International Journal of Physics and Applications.

ISSN 0974-3103 Volume 3, Number 3 (2011), pp. 229-241

© International Research Publication House
http://www.irphouse.com

Theoretical Design for Double Chirped Mirrors in

Femtosecond Pulse Lasers

Elham Jasim Mohammad and Amal Mohammed Ali Al-Hillou

Al-Mustansiriyah University, Collage of Sciences, Physics Department

E-mail: [email protected], [email protected]

Abstract

This paper is concerned with theoretical study on optoelectronics physics to

design dielectric double-chirped mirror (DCM) to achieve high reflectivity and
dispersion compensation over a broad bandwidth. Analytic expressions for
reflectivity, group delay and group delay dispersion are used. Where, most
recent research deals with Nanotechnology ( 10 −9 seconds). Our research deals
with Femtotechnology ( 10 −15 seconds), which is very advanced in laser and
fiber optics technology.
Their theoretical analysis generally relies on the well-known scattering
matrix formalism derived from the Maxwell equations. Laser performance
strongly depends on the quality of optical coatings: reflectance of high
reflectors should approach the ideal 100% value at the operation wavelengths
in order to minimize laser intracavity losses and output coupling has to be set
to specific values to ensure optimal operation. The aim is to push the limits of
pulse duration to reach femtosecond scale.
The idea of chirped mirrors is based on the quarter-wave stack concept as
a building block of dielectric high reflector.The mirror consists of 49 layers:
35 layers from dielectric materials SiO2 / TiO 2 arranged, as periodic stacks have
been used to design chirped mirrors using the Fusedsilica as a substrate and 14
layers from tow dielectric materials SiO 2 / Ta 2 O5 have been used to design anti
reflection coating AR with controlled reflectivity and dispersion in the
wavelength range 650 − 900nm , it exhibits a reflectivity of >99.99%.

Index Terms: Dispersion, double chirped mirror, group delay, group delay
dispersion, reflectivity.

Introduction
A central building block for generating femtosecond light pulses are lasers. Within
230 Elham Jasim Mohammad and Amal Mohammed Ali Al-Hillou

only two decades of the invention of the laser the duration of the shortest pulse shrunk
by six orders of magnitude from the nanosecond regime to the femtosecond regime
[1]. In a mode-locked laser for femtosecond pulse generation, the dispersion
introduced by the gain medium and other optical components in the laser cavity is
often not desirable, because it tends to broaden the generated pulses. While the
naturally occurring dispersion is usually positive (i.e., normal dispersion), the desired
dispersion may either be close to zero or even anomalous. Such dispersion values can
be achieved by introducing optical components with anomalous dispersion. Such
components are usually either special dispersive dielectric mirrors (e.g. in the form of
monolithic Gires-Tournois interferometers or chirped mirrors CM), or prism pairs
[1,2]. Dispersion compensation essentially means canceling the chromatic dispersion
of some optical elements. However, the term is often used in a more general sense of
dispersion management, meaning the control the overall chromatic dispersion of some
system. The goal can be, e.g., to avoid excessive temporal broadening of ultrashort
pulses and/or the distortion of signals. Dispersion compensation is applied mainly in
mode-locked lasers and in telecommunication systems, but also sometimes in optical
fibers transporting light e.g. to or from some fiber-optic sensor. It is the control of the
overall chromatic dispersion of a system by adding optical elements with a suitable
amount of dispersion [1].
Dispersion compensation plays a key role in generation, amplification and
propagation of femtosecond pulses. In the dispersive medium, the pulse can be
broaden or compressed depending on the sign of chirp and dispersion. To obtain the
ultrashort pulses, the pulse group delay GD should have about frequency
independence after the dispersion compensation. Especially to compress a pulse to
near the transform limit one should not only compensate the GD but also eliminate the
high order dispersion term [2,3].
Double chirped mirrors DCM are useful and a compact device for ultrashort pulse
generation in sub 10fs regime. The simulation and design presented this mirror
offering high reflectivity and controlled group delay dispersion GDD [3]. The
proposal stacks design considered in this paper has alternating layers of
SiO2 (n = 1.4716) , TiO2 (n = 2.2505) and Ta 2 O5 (n = 2.0976 ) at 800nm. The substrate material
for the design stacks is FusedSilic a ( n = 1.4520 ) . The design operates with normal
incidence and s-mode of polarization and the behavior studied at design wavelength
λ0 = 800 nm .

Principle of Operation
The principle of operation can be understood as follows:
Each interface between the two materials contributes a Fresnel reflection [4]:
rF = (n h − nl ) /(n h + nl ) (1)

Where r F Fresnel reflectivity, n h high refractive index and nl low refractive

index.
For the design wavelength, the optical path length difference between reflections
Theoretical Design for Double Chirped Mirrors 231

from subsequent interfaces is half the wavelength; in addition, the reflection

coefficients for the interfaces have alternating signs. Therefore, all reflected
components from the interfaces interfere constructively, which results in a strong
reflection. The reflectivity achieved is determined by the number of layer pairs and by
the refractive index contrast between the layer materials. The reflection bandwidth is
determined mainly by the index contrast [4].
A bandwidth of a CM and its GDD spectral shape are defined by, among other
parameters, impedance mismatch between the ambient medium and mirror stack. The
impedance mismatch can be overcome by using glass as the medium of incidence.
Unfortunately, this solution limits the highest value of GDD, because the mirror stack
should compensate an additional thin non-parallel substrate or glass wedge.
Additionally, in this case, the aperture cannot be high and the beam stability problem
may arise. There are several other approaches to reducing GDD ripples: double-
chirped mirrors (Fig.1), Brewster-angled CMs and complementary CM pairs.
Alternatively, time-domain optimization deals directly with pulse compression and
can therefore lead to the shortest pulses even in spite of large GDD oscillations [5].

Figure 1: Standard Double dielectric chirped mirror [5,6].

r t

Figure 2: is a schematic illustrating reflectivity of the double-chirped mirror of Fig.1

[5].

In the schematic above, AR coating is represented as a two-port with two

incoming waves a1 ,b1 and two outgoing waves b2 , a 2 . Assuming the multi-layer AR
coating is loss-less (no absorption), the connection between the waves at the left port
232 Elham Jasim Mohammad and Amal Mohammed Ali Al-Hillou

and the right port is described by a transfer matrix [T]. Here, r and t are the
reflectivity and transmission, respectively, at port1 assuming reflection-free
termination of port2.The black rectangle represents the back mirror, which includes
all of double-chirped mirror except AR coating. That is, back mirror includes the SiO2
substrate and all the alternating high and low refractive index layers and including
double-chirped portion. Back mirror is assumed to be perfectly matched, and has full
reflection over the total bandwidth under consideration, such that its complex
reflectivity in the wavelength range of interest [5,8].
In this paper, we shall present the transfer matrix method allowing solving
Maxwell equations in multilayer dielectric structures. We shall consider an example
of a periodical structure and derive general equations in planar structures. In the
beginning, we consider propagation of light in the normal to layer planes direction.
We shall generalize the transfer matrix approach for TE and TM linear polarizations
of light. By definition, TE-polarized (also referred to as s-polarized) light has the
electric field vector parallel to the layer planes, TM-polarized light (also referred to as
p-polarized) has the magnetic field vector parallel to the planes (see Fig.3).

Figure 3: Orientation of electric and magnetic fields in TE- and TM-polarized

incident on a planar boundary [9].

What happens to the electromagnetic field at the planar interface between two
dielectric media with different refractive indices? The answer can be found by
resolving the system of Maxwell equations independently in the two media and then
matching of the solutions for electric and magnetic fields by the Maxwell boundary
conditions at the interface. These conditions require continuity of the tangential
components of both fields. They can be microscopically justified for any abrupt
interface in the absence of free charges and free currents.
Consider a transverse light-wave propagating along the z-direction in a medium
characterized by a refractive index n that is homogeneous in the xy plane but possibly
z-dependent. The wave equation in this case becomes [1,8]:
Theoretical Design for Double Chirped Mirrors 233

∂ 2E
= − k02 n2 E (2)
∂z 2

where k 0 is the wave-vector of light in a vacuum. The general form of the solution of
Eq. (1) writes:
E = A + exp(ikz ) + A − exp(− ikz ) (3)

where k = k 0 n , A + , A − are coefficients. Using the Maxwell equation one can easily
obtain the general form of the magnetic field amplitude
B = A + n exp(ikz ) − A − n exp(− ikz ) (4)

If we consider reflection of light incident from the left side to the boundary (z = 0 )
between two semi-infinite media characterized by refractive indices n1 (left) and n 2
(right), the matching of the tangential components of electric and magnetic fields
would give:
A1+ + A1− = A2+ (5)
(A1
+
)
− A1− n1 = A2+ n2 (6)

where A1+ , A1− and A2+ are the amplitudes of incident, reflected and transmitted light,
respectively. One can easily obtain the amplitude reflection coefficient
A1− n1 − n2
r≡ = (7)
A1+ n1 + n 2

and the amplitude transmission coefficient

A2+ 2n1
t≡ + = (8)
A1 n1 + n 2

The ratio of reflected to incident energy flux (reflectivity) is given by

R= r
2
(9)

and the ratio of transmitted to incident energy flux (transmittance) is

n2 2
T= t (10)
n1

n2
In the last formula, the factor comes from the ratio of light velocities in two
n1
234 Elham Jasim Mohammad and Amal Mohammed Ali Al-Hillou

media [9].
In multilayer structures, direct application of Maxwell boundary conditions at
each interface leads to the necessity to resolve a substantial number of algebraic
equations (two per interface). A convenient method allowing reducing the number of
equations to be resolved to a strict minimum (four in general case) is the transfer
matrix method, which we are going to describe briefly here.

Let us introduce the vector

⎡ E( z) ⎤
G ⎡ E ( z )⎤ ⎢
⎥ = ⎢− i ∂E (z ) ⎥
Φ( z ) = ⎢ ⎥ (11)
⎣ B ( z ) ⎦
⎣ k 0 ∂z ⎦

where E(z), B(z) are the amplitudes of the electric and magnetic field of any light
G
wave propagating in the z direction in the structure under study. Note that Φ( z ) is
continuous at any point in the structure due to the Maxwell’s boundary conditions. In
particular, it is continuous at all interfaces where n changes abruptly.
By our definition, the transfer matrix T̂a across the layer of width a is such a
2 × 2 matrix that:
G G
Ta Φ z = 0 = Φ z = a (12)

It is easy to verify by substitution into Eq.(11) of the electric and magnetic

amplitudes (2), (3) that if n is homogeneous across the layer,
⎡ i ⎤
cos ka sin ka ⎥
Ta = ⎢ n (13)
⎢in sin ka cos ka ⎥⎦
⎣

The transfer matrix across a structure composed of m layers can be found as

 i =1 
T = ∏ Ti (14)
i =m

Where Tˆi is the transfer matrix across i-th layer. The order of multiplication in Eq.
(13) is essential. The amplitude reflection and transmission coefficients (rs and ts) of a
structure containing m layers, and sandwiched between two semi-infinite media with
refractive indices nleft , nright before and after the structure, respectively, can be found
from the relation
⎡ 1 + rs ⎤ ⎡ ts ⎤
Tˆ ⎢ ⎥=⎢ ⎥. (15)
⎣nleft − nleft rs ⎦ ⎣nright ts ⎦

One can easily obtain

Theoretical Design for Double Chirped Mirrors 235

nright t11 + nleft nright t12 − t21 − nleft t22

rs = (16)
t21 − nleft t22 − nright t11 + nleft nright t12
t12t21 − t11t22
t s = 2nleft (17)
t21 − nleft t22 − nright t11 + nleft nright t12

The intensities of reflected and transmitted light normalized by the intensity of the
incident light are given by
nright
R = rs , T = t s
2 2
(18)
nleft

Respectively, In its turn, the transfer matrix across a layer can be expressed via
reflection and transmission coefficients of this layer. If the reflection and transmission
coefficients for light incident from the right-hand side and left-hand side of the layer
are the same, and nleft = nright ≡ n (the symmetric case realized, in particular, in a
quantum well embedded in a cavity), the Maxwell boundary conditions for light
incident from the left and right sides of the structure yield:
⎡ 1 + rs ⎤ ⎡ ts ⎤
Tˆ ⎢ ⎥ = ⎢ ⎥,
⎣n − nrs ⎦ ⎣nt s ⎦
(19)
⎡ t ⎤ ⎡ 1 + rs ⎤
Tˆ ⎢ s ⎥ = ⎢ ⎥.
⎣ − nt s ⎦ ⎣ − n + nrs ⎦

This allows the matrix T̂ to be expressed as:

⎡
1 ⎢ t s 2 − rs 2 + 1 −
(1 + rs )2 − ts 2 ⎤
T =
ˆ ⎥ (20)
n
2t ⎢
(
⎣⎢n (rs − 1) − t s
2 2
) ⎥
t s − rs + 1 ⎦⎥
2 2

For a quantum well, ts = 1 + rs , and Eq. (19) becomes

⎡ 1 0⎤
TˆQW = ⎢− 2n rs ⎥
1⎥ (21)
⎢
⎣ ts ⎦

In the oblique incidence case, in the TE-polarization, one can use the basis
⎡ Eτ ( z )⎤
⎢ B ( z ) ⎥ , where Eτ , Bτ are the tangential (in-plane) components of the electric and
⎣ τ ⎦
magnetic fields of the light wave. In this case, the transfer matrix (4) keeps its form
provided that the following substitutions are made:
236 Elham Jasim Mohammad and Amal Mohammed Ali Al-Hillou

k z = k cosϕ , n → n cos ϕ (22)

where ϕ is the propagation angle in the corresponding medium ( ϕ = 0 at normal

incidence).
⎡ Bτ ( z ) ⎤
In the TM-polarization, following Born and Wolf [1] we use the basis ⎢ ⎥
⎣ Eτ ( z )⎦
which still allows the transfer matrix (12) to be used provided that the substitutions
are done:
cos ϕ
k z = k cosϕ , n → (23)
n

Note that the transfer matrices across the interfaces are still identity matrices, and
Eq. (13) for the transfer matrix across the entire structure is valid.
In the formulas for reflection and transmission coefficients (15-18) one should
replace, in the TE-polarization
nleft → nleft cos ϕleft , nright → nright cos ϕ right (24)

And in the TM-polarization

cosϕleft cos ϕ right
nleft → , nright → (25)
nleft nright

where ϕ left , ϕ right are the propagation angles in the first and last media, respectively.
The same transformations would be applied to the transfer matrices (19), (20). Note
that any two propagation angles ϕ i , ϕ j in the layers with refractive indices ni , n j are
linked by the Snell-Descartes law:
ni sin ϕ i = n j sin ϕ j (26)

which is also valid in the case of complex refractive indices, when the propagation
angles formally become complex as well.
The group delay is defined as the negative of the derivative of the phase response
with respect to frequency [10՜14], GD, also known as "Envelope Delay" [3]. In
physics and in particular in optics, the study of waves and digital signal processing,
the term group delay has the following meaning:
The rate of change of the total phase shift with respect to angular frequency
[11,12]:
dφ
GD = − (27)
dω

Through a device or transmission medium, where φ is the total phase shift in

Theoretical Design for Double Chirped Mirrors 237

radians, and ω is the angular frequency in radians per unit time, equal to 2πf , where
f is the frequency (hertz if group delay is measured in seconds).
Group delay dispersion is a ubiquitous, and often irritating, phenomenon in
ultrafast laser labs. When ultrashort pulses propagate through dispersive media, their
frequency components emerge at different times due to GDD, causing the resulting
pulse to be chirped and stretched and reducing the pulse’s peak power [13].
d 2φ
GDD = − (28)
dω 2

This effect can be compensated by using a pulse compressor, which can introduce
negative GDD [20]. The standard method for computing the GDD is to compute
complex reflection coefficients using the transfer matrix technique and then take
successive finite difference over frequency [21,22]. Group-delay dispersion of optical
elements is a critical parameter for the generation and control of femtosecond laser
pulses. GDD can either increase or decrease then pulse duration by modulating the
spectral phase of the femtosecond laser pulses. The effect of GDD becomes more
significant as the laser pulse duration gets shorter. Ideally, a femtosecond dielectric
mirror should not only have high reflectance but also low dispersion over a
sufficiently broad spectral bandwidth [24].

Experimental Results and Discussion

In this simulation, we used a quarter waves double-chirped mirror. The structure of
double-chirped mirror design consists of 49 layers, with optical thicknesses=200. The
Bragg wavelength λ B = 800nm in the rang 650 − 900nm .
Figs.4 shows the relationship between the reflective index and the wavelength for
(the FusedSilica substrate, TiO2 material, SiO2 material and Ta 2 O5 material. This
determines the dispersion properties of the structure. Fig.5 The relationship between
the refractive index and the distance from substrate. Fig.6 The relationship between
the phase and the wavelength we see the properties of phase shift in the rang between
650-900nm for DGMs.
Results given in Fig.7 reveal almost high reflector at design wavelength (
λ B = 800 nm ) when the distribution double-chirped mirror consist of alternating
periodic quarter-wavelength stack of low and high refractive index. It exhibits a
reflectivity of >99.9995%.
Fig.8 shows the relationship between the reflectance and the thickness. Where the
reflectance increase with thickness increase.
Fig.9 shows the characteristics of GD as a function of wavelength, where the
group delay is nearly constant over a bandwidth of about 740-860nm. While Fig.10
shows the dependence GDD on the optical wavelength. It is clearly seen that the
decrease of the oscillation in GDD in the rang wavelength from 670-900nm. In the
case of high reflectors, a combination of materials with the highest refractive-index
ratios nh / nl is usually preferred since the higher the ratio, the higher the theoretical
238 Elham Jasim Mohammad and Amal Mohammed Ali Al-Hillou

reflectance and bandwidth of standard quarter wave stacks.

These stacks have a peak in the reflectivity at 800nm. The group delay and group
delay dispersion has low oscillation.

Figsure 4: The relationship between the reflective index and the wavelength for: a-
(Fusedsilica) substrate. b- TiO2 material. c- SiO2 material. d- Ta2O5 material.

Figure 5: The relationship between the refractive index and the distance from
substrate.
Theoretical Design for Double Chirped Mirrors 239

Figure 6: The relationship between the phase and the wavelength.

Figure 7: The relationship between the reflectance and the wavelength.

Figure 8: The relationship between the reflectance and the thickness.

Figure 9: The relationship between the group delay and the wavelength where the
GD has a minimum oscillations in the range from 650-900nm.
240 Elham Jasim Mohammad and Amal Mohammed Ali Al-Hillou

Figure 10: The relationship between the group delay dispersion and the wavelength.
It is clearly seen that the decrease of the oscillations in the GDD in the rang
wavelength from 650-900nm.

Conclusion
In view of the results presented in this study, the main contributions of this research
can be summarized below:
• Tow materials SiO2 / Ta 2 O5 have been used to design chirped mirrors with a flat
GDD over bandwidth of 650-900nm and the reflectivity was found to be greater
than 99.995%. Therefore, chirped mirror is a good device for generating pulse
duration down to 10 fs 2 .

References
[1] Article, "Wikipedia, the free encyclopedia: laser", (2010).
[2] Frank Trager, "Springer Handbook of Lasers and Optics", Boob, Pag. 583,
Germany, (2007).
[3] "Group-Delay-Dispersive Multilayer-Mirror Structures and Method for
Designing Same", (2001).
[4] V. Pervak, A.V. Tikhonravov, M. K. Trubetskov, S. Naumov, F. Krausz, A.
Apolonski, "1.5-octave chirped mirror for pulse compression down to sub-3
fs", Appl. Phys. B 87, Pab. 5–12, (2007).
[5] Franz X. Kaertner, Erich P. Ippen, Uwe Morgner and G. Fujimoto, "Double-
Chirped Mirror Systems and Methods", United States Patent, US 6, 590,925
B1, (2003).
[6] N. Matuschek, L. Gallmann, D.H. Sutter, G. Steinmeyer and U. Keller, "Back-
side-coated chirpedmirrors with ultra-smooth broadband dispersion
characteristics", Appl. Phys. B 71, Pag. 509–522, (2000).
[7] Itzhak Shapir and Senior Member, "Suggestion for a new Formula to Calculate
Group-Delay from Frequency Domain Measurements".
[8] J. F. Woodley and M. Mojahedi, "Negative Group Velocity and Group Delay in
Left-Handed Media", American Physical Society, 70 046603-1, 2004.
[9] Applied Radio Labs, "Group Delay Explanations and Applications",
www.radiolab.com.au , (1999).
Theoretical Design for Double Chirped Mirrors 241

[10] Selcuk Akturk, Xun Gu, Mark Kimmel and Trebion, " Extremely simple
single-prism ultrashort-pulse compressor ", Optics Express, Vol. 14, No. 21,
Pag. 10101, (2006).
[11] Jonathan R. Birge, Christian Jirauschek and Franz X. Kärtner,"Efficient
Analytic Computation of Group Delay Dispersion from Optical Interference
Coatings", Optical Society of America, (2004).
[12] Elisabet Xifré Pérez, "Design, Fabriction and Characterization of Porous
Silicon Multilayer Optical Devices", University Rovira I Virgili, ISBN: 978-
84-691-0362-3/DL: Pag. 86, (2007).
[13] Alexander Petrov, "Slow light photonic crystal line-defect waveguides",
Technischen Universität Hamburg-Harburg zur Erlangung des akademischen
Grades Doktor der Naturwissenschaften genehmigte Dissertation, (2008).
[14] Tayyab Imran, Kyung-Han Hong, Tae Jun Yu and Chang Hee Nam, "
Measurement of the group-delay dispersion of femtosecond optics using white-
light interferometry", American Institute of Physics, REVIEW OF
SCIENTIFIC INSTRUMENTS VOLUME 75, Pag. 2266-2270, (2004).