region, the electron & hole
barriers which confine the charge carriers
to the active layer.
c) Variations in the refractive index; the
lower refractive index of the material in
regions 1 and 5 creates an optical barrier
around the waveguide because of the higher
band-gap energy of this material.
1.240
 (m)  [4-3]
Eg (eV)

Surface-Emitting LED
Schematic of high-radiance surface-emitting LED. The active region is limitted

to a circular cross section that has an area compatible with the fiber-core end face.

Edge-Emitting LED
Schematic of an edge-emitting double heterojunction LED. The output beam is

lambertian in the plane of junction and highly directional perpendicular to pn junction.
They have high quantum efficiency & fast response.
Light Source Material
• Most of the light sources contain III-V ternary & quaternary
compounds.
• Ga 1x Al x As by varying x it is possible to control the band-gap
energy and thereby the emission wavelength over the range of
800 nm to 900 nm. The spectral width is around 20 to 40 nm.
• In1 x Ga x As y P1 y By changing 0<x<0.47; y is approximately 2.2x,
the emission wavelength can be controlled over the range of
920 nm to 1600 nm. The spectral width varies from 70 nm to
180 nm when the wavelength changes from 1300 nm to 1600
nm. These materials are lattice matched.
Spectral width of LED types

Rate equations, Quantum Efficiency & Power of
LEDs
• When there is no external carrier injection, the excess density
decays exponentially due to electron-hole recombination.
n(t )  n0 e t / [4-4]
• n is the excess carrier density,
n0 : initial injected excess electron density

 : carrier lifetime.
• Bulk recombination rate R:
dn n
R  [4-5]
dt 
• Bulk recombination rate (R)=Radiative recombination rate +
nonradiative recombination rate
bulk recombinat ion rate ( R  1/τ ) 
radiative recombinat ion rate ( Rr  1/τ r )  nonradiati ve recombinat ion rate( Rnr  1/τ nr )
With an external supplied current density of J the rate equation for the electron-hole
recombination is:
dn(t ) J n
  [4-6]
dt qd 
q : charge of the electron; d : thickness of recombinat ion region
In equilibrium condition: dn/dt=0
J
n [4-7]
qd
Internal Quantum Efficiency & Optical Power
Rr  nr 
int    [4-8]
Rr  Rnr  r   nr  r
int : internal quantum efficiency in the active region
Optical power generated internally in the active region in the LED is:
I hcI
Pint  int h  int
q
[4-9]
q
Pint : Internal optical power,
I : Injected current to active region
External Quantum Eficiency
# of photons emitted from LED

ext  [4-10]
# of LED internally generated photons
• In order to calculate the external quantum efficiency, we need to

consider the reflection effects at the surface of the LED. If we
consider the LED structure as a simple 2D slab waveguide, only
light falling within a cone defined by critical angle will be emitted
from an LED.
c
1
ext  
4 0
T ( )(2 sin  )d [4-11]
4n1n2
T ( ) : Fresnel Transmissi on Coefficien t  T (0)  [4-12]
(n1  n2 ) 2
1
If n2  1  ext  [4-13]
n1 (n1  1) 2
Pint
LED emitted optical powr, P  ext Pint  [4-14]
n1 (n1  1) 2
Modulation of LED
• The frequency response of an LED depends on:
1- Doping level in the active region
2- Injected carrier lifetime in the recombination region, .
i
3- Parasitic capacitance of the LED
• If the drive current of an LED is modulated at a frequency of w
the output optical power of the device will vary as:
P0
P (w )  [4-15]
1  (w i ) 2
• Electrical current is directly proportional to the optical power,
thus we can define electrical bandwidth and optical bandwidth,
separately.
 p(w)   I(w) 
Electrical BW  10log    20 log  I (0) 
[4-16]
 p ( 0)   
p : electrical power, I : electrical current
 P(w )   I (w ) 
Optical BW  10 log    10 log   [4-17]
 P ( 0)   I ( 0) 

LASER
(Light Amplification by the Stimulated Emission of Radiation)
• Laser is an optical oscillator. It comprises a resonant optical

amplifier whose output is fed back into its input with matching
phase. Any oscillator contains:
1- An amplifier with a gain-saturated mechanism
2- A feedback system
3- A frequency selection mechanism
4- An output coupling scheme
• In laser the amplifier is the pumped active medium, such as
biased semiconductor region, feedback can be obtained by
placing active medium in an optical resonator, such as Fabry-
Perot structure, two mirrors separated by a prescribed distance.
Frequency selection is achieved by resonant amplifier and by
the resonators, which admits certain modes. Output coupling is
accomplished by making one of the resonator mirrors partially
transmitting.
Pumped active medium
• Three main process for laser action:
1- Photon absorption
2- Spontaneous emission
3- Stimulated emission

Lasing in a pumped active medium
• In thermal equilibrium the stimulated emission is essentially

negligible, since the density of electrons in the excited state is
very small, and optical emission is mainly because of the
spontaneous emission. Stimulated emission will exceed
absorption only if the population of the excited states is greater
than that of the ground state. This condition is known as
Population Inversion. Population inversion is achieved by
various pumping techniques.
• In a semiconductor laser, population inversion is accomplished

by injecting electrons into the material to fill the lower energy
states of the conduction band.
Fabry-Perot Resonator
Relative intensity
M1 M2 m=1
A 1 f R ~ 0.8
m=2 R ~ 0.4
 m
B
L m=8 
m - 1 m m + 1
(a) (b) (c)
Resonant modes : kL  m m  1,2,3,..
Schematic illustration of the Fabry-Perot optical cavity and its properties. (a) Reflected
waves interfere. (b) Only standing EM waves, modes, of certain wavelengths are allowed
in the cavity. (c) Intensity vs. frequency for various modes.R is mirror reflectance and
lower R means higher loss from the cavity.
(1  R) 2
I trans  I inc [4-18]
(1  R) 2  4R sin 2 (kL)
R: reflectance of the optical intensity, k: optical wavenumber
Laser Diode
• Laser diode is an improved LED, in the sense that uses stimulated
emission in semiconductor from optical transitions between distribution
energy states of the valence and conduction bands with optical
resonator structure such as Fabry-Perot resonator with both optical
and carrier confinements.

Laser Diode Characteristics
• Nanosecond & even picosecond response time (GHz BW)

• Spectral width of the order of nm or less
• High output power (tens of mW)
• Narrow beam (good coupling to single mode fibers)
• Laser diodes have three distinct radiation modes namely,

longitudinal, lateral and transverse modes.
• In laser diodes, end mirrors provide strong optical feedback in

longitudinal direction, so by roughening the edges and cleaving
the facets, the radiation can be achieved in longitudinal direction
rather than lateral direction.
DFB(Distributed FeedBack) Lasers
• In DFB lasers, the optical resonator structure is due to the incorporation
of Bragg grating or periodic variations of the refractive index into
multilayer structure along the length of the diode.

Laser Operation & Lasing Condition
• To determine the lasing condition and resonant frequencies, we
should focus on the optical wave propagation along the
longitudinal direction, z-axis. The optical field intensity, I, can be
written as:
I ( z, t )  I ( z )e j (wt  z ) [4-19]
• Lasing is the condition at which light amplification becomes

possible by virtue of population inversion. Then, stimulated
emission rate into a given EM mode is proportional to the
intensity of the optical radiation in that mode. In this case, the
loss and gain of the optical field in the optical path determine the
lasing condition. The radiation intensity of a photon at energy h
varies exponentially with a distance z amplified by factor g, and
attenuated by factor  according to the following relationship:
I ( z)  I (0) expg (h )   (h )z [4-20]
R1 n1 R2
Z=0 n2 Z=L
I (2L)  I (0) R1R2 expg (h )   (h )(2L) [4-21]
 : Optical confinemen t factor, g : gain coefficien t

2
 n1  n2 
α : effective absorption coefficien t, R   
 n1  n2 
Lasing Conditions:
I ( 2 L )  I ( 0)
[4-22]
exp(  j 2 L)  1
Threshold gain & current density
1  1 
gth    ln   [4-23]
2 L  R1R2 
Laser starts to " lase" iff : g  gth
For laser structure with strong carrier confinement, the threshold current
Density for stimulated emission can be well approximated by:
gth  J th [4-24]
 : constant depends on specific device constructi on

Optical output vs. drive current

Semiconductor laser rate equations
• Rate equations relate the optical output power, or # of photons per unit
volume,  , to the diode drive current or # of injected electrons per
unit volume, n. For active (carrier confinement) region of depth d, the
rate equations are:
d 
 Cn  Rsp 
dt  ph
Photonratestimulated emission spontaneous emission photon loss [4-25]
dn J n
   Cn
dt qd  sp
electron rate  injection  spontaneous recombination  stimulated emission
C : Coefficien t expressing the intensityof the opticalemission & absorptionprocess

Rsp :rate of spontaneous emission into the lasing mode
 ph : photonlife time
J :Injectioncurrent density
Threshold current Density & excess electron density
• At the threshold of lasing:   0, d / dt  0, Rsp  0
1
from eq. [4 - 25]  Cn   /  ph  0  n   nth [4-26]
C ph
• The threshold current needed to maintain a steady state threshold

concentration of the excess electron, is found from electron rate
equation under steady state condition dn/dt=0 when the laser is just
about to lase:
J th nth nth
0   J th  qd [4-27]
qd  sp  sp
Laser operation beyond the threshold
J  J th
• The solution of the rate equations [4-25] gives the steady state
photon density, resulting from stimulated emission and
spontaneous emission as follows:
 ph
s  ( J  J th )   ph Rsp [4-28]
qd
External quantum efficiency
• Number of photons emitted per radiative electron-hole pair

recombination above threshold, gives us the external quantum
efficiency.
i ( g th   )
ext 
g th
q dP dP (mW )
  0.8065[ m] [4-29]
E g dI dI (mA )
• Note that: i  60%  70%; ext  15%  40%

Laser Resonant Frequencies
• Lasing condition, namely eq. [4-22]:
exp(  j 2 L)  1  2 L  2m , m  1,2,3,...
2n
• Assuming  the resonant frequency of the mth
mode is: 
mc
m  m  1,2,3,... [4-30]
2 Ln
c 2
   m  m1     [4-31]
2 Ln 2 Ln
Spectrum from a laser Diode
 (  0 ) 
g ( )  g (0) exp    : spectral width [4-32]
 2 
2
Laser Diode Structure & Radiation Pattern
• Efficient operation of a laser diode requires reducing the # of

lateral modes, stabilizing the gain for lateral modes as well as
lowering the threshold current. These are met by structures that
confine the optical wave, carrier concentration and current flow
in the lateral direction. The important types of laser diodes are:
gain-induced, positive index guided, and negative index
guided.
(a) gain-induced guide (b)positive-index waveguide (c)negative-index waveguide
Laser Diode with buried heterostructure (BH)
Single Mode Laser
• Single mode laser is mostly based on the index-
guided structure that supports only the fundamental
transverse mode and the fundamental longitudinal
mode. In order to make single mode laser we have
four options:
1- Reducing the length of the cavity to the point
where the frequency separation given in eq[4-31] of
the adjacent modes is larger than the laser transition
line width. This is hard to handle for fabrication and
results in low output power.
2- Vertical-Cavity Surface Emitting laser (VCSEL)
3- Structures with built-in frequency selective grating
4- tunable laser diodes
.
VCSEL
Frequency-Selective laser Diodes:
Distributed Feedback (DFB) laser
2ne 
B  [4-33]
k
Distributed Feedback Reflector (DBR) laser
B 2
1
  B  (m  )
2ne Le 2
[4-35]
Output spectrum symmetrically distributed around Bragg wavelength in an idealized DFB laser diode
Distributed Reflector (DR) laser
Modulation of Laser Diodes
• Internal Modulation: Simple but suffers from non-linear effects.
• External Modulation: for rates greater than 2 Gb/s, more
complex, higher performance.
• Most fundamental limit for the modulation rate is set by the
photon life time in the laser cavity:
1 c 1 1  c
   ln   g th
 ph
[4-36]
n 2L R1 R2  n
• Another fundamental limit on modulation frequency is the
relaxation oscillation frequency given by:
1/ 2
1 1  I 
f    1 [4-37]
2  sp ph  I th 
Relaxation oscillation peak
Pulse Modulated laser
• In a pulse modulated laser, if the laser is completely turned off
after each pulse, after onset of the current pulse, a time
t d delay,
given by:
 Ip 
t d   ln   [4-38]
 I p  ( I B  I th ) 
 : carrier life time I p : Current pulse amplitude

I B : Bias current
Temperature variation of the threshold
current
I th (T )  I z e T / T0
Linearity of Laser
Information carrying LED or Laser diode

electrical signal s(t) modulator
Optical putput power:

P(t)=P[1+ms(t)]
Nonlinearity
x(t) Nonlinear function y=f(x) y(t)
x(t )  A cos wt
y (t )  A0  A1 cos wt  A2 cos 2wt  ...
Nth order harmonic distortion:
 An 
20 log  
 A1 
Intermodulation Distortion
x(t )  A1 cos w1t  A2 cos w 2 t 

y (t )   Bmn cos( mw1  nw 2 )t m,n  0,1,2,...
m,n
nw1 , mw 2
Harmonics:
Intermodulated Terms:
w1  w 2 ,2w1  w 2 ,w1  2w 2 ,...

Laser Noise
• Modal (speckel) Noise: Fluctuations in the distribution of

energy among various modes.
• Mode partition Noise: Intensity fluctuations in the longitudinal
modes of a laser diode, main source of noise in single mode
fiber systems.
• Reflection Noise: Light output gets reflected back from the fiber
joints into the laser, couples with lasing modes, changing their
phase, and generate noise peaks. Isolators & index matching
fluids can eliminate these reflections.
Chapter 4
Photonic Sources
Contents
• Review of Semiconductor Physics

• Light Emitting Diode (LED)
- Structure, Material,Quantum efficiency, LED Power,
Modulation
• Laser Diodes
- structure, Modes, Rate Equation,Quantum efficiency,
Resonant frequencies, Radiation pattern
• Single-Mode Lasers
- DFB (Distributed-FeedBack) laser, Distributed-Bragg
Reflector, Modulation
• Light-source Linearity
• Noise in Lasers
Review of Semiconductor Physics
k B  1.38 1023 JK -1
a) Energy level diagrams showing the excitation of an electron from the valence band to the conduction band.
The resultant free electron can freely move under the application of electric field.
b) Equal electron & hole concentrations in an intrinsic semiconductor created by the thermal excitation of
electrons across the band gap
n-Type Semiconductor
a) Donor level in an n-type semiconductor.

b) The ionization of donor impurities creates an increased electron concentration distribution.

p-Type Semiconductor
a) Acceptor level in an p-type semiconductor.

b) The ionization of acceptor impurities creates an increased hole concentration distribution

Intrinsic & Extrinsic Materials
• Intrinsic material: A perfect material with no impurities.
Eg
n  p  ni  exp(  ) [4-1]
2k BT
n & p & ni are the electron, hole & intrinsic concentrat ions respective ly.
E g is the gap energy, T is Temperatur e.
• Extrinsic material: donor or acceptor type semiconductors.
pn  ni
2 [4-2]
• Majority carriers: electrons in n-type or holes in p-type.

• Minority carriers: holes in n-type or electrons in p-type.
• The operation of semiconductor devices is essentially based on
the injection and extraction of minority carriers.
The pn Junction
Electron diffusion across a pn junction

creates a barrier potential (electric field)
in the depletion region.

Reverse-biased pn Junction
A reverse bias widens the depletion region, but allows minority carriers to move freely with the applied field.

Forward-biased pn Junction
Lowering the barrier potential with a forward bias allows majority carriers to diffuse across the junction.

Direct Band Gap Semiconductors
Indirect Band Gap Semiconductors
E E E
CB
Indirect Bandgap, Eg
Ec CB
Direct Bandgap Eg Photon CB Ec Er Ec
Ev kcb Phonon
Ev Ev
VB
VB kvb VB
–k k –k k –k k
(a) GaAs (b) Si (c) Si with a recombination center
(a) In GaAs the minimum of the CB is directly above the maximum of the VB. GaAs is
therefore a direct bandgap semiconductor. (b) In Si, the minimum of the CB is displaced from
the maximum of the VB and Si is an indirect bandgap semiconductor. (c) Recombination of
an electron and a hole in Si involves a recombination center .
Light-Emitting Diodes (LEDs)
• For photonic communications requiring data rate 100-200 Mb/s

with multimode fiber with tens of microwatts, LEDs are usually
the best choice.
• LED configurations being used in photonic communications:
1- Surface Emitters (Front Emitters)
2- Edge Emitters
Cross-section drawing of a typical
GaAlAs double heterostructure light
emitter. In this structure, x>y to provide
for both carrier confinement and optical
guiding.
b) Energy-band diagram showing the
active region, the electron & hole
barriers which confine the charge carriers
to the active layer.
c) Variations in the refractive index; the
lower refractive index of the material in
regions 1 and 5 creates an optical barrier
around the waveguide because of the higher
band-gap energy of this material.
1.240
 (m)  [4-3]
Eg (eV)

Surface-Emitting LED
Schematic of high-radiance surface-emitting LED. The active region is limitted

to a circular cross section that has an area compatible with the fiber-core end face.

Edge-Emitting LED
Schematic of an edge-emitting double heterojunction LED. The output beam is

lambertian in the plane of junction and highly directional perpendicular to pn junction.
They have high quantum efficiency & fast response.
Light Source Material
• Most of the light sources contain III-V ternary & quaternary
compounds.
• Ga 1x Al x As by varying x it is possible to control the band-gap
energy and thereby the emission wavelength over the range of
800 nm to 900 nm. The spectral width is around 20 to 40 nm.
• In1 x Ga x As y P1 y By changing 0<x<0.47; y is approximately 2.2x,
the emission wavelength can be controlled over the range of
920 nm to 1600 nm. The spectral width varies from 70 nm to
180 nm when the wavelength changes from 1300 nm to 1600
nm. These materials are lattice matched.
Spectral width of LED types

Rate equations, Quantum Efficiency & Power of
LEDs
• When there is no external carrier injection, the excess density
decays exponentially due to electron-hole recombination.
n(t )  n0 e t / [4-4]
• n is the excess carrier density,
n0 : initial injected excess electron density

 : carrier lifetime.
• Bulk recombination rate R:
dn n
R  [4-5]
dt 
• Bulk recombination rate (R)=Radiative recombination rate +
nonradiative recombination rate
bulk recombinat ion rate ( R  1/τ ) 
radiative recombinat ion rate ( Rr  1/τ r )  nonradiati ve recombinat ion rate( Rnr  1/τ nr )
With an external supplied current density of J the rate equation for the electron-hole
recombination is:
dn(t ) J n
  [4-6]
dt qd 
q : charge of the electron; d : thickness of recombinat ion region
In equilibrium condition: dn/dt=0
J
n [4-7]
qd
Internal Quantum Efficiency & Optical Power
Rr  nr 
int    [4-8]
Rr  Rnr  r   nr  r
int : internal quantum efficiency in the active region
Optical power generated internally in the active region in the LED is:
I hcI
Pint  int h  int
q
[4-9]
q
Pint : Internal optical power,
I : Injected current to active region
External Quantum Eficiency
# of photons emitted from LED

ext  [4-10]
# of LED internally generated photons
• In order to calculate the external quantum efficiency, we need to

consider the reflection effects at the surface of the LED. If we
consider the LED structure as a simple 2D slab waveguide, only
light falling within a cone defined by critical angle will be emitted
from an LED.
c
1
ext  
4 0
T ( )(2 sin  )d [4-11]
4n1n2
T ( ) : Fresnel Transmissi on Coefficien t  T (0)  [4-12]
(n1  n2 ) 2
1
If n2  1  ext  [4-13]
n1 (n1  1) 2
Pint
LED emitted optical powr, P  ext Pint  [4-14]
n1 (n1  1) 2
Modulation of LED
• The frequency response of an LED depends on:
1- Doping level in the active region
2- Injected carrier lifetime in the recombination region, .
i
3- Parasitic capacitance of the LED
• If the drive current of an LED is modulated at a frequency of w
the output optical power of the device will vary as:
P0
P (w )  [4-15]
1  (w i ) 2
• Electrical current is directly proportional to the optical power,
thus we can define electrical bandwidth and optical bandwidth,
separately.
 p(w)   I(w) 
Electrical BW  10log    20 log  I (0) 
[4-16]
 p ( 0)   
p : electrical power, I : electrical current
 P(w )   I (w ) 
Optical BW  10 log    10 log   [4-17]
 P ( 0)   I ( 0) 

Chapter 5
Laser-Fiber Connection
Content
• Launching optical power into a fiber

• Fiber-to-Fiber coupling
• Fiber Splicing and connectors
Coupling Efficiency
power coupled into the fiber PF

  [5-1]
power emitted from the sourse Ps
Ps PF
Source Optical Fiber
Radiance (Brightness) of the source
• B= Optical power radiated from a unit area of the source into a

unit solid angle [watts/(square centimeter per stradian)]
Surface emitting LEDs have a Lambertian pattern:
B( ,  )  B0 cos [5-2]

Edge emitting LEDs and laser diodes radiation pattern
1 sin  cos 
2 2
  [5-3]
B( ,  ) B0 cos  B0 cos 

T L
For edge emitting LEDs, L=1

Power Coupled from source to the fiber
As and  s : area and solid emission angle of the source

 
PF     B( As ,  s )d s dAs 
A f and  f : area and Af 
 f  [5-4]
solid acceptance angle of fiber rm 2
2  0 max 
      B( ,  ) sin dd  d s rdr
0 0 0 0 
Power coupled from LED to the Fiber
 0 max
rs 2
 
P  2B0 cos  sin d d s rdr
0  0 
0  
rs 2
 B0    0 max d s rdr
sin 2
0 0
rs 2 2
 B0   NA d rdr
s
0 0
PLED,step   rs B0 ( NA)  2 rs B0 n1 
2 2 2 2 2 2
[5-5]
Power coupling from LED to step-index fiber
• Total optical power from LED:
2  / 2
Ps  As 
0 0
B( ,  ) sin dd
 /2
Ps  rs 2B0      
2 2 2
cos sin d rs B0 [5-6]
0
Ps ( NA) 2 if rs  a 
 
PLED,step   a  2  [5-7]

  s P ( NA) 2
if rs  a 
 rs  
Equilibrium Numerical Aperture
Examples of possible lensing schemes used to improve optical source-to-fiber coupling
efficiency
Laser diode to Fiber Coupling
Fiber-to-Fiber Joint
• Fiber-to-Fiber coupling loss:
LF [dB]  10 log  F [5-8]
• Low loss fiber-fiber joints are either:

1- Splice (permanent bond)
2- Connector (demountable connection)
Different modal distribution of the optical beam emerging from a fiber lead to different degrees of
coupling loss. a) when all modes are equally excited, the output beam fills the entire output NA.
b) for a steady state modal distribution, only the equilibrium NA is filled by the output beam.
Mechanical misalignment losses
Lateral (axial) misalignment loss is a dominant

Mechanical loss.
1/ 2
d   d  
2
Acomm 2 d
 F ,step   arccos  1     [5-9]
a 2
 2a a   2a  
Longitudinal offset effect
Losses due to differences in the geometry and waveguide characteristics

of the fibers
aR
LF (a)  10 log( ) for a R  a E
aE [5-10]
NA R
LF (a)  20 log( ) for NA R  NA E
NA E
E & R subscripts refer to emitting and receiving fibers.
Experimental comparison of Loss as a function
of mechanical misalignment
Fiber end face
Fiber end defects

Fiber splicing
Fusion Splicing
V-groove optical fiber splicing
Optical Fiber Connectors
• Some of the principal requirements of a good connector design are as
follows:
1- low coupling losses
2- Interchangeability
3- Ease of assembly
4- Low environmental sensitivity
5- Low-cost and reliable construction
6- Ease of connection
Connector Return Loss
Chapter 6
Photodetectors
Content
• Physical Principles of Photodiodes
• pin, APD
• Photodetectors characteristics (Quantum efficiency,
Responsivity, S/N)
• Noise in Photodetector Circuits
• Photodiode Response Time
• Photodiodes structures
pin Photodetector
The high electric field present in the depletion region causes photo-generated carriers to
Separate and be collected across the reverse –biased junction. This give rise to a current
Flow in an external circuit, known as photocurrent.
Energy-Band diagram for a pin photodiode
Photocurrent
• Optical power absorbed,P (x )in the depletion region can be written in terms
of incident optical power, P0 :
 s (  ) x
P( x)  P0 (1  e ) [6-1]
• Absorption coefficient  s ( ) strongly depends on wavelength. The upper

wavelength cutoff for any semiconductor can be determined by its energy
gap as follows:
1.24
c ( m)  [6-2]
E g (eV)
• Taking entrance face reflectivity into consideration, the absorbed power in
the width of depletion region, w, becomes:
(1  R f ) P(w)  P0 (1  e  s ( ) w )(1  R f )
Optical Absorption Coefficient
Responsivity
• The primary photocurrent resulting from absorption is:
q
Ip  P0 (1  e  s (  ) w )(1  R f ) [6-3]
h
• Quantum Efficiency:
# of electron - hole photogener ated pairs


# of incident photons
[6-4]
IP / q

P0 / h
• Responsivity:
IP q
  [A/W] [6-5]
P0 h
Responsivity vs. wavelength
Avalanche Photodiode (APD)
APDs internally multiply the
primary photocurrent before it
enters to following circuitry.
In order to carrier multiplication
take place, the photogenerated
carriers must traverse along a
high field region. In this region,
photogenerated electrons and
holes gain enough energy to
ionize bound electrons in VB
upon colliding with them. This
multiplication is known as Optical radiation
impact ionization. The newly
created carriers in the presence of Reach-Through APD structure (RAPD)
high electric field result in more showing the electric fields in depletion
ionization called avalanche region and multiplication region.
effect.
Responsivity of APD
• The multiplication factor (current gain) M for all carriers generated in the
photodiode is defined as:
IM
M  [6-6]
Ip
• Where I M is the average value of the total multiplied output current & I P
is the primary photocurrent.
• The responsivity of APD can be calculated by considering the current gain

as:
q
 AP D  M  0 M [6-7]
h
Current gain (M) vs. Voltage for different optical
wavelengths
Photodetector Noise & S/N
• Detection of weak optical

signal requires that the
photodetector and its
following amplification
circuitry be optimized for a
desired signal-to-noise
ratio.
• It is the noise current
which determines the
minimum optical power
level that can be detected.
This minimum detectable
optical power defines the
sensitivity of
photodetector. That is the S signal power from photocurre nt
optical power that 
N photodetec tor noise power  amplifier noise power
generates a photocurrent
with the amplitude equal to
that of the total noise
current (S/N=1)
Signal Calculation
• Consider the modulated optical power signal P(t) falls on the photodetector
with the form of:
P(t )  P0 [1  ms(t )] [6-8]
• Where s(t) is message electrical signal and m is modulation index.

Therefore the primary photocurrent is (for pin photodiode M=1):
q
iph  MP (t )  I P [DC value ]  i p (t )[ AC current ] [6-9]
h
• The root mean square signal current is then:
 ip M   s
2 2 2 2 [6-9]
is
m 2 I P2
 p 
2 2
ip for sinusoidal signal [6-10]
2
Noise Sources in Photodetecors
• The principal noises associated with photodetectors are :
1- Quantum (Shot) noise: arises from statistical nature of the production
and collection of photo-generated electrons upon optical illumination. It has
been shown that the statistics follow a Poisson process.
2- Dark current noise: is the current that continues to flow through the
bias circuit in the absence of the light. This is the combination of bulk
dark current, which is due to thermally generated e and h in the pn
junction, and the surface dark current, due to surface defects, bias voltage
and surface area.
• In order to calculate the total noise presented in photodetector, we should
sum up the root mean square of each noise current by assuming that those
are uncorrelated.
• Total photodetector noise current=quantum noise current +bulk dark

current noise + surface current noise
Noise calculation (1)
• Quantum noise current (lower limit on the sensitivity):
  Q  2qI P BM F (M )
2 2 2
iQ [6-11]
B: Bandwidth, F(M) is the noise figure and generally is F ( M )  M 0  x  1.0

x
•
• Bulk dark current noise:
  DB  2qI D BM 2 F ( M )
2 2
i DB [6-12]
Note that for pin photodiode

I D is bulk dark current
M 2 F (M )  1
• Surface dark current noise: IL is the surface current.
  DS  2qI L B
2 2
i DS [6-13]
Noise calculation (2)
• The total rms photodetector noise current is:
  N  iQ  i DB  i DS
2 2 2 2 2
iN
 2q( I P  I D ) BM 2 F ( M )  2qI L B [6-14]
• The thermal noise of amplifier connected to the photodetector is:
4k BTB
 T 
2 2
iT [6-15]
RL
RL input resistance of amplifier, and k B  1.38  10 23 JK -1 is Boltzmann cte.

S/N Calculation
• Having obtained the signal and total noise, the signal-to-noise-ratio can be
written as:
2
S iP M 2
 [6-16]
N 2q( I P  I D ) BM 2 F ( M )  2qI L B  4k BTB / RL
• Since the noise figure F(M) increases with M, there always exists an
optimum value of M that maximizes the S/N. For sinusoidally modulated
signal with m=1 and F ( M )  M x :
x2 2qI L  4k BT / RL

[6-17]
M
xq( I P  I D )
opt
Photodetector Response Time
• The response time of a photodetector with its output circuit depends mainly
on the following three factors:
1- The transit time of the photocarriers in the depletion region. The transit
time t ddepends on the carrier drift velocity v d and the depletion layer
width w, and is given by:
w
td  [6-18]
vd
2- Diffusion time of photocarriers outside depletion region.
3- RC time constant of the circuit. The circuit after the photodetector acts
like RC low pass filter with a passband given by:
1
B [6-19]
2RT CT
RT  Rs || RL and CT  Ca  Cd
Photodiode response to optical pulse
Typical response time of the

photodiode that is not fully depleted
Various optical responses of photodetectors:
Trade-off between quantum efficiency & response time
• To achieve a high quantum
efficiency, the depletion layer
width must be larger than 1 /  s
(the inverse of the absorption
coefficient), so that most of the
light will be absorbed. At the
same time with large width, the
capacitance is small and RC
time constant getting smaller,
leading to faster response, but
wide width results in larger
transit time in the depletion
region. Therefore there is a
trade-off between width and
QE. It is shown that the best
is:
1/  s  w  2 /  s
Structures for InGaAs APDs
• Separate-absorption-and multiplication (SAM) APD
light
InP substrate
InP buffer layer
INGaAs Absorption layer
InP multiplication layer
Metal contact
• InGaAs APD superlattice structure (The multiplication region is composed

of several layers of InAlGaAs quantum wells separated by InAlAs barrier
layers.
Temperature effect on avalanche gain
Comparison of photodetectors
Chapter 7
Photonic Transmission
Systems (Digital & Analog)
Content
• Digital Photonic transmission System

• Digital Photonic Receiver (BER, Quantum Limit)
• Analog Photonic Transmission System
• Photonic Digital Link Analysis & Design
- Link Loss budget, Link Power budget, Rise Time budget
• System Rise Time and Information Rate
Digital Transmission System (DTS)
• The design of optical receiver is much more complicated than that of optical transmitter
because the receiver must first detect weak, distorted signals and the n make decisions
on what type of data was sent.
Error Sources in DTS
  
N  
h 0
P (t ) dt 
h
E [7-1]
e N
Pr (n)  Nn
[7-2]
n!
N is the average number of electron-hole pairs in photodetector,
 is the detector quantum efficiency and E is energy received in a time
interval  and h is photon energy, where Pr (n) is the probability

that n electrons are emitted in an interval .
InterSymbol Interference (ISI)
Pulse spreading in an optical signal, after traversing along optical fiber,

leads to ISI. Some fraction of energy remaining in appropriate time slot
is designated by  , so the rest is the fraction of energy that has spread
Into adjacent time slots.
Receiver Configuration
The binary digital pulse train incident on the photodetector can be written in the
following form:

P(t )  b h
n  
n p (t  nTb ) [7-3]
where Tb is bit period, bn is an amplitude parameter of the nth message digit

and h p (t )is the received pulse shape which is positive for all t.
• In writing down eq. [7-3], we assume the digital pulses with amplitude V
represents bit 1 and 0 represents bit 0. Thus bn can take two values
corresponding to each binary data. By normalizing the input pulse h p (t ) to
the photodiode to have unit area

h

p (t )dt  1
bn represents the energy in the nth pulse.
the mean output current from the photodiode at time t resulting from pulse
train given in eq. [7-3] is (neglecting the DC components arising from dark
current noise):
q 
i (t )  MP (t )   o M  bn h p (t  nTb ) [7-4]
h n  
Bit Error Rate (BER)
BER  Probabilit y of Error 

# of error over a certain ti me interval t
 [7-5]
total # of pulses transmitt ed during t

Ne Ne
 B  1 / Tb
Nt Bt
• Probability of Error= probability that the output voltage is

less than the threshold when a 1 is sent + probability that the
output voltage is more than the threshold when a 0 has been
sent.
vth
Probability distributions for received logical 0 and 1 signal pulses.

the different widths of the two distributions are caused by various signal
distortion effects.
v
P1 (v)   p( y | 1)dy

probablity that the equalizer output vol tage is less than v, if 1 transmitt ed

[7-6]
P0 (v)   p( y | 0)dy probablity that the equalizer output vol tage exceeds v, if 0 transmitt ed
v
Pe  q1 P1 (vth )  q 0 P0 (vth )
vth  [7-7]
 q1  p( y | 1)dy  q  p( y | 1)dy

0
vth
• Where q1 and q0 are the probabilities that the transmitter sends 0 and 1
respectively. q  1  q
0 1
• For an unbiased transmitter q0  q1  0.5

Gaussian Distribution
vth
1
vth
 (v  bon ) 2 
P1 (vth )   p( y | 1)dy 
 2  on
  exp 
  2 on 
2 dv
[7-8]
 
1  (v  boff ) 2 
P0 (vth )   p( y | 0)dy 
vth 2  off
  exp 
vth  2 off 
2 dv
mean
mean
• If we assume that the probabilities of 0 and 1 pulses are equally likely, then
using eq [7-7] and [7-8] , BER becomes:

1 1 Q 
BER  Pe (Q )   exp(  x )dx  1  erf (
2
)
 Q/ 2 2 2 
1 exp(- Q 2 /2)
 [7-9]
2 Q
vth  boff bon  vth

Q  [7-9]
 off  on
x
2
erf ( x)    2
exp( y )dy [7-10]
 0
Approximation of error function
Variation of BER vs Q,
according to eq [7-9].
Special Case
In special case when:
 off   on   & boff  0, bon  V

From eq [7-29], we have: vth  V / 2
Eq [7-8] becomes:
1 V 
Pe ( )  1  erf ( )
2 2 2 
V [7-11]
is peak signal - to - rms - noise ratio.

Study example 7-1 pp. 286 of the textbook.

Quantum Limit
• Minimum received power required for a specific BER assuming that the
photodetector has a 100% quantum efficiency and zero dark current. For
such ideal photo-receiver,
Pe  P1 (0)  exp(  N ) [7-12]
• Where N is the average number of electron-hole pairs, when the incident

optical pulse energy is E and given by eq [7-1] with 100% quantum
efficiency (  1) .
• Eq [7-12] can be derived from eq [7-2] where n=0.
• Note that, in practice the sensitivity of receivers is around 20 dB higher

than quantum limit because of various nonlinear distortions and noise
effects in the transmission link.
Analog Transmission System
• In photonic analog transmission Analog LED modulation

system the performance of the system
is mainly determined by signal-to-
noise ratio at the output of the
receiver. I
• In case of amplitude modulation the
m
IB
transmitted optical power P(t) is in
the form of:
P(t )  Pt [1  ms(t )]
where m is modulation index, and s(t)
is analog modulation signal.
• The photocurrent at receiver can be
expressed as:
is (t )   0 MPr [1  ms(t )] [7-13]

• By calculating mean square of the signal and mean square of the total
noise, which consists of quantum, dark and surface leakage noise currents
plus resistance thermal noise, the S/N can be written as:
S i s2 (1 / 2)( 0 MmPr ) 2
 2 
N iN 2q ( 0 Pr  I D ) M 2 F ( M ) B  ( 4k B TB / Req ) Ft
2 [7-14]
(1 / 2)( MmI P )

2q ( I P  I D ) M 2 F ( M ) B  ( 4k B TB / Req ) Ft
I P : primary photocurre nt   0 Pr ; I D : primary bulk dark current;

I L : Surface - leakage current; F ( M ) : excess photodiode noise factor  M x
B : effective noise bandwidth; Req : equivalent resistance of photodetec tor load and amplifier
Ft : noise figure of baseband amplifier; Pr : average received optical power
pin Photodiode S/N
• For pin photodiode, M=1:
(1 / 2)m 2  0 Pr
2 2
S (1 / 2)( I P m) 2
  Low input signal level [7-15]
N (4k BTB / Req ) Ft (4k BTB / Req ) F
S m 2
 0 Pr
 Large signal level [7-16]
N 4qB
SNR vs. optical power for photodiodes
Photonic Digital Link Analysis & Design
• Point-to-Point Link Requirement:

- Data Rate
- BER
- Distance
- Cost & Complexity
• Analysis Methods:
- Link loss & S/N analysis (link power budget analysis and loss
allocation) for a prescribed BER
- Dispersion (rise-time) analysis (rise-time budget allocation)
System Design Choices:
Photodetector, Optical Source, Fiber
• Photodetectors: Compared to APD, PINs are less expensive
and more stablewith temperature. However PINs have lower
sensitivity.
• Optical Sources:
1- LEDs: 150 (Mb/s).km @ 800-900 nm and larger than 1.5
(Gb/s).km @ 1330 nm
2- InGaAsP lasers: 25 (Gb/s).km @ 1330 nm and ideally around
500 (Gb/s).km @ 1550 nm. 10-15 dB more power. However
more costly and more complex circuitry.
• Fiber:
1- Single-mode fibers are often used with lasers or edge-emitting
LEDs.
2- Multi-mode fibers are normally used with LEDs. NA and 
should be optimized for any particular application.
Link Power/Loss Analysis
PT [dB]  Ps [dBm ]  PR [dBm ] Total Power Loss
PT  2lc [dB]   f [dB / km]  L[km]  System Margin

Receiver Sensitivities vs. Bit Rate
9
The Si PIN & APD and InGaAsP PIN plots for BER= 10 . The InGaAs APD plot is for
BER= 10 .11
Link Loss Budget [Example 8.1]
Link Power Budget Table [Example 8.2]
• Example: [SONET Component/loss Output/sensitivity Power margin
OC-48 (2.5 Gb/s) parameter /loss (dB)
link] Laser output 3 dBm
Transmitter: 3dBm
APD Sensitivity -32 dBm
@ 1550 nm;
@ 2.5 Gb/s
Receiver: InGaAs
APD with -32 dBm Allowed loss 3-(-32) dBm 35
sensitivity @ 2.5 Source connector 1 dB 34
Gb/s; loss
Fiber: 60 km long Jumper+Connect 3+1 dB 30
with o.3 dB/km or loss
attenuation; jumper Cable attenuation 18 dB 12
cable loss 3 dB each,
connector loss of 1 Jumper+Connect 3+1 dB 8
dB each. or loss
Receiver 1 dB 7(final margin)
Connector loss
Dispersion Analysis (Rise-Time Budget)
t sys  [ttx  t mod  tGVD  t rx ]
2 2 2 2 1/ 2
2 1/ 2
 2  440 Lq   350 
2

 ttx     D   L    
2 2 2
  B0   Brx  
t tx [ ns ] : transmitter rise time t rx [ ns ] : receiver rise time t mod [ n ] : modal dispersion
Brx [ MHz ]:3dB Electrical BW L[ km ]:Length of the fiber B0 [ MHz ]:BW of the 1 km of the fiber;
q  0.7 tGVD [ns]: rise- time due to group velocitydispersion
D[ ns /( km .nm )]:Dispersion   [nm]: Spectral width of the source

Two-level Binary Channel Codes
System rise-Time & Information Rate
• In digital transmission system, the system rise-time limits the

bit rate of the system according to the following criteria:
t sys  70% of NRZ bit period

t sys  35% of RZ bit period
Example
• Laser Tx has a rise-time of 25 ps at 1550 nm and spectral
width of 0.1 nm. Length of fiber is 60 km with dispersion 2
ps/(nm.km). The InGaAs APD has a 2.5 GHz BW. The rise-
time budget (required) of the system for NRZ signaling is 0.28
ns whereas the total rise-time due to components is 0.14 ns.
(The system is designed for 20 Mb/s).
Example: Transmission Distance for MM-Fiber
• NRZ signaling, source/detector: 800-900 nm LED/pin or AlGaAs
laser/APD combinations. BER  10;9 LED output=-13 dBm;fiber loss=3.5
dB/km;fiber bandwidth 800 MHz.km; q=0.7; 1-dB connector/coupling loss
at each end; 6 dB system margin, material dispersion ins 0.07 ns/(km.nm);
spectral width for LED=50 nm. Laser ar 850 nm spectral width=1 nm; laser
ouput=0 dBm, Laser system margin=8 dB;
Example:Transmission Distance for a SM Fiber
• Communication at 1550 nm, no modal dispersion, Source:Laser;
Receiver:InGaAs-APD (11.5 log B -71.0 dBm) and PIN (11.5log B-60.5
dBm); Fiber loss =0.3 dB/km; D=2.5 ps/(km.nm): laser spectral width 1
and 3.5 nm; laser output 0 dBm,laser system margin=8 dB;
Chapter 8
Photonic Networks
Contents
• Basic Networks
• SONET/SDH Standards
• Broadcast & Select WDM Networks
• Wavelength-Routed Networks
Basic Networks
• Stations or Data Equipment Terminal (DAT)
• Networks
- Local Area Networks (LAN)
- Metropolitan Area Network (MAN)
- Wide Area Network (WAN)
• Node
• Topology
- Linear Bus
- Ring
- Star
• Switching
• Routing
• Protocol
• Router
3 Network Topologies
Bus Topology
Ring Topology
Star Topology
Passive Linear Bus Topology
Simplex linear bus with uniformly spaced stations

Liner bus versus star coupler
SONET/SDH
STS-1 SONET frame

with 51.84 Mb/s
(90 bytes/row)(9 rows/frame)(8 bits/byte)

(125 microsecond/frame)=51.84 Mb/s
STS-N SONET frame

with 51.84N Mb/s
N=1,3,12,24,48,192
SONET/SDH
SONET level Electrical level Line rate (Mb/s) SDH equivalent
OC-1 STS-1 51.84 -
OC-3 STS-3 155.52 STM-1
OC-12 STS-12 622.08 STM-4
OC-24 STS-24 1244.16 STM-8
OC-48 STS-48 2488.32 STM-16
OC-96 STS-96 4976.64 STM-32
OC-12 STS-192 9953.28 STM-64

Technical Information
• For more information about SONET/SDH standards and their range,

receiver sensitivity, optical wavelength for various data rate, look at:
• http://www.itu.int (International Communication Union)
• ITU-T Recommendations G.652, G.653, G.655, G.957, G.691,
G.692
• Tables 12-3, 12-4, 12-5 in the text book give you an idea about ITU
recommendations for different photonic network configurations,
standards, and interfaces.
SONET/SDH Ring
2-Fiber UPSR (Unidirectional Path Switched Ring)
Generic two-fiber unidirectional network

with counter-rotating protection path.
Flow of primary and protection traffic

from node 1 to node 3.
SONET/SDH Networks
Add/Drop Multiplexer
WDM Network
Dense WDM deployment of n wavelengths in an OC-192 trunk ring

Broadcast & Select Networks
Bus
Star
Two physical architectures for a WDM LAN

Wavelength-Routed Networks
(Wavelength reuse, wavelength conversion & optical switching)
Optical Cross-Connects (OXC)

Optical Fiber Communications, 3e - Gerd Keiser

Enviado por

Dados do documento

Título original

Direitos autorais

Formatos disponíveis

Compartilhar este documento

Compartilhar ou incorporar documento

Opções de compartilhamento

Você considera este documento útil?

Este conteúdo é inapropriado?

Direitos autorais:

Formatos disponíveis

Optical Fiber Communications, 3e - Gerd Keiser

Enviado por

Direitos autorais:

Formatos disponíveis

E&CE 477

• Computer Simulations using Photonic Transmission Design

Optical Fiber communications, 3rd ed.,G.Keiser,McGrawHill, 2000

• The basic building block of SONET is called STS-1 (Synchronous

• The basic building block of SDH is called STM-1 (Synchronous Transport

SONET level Electrical level Line rate (Mb/s) SDH equivalent

OC-1 STS-1 51.84 -

OC-3 STS-3 155.52 STM-1

OC-12 STS-12 622.08 STM-4

OC-24 STS-24 1244.16 STM-8

OC-48 STS-48 2488.32 STM-16

OC-96 STS-96 4976.64 STM-32

OC-192 STS-192 9953.28 STM-64

Optical Fiber communications, 3rd ed.,G.Keiser,McGrawHill, 2000

Optical Fiber communications, 3rd ed.,G.Keiser,McGrawHill,

Optical Fiber communications, 3rd ed.,G.Keiser,McGrawHill, 2000

Optical Fiber communications, 3rd ed.,G.Keiser,McGrawHill, 2000

Optical Fiber communications, 3rd ed.,G.Keiser,McGrawHill, 2000

Optical Fiber communications, 3rd ed.,G.Keiser,McGrawHill, 2000

Optical Fibers: Structures,

• In engineering discipline, we should choose the appropriate & easiest

 [F/m] : Electric permittivi ty

• The solution of Maxwell’s equations in free space, through the wave

An electromagnetic wave is a travelling wave which has time

k : Wavenumber or wave propagation constant [1/m]

• Refractive index of a medium is defined as:

c velocity of light (EM wave) in vacuum 

• For non-magnetic media ( r  1) :

Examples of possible EM waves

Any two orthogonal plane waves

      2   cos   sin 2 

Optical Fiber communications, 3rd ed.,G.Keiser,McGrawHill, 2000

Optical Fiber communications, 3rd ed.,G.Keiser,McGrawHill, 2000

Reflection law: angle of incidence=angle of reflection

Snell’s law of refraction:

n1 sin 1  n2 sin 2 [2-18]

Optical Fiber communications, 3rd ed.,G.Keiser,McGrawHill, 2000

• Where (p,N) refer to the electric field components parallel or

Propagation mechanism in an ideal step-index optical waveguide.

Optical Fiber communications, 3rd ed.,G.Keiser,McGrawHill, 2000

Maximum entrance angle,  0 max is found from

n sin  0 max  n1 sin  c  n1  n2

m  0,1,2,3 (mode number)

• Most important characteristics of the EM transmission along the fiber are

n2 k   m (ω)  n1k [2-28]

• To minimize the signal distortion, the fiber is often operated in a single

Mode field diameter

Optical Fiber communications, 3rd ed.,G.Keiser,McGrawHill, 2000

Optical Fiber communications, 3rd ed.,G.Keiser,McGrawHill, 2000

Optical Fiber communications, 3rd ed.,G.Keiser,McGrawHill, 2000

Optical Fiber communications, 3rd ed.,G.Keiser,McGrawHill, 2000

• Mode propagation constant,  lm (ω), is the most important transmission

• Single mode operation is possible (Single mode fiber) when:

Only HE11 can propagate faithfully along optical fiber

• Example: A fiber with a radius of 4 micrometer and n1  1.500 & n2  1.498

Optical Fiber communications, 3rd ed.,G.Keiser,McGrawHill, 2000

• In general, a linearly polarized mode is a combination of both of the

• Signal attenuation (fiber loss) largely determines the maximum

• The parameter  p is called fiber attenuation coefficient in a units of for

P(l )[dBm ]  P(0)[dBm ]   [dB/km ]  l[km] [3-3]

• Where [dBm] or dB milliwat is 10log(P [mW]).

Optical Fiber communications, 3rd ed.,G.Keiser,McGrawHill, 2000

• Absorption is caused by three different mechanisms: