NUMERICAL ANALYSIS OF NONLINEAR PULSE

PROPAGATION IN OPTICAL FIBER

A Thesis Presented
By
SADEED M. ZAHIN #052-652 -045
M IQBAL MAZID # 052-401-045
M. A. A. WAHID RAHAT #052-339-045

Submitted to
Department of Electrical Engineering & Computer Science in partial fulfillment of the requirements for the degree
of

Bachelor of Science in Electronics and Telecommunication Engineering (ETE)

Department of Electrical Engineering & Computer Science

North South University

Dhaka, Bangladesh.

Date: 30-12-09.
NUMERICAL ANALYSIS OF NONLINEAR PULSE
PROPAGATION IN OPTICAL FIBER

Advisor:

………………………………

Mr. Zasim U. Mazumder

Lecturer

Department of Electrical Engineering & Computer Science

North South University

Dhaka

Chairman:

………………………………

Dr. Miftahur Rahman

Chairman & Professor

Department of Electrical Engineering & Computer Science

North South University

Dhaka
Abstract

In this work we study the nonlinear pulse propagation in optical fiber. Dispersive effects
during propagation causes temporal spreading of pulses and it can be compensated by nonlinear
effects. When the effects are combined they can generate stable, undistorted pulses over long
distances. Numerical analysis is carried out using the split step Fourier method on different types
of pulses including hyperbolic secant, Gaussian and super-gaussian pulses of various orders.
Higher order solitons are also included in this study. Hyperbolic secant pulses are found to
propagate with constant pulse broadening ratio and linear phase change as expected. Pulse
broadening ratio of second order solitons are found to fluctuate periodically but the other higher
order solitons disintegrate as they propagate. Gaussian pulse broadening ratio appears to
decrease and the pulse stabilizes gradually to a hyperbolic secant shape with linear phase change
in the case of zero chirp parameter. Pulse broadening ratio for super-Gaussian pulses decreases
gradually.

i
Acknowledgments
First of all we would like to thank our supervisor, Mr. Zasim U. Mozumder for
introducing us to the amazingly interesting world of Nonlinear Fiber Optics and teaching us how
to perform research work. Without his continuous supervision, guidance and valuable advice, it
would have been impossible to complete the thesis. We are especially grateful to him for
allowing us greater freedom in choosing the topic to work on, for his encouragement at times of
disappointment and for his patience with our wildly sporadic work habits. We are grateful to all
other friends for their continuous encouragement and for helping us in thesis writing. We would
like to express our gratitude to all our teachers specially Mr. Arshad Momen. Their motivation
and encouragement in addition to the education they provided meant a lot to us. Last but not
least, we are grateful to our parents and to our families for their patience, interest, and support
during our studies.

ii
To

The Chairman,

Department of Electrical Engineering & Computer Science,

North South University,

Dhaka-1217,

Bangladesh.

Summer 2009

Date: 30th December, 2009.

Subject: Approval for valid submission of project report for the course ETE499

Dear Sir,

We, M IQBAL MAZID, SADEED M. ZAHIN and M. A. A. WAHID RAHAT are doing our final year
project/thesis, ETE 499, this semester. Our topic is “NUMERICAL ANALYSIS OF NONLINEAR PULSE
PROPAGATION IN OPTICAL FIBER”.

. We declare that the work presented in this thesis entitled “NUMERICAL ANALYSIS OF NONLINEAR
PULSE PROPAGATION IN OPTICAL FIBER”is the outcome of the investigation carried out by us under the
supervision of Mr. Zasim U. Mozumder, Assistant Professor, Department of Electrical Engineering & Computer
Science, North South University Dhaka. It is also declared that neither this thesis nor any part thereof has been
submitted or is being currently submitted anywhere else for the award of any degree or diploma. We have completed
our project and will be doing our presentation on 30th December, 2009.

Under light of this letter, we hope that our report is granted for approval and is adjudged a valid submission in
fulfillment of the course ETE-499.

Sincerely,

M IQBAL MAZID

Student ID.: 052-401-045

………………………………

SADEED M. ZAHIN

Student ID.: 052-652 -045

………………………………….

M. A. A. WAHID RAHAT

Student ID.: 052-339-045

………………………………….

iii
 TABLE OF CONTENTS

1. INTRODUCTION 1
1.1 INTRODUCTION 2
1.2 THESIS OBJECTIVE 3
1.3 BACKGROUND OF THE PROJECT 4

2. LITERATURE REVIEW 7
2.1 DISPERSION 8
2.1.1 GROUP VELOCITY AND PHASE VELOCITY 9
2.1.2 GROUP VELOCITY DIPERSION 9
2.1.3 THIRD ORDER DISPERSION 10

2.2 NONLINEAR EFFECTS 11

2.2.1 NONLINEAR REFRACTIVE INDEX 11
2.2.2 TYPES OF NONLINEAR EFFECTS 12
2.2.2.1 SELF-PHASE MODULATION 12
2.2.2.2 STIMULATED RAMAN SCATTERING 13
2.2.2.3 STIMULATED BRILLOUIN SCATTERING 14

2.3 PULSE PROPAGATIONS 15

2.3.1 NONLINEAR SCHRODINGER EQUATION (NLSE) 16
2.3.1.1 GENERALIZED NLSE 16
2.3.1.2 SIMPLIFIED NLSE 17
2.3.1.3 NORMALIZED NLSE 18

iv
 2.3.2 TYPES OF PULSES 19
2.3.3 HYPERBOLIC SECANT PULSES 19
2.3.4 GAUSSIAN PULSES 19
2.3.4.1 CHIRPED GAUSSIAN PULSE 21
2.3.5 SUPER-GAUSSIAN PULSES 22

2.4 CHIRP 23
2.4.1 CHIRP DEFINATION 23
2.4.2 TYPES OF CHIRP 24
2.4.2.1 GVD INDUCED CHIRP 25
2.4.2.2 SPM INDUCED CHIRP 25
2.4.2.3 PRE INDUCED CHIRP 25
2.4.3 CRITICAL CHIRP 25
2.4.4 EFFECTS OF CHIRP 26

2.5 SOLITON 27
2.5.1 CONDTIONS FOR SOLITON 27
2.5.2 HIGHER ORDER SOLITON 27

3. DESCRIPTION OF METHOD 28
3.1 SPLIT STEP FOURIER METHOD 29
3.2 SYMMETRIZED SPLIT STEP FOURIER METHOD 32
3.3 DECISION 34

4. IMPLEMENTATION 35
4.1 NONLINEAR SCHRODINGER EQUATION (NLSE) SOLUTIONS 36
4.1.1 GENERALIZED NLSE 36
4.1.2 SIMPLIFIED NLSE 36
4.1.3 NORMALIZED NLSE 37

v
 4.2 GAUSSIAN IMPLEMENTATION 38
4.3 SOLITON IMPLEMENTATION 40
4.4 FLOW CHART 42
4.5 PULSE BROADENING RATIO 45
4.5.1 FULL WIDTH HALF MAXIMUM 45

5. ANALYSIS OF RESULTS 46
5.1 ANALYSIS PLAN 47
5.2 GAUSSIAN PULSE ANALYSIS 48
5.2.1 ANALYSIS FOR GVD VARIATION 49
5.2.2 ANALYSIS FOR NONLINEAR PARAMETER VARIATION 51
5.2.3 ANALYSIS FOR INPUT POWER VARIATION 53

5.3 SOLITON PULSE ANALYSIS 55

5.4 HIGHER ORDER SOLITON ANALYSIS 58

5.5 CHIRPING EFFECT ANALYSIS 63

5.5.1 Analysis for Un-chirped Gaussian pulse 64
5.5.2 Analysis for Gaussian pulse of Chirp = -1 66
5.5.3 Analysis for Gaussian pulse of Chirp = -2 68
5.5.4 Analysis for Gaussian pulse of Chirp = 2 70

5.6 SUPER GAUSSIAN PULSE ANALYSIS 72

5.6.1 Analysis for Supergaussian pulse for chirp =0 75
5.6.2 Analysis for Supergaussian pulse for chirp =100 76

6. CONCLUSION 77
BIBLOGRAPHY 79
MATLAB CODES 81
LIST OF SYMBOLS 92

vi
 List of Figures and Tables

Fig 2.1: Group Velocity (red) and Phase Velocity (blue).

Fig 2.2: Elastic Rayleigh scattering and Inelastic Raman Scattering

Fig 2.3: Raman Scattering

Fig2.4: Chirp

Fig 3.1: Split step Fourier method

Fig 3.2: Symmetrized split step Fourier method

Fig 4.1: Gaussian Flowchart

Fig 4.2: Soliton Flowchart

Fig 4.3: Supergaussian Flowchart

Fig 4.4: Full Width at Half Maximum

Fig5.1: Input Gaussian Pulse

Fig5.2: Pulse Broadening Ratio for different values of GVD

Fig 5.3: Pulse Broadening Ratio for different values of Nonlinearity

Fig5.4: Pulse Broadening Ratio for different values of Power

Fig5.5: Input Hyperbolic secant Pulse

Fig5.6: Soliton Pulse Propagation

Fig5.7: Pulse Broadening Ratio for Soliton Pulse Propagation

Fig5.8: Pulse propagation for higher order soliton for N=2.

Fig5.9: Pulse propagation for higher order soliton for N=4.

vii
Fig5.10: Pulse propagation for higher order soliton for N=5

Fig5.11: Pulse Broadening Ratio for Higher Order Soliton with different N

Fig5.12: Pulse Broadening Ratio of Gaussian Pulse with different chirp

Fig5.13: Pulse Propagation for chirp 0

Fig5.14: Pulse Propagation for chirp -1

Fig5.15: Pulse Propagation for chirp -2

Fig5.16: Pulse Propagation for chirp 2

Fig5.17: Input pulse for different values of m.

Fig5.18: Supergaussian pulse propagation for m=4

Fig5.19 Supergaussian pulse propagation for m=10

Fig5.20: Supergaussian pulse propagation for m=50

Fig5.21: Supergaussian pulse propagation for m=100

Fig5.22: Pulse Broadening Ratio for different values of m for chirp 0

Fig5.23: Pulse Broadening Ratio for different values of chirp for m=100

Table 4.1: Gaussian input

Table 4.2: Super-Gaussian input

Table 4.3: Hyperbolic Secant input

Table 5.1: Test Plan

viii
CHAPTER 1: INTRODUCTION

CHAPTER 1
INTRODUCTION

1|Page
CHAPTER 1: INTRODUCTION

1.1 INTRODUCTION

In this thesis numerical analysis of nonlinear pulse propagation is carried out. This is done
mainly by solving the nonlinear Schrodinger equation using the symmetrized split step fourier
method. In a nonlinear media, dispersive effects exist simultaneously with nonlinear effects.
Refractive index dependence on intensity results in optical Kerr effect which causes narrowing
of transmitted pulses by inducing self phase modulation while second order group velocity
dispersion causes the pulses to spread. In this dissertation, group velocity dispersion is discussed
in relative detail followed by self phase modulation. These individually detrimental effects are
shown to combine beneficially for propagation of pulses here. The importance of chirp is
discussed along with its various properties. Induced chirp and input chirp are differentiated and
explained using equations.

Gaussian, super-gaussian pulses and hyperbolic secant pulses are studied and propagated by
using them as input in to the nonlinear Schrodinger equation. The symmetrized split step fourier
method is described in depth. Explanation of each step is included along with the relevant
equations defining these steps. A descriptive method of how symmetrized split step fourier was
incorporated as a solution of both simplified and normalized nonlinear Schrodinger equations is
presented in the implementation section. Pulse evolutions of these pulses under varying
circumstances and parameters are presented in this body. The purpose of obtaining numerical
calculations where done by calculating the pulse broadening ratio of the various pulses
propagated under different conditions [1]. Width used is a measure of the full width half
maximum in all cases.

2|Page
CHAPTER 1: INTRODUCTION

1.2 THESIS OBJECTIVE

In this thesis, we have primarily studied the effects of non-linear pulse propagations which
include hyperbolic secant, Gaussian, Super Gaussian pulses. We shown certain effects of chirp
on these pulses and using pulse broadening (PB) ratio described their effects. We also included
Super Gaussian pulses of higher order in this thesis and shown PB ratios of higher order soliton.

3|Page
CHAPTER 1: INTRODUCTION

1.3 BACKGROUND OF THE PROJECT

From the very beginning of the human race, people wanted to convey information quickly as
well as precisely. In order to do so, they tried to transfer information using different methods and
medium. Since the evolution of modern science, there were more efficient ways to transfer
information. Recently optical communication system has been the groundbreaking development.
Optical communication system was first conceived in the late 18th century by a French Engineer
Claude Chappe who created an optical telegraph. Since the advent of telegraphy in the 1830s,
electrical communication systems became the first foremost modern communication method.
Until the early 1980s, most of fixed (non-radio) signal transmission was carried by metallic cable
(twisted wire pairs and coaxial cable) systems. However, large attenuation and limited bandwidth
of coaxial cable limited its capacity upgrade. The bit rate of the most advanced coaxial system
which was used in the United States in 1975 was 274 Mb/s [1]. At approximately the same time,
there was a need of switching from analogue to digital transmission to improve transmission
quality, which calls for further boost of transmission bandwidth. Many efforts were prepared to
surmount the limitations of coaxial cable during the 1960s and 1970s. In 1966, Kao and
Hockham proposed the use of optical fiber as a guiding medium for the optical signal [1]. Four
years later, a major innovation occurred when the fiber loss was reduced to about 20dB/km from
previous values of more than 1000dB/km. From that time, optical communication technology has
developed rapidly to achieve larger transmission capacity and longer transmission distance. As a
result the transmission capacity has been increased about 100 fold in every 10 years [1].

The first generation of optical communication was designed with multi-mode fibers and direct
bandgap GaAs light emitting diodes (LEDs) which operate at the 0.8nm- 0.9nm wavelength
range [18]. In contrast to the typical repeater spacing of coaxial system (~1km), the longer
repeater spacing (~10km) was a major motivation. Large modal dispersion of multi-mode fibers
and high fiber loss at 0.8nm (> 5dB/km) limited both the transmission distance and bit rate. In
the second generation, multi-mode fibers were replaced by single-mode fibers, and the center

4|Page
CHAPTER 1: INTRODUCTION

wavelength of light sources was shifted to1.3nm, where optical fibers have minimum dispersion
and lower loss of about 0.5 dB/km.

However, there was still a strong demand to increase repeater spacing further, which could be
achieved by operating at 1.55nm where optical fibers have an intrinsic minimum loss around
0.2dB/km. Larger dispersion in the 1.55nm window delayed moving to a new generation until
dispersion shifted fiber became available. Dispersion shifted fibers reduce the large amount of
dispersion in the1.55nm window by modifying the index profile of the fibers while keeping the
benefit of low loss at the 1.55nm window [1].

However, growing communication traffic and demand for larger bandwidth per user revealed a
significant drawback of electronic regenerator systems, namely inflexibility to upgrade. Because
all the regenerators are designed to operate at a specific data rate and modulation format, all of
them needed to be replaced to convert to a higher data rate. The difficulty of upgradeability has
finally been removed by optical amplifiers, which led to a completely new generation of optical
communication. An important advance was that an erbium-doped single mode fiber amplifier
(EDFA) at 1.55nm was found to be ideally suited as an amplifying medium for modern fiber
optic communication systems [18]. Invention of the EDFA had a profound impact especially on
the design of long-haul undersea systems.

It is highly likely that WDM systems will bring another big leap of transmission capacity of
optical communication systems. Some research groups have already demonstrated that it is
possible to transmit almost a Tbits/s of total bit rate over thousands of kilometers. In 1999, for
example, N. Bergano et al. successfully demonstrated transmission of 640 Gb/s over 7200km
using a re-circulating loop while G. Vareille et al. demonstrated the transmission capacity of
340Gb/s over 6380km on a straight-line test bed [1]. These results indeed show that remarkable
achievements have been made in recent years, and let us forecast that optical communication
systems in the next generation will have a transmission capacity of a few hundreds of Gb/s.

5|Page
CHAPTER 1: INTRODUCTION

While high capacity dense WDM systems keep heading to closer channel spacing and broader
bandwidth of optical amplifiers to fully exploit the fiber bandwidth, on the other hand, upgrading
embedded systems remains as another challenge. As of the end of 1997, about 171 million km of
fiber have been deployed worldwide, of which 69 million km is deployed in North America [1].
Unfortunately, most of the embedded fibers are conventional single-mode fibers which have a
large dispersion at the 1.55nm window. Upgrading these systems will require various dispersion
combating techniques which are highly tuned at a specific system to optimize system
performance.

6|Page
CHAPTER 2: LITERATURE REVIEW

CHAPTER 2
LITERATURE REVIEW

7|Page
CHAPTER 2: LITERATURE REVIEW

2.1: DISPERSION

Dispersion is an impediment in optical fiber transmission, especially

in case of single mode fiber. Even though the word “Chromatic” refers to color of the light
pulse, it is the wavelength of a light pulse that is more relate to this type of dispersion.
Chromatic dispersion appears in optical fiber when an incoming light pulses turn up at the
receiver site at different time instants even if the pulses travel same path along the fiber. It
may appear that the chromatic dispersion is predominant fact in multimode fiber but it is
not that much of importance as modal dispersion is the dominant factor in multimode
dispersion [2]. Now if we concentrate on single mode fiber, it is indeed a dominant factor.
One query may arise in the process that if there is only one mode in single mode fiber; then
how come chromatic dispersion is a predominant limiting factor as it is consist of material
dispersion and waveguide dispersion. In single mode fiber, we may encounter the problem
as a mode might be composed of light pulses of different colors (different wavelengths).
Pulse broadening as a result of chromatic dispersion can be evaluated as follows:

t chrom  D( ) L

Here, D (  ) =chromatic dispersion parameter measured in picoseconds (ps) per nanometer

(nm) and kilometer (km) and   spectral width of a light source in nm.

If we consider the above equation once again, then we see that dispersion parameter can be
positive or negative. When D is less than zero, then GVD parameter  2 is positive. Hence the
dispersion is positive and is known as normal dispersion region. On the contrary, if the value of
D is greater than zero, then the GVD parameter  2 is negative which is known as anomalous
dispersion region [2] [3].

In case of a single-mode optical fiber, the zero-dispersion wavelength is the wavelength for
which material dispersion and waveguide dispersion eliminates each other. In silica-based optical
fibers, minimum material dispersion occurs naturally at a wavelength of approximately 1300nm.
Again, if we consider multimode optical fiber, it is the wavelength at which the amount of
material dispersion is minimum or zero. Thus it is also known as minimum dispersion
wavelengths.

8|Page
CHAPTER 2: LITERATURE REVIEW

2.1.1 PHASE VELOCITY AND GROUP VELOCITY

Phase velocity indicates the ideal characteristics of the ideal signals unlike a practical case of
combining group of signals. Phase velocity is the velocity of a certain phase point of an
Electromagnetic (EM) wave. If velocity is v, then v=  /  ; where  and  both are time
independent. Group velocity can be obtained by differentiating  with respect to angular
frequency,  . Group velocity can be referred as the speed at which the light power propagates
along the fiber in a specific mode. Any information signal and power travel at the group not at
the phase velocity.

Fig 2.1: Group Velocity (red) and phase Velocity (blue).

2.1.2 GROUP VELOCITY DISPERSION

Different frequency components of the light pulses traveling at different velocities produces
spreading light pulses in time and this phenomenon is known as Group Velocity Dispersion
(GVD). GVD also enumerates as the group delay dispersion parameter, D, which is applicable
for only uniform medium.

 d 2n d 1 2c
D = ( ) = - 2 2
c d 2
d v g  (2.1)

9|Page
CHAPTER 2: LITERATURE REVIEW

d 2
Here,  2 = is GVD parameter which determines the amount of pulse broadening on
d 2
propagation along the fiber from the above equation it is evident that GVD is also wavelength
dependent. The frequency dependence of the group velocity leads to pulse broadening simply
because different spectral components of the pulse disperse during propagation and do not arrive
simultaneously at the fiber output.

2.1.3 THIRD ORDER DISPERSION

In most cases pulse broadening due to the lowest-order Group Velocity Dispersion (GVD) term
proportional to β2. However in some cases the third-order term proportional to β3 also plays an
important role as well. If we Consider .the basic pulse propagation equation stated bellow, then
we see β2 is the dominant case in most cases.

(2.2)

However, if the light pulse wavelength is nearly similar with the zero-dispersion wavelength,
then β2 term becomes zero. As a result the β3 term then offers the principal part to the GVD
effects. For ultra short pulses (width T0 < 1 ps), it is necessary to include the β3 term even when
β2 is not equal to 0 because the expansion parameter Δω=ω0 is no longer small enough to justify
the truncation of the expansion in the following equation after the β2 term[3].

(2.3)

One practical example of third order dispersion is in case of mode-locked lasers which have
pulse durations nearly less than 30 femtoseconds [3], it is essential to supply dispersion
compensation for the group delay dispersion as well as for the third-order dispersion and
possibly for even higher orders.

10 | P a g e
CHAPTER 2: LITERATURE REVIEW

2.2 NONLINEAR EFFECTS

2.2.1 Non Linear Refractive Index

Nonlinear effects in fiber optic are mostly generated due to nonlinear refraction. Refractive index
generally defines the density of the propagating medium thereby describing how light would
propagate and at what speed. When light of very high intensity is launched into the fiber, an
additional variable contributes to the total refractive index. The total refractive index is defined
as follows:

n  n0  n2 I
2

(2.4)

Here, n 0 is the material refractive index and it is assumed constant for a certain frequency. The

second term is n2 |I| 2 defined as the nonlinear refractive index where n2 is nonlinear index co-
efficient. When the light pulses are having short wavelengths and very high intensities (such as
output of a laser) may vary a refractive index of a medium which as a result may give rise to
nonlinear optics. If the refractive index of the medium varies nonlinearly with the field (linearly
with the intensity), then it is known as the optical Kerr effect [3] and results in phenomena such
as self-focusing and self-phase modulation. If the index varies linearly with the field then it is
known as the Pockels effect.

The value of n 2 is affected by the experimental technique used to measure it. The reason is that
two other mechanisms which is related to molecular motion (the Raman effect) and excitation of
acoustic waves through electrostriction (Brillouin scattering), also contribute to n 2 . However,
their relative contributions depend on whether the pulse width is longer or shorter than the
response time associated with the corresponding process. The electrostatic contribution
diminishes for pulses shorter than 100 ps but reaches its maximum value (_16% of total n 2 ) for
pulse widths >10 ns. In contrast, the Raman contribution does not vanish until pulse widths is
less than 50 fs and is close to 18% for pulse widths greater than 10ps. However, we should be
careful when we are measuring the value using different pulse widths.

11 | P a g e
CHAPTER 2: LITERATURE REVIEW

As silica fiber has relatively low value of n 2 , several other kinds of glasses with larger

nonlinearities have been used to make optical fibers . For a lead-silicate fiber n 2 was measured to

be 2  10 19 m 2 /W. In chalcogenide As 2 S 3 -based fibers, the measured value n 2 = 4.2  10 18 m 2

/W can be larger by more than two orders of magnitude compared with the value for silica fibers
[3]. Such fibers are attracting increasing attention for applications related to nonlinear fiber
optics in spite of their relatively high losses. Their use for making fiber gratings and nonlinear
switches has reduced power requirements considerably.

2.2.2 TYPES OF NONLINEAR EFFECTS

2.2.2.1 SELF-PHASE MODULATION

Self-phase modulation (SPM) refers to the self-induced phase shift due to an optical field during
its propagation in optical fibers. As we know the refractive index of a medium depends on the
intensity of the pulse propagating through that medium but this can be ignored most of the times
except when pulse intensity becomes large. As ultra short pulses have high intensity, thus self
phase modulation occurs when these ultra short light pulses, whose time duration is on the order
of the femtosecond (10 -15 s), are propagated through an optical fiber. When they propagate
through an optical fiber, the varying intensity across the pulse envelope will face varying
refractive index and as a result the pulse spectral components will experience varying frequency
shift. . This results in pulse broadening which causes overlapping between adjacent bits and
limits maximum bit transfer rate.

The GVD broadens optical pulses during their propagation inside an optical fiber except when
the pulse is initially chirped in the right way. More specifically, a chirped pulse can be
compressed during the early stage of propagation whenever the GVD parameter  2 and the
chirp parameter C happen to have opposite signs so that β2C is negative [2]. The nonlinear
phenomenon of SPM imposes a chirp on the optical pulse such that C > 0. Since β 2 < 0 in the
1.55- μm wavelength region, the condition β 2C < 0 is readily satisfied. Moreover, as the SPM-
induced chirp is power dependent, it is not difficult to imagine that under certain conditions,

12 | P a g e
CHAPTER 2: LITERATURE REVIEW

SPM and GVD may cooperate in such a way that the SPM-induced chirp is just right to cancel
the GVD-induced broadening of the pulse [3]. The optical pulse would then propagate
undistorted in the form of a soliton.

2.2.2.2 Stimulated Raman scattering

Raman scattering is the inelastic type of scattering which implies that the kinetic energy is not
conserved in this scattering process. When light pulse is scattered from an atom, most photons
show elastically scattered nature which means that that after scattering the photons contain the
same energy or frequency and wavelength as compared to the incident photons.

Fig 2.2: Elastic Rayleigh scattering and Inelastic Raman Scattering

This process is known as Rayleigh scattering. In case of Raman scattering, only a small fraction
of the light pulses are scattered by an excitation where the scattered photons have a different
frequency. In most cases this frequency is lower than the frequency of incident photons. It is
evident now that this scattering process results in loss of power but at low power lever, this loss
is negligible since the scattering cross sections are small enough. However, at higher power level

13 | P a g e
CHAPTER 2: LITERATURE REVIEW

this Stimulated Raman Scattering (RBS) process emerges as an important phenomenon. As the
incident power exceeds a threshold value the intensity of the scattered light raises exponentially.
Stimulated Raman Scattering can take place in both the backward direction and the forward
direction of the optical fiber [3].

Fig 2.3: Raman Scattering

2.2.2.3 Stimulated Brillouin Scattering

Stimulated Brillouin Scattering (SBS) is another instance of inelastic scattering of photons.
However, the main difference is that the optical phonons participate in Raman scattering. On the
other hand acoustic phonons participate in Brillouin scattering. At the beginning, SRS and SBS
are quite similar in nature, in case of single mode fiber different dispersion relations for acoustic
and optical phonons pinpoint following differences:

1. SBS occurs only in the backward direction; on the contrary SRS can occur in both directions.

2. The scattered light is shifted, which is known as Stokes shift, in frequency by about 10 GHz
for SBS but for SRS it is up to 13 THz.

3. The Brillouin gain spectrum possesses extremely narrow bandwidth (less than 100MHz)
compared to Raman-gain spectrum that extends over 20–30 THz [3].

14 | P a g e
CHAPTER 2: LITERATURE REVIEW

2.3 PULSE PROPAGATION

The propagation of optical fields inside fibers like all electromagnetic fields are described by
Maxwell’s equation.

B
 E  
t (2.5)

H
 E  J 
t (2.6)

 D   (2.7)

Wave equation describes propagation of light in optical fibers. The following wave equation can
be obtained by mathematical manipulation of the Maxwell’s equations in terms of E and P ,
where P is the induced polarization of the wave [2] [12].

1 2 E 2 P
 E     (2.8)
c 2 t 2 t 2
0

Where c is the speed of light in vacuum and the relation 0 0  1/ c 2

When we consider the nonlinear domain, P can be dissected to obtain linear and nonlinear parts
as shown:

P  P L  P NL (2.9)

15 | P a g e
CHAPTER 2: LITERATURE REVIEW

P L = Linear part of induced polarization

P NL = Nonlinear part of induced polarization

Using the equation …..(2.8) and equation …….(2.9)we obtain the equation :

1 2 E 2 P
 E    
c 2 t 2 t 2
0
(2.10)

This wave equation is used as the starting point for deriving the nonlinear Schrödinger equation
which will be explained in the following section.

2.3.1 NONLINEAR SCHRODINGER EQUATION (NLSE)

2.3.1.1 Generalized NLSE

A  i 2  2 A 3  3 A     A 
2
 1 
    A   i A A  ( A A  iTR
2 2
)
z  2 T 2 6 T 3 2   0 A T T 

Self Steepening

Group Velocity Dispersion Nonlinearity

Attenuation Raman Scattering

16 | P a g e
CHAPTER 2: LITERATURE REVIEW

Third Order Dispersion

A  i 2  2 A 3  3 A    2 i 1   A
2


    
 A  i A A  ( A A  TR
2
)
z  2 T 2 6 T 3 2  
  0 A T T 
 (2.11)

2.3.1.2 Simplified NLSE

A  i 2  2 A 3  3 A  
   
z  2 T 2 6 T 3 2 
 A    A
2
A  
(2.12)

A  i 2  2 A  
  
z  2 T 2
 A
2 

   A
2
A  
(2.13)

Nonlinear parameter

2n2
  ; (2.14)
Aeff

Aeff  effective core area of the fiber ;

n2 = nonlinear index coefficient ;

 = optical wavelength

17 | P a g e
CHAPTER 2: LITERATURE REVIEW

T

T0 (2.15)

z
Z
LD (2.16)

A
U
P0 (2.17)

The previous normalization parameters are used in transforming the simplified NLSE to
normalized NLSE.

2.3.1.3 Normalized NLSE

iU s  2U 3 3 A 
   
2 2
N U U A
Z 2  2
6  3
2 (2.18)

iU s  2U
  U 0
2 2
N U
Z 2  2 (2.19)

U s  2U
i  U 0
2 2
iN U
Z 2  2
(2.20)

s=sign(  2 )=  1

If s=1 Normal GVD

18 | P a g e
CHAPTER 2: LITERATURE REVIEW

If s=-1 Anomalous GVD

P0T0
N 2  P0 LD =
2 (2.21)

2.3.2TYPES OF PULSES

In this study various type of input pulses are used to analyze their propagation through nonlinear
media. Among them are the following :

1. Hyperbolic secant pulses – equations, soliton : describe soliton etc, GVD cancels SPM
2. Gaussian pulses
3. Super-Gaussian pulses

2.3.3 HYPERBOLIC SECANT PULSES

The optical field for such pulses takes the form of:

 
U (0, )  N sec h 
 To  (2.22)

   iC 2 
U (0, )  N sec h  exp   
2 
 To   2To  (2.23)

2.3.4 GAUSSIAN PULSE

The incident field for gaussian pulses can be written as

19 | P a g e
CHAPTER 2: LITERATURE REVIEW

 2 
U (0, )  Ao * exp   
2 
 2T0  (2.24)

T0 is the initial pulse width of the pulse .

In theory, gaussian pulses while propagating maintain their fundamental shape , however their
amplitude width and phase varies over given distance [2] .

Many quantitative equations can be followed to study the properties of gaussian pulses as it
propagates over a distance of z .

Amplitude of a gaussian pulse varies with the equation

 2 
exp   
T0
U (0, T ) 
T 0
2
 i 2 
1/ 2
 
 2 T 0 2  i
2  
 (2.25)

Width of a gaussian pulse varies with this equation :

2 1/ 2
  z  
 1 ( z )  T0 1    
  LD   (2.26)

Under the effect of dispersion this equation shows how the gaussian pulse width broadens over z
.

Phase of the propagating pulse can be numerically determined with this formula :

20 | P a g e
CHAPTER 2: LITERATURE REVIEW

z
s( )
LD 2  1  
 ( z , )      tan 1  z  ;
z T 2  2 L
1  ( )2  0   D  (2.27)
LD

s=sign(  2 )

The phase  changes with respect to time .The derivative of phase change is called chirp
which will be described later on.

2.3.4.1 CHIRPED GAUSSIAN PULSE

Gaussian pulses with initial chirp have the following input field

 (1  iC) 2 
U (0, )  Ao * exp   2

 2To  (2.28)

A will be shown later , the effect of initial chirp results in varied behavior of the gaussian pulse
for various chirp parameters C . The amplitude, width and phase react to the inclusion of chirp
and effects depend on the sign and magnitude of the chirp [3].

Consider a width T1 after a distance of z. The broadening of chirped gaussian pulse can be
written as

1/ 2
T1   C 2 z    2 z  
2 2

 1      
T0   T0 2   T0 2  
  (2.29)

21 | P a g e
CHAPTER 2: LITERATURE REVIEW

The broadening depends upon the signs of  2 and C.

If  2 C<0 , pulse width becomes minimum :

T0

min
T1
(1  C 2 )1 / 2 (2.30)

at a distance of

|C|
z min  LD
1 C2 (2.31)

2.3.5 SUPERGAUSSIAN PULSES

The input field for such pulse is described by

 (1  iC) 2 m 
U (0, )  Ao * exp   2

 2To  (2.32)

Here m is the order of the supergaussian pulse and determines the sharpness of the edges of the
input . The higher the value of m steeper is the leading and trailing edges of the pulse .

22 | P a g e
CHAPTER 2: LITERATURE REVIEW

As we continue increasing m , we eventually get a rectangular pulse shape which evidently has
very sharp edges. In case of m=1 we get the gaussian chirped pulse.

The sharpness of the edges plays an important part in the broadening ratio because broadening
caused by dispersion is sensitive to such a quality [12].

2.4CHIRP

2.4.1 CHIRP DEFINITION

Chirp of an optical pulse is the change of carrier frequency with respect to time in a deterministic
fashion. If the instantaneous frequency of the pulse increases with time then it is called an up-
chirp. Similarly, if the instantaneous frequency of the pulse reduces with time then it is called a
down-chirp.

Input pulse

23 | P a g e
CHAPTER 2: LITERATURE REVIEW

Fig2.4: Chirp

An input pulse with chirping can be expressed as-

(2.33)

Where, C is a chirp parameter. If a positive chirp is applied (C>0), the frequency of the pulse
will increase while propagating through the fiber. Oppositely, the frequency will decrease if
negative chirp (C<0), is applied.

We have seen before, the phase of a pulse propagating through a nonlinear medium changes
along time. Frequency of a pulse is nothing but the change of phase of that pulse. The frequency
also changes from initial frequency as a result of time dependence of phase across pulse [13].
This frequency change for a Gaussian pulse can be calculated by

(2.34)

24 | P a g e
CHAPTER 2: LITERATURE REVIEW

2.4.2 Types of Chirp:

Based on induction, there are three types of Chirp.

GVD induced Chirp

SPM induced Chirp

Pre induced Chirp

2.4.2.1 GVD induced Chirp:

GVD induced Chirp occurs due to the group velocity dispersion effect while pulse propagate
through an optical fiber. The chirp is linear and negative over a large central region. It results spreading of
pulse width from the original pulse width. It is positive near the leading edge and becomes negative near
the trailing edge of the pulse [2].

2.4.2.2 SPM induced Chirp:

SPM induced chirp is up-chirp or positive-chirp that occurs due to the nonlinear self phase
modulation effect inside an optical fiber. The chirp is linear and positive over a large central region of the
Gaussian pulse. SPM induced Chirp makes the pulse width narrower from the original pulse width as
pulses propagate through an optical fiber. It is negative near the leading edge and becomes positive near
the trailing edge of the pulse.

2.4.2.3 Pre induced Chirp:

Pre induce chirp is the chirp added to an input pulse before sending it through an optical fiber. Pre
induced Chirp can be either positive or negative. Pre induce Chirp is used to balance the GVD induced
chirp and SPM induced chirp. Also chirping is added for archiving the following advantages-

25 | P a g e
CHAPTER 2: LITERATURE REVIEW

 Suppression of four wave mixing

 Lowering pulse width fluctuations

 Spectral compression

 And generating transform limited output pulses

2.4.4 Critical Chirp:

Critical chirp is the maximum value of chirp of an input pulse for which soliton can be formed.

Chirp disturbs the balance between group velocity dispersion and self phase modulation.
Therefore chirp of the input pulse is kept to a minimum level for propagating soliton pulses. Furthermore,
higher chirp induced to an input pulse results loss of the pulse energy while forming the soliton as
dispersive waves. That’s why soliton cannot be formed if the chirping crosses a certain value. That value
of chirp is called Critical Chirp. For first order soliton the critical chirp value is 1.64.

2.4.5 Effects of Chirp:

GVD induced Chirp:

Hare effects of GVD induced chirp is only for anomalous dispersion (β₂ <0)

 When Pre Chirp is positive(C>0) : instantaneous frequency increases linearly from the leading to
the trailing edge (up chirp)

 When β₂ C<0: dispersion induced chirp(β₂) is in opposite direction to Pre Chirp (C). As a result
net chirp is reduced. Pulse width will be narrowing at first. After a distance, GVD induced chirp
dominates C and pulse spreads.

 When C<0: Instantaneous frequency decreases linearly form the leading edge to the trailing edge
of the propagated pulse (up chirp).

 When β₂ C>0: Pulse width broadens monotonically.

SPM induced Chirp:

 SPM induced frequency chirp is linear and positive over central portion of Gaussian pulse.

26 | P a g e
CHAPTER 2: LITERATURE REVIEW

 Positive chirp: (C>0): The SPM induced chirp and initial pre chirp adds together. And Results
Pulse width broadening together.

 Negative chirp (C<0): Deducts positive SPM induced chirp from negative pre chirp. The result
depends on the magnitude of the pre chirp. If the value of pre chirp is higher than the SPM
induced chirp, then the result will be pulse width broadening [13]. If the pre induce chirp is lower
than the SPM induced chirp, then the pulse width will be spread at first, but as propagating
distance becomes higher, SPM induced chirp becomes dominant and the pulse width narrows.

2.5 SOLITON:
Soliton in optical fiber generally means special type of pulse that can propagate undistorted over long
distance. Soliton is a result of dispersion and nonlinear effects cancelling each other. Red shift at the
leading edge of a pulse is caused by Self phase modulation [2.]In anomalous dispersion region blue shift
occurs at the leading edge of a pulse. As a result in anomalous dispersion these effects cancel each other
and the input pulse maintain its shape in both time domain and frequency domain.

2.5.1: CONDITIONS FOR SOLITON:

The conditions for soliton are

1) The dispersion region must be anomalous. That is <0.

2) The input pulse must be an un-chirped hyperbolic secant pulse. In our simulation we used the
following pulse-

 
U (0, )  sec h 
 To  (2.35)

27 | P a g e
CHAPTER 2: LITERATURE REVIEW

3) The pulse energy has to be-

(2.36)

4) The dispersion length must be approximately the same as the nonlinear length.
5) The GVD induced chirp should exactly cancel the SPM induced chirp.

2.5.2: Higher Order Soliton:

Higher order soliton are soliton with higher energy. More specifically, the energy of a higher order soliton
is square of an integer number times higher than a fundamental soliton.

In our simulation we used –

 
U (0, )  N sec h 
 To  (2.37)

Higher order soliton do not have a fixed pulse shape like fundamental soliton. But they gain their shape
periodically.

28 | P a g e
CHAPTER 3: DESCRIPTION OF METHOD

CHAPTER 3

DESCRIPTION OF METHOD

Page | 28
CHAPTER 3: DESCRIPTION OF METHOD

3.1 SPLIT STEP FOURIER METHOD

The split step fourier method is a pseudo-spectral numerical method for solving
partial differential equations such as the nonlinear Schrödinger equation. It is
applied because of greater computation speed and increased accuracy compared to
other numerical techniques.

Dispersion and nonlinear effects act simultaneously on propagating pulses during

nonlinear pulse propagation in optical fibers. However, analytic solution cannot
be employed to solve the NLSE with both dispersive and nonlinear terms present.
Hence the numerical split step fourier method is utilized, which breaks the entire
length of the fiber into small step sizes of length ‘h’ and then solves the nonlinear
Schrödinger equation by splitting it into two halves , the linear part (dispersive
part) and the nonlinear part over z to z + h [7]. Each part is solved individually
and then combined together afterwards to obtain the aggregate output of the
traversed pulse. It solves the linear dispersive part first, in the fourier domain
using the fast fourier transforms and then inverse fourier transforms to the time
domain where it solves the equation for the nonlinear term before combining
them. The process is repeated over the entire span of the fiber to approximate
nonlinear pulse propagation. The equations describing them are offered below [7].

 max   | Ap | 2 h,
The value of h is chosen for where max  0.05rad and Ap =

peak power of A (z, t); max = maximum phase shift

In the following part the solution of the generalized Schrödinger equation is

described using this method.

Page | 29
CHAPTER 3: DESCRIPTION OF METHOD

A  i 2  2 A  3  3 A     A 
2
 1 
    A   i A A  ( A A)  iTR
2 2

z  2 T 2 6 T 3 2    A T T 
 0 
(3.1)

A
 ( Lˆ  Nˆ ) A
t (3.2)

The linear part (dispersive part) and the nonlinear part are separated.

Linear part

 i  2 A  3  3 A  
Lˆ    2   A
 2 T
2
6 T 3 2  (3.3)

Nonlinear part

  1   A 
2
ˆ 
N  i A A 
2
( A A)  iTR
2

  A T T 
 0  (3.4)

Solution

A( z  h, T )  exp[ h( Lˆ  Nˆ )] A( z, t )
(3.5)

A( z  h, T )  exp( hLˆ ) exp( hNˆ ) A( z, t )

(3.6)

Linear part solution

The solution of the linear part exp( hLˆ ) is done in the fourier domain by using the

identity  i . Since in the fourier domain  is simply a numerical sequence
T

Page | 30
CHAPTER 3: DESCRIPTION OF METHOD

of digits, the calculation that would otherwise be complicated in the time domain
due to computation of the differential terms are mitigated in the fourier domain.

exp( hLˆ )  exp[ hLˆ (i)] and A(z,t)  A( z,   0 )

Linear Solution = FT
1
exp[ hLˆ(i)]  A( z,    )
0
(3.7)

Nonlinear part solution

A( z  h, T )  exp( hNˆ )  Linear Solution

(3.8)

In our study as will be shown later, we have considered  3 =0,   0 and higher
order nonlinear effects of self steepening and Raman scattering (reserved for ultra

A
2
 1 
( A A)  iTR  0 . We make this
2
short pulses) to be 0 i.e.
0 A T T
assumption because the pulse width chosen is of the order of picoseconds.

Fig 3.1: Split step fourier method

Page | 31
CHAPTER 3: DESCRIPTION OF METHOD

3.2 SYMMETRIZED SPLIT STEP FOURIER METHOD

The accuracy of the SSFM can be further improved by using the symmetrized
form of the method. In this technique, the linear dispersive part is solved in two
equal steps each of size h/2. To begin with, the first half of the linear part is
solved in the fourier domain over z to z +h/2 [7].It is then inverse fourier
transformed and then combined with the nonlinear part which is solved in the
time domain using step size h. The resultant outcome is then multiplied with the
second half of the linear dispersive part over z to z +h/2 in the fourier domain.
Fourier transformation is done using the fast fourier transform method as before.

Solution

If h is small, then:

h h
A( z  h, T )  [exp( Lˆ ) exp( hNˆ ) exp( Lˆ )] A( z, t )
2 2 (3.9)

If h is larger than accuracy can be improved using this equation:

A( z  h, T )  [exp( Lˆ ) exp  Nˆ ( z)dz  exp( Lˆ )] A( z, t )

h z h h
2  z  2 (3.10)

Using the trapezoid rule we have an approximation which further improves

accuracy of calculation:

z
z h h

(hNˆ )   Nˆ ( z)dz  Nˆ ( z )  Nˆ ( z  h)
2

(3.11)

h
2
h
 


h
A( z  h, T )  [exp( Lˆ ) exp  Nˆ ( z )  Nˆ ( z  h)  exp( Lˆ )] A( z, t )
2 2 (3.12)

Page | 32
CHAPTER 3: DESCRIPTION OF METHOD

Linear part solution: Step1

h h
exp( Lˆ )  exp[ Lˆ (i )] and A(z,t)  A( z,   0 )
2 2 (3.13)

1  h 
Linear Solution(1) = FT exp[ Lˆ (i )]  A( z,   0 )
 2  (3.14)

Nonlinear part solution

Nonlinear solution  exp( hNˆ )  Linear Solution(1)

(3.15)

Linear part solution: Step 2

1  h 
A(z +h ,T) = Nonlinear Solution  FT exp[ Lˆ (i )]  A( z,   0 )
 2 

(3.15)

Fig 3.2 Symmetrized split step fourier method

Page | 33
CHAPTER 3: DESCRIPTION OF METHOD

3.3 DECISIONS

We had several options before concerning which method to use for solving the
nonlinear Schrödinger equation. In the end we decided to use the symmetrized
split step fourier method. The reasons are given as follows:

 The split step fourier method(SSFM) can be used to solve the NLSE
accurately up to the second order dispersion term as relevant to our
dissertation.(See Appendix for Taylor series explanation )
 SSFM is faster because the differential terms in the dispersion (linear) part

is calculated in the fourier domain using the relation  i .
t
 The symmetrized SSFM is used because it further improves the accuracy
of the procured results depending on the selected value of the step size h.
 The FFT method is used to fourier transform back and forth in the
symmetrized SSFM which reduces computation time because FFT is the
one of the fastest methods available for transformation.

We solved two different forms of the nonlinear Schrödinger equation, the

simplified NLSE and the normalized NLSE, using two different MATLAB
codes. This is done because solution of gaussian pulses and supergaussian
pulses compared to hyperbolic secant pulses require different parameters and
conditions.

Page | 34
CHAPTER 4: IMPLEMENTATION

CHAPTER 4

IMPLEMENTATION

35 | P a g e
CHAPTER 4: IMPLEMENTATION

4.1 NONLINEAR SCHRODINGER EQUATION (NLSE)

SOLUTIONS

4.1.1 Generalized NLSE

2 3 2
A i 2 A 3 A 2 i 1 2 A
A i A A ( A A TR )
z 2 T2 6 T3 2 0 A T T

(4.1)

4.1.2 SIMPLIFIED NLSE

2
A i 2 A 2
A A A
z 2 T2 2
(4.2)

In the previous equations the generalized NLSE is simplified as the higher order
nonlinear effects of self steepening and Raman scattering are removed.

2
1 2 A
( A A) i TR 0
0 A T T
(4.3)

Since we have chosen our initial pulse width in order of picoseconds( T0 1 ps )

this assumption is feasible.

36 | P a g e
CHAPTER 4: IMPLEMENTATION

We have also considered 3 =0 .For the third order dispersion (TOD) to have
sufficient impact on nonlinear propagation , the optical wavelength must be close
to the zero-dispersion wavelength. In this case 2 0 which causes TOD to take

effect.

Hence, the criteria we have chosen ensures that 2

is the dominant dispersion
effect thereby justifying removal of the TOD from the NLSE.

4.1.3 NORMALIZED NLSE

U s 2U 2
i iN 2 U U 0
Z 2 2 (4.4)

37 | P a g e
CHAPTER 4: IMPLEMENTATION

4.2 GAUSSIAN IMPLEMENTATION

GAUSSIAN PULSE INPUT SIMPLIFIED NLSE METHOD

(1 iC)T 2 A i 2
A 2
SSSFM
2
U (0, ) Ao * exp A A A
2to 2 z 2 T2 2

Table 4.1 Gaussian input

Fourier spectrum = FT (U(0, ))

Linear dispersive part (1st half)

2
i 2 h
Linear solution(1) = FT (U(0, )) exp
2 2 2

Inverse fourier transform:

1
Linear solution(1) in time domain = FT [ Linear solution(1)]

Nonlinear part

2
Nonlinear solution in time domain= exp i Linear solution(1 ) in time domain h

Nonlinear part output =Linear solution(1) in time domain Nonlinear solution in time domain

38 | P a g e
CHAPTER 4: IMPLEMENTATION

Fourier transform

Nonlinear part output fourier domain= FT [Nonlinear part output]

Linear dispersive part(2nd half)

2
i 2 h
Linear solution(2) = Nonlinear part output fourier domain exp
2 2 2

Inverse fourier transform

1
Linear solution(2) in time domain= FT [Linear solution(2)]

SOLUTION FOR NLSE OVER 1 STEPSIZE h = Linear solution(2) in time domain

This process is repeated over the length of the fiber.

SUPER GAUSSIAN PULSE INPUT SIMPLIFIED NLSE METHOD

(1 iC)T 2 m A i 2
2
A 2 SSSFM
U (0, ) Ao * exp A A A
z 2 T2 2
2to 2

Table 4.2 Supergaussian input

The Gaussian implementation method can also be applied for supergaussian pulse.

39 | P a g e
CHAPTER 4: IMPLEMENTATION

4.3 SOLITON IMPLEMENTATION

HYPERBOLIC SECANT PULSE NORMALIZED NLSE METHOD

INPUT

2 SSSFM
U (0, ) N sec h U s U 2
To i 2
iN 2 U U 0
Z 2

Table 4.3 Hyperbolic Secant input

Fourier spectrum = FT (U(0, ))

Linear dispersive part (1st half)

2
is h
Linear solution(1) = FT (U(0, )) exp
2 2 2

Inverse fourier transform:

1
Linear solution(1) in time domain = FT [ Linear solution(1)]

Nonlinear part

2
Nonlinear solution in time domain= exp iN 2 Linear solution(1 ) in time domain h

Nonlinear part output =Linear solution(1) in time domain Nonlinear solution in time domain

40 | P a g e
CHAPTER 4: IMPLEMENTATION

Fourier transform

Nonlinear part output fourier domain= FT [Nonlinear part output]

Linear dispersive part(2nd half)

2
is h
Linear solution(2) = Nonlinear part output fourier domain exp
2 2 2

Inverse fourier transform

1
Linear solution(2) in time domain= FT [Linear solution(2)]

SOLUTION FOR NLSE OVER 1 STEPSIZE h = Linear solution(2) in time domain

This process is repeated over the length of the fiber.

41 | P a g e
CHAPTER 4: IMPLEMENTATION

4.4 FLOW CHART

42 | P a g e
CHAPTER 4: IMPLEMENTATION

43 | P a g e
CHAPTER 4: IMPLEMENTATION

44 | P a g e
CHAPTER 4: IMPLEMENTATION

4.5 PULSE BROADENING RATIO

In this thesis, analysis is done by using the pulse broadening ratio of the evolved pulses. Pulse
broadening ratio is calculated by using the Full Width at Half Maximum.

Pulse broadening ratio=FWHM of propagating pulse / FWHM of First pulse.

Pulse broadening ratio signifies the change of the propagating pulse width compared to the pulse
width at the very beginning of the pulse propagation.

4.5.1 FULL WIDTH HALF MAXIMUM

At the half or middle of the pulse amplitude, the power of the pulse reaches maximum. The
width of the pulse at that point is called full width half maximum.

Fig 4.4: Full Width at Half Maximum

45 | P a g e
CHAPTER 4: IMPLEMENTATION

4.4 FLOW CHART

START

loop = 1

INPUT
2
(1 iC)
A(0, ) Ao * exp
2To 2
INPUT FWHM

fourier.spectrum=fft(A)

for splitstep , from h to Z by h RATIO

(splitstep=h:h:Z)

2
i 2 h
fourier.spectrum fourier.spectrum * exp
2 2 2 FWHM
saved at
f = ifft(fourier.spectrum) every
interval
2
f f * exp i f h

fourier.spectrum= fft(f)
FWHM(1:loop )
2
i 2
h
fourier.spectrum fourier.spectrum * exp
2 2

SAVING PULSE AT
EVERY INTERVAL

YES loop = loop +1

STOP splitstep=Z

NO f=fft(fourier.spectrum)

Fig 4.1: Gaussian Flowchart

Page | 42
CHAPTER 4: IMPLEMENTATION

START

loop = 1

INPUT

U (0, ) N sec h
To

INPUT FWHM

fourier.spectrum=fft(U)

x=0

for splitstep, from h to Z by h RATIO

(splitstep=h:h:Z)

2
is h
fourier.spectrum fourier.spectrum * exp FWHM
2 2 2
saved at
every
f = ifft(fourier.spectrum) interval

2
f f * exp iN 2 f h

fourier.spectrum= fft(f)
FWHM(1:loop )
is 2 h
fourier.spectrum fourier.spectrum * exp
2 2 2

SAVING PULSE AT
EVERY INTERVAL
x = x +1

YES splitstep=Z loop = loop +1

STOP

NO f=ifft(fourier.spectrum)

Fig 4.2: Soliton Flowchart

43 | P a g e
CHAPTER 4: IMPLEMENTATION

START

loop = 1

INPUT
2m
(1 iC)
A(0, ) Ao * exp
2To 2
INPUT FWHM

fourier.spectrum=fft(A)

for splitstep , from h to Z by h RATIO

(splitstep=h:h:Z)

2
i h
fourier.spectrum fourier.spectrum * exp 2
FWHM
2 2 2
saved at
every
interval
f = ifft(fourier.spectrum)

2
f f * exp i f h

fourier.spectrum= fft(f)
FWHM(1:loop )
2
i 2
h
fourier.spectrum fourier.spectrum * exp
2 2

SAVING PULSE AT
EVERY INTERVAL

YES loop = loop +1

STOP splitstep=Z

NO f=fft(fourier.spectrum)

Fig 4.3: Supergaussian Flowchart

Page | 44
CHAPTER 5: ANALYSIS OF RESULTS

CHAPTER 5
ANALYSIS OF RESULTS

46 | P a g e
CHAPTER 5: ANALYSIS OF RESULTS

ANALYSIS PLAN:

NUMBER INPUT PULSE PROPERTIES

1 Gaussian Pulse Pulse broadening ratio for variable group

velocity dispersion parameters  2
2 Gaussian Pulse Pulse broadening ratio for various nonlinear
parameter 
3 Gaussian Pulse Pulse broadening ratio for various Input
Power, I
4 Fundamental Soliton, Pulse broadening ratio
Hyperbolic secant pulse
5 Higher Order Soliton, Pulse broadening ratios for various power of
Hyperbolic secant pulse the soliton order N
6 Gaussian Pulse Pulse broadening ratio for Chirp, C = 0

7 Gaussian Pulse Pulse broadening ratio for Chirp ,C = 1

8 Gaussian Pulse Pulse broadening ratio for Chirp, C = 2

9 Gaussian Pulse Pulse broadening ratio for Chirp , C = -2

10 Super Gaussian Pulse Pulse Broadening Ratio for different values of

chirp for m=4
11 Super Gaussian Pulse Pulse Broadening Ratio for different values of
chirp for m=100

Table 5.1: Test Plan

47 | P a g e
CHAPTER 5: ANALYSIS OF RESULTS

GAUSSIAN PULSE ANALYSIS

Input curve for Gaussian pulse

Fig5.1: Input Gaussian Pulse

48 | P a g e
CHAPTER 5: ANALYSIS OF RESULTS

Pulse broadening ratio for different dispersion parameter,  2 is given below:

Fig5.2: Pulse Broadening Ratio for different values of GVD

5.2.1 Analysis of Beta variation:

In this figure, the pulse broadening ratios of gaussian pulses are observed using variable group
velocity dispersion parameters  2 while keeping constant the power and nonlinear parameter  .

The values chosen for this study are  2 =  20  10 27 ,  40  10 27 ,  60  10 27  100  10 27 ,

 500  10 27 .

The purpose here is to simply observe what the effect of increasing or decreasing dispersion
parameter will have on pulse broadening and how much. It is seen, that over a length of fiber the
increasing  2 leads to a general increase in the pulse broadening ratio.

49 | P a g e
CHAPTER 5: ANALYSIS OF RESULTS

The physical meaning of this is that the high frequencies (blue shifted) of given optical pulse
travels faster than the low frequencies(red shifted) which causes GVD induced negative chirp in
the anomalous dispersion region . This essentially causes spreading of the pulse. Now increasing
the GVD parameter  2 results the high frequencies to travel faster and the low frequencies to
travel slower. So for a given length of propagation each frequency component of an optical pulse
arriving with larger and larger delay resulting in greater pulse spreading with each increasing
value of  2 .

However, as we continue increasing the GVD parameter we notice a property that the difference
of Pulse broadening ratio curves between higher values of  2 , for example  500  10 27 and

 100  10 27 is lesser than that between  60  10 27 and  40  10 27 .

50 | P a g e
CHAPTER 5: ANALYSIS OF RESULTS

Pulse broadening ratio for various nonlinear parameter  is given below:

Fig 5.3: Pulse Broadening Ratio for different values of Nonlinearity

5.2.2 Analysis of Nonlinear parameter variation:

In this figure, we study the importance of magnitude of nonlinear parameter  on nonlinear

optical fiber.

51 | P a g e
CHAPTER 5: ANALYSIS OF RESULTS

While keeping the input power and the GVD parameter constant, we generate curves for various
values of  , both greater and smaller than the ideal value used for proper GVD and SPM
balance.

The values chosen for this study are  = 0.003, 0.006, 0.009, 0.011 and 0.001 /W/m.

The purpose here is to observe what the effect of increasing or decreasing nonlinear parameter
will have on pulse broadening and how much. It is seen, that over a length of fiber the
increasing  leads to an initial decrease in the pulse broadening ratio.

For  =0.003 , the SPM induced positive chirp gradually cancels out the GVD induced negative
chirp obtaining steady width pulse propagation.

For  =0.006, the pulse appears to initially narrow more than the previous case because of
increased nonlinearity hence increased SPM effect. In this case also, the pulse width once
appears to gradually attain towards a constant pulse width. This happens because although
initially SPM effect becomes dominant , as propagation distance increases the GVD effect
increases eventually cancelling out the SPM induced positive chirp.

For  = 0.009 and 0.011, the pulse broadening ratio initially decreases to a minimum value. The
minimum value for  =0.011 is smaller than for  =0.009, which means that the pulse narrows
more for the former than the latter expected. However, as GVD effect becomes more prominent,
in these two cases the pulse does not acquire a constant pulse width further down the propagating
line.

For  =0.001, it is obvious that the SPM effect is not large enough counter the GVD effect.

From these observations, we can deduce that for a given  2 , we can increase the  value to a
certain point until which we can obtain equal GVD and SPM effect cancellation. Beyond this
value  becomes dominant and pulse shape is lost.

52 | P a g e
CHAPTER 5: ANALYSIS OF RESULTS

Figure of pulse broadening ratio for Gaussian pulse with input power =0.00005w, 0.0001w,
0.00064w and 0.001w.

Fig5.4: Pulse Broadening Ratio for different values of Power

5.2.3 Analysis for Power variation:

In this figure, we study the importance of magnitude of power on nonlinear optical fiber.

While keeping the nonlinear parameter and the GVD parameter constant, we generate curves for
various values of input power, both greater and smaller than the ideal value used for proper GVD
and SPM balance. It is observed from this plot, that the pulse broadening ratio for curves with
smaller input power is more than that with larger input power.

53 | P a g e
CHAPTER 5: ANALYSIS OF RESULTS

This property can be explained by the following equation:

1
LN  ; where, Po is the input power and LN is the nonlinear length.
Po

This equation shows that the nonlinear length is inversely proportional to the input power.

As a result LN decreases for higher values of Po . For Po =0.001W it is observed that the pulse
broadening ratio decreases meaning narrowing of pulses. This is surprising because although the
nonlinear parameter  is constant; narrowing of pulses continues to occur.

The reason is that the same amount of nonlinear effect occurs but it manifests itself over LN
which is now shorter and hence narrowing occurs faster than before.

Reducing Po results in an opposite effect. The  stays constant but LN is now larger and hence
narrowing occurs over a larger distance and hence slower. So GVD acts quicker and becomes
dominant and so spreading of pulse occurs.

54 | P a g e
CHAPTER 5: ANALYSIS OF RESULTS

5.3 SOLITON PULSE ANALYSIS

Input Soliton pulse:

Fig5.5: Input Hyperbolic secant Pulse

Pulse propagation for fundamental soliton:

Fig5.6: Soliton Pulse Propagation

55 | P a g e
CHAPTER 5: ANALYSIS OF RESULTS

Pulse broadening Ratio for soliton propagation:

Fig5.7: Pulse Broadening Ratio for Soliton Pulse Propagation

Soliton pulse Analysis:

Hyperbolic secant pulse propagation without initial chirp is simulated by analytically solving the
normalized Schrodinger equation using the split step fourier method. From Figure 5.6 we
observe that the pulse propagates at seemingly constant width. From the curve in Figure 5.7 the
pulse broadening ratio is found to be a steady, horizontal line confirming that the pulse travels at
a constant width which is equal to the input width of the pulse. This propagation of a constant
width hyperbolic secant pulse means that we have obtained soliton propagation in nonlinear
optical fiber.

56 | P a g e
CHAPTER 5: ANALYSIS OF RESULTS

Soliton propagation is possible under certain conditions. In general, the dispersion length, LD

must be approximately equal to the nonlinear length, LN . This would mean that the group
velocity dispersion would take effect over the same length as the nonlinear effects.

Under such circumstances, the GVD effect matches the SPM effects entirely and cancels each
other out to obtain steady pulse width throughout the length of the fiber. We assume that no
attenuation is present for simplification of solution.

The GVD and the SPM manifest themselves through their induced chirp effects. For GVD, the
high frequency components of the pulse travel at higher velocity than the low frequencies which
induces negative chirp causing dispersion. SPM on the other hand, induces positive chirp during
propagation causing the pulse to narrow as it evolves.

When LD  LN , the negative induced chirp resulting from the GVD and the positive induced
chirp caused by the SPM become equal in magnitude and opposite in direction.

The net outcome is that the negative induced chirp of GVD cancels out the positive induced
chirp of SPM equally producing fundamental soliton (N=1).

57 | P a g e
CHAPTER 5: ANALYSIS OF RESULTS

5.4 HIGHER ORDER SOLITON ANALYSIS

Pulse propagation for higher order soliton for N=2.

Fig5.8: Pulse propagation for higher order soliton for N=2.

58 | P a g e
CHAPTER 5: ANALYSIS OF RESULTS

Pulse propagation for higher order soliton for N=4.

Fig5.9: Pulse propagation for higher order soliton for N=4.

Pulse propagation for higher order soliton for N=4.

Fig5.10: Pulse propagation for higher order soliton for N=5.

59 | P a g e
CHAPTER 5: ANALYSIS OF RESULTS

Pulse Broadening Ratio for Higher Order Soliton Pulse Propagation with N=2, 3, 4 and 5:

Fig5.11: Pulse Broadening Ratio for Higher Order Soliton with different N

Higher order soliton Analysis:

Here, the power of the soliton order N is increased gradually to obtain pulse broadening ratios of
higher order solitons. These ratios are plotted on the same axis for comparison of pulse
propagation of each of the higher order solitons for a given length of fiber. The values chosen for
N are 2,3,4,5 and 10.

 T0 2 
N  P0 LD  P0 
2


 2 

60 | P a g e
CHAPTER 5: ANALYSIS OF RESULTS

The previous equation shows N2 which is the nonlinear factor used for the solution of the time
domain nonlinear part of the normalized nonlinear Schrodinger equation. The input to the NLSE
is:


U(0,  )=Nsech
T0

So increasing the value of N not only affects the magnitude of the input but also influences the
nonlinear solution of the NLSE resulting in the varied behavior of each higher order soliton.

The second order soliton (N=2, green line) produces a periodic outcome. It is observed that after
a certain period of traversed length the initial pulse shape is re-acquired. This pattern continues
to occur at regular intervals. The length of propagating distance after which the initial pulse
shape is re-obtained is called the soliton period and is given by the following equation:

   T0 
2

z 0  LD  
2 2   2 

The third order soliton (N=3, Blue line) also produces a periodic pattern in the pulse broadening
ratio curve. At regular intervals, the width of the pulse oscillates; however it does not seem to
re-acquire the original pulse shape but slowly heads towards acquiring a different pulse width.

The fourth (red line) and fifth order (black line) oscillates for shorter and shorter period of
length. At the end of the chosen propagating distance, the pulse broadening ratio seems to appear
less oscillatory and heading towards achieving constancy at a much lower value.

61 | P a g e
CHAPTER 5: ANALYSIS OF RESULTS

However, this event is characterized by splitting of pulses along the length of propagation.
(Waterfall figures of higher order pulse evolution attached with appendix).

The reason for the behavior of higher order soliton is that the pulse splits into several small
pulses. The net effect is the eventual destruction of the intended information due to
fragmentation of the pulses.

62 | P a g e
CHAPTER 5: ANALYSIS OF RESULTS

5.5 CHIRPING EFFECT ANALYSIS

Pulse broadening ratio of Gaussian pulse with chirp, C=2, 0, -1, -2.

Fig5.12: Pulse Broadening Ratio of Gaussian Pulse with different chirp

63 | P a g e
CHAPTER 5: ANALYSIS OF RESULTS

Figure of Pulse Propagation of Gaussian pulse for chirp 0

Fig5.13: Pulse Propagation for chirp 0

5.5.1 Analysis for Un-chirped Gaussian pulse

Here, both GVD and SPM act simultaneously on the propagating gaussian pulse with no initial
chirp. As can be seen from the evolution graph, the pulse shrinks initially for a very small period
of propagating length. After that the broadening ratio reaches a constant value and a stable pulse
is seemed to propagate as can be seen from pulse broadening ratio curve.

GVD acting individually results in the pulse to spread gradually before it loses shape. SPM
acting individually results in the narrowing of pulses and losing its intended shape.

64 | P a g e
CHAPTER 5: ANALYSIS OF RESULTS

The combined effect of GVD and SPM leads to the eventual generation of constant pulse
propagation emulating a hyperbolic secant pulse.

Initially, the positive chirp induced by SPM seems to dominate the negative chirp caused by the
GVD which accounts for the narrowing of the propagating pulse in the early steps. But after a
certain period, GVD effect becomes more prominent and it catches up to the SPM effect. At one
point, GVD induced negative chirp balances the SPM induced positive chirp, and they cancel
each other out and propagates at a narrower pulse width emulating hyperbolic secant pulse
propagation.

65 | P a g e
CHAPTER 5: ANALYSIS OF RESULTS

Figure of Pulse Propagation of Gaussian pulse for chirp -1

Fig5.14: Pulse Propagation for chirp -1

5.5.2 Analysis for Gaussian pulse of Chirp = -1

Here, both GVD and SPM act simultaneously on the gaussian pulse with initial negative chirp.
The evolution pattern shows that pulse broadens at first for a small period of length. But
gradually the rate at which it broadens slowly declines and the pulse broadening ratio seems to
reach a constant value. This means that the pulse moves at a slightly larger but constant width as
it propagates along the length of the fiber. Although the width of the pulse seems constant, it
does not completely resemble a hyperbolic secant pulse evolution. We compare the pulse
evolution of the gaussian pulse with no initial chirp in Figure 5.13 and the negative chirped

66 | P a g e
CHAPTER 5: ANALYSIS OF RESULTS

gaussian pulse evolution in Figure 5.14 to see the difference in shape and width of each of these
evolutions.

As we previously established GVD and SPM effects cancel each other out when the GVD
induced negative chirp equals the SPM induced positive chirp. But in this case the initial chirp
affects the way both GVD and SPM behave. The chirp parameter of value -1 adds to the negative
chirp of the GVD and deducts from the positive chirp of SPM causing the net value of chirp to
be negative. This means that GVD is dominant during the early stages of propagation causing
broadening of the pulse.

But as the propagation distance increases the effect of the initial chirp decreases while the
induced chirp effect of both GVD and SPM regains control. The difference between positive and
negative induced is lessened and just like in the case of gaussian pulse propagation without
initial chirp the GVD and SPM effects eventually cancels out each other to propagate at constant
width.

67 | P a g e
CHAPTER 5: ANALYSIS OF RESULTS

Figure of Pulse Propagation of Gaussian pulse for chirp -2

Fig5.15: Pulse Propagation for chirp -2

5.5.3 Analysis for Gaussian pulse of Chirp = -2

Here, both GVD and SPM act simultaneously on the gaussian pulse with initial negative chirp.
From Figure 5.15 we observe that the pulse width broadens as it propagates. From the pulse
broadening ratio curve of the Figure 5.12 we see that the pulse broadens from the very start and
continues to broaden. It is also obvious that the rate at which the pulse broadens is higher than in
previous cases.

As before we concern ourselves with anomalous dispersion region where  2 < 0.

68 | P a g e
CHAPTER 5: ANALYSIS OF RESULTS

Therefore for C= -2 we have  2C  0 which as explained previously means that the direction of
dispersion induced chirp is in the same direction to that of the initial chirp value.

The negative input chirp C adds to the negative induced chirp of the GVD. At the same time the
positive chirp caused by SPM is reduced due to deduction from the negative input chirp. Hence,
as the increasing GVD effect causes dispersion of the pulse at an enlarged rate, the curtailing of
the SPM effect results in the reduction of the rate at which the pulse narrows. The net effect is
that the rate of dispersion is much higher than the rate of pulse narrowing which explains why
the pulse broadening ratio increases at a much higher rate than previous cases.

Also, unlike in the case of chirp 0 and chirp -1, the impact of the initial chirp parameter is greater
if the magnitude of the chirp is more than a certain value. In such a case, the GVD and SPM do
not cancel each other further down the length of the propagating medium; instead the pulse
continues to broaden indefinitely. This value of the initial chirp where GVD effect grows and
eventually dominates pulse propagation is the CRITICAL chirp. Beyond such a value, the
transformation of the gaussian pulse into a constant width pulse is impossible.

69 | P a g e
CHAPTER 5: ANALYSIS OF RESULTS

Figure of Pulse Propagation of Gaussian pulse for chirp 2

Fig5.16: Pulse Propagation for chirp 2

5.5.4 Analysis for Gaussian pulse of Chirp = 2

Here, both GVD and SPM act simultaneously on the gaussian pulse with initial positive chirp.
The pulse evolution graph in Figure 5.16 describes a gaussian pulse in nonlinear media with an
initial chirp of positive 2. Here, the input pulse narrows substantially in the early goings of pulse
evolution, but only for a certain length of the fiber. Following its brief narrowing tendency
during which it seems to gain amplitude height, the pulse seems to spread as it propagates. The
pulse broadening ratio curve in Figure 5.12 confirms the aforementioned tendency and the fact
that the ratio rises sharply after its initial fall.

70 | P a g e
CHAPTER 5: ANALYSIS OF RESULTS

In this case since, C>0, so that  2C  0 . This means that the dispersion induced chirp is in
opposite direction to that of the initial chirp parameter C.

The positive input chirp deducts from the negative induced chirp and results in a decreased
dispersive effect. Simultaneously, the positive input chirp adds to the SPM induced positive
chirp. The net outcome is that the dispersive effect due GVD declines and the SPM effect rises,
explaining the initial narrowing of the pulse width.

However, as the distance of propagation increases the GVD effect gains on the SPM effect and
eventually becomes dominant. Eventually, the pulse broadening ratio begins increasing linearly.

The outcome is once again attributed to the initial chirp being greater than the critical chirp
permissible for gaussian pulses.

71 | P a g e
CHAPTER 5: ANALYSIS OF RESULTS

5.6 SUPER GAUSSIAN PULSE ANALYSIS

Input pulse of super Gaussian for m=1, 8, 40 and 100:

Fig5.17: Input pulse for different values of m.

72 | P a g e
CHAPTER 5: ANALYSIS OF RESULTS

Fig5.18: Supergaussian pulse propagation for m=4

Fig5.19 Supergaussian pulse propagation for m=10

73 | P a g e
CHAPTER 5: ANALYSIS OF RESULTS

Fig5.20: Supergaussian pulse propagation for m=50

Fig5.21: Supergaussian pulse propagation for m=100

74 | P a g e
CHAPTER 5: ANALYSIS OF RESULTS

Pulse Broadening Ratio:

Fig5.22: Pulse Broadening Ratio for different values of m for chirp 0

5.6.1 Analysis for Supergaussian pulse for chirp =0:

Supergaussian pulse broadening ratios for various powers of supergaussian pulses are studied. In
Figure: 5.22 we have considered zero initial chirp and zero nonlinearity to learn of the effect of
dispersion for higher power supergaussian pulses. From here we obtain pulse broadening ratio
curves for m = 4, 10, 25, 50 and 100. The behavior of the pulses is easily viewed. As we
continue increasing the power m of the supergaussian, the slopes of the straight lines of each of
the curves increase. This suggests that the steeper the edge of the supergaussian pulses higher are
the effects of GVD induced negative chirp. Hence, supergaussian pulses of higher power will
experience greater dispersive effects than that of lower power supergaussians.

75 | P a g e
CHAPTER 5: ANALYSIS OF RESULTS

Fig5.23: Pulse Broadening Ratio for different values of chirp for m=100

5.6.2 Analysis for Supergaussian pulse for chirp =100:

In Figure: 5.23 we consider supergaussian pulses of power m= 100 to study the effects under
various input chirp parameters. The nonlinearity is considered zero here.

In this case, chirp =0, 1, 2 and -2 are studied. The pulse broadening ratio curves reveal that as
the magnitude of chirp is increased from 0 to 1, the effect of dispersion seems to increase. But as
chirp is increased from 2 and then decreased to -2, the effect of dispersion seems to decrease and
appears to reach a constant level. This is significant because here keeping the magnitude of the
chirp relatively large for supergaussian pulses of high power can reduce dispersive effects.

76 | P a g e
CHAPTER 6: CONCLUSION

CHAPTER 6

CONCLUSION
In this dissertation we explored the combined effects of the GVD and SPM on various types of
pulses including hyperbolic secant pulses, gaussian pulses and supergaussian pulses. At first ,
we propagate a gaussian pulse and observe results for variable nonlinearity, variable group
velocity dispersion and variable input power in three separate studies. We find that for low
nonlinear parameter values the pulse regains initial shape for a given input power and  2 .But

for higher values pulse collapses. By altering  2 for same input power and nonlinear parameter,

we observe that the pulse broadening ratio increases for higher values of  2 . Increasing input

power while keeping  2 and nonlinear parameter constant , we find that nonlinear effects
increases. Decreasing input power results in decrease in nonlinear effect , so that GVD becomes
dominant and pulse disperses.

Hyperbolic pulses are propagated as a constant width pulse called soliton. The perfect
disharmonious interaction of the GVD and SPM induced chirps result in diminishing of both
dispersive and nonlinear narrowing effects and hence soliton is obtained. Gaussian pulses are
also propagated with or without pre-induced(initial) chirp to study the pattern of propagation. It
is found that in the case of chirp 0 and chirp -1 , the gaussian pulse acquires a hyperbolic secant
pulse shape and travels as a pseudo-soliton. However, higher values of initial chirp leads to
indefinite dispersion and pulse shape is not retained ; a fact that can be attributed to the critical
chirp, a chirp value beyond which no constant width pulse propagation is possible. For the super-
gaussian pulse propagation we first considered an unchirped input with zero nonlinearity
parameter to understand the effects of the power of the super-gaussian pulse m on the pulse
width and found that pulse broadening ratio curve becomes steeper for higher powered super-

Page | 77
CHAPTER 6: CONCLUSION

gaussians. We then applied initial chirp on the super-gaussian pulses and found that for values of
chirp 2 and -2 or higher , the high power super-gaussian pulse broadening steadies signifying a
decrease in dispersive effects. We also generated pulse broadening ratio curves and evolution
patterns for higher order solitons. Here , we find that as we increase the soliton order, for N=2
the pulse width periodically varies and regains the original pulse after soliton period z o .But
increasing the soliton order further results in initial periodic behaviour of pulse width before
settling at a much lower value indicating that pulse splitting has occurred.

Page | 78
BIBLOGRAPHY

Bibliography

1. Dutta, A. K. WDM Technologies - Active Optical Components, 2002

2. G. P. Agrawal, Fiber-Optic Communication Systems, Second Ed., John Wiley &

Sons, Inc. New York, 1997
3. G. P. Agrawal, Nonlinear Fiber Optics, Second Ed., Academic Press, San Diego,
1995
4. Agrawal, G. P. Applications of Nonlinear Fiber Optics, 2001.
5.Becker, P. C. Erbium-Dope Fiber Amplifiers - Fundamentals and Technology, 1997.
6. M.J. Ablowitz , H. Segur, Solitons and the inverse Scattering Transform, Siam,
Philadelphia,1998.
7. Oleg V. Sinkin, Ronald Holzlöhner, John Zweck,„„Optimization of the Split-Step Fourier
Method in Modeling Optical-Fiber Communications Systems‟‟, in Journal of lightwave
technology , January, 2003
8. Decusatis,, C. Handbook of Fiber Optic Data Communication ,2nd ed ,2002.
9. Vinay M K, „„Dispersion Managed Solitins‟‟. Department of Electrical Engineering
IIT Madras,May 2005
10. Kaminow, I. P. Optical Fiber Telecommunications III-A, 1997.
11. G. M. Gharib, „„soliton Solutions for the Unstable Nonlinear Schr¨odinger Equation which
Describe Surface of Constant Negative Curvature‟‟, in Int. Journal of Math. Analysis, Vol. 2,
2008,
12. Eran Bouchbinder, „„The Nonlinear Schr¨odinger Equation‟‟,2003.
13. Mihajlo Stefanović, Petar Spalevic, Dragoljub Martinovic, Mile Petrovic,„„Comparison of
Chirped Interference Influence on Propagation Gaussian and Super Gaussian Pulse along the
Optical Fiber‟‟, in journal of Optical Communications,2006

14. P. Lazaridis, G. Debarge, and P. Gallion, „„Optimum conditions for soliton launching from

chirped sech 2 pulses”, in optics letters, August 15, 1995

15. C. C. Mak, K. W. Chow, K. Nakkeeran, „„Soliton Pulse Propagation in Averaged Dispersion-
managed Optical Fiber System”, in Journal of the Physical Society of Japan, May, 2005,

Page | 79
BIBLOGRAPHY

16. E. Iannone, et al., Nonlinear Optical Communication Networks, John Wiley &
Sons, Inc., 1998.

17. A. Andalib, A. Rostami, N. Granpayeh, ‘‘Analytical investigation and evaluation of pulse

broadening factor propagating through nonlinear optical fibers‟‟, in Progress In
Electromagnetics Research, PIER, 2008
18. .Crisp, J. Introduction to Fiber Optics (2nd ed.), (2001).
19. “Soliton pulse propagation in optical fiber‟‟, class notes for WDM and optical Networks
course, MIT Lincoln Laboratory, December,2001.
20. Decusatis, C. Fiber Optic Data Communication - Technological Trends and Advances,
(200121. J.C. Malzahn Kampe, MATLAB Programming,2000.

Page | 80
GLOSSARY

MATLAB CODES:
GAUSSIAN CODE

clc; clear all; close all; clf;

loop=1;

i=sqrt(-1);

pi=3.1415926535;

Po=0.00064; %input pwr in watts

alpha=0; % Fiber loss value in dB/km

Ao=sqrt(Po); %Amplitude of pulse

gamma=0.003; %nonlinear parameter in /W/m

To=125e-12; %initial pulse width in seconds

beta2=-20e-27; %2nd order disp. (s2/m)

Lengthdisp=(To^2)/(abs(beta2)); %dispersion length in

meter

chirp=0; %Input chirp parameter

%----------------------------------------------------------

tau =- 4096e-12:1e-12: 4095e-12;% dt=t/to

delt=1e-12;

rel_error=1e-5;

h=1000;% step size

for Lengthprop=0.1:0.1:1.5

Page | 81
GLOSSARY

z=Lengthprop*Lengthdisp;

A=Ao*exp(-((1+i*(chirp))/2)*(tau/To).^2);

figure(1)

plot(abs(A));

title('Input Pulse');

xlabel('Time');

ylabel('Amplitude');

grid on;

hold on;

amp=max(size(A));

%---------------------------------------------------------------------
-----

input.fwhm=find(abs(A)>abs(max(A)/2));

input.fwhm=length(input.fwhm);

domega=1/(amp)/delt*2*pi;

omega=(-1*amp/2:1:amp/2-1)*domega;

A=fftshift(A);

omega=fftshift(omega);

fourier.spectrum=fft(fftshift(A)); %Pulse spectrum

for splitstep=h:h:z

Linsoln1=exp(-alpha*(h/2)+i*beta2/2*omega.^2*(h/2)) ;

fourier.spectrum=fourier.spectrum.*Linsoln1;

f=ifft(fourier.spectrum);

Nonlinsoln=exp(i*gamma*((abs(f)).^2)*(h));

f=f.*Nonlinsoln;

fourier.spectrum=fft(f);

Page | 82
GLOSSARY

Linsoln2=exp(-alpha*(h/2)+i*beta2/2*omega.^2*(h/2)) ;

fourier.spectrum=fourier.spectrum.*Linsoln2;

end

f=ifft(fourier.spectrum);

pulse.output(loop,:)=abs(f);%saving output pulse at all intervals

output.fwhm=find(abs(f)>abs(max(f)/2));

output.fwhm=length(output.fwhm);

ratio=output.fwhm/input.fwhm; %PBR at every value

pbratio(loop)=ratio;%saving PBR at every step size

loop=loop+1;

end

figure(2);

waterfall(pulse.output);

title('Pulse Evolution');

xlabel('Time');

ylabel('Distance');

zlabel('Amplitude');

figure(3)

Page | 83
GLOSSARY

plot(pbratio(1:1:loop-1),'b');

xlabel('Total number of steps:Each step is 1 SSFM cycle');

ylabel('Pulse broadening ratio');

grid on;

hold on;

SOLITON CODE:

clc; clear all; close all; clf;

loop=1;

i=sqrt(-1);

signbeta=-1;% sign of beta2

alpha=0; % Fiber loss

nonlinpar=0.003; %fiber non linearity in /W/m

N=1; %soliton order

To=125e-12; %initial pulse width in second

pi=3.1415926535;

Po=0.00064; %input pwr in watts

Ao=sqrt(Po); %Amplitude

Lengthdisp=(N^2)/(nonlinpar*Po);%dispersion length for

corresponding soliton order

beta2=-(To)^2/Lengthdisp; %2nd order disp. (s2/m)

chirp=0;

Nfactor=N.^2;

Page | 84
GLOSSARY

tau =- 4096e-12:1e-12: 4095e-12;% dt=t/to

delt=1e-12/To;

hnorm=1000;% step size

%-------------------------------------------------------------------------

%Defining interval for distance using SECH pulse

%-------------------------------------------------------------------------

for Lengthprop=0.1:0.1:1.0

z=Lengthprop*Lengthdisp;

%--------------------------------------------------------------------------

%FUNDAMENTAL SOLITON

U=N*sech(tau/To);

%--------------------------------------------------------------------------

%--------------------------------------------------------------------------

h=hnorm/Lengthdisp;

znorm=z/Lengthdisp;

amp=max(size(U));

%--------------------------------------------------------------------------

input.fwhm=find(abs(U)>abs(max(U)/2));

input.fwhm=length(input.fwhm);

fourier.spectrum=fft(fftshift(U)); %fourier domain pulse spectrum

domega=(1/amp)/delt*2*pi;

omega=(-1*amp/2:1:amp/2-1)*domega;

omega=fftshift(omega);

x=0;

Page | 85
GLOSSARY

%----------------------------------------------

%Split Step fourier method

%----------------------------------------------

for splitstep=h:h:znorm

Linsoln1=exp(-alpha*(h/2)+i*(signbeta)/2*omega.^2*(h/2));

fourier.spectrum=fourier.spectrum.*Linsoln1;

f=ifft(fourier.spectrum);

Nonlinsoln=exp(i*(Nfactor)*((abs(f)).^2)*(h));

f=f.*Nonlinsoln;

f=fftshift(f);

fourier.spectrum=fft(f);

Linsoln2=exp(-alpha*(h/2)+i*(signbeta)/2*omega.^2*(h/2)) ;

fourier.spectrum=fourier.spectrum.*Linsoln2;

x=x+1;

end

f=ifft(fourier.spectrum);

f=fftshift(f);

pulse.output(loop,:)=abs(f);%saving output pulse at all intervals

output.fwhm=find(abs(f)>abs(max(f)/2));

output.fwhm=length(output.fwhm);

ratio=output.fwhm/input.fwhm; %PBR at every value

Page | 86
GLOSSARY

pbratio(loop)=ratio;%saving PBR at every step size

loop=loop+1;

end

figure(1)

plot(abs(U));

xlabel('Time');ylabel('Amplitude');

grid on;

hold on;

figure(2);

waterfall(pulse.output);

xlabel('Time');

ylabel('Distance');

zlabel('Amplitude');

grid on;

hold on;

figure(3)

plot(pbratio(1:1:loop-1),'k');

xlabel('Total number of steps:Each step is 1 SSFM cycle');

ylabel('Pulse broadening ratio');

grid on;

hold on;

Page | 87
GLOSSARY

SUPERGAUSSIAN CODE:

clc; clear all; close all; clf;

loop=1;

i=sqrt(-1);

pi=3.1415926535;

Po=0.00064; %input pwr in watts

alpha=0; % Fiber loss value in dB/km

Ao=sqrt(Po); %Amplitude of pulse

gamma=0.003; %nonlinear parameter in /W/m

To=125e-12; %initial pulse width in seconds

beta2=-20e-27; %2nd order disp. (s2/m)

Lengthdisp=(To^2)/(abs(beta2)); %dispersion length in meter

chirp=0; %Input chirp parameter

m=4; % Power of the super gaussian pulse

%----------------------------------------------------------

tau =- 4096e-12:1e-12: 4095e-12;% dt=t/to

delt=1e-12;

rel_error=1e-5;

Page | 88
GLOSSARY

h=1000;% step size

for Lengthprop=0.1:0.1:1.5

z=Lengthprop*Lengthdisp;

A=Ao*exp(-((1+i*(chirp))/2)*(tau/To).^2*m);

figure(1)

plot(abs(A));

title('Input Pulse');

xlabel('Time');

ylabel('Amplitude');

grid on;

hold on;

amp=max(size(A));

%--------------------------------------------------------------------------

input.fwhm=find(abs(A)>abs(max(A)/2));

input.fwhm=length(input.fwhm);

domega=1/(amp)/delt*2*pi;

omega=(-1*amp/2:1:amp/2-1)*domega;

A=fftshift(A);

omega=fftshift(omega);

fourier.spectrum=fft(fftshift(A)); %Pulse spectrum

Page | 89
GLOSSARY

for splitstep=h:h:z

Linsoln1=exp(-alpha*(h/2)+i*beta2/2*omega.^2*(h/2)) ;

fourier.spectrum=fourier.spectrum.*Linsoln1;

f=ifft(fourier.spectrum);

Nonlinsoln=exp(i*gamma*((abs(f)).^2)*(h));

f=f.*Nonlinsoln;

fourier.spectrum=fft(f);

Linsoln2=exp(-alpha*(h/2)+i*beta2/2*omega.^2*(h/2)) ;

fourier.spectrum=fourier.spectrum.*Linsoln2;

end

f=ifft(fourier.spectrum);

pulse.output(loop,:)=abs(f);%saving output pulse at all intervals

output.fwhm=find(abs(f)>abs(max(f)/2));

output.fwhm=length(output.fwhm);

ratio=output.fwhm/input.fwhm; %PBR at every value

pbratio(loop)=ratio;%saving PBR at every step size

loop=loop+1;

end

Page | 90
GLOSSARY

figure(2);

waterfall(pulse.output);

title('Pulse Evolution');

xlabel('Time');

ylabel('Distance');

zlabel('Amplitude');

figure(3)

plot(pbratio(1:1:loop-1),'b');

xlabel('Total number of steps:Each step is 1 SSFM cycle');

ylabel('Pulse broadening ratio');

grid on;

hold on;

Page | 91
GLOSSARY

LIST OF SYMBOLS

A(z,T) Slowly varying pulse envelope

Ao sqrt ( Po )

Aeff Effective core area of fiber

2
A Energy of the input pulse

c Speed of light

C Initial chirp parameter

2c
D Dispersion parameter ; D =  2
2

h stepsize of each SSFM cycle

 T0 2 
LD Dispersion length, LD   
 
 2 

1
LN Nonlinear length, LN 
Po

N soliton order

 T0 2 
Nfactor Nfactor= N 2  P0 LD  P0  
 
 2 

n2 Nonlinear index co-efficient

Po Input power (W)

s sgn ( 2 )

Page | 92
GLOSSARY

T Time parameter for NLSE

T0 Initial pulse width

U(z,  ) Normalized slowly varying envelope

z propagation distance for simplified NLSE

z
Z propagation distance for normalized NLSE, Z=
LD

   T0 
2

z0 soliton period, z 0  LD  
2 2   2 


 Fiber attenuation (loss) in dB

2 Second order group velocity dispersion parameter

n 2 0
 Nonlinear parameter,  
cAeff

 (T) Induced chirp

T
 Time parameter for normalized NLSE,  =
T0

0 Center frequency

Page | 93