Design and Development of Erbium Doped Fiber

Amplifiers

A Project Report

submitted By

Laiju P.Joy
( Reg. No: 95713004 )
in partial fulfillment of the requirements
for the award of the degree of

MASTER OF TECHNOLOGY

INTERNATIONAL SCHOOL OF PHOTONICS

COCHIN UNIVERSITY OF SCIENCE AND TECHNOLOGY
COCHIN-682022
June 2015
 INTERNATIONAL SCHOOL OF PHOTONICS
COCHIN UNIVERSITY OF SCIENCE AND TECHNOLOGY
COCHIN-22

CERTIFICATE

This is to certify that the project work entitled ‘DESIGN & DEVELOPMENT OF
ERBIUM DOPED FIBER AMPLIFIERS‘ is a bonafide work done by Mr. LAIJU P.JOY
( Reg. No: 95713004 ) in partial fulfillment of the requirements for the award of the Degree
of Master of Technology in Opto Electronics & Laser Technology is carried out at
Laboratory for Electro-Optics Systems, ISRO, Bangalore during the academic year 2014-
2015.

Prof. P. Radhakrishnan Dr. M. Kailasnath

Professor Director
International School of Photonics International School of Photonics
CUSAT CUSAT
 ABSTRACT
The fiber amplifier is a key enabling technology for high speed optical communication.
The development of EDFA provide tremendous growth in communication system capacity.

The EDFA characteristics are analyzed for two different EDF length 7m and 13m, and two
different wavelengths 1550nm and 1570nm. The gain and noise characteristics of amplifier is
observed and compared with simulations.

The basic EDFA mathematical model is developed with Giles parameter by Matlab. The
Gain Master software used to simulate the EDFA with different length, input power and pump
powers.

The Optimum length of EDFA is 7m and use forward pumping configuration for reduce
effect of noise. The EDFA in 1550nm has gain 34.47dB in 20µW input with 300mW pump power
and noise figure is 4.5 dB.

The 5W amplifier use MOPA configuration. It has a EYCDFA followed by EDFA pre –
amplifier. The optimum length selected for EYCDF is 6m with pump power of 18W.
 ACKNOWLEDGEMENT

I would like to thank Mr. V.V. Laskshmi Pathi and Mrs. Lekshmi S. Rajan, Scientists,
Laboratory of Electro-Optics System (LEOS), they helped in reviewing the work progress,
results and offered valuable feedbacks in each and every stage of this project work.

I am thanking Dr. M. Kailasnath, Director, International School of Photonics, CUSAT for

giving all facilities that helped me to complete this mission.

I extend my sincere thanks to Dr. P. Radhakrishnan, Profeesor, International School of

Photonics, CUSAT for his motivation and support during my work

With deep sense of gratitude, I express my heartfelt thanks to Dr.V P N Nampoori, Emeritus
Professor, International school of photonics, CUSAT for the motivation, support in project work.

I thankful to my friends in International school of photonics, CUSAT for their support and help.

I sincerely thanks all the teachers who filled the knowledge and wisdom in me during each step
of my life. I also extend my sincere thanks to parents and God.

LAIJU P.JOY
 CONTENTS

1. INTRODUCTION 1

1.1 Basic Erbium Doped Fiber Amplifier 2

1.2 EDFA Models 4

1.2.1 GILES MODEL 4

1.2.2 Saleh –Jopson Model 5

1.2.3 Average Inversion Model 5

1.2.4 Higher Erbium Concentration Model 6

1.3 EDFA characteristics 6

1.3.1 Overlap Factor of Amplifier 6

1.3.2 Optimum Length of EDF 7

1.3.3 Small Signal Gain 7

1.3.4 Gain Saturation Region 9

1.3.5 Amplified Spontaneous Emission 10

1.4 High Power Amplifier 10

1.5 Motivations & Contributions 11

2. EDFA Components & Characterization 12

2.1 Components used in EDFA configuration 12

2.1.1 Pump Laser Diode 12

2.2.2 Super Luminescent Diode 13

2.2.3 Wavelength Division Multiplexer/De-multiplexer 14

2.2.4 Isolator 16

2.2.5 Circulator 19
 2.2.6 Fiber Bragg Grating 21

2.3 Characterization of EDFA Components 23

2.3.1 Pump Laser Diode 23

2.3.2 Super Luminescent Light Emitting Diode: SLED (Signal Source) 23

2.3.3 Isolator 24

2.3.4 Circulator 25

2.3.5 WDM Coupler 26

3. Design, Modeling & characterization of EDFA 27

3.1 MODELING OF EDFA 27

3.2 EDFA rate equations 27

3.2.1 Three level rate equations 27

3.2.2 Two level rate equation 30

3.3 Equations used for modeling 31

3.4 Amplifier Matlab Modeling with M-12 Generic Fiber using

Giles Parameters 32

3.6 Simulations using Gain Master: Giles Model 38

3.7 Optimization of EDFA Parameters 39

3.7.1 EDF Length 39

3.7.2 Signal Power 40

3.7.3 Pump Configuration 41

3.8 EDFA Experiment 43

3.8.1 EDFA Experimental setup 44

3.8.2 Amplification 46

3.8.3 EDFA Gain Characteristics 48

3.8.5 Noise Figure 54

 3.8.6 Noise Figure Characteristics 55

3.9 Experimental Data Analysis 56

4. Design of EYCDFA 57

4.1 Theory of EYCDFA 57

4.2 Experimental Setup for Simulation 58

4.3 Optimization of EYCDFA Parameters 59

4.3.1 Pump Power 59

4.3.2 Optimum Length 61

4.3.3 Signal Power 61

4.3.4 Wavelength 62

5. Conclusion 63

Future Plans 64

Bibliography 65

APPENDIX 66
 ABSTRACT
The fiber amplifier is a key enabling technology for high speed optical communication.
The development of EDFA provide tremendous growth in communication system capacity.

The EDFA characteristics are analyzed for two different EDF length 7m and 13m, and two
different wavelengths 1550nm and 1570nm. The gain and noise characteristics of amplifier is
observed and compared with simulations.

The basic EDFA mathematical model is developed with Giles parameter by Matlab. The
Gain Master software used to simulate the EDFA with different length, input power and pump
powers.

The Optimum length of EDFA is 7m and use forward pumping configuration for reduce
effect of noise. The EDFA in 1550nm has gain 34.47dB in 20µW input with 300mW pump power
and noise figure is 4.5 dB.

The 5W amplifier use MOPA configuration. It has a EYCDFA followed by EDFA pre –
amplifier. The optimum length selected for EYCDF is 6m with pump power of 18W.
 List of Figures

1.1 A basic EDFA configuration 2

1.2 Erbium ion transition 3

1.3 Erbium ion transition in different energy levels 3

1.4 Overlap between erbium ion distribution and transverse intensity profile 6

1.5 Optimum length at maximum gain 7

2.1 Diagram of simple laser diode 12

2.2 Common structure of super luminescent diode 13

2.3 The symmetric and antisymmetric mode of the combined wave guides 15

2.4 Coupling of power in wave guides 15

2.5 Polarization dependent isolator with Faraday rotator, polarizer and analyzer 18

2.6 Polarization independent isolator 19

2.7 Behavior of an optical circulator 20

2.8 Configuration of a 3 port optical circulator from port 1 to port 2 transmission 20

2.9 Configuration of a 3 port optical circulator from port 2 to port 3 transmission 20

2.10 Circulator used to drop an optical channel from a WDM system using FBG 21

2.11 FBG structure, refractive index profile and spectral response 22

2.12 Pump LD characteristics 23

2.13 SLD characteristics 24

2.14 SLED output spectrum obtained from OSA 24

2.15 Experimental setup for characterize isolator 25

2.16 Experimental setup for characterize circulator 25

2.17 Experimental setup for characterize WDM 26

3.1 EDFA as a 3 level system 28

3.2 EDFA as a 2 level system 30

3.3 Variation of signal power along the length of the fiber 35

3.4 Variation of pump power along the length of the fiber 36

3.5 Signal gain in dB along the length of the fiber 36

3.6 Forward ASE spectrum along the length of the fiber 37

3.7 Backward ASE spectrum along the length of the fiber 37

3.8 Variation of Noise Figure with pump power 38

3.9 Giles parameters 39

3.10 Variation of output signal power at different length EDF for input signal of 10μW and
pump power of 300Mw 40

3.11 Variation of output power with respect to input power 40

3.12 Different pumping configuration used in EDFA 41

3.13 Output power for different pumping configuration 42

3.14 Output ASE power for different pumping configuration 43

3.15 Block diagram of EDFA experimental setup 43

3.16 Input signal derived with FBG 44

3.17 EDFA experimental setup 45

3.18 Amplified output for 7m EDF for 1550nm 46

3.19 Amplified output for 13m EDF for 1550nm 47

3.20 Amplified output for 7m EDF for 1570nm 47

3.21 Amplified output for 13m EDF for 1570nm 48

3.22 Variation of gain for different input power in 7m EDF in 1550nm input 48

3.23 Variation of gain for different input power in 13m EDF in 1550nm input 49

3.24 Variation of gain for different input power in 7m EDF in 1570nm input 49
3.25 Variation of gain for different input power in 13m EDF in 1570nm input 50

3.26 Comparison of gain in 7m and 13m EDF for 1550nm input power of 10μW and
pump power 300mW 50

3.27 Comparison of gain at 1550nm and 1570nm input signal for 7m EDF with 10μW
input 51

3.28 Comparison of simulations and experiment for input 10μW with wavelength
1550nm 52

3.29 Forward ASE of EDF 7m without input 52

3.30 Backward ASE of EDF 7m without input 53

3.31 Forward ASE of EDF 7m with input 54

3.32 Noise figure for 1550nm signal 55

3.33 Noise figure for 1570nm signal 55

4.1 Erbium Ytterbium transitions 57

4.2 EYCDFA model configuration 58

4.3 Pump absorption of EYCDF 59

4.4 Signal absorption cross section of EYCDF 59

4.5 Variation of signal power with fiber length and pump power 60

4.6 Output Optical Power vs Length w.r.t different pump power 60

4.7 Signal output power vs length 61

4.8 Signal output power vs signal input 61

4.9 Variation of output power with wavelength 62

 List of Tables

2.1 Experimental result of circulator characterization 25

2.2 Experimental result of WDM coupler characterization 26

3.1 M-12 Generic fiber parameters used for Matlab simulation 34

3.2 EDFA output power with respect to EDF length 39

 Abbreviations

ASE Amplified Spontaneous Emission

EDF Erbium Doped Fiber

EDFA Erbium Doped Fiber Amplifier

ESA Excited State Absorption

EYCDF Erbium Ytterbium Co-Doped Fiber

EYCDFA Erbium Ytterbium Co-Doped Fiber Amplifier

FBG Fiber Bragg Grating

MOPA Master Oscillator Power Amplifier

NF Noise Figure

WDM Wavelength Division Multiplexer

 Symbols

A eff Effective erbium doped area in a fiber m2

τ Metastable life time of erbium ions s

h Plank’s constant Js

ν Frequency Hz

Δν Bandwidth Hz

g* Giles gain coefficient dB/m

α Giles loss coefficient dB/m

l Impurity and propagation losses in fiber dB/m

ζ Saturation Parameter m-3

σ (a ) Absorption cross section m-2

σ (e ) Emission cross section m-2

N Total population density of erbium ions/m3

N1 Ground state population density of erbium ions/m3

N2 Excited state population density of erbium ions/m3

λ Wavelength m

P Filed power W

I Intensity W/m2
 1

Chapter 1

INTRODUCTION
Earlier long distance optical communication systems use electronic regenerators for
amplifying the optical signals. The attenuated optical signal is amplified electronically, first
convert optical signal into electrical domain and then conversion back to optical domain. Such
regenerators are designed to operate at one optical wavelength and specific bit rate. In a WDM
communication systems carrying multiple wavelength signals through one fiber, the electronic
regeneration would be very complex and expensive.

Optical amplifier can amplify the incoming optical signals in the optical domain itself
without any conversions to the electrical domain. It do not need high speed electronics circuitry
and they are transparent to bit rate, can amplify multiple optical signals at different wavelength
simultaneously. The development of EDFA provide tremendous growth in communication
system capacity using WDM, in which multiple wavelengths carrying independent signals are
propagated through same single mode fiber.

Optical amplifiers can be used at many points in communication link. A booster amplifier
is used to boost the power of the transmit signal before it launching into the fiber link. The pre
amplifier placed just before the receiver is used to increase the receiver sensitivity. In line
amplifiers are used at intermediate points in the fiber link to overcome losses.

Today, most of the optical fiber communication system use EDFAs, due to their
advantages in terms of bandwidth, high power output, and noise characteristics.

Erbium doped fiber amplifier is a optical amplifier, it amplifies weak input optical signals
directly without any conversions pumped with a laser diode. The main application of EDFA is to
amplify signals in optical domain.

The EDFA became a key enabling technology for optical communication networks, and
have since comprised the vast majority of all optical amplifiers deployed in the field. Erbium

Design and Development of Erbium Doped Fiber Amplifiers

 2

doped fiber amplifier is most common optical amplifier, commercially available since the early
1990’s. It is a most stable optical amplifier with operating bands 1525 – 1565 nm wavelength
region. It works best in this range with gain upto 30 dB.

The main element in EDFA is Erbium doped fiber, which is developed by conventional
Silica fiber with rare earth element Erbium.

1.1 Basic Erbium Doped Fiber Amplifier

The basic EDFA configuration is shown in figure 1.1

It contains
Erbium Doped Fiber
Pump Source
Wavelength Division Multiplexer
Isolators

Figure 1.1 : a basic EDFA configuration

Erbium is generally preferred in fiber for amplification because of the inherent properties
associated with it. Erbium ions have quantum levels that can be stimulated to emit 1550nm band
with least power loss. Moreover the property of erbium is that its quantum levels allow it to get
excited by 980nm or 1480 nm pump signals. The amplification is achieved by stimulated
emission of photons from erbium ions in the doped fiber. The pump laser excites erbium ions
into a higher energy from ground level. The ions in the higher energy level will soon decay
spontaneously very fast to the metastable level, the life time of erbium ions in this level is 10ms.

Design and Development of Erbium Doped Fiber Amplifiers

 3

The stimulated decay from metastable level to ground generate light amplification by stimulated
emission.

When the Erbium is illuminated with light energy at a suitable wavelength (either 980nm
or 1480nm) it is excited to a long lifetime intermediate state (see Figure 1.2), following which it
decays back to the ground state by emitting light within the 1525-1565 nm band.

Figure 1.2 : Erbium ion transitions

Figure 1.3 : Erbium ion transitions in different energy levels

Design and Development of Erbium Doped Fiber Amplifiers

 4

When signal photon of wavelength equivalent to the band gap energy between the ground
state and metastable state is passing through the erbium doped fiber, two types of transitions
occur. First a small portion of the ions in the ground state absorb this signal photon and raise to
the metastable state known as stimulated absorption . The ions in the metastable band on
absorbing the energy from the signal photon can undergo stimulated emission and drop to the
ground level, thereby emitting a new photon of the same wavelength and same polarization that
of the input signal photon. Erbium ions can also be excited by a pump wavelength of 1480nm
but is not desirable because the pump and signal wavelengths are almost nearer and hence the
transitions between these wavelengths will lower the efficiency of the device and increase
amplifier noise. The 980nm pump source has a higher absorption cross section and hence will be
used where EDFA design demands low noise. Hence we have used a 980nm pump laser.

1.2 EDFA Models

The method of developing EDFA uses different modeling techniques. The Giles model
solves the steady state rate equation utilizing gain and absorption parameters which are
proportional to the cross sections. This model includes propagation equation which allow
modeling along the length of the fiber. Saleh-Jopson model provide analytical solutions to the
propagation equation. The higher erbium concentration model include the inhomogeneous
effects such as ion-ion interaction and ESA.

1.2.1 GILES MODEL

The simpler method of erbium doped fiber can be characterized by using amplifier
equation in terms of the erbium absorption coefficient α(λ), gain coefficient g*(γ), a fiber
saturation parameter ζ and excess loss in the fiber from scattering and impurity absorption l(λ).

These easily measured parameters allow the fiber performance evaluation in 980 nm or
1480 nm pumped optical amplifiers. Conventional fiber measurement techniques are used to
obtain these parameters, from which the amplifier performance can be calculated.

The rate equations for forward (+) and backward (-) propagating beams are

Design and Development of Erbium Doped Fiber Amplifiers

 5

±
dPK
dz
( n1
)
= α K + g K* 2 PK± + 2 g K* 2 hν K ∆ K − (α K + lK )
n n
n1
(1.1)

n2 ∑K PK hν K
=
(
n1 1 + ∑ PK± α K + g K* hν K ζ ) (1.2)
K

where

α (λ ) = σ a (λ )Γ(λ )N

g * (λ ) = σ e (λ )Γ(λ )N

ζ = Aeff N τ

1.2.2 Saleh –Jopson Model

This model developed for estimation of the pump and signal power along the length of
the EDF fiber. The Saleh-Jopson model valid for amplifiers with gain less than 20 dB, and gain
saturation by ASE can be neglected.

The pump signal absorption and signal gain in EDF can be obtained by solving
transcendental equation. When the pump and signal power propagate through fiber the change in
power in the Kth beam can be obtained by

dPK ( z , t )
dz
(( ) )
= Nµ K ΓK σ Ke + σ Ka N 2 ( z , t ) − σ Ka PK ( z , t ) (1.3)

 A(Pin − Pout ) 
PKout = PKin exp(− NΓK σ Ka L) exp 
( )
(1.4)
 ΓK σ K + σ K τ
e a


where PKin and PKout are total input and output power.

1.2.3 Average Inversion Model

This model compute the gain and noise figure at other wavelengths from a computation of the
average inversion from a measured reference spectrum and cross section ratio.

Design and Development of Erbium Doped Fiber Amplifiers

 6

1.2.4 Higher Erbium Concentration Model

The higher E r3+ ion concentration model require additional modeling terms which
account for concentration quenching or ion-ion interaction and ESA.

The concentration of E r3+ ion increase in fiber leads to undesirable effect like cooperative
up-conversion and pair induced quenching. The ion- ion interaction effects limits the
concentration of erbium ions in silica matrix.

Signal or pump excited state absorption is possible in EDF due to presence of other
energy levels in erbium energy levels. These other energy levels in erbium can absorb signal or
pump photon to higher energy level. This effect can be depletes the population inversion and also
gain.

1.3 EDFA Characteristics

1.3.1 Overlap Factor of Amplifier

Consider one dimensional model of the fiber amplifier, the overlap factor is known as
overlap between transverse intensity profile of optical mode and transverse erbium ion
distribution profile. This overlap will stimulate absorption or emission from the transitions.

If we consider a single mode cylindrical geometry optical fiber with constant area and
erbium ion density, the overlapping is shown in figure 1.4

Figure 1.4 : Overlap between erbium ion distribution and transverse intensity profile

The Overlap factor is

Design and Development of Erbium Doped Fiber Amplifiers

 7

 −R 
2


Γ = 1− e ω
2

 
  (1.5)

where R is the erbium doped radius in fiber and is the spot size of the beam. The spot size
ω will vary with frequency, the overlap factor will be depends on frequency of the mode.

1.3.2 Optimum Length of EDF

The input signal is amplified along the length of the EDF at fixed pump power upto a
specific point, after that point the gain is negative, and so the fiber should be terminated at the
point. At this point the pump power is decreased to the threshold level. The length of the fiber at
that specific point determine optimum length of EDF for EDFA. The optimum length at
maximum gain is shown in figure 1.5

Figure 1.5 : Optimum length at maximum gain

1.3.3 Small Signal Gain

The signal field and pump field with corresponding intensities Is and Ip, are propagated
simultaneously in EDF and interact with ions in the fiber. Due to this interactions variations are
occurred in signal intensity and pump intensity, they are given by

Design and Development of Erbium Doped Fiber Amplifiers

 8

σ PIP
− Γ21
dI s hν P
= σ S IS N
dz Γ + 2 σ S I S + σ P I P
hν S hν P
21
(1.6)

σ S IS
Γ21 +
dI P hν S
=− σ I N
dz σ S IS σ PIP P P
Γ21 + 2 +
hν S hν P (1.7)

If I P > I th , the threshold condition for gain, for propagation of signal field.

hν P
I P ≥ I th =
σ Pτ 2 (1.8)

where I th is the pump threshold intensity

The normalized intensities are given by

IP
I P' =
I th (1.9)

IS
I S' =
I th (1.10)

The quantity η and saturation intensity I sat (Z) as

hν P σ S
η=
hν S σ P (1.11)

1 + I P' (Z )
I sat (Z ) =
2η (1.12)

The normalized variation of signal intensity and pump intensity as

dI S' ( z )  I P' ( z ) − 1 
σ S I S' ( z )N
1
=  '
dz 1 + I S ( z ) I sat ( z )  I P ( z ) + 1 
'

(1.13)

Design and Development of Erbium Doped Fiber Amplifiers

 9

dI P (z ) 1 + ηI S' (z )
=− σ P I P' (z )N
dz 1 + 2ηI S ( z ) + I P ( z )
' '
(1.14)

The condition for small signal gain is satisfied, when I S << I sat , at this sate the signal
propagation along the length of the fiber is

I S' ( z ) = I S' (0) exp(α P z ) (1.15)

where the gain coefficient defined as

Ι 'P − 1
αP = σSN
Ι 'P + 1 (1.16)

We can see that the signal gain grows exponentially, with a coefficient proportional to the signal
emission cross section and degree of population inversion.

1.3.4 Gain Saturation Region

In gain saturation region, the signal growth is linear. This occurs when the signal I S'

grows to a large value comparable to I sat . The signal growth is then damped by the saturation

1
factor,
1 + I S' I sat

I S'
The ratio of I S' and I sat becomes large compared to unity, >> 1
I sat

The growth of signal in saturation region is determined by

dI S'  I −1
= I sat  P σ S N
dz  IP + 1 (1.17)

I sat is linearly dependent with pump power, then signal saturation is varied with pump power.

The saturation output power is inversely proportional to emission cross section of fiber,
this causes the saturation power higher at 1550 nm than at 1530 nm.

The experimentally determined saturation output power is defined as the signal output
power at which the gain has been reduced by 3 dB.

Design and Development of Erbium Doped Fiber Amplifiers

 10

1.3.5 Amplified Spontaneous Emission

It is a parasitic process, which can occur at any frequency within the fluorescence
spectrum of the amplifier transitions. The effect of ASE is to reduce the total amount of gain
available from the amplifier.

The excited ions can spontaneously relax from the upper state to ground state by emitting
a photon that is uncorrelated with the signal photons. The spontaneously emitted signal is
amplified, it travels along the fiber and it stimulates the emission of more photons from exicted
state. This process is known as ASE. The ASE power sometimes referred to as an equivalent
noise power.

For a single transverse mode fiber with two independent polarizations for a given mode at
frequency ν, the noise power in a bandwidth Δν, corresponding to spontaneous emission, is equal
to

0
PASE = 2h∆ν (1.18)

The total ASE power at a point z along the fiber is the sum of the ASE power from the
0
prevision sections of the fiber and added local noise power PASE .This local noise power will
stimulate the emission of photons from excited erbium ions.

The propagation equation for the ASE power propagating in a given direction is thus

dPASE (ν )
dz
( )
= N 2σ (e ) (ν ) − N1σ (a ) (ν ) PASE (ν ) + PASE
0
(ν )N 2σ (e ) (ν )
(1.19)

1.4 High Power Amplifier

The gain of EDF is determined by erbium ions density in the fiber, then gain is increased
by adding extra erbium ions. But, after a particular level of doping the concentration quenching
effect is occure in EDF. This is due to reduce the gain of amplifier.

We want higher amplified power in range of Watts, then use Ytterbium Co-doped with
Erbium fiber. The Yb ions around Er ions prevent the concentration quenching. Then We will

Design and Development of Erbium Doped Fiber Amplifiers

 11

get higher amplified output power. In EYCDFA the Yb ions are pumped at 800-1100nm
wavelengths. The Yb ions transfer energy to the Er ions and Er ions are excited to higher energy
state. The stimulated emission of Er ions fom metastable state give higher output power.

1.5 Motivations & Contributions

According to the previous literature survey, the Erbium doped fiber (EDF) is a reliable
gain media for lasers and amplifiers. The selection of appropriate pump wavelength sources are
very crucial.

Erbium ions exhibit a very narrow absorption bands, 10 nm at 980 nm for 3 level system
and 1480 nm for 2 level system. This is due to the selection of pump source for the amplifier is
limited to InGaAs laser diode and Titanium Sapphire (Ti:S) laser source which operate within
these wavelength region. The small absorption cross section of erbium ions is not suitable for
high power amplifications due to the limitation of ions concentrations inside fibers.

Throughout this work, the suitable parameters to operate the EDFA have been identified.
This leads to optimized output obtained by the amplifier.

Design and Development of Erbium Doped Fiber Amplifiers

 12

Chapter 2

EDFA Components & Characterization

2.1 Components used in EDFA configuration

2.1.1 Pump Laser Diode

A laser diode, or LD, is an electrically pumped semiconductor laser in which the active
laser medium is formed by a p-n junction of a semiconductor diode similar to that found in a
light-emitting diode.

A laser diode is electrically a P-i-n diode. The active region of the laser diode is in the
intrinsic (I) region, and the carriers, electrons and holes, are pumped into it from the N and P
regions respectively. All modern lasers use the double-heterostructure implementation, where the
carriers and the photons are confined in order to maximize their chances for recombination and
light generation. The goal for a laser diode is that all carriers recombine in the I region, and
produce light. Thus, laser diodes are fabricated using direct bandgap semiconductors.

Diagram of a simple laser diode shown in figure 2.1

Figure 2.1 :Diagram of simple Laser Diode

Design and Development of Erbium Doped Fiber Amplifiers

 13

The commonly used pump laser source for optical amplifiers are

980 nm – InGaAs

1480 nm – InGaAsP

2.1.2 Super Luminescent Light Emitting Diode : SLED (Signal Source)

A superluminescent diode (SLED or SLD) is an edge-emitting semiconductor light

source based on superluminescence. Its output is high power and brightness with the low
coherence, and emission bandwidth is 5–100 nm wide.

A superluminescent light emitting diode is, similar to a laser diode, based on an

electrically driven PN-junction that, when biased in forward direction becomes optically active
and generates amplified spontaneous emission over a wide range of wavelengths. The peak
wavelength and the intensity of the SLED depend on the active material composition and on the
injection current level. The basic structure of SLED is shown in figure 2.2 .

Figure 2.2 : Common structure of superluminescent diode

A SLED consists of a positive (p-doped) section and a negative (n-doped) section,

electrical current will flow from the p-section to the n-section and across the active region that is
sandwiched in between the p- and n-section. During this process, light is generated through

Design and Development of Erbium Doped Fiber Amplifiers

 14

spontaneous and random recombination of positive (holes) and negative (electrons) electrical
carriers and then amplified when travelling along the waveguide of a SLED.

The PN-junction of the semiconductor material of a SLED is designed in such a way that
electrons and holes feature a multitude of possible states (energy bands) with different energies.
Therefore, the recombination of electron and holes generates light with a broad range of optical
frequencies, i.e. broadband light.

2.1.3 Wavelength Division Multiplexer/De-multiplexer

Wavelength division multiplexing is a technology for combine number of different

wavelength signals into a single optical fiber and vice versa. This technique enables bidirectional
communication, and it effectively use the transmission capacity of optical fiber.

The WDM working principle is same as that of directional coupler. In optical fiber
directional coupler , the modal fields of a fiber extends beyond the core cladding interface. If two
fibers are close enough their modal fields overlap and there can be periodic coupling of power
between the two fibers. The mode have equal propagation constant, then power transfer can be
complete.

A directional coupler formed by two identical symmetric single mode planar waveguides
then the coupled system can be viewed as a single waveguide with two cores. The propagation
constants of the two modes of such a system would be different. Power incident on one
waveguide excites a linear combination of symmetric and antisymmetric modes. Due to
difference in their propagation constants the modes develop a phase difference. When the
accumulated phase difference is π superposition of the modes result in cancellation of the mode
in first waveguide and adding of power in the second. When the phase difference is 2π the total
power appears in the first waveguide. Thus periodic exchange of power takes place between the
waveguides.

Design and Development of Erbium Doped Fiber Amplifiers

 15

Figure 2.3 :symmetricand antisymmetric mode of the combined wave guides

Figure 2.4 : Coupling of power in waveguides

Consider the power exchange between two single mode fibers that are non-identical
supporting LP01 modes having 1 and 2 as the propagation constants. If P1(0) is the power
launched into the fiber 1 at z = 0, then at any z the power propagating in the two fibers are given
by

P1 ( z ) K2
= 1 − 2 sin 2 γz (2.1)
P1 (0 ) γ

P2 ( z ) K 2
= 2 sin 2 γz (2.2)
P1 (0 ) γ

Where γ 2 = K 2 +
1
4
(∆β 2 )

∆β = β1 − β 2

Design and Development of Erbium Doped Fiber Amplifiers

 16

The K is called coupling coefficient and is a measure of strength of the coupling between
two fibers. The value of the coupling coefficient depends on the separation between the cores,
wavelength of operation and fiber parameters. The parameter ∆β is referred to as the phase
mismatch. For ∆β = 0 , the phase matched case the minimum distance (coupling length) for
which the power is completely transferred from one fiber to the other fiber is given by

π
z = LC = (2.3)
2K

If ∆β ≠ 0 the power transfer is incomplete.

Let us consider a coupler of length L made of identical fibers and let K1 and K 2 be the
coupling coeffcients at λ1 and λ2 so that K 1 L= m π and K 2 L = m − 1 2π . So if λ1 and λ2 are
launched into the input fiber simultaneously then

P2 (λ1 , L ) = P1 sin 2 (K1 L ) = 0

Thus the light of wavelength λ1 will exit from the input fiber and that of wavelength λ2
will exit the other fiber. Such a device forms a de-multiplexer and in the opposite direction it can
form a multiplexer.

2.1.4 Isolator

In fiber optic network, most of the reflections are harmful to the stability of the system. If
back reflected and scattered light enter into the laser, the lasing process will fluctuate and the
output power of the laser will varied. This problem will be avoided with the proper isolation
between the components by use isolators.

Optical isolators are device, it transport light only in one direction and prevent the
reflections and scattered light from sensitive components, particularly lasers.

Design and Development of Erbium Doped Fiber Amplifiers

 17

The main component of the optical isolator is the Faraday rotator. The magnetic field,
applied to the Faraday rotator causes a rotation in the polarization of the light due to the Faraday
effect. The angle of rotation β , is given by

β = νBd (2.4)

where, ν is the Verdet constant of the material, and d is the length of the rotator.

Polarization dependent isolator

The polarization dependent isolator, or Faraday isolator, is made of three parts, an input
polarizer (polarized vertically), a Faraday rotator, and an output polarizer, called an analyser
(polarized at 45°).

Light traveling in the forward direction becomes polarized vertically by the input
polarizer. The Faraday rotator will rotate the polarization by 45°. The analyser then enables the
light to be transmitted through the isolator.

Light traveling in the backward direction becomes polarized at 45° by the analyser. The
Faraday rotator will again rotate the polarization by 45°. This means the light is polarized
horizontally .Since the polarizer is vertically aligned, the light will be extinguished.

Figure 2.5 shows a Faraday rotator with an input polarizer, and an output analyser.

Design and Development of Erbium Doped Fiber Amplifiers

 18

Figure 2.5 : Polarization dependent isolator with Farady rotator, polarizer and analyzer

Polarization dependent isolators are typically used in free space optical systems. This is
because the polarization of the source is typically maintained by the system. In optical fibre
systems, the polarization direction is typically dispersed in non polarization maintaining systems.
Hence the angle of polarization will lead to a loss.

Polarization independent isolator

The polarization independent isolator is made of three parts, an input birefringent wedge
(with its ordinary polarization direction vertical and its extraordinary polarization direction
horizontal), a Faraday rotator, and an output birefringent wedge (with its ordinary polarization
direction at 45°, and its extraordinary polarization direction at −45°).

Light traveling in the forward direction is split by the input birefringent wedge into its
vertical (0°) and horizontal (90°) components, called the ordinary ray (o-ray) and the
extraordinary ray (e-ray) respectively. The Faraday rotator rotates both the o-ray and e-ray by
45°. This means the o-ray is now at 45°, and the e-ray is at −45°. The output birefringent wedge
then recombines the two components.

Light traveling in the backward direction is separated into the o-ray at 45, and the e-ray at
−45° by the birefringent wedge. The Faraday Rotator again rotates both the rays by 45°. Now the

Design and Development of Erbium Doped Fiber Amplifiers

 19

o-ray is at 90°, and the e-ray is at 0°. Instead of being focused by the second birefringent wedge,
the rays diverge.

Figure 2.6 :Polarization independent isolator

Figure 2.6 shows the propagation of light through a polarization independent isolator.
The forward travelling light is shown in blue, and the backward propagating light is shown in
red.

2.1.5 Circulator

An optical circulator is a special fiber-optic component that can be used to direct the
optical signals from one port to another port and in one direction only. This action prevent the
signal from unwanted directions.

In a 3-port circulator a signal is transmitted from port 1 to port 2, another signal is

transmitted from port 2 to port 3 and, finally, a third signal can be transmitted from port 3 to port
1.

This action is shown in figure 2.8 and 2.9

Figure 2.7 Conventional figure to represent the behavior of an optical circulator.

Design and Development of Erbium Doped Fiber Amplifiers

 20

Figure 2.7 : Behavior of an optical circulator

Figure 2.8 : Configuration of a 3 port optical circulator from port 1 to 2 transmission

Figure 2.9 : Configuration of a 3 port optical circulator from port 2 to 3 transmission

Design and Development of Erbium Doped Fiber Amplifiers

 21

The input signal from port 1 is split into o-ray (00) and e-ray (900) by a birefringent plate
and the faraday rotator rotates both wave by 450 , and the birefringent plate at port 2 is
recombine the 2 components. This operation is shown in figure 2.8

The light input from port 2, that light goes through the process described before and after
the second birefringent plate, the rays are combined by using a reflector and beam splitter.

We consider the application of a Fiber-Bragg grating compensator to select a particular

wavelength signal. This can also be done using a three-port optical circulator as shown in the
following figure 2.10

Figure 2.10 : Circulator used to drop an optical channel from a WDM system using a Fiber
Bragg Grating

2.1.6 Fiber Bragg Grating (FBG)

A fiber Bragg grating (FBG) is a type of distributed Bragg reflector constructed in a short
segment of optical fiber that reflects particular wavelengths of light and transmits all others.

A fiber Bragg grating can therefore be used as an inline optical filter to block certain
wavelengths, or as a wavelength-specific reflector.

Fiber Bragg Gratings are made by laterally exposing the core of a single-mode fiber to a
periodic pattern of intense ultraviolet light. The exposure produces a permanent increase in the
refractive index of the fiber's core, creating a fixed index modulation according to the exposure
pattern. This fixed index modulation is called a grating.

Design and Development of Erbium Doped Fiber Amplifiers

 22

At each periodic refraction change a small amount of light is reflected. All the reflected
light signals combine coherently to one large reflection at a particular wavelength when the
grating period is approximately half the input light's wavelength. This is referred to as the Bragg
condition, and the wavelength at which this reflection occurs is called the Bragg wavelength.
Light signals at wavelengths other than the Bragg wavelength, which are not phase matched, are
essentially transparent. This principle is shown in Figure 2.11

Figure 2.11 : Fiber Bragg Grating structure, Refractive index profile and spectral response

Therefore, light propagates through the grating with negligible attenuation or signal
variation. Only those wavelengths that satisfy the Bragg condition are affected and strongly
back-reflected. The ability to accurately preset and maintain the grating wavelength is a
fundamental feature and advantage of fiber Bragg gratings.

The central wavelength of the reflected component satisfies the Bragg relation is

λB = 2neff Λ (2.5)

where

n eff = effective refractive index of the fiber core

Design and Development of Erbium Doped Fiber Amplifiers

 23

Λ = grating period of the index of refraction of the FBGthe wavelength of the reflected
component will also change as function of temperature and/or strain.

The parameters n eff and Λ depends to the temperature and strain,

2.2 Characterization of EDFA Components

2.2.1 Pump Laser Diode

The pump laser diode operate at 980nm wavelength. It act as a pump source in EDFA
experiment. The peak wavelength of the pump LD is 975.97nm with spectral width is 0.2nm.
The lasing action start at the drive current of 45mA. The relation between drive current versus
output power is shown in figure 2.12

Figure 2.12 : Pump LD characteristics

2.2.2 Super Luminescent Light Emitting Diode (SLED)

It is a broadband light source, operate with wavelength from 1520nm to 1580nm. The
EDFA input signals 1550nm and 1570nm are derived from this SLED source with FBG.

The lasing action of SLED start at drive current of 40mA. The characteristics curve of
input drive current versus output power is shown in figure 2.13

Design and Development of Erbium Doped Fiber Amplifiers

 24

Figure 2.13 : SLD characteristics

The output spectrum of SLED with drive current of 50mA is shown in figure 2.14

Figure 2.14 : SLED output spectrum obtained from OSA

2.2.3 Isolator

Isolator is a two port optical component, it allow propagation of light only in one
direction. It mainly used for avoid reflections from the opposite direction.

Design and Development of Erbium Doped Fiber Amplifiers

 25

Figure 2.15 : Experimental setup for characterize isolator

This figure shows the experimental setup for characterize isolator. In this configuration
isolator is connected as forward configuration, for measure insertion loss.

The isolator is connected in reverse direction, for measure isolation between port 2 and 1.

The isolation between the input and output port is 32.05dB and insertion loss of around
0.44dB.

2.2.4 Circulator

The principle of operation of circulator is same as that of isolator. It allows transmission

of signals from port 1 to 2 and port 2 to 3 only, not in other directions.

Figure 2.16 : Experimental setup for characterize circulator

The table 2.1 Shows insetion and isolation between ports in circulator

Insertion Loss (dB) Isolation (dB)

Port 1 to 2 = 0.83 Port 2 to 1 = 59.30

Port 2 to 3 = 0.80 Port 3 to 2 = 60.04

Table 2.1 : Experimental result of circulator characterization

Design and Development of Erbium Doped Fiber Amplifiers

 26

2.2.5 WDM Coupler

The WDM has two input port 1550nm signal port and 980nm pump port, and one
common output port. The below table 2.2 shows insertion loss and isolation between ports of
WDM coupler

Figure 2.17 : Experimental setup for characterize WDM

Insertion Loss (dB) Isolaton (dB)

1550nm port to common port = 0.89 1550nm port to 980nm port = 57.75

Common port to 1550nm port = 0.97 Common port to 980nm port = 24.23

Table 2.2 : Experimental result of WDM coupler characterization

Design and Development of Erbium Doped Fiber Amplifiers

 27

Chapter 3

Design, Modeling& characterization of EDFA

3.1 MODELING OF EDFA

The EDFA mathematical model was designed using Matlab with the help of Giles
parameter. The EDFA software simulation modeling and simulation done by Gain master
software. The two different EDFA configurations are analyzed using 7m and 13m EDF, and
analysis was done for two different wavelengths 1550 nm and 1570nm. The models are
described below.

3.2 EDFA Rate Equations

The Erbium doped fiber amplifier design start by considering a pure three level atomic
system. Most of the important characteristics can be obtained from this model. The rate equation
can also be made more complex by considering the effects such as excited state absorption and
the three dimensional character of the problem. For reduce the complication of EDFA model
design the basic three level system is convert to two level system, assuming that the upper pump
level and upper metastable level belongs to the same multiplet.

3.2.1 Three level rate equations

The gain model of EDFA is designed with set of coupled differential equations, because
the gain process consist of coupled atomic population and light flux propagation equations. We
will consider a pure three level atomic system for EDFA at 1550nm. The most important
characteristics of EDFA can be obtained from this model. The importance to absorption and
emission cross sections, and the difference between the two at a given transition wavelength, and
also consider the concept of overlap parameter.

Design and Development of Erbium Doped Fiber Amplifiers

 28

Figure 3.1 : EDFA as a 3 level system,

The 3 level system is shown in figure 3.1

The ground state denoted as 1, an intermediate state 3 at which energy is pumped, and
metastable state 2. Since the metastable state often a long life in the case of good amplifier.

The state 2 is upper level of amplifying transition and state 1 is the lower level. The
population of the levels are labeled as N1, N2 and N3.The population inversion between level 1
to 2 is achieved by pumping for amplification.

The one dimensional solution of coupled differential equations are obtained by assume
the pump field , signal field intensities and the Erbium ion distribution are constant in transverse
dimensions, over an effective cross sectional area of the fiber

The incident pump field due to transition from level 1 to 3 and the incident signal field
due to transition transition from level 1 to 2. The population of each level is growing from
absorption of photons from incident fields.

The Γ32 is the sum of nonradiative and radiative transition probabilities from level 3 to 2,
is mostly nonradiative. The is mostly radiative transitions, in this case erbium transitions will
occur (level 2) to (level 1). The definition of , where is the life time of erbium ions in
metastable state.

Design and Development of Erbium Doped Fiber Amplifiers

 29

The absorption cross section for level 1 to 3 denoted as and emission cross section for
level 2 to 1 is denoted as .

The rate equation for population changes as

= −Γ32 N 3 + ( N1 − N 3 )φPσ P
dN 3
(3.1)
dt

= −Γ21 N 2 + Γ32 N 3 − ( N 2 − N1 )φPσ S

dN 2
dt (3.2)

= Γ21 N 2 − ( N1 − N 3 )φPσ P + ( N 2 − N1 )φPσ S

dN1
dt (3.3)

In steady state condition, the rate equations becomes

The total population N is derived by

N=N1+N2+N3 (3.4)

The population of level 3 is obtained from equation 3.1 as

1
N3 =
1 + Γ32 φPσ P (3.5)

The value of N3 is nearly zero, because Γ32 is very large, due to fast decay from level 3 to
level 2.

The equation 3.4 used to derive the population inversion as

φPσ P − Γ21
Ν 2 − Ν1 = Ν
Γ21 + 2φSσ S + φ Pσ P (3.6)

When N2 > N1, the population inversion is achieved, and amplification of signal is
occurred in level 2 to 1 transition.

When N1 =N2, it is the threshold condition. The threshold pump flux is

Design and Development of Erbium Doped Fiber Amplifiers

 30

Γ21 1
φth = =
σP τ 2σ P (3.7)

The threshold pump intensity is

hν P Γ21 hν P
I th = =
σP σ Pτ 2 (3.8)

The conditions for lower number of pump photons to reach threshold are

1. Higner pump absorption cross section

2. The longer life time in metastable state

In erbium, the life time has large value of nearly 10ms in silica glass.

The ultra low pump threshold is one of the main advantage of erbium doped single mode
fiber amplifiers.

3.2.2 Two level rate equations

Figure 3.2 : EDFA as a 2 level system

Design and Development of Erbium Doped Fiber Amplifiers

 31

In order to model the basic characteristics of erbium doped fiber amplifier the three level
system was reduced into a 2 level system assuming that the upper pump level and the upper
amplifier level belong to same multiplet. The rate equation for multiplets 1 and 2 are

dN 2
dt
( ) (
= −Γ21 N 2 + N1σ S(a ) − N 2σ S(e ) φS − N1σ P(a ) − N 2σ P(e ) φP ) (3.8)

dN1
dt
( ) (
= Γ21 N 2 + N 2σ S( e ) − N1σ S(a ) φS − N1σ P(a ) − N 2σ P(e ) φP ) (3.9)

The Total population density N is given by

N=N 1 +N 2
(3.10)

We have,

dN1 dN
=− 2 (3.11)
dt dt

From the above rate equation we can calculate N 2 as

τσ S(a ) τσ (a ) τσ P(a )
ΓS PS + ∑ νj Γνj PΑ (νj ) + ΓΡ PΡ
Ahν S Ahν j Ahν P
N 2 (z ) =
τ (σ S(a ) + σ S(e ) ) τ (σν(ja ) + σν(je ) ) τ (σ P(a ) + σ P(e ) )
(3.12)
ΓS PS + ∑ Γνj PΑ (νj ) + ΓΡ PΡ + 1
Ahν S Ahν j Ahν P

where PA (νj ) = PA+ (νj ) + PA− (νj )

3.3 Equations used for modeling

The field propagation equations can also be written in terms of the field powers. These
equations contains overlap parameters, because overlap part between erbium ion distribution and
optical mode determine gain or attenuation. We also consider the possible intrinsic background
loss parameters in the fiber α P(a 0 ) , α S(a 0 ) and α vj(a 0 )

Design and Development of Erbium Doped Fiber Amplifiers

 32

dPΡ
dz
( )
= N 2σ Ρ(e ) − N1σ Ρ(a ) ΓΡ PΡ − α Ρ(a 0 )PΡ
(3.13)

dPS
dz
( )
= N 2σ S(e ) − N1σ S(a ) ΓS PS − α S(a 0 )PS
(3.14)

dPΑ+ (νj )
dz
( )
= N 2σν(je ) − N1σν(ja ) ΓS PΑ+ (νj ) + N 2σν(je )ΓS hν j ∆νj − αν(ja 0 )PΑ+ (νj )
(3.15)

dPΑ− (νj )
dz
( )
= − N 2σν(je ) − N1σν(ja ) ΓS PΑ− (νj ) − N 2σν(je )ΓS hν j ∆ν j + αν(ja 0 )PΑ− (νj )
(3.16)

The model becomes more accurate, when the frequency spectrum channels are chosen to
be very small about 1nm. The ASE power at spectrum ν j is sum of forward ASE and backward

ASE. The forward propagating ASE calculated with an input power of 0 at z=0 and backward
propagating ASE calculated with an input power of 0 at z=L, where L is the length of the fiber.

3.4 Amplifier Matlab Modeling with M-12 Generic Fiber using Giles Parameters

The amplifier modeling with Giles parameter can be used to calculate gain and ASE of a
simple two level system model without emission and absorption cross sections. The field
propagation equations can be write in terms of Giles parameters, such as Erbium absorption
coefficient α(λ), gain coefficient g*(λ), fiber saturation parameter ζ and excess loss in the fiber
due to scattering and impurity absorption l(λ).

The equations from 3.13 to 3.16 are rewrite according to Giles parameters

The population inversion as

αS αj α
PS ( z ) + ∑ ν PA (νj , z ) + P PP ( z )
ζ hν S ζ hν j ζhν P
N 2 ( z) =
(α S + g S )
*
(ανj + gνj )
*
(α + g P* )
PS ( z ) + ∑ PA (νj ) + P PP ( z ) + 1
ζhν S ζhνj ζhν P (3.17)

The field propagation equations are

Design and Development of Erbium Doped Fiber Amplifiers

 33

dPP
dz
=
1
N
( )
N 2 g P* − N1α P PP − l P PP
(3.18)

dPS
dz
=
1
N
( )
N 2 g S* − N1α S PS − l S PS
(3.19)

dPA+ (νj ) 1
dz
=
N
( )
N 2 gν*j − N1ανj PA+ (νj ) + 2 N 2 gν*j hν j ∆νj − lνj PA+ (νj ) (3.20)

dPA− (vj ) 1
dz
=
N
( )
N1α vj − N 2 g vj* PA+ (vj ) + 2 N 2 g vj* hν j ∆ vj − l vj PA+ (νj )
(3.21)

These modified propagation equations are used to model the mathematical model of
EDFA, we will get gain and ASE by simulate these equations with Matlab.

The field propagation equations are coupled differential equations, it is a two point
boundary value problem. Each equations have a initial conditions based on propagation
direction. Consider a EDF has length ‘L’ for amplifier modeling. So that, all fields are propagate
total distance L from z=0 to z=L. Then accumulate all field values from z=0 to z=L gives final
value of the field waves.

At the input of the EDF, we know the applied pump signal power (P p ) and input signal
power (P s ) for amplification, that is the initial values of P p and P s . The launched pump power
into the fiber is PP ( z = 0) = Pp0 and input signal power is Ps ( z = 0) = PS0 for amplification. The

forward ASE initial power at z=0 was considered as zero. Now, we know initial values of signal,
pump, and forward ASE. Then these equations are integrated from z=0 to z=L, We got final
values of all parameters at z=L. This analysis is first forward simulation. At this step don’t
consider Backward ASE, because we don’t know the value of backward ASE at z=L.

The next step is first backward simulation. At this step consider all final values, we
assume final value of backward ASE is equal to zero at z=0. The all set of equations are
integrated backward from z=L to z=0 to obtain the value of backward ASE at z=0. Then again
repeat the forward and backward integration with backward ASE values.

Design and Development of Erbium Doped Fiber Amplifiers

 34

The forward and backward simulation is repeated many times until a constant solution
was obtained.

Table 3.1 : M-12 Generic fiber parameters used in Matlab simulation

Parameters Values
A 3 × 10 −6 m2
N 1.6 × 10 25 ions/m2
σ S(e ) 40118853 × 10 −25 m2

σ P(e ) 0.0810563 × 10 −25 m2

σ S(a ) 2.910556 × 10 −34 m2

σ P(a ) 2.7876712 × 10 −25 m2

τ 10.2 × 10 −3 S
αP 2.4 dB/Km

αS 2dB/Km

δ 3100 GHz (25nm)

ζ 8.5 × 1015 m-1s-1

The table 3.1 shows parameters provided by the EDF manufacturer. These parameters
can be used for simulations. The modified field equations are used to simulate the characteristics
of EDFA. The result obtained after simulation is shown below and the Matlab code is provided
in appendix.

The result of simulation is compared with the experimental values. This simulation
neglect the Excited State Absorption (EAS), Complex effects and inhomogeneities in the erbium
ion distribution etc.

The figures 3.3 and 3.4 shows that the pump power and output amplified signal. The
pump power decreases exponentially and output signal power increases, corresponding to the
fiber length variations. The maximum amplified output signal is obtained at the EDF length 7

Design and Development of Erbium Doped Fiber Amplifiers

 35

meter. After that this EDF length, the attenuation is started, because upper level population is
reduced to low level.

The Gain in dB of the output amplified signal is shown in figure 3.5

The forward ASE spectrum is available at the output of the amplifier is shown in figure
3.6. It is considered as a unwanted broadband noise signal. It reduce the gain and efficiency of
the amplifier. The shape of the ASE spectrum depends on absorption emission cross sections and
length of the fiber.

The backward ASE his always higher than forward ASE, because it generated near the
pump source. The backward ASE is shown in figure 3.7

The figure 3.8 shows the variation of noise figure with pump power. The minimum
possible noise figure at 400mW is 4.5 dB in this simulation.

Figure 3.3 : Variation of signal power along the length of the fiber

Design and Development of Erbium Doped Fiber Amplifiers

 36

Figure 3.4 : Variation of pump power along the length of the fiber

Figure 3.5 : Signal gain in dB along the length of the fiber

Design and Development of Erbium Doped Fiber Amplifiers

 37

Figure 3.6 : Forward ASE spectrum along the length of the fiber

Figure 3.7 : Backward ASE spectrum along the length of the fiber

Design and Development of Erbium Doped Fiber Amplifiers

 38

Figure 3.8 : Variation of Noise Figure with Pump Power

3.6 Simulations using Gain Master: GILES MODEL

Fiber amplifier simulations are possible with different types of softwares. Some examples
are RP Fiber, OptSim, Gain Master (EDF forFibercore Pvt.Ltd ), OptiSystem etc..

An M-12 Generic fiber which is a product of Fibercore, the EDFA was designed with
Gain Master software, it can be used to optimize EDF length for maximum gain and pump power
for corresponding maximum gain. The Gain master also can be used to analysis different method
of pumping direction in EDF for maximum gain at optimum length.

The Gain Master used Giles Model equations and parameters for simulation. This
simulation used erbium pump absorption around 980nm and erbium signal emission around
1550nm are shown in figure 3.9

Design and Development of Erbium Doped Fiber Amplifiers

 39

Figure 3.9 : Giles Parameters

3.7 Optimization of EDFA Parameters

3.7.1 EDF Length

Length Output Power

6 76.087
7 76.402
8 74.448
9 71.434
10 67.881
Table 3.2 :EDFA output power with respect to EDF length

Design and Development of Erbium Doped Fiber Amplifiers

 40

Figure 3.10 : Variation of output signal power at different length EDF for input signal of 10µW
and pump power of 300mW.

The EDF is simulate at different length, that is 6-10 meters. The output signal power
variation along different length is obtained. We can see that the output is high for & meter long
fiber. So the optimum fiber length was fixed to be & meter. This 7 meter EDF used design
should give an output nearly 76 mW.

3.7.2 Signal Power

The figure 3.11 shows the variation of output power with respect to the change in input
power. We can see that the output power is saturated above 100µW input power.

Figure 3.11: Variation of output power with respect to input power

Design and Development of Erbium Doped Fiber Amplifiers

 41

3.7.3 Pump Configuration

Figure 3.12 : Different pumping configurations used in EDFA

The EDFA is classified according to the pumping directions, there are 3 type catagery.
They are Forward pumping (Co Pumped, Backward Pumping (Counter Pumped), and
Bidirectional pumped(Dual Pumped)

In forward pumping scheme, the WDM coupler infront of the EDF is combine the input
signal and pump signal . So the input signal and pump signal propagate through the fiber in same
direction, ie mean by co pumped.

Design and Development of Erbium Doped Fiber Amplifiers

 42

In EDF, the ions are excited due to pump signal and it transfer absorbed energy to the
input signal , the the signal is amplified. The isolators are used for prevent back reflections inside
the fiber and make sure that the signal will travel only in one direction through the fiber.

In backward pumping scheme, the input signal and pump signal are propagate in the
opposite directions inside the fiber, that is mean by counter pumped.

In bidirectional pumping schem, two pump signals pumped in both directions of input
EDF. One pump signal travel with the signal in same direction and other is opposite direction of
input signal.

The total pump power is set as 300mW and input signal is 10µW. The effect of different
pumping configuration is shown in figure 3.12

Figure 3.13 : Output power for different pumping configurations

The figure 3.13 shows that the gain is maximum in Bidirectional pump configuration
more than 90mW.

Design and Development of Erbium Doped Fiber Amplifiers

 43

Figure 3.14 : Output ASE power for different pumping configurations

The figure 3.14 shows that ASE is higher value in bidirectional pumping and minimum in
forward pumping configuration. A better amplifier require very less noisy operation, then we
will consider forward pumping configuration.

3.8 EDFA Experiment

The general blok diagram of EDFA setup is shown in figure 3.15. In this experiment use
EDF with two different length 7m and 13m, and analysis the amplification of two different input
signal wavlengths 1550nm and 1570nm.

Figure 3.15 : Block diagram of EDFA experimental setup

Design and Development of Erbium Doped Fiber Amplifiers

 44

Signal Input For EDFA

The input signals given to EDFA are at 1550nm and 1570nm. The SLED source have a
broad band spectrum ranging from 1520 nm to 1580nm . The input signal at the required
wavelengths can be derived using circulator and FBG. The configuration is shown in figure
3.16. The FBGs reflection wavelengths are 1550nm/1570nm with spectral width of 0.5nm. The
required input signal for amplification 1550nm/1570nm is available at port 3 of circulator.

Figure 3.16 : Input signal derived with FBG

3.8.1 EDFA Experimental Setup

The heart of the Erbium doped fiber amplifier is erbium doped Fiber. In this EDFA
experiment used M-12 Generic980nm Erbium doped fiber, the experiment of EDFA was
planned after the simulation and analysis of Gain Master Software which is supplied by the EDF
manufacturer Fibercore Ltd.

In this experiment the complete simulation, analysis and characterization was planned for
two different EDF length 7m and 13m,and for two different wavelength 1550nm and 1570nm.

The EDFA setup was made with other passive optical components spliced with EDF.
Then after the EDFA was characterized with different signal power, pump power, fiber length
and signal wave length.

The EDFA experimental setup is shown in figure 3.17

Design and Development of Erbium Doped Fiber Amplifiers

 45

Figure 3.17 : EDFA experimental setup

The input section of EDFA contains input isolator, WDM coupler and pump LD. The
input isolator’s output fiber pigtail and pump laser source’s output fiber pigtail is spliced to
corresponding input fiber pigtail port of WDM coupler permanently, and also ensure that the
quality of splice is good. After splicing the splice loss was measured by splicing machine, it was
about 0.05 db. The other end of the input isolator contains a optical connector for connecting
input optical signal from SLD. The main active optical component EDF was taken at the
predefined length from fiber spool and wound on a bobbin. The one end of the EDF was spliced
with the common fiber pigtail port of WDM coupler and other end spliced with output isolator’s
input fiber pigtail. The output port of output isolator is connected to optical connector for taken
amplified output signal.

The pump LD was spliced to the 980nm fiber pigtail port of WDM coupler. The pump
LD is mounted on a LD driver board. The LD driver board and other components are mounted
on a ESD pad and kept in undisturbed condition. This setup was used for other EDF length by
replacing the EDF only. This EDFA experiment was characterized by different parameters such
as signal power, signal wavelength, pump power and EDF length.

Design and Development of Erbium Doped Fiber Amplifiers

 46

3.8.2 Amplification

The input signal wavelength 1550nm with different powers are applied to EDFA,with
different EDF length 7m and 13m. This same configuration is applied to 1570nm input signal.

The amplified output Spectrum of 10µW input signal, available from the OSA are given
from figure 3.18 to figure 3.21. The observation from this OSA graph, we can see that the higher
amplification possible with EDF length 7m for 1550nm signal.

The 1550nm signal has more gain compared 1570nm signal. The peak power from the
output spectrum at wavelength 1549.82 is 3.528mW at marker1.

The noise floor (P noise ) is also obtained from this graph at markers 2 and 3.

The Gain calculation in dB is

P − Pnoise 
Gain(dB ) = 10 log signal 
 Pin  (3.22)

where P in is the input signal power

Figure 3.18 : Amplified output for 7m EDF for 1550nm

Design and Development of Erbium Doped Fiber Amplifiers

 47

Figure 3.19 : Amplified output for 13m EDF for 1550nm

Figure 3.20 : Amplified output for 7m EDF for 1570nm

Design and Development of Erbium Doped Fiber Amplifiers

 48

Figure 3.21 : Amplified output for 13m EDF for 1570nm

3.8.3 EDFA Gain Characteristics

The EDFA gain variations with respect to EDF length, input signal power, pump power
and input signal wave lengths are described below.

Figure 3.22 shows that the gain variation for different input signals with wavelength
1550nm in 7m EDF. We can obtained the maximum gain is about 35dB at 20µW input signal
power.

Figure 3.22 : Variation of Gain for different input power in 7m EDF in 1550nm input

Design and Development of Erbium Doped Fiber Amplifiers

 49

The EDF with 13m, gain variation for different input signals with wavelength 1550nm is
shown in figure 3.23. We can observe that the gain of 13m EDF used EDFA is less than 7m.

Figure 3.23 : Variation of Gain for different input power in 13m EDF in 1550nm input

The different input signals with 1570nm gain variation in 7m EDF is shown in
figure3.24. It can be seen that, the gain of 1570nm wavelength signal is less than that of 1550nm
wavelength signal.

Figure 3.24 : Variation of Gain for different input power in 7m EDF in 1570nm input

Design and Development of Erbium Doped Fiber Amplifiers

 50

The gain variation of different input signals with wavelength of 1570nm in 13m EDF is
shown in figure 3.25

Figure 3.25 : Variation of Gain for different input power in 13m EDF in 1570nm input

The figure 3.26 shows the gain variation of 7m and 13m EDF for 1550nm input signal
with power of 10µW. We can observe that the gain is higher in 7m EDF compared to 13m EDF.

Figure 3.26: Comparison of Gain in 7m &13m EDF for1550nm input power of 10µW and pump
power 300mW

Design and Development of Erbium Doped Fiber Amplifiers

 51

The gain variation related to wavelength of input signals 1550nm and 1570nm are shown
in figure 3.27. We can see that the gain is maximum available at 1550nm signals.

Figure 3.27 : Comparison of Gain at 1550nm & 1570nm input signal for 7m EDF with 10µW
input

Comparison of Simulations and Experimental values

The figure 3.28 shows the comparison of Gain Master simulation, Matlab Simulation and
experimental values, using 7m EDF for 1550nm wavelength.

The Matlab program coded with available Giles parameters and not include. The ESA
effects, excess losses in the wavelengths. The software simulation obtained by Gain Master. The
experimental data has same response with constant gain differences compared to simulations.

Design and Development of Erbium Doped Fiber Amplifiers

 52

Figure 3.28 : Comparison of EDFA output of Experimental, Matlab simulation and Gain master
simulation for 10μW input with wavelength 1550nm.

3.8.4 EDFA ASE Characteristics

The spectral response of forward ASE and backward ASE without input signal in 7m
EDF is shown in figure 3.26 and 3.27.

Figure 3.29 : Forward ASE of EDF 7m without input

Design and Development of Erbium Doped Fiber Amplifiers

 53

The forward ASE for 7m fiber pumped at 200mW without input signal. The spectrum
shape of the forward ASE is varied with pump power, after certain pump power spectrum
maintain a constant shape. Now, it has 2 peak, the maximum peak is 28.456µW at 1531.3nm
wavelength and other is 25.67µW at 1557nm wavelength.

Figure 3.30 : Backward ASE of 7m EDF without input

The backward ASE for 7m fiber pumped at 200mW without input signal. the maximum
peak is 62.55µW at 1530.6nm wavelength.

Design and Development of Erbium Doped Fiber Amplifiers

 54

Figure 3.31 : Forward ASE of EDF 7m with input

The forward ASE for 7m fiber pumped at 300mW with signal input of 10µW is shown in figure
3.27. A peak at 1550nm is observed, that is the amplified input signal.

3.8.5 Noise Figure

The noise figure (NF) calculated from the available OSA spectrum. The 3dB bandwidth of the
amplified signal spectrum is represented as Δµ.

The gain in dB from equation 3.22 denoted as ‘g’.

The noise figure calculation is

p − gPnoise
s
1
NF (dB ) = 10 noise + 
 hν∆νg G
(3.23)

S
Where Pnoise is the input signal noise floor

Design and Development of Erbium Doped Fiber Amplifiers

 55

3.8.6 Noise Figure Characteristics

The noise figure of the EDFA experiment is shown in figure 3.28 for 1550nm and 3.29
for 1570nm inputs.

The OSA measurement not give absolute power of the measured signals. For actual
measurement the output of the amplifier has to be calibrated with power meter. The OSA and
power meter measurement difference is known as offset. The offset value is calculated before the
analysis of data.

Figure 3.32: Noise figure for 1550nm signal

Figure 3.33: Noise figure for 1570nm signal

Design and Development of Erbium Doped Fiber Amplifiers

 56

3.9 Experimental Data Analysis

1. The optimum length of EDF is 7m for maximum gain.

2. The 7m EDF with forward pump configuration generate less noisy output.
3. The 1550nm wavelength has higher gain due to high emission cross section.
4. The maximum possible gain is 34.47 dB for the signal power of 20µW and pump power
of 300mw at 1550nm wavelength
5. The minimum noise figure obtained in 1550nm is 4.266 dB.

Design and Development of Erbium Doped Fiber Amplifiers

 57

Chapter 4

Modeling of EYCDFA

4.1 Theory of EYCDFA

The output power of EDFA can be increased by increase the number of erbium ions
concentration in the erbium fiber. The higher concentration of ions introduce the concentration
quench effect, which can decrease the absorption of pump light. This problem can be reduce the
gain of amplifier, increasing erbium ions above a particular level is not a solution of increasing
gain.

The amplification at 1550 nm can be increased by Er3+ is co-doped with Yb3+ ions, it can
suppress the concentration quench effect. These two kinds of ions with similar radius and low
solubility can form the ion clusters. The Yb3+ ions around Er3+ can reduce the up-conversion

probability and makes higher ion conversion. The Yb3+ ion has large pump absorption cross

section and wider range from 800nm to 1100nm.

Figure 4.1 : Erbium Ytterbium Transitions

Design and Development of Erbium Doped Fiber Amplifiers

 58

The figure 4.1 shows simplified energy level transitions for EYCDF. The Yb3+ ions absorb

the pump power and transfer to Er3+ . The energy transfer efficiency is higher than 95%, because

of large spectral overlap between Yb3+ emission spectrum and Er3+ absorption spectrum.

4.2 Experimental Setup for Simulation

Figure 4.2: EYCDFA Model Configuration.

By the virtue of co-doping Ytterbium with Erbium, Erbium can be doped to higher
concentrations without quenching effect for maximizing the gain. For higher Power output
EYDFA needs a Pre-amplifier which is normally an EDFA. EDFA amplifies weak signals to
power output of about hundreds of milli Watts enough to drive EYDFA. Isolator in the circuit
prevent back reflections and a high power WDM coupler is necessary to couple the pump and
signal powers. The pre-amplifier is a fixed gain amplifier which gives a gain of 23 dB.

The MOPA based EYCDFA is shown in figure 4.4, this can simulated by OptSim
software.

The EYDFA model was cladding pumped and the absorption and emission cross section
data for the model in OptSim is given below.

Design and Development of Erbium Doped Fiber Amplifiers

 59

Figure 4.3: Pump absorption of EYCDF

Figure 4.4: Signal absorption cross section of EYCDF

4.3 Optimization of EYDFA Parameters

4.3.1Pump Power

This configuration has to be optimized in terms of fiber length, pump power, and signal
power for an output of 5W. Initially the length was fixed and pump power was varied from 5W

Design and Development of Erbium Doped Fiber Amplifiers

 60

to 20W. This was performed for 3 cases of fiber length 5 m,

10 m and 15 m. Gain was found to be more for 5 m fiber. The output requirement of 5 W was
satisfied almost when the pump power reached 18 W. This can be shown in figure 4,4

Figure 4.5 : Variation of signal power with fiber length and pump power

The output power variations corresponds to the pump power variationis shown in figure 4.5

Figure 4.6 : Output Optical Power vs Length w.r.t different pump power

Design and Development of Erbium Doped Fiber Amplifiers

 61

4.3.2 Optimum Length

Now the pump power was fixed at 18 W and simulations for gain along the fiber.
Maximum gain was observed at 8m length of the fiber and at almost 6 m of fiber length gain
crossed 5W. The power variation corresponding to different length is shown in figure 4.6

Figure 4.7 : Signal Output Power vs Length

4.3.3 Signal Power

Now the Pump and length were fixed at 17 W and 7 m respectively and the input signal
was swept to almost 500 μW. For signal powers above 100 mW the output power condition was
met. The results are included as figures 4.6 below

Figure 4.8 : Signal Output Power vs Signal Input Power

Design and Development of Erbium Doped Fiber Amplifiers

 62

4.3.4 Wavelength

The output power variations in different wavelength with change in pump power is
shown in figure 4.8. The maximum gain is possible above the 1550nm wavelength region.

Figure 4.9:Variation of output power with wavelength

Design and Development of Erbium Doped Fiber Amplifiers

 63

Chapter 5

CONCLUSION

The EDFA with 7m fiber has better gain response compared to 13m fiber
and the noise effect is less in forward pumping configuration compared to other.
The signal input of 20μw with wavelength 1550nm and pump power of 300mW,
the gain is 34.47dB.

The 1550nm signal has higher gain and better noise figure with respect to
1570nm. The noise figure value of 13m fiber with input power of 10μw and pump
power of 300mW is 4.266 dB. It is the minimum noise figure achieved in the
experiment.

The MOPA configuration of EYCDFA model created in OptSim software

was simulated to obtain the EYCDF length is 6m for 5W output.

The first stage of MOPA configuration require minimum input signal power
for 5W output is 100μW with gain 21 dB.

The EYCDF length required is 6m for 5W output with pump power of 18W
and operating wavelength is above1550nm.

Design and Development of Erbium Doped Fiber Amplifiers

 64

Future Plans

The Matlab simulation of EDFA will improve with account of ESA and excess losses in
the wavelengths.

The experimentation and characterization of high power amplifier stage and will develop
the complete product.

Design and Development of Erbium Doped Fiber Amplifiers

 65

Bibliography

A.A.M Saleh, R.M Jopson ’Modeling of Gain in Erbium Doped Amplifiers’ IEEE Photonics
Technology Letters, 1990

Cuneyt Berkdemir, Sedat Ozsoy ‘Temperature Dependent Performance Analysis of EDFAs

pumped at 1480 nm: A more accurate propagation problem’ Optic Express, vol 13, No 13, 2005

[BECKER, P. C. 2002] “Erbium-Doped Fiber Amplifiers”, NY: Academic Press.

[Giles, C. R. and Desurvire, E. 1991] “Propagation of Signal and Noise in Concatenated Erbium
Doped Fiber Optical Amplifiers”, IEEE Journal of Light wave Technology, Vol. 9, No. 2, pp.
147-154

[Desurvire, E. 1994] “Erbium-Doped Fiber Amplifiers”, NY: John Wiley & Sons, Inc.

[Desurvire, E. 1996] “An explicit Analytical Solution for the Transcendental Equation
Describing Saturated Erbium-Doped Fiber Amplifier” Optical Fiber Technology, Vol. 2, pp.
367-377.

Wangi, ‘High power er/Yb co-doped Fiber Amplifier of Fiber Length Optimization’ IEEE, 2010

Keiichi Aiso , Yoshio Tashiro , Tsuneo Suzuki and Takeshi Yagi ‘Development of Er/Yb Co-
doped Fiber for High-Power Optical Amplifiers’, Furukawa Review, No. 20 2001

Nadir Hossain1a, V. Mishra1, A.A.R. Hairul1, F.M. Abbou1, A.R. Faidz1, S. M. M. Quadir1,
M.H. Al-Mansoori1, M.A. Mahdi2 and A.W. Naji ‘A Numerical Analysis of R-EDFA for Long
Haul Optical Fiber Communication Syste4th International Conference: Sciences of Electronic,
Technologies of Information and Telecommunications March 25-29, 2007 – TUNISIA IEEE

Design and Development of Erbium Doped Fiber Amplifiers

 66

APPENDIX

Design and Development of Erbium Doped Fiber Amplifiers

 Main Program

clc;

clear all;close all;

options = odeset('RelTol',1e-9);

Pp= 400e-3;

Ps = 10e-6;

L= 7;

% First forward Run......

[xf1,yf1] = ode45(@gain_ode_for,[0 L],[Pp Ps zeros(1,301)]);

figure(1)

plot(xf1,yf1(:,1));

title('Pump Power in first run');

figure(2)

plot(xf1,yf1(:,2));

title('signal Power in first run');

db = 10*log10(yf1(:,2)/10e-6);

figure(4)

plot(xf1,db);

title('gain in first run');

printSol(xf1,yf1,9);

m=size(yf1,1);

d = 1520e-9:0.2e-9:1580e-9;

figure(5)

plot(d,yf1(m,3:303))
title('Forward ASE in first run');

% First backward Run......

[xb1,yb1] = ode45(@gain_ode,[L 0],[yf1(m,:) zeros(1,301)]);

figure(6)

plot(xb1,yb1(:,1));

title('Pump power in first backward run');

figure(7)

plot(xb1,yb1(:,2));

title('Signal power in first backward run');

figure(3)

plot(xb1,yb1(:,3));

db = 10*log10(yb1(:,2)/10e-6);

figure(8)

plot(xb1,db);

title('Gain in first backward run');

printSol(xb1,yb1,20);

[m1 n1]=size(yb1);

figure(9)

plot(d,yb1(m1,3:303))

title('Forward ase in first backward run');

figure(10)

plot(d,abs(yb1(m1,304:604)))

title('Bacward ASE in first backward run');

% Second Forward run

[xf2,yf2] = ode45(@gain_ode,[0 L],[Pp Ps zeros(1,301) (yb1(m1,304:604))]);

figure(11)

plot(xf2,yf2(:,1));

title('Pump power in second forward run');

figure(12)

plot(xf2,yf2(:,2));

title('Signal power in second forward run');

%figure(3)

% plot(xf1,yf1(:,3));

db = 10*log10(yf2(:,2)/10e-6);

figure(14)

plot(xf2,db);

title('Gain in second forward run');

printSol(xf2,yf2,9);

m3=size(yf2,1);

figure(19)

plot(d,yf2(m3,3:303))

title('Forward ASE in second forward run');

figure(20)

plot(d,yf2(m3,304:604))

title('BAckward ase in second forward run');