Porpagation of Femtosecond Laser Pulses in

Litharge Index SF57: Time-Frequency

Mounir Khelladi α, Touaoula Hachem σ & Bouacha Abdelhafid ρ

Abstract- An efficient numerical algorithm is presented for the studied analytically, though only a few isolated attempts
numerical modeling of the propagation of ultrashort pulses have been made on numerical simulation. Most studies
with arbitrary temporal and frequency characteristics through are based analytical calculations assuming a plane
linear homogeneous dielectrics. The consequences of proper wave or TEM00 Gaussian transverse profile and a

2015
sampling of the spectral phase in pulse propagation and its
influence on the efficiency of computation are discussed in
Gaussian temporal profile for the pulse. The change in
the spectral properties of the pulse on propagation of

Year
detail. The numerical simulation presented here is capable of
analyzing the pulse in the temporal-frequency domain. As an the pulse was investigated analytically by Sheppard and
example, pulse propagation effects such as temporal and Gan (Sheppard, 1999) taking special forms of Gaussian
spectral shifts, pulse broadening effects, asymmetry and pulsed beams. Agrawal considered spatial broadening 1
chirping in dispersive media are demonstrated for wavelet of diffracted pulses assuming Gaussian transverse and

X Issue I Version I
decomposition. temporal profile. However, analytical methods have the
Keywords: ultrashort laser pulse, index refraction, limitations of not being able to handle arbitrary pulse

ew
femtosecond, propagation, time-frequency decom- profiles. Also, closed form solutions are often obtained
position, characterizations, FROG, SPIDER, VAMPIRE. after certain levels of approximations. This has motive a
few studies based on the use of numerical simulation
I. Introduction techniques in the analysis of pulse propagation. For

R
example, Kaplan introduced numerical evaluation by

A ) Volume XV
ecent developments in short pulse laser
Vi
technology have led to significant advances in fast Fourier transform to analyze pulses of arbitrary
our fundamental understanding of ultrafast temporal profile and investigated on-axis temporal
phenomena in physics, chemistry and biology, as well evolution of the pulse in the far field. In view of the recent
as stimulating the development of applications in fields advance in ultra short pulse propagation, a strong need
as diverse as optical communications, biomedical is felt for developing a numerical formalism capable of
ly

performing such a complete analysis of the issues

)
imaging and fem to second micromachining (Florenta,
involved in pulse propagation.

Global Journal of Science Frontier Research

2006).
Ultrafast laser technology has advanced over Here we introduce a numerical simulation tool
r

recent years to the point that pulses with high peak for propagation of ultra short pulses of arbitrary shape
through linear homogeneous media based on wave
Ea

power and a duration of less than 10 fs can be

generated directly from mode-locked Ti:sapphire lasers. optical field representation which enables an easy
Propagation of ultra short optical pulses in a linear evaluation for the merit functions of the pulsed field. This
optical medium consisting of free space (Chritov, 1985) allows us to analyze the pulse in the time-frequency
(Ziolkowski, 1992), (Sheppard, 1997), (Agrawal, 1998), domain at any arbitrary plane. With this tool, we
(Kpalan, 1998), dispersive media (Porras, 1999), investigate the spectral and temporal evolution of ultra
diffractive optical elements (Ichikawa, 1999),(Piestun, et. short pulses at any arbitrary propagation distance. The
Al 2001), focusing elements (Matei, 1999) and apertures propagation of the pulse is achieved in terms of its
(Liu, 2004), (Hwang, 2003) has been extensively studied spectral equivalent. Further, we introduce certain
analytically, though only a few isolated attempts have sampling rules for the spectral phase so that the phase
been made on numerical simulation. However, analytical information is sampled properly when we move from
methods have the limitations of not being able to handle one spectral component to another in the spectral
arbitrary pulse profiles. equivalent of the pulse. As a consequence, the
algorithm becomes computationally efficient since we
Propagation of ultra short optical pulses in a
only considered a small number of spectral components
linear optical medium consisting of free space (Chiritov
for simulation of pulse propagation .
1985, Agrawal 1998), dispersive media (Kaplan 1998,
Agrawal 1999), diffractive optical elements (Porras 1999, II. Ultarshort Laser Pulses Generations
Chikawa 1999), focusing elements (Piestun 2001,
Kempe 1999) and apertures has been extensively The central aim of this section is to give a
concise introduction to nonlinear optics and to provide
Author α σ ρ: University of Tlemcen, Algeria. basic information about the most-widely used tunable
e-mail: [email protected] femtosecond laser sources, in particular tunable

© 2015 Global Journals Inc. (US)

 Porpagation of Femtosecond Laser Pulses in Litharge Index SF57: Time-Frequency

Ti:sapphire oscillators and Ti:sapphire amplifiers or beam with a Gaussian profile propagating through a
optical parametric amplifiers. nonlinear medium, e.g. a Ti:sapphire crystal, which is
pumped by a cw radiation. As aforementioned, for high
a) The Ti:Sapphire Oscillator
intensity light, due to the intensity dependent refractive
In 1982, the first Ti:sapphire laser was built by
index, the Kerr lens effect occurs. When the laser
(Moulton,1982). The laser tunes from 680 nm to 1130
operates in its most usual regime (free-running laser ), it
nm, which is the widest tuning range of any laser of its
can oscillate simultaneously over all the resonance
class1. Nowadays Ti:sapphire lasers usually deliver
frequencies of the cavity. These frequencies make up
several watts of average output power and produce
the set of longitudinal modes of the laser. For the
pulses as short as 6.5 fs.
stronger focused frequencies, the Kerr lens favors a
higher amplification. Thus the self-focusing of the seed
beam can be used to suppress the cw operation,
because the losses of the cw radiation are higher.
2015

Forcing all the modes to have equal phase (mode-

Year

locking) implies that all the waves of different

frequencies will interfere (add) constructively at one
2 point, resulting in a very intense, short light pulse. The
pulsed operation is then favored and it is said that the
Figure 1. A Ti: sapphire oscillator. The compensation of laser is mode-locked. Thus the mode-locking occurs
X Issue I Version I

dispersion speed of group is ensured by both prisms, due to the Kerr lens effect induced in the nonlinear

ew
producing a negative dispersion speed of group. The medium by the beam itself and the phenomenon is
blocking of mode is ensured by Kerr effect. On the one known as Kerr-lens mode-locking.
hand, in the temporal field, the non-linear phase makes
it possible to compensate for the negative quadratic
component resulting from the dispersion of the cavity in
A ) Volume XV

linear mode, which makes it possible to have perfectly

Vi
equidistant longitudinal modes. In addition, in the space
field, self focusing by a Kerr lens makes it possible to
decrease the losses when the various longitudinal
modes are in phase, thus privileging the pulse mode.
ly

i. The Kerr Lens Mode-Locking

)

For simplicity, consider the spatial propagation

Global Journal of Science Frontier Research

of a Gaussian laser beam in a nonlinear material. The

Figure 2 : The Kerr lens effect and self-focusing. The
r

intensity profile of the beam is a function of its radius r

and of a shape parameter g [12]. At high intensities, the index of refraction varies with intensity along the beam
Ea

refractive index depends nonlinearly on the propagating diameter. Depending on the sign of the nonlinear term
field. The lowest order of this dependence can be 𝑛𝑛2 in the expression (1) the index of refraction increases
written as follows: or decreases towards the center of the laser beam. For
positive 𝑛𝑛2 the laser beam self-focuses.
1
𝑛𝑛(𝑟𝑟) = 𝑛𝑛0 2 𝑛𝑛2 𝐼𝐼(𝑟𝑟) (1)

where 𝑛𝑛2 is the nonlinear index coefficient and

describes the strength of the coupling between the
electric field and the refractive index 𝑛𝑛. The intensity is:
2
𝐼𝐼 (𝑟𝑟) = 𝑒𝑒 −𝑔𝑔𝑟𝑟 (2)
Hence, the refractive index changes with
intensity along the optical path and it is larger in the
center than at the side of the nonlinear crystal. This
leads to the beam self-focusing phenomenon, which is
known as the Kerr lens effect (see Fig.2). This process is
enhanced along the optical path because focusing the
beam increases the focal "power" of this "lens". The
increase of the focusing stops when the diameter of the
beam is small enough and the linear diffraction is large
enough to balance the Kerr effect. Consider now a seed

© 2015 Global Journals Inc. (US)

 Porpagation of Femtosecond Laser Pulses in Litharge Index SF57: Time-Frequency

adjusted to match the reciprocal cavity round trip time

𝜏𝜏𝑟𝑟𝑟𝑟𝑟𝑟 by an algorithm which records the instantaneous
repetition rate of the laser. This method, also known as
regenerative mode-locking, is used to initialize mode-
locking in the Ti:sapphire oscillator (Spectra Physics
Tsunami), where the AOM works at 80 MHz.
The passive mode-locking technique does not
require an external clock in the laser resonator. Here the
laser radiation itself generates a modulation through the
action of a nonlinear device (e.g. moving a prism into or
out of the cavity) situated in the resonator. The
modulation is automatically synchronized to the cavity

2015
round trip frequency. The Ti:sapphire oscillator
(Kapteyn-Murnane) uses the passive mode-locking

Year
technique.
Modelocking, as the word suggests, involves
Figure 3 : The Kerr Lens Mode-Locking (KLM) principle. locking a large number of longitudinal modes of a laser 3
The axial modes of a laser cavity are separated by the in phase. In a modelocked laser, the electric fields

X Issue I Version I
intermode frequency spacing 𝜈𝜈 = 𝑐𝑐/2𝐿𝐿. (a) The net gain associated with the different modes add constructively
curve (gain minus losses). In this example, from all the at one point and destructively elsewhere to create a

ew
longitudinal modes in the resonator (b), only six (c) are short spike number of frequency components with same
forced to have an equal phase. phase gives an infinite series of short pulses in time
(Rulliere, 1998). In a laser cavity, these modes are
The modes are separated in frequency by
equally spaced (with a spacing depending on the cavity
𝜈𝜈 = 𝑐𝑐/2𝐿𝐿, L being the resonator length, which also length). The electric field distribution with N such modes

A ) Volume XV
gives the repetition rate of the mode-locked lasers:
Vi
in phase (considered to be zero, for convenience) can
1 𝑐𝑐 be written as:
𝜏𝜏𝑟𝑟𝑟𝑟𝑟𝑟 = = (3)
𝑇𝑇 2𝐿𝐿
𝑒𝑒 𝑖𝑖𝑖𝑖 ∆𝑤𝑤𝑤𝑤 −𝑒𝑒 𝑖𝑖𝑤𝑤 0 𝑡𝑡
Moreover the ratio of the resonator length to the 𝐸𝐸(𝑡𝑡) = ∑𝑛𝑛𝑁𝑁−1 𝐸𝐸𝑛𝑛 𝑒𝑒 𝑖𝑖𝑤𝑤 0 +𝑛𝑛∆𝑤𝑤𝑤𝑤 ∝ (4)
𝑒𝑒 𝑖𝑖∆𝑤𝑤𝑤𝑤 −1
pulse duration is a measure of the number of modes
Where 𝑤𝑤0 is the central frequency and ∆w is the
ly

oscillating in phase. For example if 𝐿𝐿 = 1 𝑚𝑚 and the

)
emerging pulses have 100 𝑓𝑓𝑓𝑓 time duration, there are mode spacing, this appears as a carried wave with

Global Journal of Science Frontier Research

105 modes contributing to the pulse bandwidth. There frequency domain. The laser intensity is given by
r

are two ways of mode-locking a femtosecond laser: sin 2 [(2n+1)∆w .t/2]

passive mode-locking and active mode-locking. I(t) ∝ [A(t)]2 = (5)
sin 2 (∆w .t/2)
Ea

Briefly, active mode-locking implies that the

This is series of pulses with width inversely
radiation in the laser cavity is modulated by a signal
proportional to the number of modes that are locked in
coming from an external clock source (e.g. acousto-
phase of the mode spacing. The concept of mode-
optical modulator (AOM), electro-optical modulator).
locking is easier said than done.
The modulation frequency of the AOM is continuously

Figure 4 : The influence of the phase relation between oscillating modes on the output intensity of the oscillation. (a)
two modes in phase, and (b) five modes in phase

© 2015 Global Journals Inc. (US)

 Porpagation of Femtosecond Laser Pulses in Litharge Index SF57: Time-Frequency

Figure 4. shows how the time distribution of a repetition rate of about 108 1/s in an oscillator. To
laser output depends upon the phase relations between amplify these ultrashort pulses without any damage to
the modes. Fig.4.a is the resultant intensity of two the amplifier, they first are stretched. The stretched
modes in phase Fig.4.b, is the resultant intensity of five pulses now have a duration of order the of a few
modes in phase and a period repetition of a wave hundred picoseconds. These chirped pulses can then
packet from the resultant constructive interference can be amplified in one or two amplification stages (Ruiz, et
be seen. In a general continuous wave multi-mode laser, al, 2005).
the modes will oscillate independently from each other In a CPA system, there are basically three
and will have random phases relative to one another. components: a stretcher, an amplifier and a
One of the many advantages of Ti:sapphire compressor; the first and latter components are
oscillator is its high repetition rate (70-100 MHz). This basically the same for all CPA systems. For the amplifier
allows a good duty cycle where the clusters in the stage there are two commonly used designs: multi-pass
molecular beam are irradiated several times by the and regenerative (Joanna, 2007).
2015

output laser pulses of a relatively low energy (several

b) Multipass and Regenerative Amplification
Year

nJ). Thus the laser operates in the weak-field regime.

Two of the most widely used techniques for
This avoids perturbation of potential energy surfaces
amplification of fem to second laser pulses are the
4 (PES) produced by the intense laser pulse, simplifying
multipass and the regenerative amplification. In the
the interpretation of the results and theoretical
multipass amplification different passes are
description. Moreover, non-resonant two-photon
X Issue I Version I

geometrically separated (see Fig.6.a). The number of

processes are not expected to occur.
passes (four to eight) are usually limited by the

ew
ii. CPA Laser System difficulties on focussing all the passes on a single spot
Chirped pulse amplification (CPA) is a of the crystal. The amplifier bandwidth has to be broad
technique to produce ultrashort strong pulses. Although enough to support the pulse spectrum. This is why
a CPA output can be very strong, it is poorly tunable in a Ti:sapphire crystals are widely used in the amplification
small range around its output central wavelength. process.
A ) Volume XV

Various optical techniques have been innovated to

Vi
The regenerative amplification technique implies
make the output pulses tunable and/or shorter while trapping of the pulse to be amplified in a laser cavity
keeping the pulse power as high as possible. CPA is the (see Fig 2.b). Here the number of passes is not
abbreviation of chirped pulse amplification. Chirped important. The pulse is kept in the resonator until all the
pulse amplification is a technique to produce a strong energy stored in the amplification crystal is extracted.
and at the same time ultrashort pulse. The concept
ly

Trapping and dumping the pulse in and out of the

)

behind CPA is a scheme to increase the energy of an resonator is done by using a Pockel cell (or pulse-
Global Journal of Science Frontier Research

ultrashort pulse while avoiding very high peak power in picker) and a broad-band polarizer. The Pockel cell
the amplification process.
r

consists of a birefringent crystal, that can change the

polarization of a travelling laser field by applying high
Ea

voltage on it. In the regenerative amplifier the Pockel cell

is initially working like a quarter-wave plate. When the
pulse is sent to the resonator, the voltage on the Pockel
cell is switched (delay one) and it becomes the
equivalent of a half-wave plate. In this way the pulse is
kept in the cavity until it reaches saturation. Then a
second voltage is applied (delay two) and the pulse is
extracted from the resonator (Holzwarth, 2000).

Figure 5 : Schematics of a chirped pulse amplification

system, showing the duration and energy level of the Figure 6 : (a) The amplifier configuration uses two
signal at the different stages of the system spherical mirrors in a multi-pass confocal Configuration
In the CPA technique, ultrashort pulses are to make the signal pass eight times through the
generated typically at low energy~ 10-9 J, with a amplifying medium, (b) Schematic principle of a
duration around 10-12 – 10-14 seconds and at a high regenerative amplifier

© 2015 Global Journals Inc. (US)

 Porpagation of Femtosecond Laser Pulses in Litharge Index SF57: Time-Frequency

Of all potential amplifier media, titanium- doped second grating is parallel to the incident ray to the first
sapphire has been the most wide spread used. It has grating.
several desirable characteristics which make it ideal as
a high power amplifier medium such as a very high
damage threshold (~8 -10J/cm2), a high saturation
fluence, and high thermal conductivity.
c) Stretcher-Compressor

2015
Figure 8 : Grating pairs used in the control of dispersion.

Year
Figure 7 : Principle of a stretcher (a) and a compressor r and b indicate the relative paths of arbitrary long- and
(b). The stretcher setup extends the temporal duration of short-wavelength rays. γ is the (Brewster) angle of
5
the fem to second pulse, whereas the gratings' incidence at the prism face. Light is reflected in the
arrangement in the compressor will compress the time plane (p1–p2) in order to remove the spatial dispersion

X Issue I Version I
duration of the pulse. Both setups are used in fem to shown
Travelling through the third and the forth

ew
second amplifiers
By using a dispersive line (combination of gratings recombines all the components and removes
gratings and/or lenses), the individual frequencies within the spatial chirp. It is possible to compact the design
a fem to second pulse can be separated (stretched) more by inserting a mirror in the plane of symmetry of
from each other in time (see Figure 7a). the two gratings pairs. The beam will be reflected back

A ) Volume XV
and pass through the second and then the first grating.
Vi
d) Compressors To compensate a certain amount of group velocity
In normal materials, low frequency components dispersion or third order dispersion of a pulse incoming
travel faster than high frequency components; in other to the compressor we need to know how much GVD
words, the velocities of large wavelength components and TOD are produced by the compressor (Holzwarth,
are higher than that of shorter ones. These materials 2000). The group delay induced by the grating
ly

induce a positive group velocity dispersion on a

)
compressor is given
propagating pulse. To compensate the positive GVD

Global Journal of Science Frontier Research

2L λ
and rephase the dephased components a setup which Tg = �1 + � − sinγ� sinγ� (6)
2 d
r

produces negative group velocity dispersion is needed. λ

c.�1−� −sin γ�
d
Some designs have been devised to produce negative
Ea

group velocity dispersion, like grating compressor, Where, is the light wavelength, L is the distance between
prism compressor, Gires-Tournois interferometer and the gratings, d is the grating's constant and 𝛾𝛾 is the
etc. They are called compressors, because they incidence angle of the beam to the first grating.
compress a pulse (in time domain) which is broadened Dispersion of the group delay is obtained as:
due to dispersion and chirp. The grating compressor 2 −3/2
d2∅ λ3 L λ
and the prism compressor are two conventional setups GDD = dw 2 = − πc 2 d 2 �1 − �d − sinγ� � (7)
(Holzwarth, 2000).
i. Grating Compressor The third order dispersion produced in the
Four identical gratings in a sequence as shown grating compressor will be:
in Fig.8 make up a grating compressor. A pulse λ 2
impinges on the first gratings with an angle of 𝜃𝜃 . 1 d3∅ d 2 ∅ 6πλ 1+ sin γ−sin γ
TOD = L dw 3 = − dw 2 c � d λ 2 � (8)
Different wave length components of the pulse spectrum 1−� −sin γ�
d
are diffracted off with different angles from the first
grating. Orientation of the grating is such that the longer ii. Prism Compressor
wavelength path to the second grating is longer than the A prism compressor is built of four sequentially
shorter wavelengths pathes. Then from the second arranged identical prisms used in a geometry similar to
grating the spectral components in the spectrum travel Fig.9 (left); often at their minimum deviation (to decrease
together in parallel directions but with wavelength geometrical (spatial) distortion of prisms on the beam)
dependent position (spatially chirped). The gratings are and in their Brewster angle (to minimize power loss).
set in such a way that their wavelength dispersions are Because of the symmetry in the arrangement, it is
reversed which implies that the exiting ray from the possible to place a mirror after the second prism (as we

© 2015 Global Journals Inc. (US)

 Porpagation of Femtosecond Laser Pulses in Litharge Index SF57: Time-Frequency

did in the grating compressor setup) perpendicular to 𝐸𝐸(𝑡𝑡). The electric field in the time-domain is invariably
the beam propagation direction. The first prism spreads connected with its counterpart 𝐸𝐸(𝑤𝑤) in the frequency-
the pulse spectral components out in space. In the domain via a Fourier transform:
second prism the red frequencies of the spectrum must i. Pulse Duration and Spectral Width
pass through a longer length in the glass than the blue The statistical definitions are usually used in
frequencies. theoretic calculations and given as
By moving the prism in and out of the beam +∞ +∞
path, the amount of the prism glass that the beam ∫−∞ 𝑡𝑡 2 |𝐸𝐸(𝑡𝑡)|2 𝑑𝑑𝑑𝑑 ∫−∞ 𝑤𝑤 2 |𝐸𝐸(𝑤𝑤 )|2 𝑑𝑑𝑑𝑑
〈𝑡𝑡 2 〉 = +∞ ; 〈𝑤𝑤 2 〉 = +∞ (11)
passes through can be changed for exact balancing of ∫−∞ |𝐸𝐸 (𝑡𝑡)|2 𝑑𝑑𝑑𝑑 ∫−∞ |𝐸𝐸(𝑤𝑤 )|2 𝑑𝑑𝑑𝑑
the pulse dispersion. The back reflected beam will pass
where t is the intensity weighted average time. Duration
through the prisms at a height slightly higher or lower
of very complex pulse cannot be defined in the direct
than the incoming beam. To calculate the group velocity
and simple way and statistical definitions is only
dispersion and the third order dispersion produced by
2015

approximation. In case of Gaussian pulse is easy to

the prism compressor we assume that the beam inside
determine pulse duration and spectral width by applying
Year

the compressor propagates near apices of the prisms.

FWHM of intensity. One can show that these quantities
are related through the following universal inequality
6
∆𝑤𝑤∆𝜏𝜏 ≥ 1/2 (12)
X Issue I Version I

For instance, a Gaussian-shaped pulse with

pulse duration of 100 fs has a minimum spectral

ew
bandwidth of 4.41 THz. Relation (12) can only be fulfilled
for the case when ∆𝑤𝑤∆𝜏𝜏 is directly equal 1/2 with
Gaussian temporal and spectral envelopes and the
Figure 9 : Pulse compressor, with a negative GDD. pulse is said to be a Fourier-transform limited or
Longer wavelengths traverse more glass. The use of two
A ) Volume XV

unchirped pulse. In this case, the instantaneous

prisms and a mirror simplifies the device somewhat, but
Vi
frequency is a time-independent quantity.
it remains difficult to vary its GDD over a wide range and In practice, half-maximum quantities are easier
to tune it to measure. Therefore, one defines the pulse duration
The optical path of a ray propagating in the ∆𝜏𝜏 as the Full Width at Half Maximum (FWHM) of the
compressor is defined as (Holzwarth, 2000): GDD will intensity profile and the spectral width ∆𝑤𝑤 as the FWHM
ly

of the spectral intensity. The Fourier inequality is then

)

result
usually given by
Global Journal of Science Frontier Research

4𝐿𝐿𝜆𝜆 3 𝑑𝑑 2 𝑛𝑛 𝑑𝑑𝑑𝑑 2 1 𝑑𝑑𝑑𝑑 2

𝐺𝐺𝐺𝐺𝐺𝐺 = �𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 � + � � �2𝑛𝑛 − �� − 2 cos 𝛽𝛽 � � � (9)
2𝜋𝜋 2 𝑐𝑐 2 𝑑𝑑𝜆𝜆 2 𝑑𝑑𝑑𝑑 𝑛𝑛 3 𝑑𝑑𝑑𝑑 ∆𝑤𝑤∆𝜏𝜏 ≥ 𝐾𝐾 (13)
r

The third order dispersion can also be where K is a numerical constant, depending on the
Ea

evaluated in the same manner to that used above assumed shape of the pulse.
𝐿𝐿4 𝑑𝑑 2 𝑝𝑝 𝑑𝑑 3 𝑝𝑝
Gaussian pulse
𝑇𝑇𝑇𝑇𝑇𝑇 = − �3 − 𝜆𝜆 𝑑𝑑 𝜆𝜆 3 � (10) In this subsection we would like to consider an
4𝜋𝜋 2 𝑐𝑐 3 𝐿𝐿 𝑑𝑑 𝜆𝜆 2
example of Gaussian pulse, which is most commonly
e) Mathematical Description of Laser Pulses used in ultrashort laser pulse characteristics. The pulse
In order to understand the behavior of ultrashort is linearly chirped an represented by
light pulses in the temporal and spectral domain, it is
−(1+𝑖𝑖𝑖𝑖 )𝑡𝑡 2
necessary to formulate the relation between the two 𝐴𝐴(𝑡𝑡) = 𝐴𝐴0 𝑒𝑒𝑒𝑒𝑒𝑒 � 𝜏𝜏 𝑔𝑔2
� with ∆𝜏𝜏𝑝𝑝 = �2𝑙𝑙𝑙𝑙2𝜏𝜏𝑔𝑔 (14)
domains mathematically.
They are linked through a Fourier transform, The instantaneous frequency is given as
and so the modification of a laser pulse in one of the two 𝑑𝑑𝑑𝑑 (𝑡𝑡) 2𝛼𝛼
domains either by propagation due to dispersion or 𝑤𝑤 (𝑡𝑡) = 𝑤𝑤0 + 𝑑𝑑𝑑𝑑
= 𝑤𝑤0 − 𝜏𝜏 2 𝑡𝑡 (15)
𝑔𝑔
actively by pulse shaping implies a modification of its
properties in the conjugate domain as well. It is Pulse is down-chirped for a positive chirp
important to introduce the concept of the amplitude and parameter α, for negative is upchirped and when α = 0
the phase of the electric field because the generation, then the pulse is unchirped. The spectral instensity can
measurement, and shaping of ultrashort laser pulses is be derived by taking the Fourier-transform of eq.14, it
based on measuring and influencing these properties. also has the Gaussian shape (RULLIERE, 1998)
The time behavior of a laser pulse at a fixed point in If follows directly from eq.13 that the minimum
space including the envelope, the oscillations and the achievable duration is limited by the spectrum of the
phase can be fully described by the real electric field pulse. In other words, in order to produce ultrashort

© 2015 Global Journals Inc. (US)

 Porpagation of Femtosecond Laser Pulses in Litharge Index SF57: Time-Frequency

pulses a very broad spectral bandwidth is needed. The They are characterized by several parameters: temporal
shortest possible pulse, for a given spectrum, is known coherence, spatial coherence (i.e. focusing ability)
as the transform-limited pulse duration. It should be contrast, power, etc. Here the description is
noted that eq.13 is not equality, i.e. the product can very concentrated on their temporal aspects.
well exceed 𝐾𝐾. If the product exceeds 𝐾𝐾 the pulse is no ii. Time Domain Description
longer transform-limited and all frequency components Since in this paper the main emphasis is on the
that constitute the pulse do not coincide in time, i.e. the temporal dependence, all spatial dependence is
pulse exhibits frequency modulation is very often neglected, i.e., 𝐸𝐸(𝑥𝑥, 𝑦𝑦, 𝑧𝑧, 𝑡𝑡) = 𝐸𝐸(𝑡𝑡). the electric field
referred to as a chirp. 𝐸𝐸(𝑡𝑡), is a real quantity and all measured quantities are
real. However, the mathematical description is simplified
if a complex representation is used:

𝐸𝐸� (𝑡𝑡) = 𝐴𝐴̃(𝑡𝑡). 𝑒𝑒 −𝑖𝑖𝑤𝑤 0 𝑡𝑡

2015
(16)

where 𝐴𝐴̃(𝑡𝑡) is the complex envelope, usually chosen

Year
Figure 10 : Self-phase modulation. Variation in the such that the real physical field is twice the real part of
instantaneous frequency 𝑤𝑤(𝑡𝑡) of the transmitted pulse the complex field, and 𝑤𝑤0 is the carrier frequency, 7
after the propagation through a nonlinear medium with a usually chosen to the center of the spectrum. In this way
positive nonlinear index of refraction, n2 the rapidly varying is separated from the slowly varying

X Issue I Version I
Ultrashort laser pulses are coherent bursts of enveloppe �𝐴𝐴(𝑡𝑡). 𝐸𝐸� (𝑡𝑡) can be further decomposed into:

ew
electromagnetic radiation, confined in time and space.

�𝐸𝐸 (𝑡𝑡) = �𝐸𝐸� (𝑡𝑡)�. 𝑒𝑒 𝑖𝑖𝜑𝜑 0 . 𝑒𝑒 −𝑖𝑖𝑖𝑖 (𝑡𝑡) = �𝐸𝐸�(𝑡𝑡)�. 𝑒𝑒 𝑖𝑖𝜑𝜑 0 . 𝑒𝑒 −𝑖𝑖(∅(𝑡𝑡)−𝑤𝑤 0 ) (17)

𝜑𝜑(𝑡𝑡) is often to as the temporal phase of the pulse and frequency, wt, is omitted. which means that a nonlinear

A ) Volume XV
𝜑𝜑0 the absolute phase, which relates the position of the temporal phase yields a time-dependent frequency
Vi
carrier wave to the temporal envelope of the pulse (see modulation- the pulse is said to carry a chirp (illustrated
fig.11). In ∅(𝑡𝑡) the strong linear term due to the carrier in Fig.11b)
ly

) Global Journal of Science Frontier Research

r
Ea

Figure 11 : (a) The electric field of an ultra-short lasers pulse, (b) The electric field of an ultrashort lasers pulse with a
strong positive chirp
An ultrashort pulse of light will lengthen after it Phase and chrisp
has passed through glass as the index of refraction, Instantaneous phase function of 𝐸𝐸(𝑡𝑡) can be
which dictates the speed of light in the material, described as the sum of temporal phase and product of
depends nonlinearly on the wavelength of the light. The carrier frequency with time by the relation
wavelength of an ultrashort pulse of light is formed from 𝑑𝑑
the distribution of wavelengths either side of the center 𝑤𝑤(𝑡𝑡) = ∅(𝑡𝑡) + 𝑤𝑤0 𝑡𝑡
wavelength with the width of this distribution inversely 𝑑𝑑𝑑𝑑
proportional to the pulse duration. Note the frequency Carrier frequency 𝑤𝑤0 has been choosen by
variation as a function of time; at the leading edge (to minimizing of temporal variation of phase ∅(𝑡𝑡). The first
the left) the wavelength is longer than at the trailing edge deriviation of 𝑤𝑤(𝑡𝑡) is defined by temporally-dependent
consisting of only a few cycles. carrier frequency as the result of applying the derivation

© 2015 Global Journals Inc. (US)

 Porpagation of Femtosecond Laser Pulses in Litharge Index SF57: Time-Frequency

we receive relation expended in series. Then carrier does not change the shape of the pulse, but only
frequency time denotes quadratic chirp. introduces a temporal shift of the entire pulse. Therefore,
usually only the nonlinear part of the spectral phase is of
interest. Any nonlinear addition to the phase will
redistribute the frequency components and alter the
temporal shape of the pulse.
f) Ultrashort Pulse Measurement Techniques
Measuring of the pulse width and structure of
the complex laser pulses in range under 10 ps is not as
Chirp can be generated due to dispersion of routine as the measurement of their repetition rate of
materials they propagate through. Up-chirp is the signal wavelength. All techniques for determining the temporal
with increasing frequency in time and down-chirp with resolution 𝐼𝐼(𝑡𝑡) can be in general classified into direct
2015

decreasing frequency in time. Positive chirp is when and indirect methods.

leading edge of pulse is red-shifted in relation to central The most commonly used technique of
Year

wavelength and trailing edge is blue-shifted. Negative ultrashort laser pulse width measurement is concerned
chirp happens in opposite case. Linear chirp, with the study of the temporal intensity profile
8 instantaneous frequency varies linearly with time. The 𝐼𝐼(𝑡𝑡) through its second-order correlation function that is
presence of chirp results in significant different delays obtained by the second-harmonic generation. It
X Issue I Version I

between the spectrally different components of laser provides an extremely high time resolution down to tens
pulse causing pulse broadening effect and leading to a of femtoseconds. However, being an indirect technique,

ew
duration-bandwidth. the intensity autocorrelation cannot provide unequivocal
Chirps always appear when ultrashort laser information on the pulse shape and, in particular, its
pulses propagate through a medium such as air or asymmetry. Strictly speaking, higher order correlation
glass, where the spectral components of the pulse are functions are necessary for the exact determination of
subject to a different refractive index. This effect is called the pulse shape and duration. On the other hand, when
A ) Volume XV

Group Velocity Dispersion (GVD). Therefore, for pulse

Vi
performed with interferometric accuracy, second - order
durations on the order of 100 fs, the contribution of the autocorrelations provide a much more contrasted
quadratic and higher order chirps is negligible. Yet, pattern and information about the pulse shape and chirp
shorter pulses require the consideration of higher order may be extracted. However all techniques employ
chirps due to the increasing frequency bandwidth. principle of probing ultrashort pulse with time delayed
ly

iii. Lens Frequency Domain Description replica of pulse itself. Indirect methods can be
)

It is usually more convenient to represent the distinguished into self – referenced and nonself-
Global Journal of Science Frontier Research

pulse in the frequency domain rather than in the time referenced, interferometric or non-interferometric.
r

domain. The frequency representation is obtained from According to pulse retrieval that can be direct (non-
the time domain by a complex Fourier transform, iterative) or indirect (iterative) and the data can be
Ea

1 +∞ captured and analyzed in the real time or not. Ultrashort

𝐸𝐸 (𝑤𝑤) = ∫−∞ 𝐸𝐸(𝑡𝑡). 𝑒𝑒 −𝑖𝑖𝑖𝑖𝑖𝑖 . 𝑑𝑑𝑑𝑑. (18) - pulse characterization techniques, such as the
√2π
numerous variants of frequency resolved optical gating
Just as in the time domain, 𝐸𝐸� (𝑤𝑤) can be written as: (FROG) and spectral phase interferometry for direct
𝐸𝐸� (𝑤𝑤) = �𝐸𝐸� (𝑤𝑤)�𝑒𝑒𝑖𝑖𝑖𝑖 (𝑤𝑤 ) (19) electric-field reconstruction (SPIDER), fail to fully
determine the relative phases of well separated
where 𝜑𝜑(𝑤𝑤) now denotes the spectral phase. An inverse frequency components. If well-separated frequency
transform leads back to the time domain, components are also characterized gate pulses are
1 +∞ used (Joanna, 2007).
𝐸𝐸� (𝑡𝑡) = ∫ 𝐸𝐸� (𝑡𝑡). 𝑒𝑒 𝑖𝑖𝑖𝑖𝑖𝑖 . 𝑑𝑑𝑑𝑑 (20)
√2𝜋𝜋 −∞ i. Non-Interferometric Techniques
From eq.20 it is clear that 𝐸𝐸� (𝑡𝑡) can be seen as a a. Intensity Autocorrelation
superposition of monochromatic waves. The spectral Complex electric field 𝐸𝐸(𝑡𝑡) corresponds to
phase can, in the same manner as the temporal phase, intensity 𝐼𝐼(𝑡𝑡) = |𝐸𝐸(𝑡𝑡)|2 and an intensity autocorrelation
be decomposed into different parts. A common function is defined by
procedure is to employ Taylor expansion
1 𝑑𝑑 𝑛𝑛 𝜑𝜑
∅(𝑤𝑤) = ∅0 + ∑∞
𝑛𝑛=1 𝑛𝑛!
𝑎𝑎𝑛𝑛 (𝑤𝑤 − 𝑤𝑤0 )𝑛𝑛 with 𝑎𝑎𝑛𝑛 = �
𝑑𝑑𝑤𝑤 𝑛𝑛 𝑤𝑤=𝑤𝑤 0
(21)

It can be seen by inserting this Taylor expansion

into eq.21 that the first, two terms will not change the
temporal profile of the pulse. A linear phase variation

© 2015 Global Journals Inc. (US)

 Porpagation of Femtosecond Laser Pulses in Litharge Index SF57: Time-Frequency

intensity autocorrelation. However, the phase cannot be

measured (Joanna, 2007), (abdolah, 2007).
From the construction of the setup point of view
as intensity autocorrelation as also interferrometric
autocorrelation consist of the some components, only
the geometry and processing the captured data is
different. In case of intensity autocorrelation is applied
non-collinear geometry and in case of interferrometric
setup the collinear geometry of beams incomming to the
nonlinear medium.
b. FROG
Figure 12 : Setup for an interferometric autocorrelator, The technique of Frequency-Resolved Optical

2015
similar to the field autocorrelator above, with the Gating (FROG) has been introduced by Trebino and
following optics added: L: converging lens, SHG: coworkers. In FROG technique signal 𝐸𝐸1 has been

Year
second-harmonic generation crystal, F: spectral filter to temporally shifted about τ through time-delay element in
block the fundamental wavelength respect with signal 𝐸𝐸2 . Then, two signals have been in
+∞ nonlinear medium non-interferometrically overlapped. As 9
𝐴𝐴(𝜏𝜏) = ∫−∞ 𝐼𝐼 (𝑡𝑡)𝐼𝐼 (𝑡𝑡 + 𝜏𝜏)𝑑𝑑𝑑𝑑 (22) the result of SFG or DFG process (at the efficient phase

X Issue I Version I
matching conversion) one receive the FROG signal.
Two parallel beams with a variable delay are
generated, then focused into a second-harmonic-

ew
generation crystal to obtain a signal proportional to
𝐸𝐸(𝑡𝑡) + 𝐸𝐸(𝑡𝑡 + 𝜏𝜏). Only the beam propagating on the
optical axis, proportional to the cross product
𝐸𝐸(𝑡𝑡)𝐸𝐸(𝑡𝑡 − 𝜏𝜏)is retained. This signal is then recorded by

A ) Volume XV
a slow detector, which measures
Vi
+∞ +∞
𝐼𝐼(𝜏𝜏) = ∫−∞ |𝐸𝐸(𝑡𝑡)𝐸𝐸(𝑡𝑡 − 𝜏𝜏)|2 𝑑𝑑𝑑𝑑 = ∫−∞ 𝐼𝐼(𝑡𝑡)𝐼𝐼(𝑡𝑡 − 𝜏𝜏)𝑑𝑑𝑑𝑑 (23)
𝐼𝐼(𝜏𝜏)is exactly the intensity autocorrelation 𝐴𝐴(𝜏𝜏).
Both beams must be focused at the same point inside
the crystal as the delay is scanned in order for the
ly

Figure 13 : FROG reconstruction scheme. When both E1

)
second harmonic to be generated. It can be shown that

Global Journal of Science Frontier Research

the intensity autocorrelation width of a pulse is related to and E2 have been unknown, then we deal with Blind-
the intensity width. For a Gaussian time profile, the FROG problem. When 𝐸𝐸1 = 𝐸𝐸2 then we have to do with
r

autocorrelation width is longer than the width of the SHG-FROG problem

Ea

intensity, and it is 1.54 longer in the case of an In construction with process II also collinear
hyperbolic secant squared (sech2) pulse. This geometry is possible.
numerical factor, which depends on the shape of the +∞
pulse, is sometimes called the deconvolution factor. If 𝐸𝐸𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹 (𝛺𝛺, 𝜏𝜏) ∝ ∫−∞ 𝐸𝐸1 (𝑡𝑡 − 𝜏𝜏)𝐸𝐸2 (𝑡𝑡)exp (𝑖𝑖𝑖𝑖𝑖𝑖)𝑑𝑑𝑑𝑑 (24)
this factor is known, or assumed, the time duration
+∞
(intensity width) of a pulse can be measured using an ∝ ∫−∞ 𝐸𝐸�1 (𝑤𝑤)𝐸𝐸�2 (𝛺𝛺 − 𝑤𝑤)exp (𝑖𝑖𝑖𝑖𝑖𝑖)𝑑𝑑𝑑𝑑 (25)
+∞ +∞
∝ ∫−∞ 𝑑𝑑𝑑𝑑 ∫−∞ 𝐸𝐸�1 (𝑤𝑤)𝐸𝐸2 (𝑡𝑡)exp (−𝑖𝑖𝑖𝑖𝑖𝑖) exp (𝑖𝑖𝑖𝑖𝑖𝑖 + 𝑖𝑖𝑖𝑖𝑖𝑖)𝑑𝑑𝑑𝑑 (26)
The spectral intensity spectrogram, a visually intuitive transform containing
both time and frequency information. Using phase
𝐼𝐼𝐹𝐹𝐹𝐹𝐹𝐹𝐺𝐺 (𝛺𝛺, 𝜏𝜏) ∝ |𝐸𝐸𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹 (𝛺𝛺, 𝜏𝜏)|2 (27) retrieval concepts that the FROG trace yields the full
has been at the various τ by spectrometer measured intensity 𝐼𝐼(𝑡𝑡) and phase ∅(𝑡𝑡) of an arbitrary ultrashort
and 𝐼𝐼𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹 (𝛺𝛺, 𝜏𝜏) is called FROG-trace. Relations (25), pulse. As has been already mentioned, several schemes
(26) are only the mathematical representation of eq.24. and methods exist for frequency-resolved optical gating
The task of receiving unknown complex signals 𝐸𝐸1 and as a technique for the full characterization of ultrashort
𝐸𝐸2 from measured FROG trace is known as the FROG optical signals as complex electric fields. However, the
reconstruction problem uniqueness of the reconstructed fields has never been
FROG involves measuring the spectrum of the shown. There exist no proof of PCGP algorithm and
signal pulse as a function of the delay between two some papers showing a lot of trivial and non-trivial
input pulses. The resulting trace of intensity versus ambiguities, but going into the details it depends on the
frequency and delay is related to the pulse’s applied conditions (Greg, 1996),(Marco, 1996).

© 2015 Global Journals Inc. (US)

 Porpagation of Femtosecond Laser Pulses in Litharge Index SF57: Time-Frequency

ii. Interferometric Techniques SFG-Signal can be created

a. Interferometric Autocorrelation 𝐸𝐸𝑆𝑆𝑆𝑆𝑆𝑆 (𝑡𝑡) ∝ 𝐸𝐸𝑀𝑀 (𝑡𝑡){𝐸𝐸(𝑡𝑡 − 𝜏𝜏)𝑒𝑒𝑒𝑒𝑒𝑒[𝑖𝑖𝑤𝑤0 𝜏𝜏] + 𝐸𝐸(𝑡𝑡)} (32)
Setup for an interferometric autocorrelator is
similar to the field autocorrelator above, with the = 𝐸𝐸𝑀𝑀 (𝑡𝑡)𝐸𝐸(𝑡𝑡 − 𝜏𝜏)𝑒𝑒𝑒𝑒𝑒𝑒[𝑖𝑖𝑤𝑤0 𝜏𝜏] + 𝐸𝐸𝑀𝑀 (𝑡𝑡)𝐸𝐸(𝑡𝑡) (33)
following optics added: L: converging lens, SHG: As square law detectors are not sensitive to the
second harmonic generation crystal, F: spectral filter to phase, the measurement of the intensity (whether it is
block the fundamental wavelength. A nonlinear crystal spatial or spectral) is an easy task but the measurement
can be used to generate the second harmonic at the of the phase needs indirect solutions. Spectral shearing
output of a Michelson interferometer in a collinear is similar in concept to intensity autocorrelation except
geometry. In this case, the signal recorded by a slow that, instead of gating pulse with a time-shifted copy of
detector is itself, a pulse is interfered with frequency-shifted or
spectrally sheared copy of itself. SPIDER uses nonlinear
2015

mixing (usually SFG) as combination of filters to

generate signal that can be measured by slow detector.
Year

10
X Issue I Version I

Figure 14 : Setup for an interferometric autocorrelator,

similar to the field autocorrelator above, with the
ew
A ) Volume XV

following optics added: L: converging lens, SHG:

Vi
second-harmonic generation crystal, F: spectral filter to
block the fundamental wavelength Figure 15 : Main steps of SPIDER technique.
+∞
(𝜏𝜏) = ∫−∞ |𝐸𝐸(𝑡𝑡) + 𝐸𝐸(𝑡𝑡 − 𝜏𝜏 )|2 𝑑𝑑𝑑𝑑 (28)
ly

𝐼𝐼(𝜏𝜏)is called the interferometric autocorrelation.

)

It contains some information about the phase of the

Global Journal of Science Frontier Research

pulse: the fringes in the autocorrelation trace wash out

as the spectral phase becomes more complex
r

b. SPIDER
Ea

The Spectral Phase Interferometry for Direct

Electric-field Reconstruction technique (SPIDER) is
based on spectral interferometry and needs no Figure 16 : generation of two sheared replicas of the
components which has to be shifted over the input pulse by non-linear interaction with a chirped pulse
measurement process. From the signal 𝐸𝐸(𝑡𝑡) that should Sending through linear spectral phase
be characterized, the copy-signal is being generated by modulator (adds spectral shear to pulses in its arm of
beam splitter interferometer). Sending through linear temporal
modulator - adds temporal delay to the pulse in its arm,
𝐸𝐸(𝑡𝑡 − 𝜏𝜏)𝑒𝑒𝑒𝑒𝑒𝑒[𝑖𝑖𝑤𝑤0 𝜏𝜏] (29) then two pulses are recombined and send through a
The time between the signal and copy itself has spectrometer. Spectral interferometry allows us obtain
been established through fixed position at the optical difference between two spectral phases. To the spectral
delay-line. Then the copy of signal goes through phase interferometry spectrum one should apply fast Fourier
filter (dispersive medium, for instance SF10 glass), so transform and as the product one will achieve in form of
arises the signal one center peak and two sidebands lower peaks in the
time domain.
𝐸𝐸𝑀𝑀 (𝑡𝑡) = 𝐹𝐹 −1 {𝐸𝐸� (𝑤𝑤)𝑒𝑒𝑒𝑒𝑒𝑒[𝑖𝑖∅𝑀𝑀 (𝑤𝑤)]} (30) Centered peak contains only spectrum
Through the phase filter electric field 𝐸𝐸� (𝑤𝑤)get information. One filter out two peaks and from existing
additional spectral phase ∅𝑀𝑀 (𝑤𝑤), which corresponds to one can receive spectral phase difference by applied
temporal extension 𝐸𝐸(𝑡𝑡). From the signals inverse fast Fourier transform. To the main advantages
of the SPIDER method can be counted following
𝐸𝐸𝑀𝑀 (𝑡𝑡) 𝑎𝑎𝑎𝑎𝑎𝑎 𝐸𝐸 (𝑡𝑡 − 𝜏𝜏)𝑒𝑒𝑒𝑒𝑒𝑒[𝑖𝑖𝑤𝑤0 𝜏𝜏] + 𝐸𝐸(𝑡𝑡) (31) properties: pulse retrieval is direct (non-iterative),

© 2015 Global Journals Inc. (US)

 Porpagation of Femtosecond Laser Pulses in Litharge Index SF57: Time-Frequency

minimal data are required: only one spectrum yields • Wavelength range: depends on the given pulse, in
spectral phase. It naturally operates single-shot. the Vampire model a wide range from UV to IR is
Disadvantages: apparatus is very complicated (13 applicable;
alignment parameters). and unability to measure long • Input energy requirements: 10mW average power or
and complex pulses, the temporal delay τ should be 50 𝜇𝜇𝜇𝜇 single pulse (but depends on the two signals
chosen very accurate, also the zeros in complex pulses in general);
make a problems. In stretched pulse we have
• Easy alignment.
continuous waves approximately, so there are usually
weak pulses, sampling, aliasing effects in Fourier All existing techniques usually present the good
Transforms when would be too much fringes and if results by analysis simple pulse shapes. More complex
would be too not enough fringes then is not good amplitude profiles become problematic, especially when
resolved, so this issue needs a compromise (Hofmann, there are zeroes or well separated parts in temporal or
2005). spectral domain. In such experiments, iterative FROG

2015
algorithm often do not converge and give wrong output.
c. The VAMPIRE Technique
Therefore, in case of more complex signals full field

Year
Abbreviation VAMPIRE means: Very Advanced
reconstruction could not be guaranteed by conventional
Method of Phase and Intensity Retrieval of E-fields. The
FROG methods. VAMPIRE method is derived from
optical scheme of idea of VAMPIRE technique is shown 11
Blind-FROG and is based not on autocorrelation of the
in Fig.17 below. One can classified VAMPIRE as the
signal, but rather on its cross-correlation with suitable

X Issue I Version I
noninterferometric spectrographic technique. The
reference pulse. Signal and reference are also known as
conditioning filter is a temporal phase filter and a key
probe pulse and gate pulse. The VAMPIRE spectrogram

ew
element of VAMPIRE. Every spectrogram generated by
(strictly speaking, spectrochronogram) can be
VAMPIRE has special mathematical properties. For that
expressed by
reason a unique retrieval of two independent and
arbitrary ultrafast pulses is possible. Every spectrogram 𝐼𝐼(𝛺𝛺, 𝜏𝜏) ∝ �∫ 𝐸𝐸�𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔 (𝑤𝑤)𝐸𝐸�𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝 (𝛺𝛺 − 𝑤𝑤)𝑒𝑒𝑒𝑒𝑒𝑒(𝑖𝑖𝑖𝑖𝑖𝑖)𝑑𝑑𝑑𝑑�
2
(34)
generated by Vampire has special mathematical

A ) Volume XV
properties. For that reason a unique retrieval of two where 𝐸𝐸� is the complex spectral field, 𝛺𝛺 defines the
Vi
independent and arbitrary ultrahort pulses is possible specrtal axis, and 𝜏𝜏 is the delay. Vampire can guarantee
(Joanna, 2007). the uniqueness of full field reconstruction because
𝐸𝐸�𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔 can be chosen such that 𝐼𝐼(𝛺𝛺, 𝜏𝜏) does not suffer
from non-trivial ambiguities. Ideally, this would be
achieved by a well-separated double pulse structure
ly

)
with an asymmetry in duration, peak power, and chirp of

Global Journal of Science Frontier Research

the individual pulses.
A sufficient symmetry is required, but detailed
r

knowledge of the structure is definitely not. In VAMPIRE

Ea

spectrogram exists non-centrosymmetric geometry due

to existing of two spectrally dispersed signals from the
cross-correlations of the probe pulse with two different
components of the gate pulse. In comparison to
autocorrelation based techniques, this asymmetry in the
Figure 17 : General view on idea of VAMPIRE technique cross-correlation represents just that extra amount of
information that allows a unique relation between probe
Ultrashort pulses with well-separated multiple pulse and spectrogram.
frequency components cannot be retrieved by all
variants of self-referenced FROG or self-referenced III. Propagation of a Light Pulse in a
SPIDER. Ultrashort signals with well-seperated multiple Transparent Medium
frequency components are no problem for VAMPIRE.
Complex ultrashort signals cannot be retrieved by SHG- What happens to a short optical pulse
FROG, PG - FROG or Blind - FROG in general. propagating in a transparent medium? Because of its
Reconstruction scheme is presented below. The main wide spectral width and because of group velocity
features of VAMPIRE technique can be listed as follows dispersion in transparent media, it undergoes a phase
distortion inducing an increase of its duration (Khelladi,
• Unique retrieval of two independent and arbitrary
2005).
ultrafast pulses;
When an electromagnetic wave interacts with
• Very fast real-time retrieval algorithm without the bound electrons of a dielectric medium the medium
stagnation problems; response in general depends on the optical frequency
• Pulse width range: from 10fs to 1ps; of the wave. This property is referred to as chromatic

© 2015 Global Journals Inc. (US)

 Porpagation of Femtosecond Laser Pulses in Litharge Index SF57: Time-Frequency

dispersion. It manifests itself through frequency After the pulse has propagated a distance z, its
dependence of the refractive index of the medium. Its spectrum is modified to
origin is related to the characteristic resonance 𝑛𝑛 (𝑤𝑤 ).𝑤𝑤
frequencies at which the medium absorbs the 𝐸𝐸 (𝑤𝑤, 𝑧𝑧) = 𝐸𝐸(𝑤𝑤)𝑒𝑒𝑒𝑒𝑒𝑒[±𝑖𝑖𝑖𝑖(𝑤𝑤)𝑧𝑧], 𝑘𝑘 (𝑤𝑤) = 𝑐𝑐
(36)
electromagnetic radiation through the oscillation of the
bound electron. Far from resonances, the refractive where k(w) is now a frequency-dependent propagation
index of the medium is well approximated by the factor. In order to allow for a partial analytical calculation
Sellmeier equation. The frequency Fourier transform of a of the propagation effects, the propagation factor is
Gaussian pulse has already been given as rewritten using a Taylor expansion as a function of the
angular frequency, assuming that ∆𝑤𝑤 ≪ 𝑤𝑤0 (this
−(𝑤𝑤−𝑤𝑤 0 )2 condition is only weakly true for the shortest pulses).
𝐸𝐸(𝑤𝑤) = 𝑒𝑒𝑒𝑒𝑒𝑒 � � (35)
4.𝛤𝛤 Applying the Taylor expansion to eq.37, the pulse
An ultrashort Fourier limited pulse has a broad spectrum becomes.
2015

spectrum and no chirp; when it propagates a distance 1

𝑘𝑘(𝑤𝑤) = 𝑘𝑘(𝑤𝑤0 ) + 𝑘𝑘 ′ (𝑤𝑤 − 𝑤𝑤0 ) + 𝑘𝑘 ′′ (𝑤𝑤 − 𝑤𝑤0 )2 + (37)
through a transparent medium, the medium introduces
Year

2
a dispersion to the pulse inducing an increase in the 𝑑𝑑𝑑𝑑 (𝑤𝑤 ) 𝑑𝑑 2 𝑘𝑘(𝑤𝑤 )
pulse duration. To investigate and determine the where 𝑘𝑘 ′ = � � . and 𝑘𝑘 ′′ = � � ,
12 𝑑𝑑𝑑𝑑 𝑤𝑤 0 𝑑𝑑𝑤𝑤 2 𝑤𝑤 0
dispersion, we assume a Gaussian shape for the pulse.
The electric field of the pulse is given as Eq.35
X Issue I Version I

1 𝑖𝑖
𝐸𝐸(𝑤𝑤, 𝑧𝑧) = 𝑒𝑒𝑒𝑒𝑒𝑒 �−𝑖𝑖𝑖𝑖(𝑤𝑤0 )𝑧𝑧 − 𝑖𝑖𝑘𝑘 ′ 𝑧𝑧(𝑤𝑤 − 𝑤𝑤0 ) − �4𝛤𝛤 + 2 𝑘𝑘 ′′ � (𝑤𝑤 − 𝑤𝑤0 )2 � (38)

ew
The time evolution of the electric field in the This term is called the “Group Velocity
pulse is then derived from the calculation of the inverse Dispersion”. The temporal width of the pulse at point z:
Fourier transform of eq.39,
+∞
∆𝜏𝜏𝑧𝑧 = ∆𝜏𝜏0 �1 + 4. (𝛤𝛤. 𝑘𝑘′′𝑧𝑧)2 . (44)
A ) Volume XV

−𝑖𝑖𝑖𝑖𝑖𝑖
𝑒𝑒(𝑡𝑡, 𝑧𝑧) = ∫−∞ 𝐸𝐸(𝑤𝑤, 𝑧𝑧). 𝑒𝑒 𝑑𝑑𝑑𝑑 (39)
Vi
𝜆𝜆 3 𝑑𝑑 2 𝑛𝑛 2𝑙𝑙𝑙𝑙𝑙𝑙 2
with 𝑘𝑘 ′′ = 𝛤𝛤 = ,
so that 2.𝜋𝜋.𝑐𝑐 2 𝑑𝑑𝜆𝜆 2 ∆20

𝛤𝛤(𝑧𝑧) 𝑧𝑧 𝑧𝑧
2 a) Application in Litharge Index SF57 Medium
𝑒𝑒(𝑡𝑡, 𝑧𝑧) = � . 𝑒𝑒𝑒𝑒𝑒𝑒 �𝑖𝑖𝑤𝑤0 �𝑡𝑡 − �� × 𝑒𝑒𝑒𝑒𝑒𝑒 �−𝛤𝛤(𝑧𝑧) �𝑡𝑡 − � � (40)
𝜋𝜋 𝑉𝑉∅ (𝑤𝑤 0 ) 𝑉𝑉𝑔𝑔 (𝑤𝑤 0 ) In optical materials, the refractive index is
where frequency dependent. This dependence can be
ly
)

calculated for a given material using a Sellmeier

Global Journal of Science Frontier Research

𝑤𝑤 𝑑𝑑𝑑𝑑
𝑉𝑉∅ (𝑤𝑤0 ) = � �
𝑘𝑘 𝑤𝑤 0
, 𝑉𝑉𝑔𝑔 (𝑤𝑤0 ) = � �
𝑑𝑑𝑑𝑑 𝑤𝑤 0
, 1/(𝛤𝛤(𝑧𝑧) = 1/𝛤𝛤 + 2𝑖𝑖𝑖𝑖′′ 𝑧𝑧 (41) equation, typically of the form
𝐵𝐵𝑖𝑖 𝑤𝑤 𝑖𝑖2
r

In the first exponential term of eq.40, it can be 𝑛𝑛2 (𝑤𝑤) = 1 + ∑𝑚𝑚 (45)
𝑖𝑖=1 𝑤𝑤 𝑖𝑖2 −𝑤𝑤 2
observed that the phase of the central frequency 𝑤𝑤0 is
Ea

𝑧𝑧
delayed by an amount 𝑉𝑉 after propagation over a The glass containing lead-oxide (PbO) are
∅
called Litharge glasses. Litharge has a high refractive
distance z. Because the phase is not a measurable
index making it bright and a relatively soft surface
quantity, this effect has no observable consequence.
making it easy to decorate by grinding, cutting and
The phase velocity 𝑉𝑉∅ (𝑤𝑤0 ) measures the propagation
engraving. Its ability to refract light more efficiently than
speed of the plane wave components of the pulse in the
standard glass makes it an excellent material for prisms,
medium. These plane waves do not carry any
decorative objects, and artificial jewelry. It is also
information, because of their infinite duration.
favoured for electrical applications because of excellent
The second term in eq.40 shows that, after
electrical insulating characteristics, and the concave or
propagation over a distance z, the pulse keeps a
diverging half on an achromatic lens is usually made of
Gaussian envelope. This envelope is delayed by an
litharge glass.
amount 𝑧𝑧⁄𝑉𝑉𝑔𝑔 , 𝑉𝑉𝑔𝑔 being the group velocity. The second
Lead-silicate glasses are especially promising
term in eq.40 also shows that the pulse envelope is
materials for highly nonlinear fibers due to their suitable
distorted during its propagation because its form factor
combination of properties. Although their material
𝛤𝛤(𝑧𝑧), defined as
nonlinearities are lower than those of bismite, tellurite,
1/(𝛤𝛤(𝑧𝑧) = 1/𝛤𝛤 + 2𝑖𝑖𝑖𝑖′′ 𝑧𝑧 (42) and chalcogenide glasses, they provide higher thermal
and crystallization stability and less steep viscosity
Depends on the angular frequency 𝑤𝑤 through 𝑘𝑘 ′′ (𝑤𝑤), temperature curves, while exhibiting low softening
temperatures. The properties of PbO based glasses are
2
𝑘𝑘 ′′ = �𝑑𝑑 𝑘𝑘�𝑑𝑑𝑑𝑑 2 � = 𝑑𝑑𝑑𝑑 �𝑉𝑉 �
𝑑𝑑 1 summarized below.
(43)
𝑤𝑤 0 𝑔𝑔 𝑤𝑤
 High linear refractive index (1.75-2.4)
0

© 2015 Global Journals Inc. (US)

 Porpagation of Femtosecond Laser Pulses in Litharge Index SF57: Time-Frequency

 High nonlinear index coefficient (22 – 286.10-20 where 𝑤𝑤𝑖𝑖 is the frequency of resonance and 𝐵𝐵𝑖𝑖 is the
[m2/W]) amplitude of resonance. In the case of optical fibers, the
 Wide transmission range (0.4𝜇𝜇𝜇𝜇~3𝜇𝜇𝜇𝜇) parameters 𝑤𝑤𝑖𝑖 and 𝐵𝐵𝑖𝑖 are obtained experimentally by
fitting the measured dispersion curves to eq.45 with
 Transition temperature (~700 °𝐶𝐶) and melting
𝑚𝑚 = 3 and depend on the core constituents (Jong
temperature (~1100°𝐶𝐶)
Kook, 2005).
 Poor electrical conductivity
b) Parameter of Dispersion
 Highest atomic number of all stable elements
An ultrashort Fourier limited pulse has a broad
The index of litharge SF57 is given by the spectrum and no chirp; when it propagates a distance
following expression (45): through a transparent medium, the medium introduces
Table 1 : Parameters for litharge SF57 Glasses dispersion to the pulse inducing an increase in the pulse
duration. We consider dispersions of orders two. The

2015
𝐵𝐵𝑖𝑖 1.81651732 0.428893631 1.07186278 pulse broadens on propagation as a result of group
𝜆𝜆𝑖𝑖 (µ𝑚𝑚) 0.0143704198 0.0592801172 121.419942 velocity dispersion (GVD).

Year
13

X Issue I Version I
ew

A ) Volume XV
Vi
ly

)
Figure 18 : Temporal broadening of the transform- limited pulse for different values of the propagation distance z

Global Journal of Science Frontier Research

In summary, the propagation of a short optical To the first place we limited only to the order
r

pulse through transparent medium results in a delay of two of the Taylor expansion of the phase. It is noticed
Ea

the pulse, a duration broadening and a frequency chirp. that the analysis of Fourier remains valid only for
durations of pulse which are higher than ≈ 60 fs.
c) Group Velocity Dispersion
In addition media we consider all higher order
The Group Velocity Dispersion (GVD) is defined
dispersion, which completely describes the physical
as the propagation of different frequency components at
processes involved in ultrashort dispersive pulse
different speeds through a dispersive medium. This is
dynamics. The pulse broadens in time and becomes
due to the wavelength-dependent index of refraction of
asymmetric. In addition, the off axis pulse becomes
the dispersive material. GVD causes variation in the
wider than the pulse on axis (Khelladi, 2008).
temporal profile of the laser pulse, while the spectrum
remains unaltered.
𝑑𝑑𝑑𝑑 1 𝑑𝑑 2 𝜑𝜑 1 𝑑𝑑 𝑛𝑛 𝜑𝜑
𝜑𝜑 (𝑤𝑤) = 𝜑𝜑(𝑤𝑤0 ) + (𝑤𝑤 − 𝑤𝑤0 ) 𝑑𝑑𝑑𝑑 � + 2! (𝑤𝑤 − 𝑤𝑤0 )2 𝑑𝑑𝑑𝑑 2 � + ⋯ + 𝑛𝑛! (𝑤𝑤 − 𝑤𝑤0 )𝑛𝑛 𝑑𝑑𝑑𝑑 𝑛𝑛 � + 𝜃𝜃(𝑤𝑤).𝑤𝑤=𝑤𝑤 0 (46)
𝑤𝑤=𝑤𝑤 0 𝑤𝑤 =Ω

© 2015 Global Journals Inc. (US)

 Porpagation of Femtosecond Laser Pulses in Litharge Index SF57: Time-Frequency

2𝜋𝜋
⎧ 𝜑𝜑(𝜆𝜆) = 𝑛𝑛(𝜆𝜆)𝑧𝑧
𝜆𝜆
⎪ 𝑑𝑑𝑑𝑑
= −
𝜆𝜆 2
⎪ 𝑑𝑑𝑑𝑑 2𝜋𝜋𝜋𝜋
⎪ 𝑑𝑑𝑑𝑑 𝑧𝑧 𝑑𝑑𝑑𝑑
⎪ = − � − 𝑛𝑛�
𝑑𝑑𝑑𝑑 𝑐𝑐 𝑑𝑑𝑑𝑑
⎪ 𝑑𝑑 2 𝜑𝜑 𝜆𝜆 3 𝑑𝑑 2 𝑛𝑛
⎪ = + 𝑧𝑧 (47)
𝑑𝑑𝑤𝑤 2 4𝜋𝜋 3 𝑐𝑐 2 𝑑𝑑𝜆𝜆 2
𝑑𝑑 3 𝜑𝜑 𝜆𝜆 4 𝑑𝑑 2 𝑛𝑛 𝑑𝑑 3 𝑛𝑛
⎨ = − 4𝜋𝜋 2 𝑐𝑐 3 �3 2 + 𝜆𝜆 3 � 𝑧𝑧
𝑑𝑑 𝑤𝑤 3 𝑑𝑑𝜆𝜆 𝑑𝑑𝜆𝜆
⎪ 𝑑𝑑 4 𝜑𝜑
= +
𝜆𝜆 5
�12
𝑑𝑑 2 𝑛𝑛
+ 8𝜆𝜆
𝑑𝑑 3 𝑛𝑛 𝑑𝑑 4 𝑛𝑛
+ 𝜆𝜆2 4 � 𝑧𝑧
⎪ 𝑑𝑑𝑤𝑤 4 8𝜋𝜋 3 𝑐𝑐 4 𝑑𝑑𝜆𝜆 2 𝑑𝑑𝜆𝜆 3 𝑑𝑑𝜆𝜆
⎪ 𝑑𝑑 5 𝜑𝜑 𝜆𝜆 6 𝑑𝑑 2 𝑛𝑛 𝑑𝑑 3 𝑛𝑛 4
2 𝑑𝑑 𝑛𝑛 𝑑𝑑 5 𝑛𝑛
⎪ = − �60 + 60𝜆𝜆 + 15𝜆𝜆 + 𝜆𝜆3 5 � 𝑧𝑧
𝑑𝑑𝑤𝑤 5 16𝜋𝜋 4 𝑐𝑐 5 𝑑𝑑𝜆𝜆 2 𝑑𝑑𝜆𝜆 3 𝑑𝑑𝜆𝜆 4 𝑑𝑑 𝜆𝜆
⎪𝑑𝑑 6 𝜑𝜑 𝜆𝜆 7 𝑑𝑑 2 𝑛𝑛 𝑑𝑑 3 𝑛𝑛 4
2 𝑑𝑑 𝑛𝑛
5
3 𝑑𝑑 𝑛𝑛
6
4 𝑑𝑑 𝑛𝑛
2015

⎪𝑑𝑑𝑤𝑤 6 = + 32𝜋𝜋 5 𝑐𝑐 6 �360 𝑑𝑑𝜆𝜆2 + 480𝜆𝜆 𝑑𝑑𝜆𝜆3 + 180𝜆𝜆 𝑑𝑑𝜆𝜆4 + 24𝜆𝜆 𝑑𝑑𝜆𝜆5 + 𝜆𝜆 𝑑𝑑𝜆𝜆6 � 𝑧𝑧
⎩
Year

1 0 0 0 0 
14  
3 1 0 0 0 
∅(2) 12 8 1 0 0  (48)
X Issue I Version I

⎡ (3) ⎤ 
0

60 60 15 1
⎢∅ ⎥ 𝑛𝑛 
1
⎢∅(4) ⎥ = (−1)𝑛𝑛 2. 𝜋𝜋. 𝑧𝑧 � 𝜆𝜆 �

ew
360 480 180 24
⎢ (5) ⎥ 2.𝜋𝜋 .𝑐𝑐
⎢∅ ⎥
⎣∅(6) ⎦
d iϕ
A ) Volume XV

It seems to me that we can write ϕ (i ) = as a The various terms of the Taylor expansion to
Vi
dw i
order n can be written in the shape of a matrix [A],
recurrence, giving ϕ based on derivatives of order i,
(i )

which’s we can express various terms A ij.

the index of refraction. Matrix form, we can write.
𝑝𝑝 1
∅ (𝑤𝑤) = ∅(𝑤𝑤0 ) + (𝑤𝑤 − 𝑤𝑤0 )∅(1) + ∑𝑖𝑖=2 𝑖𝑖! (𝑤𝑤 − 𝑤𝑤0 )𝑖𝑖 ∅(𝑖𝑖) �𝑤𝑤=𝑤𝑤 + 𝜃𝜃 (𝑤𝑤). (49)
ly

0
)
Global Journal of Science Frontier Research

𝜆𝜆 𝑝𝑝
∅(𝑝𝑝) = (−1)𝑝𝑝 . 2𝜋𝜋. 𝑧𝑧 � � ∑𝑝𝑝𝑗𝑗=2 𝜆𝜆𝑗𝑗 −1 𝐴𝐴(𝑝𝑝 − 1, 𝑗𝑗 − 1)𝑛𝑛(𝑗𝑗 ) with 𝑝𝑝 > 2 (50)
2.𝜋𝜋.𝑐𝑐
r
Ea

Figure 19 : (a) the pulse broadens on propagation as a result of group velocity dispersion (GVD) (b) The pulse
shape is no longer Gaussian and it becomes asymmetric due to higher order dispersions
Analytically known and experimentally observed presented here. In addition, such studies can be
propagation affects such as spectral shift, pulse extended to pulses of arbitrary temporal shape without
broadening and asymmetry in dispersive media can be any further algorithmic complexity by numerical
easily brought out in the simulation using formalism simulation. Higher order dispersion effects can be

© 2015 Global Journals Inc. (US)

 Porpagation of Femtosecond Laser Pulses in Litharge Index SF57: Time-Frequency

handled easily in the numerical simulation unlike in the IV. Time-Frequency Decomposition
case of analytical calculation (Khelladi, 2008).
Discussion a) Wavelet Theory
The wavelets are very particular elementary
The Fourier theorem is the most classical
functions, these are the shortest vibrations and most
approach for describing the propagation of
elementary that one can consider. One can say that the
electromagnetic signals through dispersive media. In the
wavelet east carries out a zooming on any interesting
case of signals characterized by a slow temporal varying
phenomenon of the signal which place on a small scale
envelope, the phase is usually approximated by the
in the vicinity of the point considered (Meyer et al, 1987).
Taylor expansion in the neighborhood of the central
frequency of the input pulse. For shorter pulses, the b) Wavelet Techniques
concept of group velocity is irrelevant and the envelope Starting with a signal 𝑒𝑒(𝑡𝑡), in plane 𝑧𝑧 = 0, we
distortion is a function of the higher order terms. define wavelet centered at Ω by

2015
Ultrashort pulses less than 10 fs are now available. Their (𝑤𝑤 − Ω)2
envelope harmonic content is so high that the Taylor 𝜃𝜃(Ω) = 𝐸𝐸(𝑤𝑤). 𝑒𝑒𝑒𝑒𝑒𝑒 �− �, with 𝐸𝐸(𝑤𝑤) =
4𝛾𝛾

Year
expansion of the phase is now more possible. There is
no other way than a numerical computation of the 𝐸𝐸0 𝜋𝜋 (𝑤𝑤−𝑤𝑤 0 )2
� 𝑒𝑒𝑒𝑒𝑒𝑒 � � (51)
Fourier integral. 2.𝜋𝜋 𝛤𝛤 4.𝛤𝛤 15
However this method does not permit a

X Issue I Version I
straightforward physical understanding of the envelope
propagation and principally does not picture the fact

ew
that this is the group velocity dispersion which
generates the ultrashort pulses distortions. Such a
situation claims for another type of decomposition
involving both a time and frequency dependence of the
components.

A ) Volume XV
Numerous bidimensional representations of
Vi
acoustic and electromagnetic signals have already been
suggested. We propose here a method derived from the
Gabor transformation in order to decompose the signal
into an infinite number of elementary components
(wavelet) of same duration (much more longer than that
ly

)
of the original signal), each of them being centered at a Figure 20 : Gaussian envelope decomposed on a

Global Journal of Science Frontier Research

frequency Ω belonging to the Fourier spectrum of the number of wavelet we calculates the electric field
pulse (Khelladi, 2008). associated with the wavelet θ(Ω, z = 0).
r
Ea

𝜃𝜃(𝑡𝑡, 𝑧𝑧 = 0) = 𝑇𝑇𝑇𝑇{𝜃𝜃(Ω, 𝑧𝑧 = 0)} (52)

𝛾𝛾 −(𝑤𝑤 0 −Ω)2 𝛾𝛾𝛾𝛾 𝛾𝛾 𝑤𝑤 0 +𝛤𝛤Ω

𝜃𝜃(𝑡𝑡, 𝑧𝑧 = 0) = 𝐸𝐸0 �𝛾𝛾+𝛤𝛤 . 𝑒𝑒𝑒𝑒𝑒𝑒 � 4(𝛾𝛾+𝛤𝛤)
� . 𝑒𝑒𝑒𝑒𝑒𝑒 �−
𝛾𝛾 +𝛤𝛤
𝑡𝑡 2 � . 𝑒𝑒𝑒𝑒𝑒𝑒 �𝑗𝑗 𝛾𝛾 +𝛤𝛤
𝑡𝑡� (53)

 In time, the pulse is also Gaussian, of parameter 𝑛𝑛(𝑤𝑤). After propagation, the wavelet 𝜃𝜃(Ω, 𝑥𝑥) may be
𝜸𝜸𝜸𝜸 written as
𝜸𝜸+𝜞𝜞
.
𝐸𝐸0 (𝑤𝑤 −Ω)2
 The maximum of amplitude of the wavelet 𝜃𝜃(𝑡𝑡, 𝑧𝑧 = 𝜃𝜃(Ω, 𝑧𝑧) = 𝐸𝐸(𝑤𝑤). 𝑒𝑒𝑒𝑒𝑒𝑒 �− � . 𝑒𝑒𝑒𝑒𝑒𝑒[𝑗𝑗∅(𝑤𝑤)]. (54)
2.√𝜋𝜋𝛾𝛾 4𝛾𝛾
0) vary with Ω, center frequency of analysis on
Gaussian of parameter 𝛾𝛾 + 𝛤𝛤. As already mentioned, 𝜏𝜏𝑤𝑤𝑤𝑤𝑤𝑤𝑤𝑤𝑤𝑤𝑤𝑤𝑤𝑤 is large enough
to ensure that analyzing function has only non negligible
 The signal propagates in the positive 𝑧𝑧 direction in a values over a spectral range lying in the neighbourhood
linear dispersive and transparent medium, which fills of Ω in (Fig 20). Under these circumstances, we have
the half space 𝑧𝑧 > 0 and whose refractive index is
𝑑𝑑∅ 1 𝑑𝑑 2 ∅ 1 𝑑𝑑 𝑛𝑛 ∅
∅ (𝑤𝑤) = ∅(Ω) + (𝑤𝑤 − Ω) �+ (𝑤𝑤 − Ω)2 � + ⋯ + 𝑛𝑛! (𝑤𝑤 − Ω)𝑛𝑛 𝑑𝑑𝑑𝑑 𝑛𝑛 � + 𝜃𝜃(𝑤𝑤).𝑤𝑤 =Ω (55)
𝑑𝑑𝑑𝑑 2! 𝑑𝑑𝑑𝑑 2 𝑤𝑤=Ω 𝑤𝑤=Ω
Neglecting the higher terms in eq.50:
𝑑𝑑∅ 1 𝑑𝑑 2 ∅
∅(𝑤𝑤) = ∅(Ω) + (𝑤𝑤 − Ω) 𝑑𝑑𝑑𝑑 � + 2! (𝑤𝑤 − Ω)2 𝑑𝑑𝑑𝑑 2 � + 𝜃𝜃 (𝑤𝑤). (56)
𝑤𝑤 =Ω 𝑤𝑤 =Ω

© 2015 Global Journals Inc. (US)

 Porpagation of Femtosecond Laser Pulses in Litharge Index SF57: Time-Frequency

𝐸𝐸 𝜋𝜋 (𝑤𝑤 −𝑤𝑤 0 )2 (𝑤𝑤 −Ω)2 1

𝜃𝜃(Ω, 𝑧𝑧) = 2.√𝜋𝜋0 𝛾𝛾 �𝛤𝛤 𝑒𝑒𝑒𝑒𝑒𝑒 �− � . 𝑒𝑒𝑒𝑒𝑒𝑒 �− � . 𝑒𝑒𝑒𝑒𝑒𝑒 �𝑗𝑗∅(0) + 𝑗𝑗(𝑤𝑤 − Ω)∅(1) + 𝑗𝑗(𝑤𝑤 − Ω)2 . ∅(2) � (57)
4𝛤𝛤 4𝛾𝛾 2

1 +∞
𝜃𝜃(𝑡𝑡, 𝑧𝑧) = 2𝜋𝜋 ∫−∞ 𝜃𝜃(Ω, 𝑧𝑧). 𝑒𝑒𝑒𝑒𝑒𝑒(𝑗𝑗𝑗𝑗𝑗𝑗)𝑑𝑑𝑑𝑑 (58)

We calculates the temporal electric field associated with the wavelet 𝜃𝜃(Ω, 𝑧𝑧).
2
1 𝐸𝐸0 𝜋𝜋 �−(Ω−𝑤𝑤 0) � (0)
1
−� +
1 1 (2) 2
– 𝑗𝑗 ∅ �Ω �
(Ω−𝑤𝑤 0 )
– 𝑗𝑗 ∅(1)�Ω
𝜃𝜃 (𝑡𝑡, 𝑧𝑧) = � 𝑒𝑒 4𝛤𝛤 𝑒𝑒 �𝑗𝑗 ∅ � × 𝑒𝑒 4𝛤𝛤 4𝛾𝛾 2 . 𝑒𝑒 2𝛤𝛤
2𝜋𝜋 2. √𝜋𝜋𝛾𝛾 𝛤𝛤

1 1 1 1 1 1 (Ω−𝑤𝑤 0 )2
+∞ – 𝑗𝑗 ∅(2) �𝑤𝑤 2 − 𝑗𝑗 ∅(2) �2Ω𝑤𝑤 �− −𝑗𝑗 ∅(1) �
2015

−� + � +
× �∫−∞ 𝑒𝑒 4𝛤𝛤 4𝛾𝛾 2 . 𝑒𝑒 4𝛤𝛤 4𝛾𝛾 2 × 𝑒𝑒 2𝛤𝛤 . 𝑒𝑒 𝑗𝑗𝑗𝑗𝑗𝑗 � 𝑑𝑑𝑑𝑑 (59)
Year

The amplitude of the incident Ω wavelet is given from eq.60 by

16
2
𝐸𝐸0 𝛤𝛤(𝑧𝑧) 𝑧𝑧
𝜃𝜃(𝑡𝑡, 𝑧𝑧) = � . 𝑒𝑒𝑒𝑒𝑒𝑒�𝑗𝑗∅(0) �𝑒𝑒𝑒𝑒𝑒𝑒 �−𝛤𝛤(𝑧𝑧) �𝑡𝑡 + � �
X Issue I Version I

2. √𝜋𝜋𝛾𝛾 𝛤𝛤 𝑉𝑉𝑔𝑔 (Ω)

ew
(Ω−𝑤𝑤 0 )2 𝛤𝛤(𝑧𝑧) 𝛤𝛤(𝑧𝑧) 𝛤𝛤(𝑧𝑧) 𝑧𝑧
× 𝑒𝑒𝑒𝑒𝑒𝑒 �− �1 − �� . 𝑒𝑒𝑒𝑒𝑒𝑒 �𝑗𝑗 �1 − �Ω + 𝑤𝑤0 � �𝑡𝑡 + 𝑉𝑉 �. (60)
4𝛤𝛤 𝛤𝛤 𝛤𝛤 𝛤𝛤 𝑔𝑔 (Ω)

This wavelet is characterized by a Gaussian temporal width.

A ) Volume XV

Vi
envelope. This decomposition is valid only for the values The delay of group of the wavelet is inversely
of ∆𝑤𝑤 much larger than 𝛿𝛿𝛿𝛿 (∆𝑤𝑤 ≫ 𝛿𝛿𝛿𝛿 ). proportional to the velocity of group its envelope
𝑧𝑧 propagates without deformation (Khelladi, 2008).
The delay of group of the wavelet �𝑡𝑡 + 𝑉𝑉 � is
𝑔𝑔 (Ω)
characterized by a Gaussian envelope which is the
ly

c) Simulations
)
Global Journal of Science Frontier Research

r
Ea

© 2015 Global Journals Inc. (US)

 Porpagation of Femtosecond Laser Pulses in Litharge Index SF57: Time-Frequency

2015
Year
17
Figure 21 : (a) Initial pulse, (c) pulse after propagation of the 20 cm in litharge medium, (e) Contour of the wavelet,

X Issue I Version I
(g) the wavelet Representation, (b) Initial pulse, (d) pulse after propagation of the 20 cm in the silica medium, (f)
Contour of the wavelet, (h) the wavelet Representation (Khelladi et al, 2008)

ew
Parameters of the simulations Finally, we have demonstrated here the
Pulse initial: ∆𝜏𝜏0 = 10 𝑓𝑓𝑓𝑓 possible decomposition of an ultrashort pulse into an
Wavelength : 𝜆𝜆 = 800 𝑛𝑛𝑚𝑚 infinite number of longer Fourier transform limited
wavelets which propagate without any deformation

A ) Volume XV
Pulse of the wavelet: ∆𝜏𝜏𝑤𝑤𝑎𝑎𝑤𝑤𝑒𝑒𝑙𝑙𝑒𝑒𝑡𝑡 = 1000 𝑓𝑓𝑓𝑓 through a dispersive medium. After propagation through
Vi
Longer of the medium: 𝑧𝑧 = 20 𝑐𝑐𝑚𝑚 the medium, the pulse may be visualized in a three
To describe the propagation of the pulse, we dimensional representation by the locus of the wavelet
only consider the propagation of the maximum of each maxima. This representation permits the evaluation of
wavelet in a three dimensional representation: the broadening suffered by the pulse. For a transparent
medium, the propagation of the Ω wavelet is described
ly

Conclusion

)
V. by the convolution of the incident Ω wavelet with a 𝜃𝜃(Ω)

Global Journal of Science Frontier Research

In this article we have described powerful distribution centered at the group delay relative to Ω.
methods for crafting femtosecond optical pulses with The application to absorbing media is relatively
r

precisely controllable temporal envelopes. For optical straightforward and will be presented in a further
Ea

control the generation of light fields in a very flexible publication, as well as a generalization to nonlinear
manner is desired such that all the different open-loop media. The time-frequency representation is peculiarly
or closed-loop schemes can be implemented. In the suitable to the latter case for which the refractive index is
case of femtosecond laser pulses this means that one phenomenological time dependent. Although this
should be able to manipulate the phase, amplitude, and technique represents a vast improvement in our ability to
possibly also the polarization-state properties with a describe such pulses, they require additional effort, both
large number of degrees of freedom. Ultrashort laser in the apparatus and in the extraction of the pulse
pulses can be described in the time domain and in the intensity and phase from the experimental trace.
frequency domain which are connected via Fourier VI. Acknowledgments
transforms, thus shaping of an ultrashort laser pulse can
in principle be achieved either in the frequency or time This work has been financially supported by the
domain. Electronics, and Telecommunications laboratory of
Both phase and amplitude can be controlled, Tlemcen University, Algeria.
and great precision together with high complexity is References Références Referencias
possible. We began with a detailed discussion of
generation of the ultrashort laser pulse and shaping 1. Khelladi. M, Seddiki. O, and Bendimerad. T, Time-
methods which were largely developed by this author Frequency Decomposition of an Ultrashort Pulse:
and his coworkers and which are now becoming widely Wavelet Decomposition, Radioengineering ISSN,
adopted. We also reviewed a number of other pulse April (2008) Vol.17, pp.1210–2512, N1
characterizations techniques, involving as FROG, 2. Florenţa A.C, Dynamics of Ultra-short Laser Pulse
SPIDER, VAMPIRE. Interaction with Solids at the Origin of Nanoscale

© 2015 Global Journals Inc. (US)

 Porpagation of Femtosecond Laser Pulses in Litharge Index SF57: Time-Frequency

Surface Modification” geboren am 16.11.1971 in 19. Ruiz-de-la-Cruz.A and R. Rangel-Rojo, Multi-pass

Berca, Rumänien Cottbus (2006) confocal ultra-short pulse amplifier Revista
3. Chritov, I.P., Propagation of femtosecond light Mexicana de Fisica 51 (5) 488–493, M´exico (2005).
pulses, Opt. Commun. 53, 364- 366 20. Joanna.M, “Seutp of very advanced for phase and
4. Ziolokowski, R.W, Judkins, J.B. (1992)., Propagation amplitude reconstruction of electric field
characteristics of ultrawide bandwidth pulsed (VAMPIRE)” Master Thesis submitted to the Faculty
Gaussian beams, J. Opt. Soc. Am. A 11, 2021-2030, for the Natural Sciences and for Mathematics of the
(1985). Rostock University Germany (2007).
5. Sheppard, C.R, Xianosong, G, Free-space 21. Holzwarth, “Measuring the Frequency of Light using
propagation of femtosecond ligthpulses, Opt. Femtosecond Laser Pulses” aus Stuttgart den 21.
Commun. 133, 1-6. (1997) December (2000).
6. Agrawal, G.P, Spectrum-induced changes in 22. Hofmann.A, Bestimmung des elektrischen Feldes
2015

diffraction of pulsed beams, Opt. Commun. 157, ultrakurzer Laserpulse mit SPIDER, Teil1: Theorie,
52-56. (1998). Projektpraktikumsbericht, Physikalisches Institut,
Year

7. Kaplan, A.E., Diffraction-induced transformation of EP1, Universität Würzburg (2005).

near cycle and subcycle pulses, J. Opt. Soc. Am. B 23. Rulliere, C, Femtosecond laser pulses: principles
18 15, 951-956. (1998) and Experiments, Springer, (1998).
8. Agrawal, G.P, Far-Field diffraction of pulsed optical 24. Jong Kook K. Investigation of high nonlinearity glass
X Issue I Version I

beams in dispersive media, Opt. Commun. 167, 15- fibers for potential applications in ultrafast nonlinear
22. (1999). fiber devices, Dissertation submitted to the Faculty

ew
9. Porras,M.A, Propagation of single-cycle pulse light of the Virginia (2005).
beams in dispersive media, Phys. Rev. A 60, 5069- 25. Meyer,Y, Jaffort, S., Riol,O, wavelet analysis, edition
5073. (1998). Française de scientifique American (1987)
10. Ichikawa, H, Analysis of femtosecond-order optical 26. Khelladi,M. Ultra-short lasers pulses
pulses diffracted by periodic structure, J. Opt. Soc. Characterizations: Wavelet decomposition,
A ) Volume XV

Vi
Am. A 16, 299-304. (1999). University of Tlemcen. Algeria, 2005.
11. Piestun, R., Miller, D.B., Spatiotemporal control of 27. Moulton.P.f, Quant. Electron. Conf., Munich,
ultrashort optical pulses by refractive-diffractive- Germany, June (1982).
dispersive structured optical elements, Opt. let. 26, 28. Greg.T, Andy. R Margaret.M, Measurement of 10 fs
1373-1375. (2001). Laser Pulses. 1077-260X, IEEE (1996).
ly

12. Kempe, M., Stamm, U., Wilhelmi, B., and Rudolph, 29. Marco.A, Krumbugel.Kenneh.W,and Trebino. R,
)

W., Spatial and temporal transformation of frequency resolved optical gating measurements of
Global Journal of Science Frontier Research

femtosecond laser pulses by lenses systems, J. ultra short pulses using surface third harmonic
r

Opt. Soc. 38, 1058-1064. (1999). generation, Optics Letters, Vol21, N° 17, September
13. Matei, G. O, Gili, M.A., Spherical aberration in (1996).
Ea

spatial and temporal transformation of femtosecond 30. Chritov, I. P. Propagation of femtosecond light
laser pulses by lenses and lens systems, J. Opt. pulses. Opt. Commun., 1985, vol. 53, no. 6, pp.
Soc. Am. B9, 1158-1165. (1992). 364- 366.
14. Fuchss, U., Zeintner, U.D., and Tunnermenn (A.), 31. Agrawal, G. P. Spectrum-induced changes in
Ultrashort pulse propagation in complex optical diffraction of pulsed beams. Opt. Commun., 1998,
systems, Opt. Express 13, 3852-3861. (2005). vol. 157, pp. 52-56.
15. Leonid. S., Ferrari. A, and Bertolotti. M, Diffraction of 32. Kaplan, A. E. Diffraction-induced transformation of
a time Gaussian-shaped pulsed plane wave from a near cycle and subcycle pulses. J. Opt. Soc. Am. B,
slit, Pure. Appl. Opt5, 349-353. (1996). 1998, vol. 15, pp. 951-956.
16. Hwang. H.E, Yang, G. H., Far-Field diffraction 33. Agrawal, G. P. Far-field diffraction of pulsed optical
characteristics of a time variant Gaussian pulsed beams in dispersive media. Opt. Commun., 1999,
beam propagation from a circular aperture, Opt. vol. 167, pp. 15-22.
Eng. 41, 2719-2727. (2002). 34. Porras, M.A. Propagation of single-cycle pulse light
17. Liu,Z, Lu,B., Spectral shifts and spectral switches in beams in dispersive media. Phys. Rev. A, 1999, vol.
diffraction of ultrashort pulsed beams passing 60, no. 6, pp. 5069-5073.
through a circular aperture, Optik 115, 447-454, 35. Ichikawa, H. Analysis of femtosecond-order optical
(2004). pulses diffracted by periodic structure. J. Opt. Soc.
18. Abdolah.M.K, “Manipulation and characterization of Am. A, 1999, vol. 16, pp.299-304.
femtosecond laser pulses for cluster spectroscopy” 36. Piestun, R., Miller, D. B. Spatiotemporal control of
vorgelegt von aus Teheran (Iran) July (2007). ultrashort optical pulses by refractive-diffractive-

© 2015 Global Journals Inc. (US)

 Porpagation of Femtosecond Laser Pulses in Litharge Index SF57: Time-Frequency

dispersive structured optical elements. Opt. Let.,

2001, vol. 26, pp. 1373-1375.
37. Ssheppard, C. J. R., GAN , X. Free-space
propagation of femtosecond light pulses. Opt.
Commun., 1997, vol. 133, pp. 1-6.
38. Kempe, M., Stamm, U., Wilhelmi, B., Rudolph, W.
Spatial and temporal transformation of femtosecond
laser pulses by lenses systems. J. Opt. Soc., 1999,
vol. 38, pp. 1058-1064.

2015
Year
19

X Issue I Version I
ew

A ) Volume XV
Vi
ly

) Global Journal of Science Frontier Research

r
Ea

© 2015 Global Journals Inc. (US)

 Porpagation of Femtosecond Laser Pulses in Litharge Index SF57: Time-Frequency
2015
Year

20
X Issue I Version I

ew
This page is intentionally left blank
A ) Volume XV

Vi
ly
)
Global Journal of Science Frontier Research

r
Ea

© 2015 Global Journals Inc. (US)