Long-period gratings in chalcogenide fibers

LUCAS MAIA BEZERRA DE ARAUJO URTIGA

LONG-PERIOD GRATINGS IN CHALCOGENIDE

FIBERS

Mémoire présenté

à la Faculté des études supérieures de l'Université Laval dans le cadre du programme de maîtrise en physique pour l'obtention du grade de maître es sciences (M.Sc.)

DEPARTEMENT DE PHYSIQUE, DE GENIE PHYSIQUE ET D'OPTIQUE FACULTÉ DES SCIENCES ET DE GÉNIE

UNIVERSITÉ LAVAL QUÉBEC

2010

Chalcogenide glasses have been studied for their optical properties for over 50 years. They exhibit high nonlinearity, transparency to infrared light and photosensitivity to visible light. Recent advances on their manufacturing processes made possible the development of several applications including bio-chemical sensors, infrared laser-power delivery, mid-infrared sources and all-optical switching. This dissertation describes the research conducted over the AS1S3 chalcogenide glass Fibre, particularly on the fabrication of photo-induced and mechanically induced long-period gratings (LPGs).

First, a review of the previous works on the subject is made as the properties and applications of the chalcogenide glass fibers are introduced. Then, a brief introduction to fiber optics and the theory of LPGs are presented. Finally, the experimental results obtained in an AS2S3 chalcogenide glass fiber are shown, including the observation of photo-induced index changes, and the fabrication and characterization of mechanically induced LPGs, Mach-Zehnder interferometers based in two in-series LPGs, and a photo-induced LPG.

II

Résumé

Les verres chalcogénures sont des matériaux optiques connus depuis 50 ans. Ils sont fortement non-linéaires, transparents dans l'infrarouge et photosensibles à la lumière visible. Grâce à des percées récentes du côté des procédés de fabrication, leurs applications dans divers domaines sont devenus possibles, notamment dans les capteurs biochimiques, le guidage de sources infrarouges de haute puissance, la fabrication de sources infrarouges et dans les bascules optiques. Ce mémoire présente les résultats de travaux de recherche réalisés sur les fibres chalcogénures de AS2S3, notamment dans la conception de réseaux à longs pas (LPG), photo-induits et mécaniquement induits.

Premièrement, une revue bibliographique présentant les propriétés et les applications des fibres en verre chalcogénures est fournie. Ensuite, une brève introduction sur la théorie des fibres optiques et des réseaux LPG est développée, pour finalement présenter les résultats expérimentaux. Ces derniers portent sur les observations des effets photo-induits, et sur la fabrication et la caractérisation, à la fois, des réseaux LPG mécaniquement induits, des interféromètres Mach-Zehnder utilisant deux réseaux LPG en série, et d'un réseau LPG photo-induit.

I would like to acknowledge the technical and financial support provided by my advisors Prof. Galstian and Prof. Vallée, and to thank them for giving me the opportunity to work at COPL facilities. I would like to show my gratitude to all of the COPL personnel, particularly to M. Stéphane Gagnon, M. Marc d'Auteuil, M. Philippe Chrétien and M. Souleymane T. Bah for their patience and technical support.

To my wife Ruth, with whom Tve shared all the joy, excitement and sorrow involved in this work, and for being always there for me when I needed; To my parents, brother and sister Severino, Mabel, Breno and Marcella for making me feel always close to home despite ofthe distance.

Abstract i Résumé ii Acknowledgments iii

Table of contents v List of tables vii List of figures viii

Chapter 1: 1 Introduction to chalcogenide glass fibers 1

1.1 Fabrication 2 1.1.1 Glass fabrication 2

1.1.2 Fiber Fabrication 2 1.2 Optical transparency and refractive index 4

1.3 Refractive index of chalcogenide glasses 7

1.3 Photosensitivity 10 1.3.1 Photodarkening 11 1.3.2 Reversible modifications 11

1.3.3 Irreversible modifications 12 1.4 Applications of chalcogenide-glass fibers 13

1.4.1 Infrared power delivery 13 1.4.2 Chemical sensing 13 1.4.3 Photo-induced devices 14 1.5 Non-linear properties and applications 14

1.6 Summary 15 Chapter 2: 16 Long-period fiber gratings 16

2.1 Theory of long-period gratings 17 2.2 Numerical results on the calculation ofthe cladding effective index 19

2.4 Applications of long-period gratings 21

2.3 Summary 23 Chapter 3: 24 Observation of photo-induced changes 24

3.1 Introduction 24 3.2 Photo-induced melting and ablation in chalcogenide fibers 25

3.1.2 Photo-induced index changes in AS2S3 fibers 28

3.4 Summary 31 Chapter 4: 32 Mechanically-induced Long-period Gratings 32

4.1 Description 32 4.2 Experimental setup: 35

4.3 Experimental results 36 4.3.1 Birefringence ofthe mechanically induced LPG. 36

VI

4.4 Summary 44 Chapter 5: 45 Two-LPG Mach-Zehnder interferometer 45

5.1 Experimental results 46

5.2 Summary 51 Chapter 6: 53 Photo-induced Long-period gratings in AS2S3 fibers 53

6.1 Experimental Setup 53 6.2 Experimental Results 55 6.3 Summary 59 Conclusions: 60 Bibliography 62 Annexes 66 Annex A: Introduction to Fiber Optics 66

A.l Electromagnetic Analysis of Optical Fibers 68 A.2 First model: Two-layered step-indexed fiber 76

A.2.1 Solution in the core ofthe fiber: higher refractive index 78 A.2.2 Solution in the cladding of the fiber: lower refractive index region 79

A.2.3 Boundary conditions 83 A.2.4 Weakly-guiding fibers: LP modes 86

A.3 Second model: Three-layered step-indexed fiber 87

A.4 Numerical calculation 91

A.4.1 Core mode LP0i effective refraction index 91

2.4.2 Calculation of effective refraction indexes of HE|,M cladding modes using the

two-layered fiber model 93 2.4.3 Calculation of effective refraction indexes of HE]H cladding modes using the

three-layered fiber model 95

A.4 Summary 96 Annex B: Injection setup and fast effective injection procedure 96

Table 1.1: Basic performances of the best fibers made of high-purity arsenic chalcogenide

glasses 4 Table 1.2: The center peak of the absorption bands for the main impurities found in As-S

and As-Se Glasses, extracted from [2] 7 Table 1.3: The Cauchy coefficients found for some chalcogenide glasses 8

Table 2.1: Error introduced to the calculation ofthe effective refractive index of cladding

modes associated with the two-layered fiber model approximation 20 Table 2.2: Error introduced to the calculation of the pitch of a long-period grating generated

by the two-layered step-indexed fiber model simplification 21 Table 4.1: Optical and physical parameters ofthe AS2S3 fiber used in the experiments 33

List of figures

Fig. 1.1: Simplified schematics of the double crucible method for optical fiber fabrication. 3 Fig. 1.2: Absorption coefficient spectra for the AS2S3 extracted from [6]. Two peaks in the

transparency window associated to resonances of the impurities OH" (at 2.92fim) and

S-H (at 4.03^im) are shown 6 Fig. 1.3: Urbach tail region of the absorption spectra of both AS2S3 ( 1 ) and As2Se3 (2)

glasses extracted from [4] 7 Fig. 1.4: The refractive indexes of chalcogenide glasses are much higher than that of silica

glasses 9 Fig. 1.5: The material dispersion of AS2S3 and As2Se3 chalcogenide glasses, and of Si02

silica glass 10 Fig. 1.5: Scanning electron microscope image ofthe surface of a 50|im-thick As2S3 film

after 1000s-exposure to a lOmW focused He-Ne beam. Extracted from [11] 12 Fig. 2.1 : Transmittance of an uniform LPG written in a chalcogenide glass fiber simulated

in MathCAD. The parameters used in the simulation were: a, = 2pm ; n, = 2.4175;

a2=60.5pm; n2 = 2.4095; n3 =1 ; <r = 10"5. A = 437 fim 19 Fig. 2.1 : Schematics of the two-layered step-indexed fiber model (a) and the three-layered

step- indexed fiber model (b) 20 Fig. 2.2: Two in-series LPG acting as a Mach-Zehnder interferometer 22

Fig. 2.3: Fiber Michelson interferometer using LPG extracted from [36] 22 Fig. 2.4: Two in-series LPG acting as a Mach-Zehnder interferometer, used as

surface-plasmon-resonance sensor, extracted from [39] 23 Fig. 3.1: Urbach tail region ofthe absorption spectra of both AS2S3 (1) and As2Se3 (2)

glasses from [40] 25 Fig. 3.2: The As2Se3 fiber melted after the exposition to a focused 50mW He-Ne laser (at

633nm) 26 Fig. 3.3: Periodic laser ablation produced in a chalcogenide glass fiber 27

Fig. 3.4: The As2S3 fiber melted after the exposition to a focused 50mW He-Ne laser (at

633nm) 27 Fig. 3.5: The laser beam focus spot creates an index change in the fiber 28

Fig. 3.6: Picture of the experimental setup used to measure the refractive index changes in

the fiber 29 Fig. 3.7: Schematics of the experimental setup used to measure the refractive index changes

in the fiber 29 Fig. 3.8: Transmitted light intensity dynamics for a beam focused in the core of the fiber. 30

Fig.3.9: Transmitted light intensity dynamics for a beam focused in different portions ofthe

fiber 31 Fig. 4.1 : The clamping device used to create the long-period gratings 34

Fig. 4.2: Profile of mechanical grating 34 Fig. 4.3: Experimental setup schematics used to study the mechanically induced

long-period grating 35 Fig. 4.4: Transmittance of LPG for different clamping pressures 37

Fig. 4.5: Transmittance of LPG numerically simulated using the software IFOjGratings. 38

Fig. 4.6: Transmittance for two perpendicular polarizations 38 Fig. 4.7: Temperature variation of a single LPG for 90° polarization 40

Fig. 4.9: Thermal dependence ofthe primary peak depth for 90° polarization 40 Fig. 4.8: Extracted from [31]. Plots of the local intensity / (r) as a function of radius for the

four lowest-order / = 1 cladding modes in a typical fiber. All modes are circularly

symmetric and have been normalized to carry a power of 1 W 41 Fig. 4.10: Thermal dependence of the primary peak central wavelength for the 90°

polarization 42 Fig. 4.11 : Temperature variations of a single LPG for 0° polarization 42

Fig. 4.12: Thermal dependence of the primary peak central wavelength for the 0°

polarization 43 Fig. 4.13: Thermal dependence ofthe secondary peak central wavelength for the 0°

polarization 43 Fig. 5.1: Two in-series long period gratings acting as a Mach-Zehnder interferometer 45

Fig. 5.2: Layout of the experimental setup 46 Fig. 5.3: Transmittance ofthe week long-period grating LPG A 47

Fig. 5.4: Transmittance ofthe all-fiber, interferometer. Response of LPG A and LPG B

separated by 2cm 47 Fig. 5.5: The simulated transmittance obtained using the software IFO Gratings 48

Fig. 5.6: Closer look at the transmittance of interferometer LPG A + LPG B 48 Fig. 5.7: Transmittance of interferometer for two perpendicular polarizations 49 Fig. 5.8: No changes were detected in the transmission spectra ofthe interferometer when

the portion of fiber between both gratings was heated 50 Fig. 5.9: Peak depths changes for the 90° polarization 50 Fig. 5.10: Peak depths changes for the 0° polarization 51 Fig.6.1: Experimental setup used to produce and characterize the photo-induced LPG 54

Fig. 6.2: Transmission spectra of a photo-induced LPG 56 Fig. 6.3: Transmission spectra of a photo-induced LPG after extended exposure 56

Fig. 6.4: Numerically calculated transmission spectra the photo-induced LPG using the

software IFOjGratings 57 Fig. 6.5: Transmission spectra of LPG for two perpendicular polarizations 57

Fig. 6.6: Calculated transmission spectra of LPG for a longer exposure to light within the experimental wavelength range used to determine polarization and thermal

dependence 58 Fig. 6.7: Transmission spectra of LPG for different temperatures 58

Fig. 6.8: Transmission spectra chalcogenide fiber after thermal-bleaching, the

photo-induced LPG is erased 59 Fig. A. 1 : An optical fiber is basically constituted of three material layers, the core, the

cladding and the polymer jacket ....67 Fig. A.2: The light will only propagate in the core of the fiber if its incidence angle lies

within a certain range defined by the numerical aperture 67 Fig. A.3: Two-layered step-indexed fiber transversal section drawing 76

Fig. A.4: Plot showing that the Bessel functions of the second kind diverge at the origin. .79 Fig. A.5: Plot showing that thee modified Bessel functions of the first kind, /,.(»"), also

Fig. A.6: Fiber profile for the Three-layered step-indexed fiber model 88 Fig. A.7: Plot of functions (A.66h) through (A.66k) for n2.eff = 2.4094 90 Fig. 2.8: A plot of f(nl t f f ) = C\(n2eg ) - D\(n2elf ), a graphical solution of (A.65) 91

Fig. A.9: Extract of MathCAD code used to calculate LPoi core mode effective

refraction index 92 Fig. A. 10: Graphical solution of (A.62b). The searched eigenvalue is the value of b for

which the curves f0l(b) and h^^b) intersect each other 93

Fig. A.l 1: Fig. 12 Graphical solution of (A.59). The searched eigenvalue is the value of

^ for which the curves J^ ' and ^ ' intersect each other 94 Fig. A. 12: Extract of MathCAD code used to calculate the effective refraction indexes of

HEi.n cladding modes using the two-layered fiber model 95

Fig. B.l: Photograph of injection setup 97 Fig. B.2: Butt-coupling between chalcogenide fiber and SMF-28 silica fiber ...98

Chapter 1 :

Introduction to chalcogenide

glass fibers

The chalcogens are the elements of the group 16 of the periodic table, that is, oxygen (O), sulphur (S), selenium (Se), tellurium (Te), polonium (Po), and ununhexium (Uuh) [1]. Chalcogenide glasses are compounds based on some of these elements (sulphur, selenium, and tellurium) by the addition of germanium, arsenic, or antimony to create stable glasses, leading to a small tendency of crystallization, and devitrification. Some elements, such as P, I, CI, Br, Cd, Ba, Si or Ti, can also be added to achieve particular thermal, mechanical and optical properties [2]. The high non-linearity, mid-infrared transparency, and photosensitivity to visible light are their main features. The oxides are generally not considered chalcogenide compounds for both historical and scientific reasons. Although the oxygen is part of the group 16 of the periodic table, its physical behaviour is quite different. The oxides are often insulators, with large bandgaps of about lOeV. On the other hand, chalcogenide glasses are semiconductors having bandgaps around 2eV.

1.1 Fabrication

1.1.1 Glass fabrication

Chalcogenide glasses can be prepared as vapor-deposited thin films, or as melt-quenched bulks and rods. To fabricate the rods the compounds are heated together in sealed evacuated ampoules. The temperature and time of heating are controlled to prevent undesirable effects such as crystallization, liquation and stria formation [2]. The melting temperature ranges from 600 °C to 1100 °C depending on the composition [3]. After heating the batch for several hours the mixture is homogenized, the silica ampoules are then slowly cooled down to its transition temperature (Tg - 200 °C) by quenching the ampoules in either water or air. Annealing at transition temperature for several hours is an essential step to obtain a more resistant glass, allowing mechanical manipulation such as drilling, polishing and cutting.

One of the problems in manufacturing chalcogenide glasses is that the prepared glasses contain important amounts of hydrogen, and there is a potential danger of hydrogen bubbles blowing up during thermal treatment. That is why historically the arsenic trisulfide was prepared in an open system. The development of a closed system fabrication method was a breakthrough improvement towards the fabrication of high purity chalcogenide glasses. Other challenges are the high content of impurities (oxygen, hydrogen, and carbon) found in commercial samples of pure chalcogens and arsenic, and the fact that chalcogens can oxidize very easily in contact with room environment even at room temperature [2].

1.1.2 Fiber Fabrication

After glass rods are fabricated, the double-crucible process is used to draw the fibers. To obtain the core-cladding refractive index difference, two slightly different glasses are fabricated, for example, As4oS5gSe2 and AS^SÔO for core and cladding respectively.

The double-crucible method consists in feeding two quartz crucibles with the molten rods of both core and cladding. The crucibles are placed concentric to each other, as

placed in a furnace to control the temperature of the glasses. Each crucible is separately connected to a pressure-controlled atmosphere. The pressure, the temperature and the rate of the intake drum will control the drawing rate and the diameters of both core and cladding.

After drawing the fiber a polymer coating is immediately placed to protect its surface against scratches. The crucible method provides a high quality core/cladding interface, resulting in a higher mechanical strength.

Cladding Glass _{Core Glass}

Furnace

Fig. 1.1: Simplified schematics of the double crucible method for optical fiber fabrication.

Many important parameters are considered to determine the quality of the fibre manufacturing, including: the diameter of the core; the ratio of the core/cladding diameters; the concentricity of the core and the cladding; the numerical aperture; the mechanical strength; the minimum admissible bending radius; the continuous length of the fibre; the

optical losses. Table 1.1 lists some of those parameters obtained by the manufacturer of the chalcogenide fibre used in our experiments [2].

Fiber characteristics Parameter

Core diameter

Core/cladding concentricity

(%)

Continuous fiber length Bending strength (GPa) Working temperature (°C) Minimum optical loss at wavelength (dB Km'1)

Optical loss at laser wavelength (dB Km1)

YAG:Er3+ (X = 1.94\im) CO (X = 5.5 - 6.3 \un) C 02a = 9.2-11.3n-m) Numerical Aperture

Radiation power transmitted through the fiber

3+ YAG:Er CO laser CO2 laser laser 3-30pm for single-mode. Up to 800pm for multi-mode. 80-90 Tens-hundreds of meters, up to 1 Km for AS2S3 fibers. 0.5-1.2

Uptol00°C-120°C 23 at 2.4 pm (As2S3) 80 at 4.3 pm (As2Se3) 60at4.8pm(As2Si.5Sei.5)

160 at 6.6 pm (As2Si.5 Se]5) 160 100-200 600-1600 0.12-0.5 1.5kJ/cnr >10W 1.8W

Table 1.1: Basic performances ofthe best fibers made of high-purity arsenic chalcogenide glasses.

1.2 Optical transparency and refractive index

The transparency region of a glass is the region where the absorption coefficient have very small values (a - 10"3 cm"1). This region lies between the electronic band and the vibrational band.

electronic transitions take place. Chalcogenide glasses are semiconductors, with bandgap energy ranging between 0.8 eV and 3 eV, therefore showing opacity for visible light (2 eV < hu< 4 eV). So, the electronic band for chalcogenide glasses corresponds to the UV and visible part of the spectrum.

The vibrational band is where the vibrational transitions take place, corresponding to lower-energy infrared wavelengths. Chalcogenide glass lattice vibrations are resonant for low frequencies, so non-linear multi-phonon absorption may only occur for low-energy photons. For AS2S3 single-phonon absorption glass occurs at 29.4u.rn, while two- and three-phonon vibrational transitions occur at 19.7pm and 9.8pm respectively. This explains the wide transparency of chalcogenide glass for mid-IR light. The spectral transparency extends up to -10pm for Sulfur based glasses, and up to ~15pm for Selenide based glasses.

The optical losses caused by multi-phonon absorption, electron absorption, Rayleigh scattering due to glass density variations, and attenuation due to weak absorption tail are all intrinsic losses. The experimentally measured losses, particularly in the transparency band, are much more important. Glass impurities may cause some absorption peaks to appear corresponding to the vibrational resonance of their chemical bond. For example, we can see in Fig. 1.2 two absorption peaks in the transparence window for the AS2S3 glass. The peak at 2.92pm corresponds to the OH" radical, while the peak at 4.03pm is associated with the radical S-H. The absorption-spectra Urbach-tail region of both AS1S3 ( 1 ) and As2Se3 (2) glasses extracted from [4]

The glass contamination is a result of three basic factors. The first is that the commercial samples of As, S, Se and Te are already contaminated with metallic impurities, hydrogen, oxygen, carbon and silicon. The second is the non negligible concentrations of hydrogen in the quartz ampoules used in the glass synthesis. The third factor is the presence of remaining impurities in the vacuum atmosphere where the glass is melted, as arsenic and chalcogens have the tendency for oxidation at high temperature. A list of impurities and their absorption lines found in sulfide and selenide glasses is presented in Table 1.2 [2]. More recently it was discovered that AS2S3 thin films present photosensitivity to 1.5pm

light [5]. Although the mechanism of this effect is claimed to be unknown, it could limit the use of AS2S3 chalcogenide fibers in all-optical switching at telecommunication wavelength.

UJ

o

____. UL LU

o

o

o

w CD «a: 0.01 0.001 ^ Absorption due 10 electronic transitions Absorption due to vibrational transitions I-0.6 2.6 4.6 WAVELENGTH (um) 6.6

Fig. 1.2: Absorption coefficient spectra for the As2S* extracted from [6]. Two peaks in the transparency

window associated to resonances ofthe impurities OH (at 2.92|_tm) and S-H (at 4.03pm) are shown.

Compound or functional group leading to absorption

Position ofthe maximum of the absorption band (pm) OH S-H Se-H Ge-H As-H P H H20 Ge-O P-O C02 2.92 4.01,3.65,3.11,2.05 7.8,4.57,4.12,3.53,2.32 4.95 5.02 4.35 6.31,2.86,2.79 12.8,7.9 8.3 4.33,4.31, 15.0

CSe2

cs

2

Arsenic Oxides (different forms)

Se-O Si-O

Non-identified bands presumably due to carbon presence 7.8 6.68,4.65 15.4,12.7,9.5,8.9,7.9,7.5 10.67, 11.06 9.1-9.6 4.65,5.17,5.56,6.0

Table 1.2: The center peak ofthe absorption bands for the main impurities found in As-S and As-Se Glasses, extracted from [2].

Fig. 1.3: Urbach tail region ofthe absorption spectra of both As2S3 (1) and As2Se3 (2) glasses extracted from

[4].

1.3 Refractive index of chalcogenide glasses

Chalcogenide glasses show high refractive indexes, ranging between 2.4-2.6 for sulfide glasses, and 2.4-2.8 for selenide glasses. High refractive index yields high Fresnel refraction loss and the well-guidance of cladding modes, which are the two major problems related to the light injection into chalcogenide glass fibers. The refractive index (n) of a given glass is wavelength (k) dependent. This dependence is usually well fitted with the Sellmeier's formula Eq. 1.1 as the sum of the relevant absorption resonances for a

8

particular material. In Eq. 1.1, A, is the oscillator force and A is the resonance wavelength for a particular i,h absorption peak, whose values are not yet found in the literature. However in the transparency region, the Sellmeier's formula can be reduced to the Cauchy relation [7], in which the Cauchy coefficients A, B and C are empirical parameters Eq. 1.2. The Cauchy coefficients found for some chalcogenide glasses are shown in Table 1.3.

»

2

- l = I

A,A2 (l.i) 2 . B C n = A + — + —r A2 AA (1.2) Glass type B [pm-]

_{c [ m l}

AS2S3 As2Se3 GeSe4 GeipAsioSego 5.41 7.56 5.73 5.52 0.2 0.14 1.03 0.12 0.8 -0.18 1.23 -0.46

Table 1.3: The Cauchy coefficients found for some chalcogenide glasses.

With these relations it is possible to calculate the material chromatic dispersion, which is defined in Eq. 1.3. In Fig. 1.4 a plot of the refractive index of AS2S3 and As2Se3 chalcogenide glasses, and of SiO? silica glass are shown. There is a notable difference of refractive index between chalcogenide glasses and silica glasses. In Fig. 1.5 the material dispersion calculated using Eq. 1.3 is plotted. The material dispersion together with the waveguide dispersion will give the total dispersion introduced to a propagating light pulse. The waveguide dispersion can be adjusted to compensate the material dispersion in a desired wavelength window.

2.5" 1.5-I 1.5-I - -- -Si02 As2S3 As2Se3 i 1 I M O6 1.5-10 6 ₂₁₀ 2.5-10 Wavelength [m]

Fig. 1.4: The refractive indexes of chalcogenide glasses are much higher than that of silica glasses.

10

Material dispersion [ps/nm-Km]

1000 _

5 10 110 1.5-10 210

Wavelength [m]

Fig. 1.5: The material dispersion of As2S* and As^Se* chalcogenide glasses, and of Si02 silica glass.

1.3 Photosensitivity

Chalcogenide glasses exhibit a great variety of photo-induced transformations when exposed to near bandgap light, some permanent and some reversible, depending its intensity, the wavelength, the exposition duration and the material exposed. Among the photo-induced physical modifications are changes in density, hardness, chemical reactivity, solubility, crystallization, decomposition, as well as modifications in electrical and optical properties [8].

1.3.1 Photodarkening

Photodarkening of annealed chalcogenide glass was first reported by Neufville et al [9]. It is a reversible phenomenon, in which a shift of the absorption spectra of the sample towards longer wavelengths after exposition to near bandgap light (Eg ~ 2.4eV) is observed. Accordingly to the Kramers-Kronig relation, a variation of the absorption coefficient implies a variation of the refractive index. In fact, Photodarkening was observed in AS2S3 films exposed to both bandgap (Ar laser) and sub-bandgap (He-Ne laser) sources. A An = 0.029 index change for AS1S3 sample illuminated by a He-Ne laser at 632.8 nm was reported [10]. The reversible photo-darkening can only be observed in non-crystalline materials, and is accompanied by reversible changes in the material is structural properties, like volume, density, glass transition temperature and microhardness. Photo-darkening using polarized sources may also induce anisotropy.

Photo-darkening is reversible through thermal-bleaching, which consists of annealing the glass near softening temperature (Tg ~ 180 °C for AS2S3), and also through photo-bleaching, the reverse process of photo-darkening, which is the result of the sample exposure to sub-bandgap light.

1.3.2 Reversible modifications

Other reversible effects may also occur in annealed glasses, including photo-induced defect creation, photo-photo-induced girotropy (circular birefringence and dichroism), photo-induced anisotropy (dichroism and birefringence) with band gap or sub-band gap sources, and more. However the giant photo-expansion is the most peculiar one, being specifically found in chalcogenide glasses subjected to sub-bandgap Urbach tail illumination.

The giant photo-expansion was first observed by Hisakuni and Tanaka [11] when a sample of AS2S3 glass was illuminated by a strongly focused He-Ne laser beam. The expansion observed (4%) is ten times larger than a conventional photo-expansion

12 phenomenon (0.4%), and it increases if the sample is illuminated at lower temperatures (up to 20% of expansion). Fig. 1.5 shows a picture extracted from [11] of the surface of a 50pm-thick As2S3 film exposed for 1000s to a lOmW focused He-Ne (2 eV) beam.

10

jam

«

__WÊ_^_^m^_^_______

.*Hr

Fig. 1.5: Scanning electron microscope image of the surface of a 50u.m-thick As2S3 film after 1000s-exposure to a lOmW focused He-Ne beam. Extracted from [11].

1.3.3 Irreversible modifications

Photo-induced irreversible transformations of both physical and chemical nature may occur to chalcogenide glasses exposed to bandgap light. Among them are found: vaporization, crystallization, contraction and expansion, photo-induced softening and hardening, photo-decomposition and photo-polymerization [8].

1.4 Applications of chalcogenide-glass fibers

The infrared transparency of chalcogenide glasses directly suggests the application of chalcogenide glass fibers in infrared high power delivery, and in chemical sensing. The high photosensitivity of this material allows the fabrication of integrated photo-induced devices.

1.4.1 Infrared power delivery

High power infrared CO and CO2 lasers operating at 5.4pm and 10.6pm respectively are required in applications involving industrial welding and cutting. Free electron lasers (MFEL) operating between 2pm and 4pm and Er:YAG laser at 2.94pm are used in medical applications. Chalcogenide fibers meet the requirements for all of those laser systems due to their low losses in the infrared region and large threshold to damage Rare earth doped chalcogenide fibers are believed to have a good potential in the conception of mid-infrared laser sources. Many applications would arise from the development of these sources, especially in the development of bio-chemical sensors [12].

1.4.2 Chemical sensing

Quantitative remote detection and identification of chemical species are particularly interesting in longer wavelengths since the molecular fingerprints are stronger in the mid-infrared region. Several detection systems using chalcogenide glass fibers have been studied and applied to several materials, including liquids, gases and solids [12].

14

1.4.3 Photo-induced devices

The high photosensitivity of chalcogenide glasses to visible light is an important property to the fabrication of photo-induced devices.

Sampled Bragg gratings were fabricated in AS2S3 rib-waveguides [13] using CW frequency doubled Nd:YAG laser (at 532 nm). Relief gratings were fabricated in chalcogenide thin films, and Bragg gratings were generated in chalcogenide ridge waveguides using an Argon ion laser (at 514.5nm) [14 15]. Photo-induce Bragg gratings were fabricated in chalcogenide fibers using CW He-Ne laser (at 632.8nm) and the transverse holographic method [16], and more recently using phase masks [17].

Long-period gratings were mechanically induced in As2Se3 chalcogenide fibers [18, 19, 20]. This work is, in the best of our knowledge, the first report of the mechanically induced long-period gratings and photo-induced long-period gratings in AS2S3 chalcogenide fibers.

1.5 Non-linear properties and applications

Chalcogenide glasses are known for being highly non-linear materials. Glasses possessing high third-order non-linearity are important to the development of all-optical switches, especially for telecommunication systems. Values of X. UP t 0 400 times bigger than silica is were measured for chalcogenide glasses [21, 22]. A picosecond optical Kerr shutter using only a 48-cm-long chalcogenide fiber was demonstrated by Asobe et al [23].

High non-linearity also find applications in spectral broadening [24] and therefore chalcogenide fibers show a potential to the development of Supercontinuum sources. Second harmonic generation was observed in intrinsic and Pr-doped GaLaS glasses [25], which is phenomenon that is not a usually observed in glasses, and it is believed to be due to a local crystallization or the effect of frozen-in electric fields. Raman amplification with

gain about 300 greater than that for silica was found [26] , a promising result to the development of As-Se Raman fiber lasers and amplifiers.

1.6 Summary

A review of chalcogenide glasses, their properties and major applications were covered in this chapter. Moreover a detailed bibliographical review on chalcogenide glass fibers, including the manufacturing processes, optical properties, mechanical properties, and the most important developments and applications were presented.

16

Chapter 2:

Long-period fiber gratings

Long-period fiber gratings are structures created in the core of the fiber fundamentally designed to couple light from a core mode to co-propagating core modes, co-propagating cladding modes or leaky modes. These devices are a result of a periodic index change in the core of the fiber, with periodicities in the hundreds of micrometers that may be either photo-induced or mechanically induced.

The first long-period fiber grating was reported in 1983 by Youngquist et al. [27]. In this work the grating was mechanically induced through micro-bends on the cladding of a biréfringent polarization-maintaining fiber, the periodic pressure applied by a mechanical grating induces a periodic index change in the core creating a very biréfringent long-period grating. The purpose of this grating was to couple light propagating with one polarization to the other. An all-fiber in-line Mach-Zehnder interferometer using two in-series long-period gratings was also proposed in this work.

With the advance of externally photo-induced index-changing techniques in optical fibers, externally-written long-period gratings were first proposed by Hill et al. [28] in 1990. The fiber was operating under cutoff to allow bi-modal propagation. The grating was designed to couple LP0i core mode into higher order LPn core modes. Since both modes

are not azimuthally symmetrical, the grating had to be blazed to improve coupling efficiency between the modes. Symmetric photo-induced long-period gratings were also studied by Hill et al. [29] to couple circularly symmetrical LPoi and LP02 core modes.

The use of long-period fiber gratings in single-mode fibers to couple light from the fundamental LP01 core mode to cladding modes was first studied by Vengsarkar et al. in 1996 [30]. The light coupled to the cladding modes decay rapidly due to scattering losses and bends in the fiber, leading to the application of these devices as band-rejection filters. Since cladding modes are more sensitive to environmental conditions, long-period gratings gain another application as sensors. Our work is focused on this later application of long-period gratings.

2.1 Theory of long-period gratings

We limit our analysis considering a single-mode fiber, in which a uniform and untilted grating was written [30]. The grating consists of periodical circularly symmetrical index perturbations along the fiber axis. In this case the fundamental mode LP01, propagating in the core of the fiber, can only couple to cladding modes with azimuthal order 1=1, that is HE 1^ modes.

The phase-matching condition between the fundamental guided mode LP01 and the forward propagating cladding mode is given by Eq. 2.1, where fim is the core mode propagation constant, /?((;"' is the plh-cladding mode propagation constant, and A is the pitch of the grating.

P m . - P T = ^ y (2-D

A

The ratio power that is actually coupled to a particular pth-cladding mode to the power initially guided in the core of the fiber is given by [30]:

IS

1. f

^

2 s m '

_*"A

,+

PÏ\L) _

1 K K\u-o\ J ^01 (0)

r s \

2 i + i^ *"I^-OI y (2.2)

where ô is defined by,

<? = 1 f 2n^ (2.3).

The coupling constant K has a closed form [31] in the case of the coupling between LP0i core mode and HE|.M cladding modes and is given by :

*"l//-01 — &

(2n^

\

D

nb Zn«-,Vl + 2bA V ^ 0 " 2 x n\ 'u\ (n2 .e ir )

[uyn

2eir

)]

2

-V

2 a, 2

+ *

tr,D\<n2tir) x Ui (n2 . e ( l )JM \X U\ (n2 . e f f ) )

y

0

(vViTfe

KJ.(Vylî^b ■JQ(a.xu.(n2^)) VyJY^b (2.4)

where <r is the normalized index change. All the functions and parameters used in this chapter are detailed in Annex A. Fig. 2.1 shows the MathCAD simulated spectrum of a uniform long-period grating in a chalcogenide fiber.

-5 " î o -PdB(X.)-|5 -20 25 -LPoi«-.HE|,4_ -30 1.2 I0~ 1.3-10 LP0,~HE 1.5 LPoi^HEif, _L 1.4-10 1.5-10 1.6-10 1.7-10 1.8-10

Fig. 2.1 : Transmittance of an uniform LPG written in a chalcogenide glass fiber simulated in MathCAD. The parameters used in the simulation were: Q. = 2pm ; fl, = 2.4175; a2 = 60.5/^/t ; n2 = 2.4095;

/ i3= l ; <7=I0"5. A = 437//m

2.2 Numerical results on the calculation of the cladding effective

index

Two models are proposed to calculate the cladding mode effective refractive index. The first and simpler is a two-layered fiber model, in which the effect of the core is neglected, that is, only the interface cladding air is considered. The second model, a more complex one, is a three-layered fiber model in which the effects of the core are taken into consideration. In this section we present the error that is produced when the two-layered step-indexed fiber model, Fig. 2.1a, is used instead of using a three-layered step indexed fiber model, Fig. 2.1b. That is, we show the error produced by neglecting the effect of the core in the calculation of the cladding-mode propagation constants, using the same fiber

20

parameters. The results are shown in Table 2.1. Details on how these models are defined are discussed in the Annex A.l.

(a) (b)

Fig. 2.1: Schematics ofthe two-layered step-indexed fiber model (a) and the three-layered step- indexed fiber model (b).

From Table 2.1 we conclude that when using the simpler method we introduce a relative error of about 5x10^% to our calculations, and therefore the use of the three-layered fiber model is unnecessary. Table 2.2 shows the error introduced by the two-layered fiber model when calculating the pitch of a long-period grating. The errors are in the order of 10%, which are much high then that added to the calculation of the effective refractive index, depending on the application this error cannot be neglected.

M n2 . , f f "-■dl e{%)

Mode Two-layer Model Three-layer Model Relative error p = l 2.409472174 2,409483 4,49308X1G-4 p = 2 2.40940638 2,409424 7.31295X10-4 p = 3 2.40927154 2,409321 20,5x10 4 p = 4 2.409162459 2,409183 8,52613xl0"4 p = 5 2.408984219 2.409002 7,38106xl0"4 p = 6 2.408769779 2.40878 4,2432x10"4

Table 2.1 : Error introduced to the calculation of the effective refractive index of cladding modes associated with the two-layered fiber model approximation.

Radial _{n2,ff " 2 , 1 1} Core _{AL P G Am ;} Relative error

number _{Two-layer Three- effective Two-layer Three-layer} l n AL P G

u Model layer index Model Model £(%)

(HE,,,) Model

Three-layer Model 4 2,40911 2,40918 2,41031 959,867 889,61947 7,89641 5 2,40890 2,40900 2,41031 818,688 766,23883 6,84497 6 2,40865 2,40878 2,41031 694,936 654,84620 6,12205 7 2,40836 2,40852 2,41031 590,191 559,57325 5,47166

Table 2.2: Error introduced to the calculation ofthe pitch of a long-period grating generated by the two-layered step-indexed fiber model simplification.

2.4 Applications of long-period gratings

Long-period gratings have been successfully used as mode converters in multi-mode fibers [28, 29], as polarization couplers in biréfringent fibers (rocking filters) [27, 32], and as band-rejection filters [30]. LPGs have found applications in nonlinear optics, see e.g., [33]. Nonlinear-optical switching and pulse reshaping associated with the intensity-dependent coupling between core and cladding modes were observed.

Moreover, since cladding modes are dependent on the surrounding medium, the use of LPGs to couple light from the fundamental core mode LPoi to co-propagating cladding modes [30] is of special interest to the development of environmental sensors. Therefore a LPG can be specially designed to meet the requirements of strain, temperature and also refractive index sensors [34].

In-series long-period gratings can be set to create all-fiber Mach-Zehnder interferometers [27], a diagram is shown in Fig. 2.2. The use of LPG Mach-Zehnder interferometers in environmental sensors was also proposed as an alternative to improve the LPG-based sensors response [35]. Part of the light propagating in the core of the fiber is coupled to the cladding by LPG A, and then after a given distance L the LPG B couples' some of that light back to the core, at this point an interferometric process will happen, the effective path length depends on environmental conditions (n3). A Michelson

22

interferometer using one LPG followed by a mirror coated on the fiber ending was also suggested to improve LPG based environmental sensors [36], the schematics extracted from the cited work is shown in Fig.2.2.

LPG also found application as couplers between core modes and leaky modes [37] and [38]. In this case LPGs are coated with a material whose refractive index is higher that that of the cladding. More recently the use of a metallic-coated LPG-based Mach-Zehnder interferometer as a surface-plasmon-resonance sensor was theoretically studied by Yue-Jing He et al. [39], the idea is illustrated in Fig 2.3 extracted from the cited paper.

n

3

of environnement

• 11111 | t m 1111 rîrï 1111 ffm 111 (Tm 111 III I I [il'll II III II II I III I II M _ » II IIII II mmj„^„,™mj,,^_ ^ 71 ^ 71

^ _"-*•**" _ ^ V _

LPGA LPGB

Fig. 2.2: Two in-series LPG acting as a Mach-Zehnder interferometer.

Optical Source (LED)

Optical Spectrum Analyzer

cf-xSTtor O p f c - F l b - i r Sensor Head

<j

LPG Mirror \

( Core (HUH) Cladding

7

LPG1(A,)

C

Analyte Metal

I Core Mode Cladding

*2

Core SPW

LPG2(A2)

Fig. 2.4: Two in-series LPG acting as a Mach-Zehnder interferometer, used as surface-plasmon-resonance sensor, extracted from [39].

2.3 Summary

This chapter has introduced the fundamentals of long-period gratings. The theory and a set of ready-to program equations for the case of a uniform and untitled grating are presented, and are illustrated by simulated sketches of their transmission spectrum. Finally, a brief review of the developments and the applications of long-period grating is provided.

24

Chapter 3:

Observation of

photo-induced changes

This chapter presents the photo-induced effects that were observed in our laboratory, of both structural and optical nature. The experiments were performed using different laser sources for both selenide and sulphide fibers, with the interest of finding the optimal conditions to write diffraction gratings in the core of chalcogenide fibers.

3.1 Introduction

Strong photo-induced index changes were observed in selenide and sulphide chalcogenide glass fibers. The selenide fiber was exposed to a focused 50mW He-Ne laser at 633nm. The sulfide fiber was exposed to focused a 50mW He-Ne laser at 633nm, and to a focused Coherent Verdi laser at 532nm. A z-scan like technique was used on As2S3-based fibers.

The absorption spectra of both glasses in the Urbach tail region are shown in Fig. 3.1. The He-Ne line at 633nm corresponds 1.96eV bandgap energy. The absorption

coefficients corresponding to the He-Ne radiation are ur, =3.16cm_l and a2 ~ lOVwf1, for AS2S3 and As2Se3 fibers, respectively. The Verdi laser emission is at 532nm, which corresponds to 2.33eV and to an absorption coefficient of a. ~ lO'Vm"1 for the AsiS* glass.

Fig. 3.1 : Urbach tail region of the absorption spectra of both AsiS* ( 1 ) and AsiSe* (2) glasses from [40].

3.2 Photo-induced melting and ablation in chalcogenide fibers

Photo-induced melting was observed in As2Se3 fibers using a 50mW He-Ne laser source. The He-Ne laser beam was focused by a convex lens with f=50mm. The spot size is estimated to be 2w0 = 63pm. The idea was to focus the beam in the core of the fiber, however after 3pm of propagation length inside the fiber 95% of the laser power is already absorbed by the fiber. So in the core of the fiber (d = 60.5pm) there is not any power left.

26

The fiber was melted after a few seconds of exposition to the 50mW laser. A picture is shown in Fig. 3.2.

Fig. 3.2: The As2Se3 fiber melted after the exposition to a focused 50mW He-Ne laser (at 633nm).

Photo-induced laser etching on the chalcogenide fiber was achieved using the Coherent Verdi laser (532nm), a picture is shown in Fig. 3.3. At this wavelength the absorption coefficient of the AS2S3 glass is about a3= l 03c m ~ \ Therefore, after a propagation of 30 pm inside the fiber 96% of the energy is transformed into heat, causing the material to evaporate.

Fig. 3.3: Periodic laser ablation produced in a chalcogenide glass fiber.

Photo-induced melting was also observed in AS2S3 after a ten-minute-long exposition to a focused sub-bandgap He-Ne beam. Photo-induced melting is a particular phenomenon and is believe to be related the photo-vaporization effect. The mechanisms involved in this phenomenon are not completely understood. A picture of the melted fiber is shown in Fig. 3.4.

28

3.1.2 Photo-induced index changes in AS2S3 fibers

A z-scan-like experiment was performed to measure the local changes in the refractive indexes of the fiber induced by a focused 50mW He-Ne laser beam. This experiment was done (in analogy to the traditional z-scan technique) to study the photosensitivity of these fibers and to find the optimal conditions for grating recording.

The beam was focused at different positions before the fiber, inside the fiber and after the fiber, as shown in Fig. 3.5. The power of the central portion of the transmitted beam was continuously measured using a power meter, so we had the temporal evolution of the transmitted power intensity. The signal was acquired until it was stabilized.

The experimental setup is shown in Fig. 3.6 and Fig. 3.7. The laser beam is focused by an f=19mm convex lens. Then, the transmitted light was spatially filtered by an iris placed 10 cm away from the fiber. The power meter, situated behind the iris, was connected to a data acquisition board. The computer controlled the exposition through a shutter that was placed in front of the laser source. The computer opens the shutter, and then it starts automatically to acquire the signal from the power meter and to count the time. In the case of the exposure of AS2S3 to the He-Ne laser, there is about 98% of the laser power that reaches the core. Using a f=19mm convex spherical lens we have an intensity of about

\0*Wm 2.

» r _ • _

v*-0 1 i ■ . !

j 100 20 0 - 2 0 -100 1

Fig. 3.5: The laser beam focus spot creates an index change in the fiber.

The dynamics of the light intensity for the situation in which, as we strongly believe, the focused spot was in the core of the fiber is shown in Fig. 3.8. A remarkable

change of the transmitted power is observed. The increase of the light power is believed to be related to an increase in the refractive index (positive lensing). The attenuation of the light power in the end of the exposition time is believed to be due to negative lensing and the photodarkening effects, once that a shift of the absorption spectrum to longer wavelengths is expected. 1 ____________

1 '

ns

1

Spherical Lens Laser Beam | | 1 ____________

1 '

ns

1

Photodetector Spherical Lens Laser Beam | | 1 ____________

1 '

ns

1

MtÊ^Ê—jm^ÊM Spherical Lens Laser Beam | | 1 ____________

1 '

ns

1

•• 1 m

HI

» 1 ____________

1 '

ns

1

J_\ Optical fiber » 1 ____________

1 '

ns

1

»

Fig. 3.6: Picture ofthe experimental setup used to measure the refractive index changes in the fiber.

/ - I.9ca 10cm

k — * —

Photo.

ill'Ire lor

30 2,0-> o o Q . QJ E 00 03 1.0 0,5- 0,0-200 400 — l — 600 800 Time [s]

Fig. 3.8: Transmitted light intensity dynamics for a beam focused in the core of the fiber.

The dynamics of the light intensity for several focal spot locations with respect to the core of the fiber are shown in Fig. 3.9. The negative values of the positions indicate that the spot location lies in the region between the core of the fiber and the lens. In the other hand, the positive values of the positions indicate that the spot location lies in the region between the core of the fiber and the photo-detector. This experiment was done (in analogy to the traditional z-scan technique) to study the photosensitivity of these fibers and to find the optimal conditions for grating recording.

-120um X -60um -20um Coeu 2 C Coeur 20um ..--■ .■--. .■•■-■ .■■-•. ^ ^ C 60um 100 200 300 400 500 120um Time [s]

Fig.3.9: Transmitted light intensity dynamics for a beam focused in different portions ofthe fiber.

3.4 Summary

Several photo-induced effects were observed in chalcogenide glass fibers exposed Urbach tail sub-bandgap laser, including photo-darkening, photo-induced melting, and photo-induced etching. The dynamics of the photo-darkening process was measured using a z-scan-like method. The mechanisms involved in the photo-induced effects are not completely understood.

32

Chapter 4:

Mechanically-induced

Long-period Gratings

The successful fabrication of mechanically induced long-period gratins in AS2S3 chalcogenide fibers was achieved using an external mechanical clamp to create a periodic index change inside the fiber. The results are discussed in this chapter.

4.1 Description

A mechanical grating was clamped against the fiber using the clamping device shown in Fig. 4.1. A periodic refractive index modulation is induced in the fiber through the photoelastic effect of the periodic pressure applied on the fiber [41, 43]. In that case, it is recommended that the fiber polymer jacket not to be removed on the surface where the mechanical grating touches the fiber, in order to preserve its integrity. The fiber parameters are listed in Table 4.1.

Parameter Discription nl =2.4175 Core refractive index

n2 = 2.4095 Cladding refractive index n3 = 1.5 Polymer jacket refractive index

al = 3 pm Core diameter a2 = 121 pm Cladding diameter

Core composition AS40S60 Cladding composition AS39.5S60.5

Core density (g/cm3) 3.198

Scratch hardness 109( 1 OOg) Knoop (kg/mm2)

Young's modulus 1.585xl010 N/m2

Softening point /annealing 185 °C temperature

Termal expansion coefficient 21.4 (pm/°C)

Thermal conductivity 4xl04cal/cm

Table 4.1 : Optical and physical parameters of the As2S* fiber used in the experiments.

An additional birefringence, on, is also induced by applying a lateral pressure to the fibre [41]. This birefringence can be treated as a perturbation being additive to fibers normal birefringence if a small force is applied. The induced birefringence is given by Eq. 4.1,

_ 4n3 l + v/ /

* = = - ( . P i 2 - A i H - . (4-*)

K L Zr

where Poisson's coefficient is v'=0.16, the difference of the photo-elastic coefficients is p]2 - p . . =0.15, n is the mean refractive index of the fiber (~«2), E is the Young's modulus, / is the force per unit length applied to the fiber, and r is the fiber radius (external).

Birefringence is a property of a media in which its refractive index depends on the direction of light polarization (birefringence is naturally observed in crystals, but it may also be induced in amorphous materials by several means). To create local birefringence of

N

ôh = 10~*, a force per unit length / =6.188x105 — is required. m

The clamping device consists of two independent and identical fixation clamps and two other identical clamps where the mechanical gratings are held which will be called grating clamps. All clamps are mounted on a rail, as illustrated in Fig. 4.1, to enable adjustments in the distance between both gratings. The mechanical gratings were

34

manufactured using a high precision dicing saw. Grooves were printed on aluminum plates, which were attached to the grating clamps.

Fixation clamps

Fiber grooves

Fig. 4.1 : The clamping device used to create the long-period gratings.

7(K>nm 50nm Fiber

J L

The profile of a typical mechanical grating is sketched in Fig. 4.2. The pressure on each grating clamp is adjusted by a micrometric screw of 80 threads per inch, enabling a fine control of the coupling conditions of the induced LPG.

4.2 Experimental setup:

A collimated Supercontinuum light source (SCLS) beam passes through a spectral filter and an infrared polarizer. The output light is then injected in the core of the chalcogenide fiber. The other extremity of the chalcogenide fiber is then butt-coupled, using a high precision 6-axis positioning system, to a two-kilometer-long SMF-28 silica fiber to suppress the cladding modes. Finally the SMF-28 fiber is connected to an optical spectrum analyzer. The schematic ofthe experimental setup is shown in Fig. 4.3. Details on the injection setup, as well as the procedure used to achieve fast and efficient injection are presented in the Annex B.

As2S3 fiber Fiber support/ mechanical grating SMF-28 fiber 20x aspheric lens Polariser Spectral filter

V

Supercontinuum light source

36

4.3 Experimental results

This section presents the experimental results that were obtained with the mechanically induced long-period gratings. Gratings with different pitches (A) were fabricated. The results are presented for A = 700pm.

Results regarding the pressure dependency of the induced grating, the induced birefringence, and the grating thermal sensitivity are discussed in the remaining sub-sections of this chapter.

4.3.1 Birefringence of the mechanically induced LPG

An increase in the clamping pressure increases the coupling coefficient between two given modes. However excessive pressure may lead to an overcoupling condition and the LPG -like spectrum is no longer observed. Fig. 4.4 shows the changes in the transmitted spectrum as the clamping pressure is increased. On the green curve, for a smaller pressure, a single weak peak around 1360nm (the primary peak) appears, indicating a weak coupling between the core modes a single cladding mode. On the red curve, for a higher pressure, the primary peak is stronger and shifted of about lOnm. A weaker secondary peak around

1375nm also appears, indicating the coupling between the core mode and two cladding modes. In a 121pm-diameter cladding, hundreds of modes can be found [31]. So, for the same LPG, phase-matching conditions may occur for several modes, at different coupling coefficients. As a result the typical LPG transmission spectrum usually contains more than one peak. Evidently, the number of peaks displayed depends on the wavelength range of the acquisition. The result has good accordance to that simulated using the software IFO_Gratings, which is displayed in Fig. 4.5. According to the simulation, the primary peak corresponds to the coupling between the fundamental core mode LP0i and the cladding mode HEi.i. The secondary peak corresponds to the coupling between the fundamental core mode LP0i and the cladding mode HE1.2.

The numerical simulation was performed using the same length and pitch than the experimental setup (L = 2.8cm and A= 700pm). An index change of An = 8x10"* was imposed to obtain this result. This index change is quite strong, and in the simulations it was observed that over-coupling would occur for higher modes, which could explain why a third peak is not observed in the experimental results.

2 - i 1 -0

s. -

1

-• H -2H to g "Ë -3-| 01 c ta C . 4 - ₅ 6 --7 No grating Low pressure High pressure 1300 1350 1400 1450 1500 Wavelength [nm]

Fig. 4.4: Transmittance of LPG for different clamping pressures.

As shown in Fig. 4.6, there is a difference of the transmission spectrum for two arbitrary perpendicular polarizations. This result is in accordance with the photo-elastic effect induced by the lateral pressure from the mechanical grating, as discussed earlier.

The two perpendicular polarizations shown here were not determined based on spatial experimental parameters, such as the direction in which the clamping pressure is applied to the fiber, although it is known that these parameters are related. Instead, a wideband infrared linear polarizer was placed after the unpolarized SCLS. The axis angle of the polarized was scanned to determine the set of two perpendicular polarizations for which birefringence was found more accentuated.

38 m 2 , <__> o c CC _\ E w c ro 2 - , 0 2 6 -S JO -10 zz E CO -12 -14 1,25 1,30 — I — 1,35 — I — 1,40 1,45 1,50 — I 1,55 Wavelength [nm]

Fig. 4.5: Transmittance of LPG numerically simulated using the software IFOjGratings.

0 -I . M l - 2 -

^«P\

5T -4-__

U . v » « « i ^

8 c 6 -to j _ m 8 -—

r

t

ra __. _- -10-0° 90° -12- -14-i ■ -14-i ' -14-i ■ -14-i ■ - -14-i • 1300 1350 1400 1450 1500 Wavelength [nm]

4.3.2 Thermal sensitivity of the mechanically induced LPG

The grating transmittance was found to be sensitive to thermal variations. The metallic clamp was heated to up to 32 °C using a hot air blower. The variations of the grating transmittance were then measured for several temperatures as the system cooled down.

The same experiment was also performed for the two perpendicular polarizations determined as described in section 4.3.1. In this case birefringence was found more evident, since in addition to the grating birefringence measured in section 4.3.1, an additional birefringence associated with the thermal gradient created by the hot air blower will affect the transmission in the fiber.

The thermal dependence of the peak depth and its central wavelength were analyzed. The primary peak (1360 nm) is more sensitive to temperature variations for the 90° polarization, as seen in Fig. 4.7. On the other hand the secondary peak (1387 nm) is more affected for the 0° polarization, shown in Fig. 4.11. In fact, as shown in Fig. 4.12 and Fig. 4.13 for a same polarization, an increase of 5°C produces a change in the secondary peak of -5dB, while the primary peak changes only -ldB. Small primary-peak central wavelength changes were detected for the 90° polarization, however no significant changes were observed for the 0° polarization.

40 m Q) -5 u c ra £j E 05 C ra -15 1300 1350 1400 1450 Wavelength [nm] 1500

Fig. 4.7: Temperature variation of a single LPG for 90° polarization.

-10,0-, -10,5

-11,0-s

__ S. -■■•5-03 Q ___ ra œ -12,0-Q. -12,5--13,0 i • 1 i • 1 —i 1 r — | 1 1 1 1 1 1 1 1 1 1 1 25 26 27 28 29 30 31 32 Temperature [°C]

The fact that the higher order mode is more sensitive to environmental variations may be justified by its energy distribution along the fiber radius, as shown in Fig. 4.8, for higher-order cladding modes have more power concentrated close to the interface cladding/environment. <s 7

o

l_ s_^ _- H CO

Z

g o--^

o

v = l

v = 2

v = 3

v = 4

n • 1 • 1 ■ 1 ■

1-10 20 30 40 50

RADIAL POSITION ( \im )

60

Fig. 4.8: Extracted from [31]. Plots of the local intensity IA.r) as a function of radius for the four lowest-order / = 1 cladding modes in a typical fiber. All modes are circularly symmetric and have been normalized to carry a

1359.6- 1359,5--T- 1359,4-B -359,3 3 ra 1359,2 1359.1 1359.0 25 26 27 —r-_{28 29} l 30 31 Temperature [°C] - 1 32 42

Fig. 4.10: Thermal dependence of the primary peak central wavelength for the 90° polarization.

O-i 2 -m -4 T ) CÙ ( J c

s

-6 F UJ r TO G -8 -10--12 1300 1350 1400 1450 Wavelength [nm] 1500

-7- -8-CD S- -9 H <D T3 -_: _ra » -10-1 -11 --12 EquatW- y - a . b'x W-agM N o W a g h t m q Residual & m 0,02028 ol Squares A4. R-Square 0,97203

Value Standaid Erro PeaKI deo-hl wereeo" 3.732 1.2M44 Peak, decani Sloee 0 462 0 04503

26 27 28 29

Temperature [°C]

30 31

Fig. 4.12: Thermal dependence ofthe primary peak central wavelength for the 0° polarization.

-6-1 7 8 -a. -o % 9

a

ra ra -10 -11 -12-27 Equation y . a . b'* Weight NoWaghting Re-Kite! Sum 0,40787 ot Squares A4 R-Square 0.95189

Value Standard Emx Peak- depth I Intercept 35424 5.7S023 Peak2 depth [ Slope -1589 0.20196

28 29

Temperature [°C]

30

44

4.4 Summary

Mechanically induced long-period gratings in AS1S3 chalcogenide fibers were demonstrated by pressing a mechanical grating against the fiber. The grating was characterized using a Supercontinuum light source and an optical spectrum analyzer. The Grating birefringence and thermal dependence were studied.

A strong dependence between the grating thermal response and the light polarization was found. For a given polarization a particular peak depth of the grating spectrum may suffer more influence than another. A remarkable variation of AI = -5dB for AT=5 °C was observed for the secondary peak (at 1387nm) of the 0° polarization. A chosen polarization axis also affects the sensitivity of the central wavelength of each peak to thermal variations. While for the 90° polarization we observed a variation of AA, = 0.6nm for AT=5 °C, we found no wavelength variations for the 0° polarization. The successful demonstration mechanically induced LPGs in AS2S3 fibers led directly to their application in building an all-fiber Mach-Zehnder interferometer, which is presented in Chapter 5.

Chapter 5:

Two-LPG Mach-Zehnder

interferometer

Two in-series mechanically induced long-period gratings were produced to create an all-fiber Mach-Zehnder interferometer. The first grating couples light from the core mode LPoi to one, or more, co-propagating cladding modes. After a path length L another long-period grating couples back the cladding modes into core mode LP0|, as illustrated by Fig. 5.1. When combined, the signals will interfere since the propagation constants in the core and the cladding are different. Phase adjustments are achieved changing L. The mechanism used to induce both long-period gratings is the same shown in Fig. 4.1, only that this time two mechanical gratings are used. The experimental setup is shown in Fig. 5.2.

■ ^ H i n i ^ i i i ^ i i i ^ i n ^ m — » 1 im (JH+m (jfi i ifrffl Cn> v. ^ < . < v _

LPG A L LPG B

46

c Fiber support /

~ « 20x aspheric Spectral filter mechanical grating

\

Supercontinuum light source

Fig. 5.2: Layout ofthe experimental setup.

5.1 Experimental results

In order to build the interferometer the first step is to create one weak LPG (LPG A), as shown in Fig. 5.3. Then, another LPG is added (LPG B) about 2cm apart from LPGA, the pressure on both gratings are separately adjusted to obtain a good resolution of the interferometric fringes, shown in both Fig. 5.4 and Fig. 5.6.

The interferometer also presents birefringence. The transmission spectra of the system for two perpendicular polarizations are shown in Fig. 5.7. As in Chapter 4, the two perpendicular polarizations used to illustrate the grating birefringence were not determined based on spatial experimental parameters. The axis of a linear polarizer placed after the SCLS was scanned to determine the set of two main birefringence axis.

Numerical results have shown good accordance with the experimental ones. The results are displayed in Fig. 5.5.

-2-m 2. <B o c _ S •_ ao -3 c H -4-1 -5- -6-1300 1350 1400 1450 1500 Wavelength [nm]

Fig. 5.3: Transmittance ofthe week long-period grating LPG A.

1300 1350 1400

Wavelength [nm]

1450 1500

48

1,35 1,37 1,38

Wavelength [nm]

Fig. 5.5: The simulated transmittance obtained using the software 1F0 Gratings.

CÛ __ Q. O c ra g E <_. c TO 1350 1360 1370 1380 Wavelength [nm] 1390 1400

The thermal sensitivity of the interferometer was also studied. First only the portion of fiber between both gratings was heated, and as shown in Fig. 5.8, no significant change in the spectrum was observed. We believe that the reason for that is in the way the heating was done. In this experiment the fiber was heated by approaching the hot tip of a soldering station close to the fiber (~0.5cm). Since it was a localized heat source, we believe that there was a local index change, but not strong enough to have a significant impact on the response of the system. We believe that a better result would be obtained if the region was completely immersed in a temperature controlled environment, such as hot water.

Strong thermal sensitivity was found when one of the gratings was heated using a hot air blower. The changes in the three peaks of the interferometer fine structure (Fig. 5.6) were measured. In this, case significant changes in the peak depths were observed. The experiment was performed for two perpendicular polarizations, the results for the 90° and for the 0° polarizations are respectively shown in Fig. 5.9 and Fig. 5.10.

CO __ _a> o c E c ro 1350 1360 1370 1380 Wavelength [nm] 1390 1400

50

1350 1360 1370 1380

Wavelength [nm]

1390 1400

Fig. 5.8: No changes were detected in the transmission spectra of the interferometer when the portion of fiber between both gratings was heated.

-5,5-, -6,0 -6,5-m •D Q. -7,0 H CD Q -____ _ro 2 "7.5 H -8,0 -8,5 — Peakl [dB] — Peak2[dB] — Peak3[dB] 26 28 30 32 Temperature [oC] 34 36

-5,5-1 -6,0 -6,5 -7,0 -7,5 CQ 2- -8,0 1 -8,5 Q. -9,0 -9,5 -10,0 -10,5 — - P e a k l [dB] Peak2[dB] —-Peak3fdBl 26 28 30 32 34 Temperature [oC] 36

5.2 Summary

Fig. 5.10: Peak depths changes for the 0° polarization.

Two in-series mechanically induced long-period gratings in AS2S3 chalcogenide fibers were used to produce a Mach-Zehnder interferometer. Good accordance was found between numerical results and experimental results. The interferometer was characterized using a Supercontinuum light source and an optical spectrum analyzer. Its birefringence and thermal dependence were studied. A strong dependence between thermal response and the light polarization was also found for the interferometer. The interferometer was found less sensitive to thermal changes than a single long-period grating, which suggests that the distance L between LPG A and LPG B needs to be adjusted to obtain a more sensitive system.

52 Mechanically induced LPGs have shown strong thermal sensitivity and birefringence. Their applications as a single grating and in an interferometric setup are vast and yet to be explored. However, some negative aspects need to be taken into consideration:

• Even though the clamping device has micrometric screws to enable a fine control of the pressure applied to the fiber, it is believed that the precision is still not satisfactory. Numerical simulation revealed that the gratings were formed by index changes in the order of 10 , and at that point over-coupling was observed for higher-order cladding modes. This problem could be partially solved using a bigger duty cycle on the mechanical grating, and then reducing the pressure.

• After some time under the clamping pressure, the grating tends to relax as the polymer jacket goes beyond its elastic limit, leading to an irreversible deformation. This is reflected in a reduction of the peak depths after a few hours.

Motivated by the necessity of LPGs that are more stable and of high precision manufacturing procedures, photo-induced LPGs were produced, and are treated in the following chapter.

Chapter 6:

Photo-induced Long-period

gratings in AS2S3 fibers

As discussed on the previous chapters, chalcogenide glasses are photosensitive to visible light. This chapter presents the results of a photo-induced long-period grating written in an AS2S3 chalcogenide fiber using a He-Ne laser source. The main advantages of photo-induced gratings are the stability and the ability to have a finer control of the induced index change.

6.1 Experimental Setup

A long-period grating was produced in an AS2S3 chalcogenide glass fiber using a 50mW He-Ne laser (at 632.8nm). The laser beam was attenuated to 27mW, and a point-by-point method was used. The He-Ne laser beam was focused by a convex lens with f=50mm. The spot size is estimated to be 2w0 =63pm . The experimental setup used to produce and to characterize the photo-induced LPG is shown in Fig. 6.1.

The grating is 2.8-cm long (40 periods). The translation stage moves continuously while the shutter is periodically activated. The speed of the translation stage is controlled to