fibre bragg grating components for filtering, switching ...1805/fulltext01.pdf · fibre bragg...

KTH Information and

Communication Technology

Fibre Bragg Grating Components for Filtering, Switching and Lasing

ZHANGWEI YU

Doctoral Thesis in Microelectronics and Applied Physics

Stockholm, Sweden 2008

KTH School of Information and

TRITA-ICT/MAP AVH Report 2008:14 Communication Technology

ISSN 1653-7610 SE-164 40 Kista

ISRN KTH/ICT-MAP/AVH-2008:14-SE SWEDEN

Akademik avhandling som med tillstånd av Kungliga Tekniska Högskolan Framlägges till

offentling granskning för avläggande av teknologie doktorsexamen i datalogi fredag den 3

oktober 2008 klockan 10.00 i N2, Electrum 3, Kungliga Tekniska Högskolan,

Isafjordsgatan 28, Kista, Stockholm.

© Zhangwei Yu, August 2008

Tryck: Universitetsservice US AB

iii

Abstract

Fibre Bragg gratings (FBGs) are key components for a vast number of applications in

optical communication systems, microwave photonics systems, and optical sensors, etc.

The main topic of this thesis is fibre Bragg grating fabrication and applications in direct

microwave optical filtering, high speed switching and switchable dual-wavelength fibre

lasers.

First, a brief overview is given about the photosensitivity in optical fibre, basic FBG

fabrication techniques, the popular coupled-mode theory for describing fundamental

characteristics of FBGs and the Transfer Matrix method for the numerical simulations of

complex-structured FBGs.

An advanced FBG fabrication system based on the technique of multiple printing in

fibre (with a continuous-wave source) has been used to write complex FBGs incorporating

phase shifts, apodization and chirp.

A single double-peaked superimposed grating working in reflection can be employed

as a direct optical filter for millimetre-wave signals. Bit error rate measurements confirmed

that the filter exhibited nearly on-off behaviour in the passband with a 3-dB bandwidth of 2

GHz for a central frequency of 20 GHz, as expected from the optical reflection spectrum.

The presented technique can be used in radio-over-fibre systems or simultaneous

up-conversion of ultra-wide band signals and filtering.

This thesis focused mostly on the research of two 4-cm long Hamming-apodized

gratings written in side-hole fibres with internal electrodes. The temperature dependence

measurements showed that the birefringence of the component increased with the

temperature. Dynamic measurement has shown nanosecond full off-on and on-off switching.

During the electrical pulse action, the grating wavelength was blue-shifted for the

x-polarization and red-shifted for the y-polarization due to the mechanical stress. Both

peaks subsequently experienced a red-shift due to the relaxation of mechanical stress and

iv

the increasing core temperature transferred from the metal in many microseconds. All the

wavelength shifts of the two polarizations depend quadratically on the electrical pulse

voltage and linearly on the pulse duration. Numerical simulations gave accurate description

of the experimental results and were useful to understand the physics behind the

birefringence switching.

Finally, two switchable dual-wavelength erbium-doped fibre lasers based on FBG

feedback were proposed. In one method, an overlapping cavity for the two lasing

wavelengths and hybrid gain medium in the fibre laser were introduced. Dual-wavelength

switching was achieved by controlling the Raman pump power. The other method

employed an injection technique and the dual-wavelength switching was controlled by the

power of the injection laser. The switching time was measured to be ~50 μs. Detailed characteristics of the dual-wavelength switching in the two fibre lasers were experimentally

studied and corresponding principles were physically explained.

Keywords: fibre Bragg grating, photosensitivity, apodization, chirp, phase-shift,

microwave optical filtering, incoherent filtering, direct optical filtering, direct detection,

side-hole fibre, switching, high speed, internal electrode, birefringence, dual-wavelength,

fibre laser.

v

Contents

Abstract ⅲ

Contents ⅴ

List of Papers ⅸ

List of Acronyms ⅹⅲ

Acknowledgements ⅹⅴ

1 Introduction 1

1.1 Background ................................................................................................................ 1

1.2 Motivation and outline ............................................................................................... 2

2 Fibre Bragg Gratings Background 5

2.1 Photosensitivity in optical fibres................................................................................ 6

2.1.1 History of photosensitivity in optical fibres...................................................... 6

2.1.2 Mechanism of photoinduced refractive index change ...................................... 7

2.1.3 Enhanced photosensitivity in silica optical fibres............................................. 8

2.2 Basic techniques for fibre Bragg grating fabrication ................................................. 9

2.2.1 Interferometric fabrication technique.............................................................. 10

2.2.2 Phase-mask technique ..................................................................................... 13

2.2.3 Point-by-point technique................................................................................. 14

2.3 Advanced technique for fibre Bragg grating fabrication ......................................... 15

2.3.1 Multiple printing in fibre technique ................................................................ 15

vi

2.3.2 Sequential exposure with a continuous-wave UV laser source .......................16

2.4 Fibre Bragg grating theory........................................................................................19

2.4.1 Fundamental properties of fibre Bragg gratings ............................................19

2.4.1 Coupled-mode theory.....................................................................................20

2.4.3 Transfer matrix method..................................................................................22

3 Direct Optical Filtering of Microwave Signals 25

3.1 Introduction...............................................................................................................25

3.1.1 Incoherent filtering...........................................................................................25

3.1.2 Coherent filtering.............................................................................................28

3.1.3 Comparison between incoherent and direct optical filtering of microwave

signals ...............................................................................................................30

3.2 Direct microwave optical filtering with direct detection ..........................................31

3.3 Perspective and applications .....................................................................................35

4 High Speed Switching 37

4.1 Fabrication of the FBG component ..........................................................................37

4.2 Static experiment ......................................................................................................40

4.2.1 Temperature dependence measurement ...........................................................41

4.2.2 Discussion and calculation...............................................................................42

4.3 Dynamic experiment.................................................................................................46

4.3.1 Experimental setup...........................................................................................46

4.3.2 Result and analysis...........................................................................................46

4.4 Numerical simulation................................................................................................54

5 Switchable Dual-wavelength Fibre Lasers 63

5.1 Raman erbium-doped fibre laser...............................................................................63

5.1.1 Experimental setup...........................................................................................64

5.1.2 Results and analysis .........................................................................................64

5.1.3 Characteristics of the dual-wavelength switching ...........................................65

5.2 Tuneable and injection-switchable EDF laser...........................................................67

vii

5.2.1 Experimental setup.......................................................................................... 68

5.2.2 Results and analysis ........................................................................................ 68

5.2.3 Characteristics of the dual-wavelength switching .......................................... 69

5.2.4 Measurement of the switching time ................................................................ 72

6 Conclusions and Future Works 73

7 Summary of Papers 77

Bibliography 81

ix

List of Papers

List of Papers included in the thesis:

I. Z. Yu, A. Djupsjobacka, M. Popov, and P.-Y. Fonjallaz. Direct detection of direct

optically filtered millimeter-wave signals. Photonics Techn. Lett., 19(4):233-235,

February 2007.

II. Z. Yu, W. Marguilis, O. Tarasenko, H. Knape, and P.-Y. Fonjallaz. High-speed control

of fiber Bragg gratings. Bragg Gratings, Photosensitivity and Poling in Glass

Waveguides (BGPP), Quebec City, Canada, paper JWA44, September 2007.

III. Z. Yu, W. Marguilis, O. Tarasenko, H. Knape, and P.-Y. Fonjallaz. Nanosecond

switching of fiber Bragg gratings. Opt. Express, 15(22):14948-14953, October 2007.

IV. Z. Yu, W. Margulis, P.-Y. Fonjallaz, and O. Tarasenko. Physics of electrically

switched fiber Bragg gratings. Conference on Lasers and Electro-Optics/Quantum

Electronics and Laser Science Conference and Photonic Applications Systems

Technologies (CLEO-QELS), San Jose, USA, paper CMK1, May 2008.

V. Z. Yu, O. Tarasenko, W. Marguilis, and P.-Y. Fonjallaz. Birefringence switching of

Bragg gratings in fibers with internal electrodes. Opt. Express, 16(11):8229-8235,

May 2008.

VI. L.-E. Nilsson, Z. Yu, O. Tarasenko, H. Knap, P.-Y. Fonjallaz, and W. Margulis.

High-speed switching in fibres with electrodes. Proceedings of IEEE Conference on

Transparent Optical Networks, Rome, 232-233, July 2007.

VII. D. Chen, Z. Yu, S. Qin, and S. He. Switchable dual-wavelength Raman erbium-doped

fibre laser. Electron. Lett., 42(4):202-204, February 2006.

x

VIII. D. Chen, Z. Yu, S. Qin, and S. He. Some switchable dual-wavelength fibre lasers

based on fibre Bragg grating feedback. Proc. of SPIE 6351: 635121, 2006.

IX. D. Chen, S. Qin, Z. Yu, and S. He. Tunable and injection-switchable erbium-doped

fiber laser of line structure. Microwave and Optical Techn. Lett. 49(4):765-768,

April 2007.

List of Papers not included in the thesis, but related:

X. B. Cabon, S. Constant, G. N Guyen, Z. Yu, and P.-Y. Fonjallaz. All-optical systems

for both frequency conversion of UWB signals and filtering. Broadband Europe

Conference 2007, Antwerp, Belgium, paper We2B1, December 2007.

XI. G. N Guyen, B. Cabon, Z. Yu, and P.-Y. Fonjallaz. Simultaneous Up-Conversion to

60 GHz and Side-Band Suppression using a Fibre Bragg Grating. Accepted by

Workshop on FTTH, Wireless Communictions and their interaction, Kista, Sweden,

June 2008.

XII. G. N Guyen, B. Cabon, Z. Yu, and P.-Y. Fonjallaz. All Optical Up-Conversion of

WLAN signal in 60 GHz Range with Side-Band Supression. Submitted to IEEE

Radio and Wireless Symposium (RWS), San Diego, USA, 2009.

XIII. W. Margulis, O. Tarasenko, Z. Yu, P.-Y. Fonjallaz, and H. Knape. High-speed fiber

switches. Accepted by the first Workshop on Specialty Optical Fibers and Their

Applications (WSOF), invited, AIP Conference Proceedings Series, Campinas,

Brazil, 2008.

XIV. Z. Yu, O. Tarasenko, H. Knape, R. Koch, and W. Margulis. Versatile all-fiber

polarization switch. Accepted by the 3rd EPS-QEOD Europhoton Conference, Paris,

France, 2008.

XV. Z. Yu, P.-Y. Fonjallaz, W. Marguilis, and O. Tarasenko. High speed switching of a

DFB grating in a twin-hole fibre. Accepted by Asia-Pacific Optical

Communications (APOC), Hangzhou, China, 2008.

XVI. Z. Yu, R. Koch, W. Margulis, O. Tarasenko, H. Knape, and P.-Y. Fonjallaz.

Nanosecond All-fiber Polarization Switch. Submitted to Asia Optical Fiber

Communication and Optoelectronic Exposition & Conference (AOE), Shanghai,

China, 2008.

xi

XVII. X. Zhang, L. Shao, Z. Yu, D. Chen, R. Jin, and S. He. Automatic real-time control

for gain-flattened fiber Raman amplifiers. Opt. Commun., 239:79-84, September

2004.

XVIII. X. Zhang, and Z. Yu. Automatic control algorithm of Raman amplifiers. Proc. of

SPIE, 5625:1150-1157, 2005.

xiii

List of Acronyms

FBG Fibre Bragg grating

MWP Microwave photonics

MMW Millimetre-wave

RF Radio frequency

RoF Radio-over-fibre

HiBi High birefringence

WDM Wavelength-division multiplexing

EDF Erbium-doped fibre

SLR Sagnac loop reflector

MPF Multiple printing in fibre

CW Continuous-wave

TMM Transfer matrix method

UV Ultraviolet

GODC Germanium-oxygen deficiency centre

FWHM Full width at half maximum

SMF Single mode fibre

EDFA Erbium doped fibre amplifier

FSR Free spectral range

HF High frequency

IF Intermediate frequency

LO Local oscillator

BER Bit error rate

xiv

PC Polarization controller

LPF Low-loss filter

UWB Ultra-wide band

TLS Tuneable light source

OSA Optical spectrum analyzer

LD Laser diode

DCF Dispersion compensation fibre

OC Optical circulator

TLD Tuneable laser diode

ISO Isolator

xv

Acknowledgements

First and foremost, I am very grateful to my supervisor Dr. Pierre-Yves Fonjallaz for his

encouragement and advice throughout my time in KTH. As an expert in FBGs, he taught

me hand-in-hand in my research work. He is a precious mentor, an enthusiastic friend and a

gentle man. Thanks for the good time in London and Canada, on Sandhamn island and for

the last Midsummer party.

Especially, I would like to express my deepest appreciation to Prof. Sailing He, my

supervisor in China, for his excellent guidance and encouragement in all these six years. He

led me to the research world when I was an undergraduate student.

I would like to thank Prof. Lars Thylen to be my official supervisor in KTH. I also

want to thank China Scholarship Council (CSC) for providing me with a scholarship.

Grateful thanks to Walter Margulis for giving me the opportunity to work on these

special fibres. He is full of inspirations and really imparts me lots of experience and

knowledge. Thanks for the constant guidance, fruitful discussions and refinement on my

publications. A special thank to Oleksandr Tarasenko for his technique support in the

laboratory.

I want to show gratitude to my colleagues in Acreo’s Photonics Department: Ingemar

Petermann, Leif Kjellberg, Erik Petrini, Carola Sterner, Anders Djupsjöbacka, Mikhail

Popov, Ralf Koch, Harald Knape, Asa Claesson, Lars-Erik Nilsson, Erik Zetterlund, Jie Li,

Qin Wang, Stefan Melin, Evgeny Vanin, Marco Forzati, Jonas Mårtensson, Hanna

Lundbeck and many others. They gave me valuable help and made the department a fun

place to work at.

I also want to thank all my colleagues in FMI group, Evaa Andersson, Prof. Min Qiu,

Dr. Richard Schatz, Dr. Erik Forsberg, Prof. Urban Westergren, Prof. Eilert Berglind, Prof.

Bozena, Marek Chacinsk, and many others for their kindness and the friendly atmosphere

they conducted in the group.

xvi

Thanks to all my Chinese friends in Stockholm: Hong Xin, Jing Wang, Yaocheng Shi,

Shan Qin, Jinhua Yan, Sanshui Xiao, Ning Zhu, Yuntian Chen, Minhao Pu, Youbin Zheng,

Xingang Yu, Jintao Zhang, Chunsheng Yan, Zhili Lin, Zhichao Ruan, Liu Liu, Xiang Lv,

Junfeng Bao, Jun Song, Yi Jin, Xin Hu, Jiajia Chen, Weiquan Mei, Xuan Li, Yangjian Cai,

Jinliang Huang, Zhen Zhang, Min Yan, Wei Yan, Jie Tian, Qin Wang, Jing Wang,

Zhenzhong Zhang, Jiantong Li, Zhibin Zhang, Yan Zhou, Jun Luo, Rui Zhao, Qian Yang,

Qiao Ye, Jingshi Li, Kun Wang, Jiayue Yuan, Liyang Zhang, Shuqing Yu, Changju Wu,

Shurong Wang, Likang Yang, Jian Chen, Caihong Xu, Botao Shao, Ying Cao, Peizi Li,

Huimin She, Peng Wang, Zhi Zhang, Yi Feng, Xuan Chen, …They gave me happy

memories during these years.

I am grateful to Dr. Ruxiang Jin for providing me a training position in his company,

where I learnt valuable industry experience. I also want to thank my friends and colleagues

in China: Sulan Zhang, Ping Lv, Yanbo Li, Xiaoxi Chen, Xinhong Li, Han Chen, Ting

Chen, Fan Yu, Shengmei Zheng, Xuliang Zhang, Daru Chen, Jinfei Ding, Zexuan Qiang,

Haiyan Ou, Hongyan Fu, Jun Qian, Zuguang Guan, Meng Jiang, Liyang Shao, Yongbo

Tang, and many others. Thanks for the friendship and good collaborations.

Finally, I would like to express my sincerely appreciation to my parents, little sister

Zhangqi, grandparents, uncles, aunties and cousins for their support and encouragement

throughout my whole life. Special gratitude to my dear Geng for his love, encouragement

and all the cherished moments we shared during the past 9 years.

Zhangwei Yu

Stockholm, August 2008

1

Chapter 1

Introduction

1.1 Background

The field of fibre Bragg gratings (FBGs) is almost thirty years old, dating back to its

discovery by Hill and co-workers in Canada [1]. It grew slowly in the beginning, but an

important technological breakthrough by Meltz and co-workers 10 years later [2] renewed

worldwide interest in FBGs. People have paid a lot of attention to study the foundation of

fibre photosensitivity, to improve the FBG fabrication technology, to develop the FBG

theory and simulation methods. Evidently, FBGs have been recognized as key components

in a variety of applications [3-5]. Their unique filtering properties and versatility as in-fibre

device is illustrated by their use in optical communication systems, microwave photonic

systems and fibre sensors, etc. This thesis is focused on the applications of FBGs for

microwave optical filtering, high speed switching and switchable dual-wavelength fibre

lasers.

The area of microwave photonics (MWP) has evolved in parallel to the development

of optical communication, as optical components have become operational up to millimetre

frequencies. Referring to Ref. [6, 7], MWP can be defined as the study of photonic devices

operating at microwave frequencies and their applications in microwave and optical

systems. The interdisciplinary field of MWP covers applications within a frequency span

from MHz to THz. The term microwave will be freely used throughout this thesis to

designate microwave, millimetre-wave (MMW) or radio frequency (RF). One of the most

common MWP processing is optical filtering. FBGs are the most widely used microwave

photonic filters, with advantages of low loss, large bandwidth and electromagnetic

Chapter 1. Introduction

2

interference immunity, etc. Hence FBGs have a vast number of MWP applications,

including communications (such as radio-over-fibre (RoF) systems, wireless bidirectional

communication and broadcasting), high performance microwave and MMW signal

generation and for antennas beam-forming arrays.

In 1986, Xie et al. first reported that the side-hole fibres could be used in a simple and

sensitive fibre-optic pressure sensor [8]. Compared with high birefringence (HiBi) fibres,

side-hole fibres have the advantage of higher sensitivity to pressure and lower sensitivity to

temperature [8-14]. FBGs imprinted in side-hole fibres have been shown to be good

candidates for simultaneous sensing of hydrostatic pressure and temperature [15-17].

Side-hole fibres with internal electrode can be exploited in applications where light

propagation is controlled electrically. Besides electro-optical switching and modulation,

nanosecond polarization switching has recently been demonstrated [18]. Hence high speed

switching of a FBG written in a side-hole fibre with internal electrodes was made possible,

as it is shown in Chapter 4 of this thesis.

Wavelength-switchable fibre lasers have important applications in wavelength-division

multiplexing (WDM) fibre communication systems, fibre sensors, spectroscopy and optical

instrument testing, etc. Several techniques have been reported to achieve wavelength

switching in Erbium-doped fibre (EDF) lasers. These include the use of sampled FBGs [19],

cascaded FBGs [20-22], multimode FBGs [23, 24], a Bragg grating-based acoustooptic

superlattice modulator [25], a spectral polarization-dependent loss element [26], a Sagnac

loop reflector (SLR) [27], distributed-feedback mode oscillation [28], a multi-section HiBi

fibre loop mirror [29] and a hybrid gain medium (semiconductor optical amplifier) [30, 31].

Wavelength-switchable lasers based on some EDFs co-doped with other rare-earth irons [32]

and based on yetterbium-doped fibre [33] have also been reported. Meanwhile, Raman fibre

lasers have also been developed for high power, multiwavelength (however, non-switchable)

applications [34-36]. Most of these techniques are based on some special devices or manual

adjustments to achieve wavelength switching.

1.2 Motivation and outline

The motivation for this thesis was mainly to investigate applications of FBGs for optical

filtering of microwave signals, for high speed switching and for switchable

dual-wavelength fibre lasers. The main focus has been on experimental work, but

theoretical work and simulations have also been carried out, to better understand and

2.1. Photosensitivity in optical fibres 3

evaluate the different systems.

It is the author’s intention that this thesis be readable for an audience that has a general

knowledge in the field of physics, optics and mathematics. The introductory part of the

thesis is to give a short background of FBGs and their importance in various applications.

Chapter 2 briefly reviews the history of the photosensitivity in optical fibres, the major

accepted mechanisms of photosensitivity and techniques to enhance the fibre’s

photosensitivity. The basic external techniques for recording Bragg gratings in optical fibres

are described. An advanced fabrication system based on the technique of multiple printing

in fibre (MPF) using a continuous-wave (CW) source has been used in the work and

presented here. Chapter 2 also introduces the use of the coupled-mode theory to describe

the fundamental properties of FBGs, and the Transfer matrix method (TMM) for the

numerical simulation of complex-structured FBGs.

Chapter 3 briefly reviews the incoherent and coherent optical filtering of microwave

signals. A direct microwave optical filtering technique combined with direct detection has

been proposed, employing a superimposed FBG working in reflection as a filter. The

principle and experiment results based on this technique and its applications in RoF have

been discussed.

Chapter 4 gives an overview of the static and dynamic experimental results of two

FBGs written in side-hole fibres with internal electrodes under the excitation of nanosecond

electrical pulses. The behaviour of the wavelength shifts for the fast and slow polarization

states has been explained in details. Numerical simulations give accurate description of the

experimental results and help to understand the physics behind the birefringence switching.

Chapter 5 proposes two methods to achieve dual-wavelength switching in

erbium-doped fibre laser using FBGs. The detailed characteristics of the dual-wavelength

switching in the two fibre lasers have been experimentally studied and the principles have

been explained physically.

Conclusions of the thesis are presented in Chapter 6 followed in Chapter 7 by short

summaries of the nine included papers.

5

Chapter 2

Fibre Bragg Gratings Background

An FBG consists of a periodic modulation of the refractive index along a fibre waveguide.

It is fabricated by exposure of the fibre to an intense optical interference pattern. Light

propagating in the fibre waveguide will be reflected by each grating plane. If the Bragg

condition is satisfied, the contributions of the light reflected from each grating plane add

constructively in the backward direction within a wavelength band whose centre

wavelength is defined by the grating parameters. Otherwise, the reflected light becomes

progressively out of phase and will eventually cancel out.

The Bragg condition is simply the requirement which satisfies both energy and

momentum conservation. The first-order Bragg condition is defined as [4, 5]

2B effnλ = Λ (2.1)

where the Bragg wavelength (λB) is the free-space centre wavelength of the input light that will be reflected back from the FBG, neff is the effective refractive index of the fibre core at

the free-space centre wavelength and Λ is the period of the gratings. Fig. 2.1 illustrates a

uniform FBG, where the phase fronts are perpendicular to the fibre axis and the grating

planes have a constant period Λ . As shown in the insets of Fig. 2.1, a well defined peak

with wavelength λB exists in the reflection spectrum, resulting in a notch in the transmission spectrum.

Chapter 2. Fibre Bragg Grating Background

6

Figure 2.1. Illustration of a uniform FBG with constant index modulation amplitude and

period. Insets: reflection and transmission spectra.

2.1 Photosensitivity in optical fibres

2.1.1 History of photosensitivity in optical fibres

The photosensitivity in optical fibres refers to a permanent change in the refractive index of

the fibre waveguide induced by exposure to light radiation with characteristic wavelength

and intensity that depends on the fibre material [3, 4]. The photosensitivity of optical fibres

was first observed at the Canadian Communication Research Centre in 1978 by Hill and

co-workers during experiments using germanosilica fibre and visible argon ion laser

radiation (488 nm) [1, 37]. They explained the permanent FBG written in the fibre core as a

result of the standing wave intensity pattern due to the Fresnel reflection from the cleaved

end of the fibre (4% reflection). This particular grating had a very weak index modulation,

which was estimated to be of the order of 10-6 resulting in a narrow-band reflection filter at

the writing wavelength. This discovery led to the internal inscription technique.

In 1981, Lam and Garside [38] demonstrated that the magnitude of refractive index

change reported by Hill [1] depended on the square of the writing power at 488 nm. This

suggested a two-photon process as the possible mechanism of refractive index change. In

1987, Stone et al. [39] showed that a refractive index change could be induced in all

Λ

Bragg grating

Fibre core

Index modulation

Cladding

Transmission

Reflection

Tran

smis

sion

Wavelength

Ref

lect

ion λB

λB

Broadband source

Wavelength

2.1. Photosensitivity in optical fibres 7

germanium doped fibre, but he was still using blue-green irradiating light launched in the

fibre waveguides..

The major breakthrough came with the report on holographic writing of gratings using

single-photon absorption at 244 nm by Gerry Meltz et al. in 1989 [2]. They demonstrated

that a reflection grating in the visible range of 571-600 nm was formed when a

germanium-doped fibre was exposed to two interfering ultraviolet (UV) beams. The scheme

provided the freedom of shifting the Bragg wavelength to longer and more useful

wavelengths, determined by the angle between the interfering beams. The first reflection

gratings at around 1500 nm written into germanosilicate fibres was reported [40]. The

UV-induced refractive index change in untreated germanosilicate fibres was typical of ~

10-5-10-4.

Since then, several developments have taken place that have pushed the index change

in optical fibres up a hundredfold. Fibres doped with europium [41], cerium [42], and

erbium:germanium [43] showed varying degrees of sensitivity in a silica host optical fibre,

but none of these were as sensitive as the germanosilicate fibres. One fibre doping

producing large index modulations (of the order of 10-3) is germanium-boron codoping

[44] .

2.1.2 Mechanism of photoinduced refractive index change

The photosensitivity in optical fibres was discovered almost 30 years ago, but the

mechanism behind it is still not fully understood. Several models [3-5] have been proposed

for these photoinduced refractive index changes, such as colour centre model, dipole model,

compaction model and stress-relief model. The only common element in these theories is

that the germanium-oxygen deficiency centre (GODC) defect, Ge-Si or Ge-Ge (the

so-called “wrong bonds”) is the trigger mechanism for the photoinduced index changes.

Glasses usually contain many defects that have strong photon absorption, also called

colour centres. The peak wavelength of absorption of the well-known GODC defect is at ~

240 nm [45], with a full width at half maximum (FWHM) of ~35 nm. This absorption band

was shown to be bleached when exposed to UV radiation [46, 47]. Hand and Russell [48]

proposed a model to explain the refractive index change by relating it to the absorption

changes via the Kramer-Kronig relationship. The GODC defect, which initially absorbs the

UV light, is transformed to GeE’ with the release of an electron (free to move within the

glass until it is trapped). According to this model, the refractive index at a point is related

Chapter 2. Fibre Bragg Grating Background

8

only to the density and orientation of defects in that region and is determined by their

electronic absorption spectra. It allows for the refractive index change from 10-5 to 10-4 but

can not account for changes much greater than 10-4. The colour centre model for

photosensitivity was supported by many experiments [46, 47, 49-51]. However, it can not

completely explain all the experiment observations [52, 53]. To date, the colour centre

model is the most widely accepted model for the formation mechanism of FBGs.

2.1.3 Enhanced photosensitivity in silica optical fibres

Since the discovery of photosensitivity and the first demonstration of grating formation in

germonosilicate fibres, there has been considerable effort to understand and increase the

photosensitivity in optical fibres. Standard single mode fibres (SMFs), doped with 3%

germanium, typically display index changes of ~10-6-10-5 [3] (except one report of ~1×10-3

[54]). Increasing the dopant level or subjecting the fibre to reducing conditions at high

temperatures result in larger index changes of ~5×10-4 [4]. Over the years, hydrogen

loading, flame brushing and boron codoping have been developed to enhance the

photosensitivity of silica optical fibres. Refractive index changes of ~10-3-10-2 then become

possible.

(1) Hydrogen loading

Hydrogen loading of optical fibres is the most commonly used technique for achieving

high UV photosensitivity in germanosilica optical fibres and was first developed by

Lemaire et al. [55, 56]. Fibres are soaked in hydrogen gas at temperatures ranging from

20-75 °C and pressures from ~20 atm to more than 750 atm (typically 150 atm), which

results in diffusion of hydrogen molecules into the fibre core. Under UV laser radiation, the

hydrogen molecules react with the Ge-O-Si bonds in the glass, forming OH species which

increases the level of oxygen-deficiency in the glass matrix and hence increases the

absorption at ~240 nm. The obtainable index change in the fibre is increased to the

magnitude of 10-3-10-2. It typically takes a week of hydrogen soaking at room temperature

in order to obtain the desired photosensitivity in the fibre. The required time can be reduced

to a few days by increasing the temperature or the pressure during the loading process. It

should be pointed out that high-pressure hydrogen is dangerous and requires special

precautions.

2.2.Basic techniques for fibre Bragg grating fabrication 9

One advantage of hydrogen loading is the fabrication of FBGs in any germanosilicate

fibres, and even in germanium-free fibre when shorter wavelength UV light is used

(typically 193 nm). Additionally, permanent changes occur only in regions that are UV

irradiated. In unexposed fibre sections the molecular hydrogen does not react with the glass

matrix and diffuses out, leaving an increased absorption loss at 1.38-1.45 μm, but relatively

small excess loss at the important optical communication windows (1.3 μm and 1.5 μm).

(2) Flame brushing

Flame brushing is a simple and effective way of enhancing the photosensitivity in

germanosilicate fibres [57]. The region of the optical waveguide to be photosensitized is

brushed repeatedly by a flame fuelled with hydrogen a small amount of oxygen, reaching a

temperature of ~1700 °C. The photosensitization process takes approximately 20 mins to

complete. At these temperatures, the hydrogen diffused into the fibre core very quickly and

reacts with the germanosilicate glass to produce GODC defects, creating a strong

absorption band at 240 nm and rendering the core highly photosensitivity. This method has

been used to increase the photosensitivity of standard telecommunication fibres by a factor

greater than 10, achieving changes in the refractive index >10-3 [57]. The increased

photosensitivity in the fibre by flame brushing technique is permanent, as opposed to

hydrogen loading where the fibre losses its photosensitivity as the hydrogen diffuses out of

the fibre. However, one major drawback of this technique is that the high temperature flame

weakens the fibre which increases the likelihood of the fibre breaking.

(3) Boron codoping

Germanosilicate fibre codoped with boron has larger photosensitivity than the fibre

with higher germanium concentration and without boron [44]. In addition, saturated index

changes are higher and achieved faster in Boron codoping germanosilicate fibre. Boron

codoping increases the photosensitivity of the fibre by allowing photoinduced stress

relaxation to occur, which initiated by the breaking of the wrong bonds by UV light.

2.2 Basic techniques for fibre Bragg grating fabrication

Most FBGs are fabricated through side-exposure of the fibre to two interfering beams from

the same UV laser, also called external writing technique. Different techniques are briefly

described below.

10 Chapter 2. Fiber Bragg Grating Background

2.2.1 Interferometric fabrication technique

The interferometric fabrication technique, demonstrated by Meltz et al [2], was the first

external writing technique to form Bragg gratings in photosensitive fibres. It utilized an

interferometer that split the incoming UV light into two beams and then recombined them

to form an interference pattern. The fringe pattern was used to expose a photosensitive fibre,

inducing a refractive index modulation in the core. Bragg gratings in optical fibres have

been fabricated using both amplitude-splitting and wavefront-splitting interferometers.

(1) Amplitude-splitting interferometer

In an amplitude-splitting interferometer (see Fig. 2.2), the UV writing laser light is

divided by a beam splitter into two equal intensity beams and are later recombined after

propagation over different optical paths. This forms an interference pattern at the core of a

photosensitive fibre. Cylindrical lenses are normally placed in the interferometer to focus

the interfering beams perpendicular to the fibre axis to a fine line matching the fibre core.

This results in high intensities at the core, thereby improving the grating inscription. The

Bragg grating period Λ , which is identical to the period of the interference fringe pattern,

is given by:

2sin

UVλϕ

Λ = (2.2)

where λUV is the irradiation UV wavelength and ϕ is the half-angle between the intersecting

UV beams. Given the Bragg condition 2B effnλ = Λ , the Bragg resonance wavelength λB

can be expressed as a function of λUV and ϕ as

sineff UV

B

n λλ

ϕ= (2.3)

From Eq. 2.3 one can easily see that the Bragg grating wavelength λB can be varied either

by changing λUV and/or ϕ. The choice of λUV is limited to the UV photosensitivity region of

the fibre; however there is no restriction for the choice of the angle ϕ (limited of course to

ϕ < 90º).

2.2. Basic techniques for fibre Bragg grating fabrication 11

Figure 2.2. Setup of the amplitude-splitting interferometer for the grating fabrication [5].

The most important advantage of this technique is the ability to inscribe Bragg

gratings at various wavelengths, accomplished by simply changing the intersecting angle

between the UV beams. This method also offers a large flexibility for what concerns the

length of the gratings, which allows for adjusting the bandwidth of their spectral response.

The main disadvantage of this method is its sensitivity to mechanical vibrations.

Displacements as small as submicrons in the optical components can cause the fringe

pattern to drift and wash out the grating. Furthermore, due to long separate optical path

lengths involved in the interferometers, fluctuations of the air refractive index may also

deteriorate the formation of a stable fringe pattern. In addition, quality gratings can only be

produced with a laser source that has good spatial and temporal coherence with excellent

wavelength, output power and beam-pointing stability.

(2) Wavefront-splitting interferometer

Wavefront-splitting interferometers, not as widely used as the amplitude-splitting

interferometers, have some useful advantages. Two examples of wavefront-splitting

interferometers used to fabricate Bragg gratings in optical fibres are the prism

interferometer [40, 58] (shown in Fig. 2.3(a)) and the Lloyd’s interferometer [59] (shown in

Fig. 2.3(b)). Both of them require a UV source with good spatial coherence.

An advantage of this method is that only one optical component is used, which causes

a reduction in the sensitivity to mechanical vibrations. In addition, the short distance where

the UV beams are separated reduces the wavefront distortion induced by air currents and

12 Chapter 2. Fibre Bragg Grating Background

temperature differences between the two interfering beams. The Bragg wavelength is easily

varied by rotating the set-up around the point of intersection between the fibre and the

reflecting surface. The major disadvantage of the two wavefront methods is that the

gratings’ length is limited to half of the beam width and to the asymmetrical intensity

distribution. Nevertheless this method is suitable for fabricating short quasi-uniform FBGs

for the investigation of the fibre’s photosensitivity.

(a)

(b)

Figure 2.3. Setup of the wavefront-splitting interferometers [5] used for Bragg grating

fabrication: (a) prism interferometer; (b) Lloyd interferometer.

2.2. Basic techniques for fibre Bragg grating fabrication 13

2.2.2 Phase-mask technique

The phase-mask technique is one of the most effective methods for inscribing Bragg

gratings in photosensitive fibre [60, 61]. This method employs a diffractive optical element

(phase mask) to spatially modulate the UV writing beam. Phase masks may be formed

holographically or by electron-beam lithography [62]. The benefit of the holographic

method over the electron-beam is that there is no stitching error in the grating pattern on the

mask [63]. On the other hand, complicated patterns can be written into the electron beams

fabricated masks, such as quadratic chirps, Moiré patterns, etc. The period of phase mask

pmΛ is two times the period of the grating Λ ( 2pmΛ = Λ ) [3]. When a UV beam is

perpendicularly incident on a phase mask with a proper depth, the zero-order diffracted

beam can be nearly cancelled out and the plus and minus first-order diffracted beams are

maximized, each typically containing more than 35% of the transmitted power (see Fig.

2.4). This is used to create an interference pattern with a period of Λ in the photosensitive

optical fibre which is placed in contact with or in close proximity to the phase mask.

Figure 2.4. Setup of the phase-mask technique for Bragg grating fabrication [5].

The phase mask greatly reduces the complexity of the fibre grating fabrication system.

The simplicity of using only one optical element provides an inherently stable method for

reproducing FBGs. Since the fibre is usually placed directly behind the phase mask in the

near field of the diffracting UV beams, sensitivity to mechanical vibrations is minimized.


Low temporal coherence does not affect the writing capability due to the geometry of the

problem, opposite to the interferometric technique.

The main disadvantage of this technique is the limitation of the tuneability in the

inscribed Bragg wavelength, which is basically determined by the fixed periodicity of a

phase-mask. Prohaska et al. [64] have shown experimentally that magnifying the

phase-mask periodicity by using extra lens only achieved tuneability in the order of a few

percent.

2.2.3 Point-by-point technique

The point-by-point technique for fabricating Bragg gratings is accomplished by inducing

the refractive index change one step at a time along the core of the fibre. In a typical

experimental setup shown in Fig. 2.5, a single pulse of UV light from an excimer laser is

passed through a slit and then focused onto the fibre at one point. As a result, the refractive

index of the core in the irradiated fibre section increases locally. The fibre is then shifted by

a distance Λ , corresponding to the period of the grating, in a direction parallel to the fibre

axis. The process is repeated to form the grating structure in the fibre core. Essential to the

point-by-point fabrication technique is a very stable and precise submicron translational

system.

Figure 2.5. Setup of point-by-point technique for Bragg grating fabrication [5].

2.3. Advanced technique for fibre Bragg grating fabrication 15

The main advantage of the technique is its flexibility to alter the Bragg grating

parameters, including the grating length, pitch and response. Therefore, with the same setup

it is possible to create different grating structures without having to replace any components.

However, it requires a relatively long process time due to a step-by-step procedure. Errors

in the grating spacing due to thermal effects and/or small variations in the fibre’s strain can

occur. Therefore, the gratings are limited to rather short length and Malo et al. [65] have

demonstrated the creation of second and third order gratings [5].

2.3 Advanced technique for fibre Bragg grating fabrication

There is a rapidly growing demand for high-quality optical Bragg gratings with arbitrary

phase and index profiles, because these gratings are important elements in WDM systems,

wireless communication systems, optical sensors, etc. In the past few years, several

methods have been developed for realizing gratings with complex structures using

phase-mask based techniques [66-71]. While having the benefit of a potentially high

reproducibility, the disadvantage of all these methods is that they rely on phase masks that

have to be of very high quality with a low amount of fabrication errors (and therefore rather

expensive) to ensure high quality gratings. In many cases, new complex grating designs

will require special phase masks, which substantially lower the flexibility and raises costs.

In the following section, an advanced fabrication system in Acreo will be briefly

introduced, where no phase masks are necessary to form the grating.

2.3.1 Multiple printing in fibre technique

In 1995, Stubbe et al. [72] first proposed the multiple printing in fibre (MPF) technique for

writing long and complex fibre gratings. The idea was to consecutively expose short

subgratings consisting of a few hundred periods with a pulsed UV source while at the same

time moving the fibre. With a high precision translation of the fibre along its axis relative to

the interference pattern, the phase and position of each subgrating is precisely controlled.

The interference pattern can be generated by an interferometer with an ordinary beam

splitter. Different grating designs, such as phase shifts, chirp and apodization, are realized

by changing the relative phase between the subgratings [73, 74]. Therefore, new grating

designs can be easily implemented by mere programmatic changes in the software


synchronizing the fibre movement and UV pulse generation. There are some drawbacks in

using pulsed operation of the UV source [75], which are beyond the scope of this thesis.

One of them is that the fibre can easily be destroyed due to high peak intensities impinging

on the fibre and the not perfectly constant pulse-to-pulse energy when using excimer lasers.

2.3.2 Sequential exposure with a continuous-wave UV laser source

A fabrication scheme comprising a CW UV laser source can overcome all the problems

mentioned above. In this system, the pulses are replaced by a sawtooth movement of the

interference pattern. Apart from lower power consumption and shorter fabrication times, it

is easy to keep a constant distribution of the UV dose in the fibre and a static environment

without sudden bursts of heat, as is the case with pulsed irradiation. The basic concept of

this method was proposed in Ref. [76, 77].

Fig. 2.6 shows a schematic of the FBG fabrication system based on the MPF technique

and a CW UV source. Fig. 2.7 shows the interferometer for producing the interference

pattern during the exposure. The fibre is mounted on an air bearing linear stage. The

position of the translator stage relative to the UV interference pattern is measured either

with a heterodyne interference detection system utilizing a He-Ne laser or with an encoder

[72].

A frequency-doubled argon-ion laser emitting at 244 nm with a power of up to 100

mW is launched into the interferometer. A computer controlled shutter opens only when the

grating is to be written.

Since the fibre is translated, with a speed which is kept by the way as constant as

possible, the fringes of the CW interference pattern must follow this movement during

exposure in order to maintain the visibility of the grating. To accomplish this two mirrors c

(see Fig. 2.7) are mounted on a pair of piezo translators in the UV interferometer. A

displacement in push-pull manner of these mirrors introduces a phase shift between the

interfering beams, which shifts the fringe pattern. Driving the piezo actuators with a

sawtooth signal controlled by the fibre movement causes the pattern to move together with

the fibre over one grating period. At the end of the voltage ramp in each period, the fringes

jump back to the original position and the constructive formation of the grating continues.

During the short moment when the fringes are moved back, the fibre is evenly exposed,

giving rise to a slight dc index increase. The principle of this technique was also reported in

Ref. [76].

2.3. Advanced technique for fibre Bragg grating fabrication 17

A step motor is used to move the mirror pair d (see Fig. 2.6), which symmetrically

changes the separation of the interfering beams in the plane of incidence. This modified

separation is translated into an angular change Δα by the cylindrical lenses e, since the latter are placed at their focal distance from the fibre. The big advantage of this set-up is

that changing the Bragg wavelength of the gratings is realized in a very precise manner and

just by moving to mutually attached mirrors d. No other mechanical adjustments, e.g. the

lateral position of the fibre, are required (in the limit of the spherical aberrations of the

lenses e). The period of the interference pattern is then changed consequently. This

modification of the periodicity of the fringes can be performed during the fabrication

process of chirp gratings without introducing any phase errors in the exposed grating.

A cylindrical lenses f focuses the two interfering beams perpendicularly to the fibre

axis on to the fibre core. Longitudinally, the beam extends over ~70 μm, which roughly corresponds to 130 fringes for a Bragg wavelength of 1550 nm.

The speed of the translation stage is limited by the response time of the piezo actuators:

for large velocities, mechanical ringing will substantially reduce the visibility of the

imprinted grating. In the setup used here, writing velocities of the order of centimetres per

second have been employed, but usually it is limited to a few tenths of mm/s.

Complex structures such as apodizations, phase shifts, and chirps are realized by

controlling the shape of the sawtooth signal driving the piezo translators as described in

more detail in Ref. [75, 78]. All grating parameters are controlled by the computer software

(a Labview user interface) and new grating designs are easily implemented. The

characteristics of the gratings that can be fabricated by this advanced system are:

Bragg wavelength: ~500-1700 nm

Chirp range during fabrication: >100 nm

Spatial resolution of the profiles: ~20-100 μm (determined by the subgrating length)

Maximum grating length: ~30 cm

Maximum writing speed: ~3 cm/s

Figure 2.8 shows reflection spectra of three FBGs fabricated by this system: (a) a

20-mm long hamming-apodized phase shift grating; (b) a 60-mm long 1-nm chirp grating

with hamming apodization; (c) a 76-mm long superstructure grating with 511-chip gold

code modulated exposure over the length of the grating. The superstructure FBG can be

used as an en/decoder in 511-chip 1.25-Gbit/s optical code division multiple access

systems.


Figure 2.6. Schematic of the FBG fabrication system based on the MPF technique and a

CW UV laser source [78].

Figure 2.7. Interferometer for producing the interference pattern on the fibre core [78].

Moving fiber

UV light

2.4. Fibre Bragg grating theory 19

1544 1546 1548 1550 1552 1554 1556

-40

-20

0

1550 1551 1552 1553 1554

-30

-20

-10

(c)

Wavelength, nm

(b)

Ref

lect

ivity

, dB

m

1544.6 1544.8 1545.0 1545.2 1545.4

-20

-10

0

(a)

Figure 2.8. Reflection spectra of three FBGs: (a) a 20-mm long phase shift grating with

hamming apodization; (b) a 60-mm long 1-nm chirp grating with hamming apodization; (c)

a 76.76-mm long superstructure grating with 511-chip gold code.

2.4 Fibre Bragg grating theory

Several theories have been developed to describe the behaviour of Bragg gratings. The

coupled-mode theory [3, 79] is the most popular one, mainly due to its simplicity and

accuracy in modelling the optical properties of most FBGs of interest. For complicated

grating structures, the coupled-mode theory involves the numerical solution of two coupled

differential equations. Transfer matrix method [80] has been developed for the purpose of

grating analysis. This chapter briefly presents the coupled-mode theory to model the

properties of most FBGs and the principle of TMM. Detailed derivation of the

coupled-mode theory and TMM can be referred to the textbooks [3, 4] on FBGs and further

literature [79, 81-83].

2.4.1 Fundamental properties of fibre Bragg gratings

FBGs consist of a spatial quasi-periodic variation of the refractive index along the fibre

length. The variation of the refractive index is assumed small enough to be viewed as a


perturbation of the initial effective refractive index neff. The total effective refractive index

neff(z) at position z in the fibre takes the form:

( ) ( )eff eff effn z n n zδ= + (2.4)

If the variation is described by a cosine, the perturbation term in Eq. 2.4 can be expressed as

[84]

[ ] ( ) ( ) 1 cos ( )effeffn z n z Kz zδ ν φ= Δ + + (2.5)

where effnΔ is the effective dc index change spatially averaged over the grating period;

0 1ν≤ ≤ is the fringe visibility of the index change; 2 /K π= Λ is the spatial frequency of the grating having a central period Λ ; ( )zφ is a variable phase factor describing

grating chirp. Inserting the last equation, Eq. 2.4 can be given by

[ ]( ) ( ) cos ( )eff av effn z n n z Kz zφ= + Δ + (2.6)

Where effav effn n n= + Δ is the average refractive index, ( ) effeffn z nνΔ = Δ is the

modulation depth. The modulation depth may in general change slowly along the length of

the grating, forming an index modulation profile with apodization. Gratings with a variable spatial frequency ( ) ( )K z K K z= + Δ are called chirped gratings. Therefore, Eq. (2.6) is

able to describe both apodization and chirp of a grating.

2.4.1 Coupled-mode theory

Coupled-mode theory has often been used to accurately model the optical properties of

most gratings [3, 79]. In the presence of a dielectric perturbation associate with the FBG, a

mode of amplitude R(z) is coupled into an identical counter-propagating mode of amplitude

S(z) according to [4, 83]

( ) ( ) ( )dA z i A z i B zdz

ζ κ∧

= + (2.7)

( ) ( ) ( )dB z i B z i A zdz

ζ κ∧

∗= − − (2.8)

where ( ) ( )exp( / 2)A z R z i zδ φ= − , ( ) ( ) exp( / 2)B z S z i zδ φ= − + , 12

ddzφζ δ ζ

∧

= + − is

the general dc self-coupling coefficient ( /δ β π= − Λ is the detuning wave vector), and

κ is the ac coupling coefficient.


For a single-mode Bragg reflection grating, one finds the following simple relations

[83]:

2effnπδ δ

λ= (2.9)

effnπκ κ νδλ

∗= = (2.10)

A solution of the coupled-mode differential Eq. 2.7 and 2.8 for a uniform FBG of length L can be found assuming that (0) 1A = and ( ) 0B L = . The amplitude reflection

coefficient of a grating (0) / (0)B Aρ = and the power reflection coefficient 2r ρ= can be

shown to be [4, 83]

2 2

22 2 2 2 2

sinh ( ) ( )

sinh ( ) ( ) cosh ( ) ( )

L L

L L i L L

κ κ ζρ

ζ κ ζ κ ζ κ ζ

∧

∧ ∧ ∧

− −=

− + − −

(2.11)

2 2 2

22 2 2

2

sinh ( ) ( )

cosh ( ) ( )

L Lr

L L

κ ζ

ζ κ ζκ

∧

∧∧

−=

− + −

(2.12)

The group delay and dispersion of the reflected light can be extracted from the

associated phase θ. The time delay of the light reflected from a grating can be obtained by

differentiating this variable with respect to the angular frequency ω of the reflection mode, which results in

2

2gd d d dd d d c d

θ θ λ λ θτω λ ω π λ

= = = − (2.13)

The Dispersion Dg, the rate of change of delay with wavelength, can be obtained by

differentiating Eq. 2.13 against the wavelength, producing

''2

2g gg

d d d cDd d dτ τ ω π βλ ω λ λ

= = = − (2.14)

where ''β is defined as the group velocity dispersion parameter (second derivative of the

propagation constant with respect to frequency).


2.4.3 Transfer matrix method

Figure 2.9. A nonuniform grating of Length L is divided into a series of M uniform gratings

with different periods, represented by their respective transfer matrix Tk.

The piecewise uniform approach is considered flexible and accurate for modeling

nonuniform gratings. It is based on multiplication of 2×2 matrices for each uniform section

of the grating to obtain a single 2×2 matrix that describes the whole grating [4, 78, 80]. The

compound grating structure can be divided into M uniform matrix components, each having

a specific period kΛ , phase offset kφ and modulation depth knΔ (see Fig. 2.9). By

changing subgrating parameters from element to element, any grating apodization and

phase profile can be approximated. kA and kB are the field amplitudes after traversing

the kth uniform section travelling in z+ and z− directions, respectively. The boundary

conditions can be assumed as 0 (0) 1A A= = and 0 ( ) 0B B L= = , meaning that light with

amplitude 1 is launched at z = 0 and no light is launched or reflected beyond the grating end

at z = L. The propagation through each uniform section k is described by a matrix Tk, which

is defined such that

1 111 12

21 221 1

k k kk

k k k

A A AT TT

T TB B B− −

− −

⎛ ⎞ ⎛ ⎞ ⎛ ⎞⎛ ⎞= =⎜ ⎟ ⎜ ⎟ ⎜ ⎟⎜ ⎟⎜ ⎟ ⎜ ⎟ ⎜ ⎟⎝ ⎠⎝ ⎠ ⎝ ⎠ ⎝ ⎠

(2.15)

The elements in this matrix are given by

11 22 cosh( ) sinh( )T T dl i dlζ∧

∗= = Ω − ΩΩ

(2.16)

12 21 sinh( )T T i dlκ∗= = − ΩΩ

(2.17)

where the star (*) denotes complex conjugation, dl is the length of the kth uniform section,


ζ∧

and κ are the local coupling coefficients for the kth section, and 2 2κ ζ∧

Ω = − . The

total grating structure may be expressed by the product of the different subgrating transfer

matrices, i.e.

01 1

0

MM M k

M

AAT T T T

B B−

⎛ ⎞⎛ ⎞= ⋅⋅⋅ ⋅⋅ ⋅ ⎜ ⎟⎜ ⎟

⎝ ⎠ ⎝ ⎠ (2.18)

Moreover, the transfer matrix Tk can be easily inverted since its determinant is unity. Eq.

(2.15) can be written as

1 22 12

1 21 11

k k

k k

A AT TB T T B

−

−

−⎛ ⎞ ⎛ ⎞⎛ ⎞=⎜ ⎟ ⎜ ⎟⎜ ⎟−⎝ ⎠⎝ ⎠ ⎝ ⎠

(2.19)

The equation for the field at the front face is given by

0 22 12 11 12

10 21 11 21 21

( )0

MM

kMk

A T T A I I A LB T T B I I=

−⎛ ⎞ ⎛ ⎞ ⎛ ⎞ ⎛ ⎞⎛ ⎞= Π =⎜ ⎟ ⎜ ⎟ ⎜ ⎟ ⎜ ⎟⎜ ⎟− ⎝ ⎠⎝ ⎠ ⎝ ⎠ ⎝ ⎠⎝ ⎠

(2.20)

The reflection coefficient at the front face of the grating is given by

0 21

0 11

B IA I

ρ = = (2.21)

The reflectivity is calculated from the square of its magnitude 2ρ . In addition, the

amplitude transmission coefficient t is simply expressed as:

0 11

( ) 1A LtA I

= = (2.22)

25

Chapter 3

Direct Optical Filtering of Microwave Signals

This chapter briefly reviews the incoherent and coherent optical filtering techniques and

describes the direct microwave optical filtering technique combined with direct detection

as presented in Paper Ⅰ. The principle and experiment results based on this technique and

its applications in radio-over-fibre systems are presented.

3.1 Introduction

Microwave photonic filters [7, 85, 86] are photonic subsystems designed with the aim of

realizing tasks equivalent to those of an conventional electrical filter [87, 88] within a RF

system or link. The unique functional advantages of microwave photonic filters [89],

including low loss, large bandwidth, electromagnetic interference immunity, tuneability and

reconfigurability, have led to diverse applications. Optical filtering of microwave signals

can be divided into two main categories, incoherent and coherent, which will be reviewed

briefly below.

3.1.1 Incoherent filtering

Most of today’s microwave optical filters employ incoherent filtering. It consists of the

26 Chapter 3. Direct Optical Filtering of Microwave Signals

incoherent summing of delayed copies of a modulated optical signal at the photodetector

[90-93]. The result of this incoherent summation is a periodic microwave response and

discrete or continuous tuneability. The copies are created in a number of ways employing

FBGs, circulators, recirculating lines, couplers and even erbium doped fibre amplifiers

(EDFAs). Following the terminology used in digital signal processing, the incoherent filters

can be divided into two main groups:

Recursive: filters with an infinite impulse response (the number of delayed signal

copies is infinite).

Non-recursive or transversal: filters with a finite impulses response (the number of

delayed signal copies is finite).

The simplest incoherent microwave optical filter can be created by summing two

modulated, and delayed with respect to each other, optical signals at the photodetector. The

delay time Δt between the signals determines the frequency separation f = 1/Δt between two neighbouring maxima in the periodic microwave frequency response of the filter and is also

called the free spectral range (FSR). Employing a large number of delayed signals allows

for creating more complex microwave filter responses.

Figure 3.1. Designed and measured microwave response of the incoherent filter based on 18

FBGs with appropriate weighting [94].

The Q-factor is defined as the ratio between the filter central frequency and the

FWHM of the filter passband. The larger the number of taps, the higher the Q-factor will be.

As a rule of thumb, one may use an approximate relationship Q ≈ N, where Q is the value

of the Q-factor and N is the number of taps. Figure 3.1 shows the RF response of an

incoherent filter with 40-dB extinction ratio based on 18 taps with appropriate weighting,

3.1. Introduction 27

obtained by spectrum slicing with 18 FBGs having a separation of ~1 cm [94]. The central

passband frequency and the FWHM are 9.9 GHz and 0.8 GHz, resulting in Q = 12.4 (not 18

because of the weighting).

In non-recursive filters the number of the produced taps equals the number of

delay-forming elements (most often, FBGs). In order to obtain a large number of taps with a

small number of elements, one should use recursive filters. In these filters, a part of the signal

taken from the filter output is supplied back to the input thus creating a long sequence of

pulses. Amplification modules such as EDFAs can also be used in order to increase the

number of output taps. Q factors above 3000 have been achieved by Ortega et al. by

combining recursive filters (FBGs-EDFA) with small FSR with an additional filter based on

an electro-optic modulator (actually an electrical filter) [95].

Most of the progress in the incoherent microwave optical filtering has been observed in

the recent years in the field of so-called notch filters. In contrast to passband filters (including

high-Q filters), a notch filter is designed not to select a certain bandwidth from an aggregate

signal spectrum, but rather to suppress the desired range of frequencies. In some sense, the

notch filters are similar to band rejection filters with a relatively narrow (as a rule, a few

percent of the notch frequency) bandwidth. A simple non-tuneable notch filters consists of

two delayed copies of the modulated optical signal incoherently summed at the photodetector

[96]. In order to achieve a high rejection at the notch frequency, the amplitudes of the two

delayed signal copies should match with high accuracy.

The principle of incoherent summing does not directly allow for the accomplishment of

negative coefficients: it means that it is not possible to subtract some taps from the others.

This fact imposes limitations upon the shape of the achievable frequency responses. In

particular, using only positive coefficients, it is not possible to obtain high-pass filters or to

achieve, e.g., flat top filters. However, there are a few techniques rendering possible to obtain

negative taps in incoherent microwave optical filters. One of the ideas is to use a cross-gain

modulation in a semiconductor optical amplifier [97]. A notable feature of the two-tap filter

demonstrated in Ref. [97] is the null transmission at zero frequency, which is a direct

consequence of using a negative tap. Another idea is to use the differential detection [98, 99].

The idea is to split the positive and negative taps in the optical domain and use separate

photodetectors (inverting and non-inverting ones) for detecting the two sequences. The

responses are then summed in the electrical domain.


3.1.2 Coherent filtering

The incoherent summing of modulated optical signals at the photodetector allows for creating

positive and negative tap coefficients in the impulse response of a microwave optical filter.

However, a much greater flexibility can be achieved if the coefficients can be made complex.

This can be done by utilising information about the phase of the optical carrier. For doing this,

one should employ an optical source with the coherence length substantially longer than the

optical path lengths between the delayed copies of the modulated signal. In this case, the

summation of the signals at the photodetector is performed coherently [100]. Though very

flexible, this approach imposes extremely strict requirements of the system stability and

therefore has not been widely exploited in microwave optical filtering - in fact, there are very

few publications on the subject. In the example demonstrated in Ref. [100], a set of

waveguides is implemented on a single silicon substrate with a temperature maintained

within 0.1 oC.

An approach of coherent filtering is the direct optical filtering, proposed and

demonstrated a few years ago [101-103]. It consists of using highly spectrally selective

elements, FBGs, to select and isolate the double side bands corresponding to a certain

microwave signals and the optical carrier on which it was transferred.

(a) (b)

Figure 3.2. (a) Direct filter based on three coupled integrated ring resonators [101]. (b)

Principle of functioning and the microwave response of the presented filter [101].

Direct optical filtering using three coupled integrated ring resonators is shown in Fig.

3.2(a) [101]. The coupled resonator structure forms a series of periodic transmission peaks in

the optical domain (with the FSR of 7.5 GHz). The principle of functioning of the device as a

microwave filter as well as its microwave frequency response is given in Fig. 3.2(b). The

3.1. Introduction 29

wavelength of the optical carrier is made to be coincident with one of the maxima in the

optical transmission response of the filter. Varying the modulation frequency, one changes

the separation of the sidebands from the carrier. The modulation depth of the optical signal

after the filter depends upon whether the sidebands are within the transparency bands of the

resonator or not (see Fig. 3.2(b)), which is correspondingly reproduced in the microwave

response of the filter. The microwave response shown in Fig. 3.2(b) has the FWHM of ~1.2

GHz centred at 7.5 GHz, resulting in Q = 6.25.

Other configurations of direct microwave optical filtering have been demonstrated in

Ref. [102, 103]. In Ref. [102], the optical filter consists of four FBGs working in

transmission, creating a tuneable quasi-passband microwave response. Fig. 3.3(a) shows the

principle of this filter. Four temperature-stabilising boxes are used for stable functioning of

the filter and also for tuning the filter to achieve a wide range of microwave frequency

responses. One big advantage of this filter is the absence of an optical circulator, which

simplifies the optical part of the filter and makes it potentially cheap. In Ref. [103], the

direct microwave optical filtering is demonstrated by a superimposed FBG used in

reflection [104]. The principle of this filter is illustrated in Fig. 3.3(b). The ability of the

filter to remove out of band signals is demonstrated. Transmission over long distances after

filtering is possible with negligible deterioration of the signal quality. Contrary to the filter

working in transmission reported in Ref. [102], this filter is not able to tune the passband

central frequency and only little tuning of the spectral shape is possible. However, the

high-pass tail of the microwave response of the filter working in transmission [102]does not

exist in the filter with a grating used in reflection [103].

(a) (b)

Figure 3.3. Principles of the direct microwave optical filtering using (a) four FBGs working

in transmission [102] and (b) a superimposed grating in reflection [103].


3.1.3 Comparison between incoherent and direct optical filtering of

microwave signals

The FSR of nearly all the incoherent filters is limited by the coherence length of the optical

source, which should be shorter than optical paths existing in the filter (for a single

wavelength source configuration). The main implication of this fact is that, although their Q

factors can be very large, incoherent filters are usually more suitable for relatively moderate

passband frequencies (around a few GHz or below). Direct microwave optical filters have

no limitations on the optical source coherence length and are therefore more easily

configured for large passband frequencies. In fact, for direct optical filters, the larger the

operational frequency, the easier it becomes since the constraints on the filter design are

increasingly relaxed.

In the moderate frequency range, microwave optical filters have to compete with

relatively low-cost microwave electrical filters with rather good characteristics and large Q

factors. Most incoherent filters operating in this frequency range are usually very complex

and therefore expensive. Thus, expect for very specific cases where microwave electrical

filters might not be available (for example in terms of tuneability), it is unclear whether

incoherent optical filters of microwave signals could have a substantial commercial impact.

As explained above, incoherent filters are used exactly in the same way as microwave

electrical filters, i.e. they have both electrical input and output, optics being used

in-between just for the actual filtering operation. These filters hence bear the cost for the

electrical to optical and optical to electrical conversions which is, as everyone knows, far

from being negligible. By contrast, direct optical filters have both optical input and output

and are therefore especially suitable when microwave signals which need to be filtered are

already in the optical domain, i.e. put on an optical carrier. This implies that electrical to

optical and optical to electrical conversions are not required, making the direct optical

filters potentially cost effective.

Incoherent filters using a combination of positive and negative taps give the possibility

to design the shape of the microwave responses of the optical filter. However, the finer the

design, the more taps are required. In the direct optical filtering, the shape of the microwave

(or MMW) response is directly given by the spectral shape of the optical filter used. Since

long and complex-structured FBGs can presently be realized, there is a large degree of

freedom in designing such filters. However, further efforts in the development of direct

3.2. Direct microwave optical filtering with direct detection 31

optical filters of microwave signals need to be concentrated in achieving larger Q-factors

than what early demonstrations have shown. One promising direction is the use of

phase-shifted strong chirped FBGs. Also, direct optical filters require a very large spectral

accuracy in the adjustment of the filter with regard to the optical carrier bearing the

microwave signals to be filtered. In practice, automatic adjustment can be realized with

accuracy below 1 pm which is in the range of 100 MHz at 1550 nm.

3.2 Direct microwave optical filtering with direct detection

In direct optical filtering of microwave signals, the general case is that both optical carrier

and side-bands corresponding to the microwave carriers are preserved after filtering and

further used in the optical domain. However, when the microwave signals need to be

detected, i.e. reconverted to the electrical domain, the existence of the optical carrier

requires the down-conversion of the high-frequency (HF) signal to an intermediate

frequency (IF) signal in the detection part. This makes the detection system cumbersome

and therefore expensive. One solution, demonstrated in paper Ⅰ, is hence to remove the

optical carrier and to combine direct optical filtering with direct detection, which simplifies

the detection setup and significantly reduces its cost.

The direct detection requires the suppression of the optical carrier by the optical filter

(together with all undesired spectral components). Only the two sidebands (20 GHz in

Paper Ⅰ) of the optical carrier are supplied to the photodetector with a good extinction

ratio. The basic principle of the direct detection for the direct optical filtering of microwave

signals using a superimposed FBG in reflection is illustrated in Fig. 3.4. Two

Hamming-apodized gratings were written at the same location in a standard SMF with a

length of 10 cm. The reflection spectrum of the gratings is shown in Fig. 3.5. They have

identical spectral characteristics but different resonance frequencies and work in reflection

as passband filters for the sidebands of the microwave subcarrier to be processed. The

FWHM of the two gratings is ~16 pm corresponding to ~2 GHz. The pass function of the

filter in the microwave domain is hence expected to be about 2 GHz and a full stop effect

outside this 2 GHz range. The reflectivity of each peak in the grating is about 45% and the

wavelength separation between the two reflection peaks is about 320 pm corresponding to

40 GHz (i.e. for filtering at the subcarrier frequency of 20 GHz). The extinction ratio of the

grating is ~25 dB, guaranteeing the suppression of the optical carrier. Perfect co-location

and close to identical characteristics of the gratings imply that a negligible delay difference


between the side modes is expected after reflection from the grating filter.

Figure 3.4. Principle of the direct filtering combined with direct detection.

1534.4 1534.5 1534.6 1534.7 1534.8 1534.9

-30

-25

-20

-15

-10

-5

0

Ref

lect

ivity

, dB

Wavelength, nm

Figure 3.5. Reflection spectrum of a single double-peaked superimposed FBG.

The ability of the filter to remove out-of-band signals is tested by measuring digital

signals 155 MBit/s placed at a microwave subcarrier. Fig. 3.6 depicts the experimental

setup for bit error rate (BER) measurement. Detailed description of the functions of each

component is described in Paper Ⅰ. Contrary to the experimental setup in Ref. [103], a

local oscillator (LO), a mixer and an envelope detector are not needed in the technique

presented here. Moreover, a low-speed photodiode can be used here, additionally reducing

the cost.

The digital 155 MBit/s payload with a pseudorandom binary sequence of 231-1 was

generated by the BER tester and then used to modulate in amplitude an electrical signal

whose frequency was varied from 18.4 to 21.6 GHz. This electrical signal was then fed to

Optical carrier

Frequency

Ref

lect

ivity

Subcarrier

Modulation sidebands

3.2. Direct microwave optical filtering with direct detection 33

the optical modulator. The optical wavelength was set to the optimal 1534.64 nm

corresponding to the centre wavelength between the two reflection peaks. The gratings were

put in a temperature-stabilizing box for good function of the grating filter. The optical

power and the optical signal to noise ratio before being reflected by the FBG kept constant

as the frequency was changed. BER measurements and eye-diagrams as a function of the

frequency have been recorded. The results of these measurements at room temperature are

shown in Fig. 3.7 and 3.8.

Figure 3.6. Experimental setup used for the demonstration of the direct microwave optical

filtering combined with direct detection. IF: intermediate frequency; RF: radio frequency;

LO: local oscillator; PC: polarization controller; EDFA: erbium doped fibre amplifier; LPF:

low-pass filter.

As one can see, the passband width of the filter is about 2 GHz as expected from the

optical spectrum of the grating. The BER performance in the passband of the filter shown

by the flat part of the curve corresponds to the BER of 1.2×10-11 or better. This performance

was considered as excellent. Outside the passband the quality of the signal quickly

Pattern generator

Laser sourcetunable 1550-CW

Modulatorbandwidth 30GHz

Photodiode

Double sideband FBG

EDFA

PC

IF RF

18-26 GHz

STM-1LPF

STM-16LPF

Errordetector

Clock

Mixer

LO

Circulator

Electrical

Optical


deteriorates demonstrating an almost on-off behaviour of the filter. The eye-diagrams are of

good quality within the passband and the eyes quickly close outside the passband.

Additional experiments were realized by using only one peak of the double-peaked

grating also demonstrated efficient filtering of the signal with the bandwidth of 2 GHz,

matching well the optical bandwidth of this single peak. These experiments proved the

possibility of using the direct detection technique for optical single sideband modulated

signals and also for tuneable filtering by tuning the central wavelength of the used

reflection peak with regards to the optical carrier.

18.5 19.0 19.5 20.0 20.5 21.0 21.5

-10

-8

-6

-4

-2

0

log

BER

Frequency, GHz

Figure 3.7. BER measurements as a function of microwave signal frequency.

(a) (b)

(c) (d)

Figure 3.8. Eye-diagrams of the detected signal at a number of frequencies within and outside the passband of the filter: (a) at 19.5 GHz, (b) at 20 GHz, (c) at 21 GHz, (d) 21.5 GHz.

3.3. Perspective and applications 35

3.3 Perspective and applications

In the technique presented here, the optical carrier is spared for other subcarriers it may

carry and the composite signals (the optical carrier with unfiltered subcarriers) can be

transported further in the optical link. Therefore, the FBG filter together with the direct

detection scheme can be applied in RoF systems comprising more than one RF signal.

Because of the narrow bandwidth of the superimposed FBG presented here, cascaded

superimposed FBGs (with the same central wavelength but different wavelength

separations of the two peaks, such as 320 pm, 384 pm and 448 pm) can be used to filter

corresponding RF signals (such as 20 GHz, 24 GHz and 28 GHz, respectively). However,

the narrow bandwidth of the filter also limits the maximum data rate, which would be 800

MBit/s in the case presented here.

The direct optical filtering technique has been used for simultaneous up-conversion of

ultra-wide band (UWB) signals and filtering, as demonstrated in Papers Ⅹ-Ⅻ. In Paper Ⅹ,

the UWB signals were up-converted to 40 GHz range and the extinction ratio of the

suppressed optical carrier was improved by 35 dB using a 10-cm long Hamming-apodized

FBG in transmission. In Papers Ⅺ and Ⅻ, the UWB signals was up-converted to 60 GHz

and an extinction ratio of 28 dB was achieved at double frequency of the local carrier (60

GHz) by using a 7-cm long Hamming-apodized FBG in transmission. Due to the grating,

the remaining desired signals at 56.568 GHz is attenuated by only 2.4 dB and the undesired

side band is suppressed by more than 23 dB.

37

Chapter 4

High Speed Switching

A fibre with side-holes filled with internal electrodes can be used to electrically control the

light propagating in the fibre waveguide. Besides electro-optical switching and modulation,

nanosecond polarization switching has recently been demonstrated [18]. A FBG written in

such fibres is an excellent tool to follow the evolution of the refractive index experienced

by the light in the polarization parallel and orthogonal to the direction of the holes. In this

chapter, the research on two 4-cm long Hamming-apodized FBGs written on side-hole

fibres with internal electrodes (Papers Ⅱ–Ⅵ) is described. Nanosecond high current pulses

cause the quasi instantaneous expansion of the metallic electrodes. This increases the

birefringence of the fibre and a full on-off and off-on switching of the gratings with a

response time of ~29 ns have been achieved. From the spectral response of the grating, the

time evolution of the mechanical stress and the heat diffusion across the fibre cross section

has been studied. Numerical simulations give accurate description of the experimental

results. Without the electrical pulses excitation, the temperature dependence of the grating

characteristics is also presented.

4.1 Fabrication of the FBG component

The 125-μm thick silicate side-hole fibre, drawn at Acreo FibreLab, has two 28.8-μm diameter holes running parallel to the core. The core-hole separation (edge-to-edge) is 13.4

μm. The side-hole fibre has characteristics similar to those of a standard telecom fibre (core

38 Chapter 4. High Speed Switching

diameter 8.4 μm and Δn = 0.0056). Fig. 4.1 shows the cross section of the side-hole fibre used here. Here, the direction parallel to the two holes is defined as the x-axis; the one

perpendicular to the two holes is the y-axis.

Pieces of side-hole fibre were filled with ~8-cm long metal columns which occupied

the entire cross-section of the holes. Eutectic alloy Bi47Sn53 (melting temperature 137°C)

was pushed into the fibres in the molten state using a chamber with high pressure, and

employed at room temperature as solid electrodes, as shown in Fig. 4.2 [105]. The length of

the alloy-filled columns and the core-hole separation were such that metal-induced loss was

experimentally measured to be <0.2 dB at 1.55 μm and the polarization-dependent loss <0.1 dB. Both ends of the metal-filled fibre were free from metal to allow for convenient

low-loss splicing.

8.4 um28.8 um

13.4 um

125 um

28.8 um

13.4 um

x

y

Figure 4.1. Cross section of the side-hole fibre (SEM picture).

Figure 4.2. Filling a side-hole fibre with the alloy BiSn.

4.1. Fabrication of the FBG component 39

The metal-filled fibres were hydrogen loaded for 2 weeks at 150 bars, stripped from

the acrylate coating and exposed to UV beam in a FBG writing set-up. Frequency doubled

Ar+ laser was used as a source of UV radiation. In the work reported here, results obtained

from two 4-cm long Hamming-apodized FBG written along the middle section of

metal-filled fibres are presented. The gratings were annealed in the oven at 100 °C for 12

hours.

After annealing the gratings, the fibres were side polished to electrically access one of

the two internal electrodes at two points separated by 7 cm. A 20-μm thick gold-plated tungsten wire was inserted by ~5 mm into the electrode while melting the alloy locally to

ensure a robustly bonded electrical connection, as shown in Fig. 4.3. Then, the pieces of

fibre were mounted on an aluminium (Al) substrate and covered with heat conductive,

electrically isolating epoxy. Heat dissipation into the substrate allows switching the devices

at rates of a few kHz without significant drift of polarization or wavelength. The resistance

of the components was typically ~43 Ω for a BiSn device with 7 cm active length. This

value implies in a small impedance mismatch between the device and the 50 Ω coaxial

cable transporting the high voltage (i.e., current) pulses. Electrical SMA contacts were used

and a 0.1 Ω resistive probe was connected in series with the 43 Ω load for monitoring

purposes. Fig. 4.4 shows the photograph of the grating component.

Figure 4.3. Illustration of the fibre with electrodes contacted with 20-µm diameter Tungsten

wires.

Figure 4.4. Photograph of the grating component.

4-cm long FBG

7-cm long electrode sections filled with BiSn

0.1Ω resistance


4.2 Static experiment

Due to the intrinsic birefringence of the side-hole fibre, each grating has two reflection

peaks at different wavelength corresponding to the different effective refractive indices for

the two orthogonal polarization modes in this fibre [106]. Fig. 4.5 shows the reflection

spectrum of FBG2 at room temperature. The reflection peak at the shorter wavelength takes

place in the fast axis mode (x-axis or x-polarization, black curve in Fig. 4.5), the one at the

longer wavelength in the slow axis mode (y-axis or y-polarization, red curve in Fig. 4.5)

[107]. The identification of the axes will be discussed in the simulations below. The

information about the Bragg wavelengths of the two polarizations at room temperature, the

FWHMs and the reflectivities of two gratings are listed in Table 4.1. λx and λy denote the

Bragg wavelengths of the two polarizations for each grating. λy-λx is the Bragg wavelength

separation for the two polarizations. Δn is the refractive index difference between the two polarizations of each grating, roughly one order of magnitude lower than in typical HiBi

fibres. Δn is inferred from:

0

y xeffn n

λ λλ−

Δ = (4.1)

where λ0 = (λy+λx )/2 is the average value of the Bragg wavelengths of the two polarizations and neff = 1.445.

1547.20 1547.25 1547.30 1547.350.0

0.1

0.2

0.3

0.4

44 pm44 pm

y-polarization

Ref

lect

ion,

mW

Wavelength, nm

x-polarization43 pm

Figure 4.5. Reflection spectrum of FBG2 at room temperature.

4.2. Static experiment 41

λx

(nm)

λy

(nm)

λy-λx

(pm)

Δn FWHM

(pm)

Reflectivity

Max. index

modulation

FBG1 1546.302 1546.328 26 ~2.4×10-5 36 70% 2.7×10-5

FBG2 1547.280 1547.323 43 ~4.1×10-5 44 75% 3.1×10-5

Table 4.1. Parameters of the two gratings at room temperature.

4.2.1 Temperature dependence measurement

Figure 4.6. Setup for the temperature dependence measurement. TLS: tuneable light source;

PC: polarization controller; OSA: optical spectrum analyzer.

The experimental setup for temperature dependence measurement is depicted in Fig. 4.6.

The light source used was a tuneable light source (TLS) based on an external cavity diode

laser with linewidth < 1 pm, and synchronized with an optical spectrum analyzer (OSA) for

high resolution and high dynamic rage wavelength measurements. Different input

polarization states were selected with a polarization controller. The metal-filled fibre device

was put in a temperature-controllable oven. When the oven temperature increased, the two

internal electrodes were heated and expanded more. Consequently, the refractive indices for

both polarizations and the birefringence were observed to increase as the temperature

increased. λx and λy shifted to longer wavelengths and the splitting between them (λy-λx)

increased. Fig. 4.7 shows the temperature dependence of λx (black curve with squares), λy

(black curve with triangles) and λy-λx (blue curve with circles) of FBG1 in a steady-state situation. The insets are the reflection spectra of x- and y-polarization at ~26 °C and at ~69

°C , shown in black and red curve, respectively. From 26 °C to 69 °C, λy-λx increased from 26 pm to 87 pm. From Eq. 4.1, the refractive index difference can be calculated to increase

from ~2.4×10-5 to ~8.0×10-5. Such a state can also be reached by heating the fibre by

passing dc current in the electrode, instead of using an oven as in the characterization

TLS

PCGrating

OSA

Oven

Circulator


measurement presented here. Although the range of temperature reached is more limited

(since the heat deposited on the electrodes needs to spread over the whole fibre), heating the

electrodes with current allows for the study of the dynamics of the system. For example, the

regime can be studied when the electrodes expand because of the current but no heat

exchange has yet taken place, as discussed in section 4.3.

10 20 30 40 50 60 70 80 90

1546.4

1546.8

1547.2

1547.6

1548.0

1548.4

1548.8

1547.8 1547.9 1548.00.0

0.2

0.4

0.6

0.8

1.069 0C 87pm

Wavelength, nm

Refle

ctivi

ty, m

W

1546.28 1546.32 1546.360.0

0.2

0.4

0.6

0.8

1.0

Wavelength, nm

26pm26 0C

Refle

ctiv

ity, m

W

Wav

elen

gth

diff

eren

ce, p

m

Temperature, oC

Wav

elen

gth,

nm

λx λy λy−λx

20

40

60

80

100

120

Figure 4.7. Temperature dependence of Bragg resonance for x- and y-polarization of FBG1

in a steady-state situation. Insets: reflection spectra of x- and y-polarization of FBG1 at ~26

°C and at ~69 °C, respectively.

4.2.2 Discussion and calculation

One can note from Fig. 4.7 that the measured temperature sensitivity (~37.6 pm/°C) is

significantly larger than expected for a HiBi FBG fabricated in Panda fibre (~16.5 pm/°C)

[108]. The Bragg-reflected wavelength depends on the period of the grating and the

effective index change. By differentiating Eq. 2.1, one can find variation in Bragg centre

wavelength with respect to the effective refractive index, the grating period and their

variations..

2 2B eff effd n d dnλ = Λ + Λ (4.2)

Dividing Eq. 4.2 by Eq. 2.1, one gets the relative wavelength shift


effB

B eff

dnd dn

λλ

Λ= +

Λ (4.3)

This means that a shift in the Bragg centre wavelength can be achieved either by inducing a

gradient in the effective refractive index of the grating or by a change in the period of the

grating or both.

To find the effect of temperature, Eq. 4.2 can be differentiated with respect to

temperature T.

2( )effBeff

nn

T T Tλ ∂∂ ∂Λ

= Λ +∂ ∂ ∂

(4.4)

Thus, the whole thermal tuning scenario arises due to two mechanisms. The change in the

period of the grating due to thermal expansion or contraction and the refractive index

change due to the temperature change. Dividing Eq. 4.4 by Eq. 4.1, we get

1 1 1( )effB

B eff

nT n T Tλ

λ∂∂ ∂Λ

= +∂ ∂ Λ ∂

(4.5)

The quantity (1 )( )eff effn n T∂ ∂ represents the thermo-optic coefficient ξn whose value is

approximately 6.5×10-6 K-1 for this side-hole fibre. The term (1 )( )TΛ ∂Λ ∂ is the thermal

expansion coefficient α of the fibre, optimized to be 0.5×10-6 K-1 which will be discussed in the simulation section. It can be seen from Eq. 4.5 that the wavelength sensitivity is about ~

11 pm for a thermal gradient of 1 °C.

As discussed in Ref. [109], the temperature sensitivity can be improved by increasing

the contribution of thermal expansion factor in Eq. 4.5. This can be achieved by attaching

the FBG to a system having very-high effective thermal expansion coefficient. In the case

presented here, the FBG was written in a side-hole fibre with internal alloy and mounted on

the Al substrate, which exhibits significantly larger thermal expansion coefficient than the

one of usual silica fibres. In this case, the relative shift of the resonance wavelength of a

Bragg grating with respect to temperature can be expressed as:

1 ( ) ( )Bn F A eff S eff

B Tλ ξ α α α

λ∂

= + + +∂

(4.6)

where αF is the thermal expansion coefficient of the fibre; (αΑ)eff and (αS)eff are effective thermal expansion coefficient of the alloy and the Al substrate to the fibre, respectively. Eq.

4.6 can be rewritten as:


( ) ( )B n B F B A eff B S effT T T Tλ λ ξ λ α λ α λ αΔ = Δ + Δ + Δ + Δ (4.7)

n F A Sλ λ λ λ λΔ = Δ + Δ + Δ + Δ (4.8)

where Δλn = λBξnΔT, ΔλF = λBαF ΔT, ΔλA = λB(αΑ)effΔT and ΔλS = λB(αΑ)effΔT are the wavelength shifts due to refractive index increase, due to the length expansion of the fibre

itself, due to the alloy dilation and due to the length expansion of the Al substrate. An

estimate of the various contributions follows:

(1) Calculation of the effective thermal expansion coefficient of the alloy (αΑ)eff

A temperature change of ΔT causes the alloy expansion:

A A AL L TαΔ = Δ (4.9)

where αΑ is the thermal expansion coefficient of the used alloy (BiSn) and LA is the alloy length.

Newton’s third law gives the relation:

alloy fiber fiber alloyF F→ →= (4.10)

EAF kx xL

= − = − (4.11)

From Eq. 4.10 and 4.11, one gets:

A A A F F F

A F

L E A L E AL L

Δ Δ− = − (4.12)

where EA and EF are the Young’s modulus of the alloy and the fibre; AA and AF are the cross

sectional area of the alloy and of the fibre.

FLΔ is the fibre elongation due to the expansion of the alloy:

( )F A eff FL L TαΔ = Δ (4.13)

where LF is the effective length of the fibre.

From Eq. 4.9, 4.12 and 4.13 one can get

( ) A AA eff A

F F

E AE A

α α= (4.14)


(2) Calculation of the effective thermal expansion coefficient of the substrate

(αS)eff

Assuming that the Al substrate was perfectly attached to the fibre, (αS)eff can be calculated from:.

( ) F F F S S SS eff

F F S S

E A E AE A E A

α αα +=

+ (4.15)

where ES and AS are the Young’s modulus and the cross sectional area of the Al substrate.

Since AS is much larger than AF, Eq. 4.15 can be simplified as:

( )1

F FF S

S SS eff S

F F

S S

E AE A

E AE A

α αα α

+= ≈

+ (4.16)

(3) Calculations of Δλn, ΔλF, ΔλA and ΔλS

The parameters for calculation are listed in Table 4.2. From Eq. 4.14 and 4.16, we

obtain that (αΑ)eff = 1.1×10-6 and (αS)eff = 2.3×10-5. According to Eq. 4.7 and 4.8, the four

wavelength shifts can be calculated for a ~53 ºC temperature increase: Δλn = 0.533 nm, ΔλF, = 0.041 nm, ΔλA = 0.088 nm, ΔλS = 1.886 nm. One sees that the dominant contributions are the change in refractive index of the fibre and the expansion of the substrate (largest). The

total wavelength shift under these conditions is calculated to be 2.55 nm, which exceeds the

measured value from Fig. 4.7 of ~2.0 nm. The lower value measured is probably caused by

slippage between the substrate and the fibre [109].

Fibre Alloy (Bi47Sn53) Al substrate

Thermal expansion coeff. (K-1) αF = 0.5×10-6 αΑ =15.35×10-6 αS = 2.3×10-5

Young’s modulus (GPa) EF = 73 EA = 43 ES = 69

Cross sectional area (m2) AF = 0.4×10-7 AA = 0.5×10-8

Thermo-optic coeff.(K-1) ξn = 6.5×10-6

Bragg wavelength (nm) λB = 1547.3

Temperature increase (ºC) ΔT = 53

Table 4.2. Data of the parameters for calculation.


4.3 Dynamic experiment

4.3.1 Experimental setup

Two sources of electrical drive pulses were used for this experiment. One consists of a high

voltage dc power supply and a high speed switch. The switch is a low repetition rate spark

gap circuit delivering 2-10 ns rise-time current pulses with a maximum switching voltage of

3 kV. The pulse duration is determined by the cable length used and was varied in the

interval 10-241 ns in the following experiments. The other source is a home-made

semiconductor-based switch delivering up to 20 A with 30 ns rise-time at a repetition rate

<10 kHz and adjustable pulse length. A TLS with linewidth <1 pm, a circulator, a PC, a

3-dB splitter, a high-speed photodiode, an OSA and an oscilloscope completed the

experimental set-up, schematically illustrated in Fig. 4.8. Different input polarization states

were selected with the PC.

Figure 4.8. Experimental setup for the dynamic measurement. TLS: tunable external cavity

diode laser; PC: polarization controller; OSA: optical spectrum analyzer.

4.3.2 Result and analysis

Dynamic measurements were performed for the x-polarization by properly aligning the

input polarization and tuning TLS at different wavelengths within the grating spectrum (as

shown in Fig. 4.9(a)) before the applications of electrical pulses. Nanosecond high current

pulses heated the metal electrode and caused metal expansion rapidly. In consequence,

refractive index changed and the grating spectrum shifted. Following the electrical pulse

43Ω

Photo-diode

OscilloscopePC

Grating

OSA

Switch

High Voltage

TLS

Circulator

Splitter

OpticalElectrical

0.1Ω

4.3. Dynamic experiment 47

excitation, different switching traces with nanoseconds response time can be measured by

the oscilloscope. Fig. 4.9(b) shows the traces at different probe wavelengths for

x-polarization under the excitation of electrical pulses from the second pulse generator.

Each trace recorded the changes in reflectivity as a function of time at a certain probe

wavelength. For example, the reflectivity increased when the TLS was tuned at the short

wavelength side of the Bragg peak (such as at point A or point B), while it decreased when

the probe wavelength was longer than the grating reflection peak (such as at point F or

point G). From the traces in Fig. 4.9(b), we can conclude that the grating spectrum of the

x-polarization shifted to shorter wavelength (blue-shift) under the electrical pulse excitation.

It implies that the refractive index of x-polarization reduces due to the metal expansion.

1547.20 1547.25 1547.30 1547.35Wavelength, nm

K

J

I

H

FE

DC

BA

y-polarization

Ref

lect

ion,

a.u

.

x-polarization

G

(a)

0 200 4000

10

20

30

G

FE

DC

BA

(b)x-polarization

Am

plitu

de o

f the

sign

al, m

V

Time, ns 0 200 400 600 800

0

10

20

30

Am

plitu

de o

f the

Sig

nal,

mV

Time, ns

(c)y-polarization

D

J

I

H

Figure 4.9. (a) Reflection spectra of the x- and y-polarization of FBG1 at room temperature.

The letters (A, B, C…J, K) denote the probe wavelengths of TLS before each measurement.

(b) Different switching traces at different probe wavelengths for x-polarization following

the electrical pulse excitation from the second pulse generator. (c) Different switching

traces at different probe wavelengths for y-polarization under the electrical excitation from

the first pulse generator.


Similarly, dynamic measurements were also carried out for the y-polarization. Fig.

4.9(c) shows several switching traces at different probe wavelengths within the grating

spectrum under the electrical excitation from the first pulse generator. From these traces, we

can find that the grating spectrum of the y-polarization experienced a positive but smaller

wavelength shift (red-shift), meaning that the refractive index of y-polarization increases

because of the metal expansion. Note that, the measured traces in Fig. 4.9(c) suffered more

noise than those in Fig. 4.9(b). Due to the faster rise time, only the first pulse generator was

continued to be used in the following measurements. Electrically driven FBG switching

experiments were carried out at room temperature.

When the TLS was tuned at the short wavelength side of the x-polarization of FBG1,

full off-on switching with a response time of ~29 ns was achieved, as illustrated in Fig. 4.10.

The return of the Bragg peak to longer wavelengths is followed on a microsecond time

scale in the inset of Fig. 4.10. Full on-off switching was also obtained with a response time

of ~29 ns, as shown in Fig. 4.11, by setting the TLS wavelength to match the Bragg

wavelength of the x-polarization of FBG1. The long term recovery of the grating, shown in

the inset of Fig. 4.11, is not monotonic. The applied electrical pulse for full switching off-on

and on-off is shown in Fig. 4.12. It was a typical pulse obtained from the first source with

the pulse width ~60 ns and the amplitude ~18 A at low-repetition rate (~20 Hz). The small

step on the right flank (at ~100 ns) and the undershoot observed after the main current pulse

are caused by impedance mismatch into 50 Ω.

-100 0 100 200 300 400 500

0

10

20

30

40

-100 0 100 200 3000

10

20

30

40

τfall = 40 μs

Am

plitu

de, m

V

Time, μs

τrise = 29 ns

Time, ns

Am

plitu

de, m

V

x-polarization

Figure 4.10. Full switching off-on is accomplished with a rise time of ~29 ns for the

x-polarization of FBG1. Inset: the time evolution of signal in microseconds.


-100 0 100 200 300 400 500

0

10

20

30

40

0 1 20

10

20

30

40

Am

plitu

de, m

V

Time, ms

τfall = 29 ns

Time, ns

Am

plitu

de, m

V

x-polarization

Figure 4.11. Full switching on-off is accomplished with a rise time of ~29 ns for the

x-polarization of FBG1. Inset: the time evolution of signal in microseconds.

-200 0 200 400 600

0

5

10

15

20

τFWHM = 60 ns

Cur

rent

, A

Time, ns Figure 4.12. An typical electrical pulse for full switching off-on and on-off.

The spectral response of the grating (such as the inset of Fig. 4.11) is useful to study

the time evolution of the mechanical stress and the heat diffusion across the fibre cross

section. The grating wavelength shifts can be divided into fast part (in tens of nanoseconds)

and slow part (in hundreds of microseconds). The former is due to mechanical stress

created during the electrical pulse action, while the latter is due to the relaxation of


mechanical stress and the slow increase of temperature in the core transferred from the

metal electrode after each electrical pulse.

Fig. 4.13 schematically illustrates the grating wavelength shift for x-polarization at

various times. For the fast part (from t0 to t1), a negative wavelength shift happens during

the electrical pulse excitation and reaches its maximum absolute value ( xMλΔ ) at instant t1,

similar to the electrical pulse duration. Then the slow part takes effect: the grating gradually

moves back with the decreasing mechanical stress and the increasing temperature in the

core. The grating passes its original spectral position at instant t2 (tens of microseconds)

and red-shift continues. The maximum positive xT Mλ +Δ is reached at instant t3, where the

core temperature is near its maximum. After several milliseconds (t4), the grating returns to

its original spectral position again, since the mechanical stress is released and the

component is back to room temperature. It should be mentioned that the identification of

the behavior schematically illustrated in Fig. 4.13 (and in Fig. 4.15 below) involved

experimentally examining the response of the component for various probe wavelengths

(similar to Fig. 4.9). Fig. 4.14 shows the time evolution of the signal reflected by the

x-polarization of FBG2 when the TLS was tuned to match the Bragg wavelength of the

x-polarization. The voltage and width of electrical pulse used for this measurement are 780

V and 241 ns. If t0 is set to 0, the values of t1, t2, t3 and t4 are ~ 241 ns, ~68 μs, 270 μs and ~ 5 ms, respectively.

Similarly, schematically diagram of the grating wavelength shift for y-polarization at

various times is illustrated in Fig. 4.15. For the fast part, a positive and smaller wavelength

shift happens during the electrical pulse duration and reaches its maximum ( yMλΔ ) at

instant t1. Then, the relaxation of the mechanical stress blue-shifts the spectrum of the

grating, while the increasing temperature in the core red-shifts the grating peak. In the

beginning, the former dominates and the Bragg wavelength moves back a little until instant

t2. Afterwards, the heat reaching the core dominates, so the grating continues to red-shift.

The maximum positive spectral shift ( yT Mλ +Δ ) occurs at instant t3. Finally, the grating also

returns to its original spectral position after instant t4. Fig. 4.16 shows a typical experiment

trace for the y-polarization of FBG2 when the TLS was tuned to the grating maximum

reflectivity of the y-polarization. The electrical pulse employed here was the same as that

used for the measurement shown in Fig. 4.14. If t0 is set to 0, the values of t1, t2, t3 and t4 are

~ 241 ns, ~16 μs, 190 μs and ~ 20 ms, respectively.


1547.15 1547.20 1547.25 1547.30

t4

Ref

lect

ion,

a.u

.

Wavelength, nm

ΔλxT+M

t3

t2

t1Δλ

xM

t0λTLS

x-polarization

Figure 4.13. Schematic illustration of the wavelength shift for x-polarization at various

times. The pink arrows indicate the TLS wavelength during the measurement.

Figure 4.14. Time evolution of the signal reflected by the x-polarization of FBG2 when the

TLS was tuned to match the Bragg wavelength of x-polarization.

-0.5 0.0 0.5 1.0 1.5 2.0

0

10

20

30

40 t0

t1

t2

t3

t4

Am

plitu

de, m

V

Time, ms

x-polarization

yx


1547.25 1547.30 1547.35Wavelength, nm

Ref

lect

ion,

a.u

.

ΔλyT+M

t2

ΔλyM

t1

t0λTLS

y-polarization

t4

t3

Figure 4.15. Schematic illustration of the grating wavelength shift for y-polarization at

various times. The pink arrows indicate the TLS wavelength during the measurement.

Figure 4.16. Time evolution of the signal reflected by the y-polarization of FBG2 when the

TLS was tuned to match the Bragg wavelength of y-polarization.

0.0 0.5 1.0 1.5

0

10

20

30

40 t0 t2 t1

t3

Am

plitu

de, m

V

Time, ms

t4

y-polarization

yx


The values of ΔλM and ΔλT+M depend on the pulse voltage and the pulse duration. For a

fixed pulse duration Δt = 241 ns, experimental data of the fast and slow wavelength shifts

for two polarizations of FBG2 ( xMλΔ , x

T Mλ +Δ , yMλΔ and y

T Mλ +Δ ) at different voltages are

shown with blue triangles, red triangles, blue squares and red squares in Fig. 4.17 (d),

4.17(a), 4.17(f) and 4.17(c), respectively. The blue and red curves in these figures are the

parabolic fits with the following expressions:

5 2( ) 6.5 10 ( )xM pm U Vλ −Δ = − × (4.17)

5 2( ) 4.5 10 ( )xT M pm U Vλ −

+Δ = + × (4.18)

5 2( ) 2.1 10 ( )yM pm U Vλ −Δ = + × (4.19)

5 2( ) 7.0 10 ( )yT M pm U Vλ −

+Δ = + × (4.20)

The absolute value of xMλΔ is ~3.2 times larger than y

MλΔ , while yT Mλ +Δ is larger than

xT Mλ +Δ .

Similarly, experimental data of the fast and slow wavelength shift for x-polarizations

of FBG2 ( xMλΔ and x

T Mλ +Δ ) for a fixed voltage pulse amplitude U = 1kV (20 A pulse) at

different voltages are shown with blue triangles and red triangles in Fig. 4.17(e) and 4.17(b),

respectively. The blue and red curves in the two figures are the linear fits with the following

expressions:

( ) 0.244 ( )xM pm t nsλΔ = − × Δ (4.21)

( ) 1.66 ( )xT M pm t nsλ +Δ = + × Δ (4.22)

One concludes that the fast and slow wavelength shifts depend quadratically on the

electrical pulse voltage and linearly on the pulse duration for both polarizations states, i.e.

depend linearly on the amount of energy deposited.


0

20

40

60Δt = 241 ns

ΔλxT+M

(a) 0

20

40

60

ΔλyM

ΔλyT+M

Δt = 241 ns

(c)

0 300 600 900 1200

-60

-40

-20

0(d)

Wav

elen

gth

shift

, pm

Voltage, V0 30 60 90 120

-60

-40

-20

0

ΔλxM

(e)

ΔλxM

Pulse duration, ns0 300 600 900 1200

0

20

40

60(f)

Voltage, V

ΔλxT+M

(b)

U = 1 kV

0

20

40

60

Figure 4.17. Experimental data and fit curves of the fast and slow wavelength shifts for two

polarizations of FBG2 for 241 ns electrical pulses (a, c, d, f) and for 1 kV electrical pulses

(b, e).

4.4 Numerical simulation

All the above phenomena can be described by the following partial differential equations

for heat diffusion and displacement in x and y-direction:

( ) )(tWtTCTk =

∂∂

+∇−⋅∇ ρ (4.23)

0331211 =⎥⎦

⎤⎢⎣

⎡⎟⎟⎠

⎞⎜⎜⎝

⎛∂∂

+∂∂

∂∂

+⎥⎦

⎤⎢⎣

⎡−

∂∂

+∂∂

∂∂

xv

yuC

ybT

yvC

xuC

x (4.24)

0221233 =⎥⎦

⎤⎢⎣

⎡−

∂∂

+∂∂

∂∂

+⎥⎦

⎤⎢⎣

⎡⎟⎟⎠

⎞⎜⎜⎝

⎛∂∂

+∂∂

∂∂ bT

yvC

xuC

yxv

yuC

x (4.25)

4.4. Numerical simulation 55

where k is the thermal conductivity, ρ is the density and C is the heat capacity of the corresponding area of the cross section (BiSn and silica), W(t) is the time and temperature

dependent power density, u and v are the displacements in the x-and y-directions,

respectively. Parameters G, C11, C12, C13, C22, C33 and b can be expressed by the following

equations:

( )( )μμ 211 −+=

EG (4.26)

( )μ−= 111 GC (4.27)

μGC =12 (4.28)

( )μ−= 122 GC (4.29)

( )2

2133

μ−=

GC (4.30)

( )μα += 1Gb (4.31)

where E is the Young modulus and μ is Poisson’s ratio.

In the experiment setup, 43.2-Ω component is connected to a 50-Ω RF cable and the resistance of the component changes with temperature. Taking into account both

temperature dependence of the component resistance and that of the impedance mismatch,

W(t) can be expressed by:

( ) ( ) ( )( )[ ]2

0

02

50114

+Δ+Δ+

=TRV

TRtUtWR

R

αα (4.32)

where U(t) is the rectangular pulse with amplitude U and duration τ, LDV4

2π= is the

volume of the BiSn cylinder (D is the diameter of the metal column and L is the length of

the metal column), R0 is the resistance of the component at 24 oC, αR is the temperature coefficient of the resistance of BiSn alloy.

The above partial differential equations can be solved by the finite element technique

using the commercial software FlexPDE. Fig. 4.18 illustrates the finite element mesh used

for the presented case.


Figure 4.18. Finite element mesh. Left electrode is active, right electrode is passive.

The solutions of this set of equations allowed finding the time-dependent strain

components in the x- and y-directions (εx and εy) of the fibre, the temperature distribution over the core and to calculate the FBG wavelength shifts in x- and y-polarizations

according to the following formula [110, 111]:

( ) ( )20

1 1 1 22x x x y xn

p p Tλ λ ε ε λ α ξΔ = − + + + Δ (4.33)

( ) ( )20

11 122y y y x yn

p p Tλ λ ε ε λ α ξΔ = − + + + Δ (4.34)

where: λx and λy are the initial Bragg wavelength for x- and y-polarizations of FBG2; n0 is the initial effective refractive index of the core; p11 and p22 are the photoelastic coefficients

of the fibre; α is the thermal expansion coefficient of silica; ξ = (1/n)(dn/dT) is the

thermo-optic coefficient of fused silica; ΔT is the increased temperature in the core. The first terms of Eq. 4.33 and 4.34 stand for the wavelength shift of both polarizations due to

mechanical stress in the core, while the second terms of two equations indicate the

wavelength shifts caused by the increased temperature in the core. The values of parameters

used in the simulations are listed in Table 4.3. Note that, the exact value of the heat capacity

and the thermal expansion coefficient of the Bi47Sn53 alloy and the Ge-doped silica fibre are


unknown. They are optimized during the simulation process for excellent fitting of the

simulation results to the experimental data in Fig. 4.17.

Initial Bragg wavelength (nm) λx = 1547.280, λy =1547.323

Hole diameter(μm) D = 28.8

Core diameter (μm) 8.4

Hole-core separation (μm) 13.4

Fibre diameter (μm) 125

Component resistance at 24 oC (Ω) R0 = 43.2

Length of the metal column (cm) L = 7

Temperature coefficient of the component resistance (K-1)

αR = 3.368×10-3

Refractive Index of silica at 1.55 μm 1.445

Pulse duration (ns) Δt = 241

Silica Bi47Sn53 alloy

Young modulus (GPa) E = 73 E = 43

Poisson ratio μ = 0.17 μ = 0.40

Thermal expansion coefficient (K-1) α = 5×10-7 α = 15.35×10-6

Density (kg/m3) ρ = 2200 ρ = 8560

Thermal conductivity (W/(m.K)) 1.38 19

Heat capacity (J/(kg.K)) C = 736 C = 167

Photoelastic constants of silica [112] p11=0.121, p22 = 0.270

Thermo-optic coefficient of silica (K-1) ξ = 6.5×10-6

Table 4.3. Parameters used for numerical simulation.


Stress at the BiSn/cladding boundary increases approximately linearly during the

electrical pulse action, and then gradually decreases with time. This stress is transferred to

the core area in ~14 μm/5720 m/s = 2.34 ns, i.e. almost instantaneously compared to the pulse duration 241 ns. However, the heat at the BiSn/cladding interface needs several

microseconds to reach the core area. As a result, the heat diffusion can be neglected during

the electrical pulse action. The fast wavelength shifts xMλΔ and y

MλΔ are determined by

the first terms of Eq. 4.33 and 4.34 at t = 241 ns. The choice C = 167 J/kgK and α = 15.5×10-6 K-1 for the alloy and C = 736 J/kgK for the fibre allowed for excellent fitting of

the numerical simulations (blue lines) of xMλΔ and y

MλΔ to the experimental data from

Fig. 4. 17(d) (blue triangles) and 4.17(f) (blue squares), as shown in Fig. 4.19.

As mentioned above, the slow wavelength shift is caused by the relaxation of

mechanical stress and the increased temperature in the core. It can be calculated from Eq.

4.33 and 4.34 for both polarizations. The simulations of xT Mλ +Δ accurately reproduce the

experiment measurements when the thermal expansion coefficient for the silica fibre is

optimized to the value α = 0.50×10-6 K-1, shown by the blue line and triangles in Fig. 4.20. The simulation of the wavelength shift for the y-polarization now has no free parameters for

adjustment. The result (red line in Fig. 4.20) fits the experimental data (red squares in Fig.

4.20) relatively well. The numerical predictions are smaller than the measured values. The

small discrepancy could possibly be attributed to a temperature variation taking place

during the measurement. For example, the difference between the calculated and measured

yT Mλ +Δ at the pulse voltage 1015 V corresponds to a temperature shift of 0.8 oC, well

within the temperature uncertainty in the laboratory, which could have been rising during

the measurement. The complete simulations here agree well with a simple quadratic fit to

the data for ΔλM and ΔλT+M (for both polarizations) as a function of the pulse voltage.


0 200 400 600 800 1000

-60

-40

-20

0

20

ΔλxM

Fast

wav

elen

gth

shift

, pm

Applied voltage, V

ΔλyM

Figure 4.19. Experimental data (blue triangles and red squares) and numeric simulations

(blue and red lines) of the fast Bragg wavelength shifts of FBG2 for a 241-ns electrical

pulse.

0 200 400 600 800 1000

0

20

40

60

80

ΔλxT+M

Slow

wav

leng

th sh

ift, p

m

Applied voltage, V

ΔλyT+M

Figure 4.20. Experimental data (blue triangles and red squares) and numeric simulations

(blue and red lines) of the slow Bragg wavelength shifts of FBG2 for a 241-ns electrical

pulse.


0 50 100 150 200 250 300-2x10-4

-1x10-4

01x10-4

2x10-4

3x10-4

(a)

εy

Stra

in

Time, ns

εx

0 50 100 150 200 250 300-8x10-5

-6x10-5

-4x10-5

-2x10-5

02x10-5 (b)

Δnx

Δny

Ref

ract

ive

inde

x ch

ange

Time, ns

Figure 4.21. (a) Strain of the two polarizations at the centre of the core. (b) Refractive index

change of the two polarizations at the centre of the core.

The simulations confirm that the core area is compressed (negative) in x-direction and

stretched (positive) in y-direction during the electrical pulse action, as shown in Fig. 4.21(a).

Rectangular 241-ns pulse was applied at t = 20 ns instead of t = 0 in order to overcome

some technical difficulties related to FlexPDE program. For the pulse with applied voltage

1080 V, εx = -1.655×10-4 and εy = 2.535×10-4 at t = 261 ns. This results in a decrease of refractive index in x-direction (-7.3×10-5) and a weaker increase of refractive index in

y-direction (+2.11×10-5), as shown in Fig. 4.21(b), corresponding to 75.8xM pmλΔ = − and

23.7yM pmλΔ = + . One can conclude that the fast axis mode is in the direction parallel to the

two holes (x-direction), and the slow axis mode is in the direction orthogonal to the two

holes (y-direction).

Under a 960-V electrical pulse excitation, the wavelength shift for the x-polarization

(Δλx) of FBG2 as a function of time is shown in Fig. 4.22. From this figure, one can find

that Δλx becomes negative (-57.5 pm) following the excitation of electrical pulses. Then it

equals zero at a calculated time t2 = ~56 μs, when the two terms of Eq. 4.33 compensate

each other. Subsequently, the second term dominates and Δλx becomes positive and

increases gradually due to the increasing ΔT. The shift reaches its maximum (+40.3 pm) at a

calculated time t3 = ~250 μs.

Similarly, Fig. 4.23 shows the wavelength shift for the y-polarization (Δλy) of FBG2 under electrical pulse excitation with different voltages. As shown in this figure, it is clear

that Δλy becomes positive during the electrical pulse action. After that, it reduces because the decay of the stress dominates over the rising effect of heat. At a calculated time t2 ~17


μs the latter starts to dominate and the shift increases again until it reaches the maximum

value at a calculated time t3 ~200 μs.

0 50 100 150 200 250 300-60

-40

-20

0

20

40

60

960 Vt3

wav

elen

gth

shift

, pm

Time, μs

x-polarization

t2

Figure 4.22. Wavelength shift for the x-polarization of FBG2 under a 960-V electrical pulse

excitation.

0 50 100 150 200 250 3000

10

20

30

40

50

60

70

80

t3

746 V

855 V

960 V

Wav

elen

gth

shift

, ps

Time, μs

1080 Vy-polarization

t2

Figure 4.23. Wavelength shift for the y-polarization of FBG2 under electrical pulse

excitation with different voltages.


Under different pulse excitation, the simulation result of the time evolution of the

temperature in the core centre is illustrated in Fig. 4.24. As shown in this figure, the heat in

the metal takes ~10 μs to reach the core and increases to the maximum at ~190-220 μs.

Note that the maximum ΔT in the metal is as high as ~90 ºC for a corresponding core temperature increase of only ~6.2 ºC.

0 50 100 150 200 250 3000

1

2

3

4

5

6

7

855 V

308 V422 V528 V644 V

746 V

960 V

ΔT,

0 C

Time, μs

1080 V

Figure 4.24. Simulation results of the time evolution of the temperature in the centre of the

core under different voltage pulse excitation.

63

Chapter 5

Switchable Dual-wavelength Fibre Lasers

FBGs are ideal components for fibre lasers due to their unique advantages such as ease of

use, low cost and their extreme and designable wavelength selectivity. This chapter presents

two methods to achieve dual-wavelength switching in a fibre laser based on FBG feedback (Paper Ⅶ-Ⅸ). The detailed behaviour of the dual-wavelength switching in these two fibre

lasers was experimentally studied and the underlying physical mechanisms are explained.

5.1 Raman erbium-doped fibre laser

As is well known, an EDF can be regarded as a homogenously broadened gain medium at

room temperature [19, 31, 113]. This property allows achieving wavelength-switching in

the EDF lasers due to the gain competition between different wavelengths. Overlapping

cavities can be used for multi-wavelength lasing and wavelength switching can be achieved

by controlling the loss or the gain in different cavities [20, 22]. Here, an overlapping cavity

is designed for lasing at two wavelengths (corresponding to the Bragg wavelengths of two

FBGs) and both an EDF pump (for the EDF gain medium) and a Raman pump (for the

Raman gain medium) are employed. Dual-wavelength switching can be achieved by

controlling the power of the Raman pump. As far as know, this proposed fibre laser was the

first one combining Raman amplification and EDF amplification to achieve

multiwavelength switching.

64 Chapter 5. Switchable Dual-wavelength Fibre Lasers


The schematic configuration of the proposed switchable dual-wavelength fibre laser is

shown in Fig. 5.1. A 1480 nm laser diode (LD) with a maximum power of ~ 250 mW is

used as the EDF pump LD1, and a 1467 nm LD with a maximum power of 350 mW is

employed as the Raman pump LD2. The length of the SMF section between the two FBGs

is 10 km. Sections of SMF, dispersion compensation fibre (DCF) and 4-m long EDF are the

gain fibres in the present structure. Two FBGs are used as wavelength selective components.

The central wavelengths, FWHMs and reflectivities are (1545.13 nm, 0.093 nm, 50.0%)

and (1559.98 nm, 0.098 nm, 51.3%) for FBG1 and FBG2, respectively. An optical

circulator (OC) and an isolator are used to ensure that the ring laser cavity is a oscillating

counterclockwise.

There are two lasing cavities with wavelengths λ1 = 1545.13 nm (around the EDF gain

peak) and λ2 = 1559.98 nm (around the Raman gain peak corresponding to pump

wavelength 1467 nm). The power of EDF pump LD1 is always fixed at a certain level, and

the power of Raman pump LD2 is used to control the output wavelength and power.

Figure 5.1. Schematic configuration of the Raman erbium-doped fibre lasers. LD: Laser

diode; WDM: Wavelength division multiplexing; EDF: Erbium-doped fibre; ISO: Isolator;

OC: Optical circulator; FBG: Fibre Bragg grating; SMF: Single mode fibre; DCF:

Dispersion compensation fibre.

5.1.2 Results and analysis

In the experiments, the power of LD1 is fixed at 150 mW. When LD2 is switched off, the

output spectrum of the laser is shown by the black solid line in Fig. 5.2, from which one can

see that the structure is lasing only at λ1 with a peak power of -11.9 dBm. However, when

the drive current of LD2 is switched on to 750 mA, the structure will be lasing only at λ2

with a peak power of 1.9 dBm (see the red dashed line in Fig. 5.2). The resolution and

5.1. Raman erbium-doped fibre laser 65

sensitivity of the OSA used for all the measurements in this chapter are 0.06 nm and -65

dBm, respectively. Experimental results have shown that the laser output is rather stable.

Thus, dual-wavelength switching is achieved in this fibre laser by controlling the drive

current of LD2 (i.e., the power of the Raman pump LD2).

Figure 5.2. Output spectra of the proposed fibre laser when the drive current of LD2 is

switched off (black solid line) or switched on to 750 mA (red dashed line). The power of

the EDF pump is fixed to 150 mW.

5.1.3 Characteristics of the dual-wavelength switching

Fig. 5.3 shows the detailed behaviour of the dual-wavelength switching as the drive current

of LD2 increases (the pump power of LD1 is still fixed at 150 mW). The variation of output

power at λ1 and λ2 are indicated by the lines shown by circles and triangles, respectively.

Obviously, there are three regions where the laser is in different operational modes. In

region A (with the drive current for LD2 below switching current Ia = 70 mA), only λ1 is

lasing and its output power remains around -11.9 dBm. In region B (with the drive current

for LD2 ranging from 70 mA to 690 mA), both λ1 and λ2 are lasing. In this region, as the

drive current of LD2 increases, the output power at λ2 increases while that at λ1 decreases.

In region C (with the drive current for LD2 above switching current Ib = 690 mA), only λ2

is lasing and its output power remains around 1.9 dBm.

The operation principle of the present switchable dual-wavelength fibre laser can be

explained as follows. When the Raman pump LD2 is off, only lasing at λ1 can be obtained

since the cavity loss at λ2 is larger than that at λ1 (due to the loss of the 10-km SMF) and the


EDF amplification will occur only at λ1 (due to the gain competition). When the power of

LD2 increases gradually, the Raman gain mainly contributes to the amplification at λ2 since

λ2 is much closer to the Raman gain peak wavelength than λ1. Thus, the structure starts

lasing at λ2 when the pump power of LD2 reaches a certain level. Consequently, the power

of the lasing at λ1 decreases as the pump power of LD2 increases since the lasing at λ2 will

reduce the population inversion in the EDF. When the pump power of LD2 increases

further, lasing at λ1 disappears finally.

Figure 5.3. Output powers at the two lasing wavelengths against the drive current for LD2

for a fixed 150 mW power of LD1.

Figure 5.4. Switching currents Ia and Ib for LD2 when the power of LD1 is set to different

values.

5.2. Tuneable and injection-switchable EDF laser 67

The wavelength switching has also been studied when the pump power of LD1 is fixed

at some other values. Experimental results show that the characteristics of the wavelength

switching are similar (except the switching currents Ia and Ib) when the pump power of LD1

is set to different values. Fig. 5.4 shows the switching currents Ia (black squares) and Ib

(black dots) when the pump power of LD1 is set from 20 mW to 160 mW with an

increment of 20mW. The switching current Ia decreases as the pump power of LD1

increases. This can be explained as follows. The cavity gain for λ2 to start lasing is a

constant, which is the sum of EDF gain and Raman gain. Thus, the larger EDF pump is set,

the smaller Raman gain is needed for λ2 to start lasing (corresponding to a smaller switching

current Ia). The variation of Ia under different pump powers of LD1 is quite small, while the

variation of Ib is quite large. As shown in Fig. 6.4, the switching current Ib increases almost

linearly with the pump power of LD1. This indicates that a larger drive current for LD2 is

needed in order to let λ2 win the gain competition over λ1 when the pump power of LD1

increases.

5.2 Tuneable and injection-switchable EDF laser

Recently, Dragic et.al have reported an injection-seeded erbium-doped fibre laser [114-116],

in which an injection technique was used to achieve a wavelength switchable fibre laser. A

tunable laser (outside the cavity of the fibre laser) is used as seeding light. The shared

cavity is for self-seeded lasing or injection-seeded lasing. Switching between the

self-seeded wavelength lasing and injection-seeded wavelength lasing can be achieved by

controlling the power of the injection laser. By using this technique, one of the switching

wavelengths can be tuned among a large range since a tuneable injection laser is employed.

In the following section, a tuneable and dual-wavelength switchable EDF laser of line

structure employed an injection technique is presented. The dual-wavelength switching is

achieved by controlling the power of the injection laser.



Figure 5.5. Schematic configuration of the injection-switchable fibre laser. SLR: Sagnac

loop reflector; LD: Laser diode; WDM: Wavelength-division multiplexing; EDF:

Erbium-doped fibre; FBG: Fibre Bragg grating; TLD: Tunable laser diode; ISO: Isolator;

OC: Optical Circulator.

The experimental setup of the presented fibre laser is schematically illustrated in Fig. 5.5. A

980 nm LD with a maximum power of ~150 mW is used as an EDF pump, and a tunable

(from 1500 nm to 1590 nm) laser diode (TLD) with a maximum power of about 10 mW is

employed as the injection laser. The line structure cavity is formed by a wideband sagnac

loop reflector and an FBG. The central wavelength, FWHM and reflectivity of the FBG are

1550.48 nm, 0.096 nm and 53.2%, respectively. The seeding light is injected into the cavity

through a circulator. A section of 4-m long EDF is employed as the gain medium.

Dual-wavelength switching is achieved by controlling the power of the injection laser. One

lasing wavelength corresponds to the Bragg wavelength of the FBG (can be tuned by

changing the temperature or stress) and the other lasing wavelength can be tuned from 1535

to 1590 nm (determined by the wavelength of the injection laser).

5.2.2 Results and analysis

In the experiment for the injection-switchable fibre laser, the power of the EDF pump LD is

fixed at 16.2 mW. When the injection laser TLD is switched off, the output spectrum of the

laser is shown by the black solid line in Fig. 5.6. At this state, the structure is lasing only at

λ1 = 1550.48 nm (equal to the central wavelength of the FBG) with a peak power of -4.88

dBm. When the injection laser TLD (with tuneable wavelength λ2; for this figure λ2 = 1555

nm) is switched to -6.0 dBm (or larger), the structure will be lasing only at λ2 = 1555 nm

with a peak power of -3.16 dBm (see the red dashed line in Fig. 5.6). In this case, the

injection light is amplified by a two-pass traveling-wave amplifier. Thus, the wavelength

5.2 Tuneable and injection-switchable EDF laser 69

switching can be achieved in this fibre laser by controlling the power of the injection laser.

Figure 5.6. Output spectra of the injection-switchable EDF laser when the injection laser is

switched off (black solid line) or switched on to -6 dBm (red dashed line). The power of the

EDF pump is fixed at 16.2 mW.

5.2.3 Characteristics of the dual-wavelength switching

Fig. 5.7 shows the detailed behaviour of the dual-wavelength switching as the power of the

injection laser TLD increases (the power of 980 nm LD is fixed at 16.2 mW). The output

powers at λ1 and λ2 are indicated by the lines connected with circles and triangles,

respectively. Obviously, when the injection laser is switched off, the proposed fibre laser is

simply an EDF laser with an output wavelength of λ1. However, when the injection laser is

switched on and its output power increases, there are two regions where the laser is in

different operational modes. In region A (with the TLD power below the switching power

Ps = -6.6 dBm), both λ1 and λ2 are lasing. In this region, as the TLD power increases, the

output power at λ2 increases while that at λ1 decreases. In region B (with the TLD power

larger than the switching power Ps), the lasing occurs only at λ2 and its output power

remains around -3.16 dBm. Fig. 5.8 shows a good linear relationship between the pump

power and the switching power Ps when the wavelength of the TLD is fixed at 1555 nm.

The operation principle of the present injection-switched EDF laser can be explained

as follows. When the injection laser TLD is switched off, only lasing at λ1 can occur since

the cavity of the line structure only exists at the Bragg wavelength λ1 of the FBG. When the

TLD is switched on, the seeding light is injected into the cavity and is amplified (traveling


with a round trip on the EDF). In this case, the seeding light will reduce the population

inversion required for lasing at λ1 (the cavity gain at λ1 decreases). Consequently, the output

power at λ1 decreases as the TLD power increases. When the power of the TLD is large

enough and approaches Ps , the emission stimulated by the seeding light at wavelength λ2

consumes most of the pump power (i.e., reduces the population inversion) and consequently

the cavity gain becomes less than the cavity loss at λ1 (i.e., the lasing at λ1 becomes

impossible). This result in a drastic drop of the power at λ1 and finally only the output at λ2

can be observed.

Figure 5.7. Output power at the two wavelengths as the power of the injection laser

increases when the power of the 980 nm pump LD is fixed at 16.2 mW.

Figure 5.8. Relationship between the switching power Ps and the power of the 980 nm

pump LD when the wavelength of the TLD is fixed to 1555 nm.

5.2 Tuneable and injection-switchable EDF laser 71

The characteristics of the dual-wavelength switching have been studied further. Fig.

5.9 shows the switching powers for different wavelengths of the TLD when the power of

the 980 nm pump LD is fixed at 16.2 mW (curve connected with squares) or 19.1 mW

(curve connected with circles). From this figure, one can see that the wavelength switching

can be achieved over a wide wavelength range (about 50 nm) for the TLD. The V-shape

curves indicate that the wavelength switching is easier when the TLD wavelength is around

1560 nm and a larger switching power is needed when the TLD wavelength is far away

from 1560 nm. The saturated gain spectrum of the (one round trip) EDFA structure (just

removing the FBG in Fig. 5.5) was also measured, when the pump power is 16.2mW. This

figure shows that the gain peak is around 1560 nm, explaining why the switching power is

minimal around 1560 nm. Furthermore, from Fig. 5.9, one can see that the difference in the

switching powers for the two lines is almost a constant (4.6dBm) for different TLD

wavelengths, which is due to the linear relationship between the switching power (in mW)

and the power of the 980 nm pump LD (as shown in Fig. 5.8)

Figure 5.9. Switching power as the wavelength of the TLD increases when the pump power

is fixed at 16.2 mW (curve connected with triangles) or 19.1 mW (curve connected with

circles). Inset: The saturated gain spectrum of the one round trip EDFA at different

wavelengths of the TLD when the power is fixed at 16.2 mW.


5.2.4 Measurement of the switching time

The transient switching response of the proposed laser has also been studied when the

power of the 980 nm pump LD is 16.2 mW and the TLD wavelength is fixed at 1548.6 nm

(due to the availability of a filter in the lab for measuring the laser output at a specific

wavelength). Two InGaAs Photodetectors (PDA400, provided by Thorlabs, Ltd. UK) were

employed to measure the time delay between the moment when the injection laser was

switched on and the moment when the lasing at λ1 disappears. Form the results shown in

Fig. 5.10, the time delay (i.e., the switching time) is measured to be less than 50 μs, indicating that the proposed laser has the ability of fast wavelength-switching.

Figure 5.10. Transient switching response (measured with photodetectors) of the proposed

fibre laser when the power of the 980 nm pump LD is 16.2 mW and the TLD wavelength is

fixed at 1548.6 nm.

73

Chapter 6

Conclusions and Future Works

This thesis focuses on the applications of fibre Bragg gratings to high speed wavelength

switching, direct microwave optical filtering and switchable dual-wavelength fibre lasers.

Linear FBGs were written in side-hole fibres with internal electrodes. The gratings

exhibit two reflection peaks due to the intrinsic birefringence of the used fibre, the shorter

one associated with the fast axis. Under the excitation of nanosecond high current electrical

pulses, wavelength switching with a response time of ~29 ns has been achieved in a FBG

component. As far as know, there is no published report from other people on nanosecond

tuning FBGs. Besides, it is found that the results obtained with the FBG switch allow for

elucidating the involved physics of fibres with metal electrodes expanding under dynamic

conditions, which often deviates from intuitive. Initially, only the observations in the first

200 ns are discussed, related to the mechanical effect. Nanosecond electrical pulses heat the

metal electrode and cause metal expansion on a nanosecond scale. The mechanical stress

compresses the core in the direction of the holes (fast axis direction), and at the same time it

stretches the core in the slow axis direction. This leads to a small increase of the index

experienced by light polarized perpendicular to the direction of the holes. In contrast, light

polarized along the direction of the holes experiences a larger and negative index shift. The

results are at first sight puzzling, but can be simulated with excellent agreement with the

experiments.

74 Chapter 6. Conclusions and Future Works

Microseconds later, the relaxation of the mechanical stress and the effect of heat

reaching the core red-shifts both Bragg peaks. After a couple of hundred microseconds, the

temperature of the core reaches its maximum and the increase of refractive indices for both

polarization modes. Once again, simulations are in good agreement with measurements.

Simulations also show that although the temperature of the metal can rise by 90 °C, the

increase at the core is limited to only 6.2 °C. It is found, perhaps not surprisingly, that both

the mechanical and thermal effects depend on the square of the pulse voltage (into the 50 Ω cable), and linearly on the electrical pulse duration.

Since direct optical filtering of microwave signals is still in its infancy and has not been

studied very extensively, a review of the more general field of the optical microwave

filtering has been presented. A general comparison has been made between direct optical

filtering and the more established incoherent filtering technique. The direct optical filtering

is mostly suitable for large microwave frequencies (basically MMWs) and is potentially

cost effective and should probably be able to meet particular requirements of future MMW

communication systems. The cost-effectiveness comes from the fact that these filters have

both optical input and output, which are fully compatible with optical transmission of

MMW signals (necessary at these frequencies to reduce the attenuation of their distribution).

The concept of direct detection proposed in this thesis contributes to a substantial cost

reduction. The direct microwave optical filtering technique combined with direct detection

has been demonstrated using a single double-peaked superimposed grating working in

reflection. The grating filter exhibits nearly on-off behaviour with a FWHM of ~2 GHz

through BER and eye-diagram measurements. The presented technique can be applied in

RoF systems comprising more than one RF signal. Thermal stabilization of grating

wavelength is necessary for good function of the grating filter.

Wavelength switchable fibre lasers have a variety of important applications and FBGs

are ideal wavelength selection components for fibre lasers. Two switchable

dual-wavelength erbium-doped fibre lasers based on FBG feedback were proposed and

demonstrated in this thesis. In one proposed fibre laser, an overlapping cavity for the two

lasing wavelengths and hybrid gain medium (Raman amplification and EDF amplification)

were employed. Dual-wavelength switching was achieved by controlling the Raman pump

power. As far as know, this fibre laser is the first to combine Raman amplification and EDF

amplification to achieve multiwavelength switching. Compared with other switchable

multi-wavelength fibre lasers, which were only based on EDF amplification, the present

75

structure is simple (requiring no polarization controllers of special fibre/device) and the

separation of the dual wavelengths can be very large (since the wavelength of Raman pump

can be arbitrary, which is impossible for EDF amplification). In the other fibre laser, an

injection technique was used and the dual-wavelength switching was controlled by the

power of the injection laser. A fibre SLR and a FBG form the line structure cavity.

Compared with a fibre ring laser, this laser has the advantages of tunability, stability, low

amplified spontaneous emission noise and high injection efficiency. The detailed

characteristics of the dual-wavelength switching in the two fibre lasers were experimentally

studied and corresponding principles were physically explained.

Possible future works include:

1. Long and complex-structured FBGs can presently be realized, there is a large

degree of freedom in designing such gratings for various applications. However, a

lot of efforts still need to be put on the quality and the yield of the fabrication.

2. The nanosecond switching of the linear FBG component can have potential

application in Q-switched fibre lasers, 2x2 fibre switching, high speed sensor

readout, gated spectroscopy, and so on.

3. Research on phase shifted gratings and chirped gratings written in side-hole fibres

with electrodes has already started. These components have potential applications

in dual-wavelength single-longitudinal mode fibre lasers, microwave optical signal

generators, wavelength sweeping and radar warning systems.

4. Further efforts in the development of direct optical filtering of microwave signals

can be concentrated in achieving larger Q-factors. One promising direction is the

use of phase-shifted strong chirped FBGs.

As a general conclusion, FBGs are important components for a large number of

applications in optical communications, fibre sensors, fibre lasers, and so on. It is difficult

to think of fibre-optic systems without FBGs as it is to think of bulk optics without the

familiar laboratory mirrors.

77

Chapter 7

Summary of Papers

Paper Ⅰ Direct Detection of Direct Optically Filtered Millimeter-wave Signals.

Good performance of a direct optical microwave filter combined with direct detection consisting of a single superimposed grating used in reflection has been assessed through BER measurement. In the passband with the centre frequency of 20 GHz and a FWHM of 2 GHz (as expected from the optical reflection spectrum of the grating), the BER is below 1.2×10-11, showing nearly on-off behaviour of the filter. Such an optical filter with cost effective detection scheme can be applied in RoF systems.

Contributions of the author: Part of the original idea, all the experiment work, and the first draft of the manuscript.

Paper Ⅱ High-speed control of fiber Bragg gratings.

FBGs were written in side-hole fibres with internal alloy electrodes. Nanosecond high current pulses cause metal expansion, increase birefringence and tune the gratings. Close to full on-off and off-on switching with a response time of ~60 ns was accomplished. Such a component has potential use in Q-switching fibre lasers.

Contributions of the author: Part of the original idea, most of the experiment work, and revised the manuscript.

Paper Ⅲ Nanosecond switching of fiber Bragg gratings.

A 4-cm long Hamming-apodized FBG was written in a twin-hole fibre with internal

78 Chapter 7. Summary of Papers

alloy electrodes. Temperature dependence measurements showed that the birefringence increased as the temperature increased, due to the gradual expansion of metal column. A larger temperature sensitivity was obtained due to the fibre was attached to the Al substrate with a much larger thermal expansion coefficient. Dynamic measurements for the two polarizations under the electrical pulse excitations were also carried out. Nanosecond high current pulses heated the metal electrode and caused metal expansion rapidly. In consequence, birefringence was increased and the grating was tuned. Full on-off and off-on switching with a response time of ~29 ns was achieved for the fast axis mode. The nanosecond FBG response is attributed to a mechanical effect, while heat transporting to the core is slow and determines the recovery time of the device after every drive pulse. It is believed that the short length, low loss, all-spliced high-speed devices described here can find useful application for Q-switching fibre lasers.

Contributions of the author: Part of the original idea, most of the experiment work, and the first draft of the manuscript.

Paper Ⅳ Physics of electrically switched fiber Bragg gratings.

In this Paper, we report on the physics of electrically switched FBGs written in a twin-hole fibre with internal electrodes. For the fast axis mode, wavelength shifts due to mechanical effects (in tens of nanoseconds) and heat (in many microseconds) depend quadratically on the electrical pulse voltage and linearly on pulse duration.


Paper Ⅴ Birefringence switching of Bragg gratings in fibers with internal electrodes.

A FBG was written in a side-hole fibre with internal metal alloy electrodes. The initial geometrical birefringence of this fibre gives rise to two Bragg resonances separated by 43 pm. Nanosecond rise time current pulses of up to 23 A were applied to the metal electrode, which heated and expanded rapidly. On a nanosecond scale, the fast Bragg wavelength shift of the fast axis mode is negative and that of the slow axis mode is positive and smaller, because of the mechanical perturbation. The fast change increased the peak separation to ~ 143 pm, corresponding to an increase in birefringence from 4.0×10-5 to 1.3×10-4. Microseconds later, the relaxation of the mechanical stress and the effect of heat reaching the core red-shift both Bragg peaks.

79

All wavelength shifts depend quadratically on the electrical pulse voltage. Numerical simulations were developed to accurately and quantitatively explain the experimental observations.


Paper Ⅵ High-speed switching in fibres with electrodes.

Fibres with holes are provided with metal electrodes that fill the entire cross-section of the holes. Refractive index change and birefringence result from applying high current pulses for time intervals in the nanosecond range. High-speed polarization rotation and fibre Bragg grating wavelength tuning can be accomplished. Such electrically driven components are totally spliced with low insertion loss. They have no moving parts, are less than 10 cm long and can be used in fibre laser cavities for Q-switching.

Contributions of the author: Part of the experiment work.

Paper Ⅶ Switchable dual-wavelength Raman erbium-doped fibre laser.

A FBG feedback fibre laser with both Raman and EDF pumps is proposed. Dual-wavelength switching is achieved by controlling the power of the Raman pump. The characteristics of the dual-wavelength switching under different EDF pump levels are studied experimentally, and the mechanism is explained physically. The output power and wavelength of the present switchable dual-wavelength fibre laser are repeatable and uniquely determined by the Raman pump power.

Contributions of the author: Part of the original idea, part of the experiment work, and revised the manuscript.

Paper Ⅷ Some switchable dual-wavelength fibre lasers based on fibre bragg grating feedback.

Two methods to achieve switchable dual-wavelength fibre lasers are proposed and the corresponding fibre lasers based on FBG feedback are demonstrated in this Paper. The first proposed fibre laser employs both Raman and EDF pumps and the dual-wavelength switching is achieved by controlling the power of the Raman pump. In the second fibre laser, an injection technique is used and the dual-wavelength switching is realized by controlling the power of the injection laser. The detailed

80 Chapter 7. Summary of Papers

behaviour of the dual-wavelength switching in the two fibre lasers is experimentally studied and the principles are explained physically.

Contributions of the author: Part of the original idea, part of the experiment work, and revised the manuscript.

Paper Ⅸ Tunable and injection-switchable erbium-doped fiber laser of line structure.

A tunable and injection-switchable EDF laser is proposed based on a line structure formed by a fibre Sagnac loop reflector and a FBG. Wavelength switching is achieved by controlling the power of the tunable injection laser. The self-seeded wavelength corresponding to the Bragg wavelength of the FBG can be tuned by e.g. heating the FBG and the injection wavelength can be tuned over a wide range of about 50 nm. The characteristics of the wavelength switching for different levels of the EDF pump power and different wavelengths of the injection laser are studied experimentally. The presented fibre laser has the advantages of tunability, stability, low amplified spontaneous emission noise and high injection efficiency as compared with a fibre ring laser. The switching time is measured to be less than 50 microsecond.

Contributions of the author: Revised the manuscript.

81

Bibliography

[1]. K. O. Hill, Y. Fujii, D. C. Johnson, and B. S. Kawasaki. Photosensitivity in optical fiber waveguides: Application to reflection filter fabrication, Appl. Phys. Lett., 32(10):647-649, 1978.

[2]. G. Meltz, W. W. Morey, and W. H. Glenn. Formation of Bragg gratings in optical fibers by a transverse holographic method, Opt. Lett., 14(15):823, 1989.

[3]. R. Kashyap. Fiber Bragg Gratings, Academic Press, 1999.

[4]. A. Othonos and K. Kalli. Fiber Bragg Gratings: Fundamentals and Applications in Telecommunications and Sensing, Artech House, 1999.

[5]. A. Othonos. Fiber Bragg gratings, Rev. Sci. Instrum., 68(12):4309-4341, 1997.

[6]. A. J. Seeds. Microwave photonics, IEEE Trans. Microw.Theory Tech., 50(3):877-87, 2002.

[7]. A. J. Seeds and K. J. Williams. Microwave photonics, J. Lightwave Technol., 24(12):4628-41, 2006.

[8]. H. M. Xie, P. Dabkiewicz, R. Ulrich, and K. Okamoto. Side-hole fiber for fiber-optic pressure sensing, Opt. Lett., 11(5):333-335, 1986.

[9]. M. N. Charasse, M. Turpin, and J. P. L. Pesant. Dynamic pressure sensing with a side-hole birefringent optical fiber, Opt. Lett., 16(13):1043-1045, 1991.

[10]. W. J. Bock, M. S. Nawrocka, W. Urbanczyk, and J. Rostkowski. Highly sensitive fiber-optic sensor for dynamic pressure measurements, in Proceedings of IEEE Conference on Instrumentation and Measurement Technology Conference (IMTC 2000) 591-594, 2000.

82 Bibliography

[11]. W. J. Bock, M. S. Nawrocka, and W. Urbanczyk. Highly sensitive fiber-optic sensor for dynamic pressure measurements, IEEE Trans. Instrum. Meas., 50(5):1085-1088, 2001.

[12]. S. Kreger, S. Calvert, and E. Udd. High pressure sensing using fiber Bragg gratings written in birefringent side hole fiber, Portland, OR, USA, 355-358, 2002.

[13]. C. Jewart, K. P. Chen, B. McMillen, M. M. Bails, S. P. Levitan, J. Canning, and I. V. Avdeev. Sensitivity enhancement of fiber Bragg gratings to transverse stress by using microstructural fibers, Opt. Lett., 31(15):2260-2262, 2006.

[14]. M. Turpin, M. Brevignon, J. P. Le Pesant, and O. Gaouditz. Interfero-polarimetric fiber optic sensor for both pressure and temperature measurement, in Optical Fiber Sensors Conference, 1992. 8th, 362-365, 1992.

[15]. R. J. Schroeder, T. Yamate, and E. Udd. High pressure and temperature sensing for the oil industry using fiber Bragg gratings written onto side hole single mode fiber, in Proceedings of SPIE - The International Society for Optical Engineering, Kyongju, South Korea, 42-45, 1999.

[16]. T. Yamate, R. T. Ramos, R. J. Schroeder, and E. Udd. Thermally insensitive pressure measurements up to 300 degree C using fiber Bragg gratings written onto side hole single mode fiber, Venice, Italy, 628-631, 2000.

[17]. E. Chmielewska, W. Urbańczyk, and W. J. Bock. Measurement of Pressure and Temperature Sensitivities of a Bragg Grating Imprinted in a Highly Birefringent Side-Hole Fiber Appl. Opt., 42(31):6284-6291, 2003.

[18]. H. Knape and W. Margulis. All-fiber polarization switch, Opt. Lett., 32(6):614-616, 2007.

[19]. W. Yong, T. Swee Chuan, Y. Jian, Y. Jian Ping, H. Liming, and N. Kok Ann. Wavelength-switching fiber laser for optically controlled phased-array antenna, Opt. Commun., 211(1-6):147-151, 2002.

[20]. Q. Mao and J. W. Y. Lit. Switchable multiwavelength erbium-doped fiber laser with cascaded fiber grating cavities, IEEE Photon. Technol. Lett., 14(5):612-614, 2002.

[21]. M. Qinghe and J. W. Y. Lit. Optical bistability in an L-band dual-wavelength erbium-doped fiber laser with overlapping cavities, IEEE Photon. Technol. Lett., 14(9):1252-1254, 2002.

83

[22]. C. Zhao, X. Yang, C. Lu, H. Ng Jun, X. Guo, P. R. Chaudhuri, and X. Dong. Switchable multiwavelength erbium-doped fiber lasers by using cascaded fiber Bragg gratings written in high birefringence fiber, Opt. Commun., 230(4-6):313-17, 2004.

[23]. L. Su and C. Lu. Wavelength-switching fibre laser based on multimode fibre Bragg gratings, Electron. Lett., 41(1):11-13, 2005.

[24]. F. Xinhuan, T. Hwa-Yaw, and P. K. A. Wai. Switchable multiwavelength erbium-doped fiber laser with a multimode fiber Bragg grating and photonic crystal fiber, IEEE Photon. Technol. Lett., 18(9):1088-1090, 2006.

[25]. M. Delgado-Pinar, J. Mora, A. Diez, J. L. Cruz, and M. V. Andres. Wavelength-switchable fiber laser using acoustic waves, IEEE Photon. Technol. Lett., 17(3):552-554, 2005.

[26]. L. Yong Wook and L. Byoungho. Wavelength-switchable erbium-doped fiber ring laser using spectral polarization-dependent loss element, IEEE Photon. Technol. Lett., 15(6):795-797, 2003.

[27]. G. Das and J. W. Y. Lit. Wavelength switching of a fiber laser with a Sagnac loop reflector, IEEE Photon. Technol. Lett., 16(1):60-62, 2004.

[28]. K. Young Jae and K. Dug Young. Electrically tunable dual-wavelength switching in a mutually injection-locked erbium-doped fiber ring laser and distributed-feedback laser diode, IEEE Photon. Technol. Lett., 17(4):762-764, 2005.

[29]. S. Hu, L. Zhan, Y. J. Song, W. Li, S. Y. Luo, and Y. X. Xia. Switchable multiwavelength erbium-doped fiber ring laser with a multisection high-birefringence fiber loop mirror, IEEE Photon. Technol. Lett., 17(7):1387-1389, 2005.

[30]. H. Young-Geun, K. Gilhwan, L. Ju Han, and L. Sang Bae. Wavelength spacing tunable multiwavelength fiber laser with lasing wavelength selectivity, Opt. Commun., 256(1-3):98-102, 2005.

[31]. X. Fang, D. N. Wang, F. W. Tong, W. Jin, P. K. A. Wai, and J. M. Gong. Multiwavelength erbium-doped fiber ring laser source with a hybrid gain medium, Opt. Commun., 228(4-6):295-301, 2003.

[32]. H. L. Liu, H. Y. Tam, W. H. Chung, P. K. A. Wai, and N. Sugimoto. Wavelength-switchable La-codoped bismuth-based erbium-doped fiber ring

84 Bibliography

laser, IEEE Photon. Technol. Lett., 17(5):986-988, 2005.

[33]. F. Xinhuan, L. Yange, F. Shenggui, Y. Shuzhong, and D. Xiaoyi. Switchable dual-wavelength ytterbium-doped fiber laser based on a few-mode fiber grating, IEEE Photon. Technol. Lett., 16(3):762-764, 2004.

[34]. Y. G. Han, J. H. Lee, S. H. Kim, and S. B. Lee. Tunable multi-wavelength Raman fibre laser based on fibre Bragg grating cavity with PMF Lyot-Sagnac filter, Electron. Lett., 40(23):1475-1476, 2004.

[35]. S. K. Sim, H. C. Lim, L. W. Lee, L. C. Chia, R. F. Wu, I. Cristiani, M. Rini, and V. Degiorgio. High-power cascaded Raman fibre laser using phosphosilicate fibre, Electron. Lett., 40(12):738-739, 2004.

[36]. J. H. Lee, Y. G. Han, Y. M. Chang, and S. B. Lee. Raman amplifier based long-distance, remote FBG strain sensor with EDF broadband source recycling residual Raman pump, Electron. Lett., 40(18):1106-1107, 2004.

[37]. B. S. Kawasaki, K. O. Hill, D. C. Johnson, and Y. Fujii. Narrow-band Bragg reflectors in optical fibers, Opt. Lett., 3(2):66, 1978.

[38]. D. K. W. Lam and B. K. Garside. CHARACTERIZATION OF SINGLE-MODE OPTICAL FIBER FILTERS, Appl. Opt., 20(3):440-445, 1981.

[39]. J. Stone. Photorefractivity in GeO[sub 2]-doped silica fibers, J. Appl. Phys., 62(11):4371-4374, 1987.

[40]. R. Kashyap, J. R. Armitage, R. Wyatt, S. T. A. D. S. T. Davey, and D. L. A. W. D. L. Williams. All-fibre narrowband reflection gratings at 1500 nm, Electron. Lett., 26(11):730-732, 1990.

[41]. K. O. Hill, B. Malo, F. Bilodeau, D. C. Johnson, T. F. Morse, A. Kilian, L. Reinhart, and O. Kyunghwan. Photosensitivity in Eu2+ :Al2O3-Doped-Core Fiber: Preliminary Results and Application to Mode Converters, in Optical Fiber Communication, PD3, 1991.

[42]. M. M. Broer, R. L. Cone, and J. R. Simpson. Ultraviolet-induced distributed-feedback gratings in Ce3+-doped silica optical fibers, Opt. Lett., 16(18):1391, 1991.

[43]. F. Bilodeau, D. C. Johnson, B. Malo, K. A. Vineberg, K. O. Hill, T. F. Morse, A. Kilian, and L. Reinhart. Ultraviolet-light photosensitivity in

85

Er3+-Ge-doped optical fiber, Opt. Lett., 15(20):1138, 1990.

[44]. D. L. Williams, B. J. Ainslie, J. R. Armitage, R. Kashyap, and R. Campbell. Enhanced UV photosensitivity in Boron codoped germanosilicate fibres, Electron. Lett., 29(1):45-47, 1993.

[45]. M. J. Yuen. Ultraviolet absorption studies of germanium silicate glasses, Appl. Opt., 21(1):136, 1982.

[46]. R. M. Atkins and V. Mizrahi. Observations of changes in UV absorption bands of singlemode germanosilicate core optical fibres on writing and thermally erasing refractive index gratings, Electron. Lett., 28(18):1743-1744, 1992.

[47]. D. L. Williams, S. T. Davey, R. Kashyap, J. R. A. A. J. R. Armitage, and B. J. A. A. B. J. Ainslie. Direct observation of UV induced bleaching of 240 nm absorption band in photosensitive germanosilicate glass fibres, Electron. Lett., 28(4):369-371, 1992.

[48]. D. P. Hand and P. S. J. Russell. Photoinduced refractive-index changes in germanosilicate fibers, Opt. Lett., 15(2):102, 1990.

[49]. R. M. Atkins, V. Mizrahi, and T. Erdogan. 248 nm induced vacuum UV spectral changes in optical fibre preform cores: support for a colour centre model of photosensitivity, Electron. Lett., 29(4):385-387, 1993.

[50]. K. D. Simmons, S. LaRochelle, V. Mizrahi, G. I. Stegeman, and D. L. Griscom. Correlation of defect centers with a wavelength-dependent photosensitive response in germania-doped silica optical fibers, Opt. Lett., 16(3):141, 1991.

[51]. T. E. Tsai, E. J. Friebele, and D. L. Griscom. Thermal stability of photoinduced gratings and paramagnetic centers in Ge- and Ge/P-doped silica optical fibers, Opt. Lett., 18(12):935, 1993.

[52]. P. Niay, P. Bernage, S. Legoubin, M. Douay, W. X. Xie, J. F. Bayon, T. Georges, M. Monerie, and B. Poumellec. Behaviour of spectral transmissions of Bragg gratings written in germania-doped fibres: writing and erasing experiments using pulsed or cw uv exposure, Opt. Commun., 113(1-3):176-192, 1994.

[53]. W. X. Xie, P. Niay, P. Bernage, M. Douay, J. F. Bayon, T. Georges, M. Monerie, and B. Poumellec. Experimental evidence of two types of photorefractive effects occuring during photoinscriptions of Bragg gratings within germanosilicate fibres, Opt. Commun., 104(1-3):185-195, 1993.

86 Bibliography

[54]. H. G. Limberger, P.-Y. Fonjallaz, and R. P. Salathe. Spectral characterisation of photoinduced high efficient Bragg gratings in standard telecommunication fibres, Electron. Lett., 29(1):47-49, 1993.

[55]. P. J. Lemaire, R. M. Atkins, V. Mizrahi, and W. A. A. R. W. A. Reed. High pressure H2 loading as a technique for achieving ultrahigh UV photosensitivity and thermal sensitivity in GeO2 doped optical fibres, Electron. Lett., 29(13):1191-1193, 1993.

[56]. R. M. Atkins, P. J. Lemaire, T. Erdogan, and V. A. M. V. Mizrahi. Mechanisms of enhanced UV photosensitivity via hydrogen loading in germanosilicate glasses, Electron. Lett., 29(14):1234-1235, 1993.

[57]. F. Bilodeau, B. Malo, J. Albert, D. C. Johnson, K. O. Hill, Y. Hibino, M. Abe, and M. Kawachi. Photosensitization of optical fiber and silica-on-silicon/silica waveguides, Opt. Lett., 18(12):953, 1993.

[58]. B. J. Eggleton, P. A. Krug, L. Poladian, K. A. Ahmed, and H. F. Liu. Experimental demonstration of compression of dispersed optical pulses by reflection from self-chirped optical fiber Bragg gratings, Opt. Lett., 19(12):877, 1994.

[59]. H. G. Limberger, P.-Y. Fonjallaz, P. Lambelet, R.-P. Salathe, C. Zimmer, and H. H. Gilgen. Optical low-coherence reflectometry (OLCR) characterization of efficient Bragg gratings in optical fiber, in Photosensitivity and Self-Organization in Optical Fibers and Waveguides, Quebec City, Canada, 272-283, 1993.

[60]. Z. A. Dana, M. Victor, E. Turan, and E. W. Alice. Phase-Mask Method for Volume Manufacturing of Fiber Phase Gratings, in Conference on Optical Fiber Communication/International Conference on Integrated Optics and Optical Fiber Communication, PD16, 1993.

[61]. K. O. Hill, B. Malo, F. Bilodeau, D. C. Johnson, and J. Albert. Bragg gratings fabricated in monomode photosensitive optical fiber by UV exposure through a phase mask, Appl. Phys. Lett., 62(10):1035-1037, 1993.

[62]. D. Z. Anderson, V. Mizrahi, T. Erdogan, and A. E. A. W. A. E. White. Production of in-fibre gratings using a diffractive optical element, Electron. Lett., 29(6):566-568, 1993.

[63]. J. Albert, S. Theriault, F. Bilodeau, D. C. J. A.-D. C. Johnson, K. O. H. A.-K.

87

O. Hill, P. S. A.-P. Sixt, and M. J. R. A.-M. J. Rooks. Minimization of phase errors in long fiber Bragg grating phase masks made using electron beam lithography, IEEE Photon. Technol. Lett., 8(10):1334-1336, 1996.

[64]. J. D. Prohaska, E. Snitzer, S. Rishton, and V. A. B. V. Boegli. Magnification of mask fabricated fibre Bragg gratings, Electron. Lett., 29(18):1614-1615, 1993.

[65]. B. Malo, F. Bilodeau, D. C. Johnson, and K. O. Hill. Fiber mode converters: point-by-point fabrication of index gratings, visualization using thermoluminescence, and applications, in Optical Fiber Communication, WL3, 1991.

[66]. K. O. Hill, F. Bilodeau, B. Malo, T. Kitagawa, S. Theriault, D. C. Johnson, and J. Albert. Aperiodic In-Fiber Bragg Gratings for Optical Fiber Dispersion Compensation, in Conference on Optical Fiber Communication, PD2, 1994.

[67]. R. Kashyap, P. F. McKee, R. J. Campbell, and D. L. Williams. Novel method of producing all fibre photoinduced chirped gratings, Electron. Lett., 30(12):996-998, 1994.

[68]. J. Martin and F. Ouellette. Novel writing technique of long and highly reflective in-fibre gratings, Electron. Lett., 30(10):811-812, 1994.

[69]. W. H. Loh, M. J. Cole, M. N. Zervas, S. Barcelos, and R. I. Laming. Complex grating structures with uniform phase masks based on the moving fiber-scanning beam technique, Opt. Lett., 20(20):2051, 1995.

[70]. W. H. Loh, M. J. Cole, M. N. Zervas, and R. I. Laming. Compensation of imperfect phase mask with moving fibre-scanning beam technique for production of fibre gratings, Electron. Lett., 31(17):1483-1485, 1995.

[71]. R. Kashyap, H. G. Froehlich, A. Swanton, and D. J. Armes. Super-step-chirped fibre Bragg gratings, Electron. Lett., 32(15):1394-1396, 1996.

[72]. R. Stubbe, B. Sahlgren, S. Sandgren, and A. Asseh. Novel technique for writing long superstructured fiber Bragg gratings, in Postdeadline Papers, Photosensitivity and Quadratic Nonlinearity in Glass Waveguides: Fundamentals and Applications, PD 1, 1995.

[73]. A. Asseh, H. Storoy, B. E. Sahlgren, S. Sandgren, and R. A. H. Stubbe. A writing technique for long fiber Bragg gratings with complex reflectivity profiles, J. Lightwave Technol., 15(8):1419-23, 1997.

88 Bibliography

[74]. P. Y. Fonjallaz, B. Sahlgren, B. Grennberg, and R. Stubbe. New interferometer for complete design of UV-written fibre Bragg gratings, Williamsburg, VA, USA, 36-8, 1997.

[75]. I. Petermann, B. Sahlgren, S. Helmfrid, A. T. Friberg, and P.-Y. Fonjallaz. Fabrication of Advanced Fiber Bragg Gratings by use of Sequential Writing with a Continuous-Wave Ultraviolet Laser Source, Appl. Opt., 41(6):1051-1056, 2002.

[76]. P. Sansonetti, F. Bakhti, I. Riant, and B. Poumellec. Construction method for large size Bragg grating within optical fiber, Patent no. FR2768819,1999.

[77]. B. Sahlgren and I. Petermann. System and method for fabricating Bragg gratings with overlapping exposures, US patent no. 7046866,2006.

[78]. I. Petermann. Fibre Bragg Gratings: Realization, Characterization and Simulation, Doctoral thesis in Royal Institute of technology, 2007.

[79]. A. Yariv. Coupled-mode theory for guided-wave optics, J. of Quantum Electron., QE9(9):919-33, 1973.

[80]. M. Yamada and K. Sakuda. Analysis of almost-periodic distributed feedback slab waveguides via a fundamental matrix approach, Appl. Opt., 26(16):3474, 1987.

[81]. H. Kogelnik and C. V. Shank. Coupled-Wave Theory of Distributed Feedback Lasers, J. Appl. Phys., 43(5):2327-2335, 1972.

[82]. H. Kogelnik. Filter response of nonuniform almost-periodic structures, Bell Syst. Tech. J., 55(109-126, 1976.

[83]. T. Erdogan. Fiber grating spectra, J. Lightwave Technol., 15(8):1277-1294, 1997.

[84]. V. Mizrahi and J. E. Sipe. Optical properties of photosensitive fiber phase gratings, J. Lightwave Technol., 11(10):1513-1517, 1993.

[85]. J. Capmany, B. Ortega, and D. Pastor. A tutorial on microwave photonic filters, J. Lightwave Technol., 24(1):201-229, 2006.

[86]. R. A. Minasian. Photonic signal processing of microwave signals, IEEE Trans. Microw.Theory Tech., 54(2):832-46, 2006.

[87]. R. Olshansky, V. A. Lanzisera, and P. M. Hill. Subcarrier multiplexed

89

lightwave systems for broad-band distribution, J. Lightwave Technol., 7(9):1329-1342, 1989.

[88]. T. Darcie. Subcarrier multiplexing for multiple-access lightwave networks, J. Lightwave Technol., 5(8):1103-1110, 1987.

[89]. R. A. Minasian. Photonic signal processing of high-speed signals using fiber gratings, Opt. Fiber Technol., 6(2):91-108, 2000.

[90]. D. B. Hunter and R. A. Minasian. Tunable microwave fiber-optic bandpass filters, IEEE Photon. Technol. Lett., 11(7):874-876, 1999.

[91]. N. You and R. A. Minasian. A novel high-Q optical microwave processor using hybrid delay-line filters, IEEE Trans. Microw.Theory Tech., 47(7):1304-8, 1999.

[92]. J. Capmany, D. Pastor, and B. Ortega. Fibre optic microwave and millimetre-wave filter with high density sampling and very high sidelobe suppression using subnanometre optical spectrum slicing, Electron. Lett., 35(6):494-496, 1999.

[93]. W. Zhang, G. Yu, and J. A. R. Williams. Tap multiplexed fibre grating-based optical transversal filter, Electron. Lett., 36(20):1708-1710, 2000.

[94]. M. Popov, P. Y. Fonjallaz, and O. Gunnarsson. Compact microwave photonic transversal filter with 40-dB sidelobe suppression, IEEE Photon. Technol. Lett., 17(3):663-665, 2005.

[95]. B. Ortega, J. Mora, J. Capmany, D. Pastor, R. Garcia-Olcina, and S. Sales. Highly selective Microwave Photonic filters based on new FBGs-EDF recirculating cavities and tuned modulators, in International Topical Meeting on Microwave Photonics (MWP 2005), Seoul, South Korea, 209-212, 2005.

[96]. D. B. Hunter and R. A. Minasian. Reflectively tapped fibre optic transversal filter using in-fibre Bragg gratings, Electron. Lett., 31(12):1010-1012, 1995.

[97]. F. Coppinger, S. Yegnanarayanan, P. D. Trinh, and B. Jalali. All-optical incoherent negative taps for photonic signal processing, Electron. Lett., 33(11):973-975, 1997.

[98]. J. Capmany, J. Cascon, J. L. Martin, S. Sales, D. Pastor, and J. Marti. Synthesis of fiber-optic delay line filters, J. Lightwave Technol., 13(10):2003-2012, 1995.

90 Bibliography

[99]. S. Sales, J. Capmany, J. Marti, and D. Pastor. Experimental demonstration of fibre-optic delay line filters with negative coefficients, Electron. Lett., 31(13):1095-1096, 1995.

[100]. K. Sasayama, M. Okuno, and K. Habara. Coherent optical transversal filter using silica-based waveguides for high-speed signal processing, J. Lightwave Technol., 9(10):1225-1230, 1991.

[101]. F. Coppinger, C. K. Madsen, and B. Jalali. Photonic microwave filtering using coherently coupled integrated ring resonators, Microw. Opt. Tech. Lett., 21(2):90-93, 1999.

[102]. M. Popov, P. Y. Fonjallaz, D. Berlemont, and O. Gunnarsson. Direct microwave optical filtering: concept, configurations, and tunability issues, Budapest, Hungary, 81-4, 2004.

[103]. P.-Y. Fonjallaz, E. Yusta, A. Djupsjobacka, M. Popov, and L. Cucala. BER assessment of direct optical passband filter for 24 GHz millimetre-wave signals, Seoul, South Korea, 225-228, 2005.

[104]. M. Popov, F. Carlsson, and P.-Y. Fonjallaz. Design and fabrication of superimposed fibre Bragg gratings: towards the optimality, Brugge, Belgium, 7-15, 2003.

[105]. M. Fokine, L. E. Nilsson, Å. Claesson, D. Berlemont, L. Kjellberg, L. Krummenacher, and W. Margulis. Integrated fiber Mach-Zehnder interferometer for electro-optic switching, Opt. Lett., 27(18):1643-1645, 2002.

[106]. C. M. Lawrence, D. V. Nelson, E. Udd, and T. Bennett. A fiber optic sensor for transverse strain measurement, Exp. Mech., 39(3):202-209, 1999.

[107]. W. Zhang, J. A. R. Williams, and I. Bennion. Polarization synthesized optical transversal filter employing high birefringence fiber gratings, IEEE Photon. Technol. Lett., 13(5):523-525, 2001.

[108]. E. Chehura, C.-C. Ye, S. E. Staines, S. W. James, and R. P. Tatam. Characterization of the response of fibre Bragg gratings fabricated in stress and geometrically induced high birefringence fibres to temperature and transverse load, Smart Mater. Struct., 13(4):888-895, 2004.

[109]. J. Paul, Z. Liping, B. Ngoi, and F. Z. Ping. Bragg grating temperature sensors: Modeling the effect of adhesion of polymeric coatings, Sens. Rev., 24(4):364-369, 2004.

91

[110]. R. Gafsi and M. A. El-Sherif. Analysis of induced-birefringence effects on fiber Bragg gratings, Opt. Fiber Technol., 6(3):299-323, 2000.

[111]. J. A. Viator, S. Kreger, M. W. Winz, and E. Udd. Modeling and experimental strain measurements on a non-homogeneous cylinder under transverse load, San Diego, CA, United States, 199-205, 2004.

[112]. N. F. Borrelli and R. A. Miller. Determination of the Individual Strain-Optic Coefficients of Glass by an Ultrasonic Technique, Appl. Opt., 7(5):745-750, 1968.

[113]. H. Young-Geun, K. Gilhwan, L. Ju Han, K. Sang Hyuck, and L. Sang Bae. Lasing wavelength and spacing switchable multiwavelength fiber laser from 1510 to 1620 nm, IEEE Photon. Technol. Lett., 17(5):989-991, 2005.

[114]. P. D. Dragic. Injection-seeded Q-switched fiber ring laser, IEEE Photon. Technol. Lett., 16(8):1822-1824, 2004.

[115]. C. Mukhopadhyay and P. D. Dragic. Broad-band injection seeding of an erbium-doped fiber ring laser, IEEE Photon. Technol. Lett., 17(6):1166-1168, 2005.

[116]. P. D. Dragic. Analytical model for injection-seeded erbium-doped fiber ring lasers, IEEE Photon. Technol. Lett., 17(8):1629-1631, 2005.

Paper Ⅰ

IEEE PHOTONICS TECHNOLOGY LETTERS, VOL. 19, NO. 4, FEBRUARY 15, 2007 233

Direct Detection of Direct Optically FilteredMillimeter-Wave Signals

Zhangwei Yu, Anders Djupsjobacka, Mikhail Popov, and Pierre-Yves Fonjallaz

Abstract—Good performance of a direct optical microwave filtercombined with direct detection consisting of a single superimposedgrating used in reflection has been assessed through bit-error-rate(BER) measurement. In the passband with a 3-dB bandwidth of2 GHz, the BER is below 1:2 10

11, showing nearly ON–OFF be-havior of the filter.

Index Terms—Bit-error-rate (BER) measurement, fiber Bragggratings (FBGs), optical microwave filtering.

I. INTRODUCTION

FOR microwave signals at frequencies higher than a few gi-gahertz, optical filters [1], [2] can be an advantageous alter-

native to conventional electrical ones [3], [4]. Most microwaveoptical filters use the incoherent summing of delayed copies of amodulated optical signal at the photodetector [5]–[8]. One majorlimitation of this technique is that it requires light sources withshort temporal coherence. Direct microwave optical filtering,proposed and demonstrated a few years ago [9], [10], can, how-ever, work with coherent light sources [11], [12]. For that pur-pose, fiber Bragg gratings (FBGs) [11]–[14] are usually betterthan Fabry–Pérot etalons [9], [10] in terms of spectral charac-teristics and stability.

In this letter, we demonstrate direct optical filtering com-bined with direct detection using a single double-peaked super-imposed FBG working in reflection [15]. The direct detectionrequires the suppression of the optical carrier by the optical filter(together with all undesired spectral components) and only thetwo sidebands (20 GHz in our case) of the optical carrier witha good extinction ratio are supplied to the photodetector. Theability of the filter to remove out-of-band signals is tested bymeasuring digital signals at 155 MB/s placed on a millimeter-wave subcarrier. The direct-detection scheme excludes the ne-cessity of the down-conversion of the high-frequency signal toan intermediate frequency in the detection part and thus sparesa local oscillator, a mixer, and an envelope detector, which wereneeded in the configuration of our previous work [12]. Note thatin this case the optical carrier is also spared for other subcarriersit may carry and, therefore, the composite signal (the optical car-rier with nonfiltered subcarriers) can be transported further in

Manuscript received July 5, 2006; revised December 18, 2006. This work wassupported by the European Project LABELS (IST-2001-37435).

Z. Yu is with Joint Research Centre of Photonics of the Royal Institute ofTechnology (Sweden) and Zhejiang University (JORCEP), Zhejiang Univer-sity, Hangzhou 310058, China, and also with Kista Photonics Research Center,Department of Microelectronics and Applied Physics, Royal Institute of Tech-nology, SE-164 40 Kista, Sweden (e-mail: [email protected]).

A. Djupsjobacka, M. Popov, and P.-Y. Fonjallaz are with the Photonics De-partment, Acero AB, SE-164 40 Kista, Sweden.

Digital Object Identifier 10.1109/LPT.2007.891195

Fig. 1. Principle of the direct filtering combined with direct detection.

the optical link. The possibility for heterodyne detection and ac-tive locking [12] of the remaining subcarriers is thus preserved.

II. EXPERIMENT

The basic principle of the direct detection for the direct op-tical filtering of microwave signals using a superimposed FBGin reflection is illustrated in Fig. 1. Two gratings are written atthe same location forming what is usually called a superimposedFBG [15]. Both of them have identical spectral characteristicsbut different resonance frequencies and work as passband filtersfor the sidebands of the millimeter-wave subcarrier to be pro-cessed. Perfect colocation and close to identical characteristicsof the gratings imply that a negligible delay difference betweenthe side modes is expected after reflection from the grating filter.

The experimental setup is depicted in Fig. 2. A mixer is usedto modulate a microwave carrier at around 20 GHz with the155-Mb/s digital signal at the output of the pattern generator.This modulated signal is then amplified and connected to theradio-frequency (RF) input of the optical Mach–Zehnder mod-ulator (MZM). Light from a tunable laser is launched into themodulator after adjusting the input polarization with a polar-ization controller. The modulator is direct current biased to theinflexion point to obtain quasi-linear modulation regime. Theoutput of the modulator is first amplified with an erbium-dopedfiber amplifier before reflection on the superimposed FBG filtervia a circulator. With this configuration, the FBG will work bothas a channel selective device as well as an amplified spontaneousemission filter.

For the detection, the beating frequency between the two sub-carriers (about 40 GHz) should be completely removed by thephotodiode which had a bandwidth of 2 GHz and the first twoamplifiers which had a bandwidth of 400 MHz. Note that thebandwidth of the photodiode in the direct detection can be aslow as the bandwidth of the digital payload signal, i.e., 155 MHz

1041-1135/$25.00 © 2007 IEEE

234 IEEE PHOTONICS TECHNOLOGY LETTERS, VOL. 19, NO. 4, FEBRUARY 15, 2007

Fig. 2. Experimental setup used for the demonstration of the direct-detectionscheme.

in our case. Using a low-speed and low-cost photodiode pro-vides additional savings in the direct-detection schemes. TheSTM-1 low-pass filter (LPF) with a cutoff frequency of 75%of 155 MHz is used to shape the digital payload signal. Theamplifier after the STM-1 LPF is required to amplify the dig-ital signal to a level compatible with the error detector of thebit-error-rate tester (BERT) device. To suppress thermal noisefrom this broadband (10 GHz) amplifier, an STM-16 LPF witha bandwidth of 75% of 2.5 GHz has been put after it. In gen-eral, such a broadband performance is not required at this stageof the detection (the choice of the components was based uponavailability) and a substantially narrowband amplifier and filtercould be used here.

For this experiment, two Hamming-apodized FBGs werewritten in a standard single-mode fiber with a length of 10 cm.The reflection spectrum of the grating used in the setup isshown in Fig. 3. The 3-dB bandwidth of the two gratings wasabout 2 GHz. The pass function of the filter in the microwavedomain is hence expected to be about 2 GHz. Also, a full stopeffect is expected outside this 2-GHz range. The reflectivityof each peak in the grating is about 45% and the wavelengthseparation between the center of the two peaks is about 320 pmcorresponding to 40 GHz (i.e., for filtering at the subcarrierfrequency of 20 GHz).

A key issue of the direct-detection experiment was to have asufficient suppression of the optical carrier. At the optical MZM,the amplitude of the modulating subcarrier signal was kept suf-ficiently small to preserve the linearity of the modulation. Thisimplies that in the optical domain, the optical carrier is clearlystronger than the sidebands which represent the subcarrier withthe digital payload. On the other hand, for a proper detection ofthe digital signal, the level of the optical carrier has to be lowerthan those of the sidebands. In our case, this was guaranteed by

Fig. 3. Reflection spectrum of the direct optical microwave filter consisting oftwo superimposed FBGs.

Fig. 4. BER measurements as a function of microwave signal frequency.Inset: Eye diagrams of the detected signal at frequencies of 20, 21, and21.5 GHz, respectively.

the more than 20-dB suppression of the optical carrier providedby the grating (see Fig. 3). Special care was also taken to avoidparasitic reflections from connectors or splices.

III. RESULTS

The digital 155-Mb/s payload with a pseudorandom binarysequence of was generated by the BERT device andthen used to modulate in amplitude an electrical signal whosefrequency was varied from 18.4 to 21.6 GHz. This electricalsignal was then fed to the optical modulator. The optical wave-length was set to the optimal 1534.64 nm corresponding to thecenter wavelength between the two reflection peaks. The op-tical power and the optical signal-to-noise ratio before being re-flected by the FBG kept constant as the frequency was changed.The bit-error-rate (BER) measurements as a function of the fre-quency have been recorded and eye diagrams have been takenat a number of frequency points. The results of these measure-ments are shown in Fig. 4. Outside the given range of frequen-cies, no measurement was possible because of synchronizationloss.

YU et al.: DIRECT DETECTION OF DIRECT OPTICALLY FILTERED MILLIMETER-WAVE SIGNALS 235

As one can see, the passband width of the filter is about 2 GHzas expected from the optical spectrum of the grating. The BERperformance in the passband of the filter shown by the flat partof the curve corresponds to the BER of or better.This performance was considered as excellent and more precisemeasurements were not conducted due to time constraints. Out-side the passband, the quality of the signal quickly deterioratesdemonstrating an almost ON–OFF behavior. The eye diagrams(the inset of Fig. 4) are of good quality within the passband andquickly closing outside the passband.

Finally, we made a simple experiment for using only one peakof the double-peaked grating. To do so, the optical carrier wasplaced to a wavelength longer than that of the right-hand peak ofthe grating so that the spacing between the peak and the opticalcarrier was 19 GHz (the first test) and 20 GHz (the second test).In this way, only a single (left) sideband of the composite opticalsignal was supplied to the photodetector. In both tests, the con-figuration demonstrated efficient filtering of the signal with thebandwidth of 2 GHz which matches well the optical bandwidthof this single peak. Naturally, the energy of the detected signalin this case was only half of the signal detected using both side-bands. These experiments prove the possibility of using the di-rect-detection technique for optical single sideband modulatedsignals and also for tunable filtering of the sidebands by meansof tuning the wavelength of the grating.

The FBG filter, together with the direct-detection schemepresented in this letter, can be applied in radio-over-fiber(ROF) systems comprising more than one RF signal. Becauseof the narrow bandwidth of the superimposed FBG presentedhere, cascaded superimposed FBGs (with the same centralwavelength but different wavelength separations of the twopeaks, such as 320, 384, and 448 pm) can be used to filtercorresponding RF signals (such as 20, 24, and 28 GHz, respec-tively). However, the narrow bandwidth of the filter also limitsthe maximum data rate, which would be 800 Mb/s in the casepresented here.

IV. CONCLUSION

The good performance of a direct optical microwave filtercombined with highly simplified direct detection consisting ofa single double-peaked superimposed grating used in reflectionhas been assessed through BER measurements. The filterexhibits nearly ON–OFF behavior, where transmission is either

error-free inside the passband (with the 3-dB bandwidth of2 GHz) or basically impossible outside. Such an optical filterwith a cost-effective detection scheme can be applied in ROFsystems.

REFERENCES

[1] J. Capmany, B. Ortega, and D. Pastor, “A tutorial on microwave pho-tonic filters,” J. Lightw. Technol., vol. 24, no. 1, pp. 201–229, Jan. 2006.

[2] R. A. Minasian, “Photonic signal processing of microwave signals,”IEEE Trans. Microw. Theory Tech., vol. 54, no. 2, pt. 2, pp. 832–846,Feb. 2006.

[3] T. E. Darcie, “Subcarrier multiplexing for multiple-access lightwavenetworks,” J. Lightw. Technol., vol. LT-5, no. 8, pp. 1103–1100, Aug.1987.

[4] R. Olshansky, V. A. Lanzisera, and P. M. Hill, “Subcarrier multiplexedlightwave systems for broad-band distributuion,” J. Lightw. Technol.,vol. 7, no. 9, pp. 1329–1342, Sep. 1989.

[5] D. B. Hunter and R. A. Minasian, “Tunable microwave fiber-opticbandpass filers,” IEEE Photon. Technol. Lett., vol. 11, no. 7, pp.874–876, Jul. 1999.

[6] N. You and R. A. Minasian, “A novel high-Q optical microwave pro-cessor using hybrid delay-line filters,” IEEE Trans. Microw. TheoryTech., vol. 47, no. 7, pt. 2, pp. 1304–1308, Jul. 1999.

[7] W. Zhang, G. Yu, and J. A. R. Williams, “Tap multiplexed fibre grating-based optical transversal filter,” Electron. Lett., vol. 36, pp. 1708–1710,2000.

[8] J. Capmany, D. Pastor, and B. Ortega, “Fibre optic microwave and mil-limetre-wave filter with high-density sampling and very high sidebandsuppression using subnanometre optical spectrum slicing,” Electron.Lett., vol. 35, pp. 494–496, 1999.

[9] P. A. Greenhalgh, R. D. Abel, and P. A. Davies, “Optical prefilteringin subcarrier systems,” Electron. Lett., vol. 28, pp. 1850–1852, 1992.

[10] P. D. Sargis, R. E. Haigh, K. G. McCammon, and S. Young, “Subcarriermultiplexing with dispersion reduction,” Electron. Lett., vol. 31, pp.1769–1770, 1995.

[11] M. Popov, P.-Y. Fonjallaz, D. Berlemont, and O. Gunnarsson, “Directmicrowave optical filtering: Concept, configurations, and tunability is-sues,” in Proc. IEEE Int. Topical Meeting Microwave Photonics, Bu-dapest, Hungary, Sep. 2003, pp. 81–84.

[12] P.-Y. Fonjallaz, E. Yusta, A. Djupsjobacka, M. Popov, and L. Cu-cala, “BER assessment of direct optical passband filter for 24 GHzmillimetre-wave signals,” in 2005 Int. Topic Meeting Microwave Pho-tonics, Seoul, Korea, Oct. 2005, vol. TP-27, pp. 225–228.

[13] T. Kuri, K. Kitayama, and K. Murashima, “Error-free DWDM trans-mission of 60-GHz-band fiber-radio signals with square-like responsefiber Bragg gratings,” in 2001 Int. Topic Meeting Microwave Photonics,Jan. 2001, vol. Tu-2.6, pp. 69–72.

[14] C. Marra, A. Nirmalathas, C. Lim, M. Attygalle, D. Novak, B. Ashton,L. Poladian, W. S. T. Rowe, T. Wang, and J. A. Besley, “FBG-basedoptical interface to support a multisector antenna in a spectrally effi-cient fiber radio system,,” IEEE Photon. Technol. Lett., vol. 16, no. 1,pp. 254–256, Jan. 2004.

[15] M. Popov, F. Carlsson, and P.-Y. Fonjallaz, “Design and fabrication ofsuperimposed fibre Bragg gratings: Towards the optimality,” in Proc.SPIE—Int. Society Optical Engineering (Photonics Fabrication Eu-rope), Brugge, Belgium, Oct. 2002, vol. 4943, pp. 7–15.

Paper Ⅱ

High-Speed Control of Fiber Bragg Gratings

Zhangwei Yu1,3,4, W. Margulis2, O. Tarasenko2, H. Knape2 and P-Y. Fonjallaz1,2

1Royal Institute of Technology (KTH) and 2Acreo AB collaborating within KPRC, SE-164 40 Kista, Sweden 3Joint Research Centre of Photonics of KTH and Zhejiang University (JORCEP)

4Zhejiang University, Hangzhou 310058, China. E-mail: [email protected]

Abstract: FBGs were written in fiber with internal alloy electrodes. Nanosecond high current pulses cause metal expansion, increase birefringence and tune the gratings. High-speed wavelength switching was accomplished with potential use in Q-switching fiber lasers.

Fibers with internal electrodes can be used in a number of interesting applications. Besides electrooptical switching and modulation, another recent functionality demonstrated was nanosecond polarization switching [1]. Increased birefringence results from the application of high current electrical pulses to an internal electrode that heats up and expands. The refractive index difference between fast and slow axis of the fiber with electrodes causes polarization change, and can lead to 90° rotation with proper input polarization alignment. In the present work, we studied how the high-speed expansion of metal electrodes affects the wavelength of a fiber Bragg grating written in the core of the fiber. The grating is an excellent tool to follow the evolution of the refractive index experienced by light in the polarization parallel (P) and orthogonal (S) to the direction of the holes. Besides revealing interesting physics, the nanosecond wavelength switching obtained here has potential application in Q-switching fiber lasers.

Pieces of fiber with a pair of 27-μm diameter holes running parallel to the core were filled with internal electrodes that occupied the entire cross-section of the holes. BiSn or AuSn were used as low temperature melting alloys, pumped into the fibers as a liquid and used at room temperature as solid electrodes [2]. Electrical contact was made from the side of the fiber, thus allowing for conventional low loss splicing. The fibers were single-mode at 1.5 μm and were provided with 7-cm long electrode sections. The edge-to-edge hole separation was 32 μm. The insertion loss of the devices before FBG fabrication was as low as 0.2 dB. The fibers were H2-loaded for 2 weeks at 150 bar, stripped from the acrylate coating and exposed to UV in a FBG writing set-up. The UV laser radiation used was a frequency doubled Ar+ laser. In the work reported here, we describe results obtained for two unchirped FBG with Hamming-apodized profile and without phase shift, with reflectivities 30% and 70% and FWHM ~20 pm and 30 pm, respectively. After writing the gratings, the pieces of fiber were mounted on a metal substrate and covered with heat conductive, electrically isolating epoxy. Heat dissipation allows switching the devices at rates of a few kHz without significant drift of polarization or wavelength. Electrical SMA contacts were used and a 0.1 Ω resistive probe was inserted in series with the ~50 Ω alloy electrode for monitoring purposes.

Two sources of electrical drive pulses were used for device characterization, one consisting of a low repetition rate spark gap delivering <10 ns risetime current pulses of up to 30 A into 50 Ω and duration 10-100 ns, and the other one of a home-built semiconductor-based switch delivering up to 20 A with 30 ns rise-time at a repetition rate 10 kHz and adjustable pulse length. The energy deposited on the electrode is a few mJ at most and causes a temperature rise of only a few degrees (typically <10 °C). The light source used was a tunable external cavity diode laser (TLS) with linewidth <1 pm, and synchronized with an optical spectrum analyzer for high resolution wavelength measurements. Different input polarization states were selected with a polarization controller. A circulator, high-speed photodiode

a236_1.pdf

JWA44.pdf

mailto:[email protected]

and oscilloscope completed the characterization set-up, schematically illustrated in Fig. 1.

Tunable laser Photodiode Scope

Polarization control

HV

Fig. 1. Experiment set-up Fig 2 illustrates the Bragg wavelength associated with S and P polarization of the metal-filled fiber device at

various temperatures before the application of current pulses. The insets show the reflected intensity as a function of wavelength at ~26 °C and at ~69 °C, measured for the two orthogonal input polarization states that match the eigenstates of the fiber. Close to room temperature, the index difference inferred from δn = n δλ/λ was 2.6×10-5, one order of magnitude lower than conventional HiBi fiber. As the temperature increases, both the refractive indices for both polarizations and the birefringence increase, as shown in Fig. 2. At 69 °C the Bragg wavelength for the two polarizations are well separated (87 pm). Such a state can also be reached by heating the fiber by passing dc current through one electrode, instead of using a hot plate as in the characterization measurement here.

20 30 40 50 60 70 80 90

1546.4

1546.8

1547.2

1547.6

1548.0

1548.4

1548.8

Wav

elen

gth

diffe

renc

e (p

m)

1546.2 1546.3 1546.4

0.0

0.2

0.4

0.6

0.8

1.0

25.8 0C

Ref

lect

ivity

(mW

)

Wavelength (nm)

26pm

Temperature (0C)

Wav

elen

gth

(nm

)

wavelength of the left peak wavelength of the right peak wavelength difference

20

30

40

50

60

70

80

90

100

110

1547.8 1547.9 1548.0

0.0

0.2

0.4

0.6

0.8

1.0

68.68 0C 87pm

Ref

lect

ivity

(mW

)

Wavelength (nm)

Fig. 2. Temperature dependence of Bragg resonance for S and P polarization The dynamic measurements were performed for the S and P polarization states by tuning the TLS exactly at the

top of the grating reflection peak before the application of the current pulse and studying the drop in reflectivity as a function of time following the electrical excitation. Alternatively, the TLS was tuned outside the Bragg peak and the increase of reflectivity was followed on the oscilloscope trace following application of the current pulse. Intermediate (partial reflection) cases were also studied. Fig. 3 illustrates an example of 100 % wavelength switching with a response time ~60 ns for a drive pulse 100 ns long and amplitude ~30 A. The ripple seen on the traces is pick-up of the onset of the electrical pulse switched.

a236_1.pdf

JWA44.pdf

-300 -200 -100 0 100 200 300-5

0

5

10

15

20

25

30

35

R ise time:60ns

Time (ns)S

igna

l (m

V)

Fall time:63ns

Fig. 3. Time evolution of signal reflected by FBG for two different probe wavelengths. 100% switching is accomplished.

Three physical mechanisms have been identified in previous work on the polarization switch [1] and also

observed here. The fastest one (as rapid as <10 ns) results from the mechanical pressure due to the electrode expansion and is the most useful for application in Q-switching. It lasts for a time comparable to the electrical pulse duration. The oscillating regime is associated with acoustic modes of the fiber [3] with typical period ~40-60 ns, but the acoustic oscillations induced here were limited in amplitude. The slow response is due to the propagation of the heat wave through the fiber, which reaches the core after hundreds of nanoseconds. Even when operating above room temperature and with heat sinking, the thermal component lasts for many microseconds, limiting the maximum repetition rate of the device to tens of kHz. In the traces of Fig. 3, the acoustic oscillations are not observable and the thermal response is well outside the time interval displayed.

The experiments repeatedly showed for the two devices tested that the largest wavelength shift during the application of the electrical pulse was to shorter wavelengths (typically -25 pm for 100% switching), while for the orthogonal polarization the wavelength shift was positive, but much smaller (typically +5 pm). After the electrical pulse ends, the wavelength shifts by approximately +17 pm for both polarization states due to a 2 °C temperature rise, which is expected from the amount of energy deposited. One qualitative model to explain the FBG shift to shorter wavelengths at early times is to consider that when the core is compressed in the direction of the holes, it expands in the orthogonal direction and thus the index reduces. Simulations of the effect using the elasto-optical tensor of glass confirm that the compression of the core in one direction leads to a weak increase of the index in the direction of the force, and a stronger decrease of the refractive index for the orthogonal direction.

Previous durability tests were carried out with devices pulsed 109 times without degradation in performance [1]. The temperature increase produced here barely affects the grating and it is not expected to reduce the reliability of the wavelength switch demonstrated. It is believed that the short length, low loss, all spliced high-speed devices described here can find useful application for Q-switching fiber lasers. References 1- H. Knape and W. Margulis, Opt. Lett. 32, 614 (2007). 2- M. Fokine, L. E. Nilsson, Å. Claesson, D. Berlemont, L. Kjellberg, L. Krummenacher, W. Margulis, Opt. Lett. 27, 1643, (2002). 3- A. Gusarov, Ky Nguyen Hong, H. G. Limberger, R. P. Salathe, G. R. Fox, J. Lightwave Tech. 14, 2771 (1996).

a236_1.pdf

JWA44.pdf

Paper Ⅲ

Nanosecond switching of fiber Bragg gratings Zhangwei Yu1,3, W. Margulis2*, O. Tarasenko2, H. Knape2 and P.-Y. Fonjallaz2,3 1JORCEP [Joint Research Centre of Photonics of the Royal Institute of Technology (Sweden) and Zhejiang

University], Zhejiang University, Hangzhou 310058, China 2Acreo AB collaborating within Kista Photonics Research Center, Electrum 229, SE-164 40 Kista, Sweden 3Kista Photonics Research Center, Royal Institute of Technology, Electrum 229, SE-164 40 Kista, Sweden

*Corresponding author: [email protected]

Abstract: A FBG was written in a two-hole fiber with internal alloy electrodes. Nanosecond high current pulses cause metal expansion, increase birefringence and tune the gratings with a response time of 29 ns. This short length, low loss, all-spliced high-speed wavelength switching devices described here has potential use in Q-switching fiber laser.

©2007 Optical Society of America

OCIS codes: (060.3735) Fiber Bragg gratings; (060.4370) Nonlinear optics, fibers; (060.7140) Ultrafast processes in fibers.

References and links

1. S. M. Melle, A. T. Alavie, S. Karr, T. Coroy, K. Liu, and R. M. Measures, “A Bragg Grating-Tuned Fiber Laser Strain Sensor System,” IEEE Photon. Technol. Lett. 5, 263-266 (1993).

2. M. M. Ohn, A. T. Alavie, R. Maaskant, M. G. Xu, F. Bilodeau, and K. O. Hill, “Dispersion variable fibre Bragg grating using a piezoelectric stack,” Electron. Lett. 32, 2000-2001 (1996).

3. J. L. Cruz, A. Diez, M. V. Andres, A. Segura, B. Ortega, and L. Dong, “Fibre Bragg gratings tuned and chirped using magnetic fields,” Electron. Lett. 33, 235-236 (1997).

4. C. M. Lawrence, D. V. Nelson, E. Udd, and T. Bennett, “A fiber optic sensor for transverse strain measurement,” Exp. Mech. 39, 202- 209 (1999) .

5. M. Silva-Lopez, C. Li, W. N. MacPherson, A. J. Moore, J. S. Barton, J. D. C. Jones, D. Zhao, L. Zhang, and I. Bennion, “Differential birefringence in Bragg gratings in multicore fiber under transverse stress,” Opt. Lett, 29, 2225-2227 (2004).

6. C. Jewart, K. P. Chen, B. McMillen, M. M. Bails, S. P. Levitan, J. Canning, and I. V. Avdeev, “Sensitivity enhancement of fiber Bragg gratings to transverse stress by using microsturctural fibers,” Opt. Lett. 31, 2260-2262 (2006).

7. H. M. Xie, Ph. Dabkiewicz, R. Ulrich, and K. Okamoto, “Side-hole fiber for fiber-optic pressure sensing,” Opt. Lett. 11, 333-335 (1986).

8. S. Kreger, S. Calvert, and E. Udd, “High Pressure Sensing using Fiber Bragg Grating written in Birefringent Side Hole Fiber,” in Proceedings of OFS-15, Portland, Oregon, 355-358 (2002).

9. H. Knape, and W. Margulis, “All-fiber polarization switch,” Opt. Lett. 32, 614-616 (2007). 10. M. Fokine, L. E. Nilsson, Å. Claesson, D. Berlemont, L. Kjellberg, L. Krummenacher, and W. Margulis,

“Integrated fiber Mach-Zehnder interferometer for electro-optic switching,” Opt. Lett. 27, 1643-1645 (2002).

11. W. Zhang, J. A. R. Williams, and I. Bennion, “Polarization Synthesized Optical Transversal Filter Employing High Birefringence Fiber Gratings,” IEEE Photon. Technol. Lett. 13, 523-525 (2001).

12. E. Chehura, C.-C. Ye, S. E Staines, S. W James, and R. P Tatam, “Characterization of the response of fibre Bragg gratings fabricated in stress and geometrically induced high birefringence fibres to temperature and transverse load,” Smart Mater. Struct. 13, 888-895 (2004).

13. J. Paul, L. Zhao, B. Ngoi, and Z. Fang, “Bragg grating temperature sensors: modeling the effect of adhesion of polymeric coatings,” Sens. Rev. 24, 364-369 (2004).

14. E. Chmielewska, W. Urbanczyk, and W. J. Bock, “Measurement of pressure and temperature sensitivities of a Bragg grating imprinted in a highly birefrigent side-hole fiber,” Appl. Opt. 42, 6284-6291 (2003).

1. Introduction

High speed tuning of Fiber Bragg gratings (FBG) can be exploited in a number of fields, such as Q-switching, cavity dumping and pulse picking in a fiber laser, quickly addressing a sensor head or readjusting an optical filter. Various mechanisms have been exploited for tuning FBGs, such as bending beams [1], stretching with piezoelectric stack [2] and using magnetic fields [3]. Common to these methods are a relatively low tuning speed (microseconds) and mechanical durability issues. In recent years, there has been an increasing interest in

#88233 - $15.00 USD Received 3 Oct 2007; revised 23 Oct 2007; accepted 23 Oct 2007; published 26 Oct 2007

(C) 2007 OSA 29 October 2007 / Vol. 15, No. 22 / OPTICS EXPRESS 14948

exploiting strain transverse to the fiber [4-8]. The transverse stress distribution has been previously measured in a number of FBG sensor elements in high birefringence (HiBi) fibers [4], in multi-core fibers [5], in microstructured fibers [6] and in side-hole fibers [7,8]. The birefringence induced by the external load leads to both a shift and an increase in the splitting of the FBG peak associated with the two orthogonal polarization modes. This can be used to gauge the magnitude and the orientation of the transverse strain field. However, to the best of our knowledge there is no published report on fast tuning FBGs in these special fibers till now.

In previous work, we studied the applications of microstructured fibers with internal electrodes for controlling the light in the fiber. Besides electro-optical switching and modulation, nanosecond polarization switching has recently been demonstrated [9]. Increased birefringence results from the application of high-voltage electrical pulses to an internal electrode that heats up and expands. The refractive index difference between fast and slow axis of the fiber with electrodes causes polarization change, and can lead to 90° rotation with proper input polarization alignment.

In this Paper, we present how the high-speed expansion of metal electrodes affects the wavelength of a FBG written in the core of the two-hole fiber. The grating is an excellent tool to follow the evolution of the refractive index experienced by light in the polarization parallel (P) and orthogonal (S) to the direction of the holes. Besides revealing interesting physics, the nanosecond wavelength switching obtained here has potential application in Q-switching fiber lasers.

2. Experiment

The 125 μm thick silicate two-hole fiber, drawn at Acreo FiberLab, had characteristics similar to those of the standard telecom fiber (core diameter 8 μm and 0.0056nΔ = ) but with two ~28 μm-diameter holes running parallel to the core separated from it by 14 μm (see Fig. 1). Pieces of fiber were filled with few-centimeter-long (typically 10 cm) metal columns that occupied the entire cross-section of the holes. Eutectic alloys BiSn (melting temperature 137°C) were pumped into the fibers in the molten state and used at room temperature as solid electrodes [10]. The length of the alloy-filled holes and the core-hole separation were such that insertion loss was experimentally measured to be < 0.2 dB and the polarization-dependent loss < 0.1 dB. Both ends of the metal-filled fiber were free from metal to allow for convenient low-loss splicing.

Fig. 1. Cross section of the metal-filled fiber used here (SEM picture).

The metal-filled fibers were H2-loaded for 2 weeks at 150 bar, stripped from the acrylate

coating and exposed to ultraviolet (UV) in a FBG writing set-up. Frequency doubled Ar+ laser was used as a source of UV radiation. In the work reported here, we describe results obtained from a 4 cm-long Hamming-apodized FBG written along the middle section of metal-filled fibers. The grating was annealed in the oven at 100 °C for 12 hours.

After side polishing the fiber to electrically access one of the two internal electrodes at two points separated by 7 cm, a 20 μm thick gold-plated tungsten wire was inserted by ~5 mm into

8 um

28 um 14 um 125 um 14 um 28 um



the electrode while melting the alloy locally to ensure a robustly bonded electrical connection. Then, the pieces of fiber were mounted on an aluminum (Al) substrate and covered with heat conductive, electrically isolating epoxy. Heat dissipation into the substrate allows switching the devices at rates of a few kHz without significant drift of polarization or wavelength. The resistance of the components was typically ~50 Ω for a BiSn device with 7 cm active length. This allowed for simple impedance matching between the device and the coaxial cable transporting the high voltage pulses. Electrical SMA contacts were used and a 0.1 Ω resistive probe was connected in series with the ~50 Ω load for monitoring purposes.

A pulse generator consisting of a high voltage direct current (DC) power supply and a high speed switch was employed for the driving pulses. The switch here was a low repetition rate spark gap circuit delivering 2 ns rise-time current pulses with the maximum switching voltage 3 kV and duration 10-100 ns determined by the cable length used. The energy applied to the electrode is a few mJ at most and causes a temperature rise of only a few degrees (typically < 10 °C). The light source used was a tunable external cavity diode laser (TLS) with linewidth < 1 pm, and synchronized with an optical spectrum analyzer (OSA) for high resolution wavelength measurements. Different input polarization states were selected with a polarization controller (PC). A circulator, 3-dB coupler, high-speed photodiode and oscilloscope completed the experimental set-up, schematically illustrated in Fig. 2.

Fig. 2. (Color online) Experiment set-up. TLS: tunable laser source; PC: polarization controller; OSA: optical spectrum analyzer.

3. Results

Due to the intrinsic geometrical birefriengence of the two-hole fiber, the grating has two reflection peaks at different wavelength corresponding to the different effective refractive indices for the two orthogonal polarization modes in this fiber [4]. The reflection peak at the shorter wavelength takes place in the fast axis mode (S polarization), the one at the longer wavelength in the slow axis mode (P polarization) [11]. When the metal-filled fiber device was put in a temperature-controllable oven, the two internal electrodes were heated and expanded. Consequently, the refractive indices for both polarizations and the birefringence increased as the temperature increased. The peak Bragg wavelengths associated with S and P polarization shifted to longer wavelengths and the splitting increased, as shown in Fig. 3. The insets show the reflected intensity as a function of wavelength at ~26 °C and at ~69 °C, measured for the two orthogonal input polarization states that match the eigenstates of the fiber. From room temperature to 69 °C, the Bragg wavelength separation for the two polarizations increased from 26 pm to 87 pm. The 3-dB bandwidth and the reflectivity of each peak remained about 30pm and 70%, respectively. Close to room temperature, the index difference inferred from /n n λ λΔ = × Δ was ~2.6×10-5, approximately one order of magnitude lower than that of conventional HiBi fiber. Such a state can also be reached by heating the fiber by passing DC current, instead of using an oven as in the characterization measurement here.

TLS

Photodiode

OSA

Oscilloscope

0.1 Ω

High Voltage

Switch 50 Ω Grating

PC

Coupler Circulator



One can note from Fig. 3 that the measured temperature sensitivity (~37.6 pm/°C) is significantly larger than expected for a HiBi FBG fabricated in Panda fiber (~16.5 pm/°C) [12]. As discussed in Ref. [13], increased temperature sensitivity can be obtained when the fiber is embedded in an environment which exhibits a large /dL dT . An estimate can be made, for a ~53 ºC temperature increase, of the wavelength shift due to /dn dT (0.65 nm), due to the length expansion of a glass fiber (0.045 nm), due to the alloy dilation (0.024 nm) and due to the expansion of the Al substrate (2.1 nm). Here it is assumed that the alloy expansion is restricted by the hard glass fiber (area ratio 1:9 and Young’s modulus ratio 1:6) and that the substrate is perfectly attached to the fiber. The total wavelength shift under these conditions is calculated to be 2.8 nm, which exceeds the measured value from Fig. 3 of ~2.0 nm. The lower value measured is probably caused by slippage between the substrate and the fiber [13].

10 20 30 40 50 60 70 80 90

1546.4

1546.8

1547.2

1547.6

1548.0

1548.4

1548.8

1546.28 1546.32 1546.36

0.0

0.2

0.4

0.6

0.8

1.0

Wavelength (nm)

26pm26 oC

Ref

lect

ivity

(mW

)

Wav

elen

gth

diffe

renc

e (p

m)

Temperature ( oC)

Wav

elen

gth

(nm

)

Peak wavelength of the S polarization Peak wavelength of the P polarization Wavelength difference

20

40

60

80

100

120

1547.8 1547.9 1548.0

0.0

0.2

0.4

0.6

0.8

1.069 oC 87pm

Ref

lect

ivity

(mW

)

Wavelength (nm)

Fig. 3. (Color online) Temperature dependence of Bragg resonance for S and P polarization in a steady-state situation.

-200 0 200 400 600-5

0

5

10

15

20

τFWHM = 60 ns

Cur

rent

(A

)

Time (ns)

Fig. 4. Typical current pulse used in the dynamic experiment. Imperfect matching into 50 Ω causes the small step at 100 ns and the undershoot.



Dynamic measurements were performed for the S and P polarization states by tuning the TLS at the top of the grating reflection peak before the application of the current pulse and studying the drop in reflectivity as a function of time following the electrical excitation. Alternatively, the TLS was tuned away from the Bragg peak and the increase of reflectivity was measured on the oscilloscope trace following application of the current pulse. Intermediate (partial reflection) cases were also studied. Electrically driven FBG switching experiments were carried out at room temperature. Figure 4 illustrates a typical current pulse used here with the pulse width ~ 60 ns and the amplitude ~18 A at low-repetition rate (~20 Hz). The small step on the right flank and the undershoot observed after the main current pulse are caused by electrical mismatch.

-200 0 200 400 600

0

10

20

30

40

-100 0 100 200 300

0

10

20

30

40

τfall = 40 us

Am

plitu

de

(mV

)Time (us)

τrise = 29 ns

Time (ns)

Am

plitu

de (

mV

)

Fig. 5. Full switching off-on is accomplished with 29 ns risetime. Inset, the time evolution of signal in microseconds.

-200 0 200 400 600

0

10

20

30

40

0 1 2

0

10

20

30

40

Am

plitu

de (

mV

)

T im e (m s)

τ fa ll = 29 ns

Tim e (ns)

Am

plitu

de (

mV

)

Fig. 6. Full switching on-off is accomplished with 29 ns falltime. Inset, the time evolution of signal in microseconds.

Close to full off-on switching with a response time of ~29 ns was achieved when the TLS

was set at the short wavelength side of the S polarization, as illustrated in Fig. 5. The increase of the signal implies that the Bragg peak is shifted to shorter wavelengths following electrical pulse excitation. The return of the Bragg peak to longer wavelengths is followed on a microsecond time scale in the inset of Fig. 5. As heat spreads from the metal electrode into the glass fiber and eventually dissipates, the pressure wave dies and the reflection peak returns to its original spectral position. The process takes tens/hundreds of microseconds. The decay of the reflected signal seen in the inset of Fig. 5 is deceivingly simple, since the probing laser



wavelength does not follow the spectral evolution of the Bragg peak once the overlap disappears. Heating, however, red-shifts the grating wavelength.

Close to full on-off switching was also obtained with a response time of 29 ns, as shown in Fig. 6, by setting the TLS wavelength to match FBG peak of the S polarization. The long term recovery of the grating, shown in the inset of Fig. 6, is not monotonic. The reflected signal remains null as long as the overlap between the probing laser source and the FBG displaced to shorter wavelengths is zero. Since the FBG has a FWHM of 30 pm, full switching implies displacement > 20 pm. The time is typically a few microseconds. After this relatively short time, the peak reflectance wavelength overlaps once again the probing laser wavelength. The signal reflected by the FBG then returns to the maximum amplitude. The heat from the metal electrode finally reaches the core of the fiber, red-shifting the Bragg wavelength, and this leads to a second reduction of the reflected signal. This time, however, the FBG peak lies on the long-wavelength side of the laser source. The time scale involved in the heat component is slower. The device can be said to return to room temperature after a delay close to 2 ms. It may be noticed from the inset of Fig. 6, that the second (negative) peak in the reflected signal is not as deep as the first one, where 100% switching is achieved. This depth is a measure of the remaining overlap between the probing laser wavelength and the red-shifted FBG peak. From the relative amplitudes of the peaks, a thermal shift of +17 pm can be estimated for both polarization states, which corresponds to a ~1.3 °C rise after each high current pulse. Here, the Bragg wavelength dependence on the Al substrate expansion due to temperature is not considered because of the large thermal mass of the substrate and the short duration time and low repetition rate of the current pulse. The alloy was heated by TΔ ~10 °C with a current pulse of ~18 Amps over ~60 ns (with heat deposited ~1 mJ). Assuming for simplicity that the energy spreads uniformly into the fiber and taking into account typical values for the specific heat of the alloy and glass, a temperature increase in the core of ~1.6 °C is estimated, which is consistent with the value above.

The dynamic experiments repeatedly showed for the device tested that the largest wavelength shift during the application of the electrical pulse was to shorter wavelengths (typically -27 pm for 100% switching) for the S polarization, while for the P polarization the wavelength shift was positive, but much smaller (typically +5 pm). Similar behavior has also been reported in a HiBi side-hole fiber [14]. One qualitative model to explain the FBG shift to shorter wavelengths is to consider that when the core is compressed in the direction of the holes, it expands in the orthogonal direction and thus the refractive index reduces. Simulations of the effect using the elasto-optical tensor of glass confirm that the compression of the core in one direction leads to a weak increase of the refractive index in the direction of the force, and a stronger decrease of the refractive index for the orthogonal direction.

4. Conclusion

A FBG was written in a two-hole fiber with internal alloy electrodes, which has intrinsic geometrical birefriengence. Nanosecond high current pulses heated the metal electrode and caused metal expansion rapidly. In consequence, birefringence was increased and the gratings was tuned with a response time of 29 ns, limited by the finite time needed to deposit heat and the inductance of the alloy electrode. The fast risetime switching results from the increased pressure, while heat transport to the core is slow and determines the recovery time of the device after every drive pulse. Previous durability tests were carried out with devices pulsed 109 times without degradation in performance and high repetition rates (> 10 kHz) can be obtained by running the device above room temperature [9]. It is believed that the short length, low loss, all-spliced high-speed devices described here can find useful application for Q-switching fiber lasers.



Paper Ⅳ

Physics of electrically switched fiber Bragg gratings Zhangwei Yu1,3, W. Margulis2,3*, P.-Y. Fonjallaz2,3 and O. Tarasenko2,3

1JORCEP [Joint Research Centre of Photonics of the Royal Institute of Technology (Sweden) and Zhejiang University], Zhejiang University, Hangzhou 310058, China

2Acreo AB collaborating within Kista Photonics Research Center, Electrum 229, SE-164 40 Kista, Sweden 3Kista Photonics Research Center, Royal Institute of Technology, Electrum 229, SE-164 40 Kista, Sweden

*Corresponding author: [email protected]

Abstract: The physics of electrically switched FBGs was studied in fibers with internal electrodes. Wavelength shifts due to mechanical effects (nanoseconds) and heat (milliseconds) depend quadratically on the electrical pulse voltage and linearly on pulse duration. ©2008 Optical Society of America OCIS codes: (060.3735) Fiber Bragg gratings; (060.4005) Microstructured fibers.

1. Introduction

A fiber Bragg grating (FBG) component based on a two-hole fiber with internal electrodes has showed 29 ns full off-on and on-off switching [1]. Index change and increased birefriengence result from the application of a short duration high current pulse to an internal electrode that heats up and expands rapidly. The nanosecond FBG response is attributed to a mechanical effect. As a consequence, the grating wavelength was blue-shifted (~ -27 pm) for the polarization orthogonal (S) to the direction of the hole, while it was red-shifted for the other polarization (~ +5 pm). Besides, a much slower evolution of the index (many microseconds) follows from the diffusion of heat from the metal into the glass cladding and core. A thermal shift as high as +17 pm was obtained for both polarizations. In this Paper, we report on the physics of electrically switched FBGs written in a two-hole fiber with internal electrodes, studying the dependence on voltage and pulse duration of both mechanical and thermal effects.

2. Experiment

Pieces of fiber with a pair of 28 μm diameter holes running parallel to the core were filled with BiSn alloy that occupied the entire cross-section of the holes [2]. Fig. 1a shows the cross section of the fiber used here. A 4 cm-long hamming-apodized FBG was written along the middle section of fiber with 7 cm-long alloy. Due to the intrinsic geometrical birefriengence of the used fiber, the grating has two reflection peaks with the separation of 38 pm. The 3-dB bandwidth and the reflectivity of each peak are 44 pm and ~ 75%, respectively (see Fig. 1b). After writing the gratings, electrical contact was made from the side of the fiber, thus allowing for conventional low loss splicing. Then, the pieces of fiber were mounted on an aluminum substrate and covered with heat conductive, electrically isolating epoxy. Heat dissipation allows switching the devices at rates of a few kHz without significant drift of polarization or wavelength. Electrical SMA contacts were used and a 0.1 Ω resistive probe was inserted in series with the ~50 Ω alloy electrode for monitoring purposes.

A pulse generator consisting of a high voltage dc power supply and a high speed switch was employed for the driving pulses. The switch here was a low repetition rate spark gap circuit delivering 4 ns rise-time current pulses with the maximum switching voltage 3 kV into 50 Ω and duration 10-240 ns determined by the cable length used. A tunable external cavity diode laser (TLS), polarization controller (PC), circulator, 3-dB coupler, high-speed photodiode (PD) and oscilloscope completed the experimental set-up, schematically illustrated in Fig. 1c.

Fig. 1. (a) Cross section of the metal-filled fiber (SEM picture). (b) Reflection spectrum of the FBG. (c) Experimental set-up. TLS: tunable laser source; PC: polarization controller; PD: photodiode; OSA: optical spectrum analyzer.

TLS

OSA

Oscillo-scope

0.1 Ω

High Voltage

Switch

PC

Coupler

PD

(b)

(a)

yx

(c) Circulator

Grating50 Ω

1547.20 1547.25 1547.30 1547.350.0

0.1

0.2

0.3

Ref

lect

ion

(mW

)

Wavelength (nm)

S polarization

a1371_1.pdf

© 2008 OSA / CLEO/QELS 2008 CMK1.pdf

Fig. 2. (a) Experimental time evolution of switching on-off on a microsecond scale. (b) Illustration of the FBG shifting in the above process. (c-f) Experimental data and fit curves of the wavelength shifts due to mechanical stress and heat.

Dynamic measurements were performed for the S polarization by tuning the TLS at the grating reflection peak before the application of the current pulse and studying the changes in reflectivity as a function of time (Fig.2a). Fig. 2b illustrates how the grating wavelength shifts in the above process: blue-shifting due to mechanical stress (ΔλM) in tens of nanoseconds and red-shifting due to heat (ΔλT) in hundreds of microseconds. The values of ΔλM and ΔλT depend on the pulse amplitude (U) and the duration (Δt). Experimental data of ΔλM and ΔλT for a fixed pulse duration Δt = 241ns at different voltages are shown with black squares and blue triangles in Fig. 2c and 2d. The red curves in Fig. 2c and 2d are the parabolic fits ΔλM (pm) = -6.5×10-5 U2 and ΔλT (pm) = +4.5×10-5 U2, for the mechanical and thermal effect, respectively. The best fit to the experimental data is obtained for a power coefficient 2.08 in both cases. Experimental data of ΔλM and ΔλT for a fixed voltage pulse amplitude U=1kV (20 A pulse) under different pulse durations are shown with black squares and blue triangles in Fig. 2e and 2f. The red curves in Figs. 2e and 2f are the linear fits ΔλM(pm) = -0.244×Δt(ns) and ΔλT (pm) = +1.66×Δt(ns), respectively.

The conclusion drawn from the experiments is that the index change due to the mechanical and the thermal effects is proportional to the heat deposited in the metal electrode, which can be described by Q = U2 Δt/R = mmcmΔT, where R is the resistance of the electrode (50Ω by construction). This is valid even when no heat transport to the glass is involved, as seen in Fig. 2c and 2e. Dilation of the metal (with coefficient αm) can be related to the stress εx and εy in the direction parallel and perpendicular to the hole through εx = (Ym/Yg)αmΔT and εy = -0.17εx in a model constructed with springs of different stiffness describing the metal (Ym=12×1010 N/m2) and glass (Yg=72×1010 N/m2). The expected change in index due to mechanical stress is then ΔλM = -λ0n2(0.26εx + 0.126εy)/2 ~ ΔT ~ U2Δt, where the strain-optical constants and Poison ratio are from Ref. [3]. The model gives ΔλM (pm) = -7.8×10-5U2 (V2), in reasonable agreement with the experimental data (Fig. 2c). Simulation of the thermal effect requires solving the temperature distribution in the fiber as a function of time, following the application of the heating current pulse. A simplification was attempted of assuming an average temperature across the core at the instant of maximum index change (t3 in Fig. 2a) with the heat transferred from the metal to the core described by constant η. From this simple model ΔλT = (λ0/n) dn/dT ΔTg = (λ0/n) dn/dT (η/mgcg) U2Δt/R, i.e., the correct functional dependence ΔλT ~ U2Δt is obtained. Using one point of Fig 2d to fit the wavelength shift to the applied voltage, η is estimated to be 0.12, corresponding to a core temperature increase limited to ~ 5% of the maximum metal electrode temperature increase. The value used for the expansion coefficient of the metal was αm = 16 ×10-6K-

1 and for the specific heat cm=167 J/(kgK) (metal) and cg=703 J/(kgK) (glass).

References [1] Z. Yu, W. Margulis, O. Tarasenko, H. Knape, P.-Y. Fonjallaz, “Nanosecond switching of FBG”, Opt. Express 15, 14948-14953 (2007). [2] M. Fokine, L. E. Nilsson, Å. Claesson, D. Berlemont, L. Kjellberg, L. Krummenacher, and W. Margulis, “Integrated fiber Mach-Zehnder interferometer for electro-optic switching,” Opt. Lett. 27, 1643-1645 (2002). [3] N. F Borrelli and R. A. Miller, “Determination of the Individual Strain-Optic Coefficients of Glass by an Ultrasonic Technique”, Appl. Opt. 7, 745-750 (1968).

0 1 2

0

10

20

30

40 t4

t3

t2

t1

Am

plitu

de (m

V)

T im e (m s)

t0

(a )

1547.20 1547.24 1547.28 1547.32

λTLS

t4

Ref

lect

ion

(a. u

.)

Wavlength (nm)

26 pm t3

t2

t1

- 40 pm

t0 (b)0

20

40

605 2( ) 4.5 10T pm Uλ −Δ = ×

5 2( ) 6.5 10M pm Uλ −Δ = − ×

241t nsΔ =

(d)

thermal

0

10

20

30

thermal

1U kV=

(f)

0 200 400 600 800 1000

-60

-40

-20

0

mechanicalW

avel

engt

h sh

ift (p

m)

Voltage (V)

(c)

0 20 40 60 80 100-30

-20

-10

0

( ) 0.244 ( )M pm t nsλΔ = − ×Δ

( ) 0.166 ( )T pm t nsλΔ = ×Δ

mechanical

(e)

Pulse duration (ns)

a1371_1.pdf

© 2008 OSA / CLEO/QELS 2008 CMK1.pdf

Paper Ⅴ

Birefringence switching of Bragg gratings in fibers with internal electrodes

Zhangwei Yu1,3, O. Tarasenko2,3, W. Margulis2,3* and P. -Y. Fonjallaz2,3 1JORCEP [Joint Research Centre of Photonics of the Royal Institute of Technology (Sweden) and Zhejiang

University], Zhejiang University, Hangzhou 310058, China 2Acreo AB, Electrum 229, SE-164 40 Kista, Sweden

3Kista Photonics Research Center, Royal Institute of Technology, Electrum 229, SE-164 40 Kista, Sweden *Corresponding author: [email protected]

Abstract: A fiber Bragg grating was written in a side-hole fiber with internal metal alloy electrodes. The initial geometrical birefringence of this fiber gives rise to two Bragg resonances separated by 43 pm. Nanosecond risetime current pulses of up to 23 A were applied to the metal electrode, which heated and expanded rapidly. This caused mechanical stress in the fiber on a nanosecond scale, resulting in a negative shift of the Bragg wavelength peak for the fast axis mode, and positive but smaller shift for the slow axis mode. The fast change increased the peak separation to ~ 143 pm, corresponding to an increase in birefringence from 4.0×10-5 to 1.3×10-4. Both peaks subsequently experienced a red-shift due to the relaxation of mechanical stress and the increasing core temperature transferred from the metal in many microseconds. Simulations give accurate description of the experimental results.

©2008 Optical Society of America

OCIS codes: (060.3735) Fiber Bragg gratings; (060.4005) Microstructured fibers.

References and links

1. H. M. Xie, Ph. Dabkiewicz, R. Ulrich, and K. Okamoto, “Side-hole fiber for fiber-optic pressure sensing,” Opt. Lett. 11, 333-335 (1986).

2. M. Turpin, M. Brevignon, J. P. Le. Pesant, and O. Gaouditz, “Interfero-polarimetric fiber optic sensor for both pressure and temperature measurement,” in Proceedings of IEEE Conference on Optical Fiber Sensors (Institute of Electrical and Electronic Engineers, New York, 1992), pp. 362-365.

3. W. J. Bock, M. S. Nawrocka, and W. Urban´ czyk, “Highly sensitive fiber-optic sensor for dynamic pressure measurements,” IEEE Trans. Instrum. Meas. 50, 1085-1088 (2001).

4. S. Kreger, S. Calvert, and E. Udd, “High pressure sensing using fiber Bragg grating written in birefringent side hole fiber,” in Proceedings of IEEE Conference on Optical Fiber Sensors, (Portland, 2002), pp. 355-358.

5. R. J. Schroeder, T. Yamate, and E. Udd, “High pressure and temperature sensing for the oil industry using fiber Bragg gratings written onto side hole single mode fiber,” Proc. SPIE 3746, 72-74 (1999).

6. T. Yamate, R. T. Ramos, R. J. Schroeder, and E. Udd, “Thermally insensitive pressure measurements up to 300 degrees C using fiber Bragg gratings written onto side-hole single mode fiber,” Proc. SPIE 4185, 628-631 (2000).

7. E. Chmielewska, W. Urbanczyk, and W. J. Bock, “Measurement of pressure and temperature sensitivities of a Bragg grating imprinted in a highly birefrigent side-hole fiber,” Appl. Opt. 42, 6284-6291 (2003).

8. C. Jewart, K. P. Chen, B. McMillen, M. M. Bails, S. P. Levitan, J. Canning, and I. V. Avdeev, “Sensitivity enhancement of fiber Bragg gratings to transverse stress by using microsturctural fibers,” Opt. Lett. 31, 2260-2262 (2006).

9. L. -E. Nilsson, Z. Yu, O. Tarasenko, H. Knape, P. -Y. Fonjallaz, and W. Margulis, “High-speed switching in fibres with electrodes,” in Proceedings of IEEE Conference on Transparent Optical Networks (Rome, 2007), pp. 232-233.

10. Z. Yu, W. Margulis, O. Tarasenko, H. Knape, and P.-Y. Fonjallaz, “High-speed control of fiber Bragg gratings,” in Bragg Gratings, Photosensitivity, and Poling in Glass Waveguides, OSA Technical Digest (CD) (Optical Society of America, 2007), paper JWA44.

11. Z. Yu, W. Margulis, O. Tarasenko, H. Knape, and P. -Y. Fonjallaz, “Nanosecond switching of fiber Bragg gratings,” Opt. Express 15, 14948-14953 (2007).

12. Z. Yu, W. Margulis, P.-Y. Fonjallaz, and O. Tarasenko, “Physics of electrically switched fiber Bragg gratings,” in Conference on Lasers and Electro-Optics/Quantum Electronics and Laser Science Conference and Photonic Applications Systems Technologies, OSA Technical Digest Series (CD) (Optical Society of America, 2007), paper CMK1.

(C) 2008 OSA 26 May 2008 / Vol. 16, No. 11 / OPTICS EXPRESS 8229#94010 - $15.00 USD Received 18 Mar 2008; revised 19 May 2008; accepted 19 May 2008; published 21 May 2008

13. M. Fokine, L. E. Nilsson, Å. Claesson, D. Berlemont, L. Kjellberg, L. Krummenacher, and W. Margulis, “Integrated fiber Mach-Zehnder interferometer for electro-optic switching,” Opt. Lett. 27, 1643-1645 (2002).

14. J. A. Viator, S. Kreger, M. W. Winz, and E. Udd, “Modeling and experimental strain measurements on a non-homogeneous cylinder under transverse load,” Proc. SPIE 5384, 199-205 (2004).

15. R. Gafsi and M. A. El-Sherif, “Analysis of Induced-Birefringence Effects on Fiber Bragg Gratings,” Opt. Fiber Technol. 6, 299-323 (2000).

16. N. F Borrelli and R. A. Miller, “Determination of the individual strain-optic coefficients of glass by an ultrasonic technique,” Appl. Opt. 7, 745-750 (1968).

17. H. Knape and W. Margulis, “All-fiber polarization switch,” Opt. Lett. 32, 614-616 (2007).

1. Introduction

Side-hole fibers have the advantage of high sensitivity to pressure and low sensitivity to temperature [1-4]. Fiber Bragg gratings (FBG) imprinted in side-hole fibers have been shown to be good candidates for simultaneous sensing hydrostatic pressure and temperature [5-7]. Recently Kreger, et al., [8] has reported that when the static transverse load was applied perpendicular or parallel to the direction of the two air holes, the direction orthogonal to the two holes suffered a much larger stress or strain, and the polarization parallel to the two holes has a larger wavelength shift. However, to the best of our knowledge, there is no published report on high-speed dynamic measurements of pressure and temperature in these special fibers until now. The potential application of high-speed polarization/wavelength switching in fiber lasers makes these studies particularly relevant.

In our previous work [9-11], we reported that a FBG component based on a side-hole fiber with internal electrodes has shown nanoseconds full off-on and on-off switching. Nanosecond high current pulses heated the metal electrodes and caused rapid metal expansion. As a consequence, the grating wavelength was blue-shifted (~-27 pm) for the fast axis mode, while it was red-shifted for the slow axis mode (~+5 pm). Besides, a much slower wavelength red-shift (many microseconds) as high as ~+17 pm was obtained for both polarizations. Furthermore, we demonstrated that the fast and slow wavelength shifts depend quadratically on the electrical pulse voltage and linearly on the pulse duration for the fast axis mode [12], i.e. depend linearly on the amount of energy deposited.

In this Paper, we further investigate the birefringence dynamic changes of side-hole fibers with internal electrodes. The FBG used show to be an excellent tool to reveal the amplitude and sign of the mechanical and thermal perturbations, and the ensuing time evolution of the birefringence. The simulation results are consistent with the experimental dynamic measurements.

2. Experiment

The 125 μm-thick silicate side-hole fiber had characteristics similar to those of the standard telecom fiber (core diameter 8.4 μm and Δn = 0.0056) but with two 28.8 μm-diameter holes running parallel to the core. The core-hole separation (edge-to-edge) was 13.4 μm. Figure 1(a) shows the cross section of the fiber used. Pieces of fiber were filled with 7 cm-long BiSn alloy (melting temperature 137°C) that occupy the entire cross-section of the holes [13]. The length of the alloy-filled holes and the core-hole separation were such that insertion loss was experimentally measured to be < 0.2 dB and the polarization-dependent loss < 0.1 dB. Both ends of the metal-filled fiber were free from metal to allow for convenient low-loss splicing.

A 4 cm-long Hamming-apodized FBG was written along the middle section of BiSn-filled fiber, with the holes aligned perpendicular to the UV beam incidence direction [11]. The grating was annealed in the oven at 100 °C for 12 hours. Due to the intrinsic geometrical birefringence of the fiber with internal metal electrodes, the grating has two reflection peaks at different wavelengths corresponding to the differing effective refractive indices for the two orthogonal polarization modes in this fiber. The reflection peak at the shorter wavelength takes place in the fast axis mode (x-axis), the one at the longer wavelength in the slow axis mode (y-axis). The identification of the axes is discussed in the simulations below. At room temperature, the Bragg wavelengths of the two polarizations are 1547.230nm and 1547.273nm, with a separation of 43 pm corresponding to a refractive index difference of


~4.0×10-5. This is roughly one order of magnitude lower than in typical high-birefringence fibers. The 3-dB bandwidth and the reflectivity of both peaks are 44 pm and 75%, respectively.

After writing the gratings, electrical contacts were made from the side of the fiber at both extremities of the electrode. The grating was mounted on an aluminum substrate and covered with heat conductive, electrically isolating epoxy. Heat dissipation allows switching the devices at a few kHz with minor average drift of polarization. The resistance of the components was typically 43 Ω for a BiSn device with 7 cm active length. This value implies in a small impedance mismatch between the device and the 50 Ω coaxial cable transporting the high voltage (i.e., current) pulses. Electrical SMA contacts were used and a 0.1Ω resistive probe was connected in series with the 43 Ω loads for monitoring purposes.

A pulse generator consisting of a high voltage dc power supply and a high speed switch was employed for the driving pulses. The switch here was a low repetition rate spark gap circuit delivering 2 ns rise-time current pulses with the maximum switching voltage 3 kV and duration 10-241 ns determined by the cable length used. The light source was a tunable external cavity diode laser with which 1 pm wavelength resolution measurements are possible. Different input polarization states were selected with a polarization controller. A circulator, 3-dB splitter, high-speed photodiode, Ando AQ6317B optical spectrum analyzer and oscilloscope completed the experimental set-up, schematically illustrated in Fig. 1(b).

Fig. 1. (Color online) (a) Cross section of the two-hole fiber used here (SEM picture). (b) The schematic of the system for dynamic measurements of the two polarization states.

Dynamic measurements were performed for the fast axis mode by properly aligning the input polarization and by tuning the external cavity laser source to the grating reflectivity maximum before the application of the current pulse. The changes in reflectivity as a function of time can be studied from the oscilloscope trace following the electrical excitation (Fig. 2(a)). Figure 2(b) schematically illustrates that the grating wavelength shifts in the above process can be divided into fast part (in tens of nanoseconds) and slow part (hundreds of microseconds). The former is due to mechanical stress created during the electrical pulse action, while the latter is due to the relaxation of mechanical stress and the slow increase of temperature in the core transferred from the metal electrode after each electrical pulse. The fast part corresponds to the initial reduction of the reflection of the grating at the probe wavelength (from t0 to t1). It reaches its maximum absolute value (ΔλM

x) at instant t1 ~241 ns, equal to the current pulse duration. Then the slow part takes effect: the grating gradually moves back with the decreasing mechanical stress and the increasing temperature in the core. The grating passes its original spectral position at instant t2 ~68 μs and red-shift continues. At t3 ~270 μs, the maximum positive ΔλT+M

x is reached, where the core temperature is near its maximum. After several milliseconds (t4), the grating returns to its original spectral position again, since the mechanical stress is released and the component is back to room temperature. It should be mentioned that the identification of the behavior schematically illustrated in Fig. 2(b) (and in Fig. 2(d) below) involved experimentally examining the response of the device for various other probe wavelengths, as discussed in Ref. [11].

y

x

Tunable external cavity

diode laser

Pol. control

Splitter

(a)

(b) Circulator

Photodiode

Optical Spectrum Analyzer

0.1Ω

FBG

HV

Switch

Oscilloscope

43 Ω


Similarly, dynamic measurements were also carried out for the slow axis mode, shown in a typical experimental trace (Fig. 2(c)) and the schematic diagram (Fig. 2(d)). Note that, now a positive and smaller wavelength shift happens during the electrical pulse duration and reaches its maximum (ΔλM

y) at instant t1 ~241 ns. Then, the relaxation of the mechanical stress blue-shifts the spectrum of the grating, while the increasing temperature in the core red-shifts the grating peak. In the beginning, the former dominates and the Bragg wavelength moves back a little until instant t2 is attained at ~16 μs. Afterwards, the heat reaching the core dominates, so the grating continues to red-shift. The maximum positive spectral shift (ΔλT+M

y) occurs at instant t3 ~190 μs. Finally, the grating also returns to its original spectral position after instant t4 ~20 ms. Full (100%) on-off switching is not achieved for this axis even with the highest current level (~23A) applied here, corresponding to an electrical pulse energy ~5 mJ/pulse.

The values of ΔλM and ΔλT+M depend on the pulse voltage and the pulse duration. Experimental data of the fast wavelength shift for both polarization states (ΔλM

x and ΔλMy) at

different voltages are shown with blue and red squares in Fig. 3(a), for pulses with duration 241 ns. The absolute value of ΔλM

x is ~3.2 times larger than ΔλMy. Similarly, experimental

data of the slow wavelength shift (ΔλT+Mx and ΔλT+M

y) at different voltages are shown in Fig. 3(b) with blue and red squares. Here, ΔλT+M

y is larger than ΔλT+Mx.

Fig. 2. (Color online) (a) Time evolution of the signal reflected by the FBG for probe light polarized along the x-axis. (b) Schematic illustration of the wavelength shift at various times. (c) Time evolution of the signal reflected by the FBG for probe light polarized along the y-axis. (d) Schematic illustration of the wavelength shift at various times. The red arrows in Figs. (b) and (d) indicate the wavelength of the tunable laser source during the measurement.

1547.10 1547.15 1547.20 1547.25 1547.30

t4 = ~5 ms

Ref

lect

ion

(a. u

.)

Wavlength (nm)

ΔλT+Mx

t3 = ~270 us

t2 = ~68 us

ΔλMx t1 = ~241 ns

t0 = 0

(b)

x-axis

0.0 0.5 1.0 1.5

0

10

20

30

40

t4

t3

t2t1

t0

Am

plitu

de (

mV

)

Time (ms)

(c)

y-axis

yx

-0.5 0.0 0.5 1.0 1.5 2.0 2.5

0

10

20

30

40 t4

t1

t3

t2

Am

plitu

de (

mV

)

Time (ms)

t0x-axis

(a)

1547.20 1547.25 1547.30 1547.35

t4 = ~20 ms

Ref

lect

ion

(a. u

.)

Wavlength (nm)

ΔλT+My

t3 = ~190 us

t2 = ~16 us

ΔλMy

t1 = ~241 ns

t0 = 0

(d)

y-axis

y

x


0 200 400 600 800 1000

-60

-40

-20

0

20

ΔλMx (x-axis)

Fas

t wav

elen

gth

shift

(pm

)

Applied voltage (V)

ΔλMy (y-axis)

(a)

0 200 400 600 800 1000

0

20

40

60

80

ΔλT+Mx (x-axis)

Slo

w w

avle

ngth

shi

ft (p

m)

Applied voltage (V)

ΔλT+My (y-axis)

(b)

Fig. 3. (Color online) Experimental data (squares) and numeric simulations (solid lines) of Bragg wavelength shift for a 241 ns electrical pulse. (a) Fast wavelength shift for both polarization states measured at instant t1. (b) Slow wavelength shift measured at instant t3.

3. Simulation

To better study the physics of the Bragg wavelength shift for both polarizations, the partial differential equations describing heat diffusion and displacement in x and y-direction were solved by the finite element technique using the commercial software FlexPDE. The solutions allowed finding the time-dependent strain components in the fiber in the x- and y-directions (εx and εy), the temperature distribution over the core and to calculate the FBG wavelength shift in x- and y-polarizations according to the following formula [14-15]: ( ) ( )

20

1 1 1 22x x x y x

np p Tλ λ ε ε λ α ξΔ = − + + + Δ (1)

( ) ( )20

11 122y y y x y

np p Tλ λ ε ε λ α ξΔ = − + + + Δ (2)

where: λx=1547.23nm, λy=1547.273nm are the initial Bragg wavelength for x- and y-polarizations; n0 = 1.445 is the initial effective refractive index of the core; p11 = 0.121 and p22 = 0.270 are the photoelastic coefficients of the fiber [16]; α is the thermal expansion coefficient of silica; ξ = (1/n)(dn/dT) = 6.5×10-6 K-1 is the thermo-optic coefficient of fused silica; ΔT is the increased temperature in the core. The first terms of Eqs. (1) and (2) stand for the wavelength shift of both polarizations due to mechanical stress in the core, while the second terms of two equations indicates the wavelength shift caused by the increased temperature in the core.

Stress at the BiSn/cladding boundary increase approximately linearly during the electrical pulse action, and then gradually decrease with time. This stress is transferred to the core area in ~14 μm/5720 m/s = 2.45 ns, i.e. almost instantaneously compared to the pulse duration 241 ns. However, the heat at the BiSn/cladding interface needs several microseconds to reach the core area. As a result, the heat diffusion can be neglected during the pulse action. The fast wavelength shifts ΔλM

x and ΔλMy are determined by the first terms of Eqs. (1) and (2) at t =

241 ns. The exact value of the heat capacity and the expansion coefficient of the BiSn alloy electrode are not known, nor the exact heat capacity for the Ge-doped silica fiber used. The choice Cv = 167 J/kgK and Δl/l = 15.5×10-6 K-1 for the alloy and Cv = 736 J/kgK for the fiber allowed for excellent fitting of the theoretical equations (1) and (2) to the experimental data, as seen in Fig. 3(a). The simulations confirm that the core area is compressed (negative) in x-direction and stretched (positive) in y-direction under the applied pulses. For the pulse with applied voltage 1080 V, εx = -1.655×10-4 and εy = 2.535×10-4 at t = 241 ns. This results in a decrease of refractive index in x-direction (-7.3×10-5) and a weaker increase of refractive index in y-direction (+2.11×10-5), corresponding to ΔλM

x = -75.8 pm and ΔλMy = +23.7 pm.

We can conclude (in opposition to [9, 12]) that the fast axis mode is in the direction parallel to the two holes (x-direction), and the slow axis mode is in the direction orthogonal to the two holes (y-direction).


As mentioned above, the slow wavelength shift is caused by the relaxation of mechanical stress and the increased temperature in the core. It can be calculated from Eqs. (1) and (2) for both polarizations. For the x-polarization, Δλx equals zero at a calculated time t2 = ~56 μs, when the two terms of Eq. (1) compensate each other. Subsequently, the second term dominates and Δλx becomes positive and increases gradually due to the increasing ΔT. The shift reaches its maximum at a calculated time t3 = ~250 μs. For the y-polarization, initially Δλy reduces because the decay of the stress dominates over the rising effect of heat. At a calculated time t2 ~17 μs the latter starts to dominate and the shift increases again until it reaches the maximum value at a calculated time t3 ~ 200 μs.

The simulations of ΔλT+Mx accurately reproduce the experiment measurements when the

thermal expansion coefficient for the silica fiber is optimized to the value α = 0.50×10-6 K-1. This is shown by the blue line and squares in Fig. 3(b). The simulation of the wavelength shift for the y-axis now has no free parameters for adjustment. The result (red line in Fig. 3(b)) fits the experimental data (squares) relatively well. The numerical predictions are smaller than the measured values. The small discrepancy could possibly be attributed to a temperature variation taking place during the measurement. For example, the difference between the calculated and measured ΔλT+M

y at the pulse voltage 1015 V corresponds to a temperature shift of 0.8 oC, well within the temperature uncertainty in the laboratory, which could have been rising during the measurement.

The complete simulations here agree well with a simple quadratic fit to the data for ΔλM and ΔλT+M (for both polarizations) as a function of the pulse voltage [12].

Figure 4 illustrates the simulation results of the time evolution of the temperature in the core center under different pulse excitation. The voltages shown in Fig. 4 are obtained multiplying the applied voltage by 2×43/(43+50) to take into account the impedance mismatch between the device and the 50 Ω coaxial cable. As shown in Fig. 4, the heat in the metal takes ~10 μs to reach the core and increases to the maximum at ~190-220 μs. Note that the maximum ΔT in the metal is as high as ~90 ºC for a corresponding core temperature increase of only ~6.2 ºC.

0 50 100 150 200 250 3000

1

2

3

4

5

6

7

792 V

285 V

391 V489 V597 V

691 V

890 V

ΔT (

0C

)

Time (μs)

1001 V

Fig. 4. (Color online) Simulation results of the time evolution of the temperature in the center of the core under different voltage pulse excitation.

4. Conclusion

A FBG was written in a side-hole fiber with internal metal alloy electrodes, resulting in two reflected peaks separated by 43 pm due to the intrinsic geometrical birefringence. Nanosecond risetime current pulses of up to 23 A were applied to the metal electrode, which heated and expanded rapidly. On a nanosecond scale, the fast Bragg wavelength shift of the fast axis mode is negative and that of the slow axis mode is positive and smaller, because of the mechanical perturbation. Microseconds later, the relaxation of the mechanical stress and the effect of heat reaching the core red-shift both Bragg peaks. All wavelength shifts depend quadratically on the electrical pulse voltage. Numerical simulations were developed to accurately and quantitatively explain the experimental observations. Similar components to the one used here have been tested to 109 pulses without degradation of performance [17].


This component has potential applications for Q-switching fiber lasers and RF signal generation.


Paper Ⅵ

We.C1.4 232 ICTON 2007

1-4244-1249-8/07/$25.00 ©2007 IEEE

High-Speed Switching in Fibres with Electrodes L-E. Nilsson1, Z. Yu2,3, O. Tarasenko1, H. Knape1, P-Y. Fonjallaz1,2, W. Margulis1

1 Acreo, Electrum 236, 16440 Stockholm, Sweden 2 Royal Institute of Technology, Electrum 239, 16440 Stockholm, Sweden

3 Zhejiang University, Zijingang campus, Zhejiang University, Hangzhou 310058, China Tel: (4670) 7603843, Fax: (468) 6327710, e-mail: [email protected]

ABSTRACT Fibres with holes are provided with metal electrodes that fill the entire cross-section of the holes. Refractive index change and birefringence result from applying high current pulses for time intervals in the nanosecond range. High-speed polarization rotation and fibre Bragg grating wavelength tuning can be accomplished. Such electrically driven components are totally spliced with low insertion loss. They have no moving parts, are < 10 cm long and can be used in fibre laser cavities for Q-switching. Keywords: Fibre component, microstructured fibre, internal electrodes, high-speed switching, electrical control,

fibre lasers.

Highly functional optical fibre components exploit the intrinsic advantages that optical fibres have of low loss, high power damage threshold, the stability of silica as base material and above all the potential of low cost. Over the last few years, our group has carried out research on fibre components based on electrical excitation to control the light in the fibre. To this end, we have used fibres with holes disposed longitudinally parallel to the fibre core. These holes are filled with metal by means of a technique [1] briefly summarized with help of Fig. 1. A low melting temperature alloy such as BiSn (Tm = 137 °C) is pumped at high pressure into the fibre orifices as a liquid, and upon solidification used as electrode. Fibre sections of length ranging from a few centimetres to a few meters can be filled in this way. The final position of the metal column along the piece of fibre is also adjusted with high pressure, so that the fibre ends are free from metal. Electrical contact to the metal is made by side-polishing the fibre, so that the ends can be spliced with low loss to subsequent optical fibre. Besides allowing for the fabrication in parallel of a number of devices and thus reducing the final cost of the components, the filling technique we use also guarantees a good physical contact between the metal column and the walls of the holes. This in turn allows us to exploit the expansion of the electrodes under the application of current - be it dc current or short pulses - to induce electrically controlled birefringence. By adjusting the current one can impose an adjustable degree of polarization rotation. Typically, the devices developed are driven with tens of milliamps, and for an electrode with resistance ~15 - 50 Ω this corresponds to voltages 0 - 5 V. In such application, the electrically driven section of the fibre is generally 2 - 5 cm long. Full three-stage polarization controllers with drive and control electronics have been constructed with insertion loss 0.5 dB, and can reach any possible state of polarisation (SoP) for an arbitrary input SoP [2].

Figure 1. Filling a fibre with metal electrodes. The inset shows a SEM picture of one fibre used [1].

Since the devices rely on the heating of the electrode to cause heat expansion, it may be assumed that their operation is limited to the range of millisecond or slower, but this is not the case. Heat dissipation indeed determines the fall time of such components. With proper heat dissipation and operating above room temperature, the fall time can be brought to as low as ~20 µs (50 kHz). More importantly, the risetime is only limited by the time required to deposit heat into a metal wire, and with high current this can be a few nanosecond or less. We have demonstrated 90° polarization rotation in 7-cm long elements with risetime ~4 ns, driven by 20 A current pulses [2]. The temperature rise associated with every drive pulse is <10 °C. A typical example of the response achieved is illustrated in Fig. 2. The trace on the left represents the time evolution of

Fibre with holes

High pressure

Liquid metal

High temperature

ICTON 2007 233 We.C1.4

0 100 200 300-0.2

0.0

0.2

0.4

0.6

0.8

1.0

1.2

Ele

ctric

al P

ulse

(kV

)

Time (ns)0 100 200 300

0.0

0.2

0.4

0.6

0.8

1.0

Acoustical Oscillations

Opt

ical

Res

pons

e (%

)Time (ns)

the electrical signal delivered by a spark gap to a device with electrode resistance 50 Ω. The trace on the right shows the optical signal measured by placing a polarisation switch between the polarised input beam at 1.55 µm and an analysing crossed polariser. Approximately 100% switching is obtained. Three physical contributions are identified, namely a mechanical component that is responsible for the fast switching shown in Fig. 2, an acoustical component that causes ringing in the nanosecond time scale if the driving pulse is short, and a thermal component that operates on the microsecond regime and is responsible for the slow decay of the signal (not apparent in the limited time range of Fig. 2).

Figure 2. Time evolution of electrical drive pulse (left) and optical signal switched between crossed polarisers (right). The mechanical (fast) and acoustic (oscillating) contributions are clearly identified.

The fibre is affected mechanically for as long as the metal electrode is heated, even if the temperature of the core is unchanged. Therefore, the switch remains “on” for a long time interval compared to the risetime (see Fig. 2). The components thus constructed find applications in Q-switching fibre lasers, where the fast risetime and long switch off time provide for sufficient delay to exhaust the laser cavity from population inversion. Other useful characteristics of such components are the low insertion loss (< 0.2 dB) and short length (does not affect the cavity roundtrip significantly).

The induced birefringence is associated to a refractive index change in the fibre core. If a Bragg grating (FBG) is recorded in the fibre section with electrodes, it is possible to shift the wavelength of the grating by means of the electrical drive pulse. Also here the risetime can be in the nanosecond range. Preliminary experiments were carried out with a 4-cm long unchirped FBG with ~20 pm bandwidth. The SoP along the direction of the holes and the one orthogonal to that experience a slightly different refractive index. This causes a splitting of the reflected light into two wavelength peaks, each one associated with one polarisation. At room temperature this splitting is only a few pm, and the two peaks overlap, but at 67 °C they separate by as much as 80 pm. We found that upon application of the nanosecond electrical pulse, the spectral separation between peaks increases, and while one peak is shifted to shorter wavelengths the other one shifts to longer. This is tentatively attributed to the compression of the core in the direction of the holes due to metal expansion, and to a distension of the core in the orthogonal direction. Such behaviour is expected, according to the simulations carried out to evaluate the elasto-optical effect throughout the cross-section of the fibre.

The wavelength switch described here and embodied by an electrically tuned FBG in a fibre with internal electrodes – like the polarisation switch – finds use in Q-switching of fibre lasers.

ACKNOWLEDGEMENTS The authors would like to acknowledge partial financial support from Fiberlogix Ltd.

REFERENCES [1] M. Fokine, L.E. Nilsson, Å. Claesson, D. Berlemont, L. Kjellberg, L. Krummenacher, and W. Margulis:

“Integrated fiber Mach-Zehnder interferometer for electro-optic switching”, Opt. Lett. 27, 1643 (2002). [2] O. Tarasenko, Å. Claesson, W. Margulis, “All-fibre polarisation control”, in Proc. ACOFT 05, Sydney,

Australia, July 2005.

Paper Ⅶ

Switchable dual-wavelength Ramanerbium-doped fibre laser

D.R. Chen, Z.W. Yu, S. Qin and S.L. He

A fibre Bragg grating feedback fibre laser with both Raman and

erbium-doped fibre pumps is proposed. Dual-wavelength switching is

achieved by controlling the power of the Raman pump. The char-

acteristics of the dual-wavelength switching are studied experimen-

tally, and the mechanism is explained physically.

Introduction: Wavelength-switchable fibre lasers have attracted

considerable attention recently owing to their important applications

in WDM fibre communication systems, fibre sensors, spectroscopy

and optical instrument testing, etc. Several techniques have been

reported to achieve wavelength switching in erbium-doped fibre

(EDF) lasers, such as the use of cascaded fibre Bragg gratings

(FBGs) [1], spectral polarisation-dependent loss elements [2], the

Sagnac loop reflector [3], few-mode fibre Bragg gratings [4],

multimode fibre Bragg gratings [5], acoustic waves [6], distributed-

feedback (DFB) mode oscillations [7] and the multisection

high-birefringence fibre loop mirror [8]. Wavelength-switchable

lasers based on some EDFs co-doped with other rare-earth irons

have also been reported [9]. Meanwhile, Raman fibre lasers have also

been developed for high power, multiwavelength (however, non-

switchable) applications (see e.g. [10–12]). Most of these techniques

are based on some special devices or manual adjustments to achieve

wavelength switching.

In this Letter, we propose a novel switchable dual-wavelength Raman

erbium-doped fibre laser. Dual-wavelength switching is achieved by

controlling the power of the Raman pump. To the best of our know-

ledge, the proposed fibre laser is the first to combine Raman amplifica-

tion and EDF amplification to achieve multiwavelength switching.

Operation principles and experimental results: Fig. 1 shows the

schematic configuration of the proposed switchable dual-wavelength

fibre laser. A 1480 nm laser diode (LD) with a maximum power of

about 250 mW is used as the EDF pump LD1, and a 1467 nm LD with

a maximum power of 350 mW (when the drive current is 950 mA) is

employed as the Raman pump LD2. The length of the singlemode

fibre (SMF) section (with an attenuation coefficient of 0.19 dB=km at

1550 nm) between the two FBGs is 10 km. Sections of dispersion

compensation fibre (DCF), SMF and EDF are the gain fibres in the

present structure. The DCF section (with total dispersion of

986 ps=nm and total loss of 5.7 dB) used here can compensate

60 km singlemode fibre. The length, the numerical aperture, the cutoff

wavelength and the peak absorption (at 1531 nm) of the EDF section

are 4 m, 0.25, 950 nm and 19.2 dB=m, respectively. Two FBGs are

used as wavelength selection components. The central wavelengths,

the full-widths at half maximum (FWHM) and the reflectivities are

(1545.13 nm, 0.093 nm, 50.0%) and (1559.98 nm, 0.098 nm, 51.3%)

for FBG1 and FBG2, respectively (these parameters are provided by

the manufacturer). A circulator and an isolator are used here to ensure

that the lasing cavity is a counter-clockwise ring.

Fig. 1 Schematic configuration of proposed fibre laser

There are two lasing cavities with wavelengths l1¼ 1545.13 nm

(around the EDF gain peak) and l2¼ 1559.98 nm (around the Raman

gain peak corresponding to pump wavelength 1467 nm). The power of

EDF pump LD1 is always fixed to a certain level, and the power of

Raman pump LD2 is used to control the output wavelength and power.

In our first experiment, the power of EDF pump LD1 is fixed to

150 mW. When Raman pump LD2 is switched off, the output spectra of

the laser is shown by the solid line in Fig. 2a, from which one sees that

the structure is lasing only at l1 with a peak power of 11.9 dBm.

However, when the drive current of LD2 is switched on to 750 mA, the

structure will be lasing only at l2 with a peak power of 1.9 dBm (see

dashed line in Fig. 2a). The solution and sensitivity of the optical

spectrum analyser (OSA) used in the measurement are 0.06 nm and

65 dBm, respectively. Our experimental results show that the laser

output is rather stable. Thus, wavelength switching is achieved in this

fibre laser by controlling the drive current of LD2 (i.e. the power of the

Raman pump LD2).

100

-10-20-30-40-50-60

outp

ut p

ower

, dB

m

drive current for LD2, mA0 100 200 300 400 500 600 700 800 900

b

100

-10-20-30-40-50-60ou

tput

pow

er, d

Bm

l, nm1540 1545 1550 1555 1560 1565 1570

a

A B

l1l2

C

Ia Ib

Fig. 2 Output spectra of proposed fibre laser, and output powers at twolasing wavelengths against drive current for LD2

Fig. 2b shows the detailed behaviour of the dual-wavelength switch-

ing as the driving current of LD2 increases (the pump power of LD1 is

fixed to 150 mW). The output powers at l1 and l2 are indicated by the

lines connected with squares and circles, respectively. Obviously, there

are three regions where the laser is in different operation modes. In

region A (with the drive current for LD2 below switching current

Ia¼ 70 mA), only l1 is lasing and its output power remains around

11.9 dBm. In region B (with the drive current for LD2 ranging from

70 to 690 mA), both l1 and l2 are lasing. In this region, as the drive

current of LD2 increases, the output power at l2 increases while the

output power at l1 decreases. In region C (with the drive current for

LD2 above switching current Ib¼ 690 mA), only l2 is lasing and its

output power remains around 1.9 dBm.

The operation principle of the present switchable dual-wavelength

fibre laser can be explained as follows. When the power of Raman

pump LD2 is zero, only lasing at l1 can be obtained since the cavity

loss at l2 is larger than that at l1 (owing to the loss of the 10 km SMF)

and the EDF amplification will occur only at l1 (owing to the gain

competition). When the power of LD2 increases gradually, the Raman

gain mainly contributes to the amplification at l2 since l2 is much

closer to the Raman gain peak wavelength than l1. Furthermore, when

the LD2 pump power is large enough, the Raman gain in the SMF

section can also increase the cavity gain at l2. Thus, the structure starts

lasing at l2 when the pump power of LD2 reaches a certain level.

Consequently, lasing at l1 decreases as the pump power of LD2

increases since the lasing at l2 will reduce the population inversion

in the EDF. When the pump power of LD2 increases further, lasing at l1

disappears finally.

Next we study the wavelength switching when the pump power of

LD1 is fixed to some other values. The inset of Fig. 3 shows the output

powers at the two lasing wavelengths as the drive current for LD2

increases when the pump power of LD1 is fixed to 50 or 100 mW. One

sees that the characteristics of the wavelength switching are similar

(except the switching currents Ia and Ib) when the pump power of LD1

is set to a different value. Fig. 3 shows the switching currents Ia and Ib

when the pump power of LD1 is set from 20 to 160 mW with an

increment of 20 mW. The switching current Ia decreases as the pump

power of LD1 increases. This can be explained as follows. The cavity

gain for l2 to start lasing is a constant, which is the sum of EDF gain

ELECTRONICS LETTERS 16th February 2006 Vol. 42 No. 4

and Raman gain. Thus, the larger EDF pump is set, the smaller Raman

gain is needed for l2 to start lasing (corresponding to smaller switching

current Ia). The variation of Ia under different pump powers of LD1 is

quite small, while the variation of Ib is quite large. As shown in Fig. 3,

the switching current Ib increases almost linearly with the pump power

of LD1. This indicates that a larger drive current for LD2 is needed to

let l2 win the gain competition over l1 when the pump power of LD1

increases.

ABCDEF

IaIb

900

800

700

600

500

400

300

200

100

0

switc

hing

cur

rent

, mA

pump power of LD1, mW0 20 40 60 80 100 120 140 160 180

0

-10

-20

-30

-40

-50

-600 200 400 600 800 1000

outp

ut p

ower

, dB

m

drive current for LD2, mA

Fig. 3 Switching currents Ia and Ib for LD2 when power of LD1 set todifferent values

Inset: Output powers at two lasing wavelengths as drive current for LD2 increaseswhen pump power of LD1 set to 50 mW (curves A, B), 100 mW (curves C, D),150 mW (curves E, F)

Conclusion: We have proposed and demonstrated a switchable dual-

wavelength Raman erbium-doped fibre laser. The wavelength switch-

ing is controlled by the power of the Raman pump and the operation

principle has been explained. The characteristics of the wavelength

switching under different EDF pump levels have been studied experi-

mentally. Unlike the other switchable multiwavelength fibre lasers,

which are based on only EDF amplification, the present structure is

simple (requiring no polarisation controller or special fibre=device)

and the separation of the dual wavelengths can be very large (since the

wavelength of the Raman pump can be arbitrary, which is impossible

for EDF amplification). The output power and wavelength of the

present switchable dual-wavelength fibre laser are repeatable and

uniquely determined by the Raman pump power (without any

hysteresis characteristic, which exists in some other switchable

multiwavelength fibre lasers reported in e.g. [1, 13]).

# IEE 2006 16 November 2005

Electronics Letters online no: 20064019

doi: 10.1049/el:20064019

D.R. Chen, Z.W. Yu, S. Qin and S.L. He (Center for Optical and

Electromagnetic Research, Zhejiang University, Joint Research

Center of Photonics of the Royal Institute of Technology, Sweden)

D.R. Chen, Z.W. Yu, S. Qin and S.L. He: Also with Zhejiang

University, East Building No. 5, Zijingang Campus, Zhejiang Univer-

sity, Hangzhou 310058, People’s Republic of China

S.L. He: Also with Alfven Laboratory, Royal Institute of Technology,

S-100 44 Stockholm, Sweden

E-mail: [email protected]

References

1 Mao, Q.H., and Lit, J.W.Y.: ‘Switchable multiwavelength erbium-dopedfiber laser with cascaded fiber grating cavities’, IEEE Photonics Technol.Lett., 2002, 14, pp. 612–614

2 Lee, Y.W., and Lee, B.: ‘Wavelength-switchable erbium-doped fiberring laser using spectral polarization-dependent loss element’, IEEEPhotonics Technol. Lett., 2003, 15, pp. 795–797

3 Das, G., and Lit, J.W.Y.: ‘Wavelength switching of a fiber laser with aSagnac loop reflector’, IEEE Photonics Technol. Lett., 2004, 16,pp. 60–62

4 Feng, X.H., et al.: ‘Switchable dual-wavelength ytterbium-doped fiberlaser based on a few-mode fiber grating’, IEEE Photonics Technol. Lett.,2004, 16, pp. 762–764

5 Su, L., and Lu, C.: ‘Wavelength-switching fibre laser based on multimodefibre Bragg gratings’, Electron. Lett., 2005, 41, pp. 11–13

6 Delgodo-Pinar, M., Dıez, A., Cruz, J.L., and Andres, M.V.: ‘Wavelength-switchable fiber laser using acoustic waves’, IEEE Photonics Technol.Lett., 2005, 17, pp. 552–554

7 Kim, Y.J., and Kim, D.Y.: ‘Electrically tunable dual-wavelengthswitching in a mutually injection-locked erbium-doped fiber ring laserand distributed-feedback laser diode’, IEEE Photonics Technol. Lett.,2005, 17, pp. 762–764

8 Hu, S., et al.: ‘Switchable multiwavelength erbium-doped fiber ring laserwith a multisection high-birefringence fiber loop mirror’, IEEEPhotonics Technol. Lett., 2005, 17, pp. 1387–1389

9 Liu, H.L., et al.: ‘Wavelength-switchable La-codoped bismuth-basederbium-doped fiber ring laser’, IEEE Photonics Technol. Lett., 2005,17, pp. 986–988

10 Han, Y.G., et al.: ‘Tunable multi-wavelength Raman fibre laser based onfibre Bragg grating cavity with PMF Lyot-Sagnac filter’, Electron. Lett.,2004, 40, pp. 1475–1476

11 Sim, S.K., et al.: ‘High-power cascaded Raman fibre laser usingphosphosilicate fibre’, Electron. Lett., 2004, 40, pp. 738–739

12 Lee, J.H., et al.: ‘Raman amplifier based long-distance, remote FBGstrain sensor with EDF broadband and source recycling residual Ramanpump’, Electron. Lett., 2004, 40, pp. 1106–1107

13 Mao, Q.H., and Lit, J.W.Y.: ‘Optical bistability in an L-band dual-wavelength erbium-doped fiber laser with overlapping cavities’, IEEEPhotonics Technol. Lett., 2002, 14, pp. 1252–1254

ELECTRONICS LETTERS 16th February 2006 Vol. 42 No. 4

Paper Ⅷ

Some Switchable Dual-wavelength Fibre Lasers Based on Fibre Bragg Grating Feedback

Daru Chen*, Zhangwei Yu, Shan Qin, and Sailing He

Centre for Optical and Electromagnetic Research, Department of Optical Engineering, Zhejiang University, Zijingang, Hangzhou 310058, China;

ABSTRACT

Two methods to achieve dual-wavelength switching in a fibre laser are proposed and two corresponding switchable dual-wavelength fibre lasers based on fibre Bragg grating (FBG) feedback are demonstrated in this paper. In one proposed fibre laser, both Raman and Erbium-doped fibre (EDF) pumps are employed and the dual-wavelength switching is achieved by controlling the power of the Raman pump. In the other proposed fibre laser, an injection technique is used and the dual-wavelength switching is realized by controlling the power of the injection laser. The detailed behavior of the dual-wavelength switching in the two fibre lasers is experimentally studied and the principle is explained physically.

Keywords: switchable, dual-wavelength, fibre Bragg grating, fibre laser

1. INTRODUCTION Wavelength-switchable fibre lasers have attracted considerable attention recently due to their important applications in wavelength-division multiplexer (WDM) fibre communication systems, fibre sensors, spectroscopy and optical instrument testing, etc. Several techniques have been reported to achieve wavelength switching in Erbium-doped fibre (EDF) lasers, such as the use of cascaded fibre Bragg gratings (FBGs) [1,2], few-mode fibre Bragg gratings [3], multimode fibre Bragg gratings [4, 5], spectral polarization-dependent loss elements [6], Sagnac loop reflector [7], acoustic waves [8], distributed-feedback (DFB) mode oscillations [9] and multisection high-birefringence fibre loop mirror [10]. Wavelength-switchable lasers based on some EDFs co-doped with other rare-earth irons have also been reported [11]. Most of these techniques are based on some special devices or manual adjustments to achieve wavelength switching.

As we all know, an EDF can be regarded as a homogenous gain media at the room temperature due to the homogenous line broadening of Erbium ions [12, 13]. This promises possible methods to achieve wavelength-switching in the EDF laser due to the gain competition between different wavelengths. In this paper, two methods to achieve wavelength-switching are proposed and the corresponding fibre lasers are demonstrated. One method is to introduce overlapping cavity (for two lasing wavelengths) and hybrid gain mechanism in a fibre laser. Lasing wavelength can be switched from one to the other when different pump condition is employed. In the corresponding fibre laser, an overlapping cavity is designed for lasing at two wavelengths (corresponding to the central wavelengths of two FBGs) and both EDF pump (for the EDF gain medium) and Raman pump (for the Raman gain medium) are employed. Dual-wavelength switching can be achieved by controlling the power of the Raman pump. The other method is to employ an injection technique in a normal EDF laser with FBG feedback. The lasing at the central wavelength of the FBG can be suppressed by the injection light (at a certain range of wavelength) with enough power into the cavity due to the gain competition between the lasing light and the injection light. In the corresponding fibre laser, a fibre Sagnac loop reflector and an FBG are employed to form a line structure cavity (like a Fabry–Perot cavity). The injection light is injected into the cavity through an optical circulator. Dual-wavelength switching is achieved by controlling the power of the injection laser. It is worthewhile to note that in the two proposed methods, wavelength switching is achieved by controlling the power of the pump or the injection laser and therefore this type of wavelength switching can be called optically controlled wavelength switching, which can be expected to be one type of fast wavelength switching. Our experimental results in the proposed fibre laser with an injection technique have shown the switching time is less than 50 microsecond.

Passive Components and Fiber-based Devices III, edited by Sang Bae Lee, Yan Sun, Kun Qiu, Simon C. Fleming, Ian H. White,

Proc. of SPIE Vol. 6351, 63512I, (2006) · 0277-786X/06/$15 · doi: 10.1117/12.688319

Proc. of SPIE Vol. 6351 63512I-1

WDM1

LD2

LD1

WDM2

EDF

DCF

SMF

FBG2 FBG1

Output ISO OC

1480nm

1467nm

SLR

980nm LD

WDM

TLD

EDF FBG OutputOC

ISO

( a )

( b )

Fig.1. Schematic configurations of (a) the first and (b) the second proposed fibre lasers. LD: Laser Diode. TLD: Tunable Laser Diode. WDM: Wavelength Division Multiplexing. FBG: Fibre Bragg Grating. OC: Optical Circulator. ISO: Isolator. EDF: Erbium-Doped Fibre. SLR: Sagnac Loop Reflector. SMF:Single Mode Fibre. DCF: Dispersion Compensation Fibre.

2. EXPERIMENTAL STUDY AND ANALYSIS

2.1 Experimental setup

The schematic configurations of the proposed switchable dual-wavelength fibre lasers are shown in Fig. 1. For the first proposed fibre laser (shown in Fig. 1(a)), a 1480 nm laser diode (LD) with a maximum power of about 250 mW is used as the EDF pump LD1, and a 1467 nm LD with a maximum power of 350 mW (the corresponding drive current is 950 mA) is employed as the Raman pump LD2. The length of the SMF (single mode fibre) section (with an attenuation coefficient of 0.19 dB/km at 1550 nm) between the two FBGs is 10 km. Sections of DCF (dispersion compensation fibre), SMF and EDF are the gain fibres in the present structure. The DCF section (with a total dispersion of -986 ps/nm and a total loss of 5.7 dB) used here can compensate 60 km single mode fibre. The length, numerical aperture, cutoff wavelength and peak absorption (at 1531nm) of the EDF section are 4 m, 0.25, 950 nm and 19.2 dB/m, respectively. Two FBGs are used as wavelength selection components. The central wavelengths, the full-widths at half maximum (FWHM) and the reflectivities are (1545.13 nm, 0.093 nm, 50.0%) and (1559.98 nm, 0.098 nm, 51.3%) for FBG1 and FBG2, respectively (these parameters are provided by the manufacturer). An optical circulator and an isolator are used here to ensure that the lasing cavity is a counterclockwise ring. For the second proposed fibre laser (shown in Fig. 1(b)), a 980 nm LD with a maximum power of about 150 mW is used as an EDF pump, and a tunable (from 1500 nm to 1590 nm) laser diode (TLD) with a maximum power of about 10 mW is employed as the injection laser. The line structure cavity is formed by a wideband Sagnac loop reflector (a wideband 3dB 2x2 coupler with two ports connected with each other by a fibre) and an FBG whose central wavelength, FWHM and reflectivity are 1550.48 nm, 0.096 nm and 53.2%, respectively (these parameters are provided by the manufacturer). A section of EDF (the same one used in the first proposed fibre laser) is employed as the gain medium.


3m

Out

put

pow

er,

dE

:1ffIH1540 1545 1550 1555 1560 1565

Wavelength, nm

10

0

- -10

Out

put

pow

er,

dE

c,

c,

c,

0 0 I 0 c / 0

-50

-60 I I I

0 Ia 200 400 600 lb 800 1000

Drive current for LD2, mA

Fig. 2. Output spectra of the first proposed fibre laser when the drive current of LD2 is switched off (solid line) or

switched on to 750 mA (dashed line). The power of the EDF pump is fixed to 150mW.

2.2 Experimental results and analysis for the first fibre laser

In our experiment for the first proposed fibre laser, the power of EDF pump LD1 is fixed to 150 mW. When the Raman pump LD2 is switched off, the output spectra of the laser is shown by the solid line in Fig. 2, from which one sees that the structure is lasing only at 1λ with a peak power of -11.9dBm. However, when the drive current of LD2 is switched on to 750 mA (note that we read the current of the driver of the LD2 directly), the structure will be lasing only at 2λ with a peak power of 1.9 dBm (see the dashed line in Fig. 2). Note that the resolution and sensitivity of the optical spectrum analyzer (OSA) used in the all measurements in this paper are 0.06 nm and -65 dBm, respectively. Our experimental results have shown that the laser output is rather stable. Thus, the wavelength switching is achieved in this fibre laser by controlling the drive current of LD2 (i.e., the power of the Raman pump LD2).

Fig. 3. Output power at the two lasing wavelengths against the drive current for LD2.


900

800 L

< 700L1

3witc

hing

cur

rent

,

- (_

n 0)

C

) C

) C

) C

) C

) C

) C

) C

) I

I I

I

UUF-

100

0 I I I0 20 40 60 80 100 120 140 160 180

Pump power of LD1, mW

Figure 3 shows the detailed behavior of the dual-wavelength switching as the drive current for LD2 increases (the pump power of LD1 is fixed to 150 mW). The variation of output power at 1λ and 2λ are indicated by the lines connected with circles and triangles, respectively. Obviously, there are three regions where the laser is in different operational modes. In region A (with the drive current for LD2 below switching current aI = 70 mA), only 1λ is lasing and its output power remains around -11.9dBm. In region B (with the drive current for LD2 ranging from 70mA to 690mA), both 1λ and 2λ are lasing. In this region, as the drive current of LD2 increases, the output power at 2λ increases while the output power at 1λ decreases. In region C (with the drive current for LD2 above switching current bI = 690 mA), only 2λ is lasing and its output power remains around 1.9 dBm.

The operation principle of the present switchable dual-wavelength fibre laser can be explained as follows. When the Raman pump LD2 is off, only lasing at 1λ can be obtained since the cavity loss at 2λ is larger than that at 1λ (due to the loss of the 10 km SMF) and the EDF amplification will occur only at 1λ (due to the gain competition). When the power of LD2 increases gradually, the Raman gain mainly contributes to the amplification at 2λ since 2λ is much closer to the Raman gain peak wavelength than 1λ . Furthermore, when the LD2 pump power is large enough, the Raman gain in the SMF section can also increase the cavity gain at 2λ . Thus, the structure starts lasing at 2λ when the pump power of LD2 reaches a certain level. Consequently, the power of the lasing at 1λ decreases as the pump power of LD2 increases since the lasing at 2λ will reduce the population inversion in the EDF. When the pump power of LD2 increases further, lasing at 1λ disappears finally.

Fig. 4. Switching currents aI and bI for LD2 when the power of LD1 is set to different values.

We also study the wavelength switching when the pump power of LD1 is fixed to a different value. Experimental results show that the characteristics of the wavelength switching are similar (except the switching currents aI and bI ) when the pump power of LD1 is set to different values. Figure 4 shows the switching currents aI and bI when the pump power of LD1 is set from 20 mW to 160 mW with an increment of 20mW. The switching current aI decreases as the pump power of LD1 increases. This can be explained as follows. The cavity gain for 2λ to start lasing is a constant, which is the sum of EDF gain and Raman gain. Thus, the larger EDF pump is set, the smaller Raman gain is needed for 2λ to start lasing (corresponding to smaller switching current aI ). The variation of aI under different pump powers of LD1 is quite small,


0

Out

put p

ower

, dE

I.j

j::1545 1550 1555 1560

Wavelength, nm

while the variation of bI is quite large. As shown in Fig. 4, the switching current bI increases almost linearly with the pump power of LD1. This indicates that a larger drive current for LD2 is needed in order to let 2λ win the gain competition over 1λ when the pump power of LD1 increases.

Fig. 5. Output spectra of the second proposed fibre laser when the injection laser is switched off (solid line)

or switched on to -6 dBm (dashed line). The power of the EDF pump is fixed to 16.2 mW.

2.3 Experimental results and analysis for the second fibre laser

In our next experiment for the second proposed fibre laser, the power of the EDF pump LD is fixed to 16.2 mW. When the injection laser TLD is switched off, the output spectra of the laser is shown by the solid line in Fig. 5, from which one sees that the structure is lasing only at 1λ = 1550.48 nm (corresponding to the central wavelength of the FBG) with a peak power of -4.88 dBm. When the injection laser TLD (with tunable wavelength 2λ ; for this figure 2λ = 1555 nm) is switched to -6.0 dBm (or larger), the structure will be lasing only at 2λ = 1555 nm with a peak power of -3.16 dBm (see the dashed line in Fig. 5). Thus, the wavelength switching can be achieved in this fibre laser by controlling the power of the injection laser TLD.


0

-10

Out

put

pow

er,

d V

___0

-40

—o-oo

-50 I I I I I I-16 -14 -12 -10 -8 Ps-6 -4

Power of the injection laser, dBm

Sw

itchi

ng p

ower

,

(_n

--770

0.0 ___I I I

12 14 16 18 20 22 24 26 28 30Pump power, mW

Fig. 6. Output power at the two wavelengths as the power of the injection laser increases when the power of

the 980 nm pump LD is fixed to 16.2 mW.

Figure 6 shows the detailed behavior of the dual-wavelength switching as the power of the injection laser TLD increases (the power of 980 nm LD is fixed to 16.2 mW). The output power at 1λ and 2λ are indicated by the lines connected with circles and triangles, respectively. Obviously, when the injection laser is switched off, the proposed fibre laser is simply an EDF laser with an output wavelength of 1λ . However, when the injection laser is switched on and its output power increases, one sees there are two regions where the laser is in different operational modes. In region A (with the TLD power below the switching power sP = -6.6 dBm), the laser outputs at both 1λ and 2λ . In this region, as the TLD power increases, the output power at 2λ increases while the output power at 1λ decreases. In region B (with the TLD power larger than the switching power sP ), the lasing occurs only at 2λ and its output power remains around -3.16 dBm. Figure 7 shows a good linear relationship between the pump power and the switching power sP when the wavelength of the TLD is fixed to 1555 nm.

Fig. 7. Relationship between the switching power sP and the power of the 980 nm pump LD when the

wavelength of the TLD is fixed to 1555 nm.


9

E6ca \ 0

In I.-._lAI

witc

hing

pow

er, d

o o / 0/

0 /

7-6

_9F. I I I1530 1540 1550 1560 1570 1580 1590

Wavelength, nm

The operation principle of the present injection-switched EDF laser can be explained as follows. When the injection laser TLD is switched off, only lasing at 1λ can occur since the cavity of the line structure only exists at the Bragg wavelength

1λ of the FBG. When the TLD is switched on, the seeding light is injected into the cavity and is amplified (traveling with a round trip on the EDF). In this case, the seeding light will reduce the population inversion required for lasing at

1λ (the cavity gain at 1λ decreases). Consequently, the output power at 1λ decreases as the power of the TLD increases. When the power of the TLD is large enough and approaches sP , the emission stimulated by the seeding light at wavelength 2λ consumes most of the pump power (i.e., reduces the population inversion) and consequently the cavity gain becomes less than the cavity loss at 1λ (i.e., the lasing at 1λ becomes impossible). This result in a drastic drop of the power at 1λ and finally only the output at 2λ can be observed.

Fig. 8. Switching power as the wavelength of the TLD increases when the pump power is fixed at 16.2 mW

(line connected with triangles) or 19.1 mW (line connected with circles).

We go further to study the characteristics of the wavelength switching. Figure 8 shows the switching powers for different wavelengths of the TLD when the power of the 980 nm pump LD is fixed to 16.2 mW (line connected with squares) or 19.1 mW (line connected with circles). From this figure one sees that the wavelength switching can be achieved over a wide wavelength range (about 50 nm) for the TLD. The V-shape lines indicate that the wavelength switching is easier when the TLD wavelength is around 1560 nm and a larger switching power is needed when the TLD wavelength is far away from 1560 nm. We have measured the saturated gain spectrum of the (1 round trip) EDFA structure (just removing the FBG in Fig. 1(b)) when the pump power is 16.2mW, from which we find the gain peak is around 1560 nm, and this is the reason why the switching power is minimal around 1560 nm as shown in Fig. 8. Furthermore, from Fig. 8 one sees that the difference in the switching powers for the two lines is almost a constant (4.6dBm) for different TLD wavelengths, and this is due to the linear relationship between the switching power (in mW) and the power of the 980 nm pump LD. The transient switching response of the proposed laser is also studied when the power of the 980 nm pump LD is 16.2 mW and the TLD wavelength is fixed at 1548.6 nm (due to the availability of a filter in our lab for measuring the laser output at a specific wavelength). We employ two InGaAs Photo-detectors (PDA400, provided by Thorlabs, Ltd. UK) to measure the time delay between the moment when the injection laser is switched on and the moment when the lasing at

1λ disappears, and the results are shown in Fig. 9. From this figure one sees that the time delay (i.e., the switching time) is less than 50 microsecond, which indicates the proposed laser has the ability of fast wavelength-switching (much faster than the switching time reported in Ref. [14]). The switching time from the self-seeded wavelength to the injection


120

100

Vot

age,

mV

-

0)

co

T'T TTH-200 -150 -100 -50 0 50 100 150 200

Time, us

wavelength is expected to be about several round-trip time of the laser cavity. However, we have not reached the limit of the switching time due to the limitation of our measurement equipment.

Fig. 9. Transient switching response (measured with photo-detectors) of the proposed fibre laser when the power of the

980 nm pump LD is 16.2 mW and the TLD wavelength is fixed at 1548.6 nm.

3. CONCLUSION In summary, we have proposed two methods to achieve wavelength switching in fibre lasers and demonstrated two switchable dual-wavelength fibre lasers based on FBG feedback. For the first method, both an overlapping cavity and the hybrid gain medium are employed in the fibre laser. The lasing wavelength can be switched by controlling the gain at different wavelengths. As an example, in our first proposed fibre laser, an overlapping cavity has been designed combining both the EDF gain medium and the Raman gain medium. The wavelength switching is controlled by the power of the Raman pump and the operation principle has been explained. The characteristics of the wavelength switching under different EDF pump levels have been studied experimentally. Unlike the other switchable multi-wavelength fibre lasers, which are only based on EDF amplification, the present structure is simple (requiring no polarization controller or special fibre/device) and the separation of the dual wavelengths can be very large (since the wavelength of the Raman pump can be arbitrary, which is impossible for EDF amplification). The output power and wavelength of the present switchable dual-wavelength fibre laser are repeatable and uniquely determined by the Raman pump power (without any hysteresis characteristic, which exists in some other switchable multi-wavelength fibre lasers reported in e.g. [1, 2]). To extend this method, it is expected that switchable dual-wavelength fibre laser can also been achieved by employing other hybrid gain (i.e., by using EDFA and SOA). For the second method, an injection technique is employed. As an example, an injection-switchable Erbium-doped fibre laser of the line structure has been proposed. The line structure cavity is formed by a fibre Sagnac loop reflector and an FBG. The dual-wavelength switching is achieved by controlling the power of the injection laser and the operational principle has been explained in detail. The characteristics of the wavelength switching for different levels of the EDF pump laser and different wavelengths of the tunable injection laser have been studied experimentally. The switching time from one wavelength to the other is less than 50 microsecond. Furthermore, we can also try to achieve wavelength switching in a fibre ring laser by employing an injection technique.


ACKNOWLEDGEMENT

This work was partially supported by Zhejiang provincial government (under grant Nos. 20044131095 and R104154) and Hangzhou municipal government (under grant No. 20051321B14).

REFERENCES

1. Q. H. Mao and J. W. Y. Lit, "Switchable multiwavelength erbium-doped fiber laser with cascaded fiber grating cavities," IEEE Photonics Technol. Lett., 14(5), 612–614 (2002). 2. Q. H. Mao and J. W. Y. Lit, "Optical bistability in an L-band dual-wavelength erbium-doped fiber laser With Overlapping Cavities," IEEE Photonics Technol. Lett., 14(9), 1252–1254 (2002). 3. X. H. Feng, Y. G. Liu, S. G. Fu, S. Z. Yuan, and X. Y. Dong, "Switchable Dual-Wavelength Ytterbium-Doped Fiber Laser Based on a Few-Mode Fiber Grating," IEEE Photonics Technol. Lett., 16(3), 762–764 (2004). 4. L. Su and C. Lu, "Wavelength-switching fibre laser based on multimode fibre Bragg gratings," IEE Electron. Lett., 41(1), 11–13 (2005). 5. X. H. Feng, H. Y. Tam and P. K. A. Wai, "Switchable multiwavelength erbium-doped fiber laser with a multimode fiber Bragg grating and photonic crystal fiber," IEEE Photonics Technol. Lett., 18(9), 1088–1090 (2006). 6. Y. W. Lee and B. Lee, "Wavelength-switchable erbium-doped fiber ring laser using spectral polarization-dependent loss element," IEEE Photonics Technol. Lett., 15(6), 795–797 (2003). 7. G. Das and J. W. Y. Lit, "Wavelength switching of a fiber laser with a Sagnac loop reflector," IEEE Photonics Technol. Lett., 16(1), 60–62(2004). 8. M. Delgodo-Pinar, A. Ciez, J. L. Gruz and M. V. Andres, "Wavelength-switchable fiber laser using acoustic waves," IEEE Photonics Technol. Lett., 17(3), 552–554 (2005). 9. Y. J. Kim and D. Y. Kim, "Electrically tunable dual-wavelength switching in a mutually injection-locked erbium-doped fiber ring laser and distributed-feedback laser diode," IEEE Photonics Technol. Lett., 17(4), 762–764 (2005). 10. S. Hu, et al., "Switchable Multiwavelength Erbium-Doped Fiber Ring Laser With a Multisection High-Birefringence Fiber Loop Mirror," IEEE Photonics Technol. Lett., 17(7), 1387–1389 (2005). 11. H. L. Liu, H. Y. Tam, W. H. Chung, P. K. A. Wai and N. Sugimoto, "Wavelength-Switchable La-Codoped Bismuth-Based Erbium-Doped Fiber Ring Laser," IEEE Photonics Technol. Lett., 17(5), 986–988 (2005). 12. Y. G. Han, G. Kim, J. H. Lee, S. H. Kim, and S. B. Lee, "Lasing wavelength and spacing switchable multiwavelength fiber laser from 1510 to 1620 nm," IEEE Photonics Technol. Lett., 17(5), 989–991 (2005). 13. D. N. Wang, F.W. Tong, X. Fang, W. Jin, P. K. A.Wai, and J. M. Gong, "Multiwavelength erbium-doped fiber ring laser source with a hybrid gain medium," Opt. Commun. 228, 295-301 (2003). 14. Q. H. Mao and W. Y. L. John, "L-band fiber laser with wide tuning range based on dual-wavelength optical bistability in linear over lapping grating cavities," J. Lightwave Technol., 39(10), 1252-1259 (2003).


Paper Ⅸ

this condition, the excited surface currents in the system groundplane is smaller for the proposed antenna operated in the self-balanced resonant mode.

REFERENCES

1. K.L. Wong, Planar antennas for wireless communications, Wiley, NewYork, 2003, Chapter 2.

2. H. Morishita, Y. Kim, and K. Fujimoto, Design concept of antennas forsmall mobile terminals and the future perspective, IEEE AntennasPropag Magn 44 (2002), 30–43.

3. H. Morishita, H. Furuuchi, and K. Fujimoto, Performance of balanced-Fed antenna system for handsets in the vicinity of a human head orhand, IEE Proc Microwave Antennas Propag 149 (2002), 85–91.

4. S. Hayashida, T. Tanaka, H. Morishita, Y. Koyanagi, and K. Fujimoto,Built-in folded monopole antenna for handsets, Electron Lett 39 (2004),1514–1516.

5. S.L. Chien, F.R. Hsiao, Y.C. Lin, and K.L. Wong, Planar inverted-Fantenna with a hollow shorting cylinder for mobile phone with anembedded camera, Microwave Opt Technol Lett 41 (2004), 418–419.

6. K.L. Wong, S.L. Chien, C.M. Su, and F.S. Chang, An internal planarmobile phone antenna with a vertical ground plane, Microwave OptTechnol Lett 47 (2005), 597–599.

7. S. Gabriel, R.W. Lau, and C. Gabriel, The dielectric properties ofbiological tissues. III. Parametric models for the dielectric spectrum oftissues, Phys Med Biol 41 (1996), 2271–2293.

8. C.M. Su, C.H. Wu, K.L. Wong, S.H. Yeh, and C.L. Tang, User’s handeffects on EMC internal GSM/DCS dual-band mobile phone antenna,Microwave Opt Technol Lett 48 (2006),1563–1569.

© 2007 Wiley Periodicals, Inc.

TUNABLE AND INJECTION-SWITCHABLE ERBIUM-DOPED FIBERLASER OF LINE STRUCTURE

Daru Chen,1,2 Shan Qin,1,2 Zhangwei Yu,1,2 and Sailing He1,2,3

1 Centre for Optical and Electromagnetic Research, ZhejiangUniversity, Hangzhou 310058, China2 Joint Research Center of Photonics of the Royal Institute ofTechnology (Sweden) and Zhejiang University, East Building No. 5,Zhejiang University, Hangzhou 310058, China3 School of Electrical Engineering, Royal Institute of Technology

Received 28 August 2006

ABSTRACT: A tunable and injection-switchable erbium-doped fiber(EDF) laser is proposed based on a line structure formed by a fiberSagnac loop reflector and an fiber Bragg grating (FBG). Wavelengthswitching is achieved by controlling the power of the tunable injectionlaser. The self-seeded wavelength corresponding to the Bragg wave-length of the FBG can be tuned by, for example, heating the FBG, andthe injection wavelength can be tuned over a wide range of about 50nm. The characteristics of the wavelength switching for different levelsof the EDF pump power and different wavelengths of the injection laserare studied experimentally. The present fiber laser has the advantages oftunability, stability, low amplified spontaneous emission noise, and highinjection efficiency when compared with a fiber ring laser. Rapid wave-length switching is expected and the transient switching response of thelaser is also studied. © 2007 Wiley Periodicals, Inc. Microwave OptTechnol Lett 49: 765–768, 2007; Published online in Wiley InterScience(www.interscience.wiley.com). DOI 10.1002/mop.22262

Key words: tunable; injection; switchable; line structure; Erbium-dopedfiber laser

1. INTRODUCTION

Wavelength-switchable and tunable fiber lasers have importantapplications in WDM fiber communication systems, fiber sen-

sors, and so forth. Several techniques have been reported toachieve wavelength switching in fiber lasers, such as the use ofsampled fiber Bragg gratings (FBG) [1], cascaded FBGs [2, 3],a Sagnac loop reflector [4], a few-mode fiber grating [5], aslanted multimode FBG [6], a Bragg grating-based acoustoopticsuperlattice modulator [7], a hybrid gain medium [8], a multi-section high-birefringence fiber loop mirror [9], a programma-ble electric-actuated polarization controller [10], and hybridpumps [11]. Overlapping cavities can be used for multi-wave-length lasing and wavelength switching can be achieved bycontrolling the loss or the gain in different cavities [2, 3, 11].Recently, Dragic and coworkers have reported an injection-seeded erbium-doped fiber (EDF) ring laser [12–14], in whichan injection technique was used to achieve a wavelength-switchable fiber laser. A tunable laser (outside the cavity of thefiber laser) is used as seeding light. The shared cavity is forself-seeded lasing or injection-seeded lasing. Switching be-tween the self-seeded wavelength lasing and injection-seededwavelength lasing can be achieved by controlling the power ofthe injection laser. By using this technique, one of the switchingwavelengths can be tuned among a large range since a tunableinjection laser is employed.

In this article we propose a novel tunable and switchable EDFlaser of line structure [see Fig. 1(a)]. Here a fiber Sagnac loopreflector and an FBG are employed to form a line cavity (like aFabry-Perot cavity). The seeding light is injected into the cavitythrough an optical circulator. Dual-wavelength switching isachieved by controlling the power of the injection laser. Onewavelength corresponds to the Bragg wavelength of the FBG (canbe tuned by changing the temperature or stress) and the otherwavelength can be tuned from 1535 to 1590 nm (determined by thewavelength of the injection laser).

Figure 1 The schematic configuration of (a) the proposed tunable andinjection-switchable fiber laser of line structure and (b) a fiber ring lasersimilar to the one proposed in Ref. 13. LD, laser diode; TLD, tunable laserdiode; WDM, wavelength division multiplexing; FBG, fiber Bragg grating;OC, optical circulator; ISO, isolator; EDF, erbium-doped fiber; SLR,Sagnac loop reflector

DOI 10.1002/mop MICROWAVE AND OPTICAL TECHNOLOGY LETTERS / Vol. 49, No. 4, April 2007 765

2. EXPERIMENTAL SETUP AND RESULTS

In our experiment, a 980 nm laser diode (LD) with a maximumpower of about 150 mW is used as an EDF pump, and a tunable(from 1500 to 1590 nm) LD with a maximum power of about 10mW is employed as the injection laser. The line cavity is formedby a wideband Sagnac loop reflector (a wideband 3 dB 2 2coupler with two ports connected with each other with a fiber)and an FBG whose central wavelength, full-width at half max-imum, and reflectivity are 1550.48 nm, 0.096 nm, and 53.2%,respectively (these parameters are provided by the manufac-turer). A section of EDF is employed as the gain medium, andits length, numerical aperture, cutoff wavelength, and peakabsorption (at 1531 nm) are 4 m, 0.25, 950 nm, and 19.2 dB/m,respectively.

In our first experiment, the power of the EDF pump LD isfixed to 16.2 mW. When the injection laser (TLD) is switchedoff, the output spectrum of the laser is shown by the solid linein Figure 2, from which one sees that the structure is lasing onlyat 1 1550.48 nm (corresponding to the central wavelength ofthe FBG) with a peak power of 4.88 dBm. When the injectionlaser (TLD) (with tunable wavelength 2; for this figure 2 1555 nm) is switched to 6.0 dBm (or larger), the structureoutputs only at 2 1555 nm with a peak power of 3.16 dBm(see the dashed line in Fig. 2). The inset of Figure 2 shows thespectrum of the injection laser (TLD), which is similar to theoutput spectrum of the proposed fiber laser shown by the dashedline in Figure 2. In this case, the injection light is amplified bya two-pass traveling-wave amplifier. Note that the resolutionand sensitivity of the optical spectrum analyzer (OSA) used inthe measurement are 0.06 nm and 65 dBm, respectively.Thus, the wavelength switching can be achieved in this fiberlaser by controlling the power of the injection laser (TLD).

3. COMPARISON WITH A FIBER RING LASERS

For comparison, we have also made a fiber ring laser [see Fig.1(b)] for the schematic configuration (similar to the one pro-posed recently in Ref. 13). We use a 980 nm pump LD and a

section of EDF (same as the one used for Fig. 2) to form anEDFA. A 3-dB coupler and an isolator are employed in the ringcavity. The injection efficiency is rather low for the fiber ringlaser shown in Figure 1(b) (only half of the TLD power can beinjected into the ring cavity), while our proposed fiber laser hasa high injection efficiency (up to 100% if an ideal opticalcirculator is used). The measured output spectra of this fiberring laser are shown in Figure 3. One can see that the amplifiedspontaneous emission (ASE) noise of this fiber ring laser ismuch larger than that of the proposed fiber laser of line struc-ture. The ASE noise is larger for the fiber ring laser reported inRef. 13 since a 90/10 coupler is used. One may suggest toreplace with a coupler of low power ratio (e.g., a 1/99 coupler)to reduce the ASE noise. However, it will drastically reduce theinjection efficiency (only 1% for a 1/99 coupler). Unlike thepresent fiber laser of line structure whose lasing wavelength canbe tuned by, for example, heating the FBG (shown in Fig. 4),the lasing wavelength of the fiber ring laser cannot be tuned

Figure 2 The output spectra of the proposed fiber laser. The power of theEDF pump LD is fixed to 16.2 mW. The solid line is the output spectrumof the fiber laser when the injection laser is switched off, and the dashedline is the output spectrum of the fiber laser when the power of the injectionlaser (at a wavelength of 1555 nm) is switched on and fixed to 6 dBm.The inset shows the spectrum of the injection laser (TLD)

Figure 3 The output spectrum of the fiber ring laser shown in Figure1(b). The inset shows the spectra measured 6 times with an interval of 2min

Figure 4 The output spectra of the fiber ring laser when the temperatureof the FBG is different (when the injection laser TLD is switched off). Theinset shows the spectra measured 6 times with an interval of 2 min

766 MICROWAVE AND OPTICAL TECHNOLOGY LETTERS / Vol. 49, No. 4, April 2007 DOI 10.1002/mop

since there is no tunable wavelength-selection component in thering cavity. Furthermore, the wavelength of the fiber ring laseris unstable (see the inset of Fig. 3 for the spectra measured 6times with an interval of 2 min) since the output wavelength isdetermined by the length of the ring cavity, which is sensitiveto an environmental (e.g. temperature and stress) change. How-ever, the lasing wavelength of the proposed fiber laser of linestructure is very stable (since it is determined by the Braggwavelength of the FBG) and we cannot see any spectrum shiftin our OSA with the resolution of 0.06 nm (see the inset of Fig.4 for the spectra measured 6 times with an interval of 2 min).

4. CHARACTERISTICS OF WAVELENGTH SWITCHING

Figure 5 shows the detailed behavior of the dual-wavelengthswitching as the power of the injection laser TLD increases (thepower of 980 nm LD is fixed to 16.2 mW) in the proposedinjection-switchable fiber laser. The output powers at 1 and 2

are indicated by the lines connected with squares and circles,respectively. Obviously, when the injection laser is switchedoff, the proposed fiber laser is simply an EDF laser with anoutput wavelength of 1. However, when the injection laser isswitched on and its output power increases, one sees there aretwo regions where the laser is in different operational modes. InRegion A (with the TLD power below the switching power Ps

6.6 dBm), the laser outputs at both 1 and 2. In thisregion, as the TLD power increases, the output power at 2

increases, while the output power at 1 decreases. In Region B(with the TLD power larger than the switching power Ps), thestructure has the output only at 2 and its output power remainsaround 3.16 dBm. The inset of Figure 5 shows a good linearrelationship between the pump power and the switching powerPs when the wavelength of the TLD is fixed to 1555 nm. Theoperation principle of the present tunable and injection-switch-able EDF laser can be explained as follows. When the injectionlaser (TLD) is switched off, only lasing at 1 can occur sincethe cavity of the line structure only exists at the Bragg wave-length 1 of the FBG. When the TLD is switched on, theseeding light is injected into the line cavity and is amplified

(traveling with a round trip on the EDF). In this case, theseeding light will reduce the population inversion required forlasing at 1 (the cavity gain at 1 decreases). Consequently, theoutput power at 1 decreases as the power of the TLD increases.When the power of the TLD is large enough and approaches Ps,the emission stimulated by the seeding light at wavelength 2

consumes most of the pump power (i.e., reduces the populationinversion) and consequently the cavity gain becomes less thanthe cavity loss at 1 (i.e., the lasing at 1 becomes impossible).It results in a drastic drop of the power at 1 and finally only theoutput at 2 can be observed.

Figure 6 shows the switching powers for different wavelengthsof the TLD when the power of the 980 nm pump LD is fixed to16.2 mW (lines connected with squares) or 19.1 mW (lines con-nected with circles). From this figure one sees that the wavelengthswitching can be achieved over a wide wavelength range (about 50nm) for the TLD. The V-shape curves indicate that the wavelengthswitching is easier when the TLD wavelength is around 1560 nmand a larger switching power is needed when the TLD wavelengthis far away from 1560 nm. The inset of Figure 6 shows thesaturated gain spectrum of the (one round trip) EDFA structure(just removing the FBG in Fig. 1) when the pump power is 16.2mW. From this inset one sees the gain peak is around 1560 nm,and this is the reason why the switching power is minimal around1560 nm as shown in Figure 6. Furthermore, from Figure 6 onesees that the difference in the switching powers for the two curvesis almost a constant (4.6 dBm) for different TLD wavelengths, andthis is due to the linear relationship between the switching power(in mW) and the power of the pump 980 nm LD (see the inset ofFig. 5).

5. MEASUREMENT OF SWITCHING TIME

The transient switching response of the proposed laser is alsostudied when the power of the 980 nm pump LD is 16.2 mW andthe TLD wavelength is fixed at 1548.6 nm (due to the availabilityof a filter in our lab for measuring the laser output at a specificwavelength). We employ two InGaAs Photo-detectors (PDA400,provided by Thorlabs, UK) to measure the time delay between themoment when the injection laser is switched on and the momentwhen the lasing at 1 disappears, and the results are shown in

Figure 5 The output powers at the two wavelengths as the power of theTLD increases. The power of the 980 nm pump LD is fixed to 16.2 mW.The inset shows the relation between the switching power Ps and the powerof the 980 nm pump LD when the wavelength of the TLD is fixed to 1555nm

Figure 6 The switching powers as the wavelength of the TLD increaseswhen the pump power is fixed at 16.2 mW (lines connected with squares)or 19.1 mW (lines connected with circles)

DOI 10.1002/mop MICROWAVE AND OPTICAL TECHNOLOGY LETTERS / Vol. 49, No. 4, April 2007 767

Figure 7. From this figure one sees that the time delay (i.e., theswitching time) is less than 50 s, which indicates the proposedlaser has the ability of fast wavelength-switching (much faster thanthe switching time reported in Ref. 15). The switching time fromthe self-seeded wavelength to the injection wavelength is expectedto be about several round-trip time of the laser cavity. However,we have not reached the limit of the switching time because of thelimitation of our measurement equipment.

6. CONCLUSIONS

In conclusion, we have proposed and demonstrated a tunable andinjection-switchable EDF laser of line structure. The line cavity isformed by a fiber Sagnac loop reflector and an FBG. The dual-wavelength switching is achieved by controlling the power of theinjection laser and the operational principle has been explained indetail. Both wavelengths can be tuned by adjusting the injectionwavelength of the tunable laser and the Bragg wavelength of theFBG. The characteristics of the wavelength switching for differentlevels of the EDF pump laser and different wavelengths of thetunable injection laser have been studied experimentally. Ourexperimental results have shown that the present fiber laser is morestable and has a much lower ASE noise and a higher injectionefficiency when compared with a fiber ring laser. The switchingtime from one wavelength to the other wavelength is less than50 s.

ACKNOWLEDGMENT

This work was partially supported by a research grant (No.R104154) of Natural Science Foundation of Zhejiang Province(China).

REFERENCES

1. Y. Wang, S.C. Tjin, J. Yao, J.P. Yao, L. He, and K.A. Ngoi, Wave-length-switching fiber laser for optically controlled phased-array an-tenna, Opt Commun 211 (2002), 147–151.

2. Q.H. Mao and J.W.Y. Lit, Switchable multiwavelength erbium-dopedfiber laser with cascaded fiber grating cavities, IEEE Photon TechnolLett 14 (2002), 612–614.

3. C.L. Zhao, X.F. Yang, C. Lu, J.H. Ng, X. Guo, R.C. Partha, and X.Y.Dong, Switchable multi-wavelength erbium-doped fiber lasers by us-

ing cascaded fiber Bragg gratings written in igh birefringence fiber,Communication 230 (2004), 313–317.

4. G. Das and J.W.Y. Lit, Wavelength switching of a fiber laser with aSagnac loop reflector, IEEE Photon Technol Lett 16 (2004), 60–62.

5. X.H. Feng, Y. Liu, S.G. Fu, S.Z. Yuan, and X.Y. Dong, Switchabledual-wavelength ytterbium-doped fiber laser based on a few-modefiber grating, IEEE Photon Technol Lett 16 (2004), 762–764.

6. C.L. Zhao, Z.H. Li, M.S. Demokan, X.F. Yang, and W. Jin, Switchablemultiwavelength SOA-fiber ring laser based on a slanted multimodefiber Bragg grating, Opt Commun 252 (2005), 52–57.

7. M. Delgado-Pinar, J. Mora, A. Diez, J.L. Cruz, and M. V. Andres,Wavelength-switchable fiber laser using acoustic waves, IEEE PhotonTechnol Lett 17 (2005), 552–554.

8. Y.G. Han, G. Kim, J.H. Lee, and S.B. Lee, Wavelength spacingtunable multiwavelength fiber laser with lasing wavelength selectivity,Opt Commun 256 (2005), 98–102.

9. S. Hu, L. Zhan, Y.J. Song, W. Li, S.Y. Luo, and Y.X. Xia, Switchablemultiwavelength erbium-doped fiber ring laser with a multisectionhigh-birefringence fiber loop mirror, IEEE Photon Technol Lett 17(2005), 1387–1389.

10. H.L. Liu, H.Y. Tam, W. H. Chung, P.K.A. Wai, and N. Sugimoto,Wavelength-switchable La-codoped bismuth-based erbium-doped fi-ber ring laser, IEEE Photon Technol Lett 17 (2005), 986–988.

11. D.R. Chen, Z.W. Yu, S. Qin, and S.L. He, Switchable dual-wavelengthRaman erbium-doped fibre laser, IEE Electron Lett 42 (2006), 202–203.

12. Mukhopadhyay and P.D. Dragic, Broad-band injection seeding of anerbium-doped fiber ring laser, IEEE Photon Technol Lett 17 (2005),1166–1168.

13. P.D. Dragic, Injection-seeded Q-switched fiber ring laser, IEEE Pho-ton Technol Lett 16 (2004), 1822–1824.

14. P.D. Dragic, Analytical model for injection-seeded erbium-doped fiberring lasers, IEEE Photon Technol Lett 17 (2005), 1629–1631.

15. Q.H. Mao and J.W.Y. Lit, L-band fiber laser with wide tuning rangebased on dual-wavelength optical bistability in linear overlappinggrating cavities, J Lightwave Technol 39 (2003), 1252–1259.

© 2007 Wiley Periodicals, Inc.

FIBER OPTIC DISPLACEMENT SENSINGMONITORED BY AN OTDR ANDREFERENCED BY FRESNELREFLECTION AND BY FIBER BRAGGGRATINGS

S. Mendonca,1 O. Frazao,2 J. M. Baptista,2,3 andJ. L. Santos1,2

1 Departamento de Fısica, Faculdade de Ciencias da Universidade doPorto, Rua do Campo Alegre, 687, 4169–007 Porto, Portugal2 INESC Porto, Rua do Campo Alegre, 687, 4169–007 Porto,Portugal3 Departamento de Matematica e Engenharias, Universidade daMadeira, Caminho da Penteada, 9000–390 Funchal, Portugal

Received 28 August 2006

ABSTRACT: This work presents a study of displacement sensing basedon a fiber eight geometry shape monitored by an optical time domainreflectometer. The displacement sensor is referenced by Fresnel reflec-tion or by a fiber Bragg grating structure located after the sensor. Oneof the advantages of the fiber Bragg grating as referencing device is itscapability to be multiplexed by a number of displacement sensors inseries along the optical fiber. This concept is also demonstrated usingtwo Bragg gratings. © 2007 Wiley Periodicals, Inc. Microwave OptTechnol Lett 49: 768–770, 2007; Published online in Wiley InterScience(www.interscience.wiley.com). DOI 10.1002/mop.22261

Figure 7 The transient switching response (measured with photo-detec-tors) of the proposed fiber laser when the power of the 980 nm pump LDis 16.2 mW and the TLD wavelength is fixed at 1548.6 nm

768 MICROWAVE AND OPTICAL TECHNOLOGY LETTERS / Vol. 49, No. 4, April 2007 DOI 10.1002/mop

fibre bragg grating components for filtering, switching ...1805/fulltext01.pdf · fibre bragg...

Documents