photonics and nanotechnology, 2008, p.116

&
Photonics
Nanotechnology
This page intentionally left blankThis page intentionally left blank
N E W J E R S E Y
•
L O N D O N
•
S I N G A P O R E
•
B E I J I N G
•
S H A N G H A I
•
H O N G K O N G
•
TA I P E I
•
C H E N N A I
World Scientific
&
Photonics
Nanotechnology
Proceedings of the International Workshop and Conference on ICPN 2007
Pattaya, Thailand 16 – 18 December 2007
edited by
Preecha Yupapin
King Mongkut’s Institute of Technology Ladkrabang, Thailand
Prajak Saeung
Department of Physics,

Faculty of Science Ramkhamhaeng University, Thailand
British Library Cataloguing-in-Publication Data
A catalogue record for this book is available from the British Library.
For photocopying of material in this volume, please pay a copying fee through the Copyright Clearance Center, Inc., 222
Rosewood Drive, Danvers, MA 01923, USA. In this case permission to photocopy is not required from the publisher.
ISBN-13 978-981-277-971-7
ISBN-10 981-277-971-X
All rights reserved. This book, or parts thereof, may not be reproduced in any form or by any means, electronic or
mechanical, including photocopying, recording or any information storage and retrieval system now known or to be
invented, without written permission from the Publisher.
Copyright © 2008 by World Scientific Publishing Co. Pte. Ltd.
Published by
World Scientific Publishing Co. Pte. Ltd.
5 Toh Tuck Link, Singapore 596224
USA office: 27 Warren Street, Suite 401-402, Hackensack, NJ 07601
UK office: 57 Shelton Street, Covent Garden, London WC2H 9HE
Printed in Singapore.
PHOTONICS AND NANOTECHNOLOGY
Proceedings of the International Workshop and Conference on ICPN 2007
Benjamin - Photonics and Nanotechnology.pmd 10/31/2008, 11:50 AM1
v

PREFACE


Nonlinear optical physics has been to optical engineers and scientists a very interesting subject, especially, when
the nonlinear behaviors of light in optical devices generate benefits in some cases. This book is entitled ‘Photonics
and Nanotechnology’, where the papers were selected by the International Conference on Photonics and
Nanotechnology (ICPN) committees during the conference in the year 2007, which was held in Pattaya, Thailand,
from December 16–18, 2007. The conference was organized by the Department of Applied Physics, Faculty of

Science, King Mongkut’s Institute of Technology Ladkrabang (KMITL), Thailand. Twelve papers out of sixty were
selected, and the extended versions were slightly different from the conference versions. All papers concern optical
devices and materials, especially, the nonlinear behaviors and their benefits. The conference was partially supported
by the Department of Applied Physics, Faculty of Science, KMITL, Optical Society of America (OSA), The
Institute of Optical Engineering Society (SPIE), IEEE-LEOS (Thailand), National Electronics and Computer
Technology Center (NECTEC), Thailand and Ch. Karnchang (Thailand). There were some keynote and invited
talks involved from the United States of America, Europe and Japan. Professor Yupapin from KMITL was the
general chair of the conference. He had pushed a lot of effort and contributions to make the conference a success.
Finally, we expect that the proceedings volume will be useful to the optical researchers and society.
Preecha P. Yupapin


CONTENTS


Preface v

Capacitance-Voltage Characteristics of InN Quantum Dots in AlGaN/GaN Heterostructure 1
A. Asgari and M. Afshari Bavili

A Comparison of Different Coherent Deep Ultraviolet Generations Using Second Harmonic Generation 7
with the Blue Laser Diode Excitation
C. Tangtrongbenjasil and K. Konaka

Application of Reflection-Spectrum Envelope for Sampled Gratings 18
X. He, D.N. Wang, D. Huang and Y. Yu

Temperature-Dependent Photoluminescence Investigation of Narrow Well-Width InGaAs/InP Single 24
Quantum Well
W. Pecharapa, W. Techitdheera, P. Thanomngam and J. Nukeaw

Shooting Method Calculation of Temperature Dependence of Transition Energy for Quantum 31
Well Structure
B. Jukgoljun, W. Pecharapa and W. Techitdheera

Design of Optical Ring Resonator Filters for WDM Applications 35
P. Saeung and P.P. Yupapin

Chaotic Signal Filtering Device Using the Series Waveguide Micro Ring Resonator 42
P.P. Yupapin, W. Suwancharoen, S. Chaiyasoonthorn and S. Thongmee

An Alternative Optical Switch Using Mach Zehnder Interferometer and Two Ring Resonators 48
P.P. Yupapin, P. Saeung and P. Chunpang

Entangled Photons Generation and Regeneration Using a Nonlinear Fiber Ring Resonator 52
S. Suchat, W. Khunnam and P.P. Yupapin

Nonlinear Effects in Fiber Grating to Nano-Scale Measurement Resolution 60
P. Phipithirankarn, P. Yabosdee and P.P. Yupapin

Quantum Chaotic Signals Generation by a Nonlinear Micro Ring Resonator 65
C. Sripakdee, W. Suwancharoen and P.P. Yupapin

Investigation of Photonic Devices Pigtailing Using Laser Welding 72
M.M.A. Fadhali

A Soliton Pulse in a Nonlinear Micro Ring Resonator System: Unexpected Results and Applications 81
P.P. Yupapin, S. Pipatsart and N. Pornsuwancharoen

Author Index 109

October 8, 2008 15:54 WSPC - Proceedings Trim Size: 11in x 8.5in 01˙Asghar
1
CAPACITANCE-VOLTAGE CHARACTERISTICS OF InN QUANTUM DOTS IN
AlGaN/GaN HETEROSTRUCTURE
A. ASGARI
1 ,2
, M. AFSHARI BAVILI
1
1
Photonics-Electronics Group, Research Institute for Applied Physics, University of Tabriz,
Tabriz 51665-163, Iran
2
School of Electrical, Electronic and Computer Engineering, The University of Western Australia,
Crawley, WA 6009, Australia
E-mail:
In this paper the capacitance-voltage characteristics of InN quantum dots embedded in AlGaN/GaN heterostructure
has been studied. This work has been done for the InN quantum dots with different quantum dot size, and energy
dispersion and for the AlGaN/GaN heterostructures with different Al mole fraction and number of quantum wells in
different temperatures. The presence of InN quantum dot will cause Gaussian shape in capacitance-voltage charac-
teristics approximately, where the peak of curves can evidence the position of quantum dots in the structures. Our
calculation results show the Gaussian shape (or negative differential capacitance) is much higher at low temperature
and for quantum dots with low energy and higher size dispersion.
Keywords: AlGaN/GaN Heterostructures; InN Quantum Dots; capacitance.
1. Introduction
The progress of epitaxial growth technology has been responsible for many new structures based on low-
dimensional system where quantum effects were clearly observed. Self assembled systems like the quantum
dots (QDs) are very important examples of this advance. Recently, considerable interest has been focused on
electronic devices based on semiconductor heterostructures containing quantum dots, in which the motion of
quasi-particles is quantized along all three coordinates. In order to develop the application of these devices,
it is necessary to investigate the influence of quantum dot on electronic properties of these semiconductor

structures.
1
Capacitance spectroscopy is a highly efficient method for studying of electrical properties of these
structures. Recently, several research groups reported some results about AlGaAs/GaAs heterostructures
containing a layer of self-organized InAs QDs.
2
But Nitride based nanostructures have significantly different
properties as compared to GaAs based quantum wells and QDs. GaAs has zinc blend crystal structure, but
III-V nitrides are available in both zinc blend and wurtzite crystal structure which leads to strong built in
piezoelectric fields in heterostructures. This can induce red shift in GaN/AlN self-organized QDs.
3
In this
paper we present the results of capacitance-voltage studies of the AlGaN/GaN heterostructures containing
a layer of InN QDs.
2. Model Description
Consider the modeled sample structure as Fig. 1. The capacitance of the structure is the sum of the bulk
capacitance and QDs capacitance. In order to determine the capacitance, one has to know the conduction
band profile and all quantized state to calculate the electron density function and Fermi energy level using
self consistent solution of Schrdinger and Poisson equations. It has been done in this article using numerical
Numerov’s method.
4–6
To calculate the charge density in the structure, it has been assumed that the plane
containing the QDs acts like an equipotential surface and also only the ground state of quantum dots has
been occupied. Also, we consider the plane containing QDs and the highly doped buffer layer are near the
electrostatic equilibrium.
7
The capacitance in devices as Schottky device is directly related with the charge
inside the depletion region and can be expressed by C =
∂Q
∂V

where
Q = Q
bulk
+ Q
QD
= qS(N
D
W − N
QD
) (1)
October 8, 2008 15:54 WSPC - Proceedings Trim Size: 11in x 8.5in 01˙Asghar
2
Fig. 1. The modeled sample structure and Schematic conduction band profile.
And S is the Schottky contact area, ND is the bulk doping density, W is the width of depletion region and
equal to in the Fig. 1. Solving Poisson equation for the different applied voltage the capacitance can be
obtained as:
7
C
bulk
=
S
d
SL
+ S

qN
D

2(φ
B

− V )
(2)
C
QD
= qS

d
t

·
d
dV


∞
0
D(E, V )f (E, V )dE

(3)
where
D(E, V ) =
N
QD
√
π∆E
exp


−2


E + E
QD
+ q
d
t
V
∆E

2


(4)
and f(E, V ), the Fermi-Dirac energy distribution function, is
f(E, V ) =
1
1 + exp

q(E − qV )
kT

(5)
Also  is the GaN dielectric constant, t and d are the structure width as expressed in Fig. 1, φ
B
is the
Schottky barrier high, ∆E is the energy dispersion characteristics which expresses the effects of the dots
size dispersion.
2
E
QD
is the electron level within the QDs. To calculate the QD energy levels, it has been

assumed the QDs have spherical shape and the Fermi level in the dots was the same as in highly doped
substrate.
8
To find the C
QD
, the integral in Eq. (2.3) has been solved numerically.
3. Results and Discussion
The device structure as shown in Fig. 1 contains an AlGaN/GaN superlattice of N quantum wells with 1.5 nm
thickness of GaN and 3 nm of Al
x
Ga
1−x
N barrier width, a layer containing InN QDs, a 10 nm GaN layer
between QDs layer and superlattice, and a 20 nm Si-doped GaN layer with doping density of 8 × 10
4
cm
−3
.
The QDs layer includes a carrier density of 6.5 × 10
10
cm
−2
. The Schottky contact area and barrier high is
2 × 10
−7
cm
2
and φ
B
= 1.3x + 0.84 (eV), respectively, where x is the Al mole fraction in the barrier.

The capacitance-voltage characteristics of these structures have been analyzed in different physical sit-
uations. For the applied voltage range from −1 to +1 V, the dominant behavior of capacitance comes from
October 8, 2008 15:54 WSPC - Proceedings Trim Size: 11in x 8.5in 01˙Asghar
3
Fig. 2. The variation of Capacitance of AlGaN/GaN heterostructures with InN QDs as function of applied voltage at different
temperature. The number of quantum well in super lattice is n = 30 and Al mole fraction is x = 0.3.
Fig. 3. The variation of Capacitance of AlGaN/GaN heterostructures with InN QDs as function of applied voltage at T =
100 K, E
QD
= 80 meV and for different energy dispersion characteristic. The number of quantum well in super lattice is n = 30
and Al mole fraction is x = 0.3.
the QD capacitance. As one can see from Figs. 2, 3, and 4, the total capacitance increases with increasing
of applied voltage from −1 V, due to the filling of the quantum dots, showing a peak at voltage range from
V = −0.25 to 0.25. The peak broadening is due to the fluctuations in the dot sizes. If the voltage increases,
the total capacitances decreases and for further increases trend to bulk capacitance because the dots are
completely discharge. Figure 2 shows the variation of the capacitance as function of applied voltage at dif-
ferent temperature. As evident from the figure, the QD capacitance for low temperature is higher than the
October 8, 2008 15:54 WSPC - Proceedings Trim Size: 11in x 8.5in 01˙Asghar
4
Fig. 4. The variation of C-V characteristics of AlGaN/GaN heterostructures with InN QDs at T = 100 K, ∆E = 110 meV
and for different quantum dot energy. The number of quantum well in super lattice is n = 30 and Al mole fraction is x = 0.3.
high temperatures capacitance. Also with the increasing of the temperature, the peak positions shift toward
the low voltages. This is caused by the dynamical process involving the capture/emission rate of the dots.
The energy dispersion characteristics, ∆E, is 110 meV and E
QD
is the 80 meV in these calculations.
Figure 3 shows the variation of the capacitance as function of applied voltage at different energy dis-
persion characteristics, in T = 100 K and for dots with E
QD
is the 80 meV. As evident from the figure,

Fig. 5. The variation of C-V characteristics of AlGaN/GaN heterostructures with InN QDs at T = 100 K, ∆E = 110 meV,
E
QD
= 80 meV and for different Al mole fraction. The number of quantum well in super lattice is n = 30.
October 8, 2008 15:54 WSPC - Proceedings Trim Size: 11in x 8.5in 01˙Asghar
5
Fig. 6. The variation of C-V characteristics of AlGaN/GaN heterostructures with InN QDs at T = 100 K, ∆E = 110 meV,
E
QD
= 80 meV and for different number of quantum well in super lattice, the Al mole fraction is x = 0.3.
the QD capacitance for low energy dispersion characteristic is very small and total capacitance behave as
bulk capacitance. With increasing the energy dispersion characteristic, the QD show a higher negative dif-
ferential capacitance. In this case there is only a small change in the peak positions which shift toward the
high voltages. Also, as shown in Fig. 4, the total capacitance as function of applied voltage at different QD
energy level, in T = 100 K and for the quantum dots with energy dispersion characteristic of 110 meV
is expressed. As evident from the figure, the QD capacitance for dots with high energy level is very small
and total capacitance behaves as same as bulk capacitance. With decreasing the QD energy level, the QD
capacitance shows a higher negative differential capacitance. In this case there is not any change in the peak
positions along the voltage axes. To calculate the C-V characteristic in Figs. 3, 4, and 5, 30 quantum wells
has been taken into account in super lattice and Al mole fraction in the barriers is x = 0.3. The effects of
AlGaN/GaN heterostructures on C-V characteristics have been analyzed too. The variation of Capacitance
as function of applied voltage for the structures with Al mole fraction of 0.1 to 1 in the super lattice barriers
are shown in Fig. 5. As evident from the figure and Eq. (2.2), the capacitance decreases with increasing the
Schottky barrier potential which varies linearly with Al mole fraction. So, for the structures with higher Al
mole fraction, the quantum dots show more negative differential capacitance. Also the calculation has been
done for the structures with different number of quantum well in the superlattice. Figure 6 shows the results
of this calculation for the number of quantum well of n = 5, 10, 20, 30, 40, and 50. It is clearly known that
with increasing the distance between the capacitor plates, the electrical capacitance decreases. So to see the
quantum dot capacitance effect, it’s better to have the structures with low bulk capacitance.
4. Conclusions

In summary, this paper presented a study of the capacitance-voltage characteristics in the InN quantum
dots system embedded in a GaN matrix in AlGaN/GaN heterostructure. The proposed is based on the
analysis of the solution of the Poisson and Schrdinger equations and in the well defined relationship between
capacitance and density of sates. The calculation results shoe that the presence of InN quantum dot will cause
a negative differential capacitance which can evidence the position of quantum dots in the structures. Also,
our calculation results show that the negative differential capacitance is much higher at low temperature and
for quantum dots with low energy and higher size dispersion.
October 8, 2008 15:54 WSPC - Proceedings Trim Size: 11in x 8.5in 01˙Asghar
6
References
1. A. A. J. Chiquito, et al., Phys. Rev. B 61, 5499 (2000).
2. P. N. Brounkov, et al., Semiconductors. 32, 1096 (1998).
3. A. Bagga, et al., Phys. Rev. B 68, 155331 (2003).
4. A. Asgari, Study of transport properties of AlGaN/GaN Heterostructure, Physics Faculty, University of Tabriz,
Ph.D. Thesis. , 84 (2003).
5. A. Asgari, et al., J. Appl. Phys. 95, 1185 (2004).
6. A. Asgari, et al., Materials Science and Engineering C 26, 898 (2006).
7. Ph. Lelong, et al., Physica E 2, 678 (2006).
8. C. E. Pryor, et al., Phys. Rev. B 72, 205311 (2005).
7

A COMPARISON OF DIFFERENT COHERENT DEEP ULTRAVIOLET GENERATIONS
USING SECOND HARMONIC GENERATION WITH BLUE LASER DIODE EXCITATION

C. TANGTRONGBENCHASIL

AND K. NONAKA

Department of Electronic and Photonic Systems Engineering, Frontier Engineering Course,
Kochi University of Technology, Tosayamada, Kami City, Kochi Prefecture 782-8502, Japan

Nano-focus beam applications of short wavelength approximately 220 nm now play important roles
in engineering and industrial sections. At present, light sources at approximately 220 nm are
commercially available but large size, difficult to maintain, and expensive. Compact wavelength
tunable and cost effective light sources at approximately 220 nm are required. Laser diode with
sum-frequency generation methods are employed to generated the shorter wavelength
approximately 220 nm. This paper presents comparison of second harmonic generation schemes
using a nonlinear optic crystal and two types of laser diode, which are a 440 nm single mode blue
laser diode and a 450 nm multimode Fabry-Perot blue laser diode, has potential to generate wide
tunable coherent deep ultraviolet-c at approximately 220 nm. Using the blue laser diode with the
sum-frequency technique, a high second harmonic power is hardly observed due to low conversion
efficiency. The best performance of second harmonic generation using blue laser diode, nonlinear
optic crystal, and an high-Q external cavity laser diode was observed as 1.1 µW second harmonic
ultraviolet-c power at 224.45 nm ultraviolet-c wavelength and 5.75 nm ultraviolet wavelength
tunability. In addition, the improvement of increasing second harmonic power approximately
220 nm and the limitation of wavelength tuning of short wavelength are also theoretically discussed
in this paper.

1. Introduction
Coherent short wavelength ultraviolet C (UV-C) approximately 220 nm is very useful for nano-focus beam
applications such as beam lithography for very large scale integrated circuit (VLSI) and molecular spectroscopy.
Excimer lasers and sum-frequency from solid state lasers, which are able to generate very high power
1
but these
lasers have very large bodies, complex structures, fixed wavelength, high manufacturing costs, and high
maintenance costs, are conventional coherent UV sources at the short wavelength approximately 220 nm. Due to
these disadvantages of excimer UV lasers and solid state UV lasers, compact, simple to fabricate, cost effective, and
coherent wavelength tunable flexibility of UV sources are in demanded. One of the possible solutions is second
harmonic generation (SHG) with nonlinear optic crystal and laser diode (LD). Due to advance technology in opto-

electronics, small size LD can generate high optical power as 300 mW for continuous wave at 20ºC
2,3
. The SHG
researches have been reported for 4 decades
4-6
. Most of SHG researches were implemented with gas laser at
wavelength longer than 780 nm resulting fixed wavelength
4-6
. Only a few researchers reported the SHG for short
wavelength approximately 220 nm, due to very low conversion efficiency, a complex setup, insufficient LD power,
oscillation quality, and crystal efficiency
7-10
. External cavity diode laser (ECDL) with nonlinear optic crystal, that is
one of the solutions for SHG researches, has been reported
7-10
.
This paper present a performance comparison of SHG using a 440 nm single mode blue LD with a BBO
nonlinear optic crystal and a 450 nm multimode Fabry-Perot blue LD with a BBO nonlinear optic crystal. The
mathematical estimation of SH power and the improvement of SHG conversion efficiency are also discussed.
2. Theoretical Background
Dmitriev, et al. and Mills published simple SHG mathematical estimations when uniform beam is employed
4-5
.
However, Dmitriev, et al. and Mills’ equations are not able to estimate properly the SH power. Boyd and Kleinman
8

published a SHG mathematical model including phase mismatch factor, focal position factor, strength of focusing
factor, birefringence factor, and absorption factor, that is suitable to estimate the SH power when focusing Gaussian
beam is employed
6

. In this paper, a 440 nm fundamental single mode wavelength blue LD with a BBO nonlinear
optic crystal and a 445 nm and fundamental wavelength multimode Fabry-Perot blue LD with a BBO nonlinear
optic crystal were implemented to generate tunable coherent deep UV-C. The shortest usable wavelength of the
BBO crystal is 205 nm, due to phase matching angle limitation of fundamental and SH waves
11
. Using Sellmeier’s
equations
4,5
, operating refractive index, phase matching angle, walk-off angle, and effective conversion coefficient
can be theoretically obtained. The SH output power (
2
P
ω
) can estimated by
6
,

2 2
2 2
(1 ) (1 )
[ ( ) ( ) ( )]
2
2
2 3 ( )]
(1 ) (1 )
2
1
4
i
eff F

i
o oF eUV
d k
e
P P L d d
n n c e
ξ µ ξ µ
β τ τ κ τ τ σ τ τ
ω ω
τ τ
ξ µ ξ µ
ω
τ τ
ξπε
+ +
′ ′ ′
− − + − − −
′
− −
− − − −
′
= ⋅
∫ ∫
(1)
where
P
ω
is fundamental power [W],
L
is crystal length [m],

d
eff
is conversion efficiency,
k
F
is fundamental wave
propagation constant,
ε
0

is Planck’s constant
12
A s
V m
8.854 10
−
= ×
 
 
i
i
,
c
is light speed in free space
[
]
8
m
s
3 10= ×

,
n
oF
is
ordinary fundamental wave refractive index,
n
eUV

is extraordinary SH wave refractive index, b is confocal
parameter,
ξ

is strength of focus
L
b
=
,
µ

is focal position,
β

is birefringent parameter
1
2
F
Lk
ρ
ξ
=

,
ρ

is walk off angle
[radian],
κ

is absorption factor, and
σ

is phase mismatch. When

σ
= 0, focal position is at the center of the BBO
crystal or

µ
= 0, and no absorption or

κ
= 0, the optimized SH output power (
2
P
ω
) can be calculated as
6
,

2 2
2 2

[ ( ) ]
2
2
2 3 ( )]
2
1
4
eff F
i
o oF eUV
d k
e
P P L d d
n n c e
ξ ξ
β τ τ
ω ω
τ τ
ξ ξ
ω
τ τ
ξπε
′
− −
′
− −
− −
′
= ⋅
∫ ∫

, (2)
β
= 0

if and only if
ρ
= 0 that is invalid at either short fundamental wavelength as 440 nm or 445 nm.
ρ
are equal
to 0.067 radian and 0.073 radian, at 440 nm fundamental wavelength and 445 nm fundamental wavelength,
respectively. Consequently,
β
are equal to 13.06 radian and 14.05 radian when 100 mm focal length of focusing lens
was employed for 440 nm fundamental wavelength and 445 nm fundamental wavelength, respectively. Moreover,
effective focal length, which is a very important factor to optimized the SH conversion efficiency, is required to
estimate but it is not included in Eq. (1) and Eq. (2). The effective focal length can be estimated by
6
,

2
2
oF
eff
c
n f
b
L
w
π
π

= =
, (3)
where
f
is operating focal length of focusing lenses and
w
c
is beam radius of the collimated input beam. The
effective focal length is directly proportional to the confocal parameter. Equation 3 implies that there is an optimum
effective focal length of any arbitrary confocal parameter depending on the focal length of focusing lens. By the
symmetry of focusing and defocusing with the identical focal lengths of focusing lenses, consequently the optimized
crystal length is equal to 2
L
eff
. In addition, the confocal parameter is directly proportional to the operating focal
length, so longer focal length requires longer crystal length or longer SH interaction length to optimize the
conversion efficiency. Figure 1 shows simulations of SH output power vs fundamental input power with various
focal lengths and optimum nonlinear optic crystal lengths when confocal parameter = 0.084. Even confocal
parameter is fixed; the conversion efficiency can be improved by implementing long focal length and long nonlinear
optic crystal length. Moreover, it implies that long focal length requires long interaction length or long crystal length
to optimize SH output power. Figure 2 shows simulations of SH output power vs fundamental input power with
various confocal parameters when focal length of focusing lenses is 100 mm and nonlinear optic crystal length is
10 mm. Even focal length of focusing lens and interaction length are fixed; the conversion efficiency can be
improved by constructing the higher confocal parameter system. However, if the higher value of confocal parameter
is required, the optic size including lens diameter, nonlinear optic crystal length, nonlinear optic crystal cross-
sectional area, and optical operating distance must be enlarged.
9


Fig. 1. Simulations of SH output power vs fundamental input power with various focal length of focusing lenses

and optimum nonlinear optic crystal lengths when confocal parameter = 0.084.



Fig. 2. Simulations of SH output power vs fundamental input power with various confocal parameters when
focal length of focusing lenses is 100 mm and nonlinear optic crystal length is 10 mm.

On the other hand, narrow wavelength tolerance or single longitudinal mode oscillation is one of requirements to
realize theoretical SHG efficiency that must be enhanced. The wavelength tolerance must be control as narrow as
possible by oscillation wavelength selection system e.g. grating and feedback mirror, etc
12-16
. If the fundamental
10

wavelength is single longitudinal mode oscillation, consequently the SH wave is also single longitudinal mode
oscillation. The SH wavelength tunablility depends on the angle and position of feedback fundamental light passing
through grating back to LD. The feedback angle of fundamental light must be set as close as possible to the
polarization plane of LD, so that narrow single mode fundamental wavelength can be realized. To tune fundamental
wavelength, the position shift with respect to the orthogonal of polarization plane must be tuned. In contrast,
narrowing wavelength tolerance can cause phase mismatch. To overcome this problem, the nonlinear optic crystal
angle must be properly adjusted to matching angle.
3. Coherent Deep UV-C Generation Setups and Experimental Results
In this section, 3 experimental setups of coherent deep UV-C generations approximately 220 nm basing on SHG
scheme are discussed. A single mode blue LD approximately 440 nm and a multimode Fabry-Perot blue LD
approximately 450 nm are employed as fundamental wavelength light sources. The single mode blue LD 440 nm
can provide maximum continuous fundamental wave only 60 mW. To enhance higher continuous fundamental
power, the multimode Fabry-Perot blue LD 450 nm, which is able to provide up to 300 mW when LD temperature is
proper controlled at 25ºC
3
, was implemented instead of the 440 nm single mode blue LD. The detail performance

and comparison will be discussed in later section.

3.1. SHG with Feedback Grating as a Wavelength Selector Configuration
The simple and compact coherent deep UV-C generation approximately 220 nm is shown in Fig. 3. The single mode
blue LD approximately 440 nm was employed as fundamental light source that can provide maximum continuous
fundamental wave at 60 mW. The single mode blue LD was installed in a mount that can control the LD
temperature constantly and also provide a quasi-collimating beam. LD waveguide rear-end has approximately 90%
high reflection (HR) coating but LD waveguide front-end has coating reducing a few percentage of reflectivity
decreases the catastrophic optical damage (COD) damaging. The LD was controlled at 20ºC. The quasi-collimating
beam has beam profile of the effective parallel and perpendicular beam axes are 3.5 mm and 1.5 mm, respectively.
The quasi-collimating beam was focused at the center of 10 mm length BBO crystal by a 100 mm bi-convex lens.
Consequently, the effective parallel and perpendicular beam waist at the focusing region are 74.68 µm and
32.01 µm, respectively. The maximum average power of the fundamental wavelength inside the cavity was
64.83 mW. Thus, a 3.45 kW/cm
2
excitation is expected at around focus region. The output radiation from BBO optic
crystal consisting of approximately 440 nm fundamental wavelength and approximately 220 nm SH wavelength
were, consequently, collimated and reflected by a 100 mm concave mirror to obtain the similar quasi-collimating
beam profile as launching from LD mount. Then, the approximately 440 nm fundamental wavelength and the
approximately 220 nm SH wavelength were completely separated by prism at Brewster angle for the 440 nm
fundamental wave. To stabilize and narrow wavelength tolerance, the 440 nm fundamental wave was launched to
reflection grating that has 40% transmission and 60% reflection. The 40% transmission from the reflection grating
was employed to monitor the fundamental wavelength tolerance. To enhance the external high-Q ECLD and to
narrow the fundamental wave, the 60% reflection from the reflection grating must be fed to the same path back to
the LD mount. In addition, if the fundamental wave is narrow and single mode, the SH wave would also narrow and
single mode. Consequently, the conversion efficiency of narrow and single mode wave is better than wide and
multimode wave. To tune the wavelength, slight adjusting the angle of the reflection grating with respect to LD
polarization plane can stabilize and tune fundamental wavelength and SH wavelength. Generated UV light was
measured by the photomultiplier tube with transimpedance amplifier as shown in Fig. 3.
Figure 4 shows experimental results of SHG with feedback grating as the wavelength selector and examples of

operating fundamental wavelength stabilities. 220-nm range deep UV-C is too close to the limiting edge of crystal
matching condition and BBO crystal absorption band. Consequently, the conversion efficiency is lower than near
UV wavelength. The maximum generated SH power was obtained as 0.165 µW at 218.45 nm SH wavelength when
11


Fig. 3. SHG experimental setup with internal wavelength separator.



Fig. 4. Experimental results of SHG with internal wavelength separator and examples of operating fundamental
wavelength stabilities.

64.83 mW fundamental power was enhanced. The SH tunability is 1.45 nm in range of 218.45 nm – 219.9 nm. The
3-dB spectrum widths (
∆λ
) of 436.9 nm fundamental wave and 439.8 nm fundamental wave are 0.035 nm and
0.039 nm, respectively. In addition, the extinction ratio of 436.9 nm fundamental wave and 439.8 nm fundamental
wave are 24.10 dB and 29.56 dB, respectively. The wavelength separator was located inside the SHG cavity, it
caused the difficulty of minimizing phase mismatch. When operating fundamental wavelength was changed, the
crystal angle must be re-tuned. In addition, the position of optical spectrum analyzer (OSA) must be re-tuned to
monitor the stability of operating fundamental wavelength. Moreover, 1.45-nm narrow SH wavelength tunability
wave was observed.

12

3.2. SHG with Transmission Grating as a Wavelength Selector Configuration
Due to the difficulty of re-tuning of SHG cavity and OSA position in section 3.1, the wavelength separator (prism)
should be located outside the SHG cavity. Because of the polarizations of fundamental wavelength and SH
wavelength differ by 90º, so a 220 nm dichroic mirror was employed to separate approximately 220 nm wave and

approximately 440 nm wave. In addition the feedback light in section 3.1 is only 60%, so a transmission grating and
a 440 nm high reflection (HR) flat mirror were implemented as wavelength selector. In addition the transmission
grating and the 440 nm HR flat mirror were also employed as an external high-Q ECDL enhancement. The
transmission grating has splitting ratio of 0
th
order and 1
st
order by 5% and 95% of incident wave, respectively. The
0
th
order from the transmission grating was employed to monitor the stability of the operating fundamental wave.
The OSA can be fixedly place to monitor the stability of the operating fundamental wave because the position of 0
th

order does not depend on the transmission grating angle. Adjust the angle of feedback mirror is able to tune and
stabilize the operating wavelength in this setup. In addition, employing the transmission grating with the 440 nm HR
feedback mirror is able to improve the Q factor of ECLD. Figure 5 shows SHG experimental setup with
transmission grating as the wavelength selector configuration. In practice, the 220 nm dichroic mirror is not able to
reflect only 220 nm wave but a few percentages of 440 nm wave is also reflected. To separate 220 nm wave out of
440 nm wave completely, the prism must be employed. With the similar of beam profile as in section 3.1 and the
maximum average power of the fundamental wavelength inside the cavity was 64.36 mW. Thus, a 3.43 kW/cm
2

excitation is expected at around focus region.



Fig. 5. SHG experimental setup with external wavelength separator.

Figure 6 shows experimental results of SHG with external wavelength separator and examples of operating

fundamental wavelength stabilities. The maximum generated SH power was obtained as 0.194 µW at 218.25 nm
SH wavelength when 64.36 mW fundamental power was enhanced. The SH tunability is 1.85 nm in range of
218.25 nm – 220.1 nm. The 3-dB
∆λ
of 436.6 nm fundamental wave, 438.2 nm fundamental wave, and 440.2 nm
fundamental wave are 0.040 nm, 0.039 nm, and 0.042 nm, respectively. In addition, the extinction ratio of 436.6 nm
fundamental wave, 438.2 nm fundamental wave, and 440.2 nm fundamental wave are 25.73 dB, 23.55 dB, and
29.47 dB, respectively. This configuration can be improved by enhancing higher fundamental power. Because of
high-Q ECDL feed operating fundamental wave back to LD mount with similar beam profile, so SH power can also
be detected in front of LD mount.


13


Fig. 6. Experimental results of SHG with external wavelength separator and examples of operating fundamental
wavelength stabilities.

3.3. Symmetry SH Detection Configuration with Multimode Blue LD
The enhancement of higher fundamental wave power is one of the important factors to generate high SH power, so a
450 nm multimode Fabry-Perot blue LD was employed. The 450 nm multimode Fabry-Perot blue LD is able to
generate continuous power up to 300 mW when LD temperature is controlled at 25ºC
3
. However, the number of
oscillation mode increases when the injection current increases. Multimode oscillation reduces the SHG efficiency
due to narrow crystal matching tolerance. Moreover, the SHG from section 3.2 can be improved by bi-directional
detection; in front of LD mount and before the transmission grating (see Fig. 7). However, to simplify the
experimental setup of bi-directional detection, symmetry configuration is extremely required. Figure 7 shows
symmetry SH detection configuration multimode blue LD. In this section, two 100 mm plano-convex lenses were
employed to focus quasi-collimating fundamental wave from LD mount and defocus to re-collimate and to obtain

similar quasi-collimating fundamental wave beam profile as launching from LD mount. The polarizations of
fundamental wavelength and SH wavelength differ by 90º as mentioned in section 3.2. Two dichroic mirrors were
placed before (forward detection) and after (backward detection) plano-convex lenses (see Fig. 7) to separate the
225 nm SH wave out of SHG cavity and to maintain the 450 nm fundamental wave in the SHG cavity. In practice,
the reflected 225 nm wave from the dichroic mirror always contains a few percentage of the 450 nm wave, even the
dichroic mirrors are exactly placed at the Brewster angle. So the reflected wavelength can be completely separated
by prisms that were set to Brewster angle for the 225 nm SH wavelength transparency to obtain pure 225 nm
coherent deep UV-C. With the similar of beam profile as in section 3.1 and 3.2, the maximum average power of the
fundamental wave inside the cavity of multimode Fabry-Perot blue LD can be obtained 103.30 mW, due to
imperfect of temperature controller. Thus, a 5.50 kW/cm
2
excitation is expected at around focus region.
Figure 8 shows experimental results of SHG at 448.9 nm fundamental wavelength and the variation of
fundamental power vs. fundamental wavelength. The maximum generated SH power was obtained as 1.1 µW at
224.45 nm SH wavelength when 103.30 mW fundamental power was enhanced. The 1.1 µW was the total
detections of 0.67 µW forward detection and 0.34 µW backward detection. Using bi-directional detection technique,
an approximately 50% of SH power was obtained at backward detection, due to surface loss and scattering from

14


Fig. 7. Symmetry configuration of SHG with external wavelength separator.



Fig. 8. Experimental results of SHG at 448.9 nm and the variation of fundamental power vs. fundamental
wavelength.

fundamental wave, and 455.4 nm fundamental wave were small as 0.040 nm and 0.050 nm, respectively. In
addition, the average extinction ratio was approximately 25 dB. From this setup, the transmission grating was set

and fixed at 60º that maximized the splitting ratio of 0
th
order and 1
st
order of transmission. Only adjusting the angle
of the 440 nm HR flat mirror can tune the operating fundamental wavelength. The difference position and angle of
feedback fundamental light can cause wavelength tunability and oscillation mode suppression of fundamental light
of multimode Fabry-Perot blue LD. Consequently, fundamental power of nearby wavelength is decreased, due to
mode suppression. In addition, the limitation of wavelength tuning range is gain profile LD waveguide. At
the shortest wave and longest wave, the operating fundamental power was decreased by 2.1 dB comparing with
448.9 nm fundamental wave that has the maximum fundamental power at 103.30 mW. Consequently, the SH power
at the shortest wave and longest wave in this setup were decreased as 0.6 dB comparing with 224.45 nm SH wave
15

that has the maximum generated UV power at 1.1 µW. The total SH powers of shortest wave and longest wave in
this system were obtained as 0.101 µW and 0.104 µW, respectively.
4. Discussion
The 3 different coherent deep UV-C generation experimental setups and experimental results were explained.
Table 1 shows performance comparison of 3 experimental SHG setup. The SHG with feedback grating as a
wavelength selector configuration can generate the maximum UV power only 0.165 µW at 218.45 nm SH
wavelength when 64.83 mW fundamental power was enhanced by the 440 nm single mode blue LD. The SH
tunability was only 1.45 nm in range of 218.45 nm – 219.9 nm. Because of the SH wavelength tunability was too
narrow as 1.45 nm, the wavelength separator (prism) should be located outside SH cavity by inserting the 220 nm
dichroic mirror to reflect approximately 220 nm wave outside the SH cavity and maintain the fundamental wave
inside the SH cavity. In addition, to improve the fundamental feedback power, the transmission grating, which has
the extinction ratio of 5% and 95% of 0
th
order and 1
st
order, and the 440 nm HR flat mirror were employed

to reduce feedback loss and to enhance high-Q ECLD. From this improvement, the maximum UV power can
be generated as 0.194 µW at 218.25 nm SH wavelength when 64.36 mW fundamental power was enhanced by the
440 nm single mode blue LD. The SH tunability was improved by 0.5 nm to be 1.85 nm in range of 218.25 nm –
220.1 nm. The UV generation basing on SHG scheme can generate UV in both of forward direction that was
enhanced directly from LD and backward direction that was enhanced indirectly from LD but from feedback
fundamental light from the 440 nm HR flat mirror. To assure the experimental high-Q cavity symmetry for
bi-directional UV detection, two of the 220 nm dichroic mirrors were employed to separate the SH wave outside the
SHG cavity and maintain fundamental wave inside the SHG cavity. The maximum generated UV power was

Table 1. Performance comparison of 3 experimental SHG setups.
SHG setup types

Comparison topics
SHG with feedback
grating as the wavelength
selector
SHG with transmission
grating as the
wavelength selector
Symmetry SH detection with multimode
blue LD
Type of LD 440 nm Single mode blue LD 450 nm multimode Fabry-Perot blue LD
Maximum enhanced
fundamental power
inside SHG cavity
64.83 mW 64.36 mW 103.30 mW
Maximum generated
SH power
0.165 µW at 218.45 nm 0.194 µW at 218.25 nm 1.1 µW at 224.45 nm

Weak points of
generated SH power
1. Too low enhanced fundamental power low
2. Low confocal parameter
1. Difficulty of LD temperature control
2. Low confocal parameter
SH wavelength
tunability
1.45 nm in range of
218.45 nm – 219.9 nm
1.85 nm in range of
218.25 nm – 220.1 nm
5.75 nm in range of
221.95 nm – 227.7 nm
Type of wavelength
selection and feedback
Only reflection grating Transmission grating and flat mirror
Wavelength tuning
technique
Adjust the angle of
refection grating
Adjust only the angle of feedback mirror without adjust the angle of
transmission grating
Weak point of
wavelength tuning
technique
Low fundamental
feedback light causes
difficulty in wavelength
tuning

Waveguide
characteristic of single
mode blue LD limits
wavelength tuning
Waveguide characteristic of multimode
Fabry-Perot blue LD limits wavelength
tuning
16

obtained as 1.1 µW at 224.45 nm SH wavelength when 103.30 mW fundamental power was enhanced by the
450 nm multimode Fabry-Perot blue LD. The 1.1 µW was the total detections of 0.67 µW forward detection and
0.34 µW backward detection. Using bi-directional detection technique, an approximately 50% of SH power was
obtained at backward detection, due to surface loss and scattering from lenses, BBO crystal, transmission grating,
and flat mirror. Using the multimode LD, the SH tunability was extremely improved by 3.9 nm to be 5.75 nm in
range of 221.95 nm – 227.7 nm.
On the other hand, there are 2 possibilities to improve the SHG conversion efficiency; 1) increase the enhanced
fundamental power or 2) increase the confocal parameter. Increasing the enhanced fundamental power is easy
method but it consumes a lot of energy whereas increasing the confocal parameter requires beam expander and
beam reducer that causes enlarge optical size. To improve the wavelength tunability, the high feedback fundamental
power is required to reduce wavelength tolerance and suppress nearby wavelength. By this technique, the high
performance grating and high refection fundamental power are required.
5. Conclusions
In summary, using the single mode blue LD with flexible wavelength tunable and high-Q ECLD can observed
similar level of fundamental power at every tuned wavelength resulting similar level of SH generated UV power can
also be obtained. In contrast, using the multimode Fabry-Perot blue LD with wide wavelength tunability and single
mode oscillation high-Q ECLD cannot provide the similar level of fundamental power at every tuned wavelength.
The maximum fundamental power was observed as 103.30 mW at 448.9 nm whereas the maximum fundamental
power of the shortest and the longest wavelength were observed as 55 mW which was approximately 2.1 dB
decreasing resulting different levels of SH generated UV power were observed which was approximately 0.6 dB
difference. The experimental results of our experimental setups were well matched to the Boyd and Kleinmann

model estimation. To improve the SH conversion efficiency, higher enhanced fundamental power is required but it
consumes a lot of energy. Moreover, the increasing of confocal parameter and the crystal length are another possible
solution. However, there is a trade-off between optic size and conversion efficiency. If the high conversion
efficiency is required, the optical system size must be increased. In contrast, if the compactness is required, the
conversion efficiency is low. On the other hand, the main parameters; phase matching angle, walk-off angle, and
effective coefficient must be carefully controlled due to very slightly change of these parameters cause suddenly
change of SH efficiency. Up to present, the best performance of wavelength tuning is implementation of the
multimode Fabry-Perot blue LD with transmission grating and feedback mirror that is able to tune as wide as
5.75 nm. The limitation of wavelength tuning of this setup is from the waveguide characteristic and gain profile of
the multimode Fabry-Perot blue LD.
This paper showed the sufficient of wavelength tunability and compactness comparing with the conventional
excimer and YAG laser. Moreover, this system has potential to focus and achieve the higher power density than
bulk laser at the selected area.
Acknowledgment
This research was supported by JST research foundation and NICHIA Corporation foundation.
References
1. W. L. Zhou, Y. Mori, T. Sasaki, and S. Nakai, Optics Communications 123, pp. 583-586, 1996.
2. NICHIA Corp., “Blue Violet Laser Diode, NDHU110APAE2”.
3. NICHIA Corp., “Fabry-Perot Multimode Blue Laser Diode, NDHB220APAT1”.
4. V. G. Dmitriev, G. G. Gurzadyan, and D.N. Nikogosyan, Handbook of Nonlinear Optical Crystals, Springer
Series in Optical Sciences Volume 64.
5. D.L. Mills, Nonlinear Optics: Basic Concepts, 2
nd
edition, ed. (Springer, New York, 1998).
17

6. G. D. Boyd and D.A. Kleiman, “Parametric Interaction of Focused Gaussian Light Beam,” Journal of Applied
Physics, Vol. 39, No. 8, pp 3597–3639, 1968.
7. K. Ohara, M. Sako, and K. Nonaka, “210 nm ultraviolet generation using blueviolet laser diode and BBO SHG
crystal,” CLEO Pacific RIM conference, Taipei, 2003.

8. K. Ohara K. Nonaka, and P. Vesarach, “0.2
µ
m Deep UV Generation using 0.4
µ
m Blue Laser Diode with
Wavelength Tunable Cavity,” CLEO Pacific RIM conference, Tokyo, 2005.
9. C. Tangtrongbenchasil, K. Ohara, T. Itagaki, P. Vesarach, and K. Nonaka, “219-nm Ultra Violet Generation
Using Blue Laser Diode and External Cavity,” Japanese Journal of Applied Physics, Vol. 45, No. 8A, pp. 6315–
6316, 2006.
10. C. Tangtrongbenchasil, K. Nonaka, and K. Ohara, “220-nm Ultra Violet Generation Using an External Cavity
Laser Diode with Transmission Grating,” MOC 2006, Sep. 2006, Seoul, Korea, Vol. 2, pp. 5–8.
11. CASIX Co., Ltd., “Product Catalog 2004”.
12. T. Laurila, T. Joutsenoja, R. Hernberg, and M. Kuittinen, “Tunable external-cavity diode laser at 650 nm based
on a transmission diffraction grating,” Applied Op., Vol. 41, No. 27, pp. 5632–5637, 2002.
13. H. Patrick and C.E. Wieman, “Frequency stabilization of a diode laser using simultaneous optical feedback
from a diffraction grating and narrowband Fabry-Perot cavity,” Rev. Sci. Instrum., Vol. 62, No. 11, pp. 2593–
2595, 1991.
14. A. Wicht, M. Rudolf, P. Huke, R. Rinjkeff, and K. Danzmann, “Grating enhanced external cavity diode laser,”
Appl. Phys. B, 2003.
15. M. W. Flemming and A. Mooridian, “Spectral Characteristics of External-Cavity Controlled Semiconductor
Lasers,” IEEE J. Quantum Electron, Vol. QE-17, No. 1, pp. 44–59, 1981.
16. K. Hayasaka, “Frequency stabilization of an extended-cavity violet diode laser by resonant optical feedback,”
Optics Comm. 206, pp. 401–409, 2002.
18

APPLICATION OF REFLECTION SPECTRUM ENVELOP
FOR SAMPLED GRATINGS

XIAOYING HE
1,2

, D.N.WANG
2*
, DEXIU HUANG
1
AND YONGLIN YU
1

1
Wuhan National Laboratory for Optoelectrons, Huazhong University of Science and Technology, Wuhan,
Hubei,430074, P.R.China
2
Department of Electrical Engineering, The Hong Kong Polytechnic University, Hung Hom, Kowloon,
Hong Kong, P.R.China



*
Corresponding author email:
Analytical expression is proposed for evaluating the performances of sampled gratings. Accuracy of this expression
has been verified by simulated reflectivity spectrum with the transfer matrix method. A new technique of multiplex
reflection-spectrum envelope concatenation is introduced to demonstrate a 23-channel grating with uniform
characteristics in all channels. The proposed technology can densify sampled grating both in spectral channels
number and in spatially physical corrugation.

1. Introduction
Sampled gratings (SGs) are naturally attractive for wide applications in optical communications and optical sensor
systems such as tunable semiconductor reflectors [1, 2], multi-channel dispersion compensators [3, 4], multi-channel
multiplexers-demultiplexers [5], repetition rate multiplication [6], etc. Particular interests that have been shown in
the performances of sampled gratings include the envelope-top flatness and 3dB envelope bandwidth of the
reflection spectrum. Especially, the multi-channel gratings with broad flat-top spectrum envelopes, as tunable

semiconductor reflectors, will significantly improve the performances of laser over a wide tuning range. A number of
techniques proposed for this purpose, including Sinc-apodization [7], multiple-phase shift technique [8], and
interleaved technique [9]. Moreover, the transfer matrix method cannot convey the relation of grating parameters and
the top-flatness and width of reflection-spectrum envelope (RSE). Simulation employing transfer matrix method is a
time-consuming task especially for long gratings. Therefore, it is necessary to propose an analytical expression of
RSEs for conventional sampled gratings to study the impacts of grating parameters on RSEs.
In this paper, an accurate analytical expression of the RSE for sampled grating is proposed and demonstrated.
Based on this analytical expression, the new multiple reflection-spectrum envelope concatenation (MRSEC)
technology is employed to design multi-channel gratings with broad flat-top reflection spectra.
2. Analytical Expression of Reflection-spectrum Envelope
2.1. Theory
The main spectral features of sampled grating, whether photo-refractive grating or etched grating, can be derived
from the modulation of the effective refractive index. The effective refractive-index profile can be governed by:


,
( ) ( ) ( ) ( ) ( ) cos ,
eff eff eff
i
n z n n z f z z iZ g z z
∞
=−∞
 
 
 
 
= + ⋅ ∗ − ⋅ ⋅ +
 
 
 

 
 
Λ
 
 
 
 
∑
0 0
2
1
π
δ δ υ
(1)