Embed Size (px)
Citation preview
_____________________________________
Photosensitivity, chemical compositiongratings, and optical fiber based
components
Michael Fokine
Doctoral Thesis
Department of Microelectronics and Information TechnologyRoyal Institute of Technology
Stockholm 2002
Photosensitivity, chemical composition gratings, and optical fiber basedcomponents
Michael FokineISBN 91-7283-397-1
© Acreo AB and Michael Fokine, 2002
Doktorsavhandling vid Kungliga Tekniska HögskolanTRITA-FYS 2239ISSN 0280-316X
Royal Institute of TechnologyDepartment of Microelectronics and Information Technology, Optics SectionElectrum 229SE-164 40 Kista
Acreo ABElectrum 236SE-164 40 Kista
Cover: ToF SIMS imaging of the fluorine distribution in the core and inner claddingof two optical fibers.
Printed by Universitetsservice US-AB, Kista, 2002.
I
Michael FokinePhotosensitivity, chemical composition gratings, and optical fiber based components.Department of Microelectronics and Information Technology, Royal Institute ofTechnology, SE-164 40 Kista, Sweden. TRITA-FYS 2239.
Abstract
The different topics of this thesis include high-temperature stable fiber Bragggratings, photosensitivity and fiber based components.
Fiber Bragg gratings (FBG) are wavelength dispersive refractive index structuresmanufactured through UV exposure of optical fibers. Their applications range fromWDM filters, dispersion compensators and fiber laser resonators fortelecommunication applications to different types of point or distributed sensors for avariety of applications.
One aim of this thesis has been to study a new type of FBG referred to as chemicalcomposition grating. These gratings differ from other types of FBG in that theirrefractive index structure is attributed to a change in the chemical composition.Chemical composition gratings have shown to be extremely temperature stablesurviving temperatures in excess of 1000 oC. Photosensitivity of pure silica andgermanium-doped core fibers in the presence of hydroxyl groups has also beenstudied and different types of fiber based components have been developed.
The main result of the thesis is a better understanding of the underlying mechanism ofthe formation of chemical composition gratings and their decay behavior at elevatedtemperatures. The refractive index modulation is caused by a periodic change in thefluorine concentration, which has been verified through time-of-flight secondary-ion-mass spectrometry and through studies of the decay behavior of chemical compositiongratings. A model based on diffusion of dopants has been developed, whichsuccessfully predicts the thermal decay at elevated temperatures. Studies of thedynamics of chemical composition grating formation have resulted in a manufacturingtechnique that allows for reproducible grating fabrication.
The main results regarding photosensitivity is a method to significantly increase theeffect of UV radiation on standard telecommunications fiber. The method, referred toas OH-flooding, has also been applied to pure-silica core fibers resulting in the firstreport of strong grating formation in such fibers.
Finally, research into different schemes for developing fiber-based components hasresulted in two types of single fiber integrated Mach-Zehnder interferometers; onepassive interferometer that can be used as an optical filter and one activeinterferometer controlled with internal metal electrodes.
Keywords: optical fibers, fiber Bragg gratings, photosensitivity, thermal stability,fiber sensors, chemical composition gratings, fiber components, Mach-Zehnderinterferometer, optical switch, optical modulator.
II
III
Preface
The work in this thesis was carried out at Acreo AB in Kista, Sweden, within theIndustrial Research School in optics in collaboration with the Department ofMicroelectronics and Information Technology (IMIT) at the Royal Institute ofTechnology (KTH).
The thesis has been made possible by the financial support from the KK-foundationand industry. I would like to pay special gratitude to the industrial partners, EricssonNetwork Technologies AB and Telia AB, for their financial support.
I would like to thank the board members of the Industrial Research school in Optics,Dr. Magnus Breidne (Acreo AB), Prof. Ari Friberg (KTH), Dr. Nils Artlöve (TeliaAB), Dr. Bertil Arvidsson (Ericsson Network Technologies AB) and Dr. LeifStensland (Ericsson Components AB) for their involvement during my studies.
Part of the work was completed within the framework of different industrial projectsat Acreo AB.
IV
V
List of publications
Paper I.
Formation of thermally stable chemical composition gratings in optical fibers.
M. Fokine, Journal of the Optical Society of America B, Vol. 19, 1759-1765 (2002).
Paper II.
Thermal stability of chemical composition gratings in fluorine-germanium-dopedsilica fibers.
M. Fokine, Optics Letters, Vol. 27, 1016-1018 (2002).
Paper III.
Growth dynamics of chemical composition gratings in fluorine-doped silica opticalfibers.
M. Fokine, Optics Letters, Vol. 27, 1974-1976 (2002).
Paper IV.
High temperature miniature oven with low thermal gradient for processing fiberBragg gratings.
M. Fokine, Review of Scientific Instruments, Vol. 72, 3458-3461 (2001).
Paper V.
ToF-SIMS imaging of dopant diffusion in optical fibers.
M. Hellsing, M. Fokine, Å. Claesson, L.-E. Nilsson, W. Margulis, Applied SurfaceScience, Article in press (2002).
VI
Paper VI.
Large increase in photosensitivity through massive hydroxyl formation.
M. Fokine, W. Margulis, Optics Letters, Vol. 25, 302-304 (2000).
Paper VII.
Grating formation in pure silica-core fibers.
J. Albert, M. Fokine, W. Margulis, Optics Letters, Vol. 27, 809-811 (2002).
Paper VIII.
Mach-Zehnder interferometer using single standard telecommunication opticalfibre.
F. C. Garcia, M. Fokine, W. Margulis, R. Kashyap, Electronics Letter,Vol. 37, 1440-1442 (2001).
Paper IX.
Integrated fiber Mach-Zehnder interferometer for electro-optic switching.
M. Fokine, L. E. Nilsson, Å. Claesson, D. Berlemont, L. Kjellberg, L.Krummenacher, W. Margulis, Optics Letters, Vol. 27, 1643-1645 (2002).
VII
Other publications and conference contributions by the authorrelevant to the subject, but not included in the thesis.
A. G. Heiberg, J. Skaar, M. Fokine, L. Arnberg, ”A new method for temperature measurement insolidifying aluminum alloys by use of optical fiber Bragg grating sensors,” 106th America FoundrySociety Casting Congress, Kansas, May (2002).
B. M. Fokine, L-E. Nilsson, Å- Claesson, D. Berlemont, L. Kjellberg, L. Krummenacher, W.Margulis, "Towards fiber electroopic switches: Considerations on electrodes, fiber design andtraveling wave arrangements", OSA Technical Digest (Optical society of America, Washington,DC, 2001), Post deadline paper, Stresa, Itally (2001).
C. M. Hellsing, M. Fokine, Å. Claesson, L-E. Nilsson and W. Margulis, "ToF SIMS imaging ofdopant diffusion in optical fibers,” 13th International Conference on Secondary Ion MassSpectrometry and Related Topics, Nara, Japan, November 2001.
D. J. Albert, M. Fokine, W. Margulis, “Grating formation in pure silica fibers,” Bragg Gratings,photosensitivity, and Poling in Glass Fibers and Waveguides, Applications and Fundamentals,OSA Technical digest series, (Optical society of America, Washington, DC, 2001), BThC9 (2001),(BGPP´01).
E. M. Fokine, W. Margulis, “Photoinduced refractive index changes in frequency doubling fibers”,Trends in Optics and Photonics, Bragg gratings, photosensitivity, and poling in glass waveguides,TOPS Vol. 33, pp. 385-262, (2000).
F. F. C. Garcia, R. Kashyap, M. Fokine, W. Margulis, "Highly Stable Mach-Zehnder Interferometerusing a Single Standard Telecommunication Optical Fibre", presented at Conference on Lasers &Electro-Optics, San Francisco, CThI7 (2000).
G. M. Fokine, W. Margulis, “Large Increase in Photosensitivity through Massive HydroxylFormation”, Bragg Gratings, photosensitivity, and Poling in Glass Fibers and Waveguides, 1999,Stuart, Florida, (BGPP´99) Post-deadling papers, PD4, pp. PD4/1-PD4/3.
H. M. Fokine, B.E. Sahlgren, R. Stubbe, “A novel approach to fabricate high-temperature resistantfiber Bragg gratings,” Bragg Gratings, Photosensitivity, and Poling in Glass Fibers Waveguides:Applications and Fundamentals, Vol. 7 of OSA Technical digest series, (Optical society ofAmerica, Washington, DC, 1997), 58-60 (1997).
I. M. Fokine, “Fabrication of High-Temperature Resistant Bragg Gratings in Optical Fibers”, 46th
International Wire and Cable Symposium, IWCS´97, 17-20 November, Philidelphia, pp. 64-67(1997).
J. M. Fokine, “Fabrication of high-temperature resistant Bragg gratings by post-alteration of thechemical composition of the fiber,” European Symposium on Lasers and Optics in Manufacturing,16-20 June, Munich, Paper 3099-45 (1997).
K. M. Fokine, B.E. Sahlgren, R. Stubbe, “High temperature resistant Bragg gratings fabricated insilica optical fibres,” Australian Conference on Optical Fibre Technology (ACOFT´96), 1-4December, Gold Coast, Post dealine paper PD2 (1996).
Patents relevant to the subject, but not included in the thesis.
L. WO99/50696, M. Fokine, An optical body having modifiable light guiding properties.
M. WO 98/12586, M. Fokine. Optical means.
VIII
Thesis structureChapter 1 is a short introduction to the thesis, motivation and evolution of theresearch.
Chapter 2 is an overview of hydrogen-glass interactions. As well as being a buildingblock of the universe, hydrogen has played a central role in the thesis and deserves itsown chapter.
Chapter 3 is an introduction to fiber Bragg gratings and photosensitivity and isintended to give the reader a general understanding of different areas in the field. Thischapter deals with definitions of Bragg gratings and their applications, how they aremade and a short summary of the different theories of the mechanisms responsible forthe change in refractive index. Included is a literature review of materialconsiderations for photosensitive fibers.
Chapter 4 is an overview of the issues regarding the thermal stability of fiber Bragggratings.
Chapter 5 is a separate chapter on chemical composition gratings regardingbackground, theory, manufacturing and models for their thermal stability.
Chapter 6 is a short description of different types of fiber-based components andtheir applications. This chapter deals with fibers containing internal metal electrodesand techniques for inserting such electrodes. This chapter serves as background toPaper VIII and IX.
Chapter 7 is a short discussion of the different papers included in the thesis.
Chapter 8 concludes the thesis with a discussion of different possible applications,field trials and implications of research results presented in the thesis.
Reproduction of thesis papers I through IX.
IX
AcknowledgementsI would like to thank Dr. Magnus Breidne for giving me the opportunity to performthe research and studies at Acreo, for being the coordinator for this project andpermitting me to visit different research-labs around the world. I would also like tothank Prof. Ari Friberg for accepting me as a PhD student at IMIT and for pushing meforward in my studies.
My colleagues at Acreo have had a strong effect on my work and me. I wouldespecially like to thank Lars-Erik Nilsson whose knowledge and efficiency is trulyinspiring, Åsa Claesson for all the discussions, her involvement and for all the fun,Leif Kjellberg for his magic solutions to just about any type of problem, NiklasMyrén for his good spirit, Ola Gunnarson for taking care of my bike, and all mycolleagues that I have had the pleasure to work with. I would also like to thank threeformer colleagues who have had a strong influence on my work and me. BengtSahlgren, for all the long discussions and many theories, Feliciano de Brito, whoshowed me the art of making preforms and fiber drawing and for the many hours infront of the preform lathe, and the ski master Håkan Karlsson for past and futureslopes.
I would also like to thank all the people who have taken very well care of me duringmy visits abroad. Many thanks go to Prof. Akira J. Ikushima, Prof. Kazuya Saito, Dr.Hiroshi Kakiuchida and the group at the Frontiers Materials Laboratory of the ToyotaTechnological Institute in Nagoya, Japan. I would also like to thank Prof. YounésMessadeq, Prof. Sidney J.L. Ribeiro and Dr. Édison Pecoraro at the Instituto deQuímica of the Universidade Estadual Paulista in Araraquara, Brazil. Many thanks goto Prof. Isabel C. S. Carvalho and her optoelectronics group at the physics departmentof Pontifícia Universidade Católica in Rio de Janeiro, Brazil. A better location for aUniversity is difficult to find.
Special thanks for the support and patience from the Fokine clan and my friends.Special thanks to my girlfriend, Åsa Uhrwing, who has stood by me during theseyears, accepting the late nights and weekends in the lab, listening to my problems,motivating me and for trying to understand what I'm actually doing.
I would also like to thank the many people that I have had the honor to collaboratewith during my studies. My thanks go to Dr. Raman Kashyap, Dr. Fatima C. Garcia,Dr. Jacques Albert, Dr. Jon Thomas Kringlebotn, Dr. Magnus Hellsing, and Prof.Johannes Skaar and his group.
Finally, I would like to thank my supervisor, Dr. Walter Margulis, for making my lifeeasy at times and difficult at others, for inspiring me, for pushing me into the lab anddo the experiments instead of just talking about them, for listening, for all thediscussions (of which some resulted in very strange ideas), for making my articlesunderstandable to others than myself, for introducing me to people and then sendingme off to various corners of the world, and for being a very good friend.
Many thanks to all of you!
Michael Fokine
X
XI
ContentsAbstract I
Preface III
List of publications V
Thesis structure VIII
Acknowledgements Error! Bookmark not defined.
1. Introduction, motivation and evolution of the thesis 1
1.1. Introduction and motivation of optical fiber Bragg gratings and fiber components 1
1.2. Evolution of the thesis 2
1.3. References 4
2. Hydrogen interactions with silica glass 5
2.1. Optical losses 5
2.2. Diffusion of hydrogen in silica 6
2.3. Diffusion of hydrogen in optical fibers 7
2.4. Hydrogen interactions with silica glass 82.4.1. Thermally induced reactions 82.4.2. UV induced reactions 102.4.3. Diffusion of water and hydroxyl groups in silica 102.4.4. Reduction of hydroxyl species in fluorine doped silica glass 11
2.5. References 13
3. Fiber Bragg gratings and photosensitivity 15
3.1. Fiber Bragg gratings 15
3.2. History 16
3.3. Classification of fiber Bragg gratings 173.3.1. Classification by coupling characteristics 173.3.2. Classification by growth characteristics 17
3.4. Mechanisms of photosensitivity 183.4.1. Color-center model 183.4.2. Stress relaxation model 203.4.3. Densification-compaction model 20
3.5. Increasing photosensitivity 213.5.1. Hydrogen treatment 213.5.2. Thermal treatment 233.5.3. Mechanical treatment 233.5.4. Preform manufacturing 23
3.6. Literature review of material considerations 23
3.7. References 30
XII
4. Thermal stability 37
4.1. Decay models 37
4.2. Methods to increase thermal stability 394.2.1. Thermal annealing 394.2.2. Pre and post exposure 39
4.3. Gratings with intrinsically high thermal stability 39
4.4. References 41
5. Chemical composition gratings 43
5.1. Background 43
5.2. Typical manufacturing procedure of chemical composition gratings 44
5.3. Hydroxyl assisted fluorine diffusion 45
5.4. Decay behavior of chemical composition gratings 485.4.1. Modeling the decay behavior of CCGs 495.4.2. Modeled and experimental decay of CCGs 50
5.5. Dynamics of CCG formation 52
5.6. Ultra-stable chemical composition gratings 55
5.7. Miniature oven for processing chemical composition gratings 56
5.8 References 59
6. Extended functionality of optical fibers 61
6.1. Background 61
6.2. Mach-Zehnder interferometers for filtering and switching 61
6.3. Single core bimodal Mach-Zehnder interferometer using OH flooding for refractive indexcontrol 62
6.4. Dual core Mach-Zehnder interferometer for electrooptic switching 646.4.1. Electrode insertion technique 656.4.2. Evaluation of dual core electrode filled Mach-Zehnder interferometer 67
6.5. References 69
7. Description of original work and author contribution 71
8. Conclusions and outlook 75
Reproduction of Papers I through IX 77
1
1. Introduction, motivation and evolution of the thesis
The following section is intended as an introduction to the readers not familiar to thefield of fiber Bragg gratings and photosensitivity and serves as motivation for theresearch performed and presented in the thesis. The second section briefly explainsthe evolution and chain of events leading to the nine papers included in the thesis.
1.1. Introduction and motivation of optical fiber Bragg gratingsand fiber components
The wide use of silica based optical fibers in telecommunications has made it possiblefor millions of people to have access and to share enormous amounts of information.Services such as email, pictures, games, banking, TV and radio channels can betransmitted or received within a very short time span. To provide for the increasingdemand of capacity, and maintaining sufficiently fast transfer of information,telecommunication systems are continuously being upgraded with increasingcomplexity. One important technology that has had a large impact ontelecommunications systems today is fiber Bragg gratings (FBGs). FBGs arewavelength dispersive elements that can be fabricated within the light-guiding core ofoptical fibers. There are a number of different FBG based components that have beentailor-made to meet the stringent requirements of telecommunications systems. Sometypical examples of such components are; narrow band reflectors for wavelength-division-multiplexing (WDM) and dense-WDM (DWDM) applications, fiber lasers,dispersion compensators, loss filters for gain equalization, add-drop filters and manymore [1]. Although the telecommunications market is by far the dominant market inboth volume and revenue, FBGs are increasingly being used for sensing, measuringe.g. temperature and strain [2]. Advantages of FBG as sensors are that they areimmune to electrical disturbances, are small in size allowing for integration intodifferent materials and structures, and the actuator is already integrated in aninformation-transmissive medium.
Bragg gratings are refractive index structures manufactured by exposing the core ofan optical fiber to intense periodic ultraviolet radiation. The ability to change therefractive index with radiation is referred to as photosensitivity. Althoughphotosensitivity in optical fiber has been extensively studied over several decades,there is still no unified theory for the physical and chemical mechanisms responsiblefor the changes in refractive index. One explanation of the lack of a unified theory isthat the photosensitive response may differ significantly depending on several factorssuch as the type of fiber, prior treatment of the fiber, UV writing wavelength and theUV laser writing power (chapter 3).
In most cases, the ideal situation would be to inscribe gratings into the low-lossstandard telecommunications fiber that are used in optical networks today. This wouldenable 100 % fiber compatibility and the use of low-cost mass-produced fibers.However, standard telecommunications fibers used today have a very lowphotosensitivity. Solutions to this problem have been to design special photosensitive
2
fibers and/or methods to make non-photosensitive fibers more sensitive to UVradiation.
Another issue that has been a concern for the deployment of fiber Bragg gratingdevices has been the lifetime of the grating (chapter 4 and 5). It was found that whenincreasing the ambient temperature that the grating reflectivity started to decay anddid not completely stabilize even for long treatment times. Additional increase intemperature resulted in an even larger decrease in reflectivity. Although the gratingsappeared to be stable at room temperature, this behavior placed significant limitationson deployment, considering the required lifetime of some components in thetelecommunications industry exceed several decades. The problems facing sensorapplications are even more stringent, as the temperature of the sensor may very wellexceed the typical –40 to +80 oC temperature requirements for telecommunications.
1.2. Evolution of the thesis
In 1996, research performed at the Institute of Optical Research (now Acreo AB)resulted in a novel type of fiber Bragg grating. These fiber gratings showed superiortemperature stability compared to other types of known gratings. The refractive indexchanges of these gratings were believed to result from a redistribution of dopantswithin the core, fluorine in particular, resulting in a periodic change of the refractiveindex. Although these gratings were soon used in field trials for different projectsinvolving temperature and strain monitoring at elevated temperatures, little wasknown about these gratings at the time. The aim of the thesis was therefore do studythese gratings in order to gain insight of the underlying mechanisms of refractiveindex change and their behavior at elevated temperatures (Paper I, II and III).
The fabrication of these gratings, referred to as chemical composition gratings,involves chemical reactions with in-diffused molecular hydrogen and requires thermalprocessing at temperatures near 1000 oC. As the grating is a phase structure,temperature variations within the heating zone need to be small. In addition, tomaintain the mechanical strength of the heat-treated fiber, the heated length should bekept minimal and preferably be treated in an inert atmosphere. For these reasons,traditional ovens could not be used and a specially designed process oven was built(Paper IV). To study the formation of chemical composition gratings, in particularthe effect of hydrogen interactions on the redistribution of fluorine in optical fibers, ananalysis technique to measure dopant distribution in optical fibers was required.Traditional analysis techniques have limitations regarding sensitivity or resolution dueto the light fluorine atom and the small dimensions of the core of optical fibers. Forthis reason, time-of-flight secondary ion mass spectrometry (ToF-SIMS) wasevaluated concerning dopant profiling of optical fibers (Paper V).
As a step towards understanding the formation of chemical composition gratings,studies of hydrogen interactions within hydrogen loaded optical fibers whenprocessed at temperatures up to 1000 oC were performed. The process, referred to asOH flooding, showed that a large and controlled amount of hydroxyl species could beformed within the core of the fiber. As photosensitivity in optical fibers has beenassociated with different types of defects, this heating process was applied to a
3
standard telecommunications fiber in order to increase the photosensitivity. Theresults showed a significant increase in the photosensitive response (Paper VI).
Photosensitivity is generally associated with fibers containing germanium andgermanium-related defects although there has been a great deal of research effort forfinding new dopants to further increase the photosensitivity. To further evaluate theeffect of OH flooding on photosensitivity tests were performed using pure-silica corefibers. Although containing no germanium or other dopants in the core, stronggratings were successfully manufactured in these fibers (Paper VII).
In addition to a large increase in photosensitivity, OH flooding results in a large andcontrollable change in the refractive index of the core. This effect was used to make aMach-Zehnder interferometer integrated into a single piece of standardtelecommunications fiber (Paper VIII).
The last paper of the thesis (Paper IX) differs slightly from the previous publications.The paper deals with fiber based components which can be viewed as a combinationof several different technologies. The main aim of this work was to extend thefunctionality of optical fibers. As telecommunication systems become more complex,the use of tailor-made materials, special processing techniques, micro-mechanicaldesigns, Bragg gratings and specialty fibers are combined to enable extraction ofspecific optical functions from fiber components. There is a vast number ofpossibilities and combinations to enable fiber-based components to perform a specifictask. In addition, the increasing complexity of optical networks also requires activeoptical components to enable flexibility and adaptability to changing conditions. Afew examples of active components are switches, modulators, variable attenuators andamplifiers. By combining a special method for inserting metal electrodes into side-hole fibers with multiple cores a 1 meter long single-fiber electric-field controllableMach-Zehnder interferometer was designed and evaluated (Paper IX).
4
1.3. References
1. Raman Kashyap, Fiber Bragg Gratings, Academic Press (1999).2. A. Othonos and K. Kalli, Fiber Bragg Gratings: Fundamentals and Applications
in Telecommunications and Sensing, Artech House, (1999).
5
2. Hydrogen interactions with silica glass
There has been a wide interest in hydrogen interactions with silica glass, and glass ingeneral, due to the effects on fatigue (weathering), mechanical strength, resistance toionizing radiation, optical absorption and defects generation (see e.g. [1]). Fortelecommunications, the main issues regarding hydrogen interactions with silica basedfibers are the detrimental effects on signal transmission and the use of hydrogenloading for increased photosensitivity (see chapter 3.5). The aim of this section is togive a general overview of hydrogen interactions with silica glass and silica basedfibers regarding optical losses, diffusion properties and chemical reactions.
The chemical reactions discussed in this chapter, especially figure 2.9, have been thebasis of the understanding and development of the chemical composition gratingsdescribed in chapter 5.
2.1. Optical losses
Today the most critical loss mechanism in modern telecommunications fibers ishydrogen contamination. Typical sources of hydrogen are ambient water, hydrogengas from oxy-hydrogen torches used in manufacturing or contamination from theoriginal material such as the deposition tube used in modified chemical vapordeposition (MCVD) [2]. The main hydrogen and deuterium related absorption bandsare shown in table 2.1. Deuterium is often used to reduce the absorption in thetelecommunications windows. When the hydroxyl groups is attached to germanium,the first overtone is shifted to ~1.41 µm and for phosphorous the band is shifted to 1,6µm (see table 2.2). Typically a loss figure of 50 dB/km/ppm wt OH measured at~1.37 µm can be expected while for the fundamental absorption peak (at 2,7 µm) thevalue is 0.1 dB/cm/ppm wt OH [3].
Si-OH Si-OD2.7 µm fundamental 3.7 µm fundamental1.37 µm 1st overtone 1.87 µm 1st overtone1.27 µm combination tone ~1.67µm combination tone0.95 µm 2nd overtone 1.26 µm 2nd overtone0.66 µm 3rd overtone
Table 2.1 Absorption bands of Si-OH and Si-OD [3].
Hydroxyl 1st overtoneSi-OH 1.37 µmGe-OH 1.41 µmP-OH 1.6 µm
Table 2.2 Absorption bands of hydroxyl species for different dopants [3].
6
The hydrogen molecule (H2) is not infrared active due to the symmetry of themolecule. However when incorporated into the glass, interaction with the silica latticeresults in several absorption features (see table 2.3). These absorption peaks, whichhave much lower absorption than OH-groups, have their first overtones near 1,2 µm,with the largest peak located at 1,24 µm. The spectrum of these absorption peaksdepends somewhat on co-dopants [4], but an approximate method to calculate theconcentration of molecular hydrogen in fibers is to measure the hydrogen inducedabsorption at 1,24 µm according to the following equation,
24.133.02
α⋅≈HC , Eq. (2-1)
where CH2 is the concentration of molecular hydrogen in mol% and α1.24 is theabsorption at 1.24 µm measured in dB/m [5].
H2 in silica D2 in silica2,4 µm fundamental 3,36 µm fundamental
1,245 µm 1st overtone 1,716 µm 1st overtone0.881 µm 2nd overtone 1.244 2nd overtone
Table 2.3. Main absorption bands for H2 and D2 in silica [3].
2.2. Diffusion of hydrogen in silica
When placing silica glass in a hydrogen containing ambient, hydrogen moleculesdiffuse into the glass network. Diffusion of hydrogen takes place without anysignificant amount of chemical reactions as long as the temperature is low enough.The hydrogen molecules are located in interstices in the silica lattice and the diffusionin silica follows an Arrhenius behavior, as described by the following equation.
��
���
�−=RTEDD aexp0 , Eq. (2-2)
where D0, the pre-exponential constant, can be considered independent of temperatureand Ea is the activation energy for the diffusion process. The following table givestypical values for hydrogen and deuterium diffusion in silica.
H2 in silica D2 in silicaD0 5.65·10-4 cm2s-1 D0 5·10-4 cm2s-1
Ea 43.54 kJ/mole Ea 43.96 kJ/mole
Table 2.4. Diffusion parameters for H2 and D2 in silica [3].
7
2.3. Diffusion of hydrogen in optical fibers
Solving the diffusion equation for a cylindrical structure results in the equationsshown below. The equations describe the normalized radial concentration for in-diffusion (Eq. 2-3) and out-diffusion (Eq. 2-4) of a mobile substance using theboundary conditions of constant radial concentration at time equals zero, i.e. C=0 or 1depending on in-diffusion or out-diffusion [6].
( ) ( )( )�
∞
=
−−=1 1
02exp21
n nn
nn
aJrJtD
aC
αααα (2.3)
for in-diffusion and
( ) ( )( )�
∞
=
−=1 1
02exp2
n nn
nn
aJrJtD
aC
αααα (2.4)
for out-diffusion where C is the concentration, a is the fiber radius, D is the diffusioncoefficient, t is time and αn are the roots of J0(aαn). J0 and J1 are the Bessel function oforder zero and one respectively. In figure 2.1, a simulation of the time dependence ofhydrogen in-diffusion is shown. The figure shows the normalized concentration ofhydrogen versus normalized radius for different times at a temperature of 50 oC. Thefiber radius used in the simulation was 62.5 µm, i.e. standard fiber dimensions.
0,00 0,25 0,50 0,75 1,000,00
0,25
0,50
0,75
1,00Temperature = 50 oC
Time [hrs] 0.1 1 5 15 25 40 60
Norm
aliz
ed c
once
ntra
tion
Normalized radius
Figure 2.1 Normalized hydrogen concentration vs. normalized radius for differenttimes at a temperature of 50 oC. The fiber radius used in the simulation was 62.5 µm.
8
2.4. Hydrogen interactions with silica glass
2.4.1. Thermally induced reactions
If hydrogen loaded fibers are exposed to temperature above ~150-250 oC, thermallyinduced reactions can occur. Hydrogen is assumed to react with a limited number ofpre-existing sites [7,8,9]. In thermally treated hydrogen-loaded germanium dopedsilica samples Awazu et al [10] found a linear dependence between the thermallyinduced hydroxyl concentration and the absorption at 5.14 eV (242 nm). Theyproposed the model for the reaction as shown in figure 2.2.
Figure 2.2 Proposed model for thermally induced hydroxyl groups and Ge-defects inHydrogen loaded bulk glass. T is either Si or Ge and GLPC-Germanium Lone Pair
Center [10].
The maximum achievable concentration of hydroxyls species is generally higher if thesilica glass is doped with germanium or phosphorous while lower if doped withfluorine [3]. Examples of the dynamics of hydroxyl formation when rapidly heatinghydrogen-loaded fibers at elevated temperatures is shown in figure 2.3 and 2.4. Thedata are recorded by monitoring the absorption change of the first overtone of thehydroxyl band using a standard telecommunications fiber (figure 2.3) and pure-silicacore fiber (figure 2.4). In both situations, the hydrogen concentration wasapproximately 0.75 mol% H2. As can be seen, the maximum concentration ofhydroxyl species is much higher in the germanium-doped fiber. The reason for thismay be the weaker Ge-O bond or a higher concentration of defects. The decay of thehydroxyl absorption which follows continued heating is a result of diffusion ofhydroxyl species or possibly molecular water out from the core (see e.g. Paper VI)
O O T
Ge
O O T
+ H2
O
Ge
O (GLPC)
2 HO T
9
Figure 2.3 Dynamics of optical loss due to hydroxyl formation at differenttemperatures in standard telecommunications fiber [11].
Figure 2.4 Dynamics of optical loss due to hydroxyl formation at 1000 oC in puresilica core fiber [Thesis paper VI].
10
In addition to hydroxyl formation, hydride formation may also take place (see e.g.[12]). The proposed model for formation of hydride species is shown in figure 2.5.
Figure 2.5 Hydroxyl and hydride formation in silica glass [12].
2.4.2. UV induced reactions
UV exposure of hydrogen loaded fibers generally results in the same hydrogen relatedspecies as those formed in thermally induced reaction. A proposed model of the UVinduced hydroxyl groups in germanium doped silica is that the Ge-O bond is exitedthrough absorption of a photon. A nearby hydrogen molecule can then react with theexcited bond to form a hydroxyl group, germanium-defect (GeE´center) and atomichydrogen [13,14]. The two steps in the process are shown in figure 2.6. Hydridespecies have also been observed after UV exposure, although with much lowerconcentrations compared to that of hydroxyl. The suggested mechanism for hydrideformation is an additional step where hydrogen reacts with the GeE´center [14, 15].This second process could be thermally driven as GeE´centers reacts with hydrogenabove room temperature [16].
Figure 2.6 Proposed model for UV induced hydroxyl groups and Ge-related defects[14,15,16].
2.4.3. Diffusion of water and hydroxyl groups in silica
Molecular water diffuses through interstitial diffusion and reacts with the silica latticeto form hydroxyl groups (OH) [17,18]. The diffusion mechanism is depicted in figure2.7. Due to the reaction-diffusion mechanism, measured diffusion coefficients for OHdiffusion is the effective diffusion constant for OH and water. The equilibrium ofmolecular water and hydroxyl groups depends on both concentration and temperature.
When molecular water diffuses into silica and reacts with the glass network, severaldifferent processes occur. It has been shown that water accelerates structuralrelaxation and efficiently annihilates oxygen vacancies [19]. The later process isschematically shown in figure 2.8.
Ge O Si + Ge . . . O Sihν
Ge . . . O Si + H2
Ge H O Si +H
Si - O - Si + H 2
Si - OH + H - Si
11
Figure 2.7 Schematic mechanism of water diffusion in silica glass [18,19].
Figure 2.8 Annihilation of oxygen vacancies from diffusion of molecular water [20].
2.4.4. Reduction of hydroxyl species in fluorine doped silica glass
It is well known that co-doping silica fibers with fluorine reduce the hydroxylabsorption peak [21]. The reduction of hydroxyl groups (removal of the proton) insilica during manufacturing has been suggested to result from HF formation accordingto the reaction shown in figure 2.9 [22].
Figure 2.9 Schematic of the suggested mechanism for the reduction of hydroxylgroups in fluorine doped optical fibers [21].
Si - OH + F-Si Si - O - Si + HF
Si
O
Si
Si
OH
OH
Si
Si
O
Si
Si
O
Si
H O
Si
OH
OH
Si
Si
O
Si
O O O
O O O
Si - Si + H O 2
Si - OH + H - Si
12
If one considers the equilibrium between hydroxyl groups and molecular water(shown in figure 2.7), an additional and similar path resulting in the formation of HFmolecules may be possible as shown in figure 2.10. Here fluorine acts as a transportion for a proton. The process can be viewed as an ion exchange between the F ¯ andOH ¯.
Figure 2.10 Schematic of suggested mechanism for hydrogen reduction in fluorinedoped silica.
Si - F + H O Si - OH + HF2
13
2.5. References
1. Glass Science, 2nd Ed. by R. H. Doremus, John-Wiley & Sons Inc. (1994).2. S. R. Nagel, J. B. MacChesney, K. L. Walker, “An overview of the modified
chemical vapor deposition (MCVD) process and performance,” J. Quant.Electron., QE-18, 459-476 (1982).
3. J. Stone, “Interactions of hydrogen and deuterium with silica optical fibers: Areview,” J. Lightwave Technol., LT-5, 712-733 (1987).
4. M. Kuwazuru, K. Mochizuki, Y. Namihira, Y. Iwamoto, “Dopant effect ontransmission loss increase due to hydrogen permeation,” Electron. Lett., 20,115-116 (1984).
5. J.-L. Archambault, “Photorefractive gratings in optical fibres,” Doctoral thesis,Dept. Electronics and Computer Science, University of Southampton (1994).
6. J. Crank, in The Mathematics of Diffusion, chapter 5, Oxford University Press,1983.
7. S. Tanaka, M. Kyoto, M. Watanabe, H. Yokota, “Hydroxyl group formationcaused by hydrogen diffusion into optical glass fibre,” Electron. Lett., 20, 283-284 (1984).
8. J. Stone, J. M. Wiesenfeld, D. Marcuse, C. A. Burrus, S. Yang, “Formation ofhydroxyl due to reaction of hydrogen with silica optical fiber preforms,” Appl.Phys. Lett., 17, 328-330 (1985).
9. J. M. Wiesenfeld, J. Stone, D. Marcuse, C. A. Burrus, S. Yang, “Temperaturedependence of hydroxyl formation in the reaction of hydrogen with silica glass,”J. Appl. Phys., 61, 5447-5454 (1987).
10. K. Awazu, H. Kawazoe, M. Yamane, “Simultaneous generation of opticalabsorption bands at 5.14 and 0.452 eV in 9SiO2:GeO2 glasses heated under an H2atmosphere,” J. Appl. Phys, 68, 2713-2718 (1990).
11. M. Fokine, unpublished background work relating to Paper I and VI.12. J. E. Shelby, “Reaction of hydrogen with hydroxyl-free vitreous silica,” J. Appl.
Phys. 51, 2589-2593 (1980).13. D. L. Williams, B. J. Ainslie, R. Kashyap, G. D. Maxwell, J. R. Armitage, R. J.
Campbell, R. Wyatt, in SPIE proc., 2044, 55 (1993).14. V. Grubsky, D. S. Starodubov, J. Feinberg, “Photochemical reaction of hydrogen
with germanosilicate glass initiated by 3.4-5.4-eV ultraviolet light,” Opt. Lett. 24,729-731 (1999).
15. P. Cordier, C. Dalle, C. Depecker, P. Bernage, M. Douay, P. Niay, J.-F. Bayon, L.Dong, “UV-induced reactions of H2 with germanosilicate and aluminosilicateglasses,” J. Non-Cryst. Solids, 224, 277-282 (1998).
16. T.-E. Tsai, G. M. Williams, E. J. Friebele, "Index structure of fiber Bragg gratingsin Ge-SiO2 fibers," Opt. Lett., 22, 224-226 (1997).
17. R. H. Doremus, “The diffusion of water in fused silica”, in Reactivity of Solids,Edited by J. W. Mitchell, R. C DeVries, R. W. Roberts, and P. Cannon, Wiley,New York, 667-673 (1969).
18. M. Tomozawa, “Concentration dependence of the diffusion coefficient of water inSiO2 glass,” J. Am Chem. Soc. 68, C251-252 (1985).
19. R. H. Doremus, “Diffusion of water in silica glass,” J. Mater. Res., 10, 2379-2389(1995).
20. M. Tomozawa, H. Li, K. M. Davis, “Water diffusion, oxygen vacancyannihilation and structural relaxation in silica glasses,” J. Non-Cryst. Solids, 179,162-169 (1994).
14
21. P. Bachmann, P. Geittner, D. Leers, M. Lennartz, H. Wilson, “Low OH excessloss PCVD fibres prepared by fluorine doping,” Electron. Lett., 20, 35-36 (1984).
22. J. Kirchhof, S. Unger, H-J. Pißler, and B. Knappe, ”Hydrogen-induced hydroxylprofiles in doped silica layers”, OFC’95 Vol. 8, OSA Tech. Dig. Series, paperWP9 (1995).
23. D. K. Lam, B. K. Garside, “Characterization of single-mode optical fiber filters,”Appl. Opt., 20, 440-445 (1981).
24. Russell, Archambault, Reekie, “Fibre gratings,” Phys. World, pp. 41-46, Oct.(1993).
15
3. Fiber Bragg gratings and photosensitivity
Bragg gratings are refractive index structures manufactured by exposing the core ofan optical fiber to intense periodic ultraviolet radiation. The ability to change therefractive index with radiation is referred to as photosensitivity. This chapter is anintroduction to fiber Bragg gratings and photosensitivity and is intended to give thereader a general understanding of different areas in the field. The chapter deals withdefinitions of Bragg gratings and their applications, how they are made and a shortsummary of the different theories of the mechanisms responsible for the change inrefractive index. Included is a literature review of material considerations forphotosensitive fibers.
3.1. Fiber Bragg gratings
The simplest fiber Bragg grating structure in optical fibers is an axial and periodicchange of the refractive index of the core, as shown schematically in figure 3.1, with arefractive index modulation given by:
��
���
�+=Λ
z2 cos∆n nn(z) avgπ , Eq. (3-1)
where navg is the average refractive index of the structure, ∆n is the modulationamplitude, z is the axial position and Λ is the geometrical grating period.
Figure 3.1 Schematic of refractive index modulation and effective refractive index ofthe grating structure.
The reflected wavelength of such a structure is given by Λ=λ avgB n2 , where navg isthe average, or effective index of core along the grating. The spectral reflectivity ofthe grating, solved using coupled-mode theory, is given by:
( ) ��
���
� ∆= ηλ
πλ nLLR 2tanh, , Eq. (3-2)
Axial position
n avg -
ave
rage
refra
ctiv
e in
dex
Axial position
n - R
efra
ctiv
e in
dex
16
where L is the grating length, λ is the reflected wavelength and η is the overlapbetween the fundamental mode and core [1]. Typical values of η are η = 0.8-0.9. Thespectral full-width half-maximum of the grating is given by:
22
0
12
��
���
�+���
����
� ∆=∆Nn
nsBλλ , Eq. (3-3)
where N is the number of fringes (typically N~20 000 for a 10 mm long grating) andthe pre-factor s is s~0.5 for weak gratings and s~1 for strong gratings [2]. A simulatedreflection spectrum of a 10 mm long grating with a uniform index modulation of∆n=1·10-4 is shown in figure 3.2. By controlling the refractive index modulation andaverage refractive index along the grating the spectral properties of the grating can betailored to give desired properties (see e.g. [3]).
Figure 3.2 Simulation of the reflection spectra of a 10 mm long uniform grating as afunction of wavelength (∆n=1·10-4).
3.2. History
In 1978 Hill and coworkers reported the first observation of photosensitivity in opticalfibers when exposing a germanium doped silica core fiber with coherent counterpropagating light from an argon-ion laser at 488 nm wavelength [4]. The result was aperiodic change in the refractive index corresponding to the period of the interferencepattern generated by the two beams. As the light reflected from the gratings is thesame wavelength as that used to write the grating, this technique is limited toapplications using wavelengths at or near the writing wavelength. It was not until1989 when Meltz et al [5] presented the transverse holographic method using awriting wavelength of 244 nm that it was possible to write gratings at wavelengthsother than the writing wavelength. Previous studies of the growth dynamics of the
17
grating strength, when using counter propagating waves, showed that thephotosensitivity using 488 nm wavelength was a two-photon process [1]. The writingwavelength of 244 nm used by Meltz et. al. coincided with a germanium oxygen-vacancy defect band and the second harmonic of the wavelength used by Hill. Sincethe discovery of photosensitivity in optical fiber by Hill and the developments of thetransverse holographic writing method by Melts, thousands of articles have beenpublished concerning photosensitivity and Bragg gratings. Articles include differentgrating writing schemes and grating structures, the underlying mechanisms ofphotosensitivity, defect absorption and luminescence and numerous applications offiber Bragg gratings (see e.g. [3, 6] ).
3.3. Classification of fiber Bragg gratings
Several different types of fiber Bragg gratings have been reported. The followingsections briefly describe two different classification schemes based on the couplingcharacteristics and the growth behavior of the grating during manufacturing [3,6].
3.3.1. Classification by coupling characteristics
There are three different types of basic gratings depending on the coupling function.The Bragg gratings described previously are usually referred to as short periodgratings. The grating period is typically 0.25-0.5 µm with the light coupled into thebackward propagating direction, reflection. By tilting the fringes of short periodgratings, it is possible to couple light out from the core into backward propagatingradiation modes. These loss gratings are usually referred to as slanted or tiltedgratings. Such gratings have been used for gain equalization in erbium-doped fiberamplifiers [7].
A third type of gratings is referred to as long-period gratings [8]. These gratings havea period that is typically 100-500 µm and the light is coupled into forwardpropagating cladding modes. Acting as loss filters, these gratings are typically usedfor gain equalization. Due to the long period of the grating, they can be successfullymanufactured using point-by-point writing with either UV exposure or heat. For localheating of the fiber, a CO2-laser or an electric-arc discharge can bee used [9].
3.3.2. Classification by growth characteristics
There is also a classification scheme used depending on the growth behavior of thegrating during inscription. This scheme is mainly used to describe short-periodgratings. Prior the discovery of photosensitivity in fibers, surface relief gratings wasused for some applications. Here a surface corrugation/modification of cladding nearthe core results in interacting with the evanescent field causing strong reflection.These gratings were typically manufactured through etching or polishing [10].
Type 0 gratings or Hill gratings, are the self-organized gratings, discovered by Hill et.al. [4], formed by launching light in the fiber from which partial reflection at the
18
cleaved fiber end-face creates the periodic interference pattern. As the grating isformed more light will be reflected within the fiber and hence increase the growth rateof the grating. These gratings have limited use, as the writing wavelength is also theBragg wavelength of the grating.
Type I gratings refers to the most common gratings characterized by a monotonousgrowth.
Type II gratings are high power single-pulse “damage” gratings characterized bylarge losses on the short wavelength side of the Bragg wavelength [11]. The damageis often localized at the core-cladding interface.
Type IIa gratings are characterized by the fact that the reflection initially grows as fortype I gratings then decreases followed by a subsequent growth. Also referred to asnegative index gratings, these gratings probably contain two components; one positiveindex grating (type I) and one negative index grating [12].
Chemical composition gratings (CCGs) (thesis papers I, II and II) do not clearly fitinto any of the grating types defined above. The optical properties of the final gratingdo not differ from type I and type IIa, however the manufacturing procedure, growthof refractive index and thermal properties differ significantly. The refractive indexmodulation is ascribed to a periodic variation of one or several dopants in the core.The decay mechanism of chemical composition gratings requires diffusion of themodulated dopants and therefore these gratings show exceptional thermal stability(see chapter 5).
3.4. Mechanisms of photosensitivity
This section is intended as a short summary of the main models to clarify differentsuggested mechanisms for photosensitivity. For further reading, see e.g. [3,6,13] andreferences within.
3.4.1. Color-center model
Photosensitivity in germanium doped fibers was early on associated to an absorptionband peaking at ~240 nm [5] which was attributed to germanium-oxygen deficiency[15]. Irradiating the core, using near 240 nm wavelength, results in bleaching of the240 nm band and growth of absorption bands located near 190 nm. An example ofchanges in the absorption of germanium doped silica after 248 nm UV irradiation isshown in figure 3.3 [16]. Using the Kramers-Kronig relation, Hand et al [17] and laterDong et al [18] linked these absorption changes in the UV region to the change inrefractive index. A number of different defects have been observed of which the mostcommonly discussed are schematically shown in figure 3.4. Associated absorptionbands for some of the main defects are given in table 3.1. Generally oxygen deficientcenters (ODC) are precursors and GeE´, Ge(1) and Ge(2) are products of thephotochemical processes although there are numerous different photochemicalpathways, which have been presented to explain transformation of defects and theglass structure.
19
Figure 3.3 Change in UV absorption spectra after 248 nm exposure(from [16]).
SiE´
Si Si
GeE´
Ge Si
NBOHC
Si O
GODC-wrong bond
Ge Si
Peroxy-radical
Si OO
GODC - 2 coordinated Ge
Si O Ge O Si
Ge(1)
Ge
Si Si
Si Si
O O
OO
Ge(2)
Ge
Si Si
Ge Si
O O
OO
DID
Ge Si
= Trapped electron
Figure 3.4 Schematic structure of some defects related to photosensitivity (GODC –Germanium oxygen deficient center, NBOHC – Non bonding oxygen hole center, DID
– Drawing induced defect). See e.g. [6,19,20].
20
Defect absorption peak Ref.GODC – wrong bond 240 nm/ 5 eV [6,21]
GeE´ 195 nm / 6.4 eV [6,15]Ge(1) 281 nm/ 4.4 eV [6,15]Ge(2) 213 nm/ 5.8 eV [6,15]
Table 3.1 Associated absorption bands for the most common defects related tophotosensitivity.
3.4.2. Stress relaxation model
As the refractive index of silica glass changes due to the stress-optic effect, relaxationof stress will consequently result in a change in refractive index [19]. Due to materialproperties and manufacturing procedures, optical fibers may have highly stressedregions. The residual stress arise from difference in the thermal expansion betweencore and cladding regions (thermoelastic stress) and due to difference in transitiontemperature (Tg) in combination with the applied tension during fiber drawing(mechanical stress).
In a fiber with a core having a higher thermal expansion coefficient than to thecladding (αT-core > αT-clad), the contraction of the core as the fiber cools down will berestricted by the cladding glass. The situation will be the opposite if the core has alower thermal expansion than to the cladding. As the stress integrated over the fiberwill be zero, the residual stress in the different regions depends on the ratio of theirarea. The mechanical stress is a result of the tension used to extract the fiber from thedrawing furnace. The drawing tension will essentially be applied to the region thatsolidifies first while the remaining glass with the lower transition temperature (Tg)will solidify once the temperature has decreased sufficiently. When the drawingtension is released, the fiber will contract resulting in a compressive stress in theregions with a lower Tg. For a more complete treatment of residual stress in opticalfibers, see e.g. [22]. Estimated refractive index change due to stress relaxation inhighly stressed fiber is in the order of ∆n~10-3 [23]. There is however, some debate onwhether UV induced stress relaxation is a mechanism involved in photosensitivity ornot (se following section on densification). For fabrication of long-period gratingsusing point-by-point heat exposure, the mechanism of stress relaxation is moreobvious. LPG’s have successfully been fabricated in pure silica core fibers [24, 25]and boron-germanium doped fibers [26,27] by thermal relaxation of drawing inducedstress using a CO2-laser or arc-discharge.
3.4.3. Densification-compaction model
Exposing amorphous silica to ultraviolet irradiation is known to result incompaction/density changes of the glass matrix [28]. In a study of silica glasses forlithography applications it was determined that UV induced densification wasproportional to the softening temperature for a series of different silica based glasses[29]. UV induced densification in germanium doped silica preforms and fibers have
21
been observed e.g. using atomic force microscope [30], transmission electronmicroscope [31] and indirectly through changes in the Raman spectra [32] andchanges in core tension [33]. For the latter, an increase in refractive index wascorrelated to an increase in tensile stress in the core, which could be linked to acompaction of the core. For a wider discussion on photosensitivity and densification,see e.g. ref. [34].
3.5. Increasing photosensitivity
The following section is a short description of the main methods used for increasingthe photosensitive response of optical fibers. The methods include hydrogentreatment, thermal treatment, mechanical treatment, and preform manufacturing.
3.5.1. Hydrogen treatment
High temperature hydrogen treatment of preform
The absorption at 5 eV, corresponding to oxygen-deficient defect absorption, wasshown to increase by heating a highly germanium doped preform in hydrogenatmosphere [35]. The 1.2 cm long preform was heated in a hydrogen atmosphere at610 oC for 75 hrs. Fibers drawn from this preform showed significantly higherphotosensitivity compared to untreated preform.
Low-temperature hydrogen loading
When placing a fiber in a high-pressure hydrogen atmosphere at room temperature,hydrogen molecules will inertly diffuse into the glass network. Loading the fiber withhydrogen prior to UV exposure significantly increases the photosensitivity [36]Refractive index changes as high as ∆n~6·10-3 were reported for standard hydrogensensitized telecommunications fiber. The increase in photosensitivity was suggestedto come from thermally induced reactions, resulting in Si-OH and bleachable GODCsand photolytically driven reactions resulting in Si-OH and germanium related defects.Using hydrogen loading removes the requirement of using UV wavelengthscoinciding with defect absorption bands for accessing photosensitivity [37]. Low-temperature high-pressure hydrogen-loading or simply hydrogen loading, is the mostcommon method used for increasing the photosensitivity.
Flame brushing
Flame brushing involves localized heating of fibers and waveguides using a hydrogenrich flame [38]. The technique results in a large increase in the UV absorption spectraincluding the 240 nm absorption band. The increased photosensitivity is likely due toin-diffusion of hydrogen and hydroxyl species into the core resulting in increased
22
defect absorption. The drawback of this method is the mechanical degradation of thefiber after prolonged processing. For example, for an estimated temperature of ~1700oC, the process took roughly 20 minutes for maximum efficiency when sensitizing astandard telecommunications fiber.
Writing at elevated temperatures
Heating hydrogen loaded fiber during grating fabrication results in enhancedphotosensitivity compared to hydrogen loading only. The photosensitivity ofhydrogen loaded GeO2 and P2O5 doped fibers increased significantly when heated at250-400 oC [39]. The moderate heating in combination with UV exposure duringwriting results in a dramatic increase in hydroxyl formation whereas moderate heatingalone does not.
UV pre-exposure followed by hydrogen out-diffusion
Using fringeless UV exposure of hydrogen loaded fibers, a permanent andcontrollable increase in photosensitivity can be obtained even after the remaininghydrogen has been removed [40,41]. This result can be partially understoodconsidering the previous sections where heat or UV exposure results in both hydroxylformation and defect generation. By controlling the exposure conditions, axialvariations of the photosensitive response can be produced. This technique has alsoshown to render phosphorous doped fibers photosensitive. However, in the case ofUV pre-exposure of phosphorous doping the mechanism for enhancedphotosensitivity is suggested to be different. For phosphorous doped fibers a similartechnique has been used where the UV exposure is replaced by thermal treatment at~80 oC. [42]
OH flooding
A large increase in photosensitivity can also be obtained by significantly increasingthe hydroxyl concentration in the core of the fiber. The process, referred to as OHflooding (Paper VI), is achieved by rapid (<1 sec) heat treatment at 1000 oC ofhydrogen loaded fibers. The technique can be compared to high-temperaturehydrogen-treatment of the preform and flame brushing. The advantage of OHflooding is that hydrogen diffusion takes place at low temperatures while the high-temperature treatment is minimal. The result is that very high concentrations ofhydroxyl groups can be formed in the core while effects in the cladding are minimal.
23
3.5.2. Thermal treatment
CO2 laser treatment
Exposure of germanium doped silica films [43] and fibers [44] to CO2 laserirradiation has shown to result in an increase of the 240 nm absorption band. Inaddition to an increase in photosensitivity, the refractive index also increases.
3.5.3. Mechanical treatment
Applied strain on photosensitivity
Applying strain (∆L/L = 0- 6.7·10-3) during grating writing to highly germaniumdoped fibers (11.5 and 28 mol% GeO2) the reflectivity of the type I grating spectradecreased while the onset, rate and amplitude of the type IIa grating increased [45].Similar observations have been made showing an approximately linear decrease intype I grating saturation index with applied strain [46].
On the contrary, by applying large strains, 3 to 3.3 %, on different fibers (AT&Tstandard fiber, B2O3-GeO2 fiber (Fibercore) and high GeO2 containing fiber(Fibercore)) during UV irradiation a significant increase in type I photosensitive wasreported [47]. Measurements of the residual stress in the core after UV-exposureshowed opposite sign for fibers with and without applied strain. In the unstrained fiberthe core, which was initially in compression, relaxed during UV exposure. For thestrained fiber the core had a higher degree of compression, measured after releasingthe strain, compared to unexposed regions.
3.5.4. Preform manufacturing
Reduced atmosphere
With photosensitivity closely associated with GODCs, attempts to increase thesedefects during preform manufacturing have been made. The formation of GODC isgoverned by the equilibrium GeO2 ↔ GeO + ½O2, and using a reducing atmosphereduring preform manufacturing a large increase in defect absorption can be achieved.[48]. This technique is commonly used when manufacturing photosensitive fiberstoday.
3.6. Literature review of material considerations
The following section gives a short summary of some different dopants and theirreported effects on photosensitivity. Considering the numerous articles published onthe topic, the dopants discussed are limited and the list is in no way complete. Forinstance, rare earth doped glass and multi-component glasses are not included here.
24
B – Boron
Co-doping silica fibers with B2O3 and GeO2 results in highly photosensitive fibers[49]. Addition of B2O3 to GeO2 doped silica does not alter the 240 nm UV absorptionband characteristic for germanium doped fibers. However, for B2O3 concentrationsmore than 10 mol% the 240 nm absorption band is reduced (17-18 mol% GeO2). InB2O3-SiO2 glass, no absorption band is present at 240 nm and the UV absorptionstarts to increase at ~190 nm. [48] The enhancing mechanism of B2O3-codoping hasbeen attributed to densification enhancement due to stress effects [50], to soften theglass making more plausible a structural rearrangement of defects and consequentlyof the glass network. The germanium-oxygen deficient centers (GODC) are proposedto be transformed into drawing induced defects (DID’s) [51].
The drawback of B2O3-codoping is a lower thermal stability compared to GeO2doping alone (see e.g. [52,53]). Large changes in refractive index due to fiber drawingconditions also impose a problem on the thermal stability of these gratings, especiallyconcerning wavelength stability. In ref [54] the wavelength stability of gratings inB2O3-GeO2 doped fibers was studied and found to be dependant on two components.One component was the thermal decay of the UV induced refractive index modulationcausing a negative wavelength shift, while the second component was due to arelaxation of the B2O3-GeO2 silica glass causing a positive wavelength shift (∆n>0).The second component appears to be active at temperatures as low as 100 oC.
Enhanced photosensitivity in B2O3-GeO2 silica fibers with increased fiber drawingtension has been reported [55] and thermal release of drawing induced stress has beenused for fabricating long period gratings (LPG’s) [56].
N – Nitrogen
In 1997 Dianov et. al. [57] presented highly photosensitive nitrogen dopedgermanosilicate fibers (7 mol% Ge – 0.1 at% N silica ). Refractive index changes of∆n=2.8·10-3 were achieved for an irradiation dose of 75 kJ/cm2 at 244 nm wavelengthusing fibers without hydrogen loading. When hydrogen loaded, the result was∆n=1·10-2. A large amount of hydrogen in N-doped fibers could pose a problem, asthe N-H bond has an absorption in the third telecommunications window at awavelength of λ~1506 nm (1st overtone) [58,59]. Grating formation has also beenreported in Ge free N doped fibers using 193 nm wavelength (∆n=8.4·10-4, no H2-loading) [60]. The grating formation dynamics follows type IIa behavior, and showhigh thermal stability (>1000 oC). Gratings grow according to type I grating behaviorwhen hydrogen loaded, but show no improvement in photosensitivity (∆n=3.5·10-4 forirradiation dose of 15 kJ/cm2).
“Ultra thermally stable” LPG’s have been formed in N doped silica fibers by CO2laser or arc discharge [61]. The thermal stability of these gratings exceed 1000 oC andare believed to be of the type chemical composition gratings. The mechanism behindthe thermo-diffusion was believed to be the weaker bond of Si-N compared to Si-Oresulting in decomposition of silicon oxynitride glass at a temperature of about 1800oC.
25
F – Fluorine
Fluorine in not generally used as a codopant for enhancing photosensitivity possiblybecause of the decrease of the GODC absorption band at 240 nm [48]. Fluorine hasbeen used as a refractive index decreasing dopant to manufacture fibers withphotosensitive core and cladding [62]. For SiO2 glass, Hosono et al [63] show asignificant decrease in UV absorption (blue shift) and in concentration of SiE´ centerswhen doping with up to ~1 mol% fluorine. The explanation of the shift in UVabsorption (shift in bandgap) is attributed to a reduction of strained Si-O-Si bonds bya reduction of 4- and 3-membered rings, suggesting a more open structure for fluorinedoped silica. Raman spectroscopy showed that fluorine doping resulted in a decreasein the D1 (495 cm-1) and D2 (606 cm-1) band corresponding to the 4- and 3- memberedring structures. It is also believed that the strong Si-F bond results in increasedresistance to ionizing radiation (see e.g. [64,65]).
Al – Aluminum
Absorption measurements in MCVD preforms show that Al2O3 does not cause anysignificant absorption at 240 nm. However, an increase in the absorption atwavelengths lower than 220 nm is evident. Addition of Al2O3 to GeO2 doped silicareduces the 240 nm absorption band and results in an absorption band centered at 205nm [48].In ref [66], a cladding-mode-suppressing fiber is designed with a constantphotosensitivity across core and inner cladding. Here Al2O3 is used in the core as anindex modifying dopant with negligible effect on photosensitivity while a constantconcentration of B2O3-GeO2 in core and cladding provides a homogeneousphotosensitivity with a cladding refractive index matched to silica.
Photosensitivity of Al2O3-SiO2 fibers is often reported in combination withlanthanide’s such as Ce, Tb, Er, Tm, Yb (see e.g. [13]).
SiO2 – Silica
Formation of long period gratings in pure silica core fibers has been reported [67].Here periodic residual stress relaxation was achieved using an arc discharge resultingin an increase in refractive index. With a pure silica core and fluorine depressedcladding, an increase in drawing tension results in a tensile stress of the core andtherefore a reduction in refractive index.
Densification and generation of SiE´ center in pure SiO2 have been reported [68]yielding index changes in the order of ∆n~5·10-5. Increased photosensitivity using 193nm wavelength in high-purity fused silica after Si2+ ion implantation has also beenreported [69] and correlated through absorption changes with the Kramers-Kronigrelation. In [70] Kannan et. al. identify a distribution of non-bonding oxygen-holecenters (NBOHC’s) to be located in the core-cladding region in dry silica core fiberswith F-depressed cladding.
26
There is currently only one report of strong grating formation (∆n~5·10-4) in puresilica core fibers, using a writing wavelength of 193 nm (Paper VII). It is believed thatdrawing induced stress may play an important role in the large index change in thiscase.
P – Phosphorous
Phosphorous doping does not result in any pronounced absorption features at 240 nm,but produces a weak band centered at 210 nm. The absorption peak at 240 nm ingermanium doped silica co-doped with phosphorous is significantly reduced [48].Photosensitivity in phosphorous doped silica has been observed in hydrogen ordeuterium loaded fibers and waveguides after long exposure times [71]. It has alsobeen shown that the photosensitivity in phosphorous doped fibers increases withtemperature (∆n=3·10-4 for room temp, ∆n=4.5·10-4 for 250 oC, ∆n=7·10-4 for 400 oCusing 248 nm) [39]. Transient gratings have been observed in phosphorous dopedfibers and permanent gratings only when hydrogen loaded (193 nm irradiation) [72].Increase in photosensitivity is achieved by different hydrogen treatment techniquesincluding UV pre-exposure followed by hydrogen out-diffusion [73] and thermaltreatment (~80 oC) of hydrogen loaded fibers followed by hydrogen out-diffusion[74].
LPG’s have been also been fabricated in phosphorous doped silica fibers bythermoactivated diffusion using a CO2 laser [75].
S – Sulfur
LPG formation in 0.4 wt% S – 2 wt% Cl doped silica using electric arc technique, hasbeen achieved, however no grating formation using 248 nm KrF laser exposure couldbe detected [76]. The “electric arc photosensitivity” was increased by annealing thefibers at ~1000 oC and the suggested mechanism for refractive index change was achange in fictive temperature resulting from high cooling rate.
Cl – Chlorine
Chlorine, which has an index raising effect on silica, is usually incorporated in lowconcentrations during fiber manufacturing. For MCVD fiber, the source is thechlorine containing dopant precursors while for OVD and VAD fiber chlorine isincorporated during the drying process (heated in Cl2 gas). Chlorine is incorporated asSi-Cl and results in a decrease of the D2 Raman band at 606 cm-1, associated withthree-membered ring structures [77]. However, increasing the concentration above 2mol% results in an increase in the D2 band. Defect absorption studies of chlorinedoped silica show an absorption band at 3.8 eV, corresponding to Cl2 molecules, andpossibly, an absorption tail around 6-7 eV. [78] Chlorine is also found to increasestructural relaxation rates of silica, which in turn may influence the refractive indexthrough lower fictive temperature [79].
27
A radial, periodic structure of alternating concentration of Ge and Cl was found inpreforms manufactured by low-pressure plasma chemical vapor deposition [80]. Thevariation is result of different reaction kinetics of the dopant precursors. By thermaltreatment of such preforms, a significant increase in 240 nm absorption was achieved.The increase in absorption was attributed to diffusion of germanium and liberation ofchlorine.
Ti – Titanium
Fibers with titanium-doped (TiO2) outer cladding are often used for increasedmechanical strength. In ref. [81], an enhanced growth rate of Bragg gratings wasobserved in fibers with TiO2-doped outer cladding, and hydrogen-loading times werereduced to 1 day at room temperature. This should be compared to >5 days for thesame fiber but without the TiO2 outer cladding. The explanation for the enhancementwas ascribed to temperature effects. The TiO2 layer absorbs 240 nm light quiteeffectively causing heating of the fiber (200 oC at 40 mW and 244 nm). The increasein temperature increases the diffusion rate of hydrogen resulting in a higherconcentration of hydrogen in the core. In addition, the increase in temperature leads toan increased photosensitivity.
Ge – Germanium
The first gratings were written in GeO2 doped silica fibers, and GeO2 appears still tobe the most prominent dopant for highly photosensitive fibers. Under similarmanufacturing conditions, the photosensitivity grows linearly with GeO2concentration. The photosensitivity in GeO2 doped fibers and preforms has long beenassociated with defect absorption near 240 nm and using reducing atmosphere duringmanufacturing significantly increases the absorption [48]. Codoping with B2O3 orSnO2 significantly increases the photosensitivity (see respective dopant).
Sn – Tin
Photosensitivity concerning Sn-doping has been reported in combination with GeO2-,P2O5-, Na2O doping or in binary silicate. The first report of photosensitive fibers withSnO2 doping was in combination with GeO2
[82]. The main advantages pointed outare lower losses at 1.5 µm and higher thermal stability compared to B2O3-GeO2 co-doped fibers. No absorption features in the 400 nm to 2 µm range were observed. The240 nm absorption band was slightly shifted towards 250 nm and increased after 248nm irradiation contrary to the behavior of GeO2-doped silica. A refractive indexchange of ∆n~1.4·10-3 was achieved for a total fluence of 0.3 kJ/cm2. In ref. [55]increased tension during fiber drawing of SnO2-GeO2 doped silica fiber is shown toreduced the photosensitivity and gratings in this fiber grow according to a type IIabehavior.
In [83], highly photosensitive SnO2-P2O5 doped SiO2 was reported for both type I andtype II gratings. Refractive index change of ∆n~1.2·10-3 (248 nm, 0.4 J/cm2 per pulse,
28
total fluence 0.3 kJ/cm2) was observed and the damage threshold (onset for type IIgrating formation) was 0.5 J/cm2. The MCVD fiber (0.6 mol% SnO2-9 mol% P2O5)had an intrinsic loss of 40 dB/km at 1.5 µm and no absorption peak at 240 nm. Thefibers showed a large transient loss measured at 633 m and large refractive indexdecay immediately after UV exposure.
Ref. [84], reports on refractive index modulations of ∆n~2.8·10-4 (3.7 hr exposure,total fluence 21.3 kJ/cm2) written in 0.15 mol% SnO2 doped silica fiber using 248 nmexposure. Difficulties in preform fabrication are volatile SnO2 burning off duringpreform collapse and crystallization at levels of ~1mol% SnO2. Measurements of theUV induced absorption contribution to refractive index (using Kramers-Kronigrelation) show a negative index contribution.
To increase the concentration of SnO2 while avoiding crystallization, Brambilla et al[85] used 75 SiO2: 5 SnO2: 20 Na2O (in mol%) glass made by melting glass powdersand collapsing the rod with the cladding tube in the drawing tower. The refractiveindex of the core was n~1.52 and a UV induced refractive index modulation of∆n~6.2·10-4 (248 nm, 140 min exposure, total fluence 35.2 kJ/cm2) was achieved.
VUV absorption studies in SnO2-doped silica resulted in the estimation of a refractiveindex change of ∆n ~ -3·10-4 (Kramers-Kronig model), while the measured refractiveindex was positive [86]. Optical absorption and EPR studies of SnO2 doped silicashow that UV absorption is not the primary mechanism for the refractive indexchange. [87]. Structural compaction resulting from reduction in ring size is suggested.Densification is suggested to be related to the frozen stress reservoir in the core as anincrease in SnO2 content increases the strain on the Si-O bonds resulting in a weakerbond [88]. Compaction (using 248 nm laser) in the core of a 1 mol% SnO2-20 mol%GeO2 fiber is also verified using transmission electron microscopy (TEM) [31]. Phaseseparation has also been reported for highly SnO2-doped GeO2 or P2O5 preform slicesand fibers [89].
Sb – Antimony
Photosensitive Sb-doped fibers have been reported addressing three problems withphotosensitive fibers: splice loss, thermal stability and ability to write through thecoating [90]. A 257 nm CW laser was used to write through a polymer coating withrapid growth necessary due to photo-darkening of the coating. Sb-doping causesstrong absorption below 250 nm and in deuterium loaded fibers an index change of∆n~4·10-4 was achieved. The photosensitivity was comparable to that of a germaniumdoped fiber fabricated under reduced atmosphere. Prolonged exposure results in alarge increase of the UV absorption tail into the visible. Some improvement in thethermal stability below 200 oC was observed.
Ta – Tantalum
Dong et al [91] reports on photosensitivity in 1.5 mol% tantalum-doped silica, wherepulsed UV at 248 nm wavelength was used with resulting refractive index changes of
29
the order of ∆n~2·10-6. The grating reflectivity saturates after only a few pulses andcauses relatively large losses in the visible and the NIR.
Pb – Lead
Photosensitivity in PbO-doped silica is mainly reported for multicomponent glasses.UV induced surface-relief gratings with refractive index changes as high as ∆n~0.09has been reported [92]. Color center formation has been reported to occur in leadglass fibers [93].
30
3.7. References
1. D. K. Lam, B. K. Garside, “Characterization of single-mode optical fiber filters,”Appl. Opt., 20, 440-445 (1981).
2. P. St. J. Russell, J. L. Archambault, L. Reekie, "Fibre gratings," Physics World,pp. 41-46 October (1993).
3. Raman Kashyap, Fiber Bragg Gratings, Academic Press (1999).4. K. O. Hill, Y. Fujii, D. C. Johnson, B. S. Kawasaki, “Photosensitivity in optical
fiber waveguides: Application to reflection filter fabrication,” Appl. Phys. Lett.,32, 647-649 (1978).
5. G. Meltz, W. W. Morey, W. H. Glenn, “Formation of Bragg gratings in opticalfibers by transverse holographic method,” Opt. Lett., 14, 823-825 (1989).
6. A. Othonos and K. Kalli, Fiber Bragg Gratings: Fundamentals and Applicationsin Telecommunications and Sensing, Artech House, (1999).
7. I. Riant, “UV-photoinduced fibre gratings for gain equalisation,” Opt. FiberTechnol., 8, 171-194, (2002).
8. M. Vengsarkar, P. J. Lemaire, J. B. Judkins, V. Bhatia, T. Erdogan, J. E. Sipe,“Long-period fiber gratings as band-rejection filter,” J. Lightwave Technol., 14,58-65 (1996).
9. V. I. Karpov, M. V. Grekov, E. M. Dianov, K. M. Golant, S. A. Vasiliev, O. I.Medvedkov, R. R. Khrapko, "Thermo-induced long-period fibre gratings",ECOC'97, 22-25 Sept. 1997, Edinburg, published by IEE, ConferencePublications No. 448, vol.2, pp.53-56, (1997).
10. J. Rowe, I. Bennion, D. C. J. Reid, “High-reflectivity surface relief gratings insingle-mode optical fibers,” IEE Proc., Vol. 134, 197-202 (1987).
11. J.-L. Archambault, L. Reekie, P. St. J. Russel, “100% reflectivity Bragg reflectorsproduced in optical fibres by single excimer laser pulses,” Electron. Lett., 29, 453-455 (1993).
12. L. Dong, W. F. Liu, “Thermal decay of fiber Bragg gratings of positive andnegative index changes formed at 193 nm in a boron-codoped germanosilicatefiber,” Appl. Opt., 36, 8222-8226 (1997).
13. M. Douay, W. X. Xie, T. Taunay, P. Bernage, P. Niay, P. Cordier, B. Poumellec,L. Dong, J. F. Bayon, H. Poignant, E. Delevaque, “Densification involved in theUV-based photosensitivity of silica glasses and optical fibers,” J. LightwaveTechnol., 15, 1329-1342 (1997).
14. B. Poumellec and F. Kherbouche, “The Photorefractive Bragg Gratings in theFibers for Telecommunications,” J. Phys. III France, 6, 1595-1624 (1996).
15. H. Hosono, Y. Abe, D. L. Kinser, R. A. Weeks, K. Muta, H. Kawazoe, ”Natureand origin of the 5-eV band in GeO2:SiO2 glasses,” Phys. Rev. B 46, 11445-11451(1992).
16. R. M, Atkins, V. Mizrahi, T. Erdogan, “248 nm induced vacuum UV spectralchanges in optical fibre preform cores: support for a color centre model ofphotosensitivity,” Electron. Lett., 29, 385 –387 (1993).
17. P. Hand, P. St. J. Russel, “Photoinduced refractive-index changes ingermanosilicate fibers,” Opt. Lett., 15, 102-104 (1990).
18. L. Dong, J. L. Archambault, L. Reekie, P. St. J. Russel, D. N. Payne,“Photoinduced absorption change in germanosilicate preforms: evidence for thecolor-center model of photosensitivity,” Appl. Opt., 34, 3436-3440 (1995).
19. M. G. Sceats, G. R. Atkins, S. B. Poole, “Photolytic index changes in opticalfibers,” Annu. Rev. Mater. Sci., 23, 381-410 (1993).
31
20. L. Skuja, "Optically active oxygen-deficiency-related centers in amorphoussilicon dioxide," J. Non Cryst. Solids, 239, 16-48 (1998).
21. E. J. Friebele, D. L. Griscom, “Color centers in glass optical fiber waveguides,”Materials Research Society Symposium, 61, 319-331 (1986).
22. Y. Park, K. Oh, U. C. Paek, D. Y. Kim, and C. R. Kurkjian, “Residual Stresses ina Doubly Clad Fiber with Depressed Inner Cladding (DIC),” J. LightwaveTechnol., 17, pp. 1823-1834, 1999.
23. M. G. Sceats, P. A. Krug, “Photoviscous annealing – dynamics and stability ofphotorefractivity in optical fibres,” in SPIE proc., 2044, 113-120 (1993).
24. M. Akiyama, K. Shima, A. Wada, R. Yamauchi, “A novel long-period fibergrating using periodically released residual stress of pure-silica core fibers,”OFC´98, Paper ThG1 (1998).
25. T. Enomoto, M. Shigehara, S. Ishikawa, T. Danzuka, H. Kanamori, “Long-periodfiber gratings in a pure silica-core fiber written by residual stress relaxation,”OFC´98, Paper ThG2 (1998).
26. B. H. Kim, Y. Park, T. –J. Ahn, D. Y. Kim, B. H. Lee, Y. Chung, U. C. Peak, andW. –T. Han, “Residual stress relaxation in the core of optical fiber by CO2 laserirradiation,” Opt. Lett., 26, 1657-1659 (2001).
27. C.-S. Kim, Y. Han, B. H. Lee, W.-T. Han, U.-C. Paek, Y. Chung, “Induction ofthe refractive index change in B-doped optical fibers through relaxation of themechanical stress,” Opt. Comm. 185, 337-342 (2000).
28. C. Fiori, R. A. B. Devine, “Ultraviolet irradiation induced compaction andphotoetching in amorphous, thermal SiO2,” Mat. Res. Soc., 61, 187-195 (1986).
29. N. F. Borrelli, C. M. Smith, D. C. Allan, “Excimer-laser-induced densification inbinary silica glasses,” Opt. Lett., 24, 1401-1403 (1999).
30. B. Poumellec, P. Niay, M. Douay, J. F. Bayon, “The UV-induced refractive indexgrating in Ge:SiO2 preforms: additional CW experiments and the macroscopicorigin of the change in index,” J. Phys. D: Appl. Phys., 29, 1842-1853 (1996).
31. P. Cordier, S. Dupont, M. Douay, G. Martinelli, P. Bernage, P. Niay, J. F. Bayon,L. Dong, “Evidence by transmission electron microscopy of densificationassociated to Bragg grating photoimprinting in germanosilicate optical fibers,”Appl. Phys. Lett., 70, 1204-1206 (1997).
32. E. M. Dianov, V. G. Plotnichenko, V. V. Koltashev, Yu. N. Pyrkov, N. H. Ky, H.G. Limberger, R. P. Salathe, “UV-irradiation induced structural transformation ofgermanosilicate glass fiber,” Opt. Lett., 22, 1754-1756 (1997).
33. P. Y. Fonjallaz, H. G. Limberger, R. P. Salathé, “Tension increase correlated torefractive-index change in fibers containing UV-written Bragg gratings,” Opt.Lett., 20, 1346-1348 (1995).
34. M. Douay, W. X. Xie, T. Taunay, P. Bernage, P. Niay, P. Cordier, B. Poumellec,L. Dong, J. F. Bayon, H. Poignant, E. Delevaque, “Densification involved in theUV-based photosensitivity of silica glasses and optical fibers,” J. LightwaveTechnol., 15, 1329-1342 (1997).
35. G. Meltz, W. W. Morey, “Bragg grating formation and germanosilicate fiberphotosensitivity,” in International workshop on photoinduced self-organizedeffects in fiber, in SPIE proc., 1516, 185-199 (1991).
36. P. J. Lemaire, R. M. Atkins, V. Mizrahi, AW. A. Reed, “High pressure H2 loadingas a technique for achieving ultrahigh UV photosensitivity and thermal sensitivityin GeO2 doped optical fibres,” Electron. Lett., 29, 1191-1193 (1993).
32
37. V. Grubsky, D. S. Starodubov, J. Feinberg, “Photochemical reaction of hydrogenwith germanosilicate glass initiated by 3.4-5.4-eV ultraviolet light,” Opt. Lett., 24,729-731 (1999).
38. F. Bilodeau, B. Malo, J. Albert, D. C. Johnson, K .O. Hill, Y. Hibino, M. Abe, andM. Kawachi, “Photosensitization of optical fiber and silica-on-silicon/silicawaveguides,” Opt. Lett., 18, 953-955 (1993).
39. P. J. Lemaire, A. M. Vengsarkar, W. M. Reed, D. J. DiGiovanni, “Thermallyenhanced ultraviolet photosensitivity in GeO2 and P2O5 doped optical fibers,”Appl. Phys. Lett., 66, 2034-2036 (1995).
40. G. E. Konkhe, D. W. Nightingale, P. G. Wigley, “Photosensitization of opticalfibers by UV preexposure of hydrogen loaded fibers,” OFC´99, PD201-3 (1999).
41. M. Åslund, J. Canning, G. Yoffe, “Locking in photosensitivity within optical fiberand planar waveguides by ultraviolet preexposure,” Opt. Lett., 24, 1826-1828(1999).
42. J. Canning, P.- F. Hu, “Low temperature hypersensitisation of phosphosilicatefibers in hydrogen,” in proceedings of Bragg Gratings, Photosensitivity, andPoling in Glass waveguides, OSA, Stresa, Italy Paper BThA4 (2001).
43. E. M. Starr, N. Nourshargh, J. S. McCormack, “The effect of laser heating onoptical properties of germania doped silica optical waveguides,” Mat. Res. Soc.Symp. Proc., 88 (1987).
44. G. Brambilla, V. Pruneri, L. Reekie, D. N. Payne, “Enhanced photosensitivity ingermanosilicate fibers exposed to CO2 laser radiation,” Opt. Lett., 24, 1023-1025(1999).
45. T. Taunay, P. Niay, P. Bernage, M. Douay, W. X. Xie, D. Pureur, P. Cordier, J. F.Bayon, H. Poignant, E. Delevaque, and B. Poumellec, “Bragg grating inscriptionswithin strained monomode high NA germania-doped fibers: part I.Experimentation,” J. Phys. D: Appl. Phys., 30, 40-52 (1997).
46. I. Riant, F. Haller, “Study of the photosensitivity at 193 nm and comparison withphotosensitivity at 240 nm Influence of fiber tension: Type IIa aging,” J.Lightwave Technol., 15, 1464-1469 (1997).
47. E. Salik, D. S. Starodubov, and J. Feinberg, “Increase of photosensitivity in Ge-doped fibers under strain,” Opt. Lett., 25, 1147-1149 (2000).
48. L. Dong, J. Pinkstone, P. St. J. Russell, D. N. Payne, “Ultraviolet absorption inmodified chemical vapor deposition preforms, ”J. Opt. Soc. Am. B, 11, 2106-2111(1994).
49. 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, 45-47 (1993).
50. I. Camlibel, D. A. Pinnow, and F. W. Dabby, “Optical aging characteristics ofborosilicate clad fused silica core fiber optical waveguides,” Appl. Phys. Lett., 26,185-187 (1975).
51. E. M. Dianov, D. S. Starodubov, "Microscopic mechanisms of photosensitivity ingermanium- doped silica glass," in SPIE proc., 2777, 60-70 (1995).
52. D. L. Williams, and R. P. Smith, “Accelerated lifetime tests on UV written intra-core gratings in boron germania codoped silica fibre,” Electron. Lett., 31, 2120-2121 (1995).
53. S. R. Baker, H. N. Rourke, V. Baker, and D. Goodchild, “Thermal decay of fiberBragg gratings written in boron and germanium codoped silica fiber,” J.Lightwave Technol., 15, 1470-1477 (1997).
33
54. K. E. Chisholm, K. Sugden, and I. Bennion, “Effect of thermal annealing onBragg fibre gratings in boron/germania co-doped fibre,” J. Phys. D: Appl. Phys.,31, 61-64 (1998).
55. N. H. Ky, H. G. Limberger, R. P. Satathé, F. Cochet, and L. Dong, “Effects ofdrawing tension on the photosensitivity of Sn-Ge and B-Ge-codoped core fibers,”Opt. Lett., 23, 1402-1404 (1998).
56. C.-S. Kim, Y. Han, B. H. Lee, W.-T. Han, U.-C. Paek, Y. Chung, “Induction ofthe refractive index change in B-doped optical fibers through relaxation of themechanical stress,” Opt. Comm., 185, 337-342 (2000).
57. E. M. Dianov, K. M. Golant, V. M. Mashinsky, O. I. Medvedkov, I. V. Nikolin,O. D. Sazhin, and S.A. Vasiliev, “Highly photosensitive nitrogen-dopedgermanosilicate fibre for index grating writing,” Electron Lett., 33, 1334-1336(1997).
58. G. Grand, J. P. Jadot, H. Denis, S. Valette, A. Fournier, A. M. Grouillet, “Lowloss PECVD silica channel waveguides for optical communications,” Electron.Lett., 26, 2135-2137 (1990).
59. E. M. Dianov, K. M. Golant, R. R. Khrapko, A. S. Kurkov, A. L. Tomashuk,“Low-hydrogen silicon oxynitride optical fibers prepared by SPCVD,” J.Lightwave Technol., 13, 1471-1474 (1995).
60. E. M. Dianov, K. M. Golant, R. R. Khrapko, A. S. Kurkov, B. Leconte, M.Douay, P. Bernage, and P. Niay, “Grating formation in a germanium free siliconoxynitride fibre,” Electron. Lett., 33, 236-238 (1997).
61. V. I. Karpov, M. V. Grekov, E. M. Dianov, K. M. Golant, R. R. Khrapko, “Ultra-thermostable long-period gratings written in nitrogen-doped silica fibers”, Mat.Res. Soc. Symp. Proc., 531, 391-396 (1998).
62. E. Delevaque, S. Boj, J. –F. Bayon, H. Poignant, J. Le Mellot, M. Monerie, P.Niay, and P. Bernage, “Optical fiber design for strong gratings photoimprintingwith radiation mode suppression,” Opt. Fiber Commun. Conf., 1995, Paper PD 5.
63. H. Hosono, M. Mizuguchi, L. Skuja, and T. Ogawa, “Fluorine-doped SiO2 glassesfor F2 excimer laser optics: fluorine content and color-center formation,” Opt.Lett., 24, 1549-1551 (1999).
64. S. Shibata, and M. Nakahara, “Fluorine and chlorine effects on radiation-inducedloss for GeO2-doped silica optical fibers,” J. Lightwave Technol., LT-3, 860-863(1985).
65. K. Awazu, H. Kawazoe, K. Harada, K. Kido, and S. Inoue, “Precursor toparamagnetic centers induced in gamma-irradiated doped silica glasses,” J. Appl.Phys., 73, 1644-1649 (1993).
66. J. M. Kim, K. Oh, T. S. Park, C. S. Kim, K. Jeong, “Suppression of cladding-mode coupling loss in fiber Bragg gratings by independent control of therefractive index and photosensitivity profiles in a single-mode optical fibers,”IEEE Photon. Technol. Lett., 12, 1504-1506 (2000).
67. T. Enomoto, M. Shigehara, S. Ishikawa, T. Danzuka, H. Kanakori, “Long-periodfiber gratings in pure-silica-core fiber written by residual stress relaxation,” inProceedings of Optical Fiber Communications (OFC ´98), paper ThG2 (1998).
68. M. Rothschild, D. J. Ehrlich, and D. C. Shaver, “Effects of excimer laserirradiation on the transmission, index of refraction, and density of ultravioletgrade fused silica,” Appl. Phys. Lett., 55, 1276-1278 (1989).
69. M. Verhaegen, J. L. Brebner, and L. B. Allard, J. Albert, “Ion implantation-induced strong photosensitivity in high-purity fused silica: correlation of indexchanges with VUV color centers,” Appl. Phys. Lett., 68, 3084-3086 (1996).
34
70. S. Kannan, M. E. Fineman, J. Li, and G. H. Siegl, Jr., “Nonuniform distribution ofoxygen hole centers in silica optical fibers,” Appl. Phys. Lett., 63, 2440-3442(1993).
71. T. A. Strasser, A. E. White, M. F. Yan, P. J. Lemaire, T. Erdogan, “Strong Braggphase gratings in phosphorous-doped fiber induced by ArF excimer radiation,” inProc. Optical Fiber Conference OFC´95, paper WN3 (1995).
72. J. Canning, M. G. Sceats, H. G. Inglis, P. Hill, “Transient and permanent gratingsin phosphosilicate optical fibers produced by the flash condensation technique,”Opt. Lett., 20, 2189-2191 (1995).
73. J. Canning, M. Åslund, “The photosensitisation of optical waveguides,” in Trendsin Optics and Photonics series TOPS vol 33, Optical Society of America,Washington D. C. USA (1999).
74. J. Canning, P. –F. Hu, “Low-temperature hypersensitization of phosphosilicatewaveguides in hydrogen,” Opt. Lett., 26, 1230-1232 (2001).
75. E. M. Dianov, V. I. Karpov, A. S. Kurkov, M. V. Grekov, “Long period fibergratings and mode-field converters fabricated by thermodiffusion inphosphosilicate fibers,” in Proc. ECOC´98, 395-396, Madrid (1998).
76. G. Rego, O. Okhotnikov, E. Dianov, V. Sulimov, “Long-period fiber gratingsstable at very high temperatures,” In proceedings of Bragg Gratings,Photosensitivity, and Poling in Glass waveguides, OSA, Stresa, Italy (BGPP´01)BFB4-1 (2001).
77. A. Chmel, V. N. Svetlov, “Si-Cl groups in chlorine-impregnated silica,” J. Non-Cryst. Solids, 195, 176-179 (1996).
78. K. Awazu, H. Kawazoe, K. –I. Muta, T. Ibuki, K. Tabayashi, and K. Shobatake,“Characterization of silica glasses sintered under Cl2 ambients,” J. Appl. Phys.,69, 1849-1852 (1991).
79. K. Saito, A. J. Ikushima, “Structural relaxation enhanced by Cl ions in silicaglass,” Appl. Phys. Lett., 73, 1209-1211 (1998).
80. K. M. Golant, “Bulk silicas prepared by low-pressure plasma CVD: formation ofstructure and point defects,” in Proceedings of the NATO advanced StudyInstitute on Defects in SiO2 and related dielectrics: Science and Technology,Eriche, 427-452 (2000).
81. Y. Wang, L. Phillips, C. Yelleswarapu, T. George, A. Sharma, S. Burgett, P.Ruffin, “Enhanced growth rate for Bragg gratings formation in optical fibers withtitania-doped outer cladding,” Optics Communications, 163, 185-188 (1999).
82. Dong, J. L. Cruz, L. Reekie, M. G. Xu, and D. N. Payne, “Enhancedphotosensitivity in tin-codoped germanosilicate optical fibers,” IEEE Photon.Technol. Lett., 7, 1048-1050 (1995).
83. L. Dong, J. L. Cruz, J. A. Tucknott, L. reekie, and D. N. Payne, “Strongphotosensitive gratings in tin-doped phosphosilicate optical fibers,” Opt. Lett., 20,1982-1984 (1995).
84. G. Brambilla, V. Pruneri, and L. Reekie, “Photorefractive index gratings inSnO2:SiO2 optical fibers,” Appl. Phys. Lett., 76, 807-809 (2000).
85. G. Brambilla, V. Pruneri, L. Reekie, C. Contardi, D. Milanese, and M. Ferraris,“Bragg gratings in ternary SiO2:SnO2:Na2O optical glass fibers,” Opt. Lett., 25,1153-1155 (2000).
86. A. Anedda, C.M. Carbonaro, A. Serpi, N. Chiodini, A. Paleari, R. Scotti, G.Spinolo, G. Brambilla, V. Pruneri, “Vacuum ultraviolet absorption spectrum ofphotosensitive Sn-doped silica fiber preforms,” J. Non-Cryst. Solids, 280, 287-291(2001).
35
87. N. Chiodini, S. Ghidini, and A. Paleari, “Photoinduced processes in Sn-dopedsilica fiber preforms,” Appl. Phys. Lett., 77, 3701-3703 (2000).
88. N. Chiodini, F. Morazzoni, A. Paleari, R. Scotti, G. Spinolo, “Sol-gel synthesis ofmonolithic tin-doped silica glass,” J. Mater. Chem., 9, 2653-2658 (1999).
89. G. Brambilla, P. Hua, “Phase separation in highly-photosensitive tin-doped andcodoped silica optical fibers and preforms,” in Proc. of Bragg Gratings,Photosensitivity, and Poling in Glass waveguides, OSA, Stresa, Italy (BGPP´01)Paper BThA1 (2001).
90. K. Oh, P. S. Westbrook, R. M. Atkins, P. Reyes, R. S. Windeler, W. A. Reed, T.E. Stockert, D. Brownlow, D. DiGiovanni, “Observation of rapid, thermally stableUV photosensitive response in an Sb doped optical fiber,” BGPP 2001, Postdeadline paper PD2, Stresa, Italy (2001).
91. L. Dong, J. L. Archambault, E. Taylor, M. P. Roe, L. Reekie, P. St. J. Russell,“Photosensitivity in tantalum-doped silica optical fibers,” J. Opt. Soc. Am. B 12,1747-1750 (1995).
92. X. –C. Long, S. R. J. Brueck, “Large photosensitivity in lead-silicate glasses,”Appl. Phys. Lett., 74, 2110 –2112 (1999).
93. K. W. DeLong, V. Mizrahi, G. I. Stegeman, M. A. Saifi, M. J. Andrejco, “Color-center dynamics in a lead glass fiber,” J. Opt. Soc. Am. B, 7, 2210 (1990).
36
37
4. Thermal stability
Although fiber Bragg gratings are often referred to as permanent refractive indexstructures, exposure to increasing temperatures usually results in a decay of therefractive index modulation. Complete erasure takes place at sufficiently hightemperatures. If annealed at a fixed temperature for a sufficiently long time, the decayof the grating structure stabilizes. Afterwards, if operated at temperatures lower thanthe annealing temperature, the grating generally does not experience further thermaldecay [1]. This chapter gives a general overview of issues regarding thermal stabilityof FBGs and is intended to serve as background to chapter 5.
4.1. Decay models
The thermal stability of gratings depends on several factors such as the type of fiberused [2], use of hydrogen loading [3,4] and the writing wavelength [5]. To be able topredict the thermal stability of the refractive index after thermal annealing twoapproaches were presented by Erdogan et al [6]. The power-law model and theaccelerated aging model. Both models are based on the assumption that UV induceddefects, responsible for the change in refractive index, are thermally reversible havinga distribution of activation energies. A schematic of the model is shown in figure 4.1.In the power-law approach decay measurements are fitted to the following equation,
( )α+=η
1t/tA11 , Eq. (4-1)
where η is the normalized index change, t is time, t1 is added to keep dimensionsconsistent and the fitting parameters A and α are proportional to exp(T) and Trespectively. Once the parameters (A and α) have been determined, the decay for anytemperature and time can be calculated using equation 4-1. The difficulty with thisapproach is to determine reliable fitting parameters for different types of fibers [6,7].In the accelerated aging approach the decay in refractive index is characterized forany time and temperature combination using an aging parameter referred to as thedemarcation energy, Ed, given by
)tln(kTEd ν= , Eq. (4-2)
where k is the Boltzman's constant, T is temperature, t is time and ν is a fittingparameter, which is determined from a series of decay measurements. An example ofan aging curve is shown in figure 4.2, where the change in refractive index is plottedas a function of demarcation energy. From such a plot, the change in refractive indexcan be determined for any combination of time and temperature by using equation4-2.
From decay experiments, it is also possible to determine the energy distribution of thedefects responsible for the refractive index change by differentiating the aging curve[6,7,8]. Results for gratings written in hydrogen loaded fibers compared to non-loaded
38
fibers show a much wider energy distribution in support of the lower thermal stabilityoften found in hydrogen loaded fibers [8,9].
Figure 4.1 Schematic model for proposed decay of refractive index through thermalexcitation of trapped electrons (from [6]).
Figure 4.2 Example of the change in refractive index vs. demarcation energy neededto determine the decay behavior for arbitrary time and temperature combination
(from [7]).
39
4.2. Methods to increase thermal stability
In the following section, different approaches to form FBGs with increased thermalstability are summarized. These include accelerated aging, UV exposure, specialwriting conditions and new dopants.
4.2.1. Thermal annealing
By thermal annealing, as briefly described in the introduction to this chapter,accelerated aging removes the lesser stable contributions to the index change. Theresulting grating, although having lower refractive index modulation, has increasedthermal stability as long as the operation temperature is lower than the temperatureused for annealing.
4.2.2. Pre and post exposure
In [10], Salik et al show that fringeless pre- or post-exposure of the grating results inincreased thermal stability. A model is proposed assuming an energy barrier for defectformation and structural changes, similar to that of Erdogan [6]. Due to a distributionof energy barriers for defect formation and structural changes, some will be easilytransformed while others, with high higher energy barriers, will be more difficult totransform. The reverse is also assumed. Thermal annealing will initially relax thosemost easily transformed. By fringeless pre- or post exposure, the grating will containa dc-term of less stable refractive index and a refractive index modulation with higherthermal stability. When annealing such a grating the dc-term will decay faster,resulting in a shift in wavelength, while the refractive index modulation will remainunchanged or at least decay slower. The model is also used to explain the increasedthermal stability of gratings written using a phase mask. Here the 0th order from thephase mask creates the dc-term with lower thermal stability. A similar effect is foundfor pre-exposure of hydrogen loaded fibers [11].
4.3. Gratings with intrinsically high thermal stability
Type II gratings are manufactured through high-power single-pulse exposure resultingin localized damage at the core-cladding interface [12]. These gratings have very hightemperature resistance as can be seen from figure 4.3.
In the case of type IIa grating inscription, shown in the inset in figure 4.4, the finalgrating may consist of two superimposed gratings. The initial grating, associated totype I grating formation, has a lower thermal stability compared to the second grating,which has a negative refractive index change [13]. As the two gratings may counteractin refractive index modulation, the thermal stability is dependent on the level at whichgrating writing is stopped as indicated by the levels A, B, C etc. in figure 4.4.
In the case of special dopants, two alternatives have shown promise regardingincreased thermal stability. These are SnO2 doping and N-doping. For gratings written
40
in SnO2 doped fibers the thermal stability is increased compared to GeO2 doped orB2O3-GeO2 doped fibers. [14,15] For type IIa gratings written in N-doped fibers,thermal stability in excess of 1000 oC has been reported [16,17], as discussed inchapter 3.
Figure 4.3 Decay behavior of type II gratings (from [12]).
Figure 4.4 Decay behavior of type IIa gratings (from [13]). The letters correspond towhen the grating writing was stopped, as indicated in the inset.
41
4.4. References
1. G. Meltz, W. W. Morey, “Bragg grating formation and germanosilicate fiberphotosensitivity,” in SPIE proc., 1516, 185-199 (1991).
2. D. L. Williams, R. P. Smith, “Accelerated lifetime tests on UV written intra-coregratings in boron germania codoped silica fibre,” Electron. Lett., 31, 21202121(1995).
3. V. Grubsky, D. S. Starodubov, J. Feinberg, “Effect of molecular water on thermalstability of gratings in hydrogen-loaded optical fibers,” in Technical Digest serieson Conference on Optical Communications, OFC´99, Paper ThD2, 53-55 (1999).
4. S. Kannan. P. J. Lemaire, “Thermal reliability of Bragg gratings written inhydrogen-sensitized fibers,” OFC´96, Paper TuO4 (1996).
5. D. S. Starodubov, V. Grubsky, J. Feinberg, B. Kobrin, S. Juma, “Bragg gratingfabrication in germanosilicate fibers by use of near-UV light: a new pathway forrefractive index changes,” Opt. Lett., 22, 1086-1088 (1997).
6. T. Erdogan, V. Mizrahi, P. J. Lemaire, and D. Monroe, “Decay of ultraviolet-induced fiber Bragg gratings,” J. Appl. Phys., 1, 73-80 (1994).
7. S. Kannan, J. Z. Y. Guo, P. J. Lemaire, “Thermal stability analysis of UV-inducedfiber Bragg gratings,” J. Lightwave Technol., 15, 1478-1483 (1997).
8. J. Rathje, M. Kristensen, “Continous anneal method for characterizing the thermalstability of ultraviolet Bragg gratings,” J. Appl. Phys., 88, 1050-1055 (2000).
9. G. Robert, I. Riant, “Demonstration of two distributions of defects centers inhydrogen-loaded high germanium content fibers,” in Proceedings of Optical FiberCommunications, Paper WL18, 180-181 (1997).
10. E. Salik, D. S. Starodubov, V. Grubsky, J. Feinberg, “Thermally stable gratings inoptical fibers without temperature annealing,” OFC´99 Technical digest series,Vol. 3, 56-58 (1999).
11. M. Åslund, J. Canning, “Annealing properties of gratings written into UV-presensitized hydrogen-outdiffused optical fiber,” Opt. Lett., 25, 692-694 (2000).
12. J.-L. Archambault, L. Reekie, P. St. J. Russel, “100% reflectivity Bragg reflectorsproduced in optical fibres by single excimer laser pulses,” Electron. Lett., 29, 453-455 (1993).
13. L. Dong, W. F. Liu, “Thermal decay of fiber Bragg gratings of positive andnegative index changes formed at 193 nm in a boron-codoped germanosilicatefiber,” Appl. Opt., 36, 8222-8226 (1997).
14. L. Dong, J. L. Cruz, L. Reekie, M. G. Xu, D. N. Payne, “Enhancedphotosensitivity in tin-codoped germanosilicate optical fibers,” Photon. Technol.Lett., 7, 1048-1050 (1995).
15. G. Brambilla, V. Pruneri, “Enhanced photorefractivity in tin-doped silica opticalfibers (review),” J. Select. Topics in Quant. Electron., 7, 403-408 (2001).
16. E. M. Dianov, K. M. Golant, R. R. Khrapho, A. S. Kurkov, B. Leconte, M.Douay, P. Bernage, P. Niay, “Grating formation in germanium free siliconoxynitride fibre,” Electron. Lett., 33, 236-238 (1997).
17. O. V. Butov, K. M. Golant, I. V. Nikolin, “Ultra-thermo-resistant Bragg gratingswritten in nitrogen-doped silica fibres,” Electron. Lett., 38, 523-525 (2002).
42
43
5. Chemical composition gratings
5.1. Background
The origin of the UV induced refractive index changes, as described in previouschapters, has been attributed to the creation of defects, stress relaxation or compactionof the glass structure. The temperature limit for thermal stability of UV induceddefects can be roughly estimated to be 100-600 oC, depending on the type of fiber.The decay mechanism can be associated to the recombination of defects, as describedby Erdogan (see chapter 4.1). Likewise, the upper limit of thermal stability forrefractive index changes attributed to UV induced stress relaxation or compaction,which is generally higher compared to UV induced defects, can be estimated to be400-1000 oC, depending on the type of fiber. The decay mechanism is this case couldbe linked to a temperature at which the viscosity of the glass allows strained bondsand the glass structure to relax towards equilibrium [1]. To improve the temperaturestability of fiber Bragg gratings, new dopants could be identified that result in morestable UV induced defects. Another possibility could be to increase the viscosity ofthe glass.
Other possible way to record more permanent gratings could be to cause UV inducedphase separation or crystallization, which has been observed in GeO2 doped silicafilms [2] and SnO2 doped fibers [3]. However, as the non-exposed regions may alsoundergo phase separation, when exposed to high temperatures, such gratings wouldnot necessarily result in higher thermal stability of the refractive index modulation.Even if non-exposed regions do not undergo phase separation, exposed areas mayexperience continued phase separation or crystal growth, resulting in non-stablegrating characteristics (increasing or decreasing reflectivity).
If one considers recombination of defect and structural relaxation to be twomechanisms associated with increasing thermal stability of FBGs, the nextmechanism of even higher thermal stability would be that limited by viscous flow ordiffusion of dopants. Reaching temperatures where dopants diffuse within the fiberwould not only limit the stability of gratings, but ultimately limit the thermal stabilityof the fiber itself.
In 1996, a method was proposed to create high-temperature stable Bragg gratings byperiodically altering the concentration of fluorine in the core of the fiber [Paper K andM]. A change in the fluorine concentration would subsequently result in a change inthe refractive index. The limiting thermal stability of such a refractive index structurewould depend on the diffusing properties of fluorine. The proposed method to alterthe fluorine concentration in the core was based on the chemical equation shown infigure 5.1. The equation was proposed by Kirchhof et al for the reduction of hydroxylgroups in fluorine doped silica [4]. The reduction of hydroxyl groups was attributed toformation and diffusion of hydrogen fluoride, which was assumed to have a highermobility than bound fluorine. A different way of looking at the equation, is that thereis a reduction of fluorine in hydroxyl doped regions, as discussed in section 2.4.4.
44
Introduction of hydroxyl groups in the core of the fiber can be achieved eitherthermally or by UV exposure of hydrogen loaded fibers (see chapter 2). By usingfringed UV exposure, hydroxyl formation can be achieved locally with the periodicityof the interference pattern, creating sub-micron structures. Fiber Bragg gratings inwhich the refractive index modulation is a result of a change in the chemicalcomposition are here referred to as chemical composition gratings (CCGs).
Figure 5.1 Schematic of the suggested mechanism for the reduction of hydroxylgroups in fluorine doped optical fibers [4].
Due to the quadratic dependence on distance in diffusion processes, the grating periodof the modulation will also have an effect on the thermal stability. For examplegratings reflecting at 800 nm (period ~0.25 µm) will decay approximately four timesfaster than gratings at 1550 nm (period ~0.5 µm). For long period gratings withperiods typically > 100 µm, diffusion from the core to the cladding dominates, andthus the core size (~8-9 µm for standard telecommunications fiber) will determine themaximum thermal stability. Fabrication of long period CCGs can, for the abovereason, result in gratings with extremely high thermal stability [5]. For gratings withsufficiently long period (fringes ≥ 125 µm) it is possible to separately heat each fringeand cause local dopant diffusion e.g. using a CO2-laser, electric discharge or otherlocal heating sources. For short period gratings, the geometry of the fiber prohibitsseparate fringe heating.
5.2. Typical manufacturing procedure of chemical compositiongratings
Chemical composition gratings have typically been manufactured by hydrogenloading a germanium-fluorine core doped fiber and exposing the fiber to fringed UVradiation followed by a thermal developing process. The UV exposure results in anormal, type I grating but also creates a periodic concentration of hydroxyl groupswith the same periodicity as the grating. The grating has then been inserted into a tubefurnace and the temperature increased slowly (typically 25 oC/min) to 1000 oC. Thedynamics of the grating reflectivity during this thermal treatment is shown in figure5.2. The increase in temperature results in the decay of the type I grating, which isconsequently erased at ~40 minutes. Following this erasure, a much more stablegrating grows and then saturates. This second grating is identified as the CCG, whichis believed to be formed by the formation and out-diffusion of mobile HF molecules.
Si - OH + F-Si Si - O - Si + HF
45
0 10 20 30 40 50 60 700,0
0,2
0,4
0,6
0,8
1,0
Time [min]
Norm
aliz
ed re
flect
ivity
0
200
400
600
800
1000
CCG
Initial grating
Temperature [ oC]
Figure 5.2 Typical behavior of grating reflectivity during CCG formation.
5.3. Hydroxyl assisted fluorine diffusion
An analytical tool was required to verify the effects of hydroxyl groups on thebehavior of fluorine diffusion. Dynamic secondary ion mass spectrometry (SIMS) hasbeen frequently used for dopant profiling of fiber preforms. However, for opticalfibers with a typical core diameter of 8-10 µm the resolution provided byconventional SIMS is not sufficient. It is in some cases possible to use SEM/EDS toanalyze fiber cross sections but the results may be impaired by sample chargingeffects, insufficient lateral resolution and low sensitivity for light elements such asfluorine. For these reasons time-of-flight secondary ion mass spectrometry (ToFSIMS) [6,7] was evaluated (Thesis papers I and V).
Experiments were performed on the same fibers used for manufacturing CCGscontaining fluorine and germanium in the core. In these experiments, high hydroxylconcentration was achieved in the core of the fibers through OH flooding (Paper VI),i.e. high-temperature treatment of hydrogen loaded fibers. Four samples wereexamined in this study: (a) pristine fiber, (b) pristine fiber heated for 10 minutes at1000 °C, (c) OH flooded fiber (containing high levels of hydroxyl) by heating for 1second at 1000 °C and (d) OH flooded fiber heated for 10 minutes at 1000 °C.
The results of the ToF-SIMS profiling of the fluorine concentration for samples (a)through (d) are shown in figure 5.3 (a) through (d) and figure 5.4 (a) through (d)respectively. The dimensions are indicated in the figures. The results are shown ascolored surface plots of the fluorine concentration of cleaved fibers (figure 5.3) andcorresponding line-scans through the center of the core (figure 5.4). The measuredfluorine distribution of untreated reference sample is shown in figure 5.3 (a). Thefluorine distribution corresponds to the refractive index structure as follows: thecentral ring-structure corresponds to the core region with a central dip in fluorineconcentration, characteristic for MCVD manufactured fibers. The central region alsocontains germanium (shown as a dashed line in figure 5.4 (a)) and the surrounding
46
structure consists of the deposited cladding layers, the inner region of which has aslightly higher fluorine concentration than the outer region. The inner cladding regionis also phosphorous-free. Surrounding the deposited cladding layers is the silicapreform tube. The ToF-SIMS analysis of sample (b), shown in figure 5.4 (b), showsthe fluorine distribution of the fiber heated to 1000 °C for 10 min. The trace is quitesimilar to that of figure 5.4 (a), implying that little apparent diffusion occurs by justheating the fiber without further treatment. Samples (c) and (d) show the fluorinedistribution of OH flooded fibers heated to 1000 ºC for 1 second and for 10 min,respectively. In this case, there is a significant redistribution of fluorine even after aheating time of only 1 second. After heating the OH flooded fiber (d) for 10 minutesthe central dip in the fluorine distribution has been removed completely. No change ingermanium distribution was observed in any of the samples (b), (c) or (d).
(a) (b)
(c) (d)
Figure 5.3 ToF-SIMS measurement of fluorine distribution of (a) untreated referencefiber (sample a), (b) fiber heated at 1000 oC for 10 minutes (sample b), (c) in OH
flooded fiber heated at 1000 oC for 1 second (sample c), and (d) in OH flooded fiberheated at 1000 oC for 10 minutes (sample d).
47
(a) (b)
(c) (d)
Figure 5.4 ToF-SIMS measurement of fluorine distribution of (a) untreated referencefiber (sample a), (b) fiber heated at 1000 oC for 10 minutes (sample b), (c) in OH
flooded fiber heated at 1000 oC for 1 second (sample c), and (d) in OH flooded fiberheated at 1000 oC for 10 minutes (sample d). The germanium distribution is shown in
figure 5.4 (a) for comparison.
A second set of experiments with ToF SIMS was also performed to assess the abilityto measure compositional profiles in Er3+, Yb3+ and P2O5 doped silica fiber. Thedopant levels for this fiber were not available. The results are shown in figure 5.5. Allfibers used in the experiments were produced using modified chemical vapordeposition (MCVD). In erbium doped fibers the erbium concentration wasincorporated using solution doping of MCVD prepared preforms [8] while for thefluorine diffusion studies the central core was doped with germanium and fluorineusing GeCl4 and C2F6 as precursors.
The results show that the ToF-SIMS instrument can be used for directly analyzingcross-sections of optical fibers, with high sensitivity and high lateral resolution. Thelateral resolution was estimated to be 0.5 µm.
48
0
5
10
15
20
25
- 20 -10 0 10 20
Distance from fiber center (µm)
No.
of i
ons
(arb
. uni
ts)
Er+ x5
Yb+
P-
Figure 5.5 Line scans from ToF-SIMS measurement of Er, Yb and P2O5 doped silicafiber (Paper V).
5.4. Decay behavior of chemical composition gratings
As the grating reflectivity is caused by a periodic change in the chemical composition,the thermal stability will be limited by fringe-to-fringe diffusion of the modulateddopants as schematically shown in figure 5.6. Here the periodic variation ofconcentration, or refractive index, is shown for different degrees of dopant diffusion.
-1,0
-0,8
-0,6
-0,4
-0,2
0,0
0,2
0,4
0,6
0,8
1,0
(c)
(b)
(a)
Mod
ulat
ion
ampl
itude
[a.u
.]
Axial position along grating [a.u.]
Figure 5.6 Schematic of decay of refractive index modulation or change in dopantdistribution due to fringe-to-fringe diffusion shown as (a) initial distribution, (b)
partially erased and (c) erased.
49
5.4.1. Modeling the decay behavior of CCGs
Due to the nature of CCGs, the thermal stability should be limited by the diffusionproperties of the modulated dopants. A model was developed to better understand theproperties and behavior of CCGs (Thesis paper II). The model was based on diffusionof dopants in a periodic structure consisting of parallel infinitely wide layers ofalternating concentration. The model was based on the following assumptions:
(1) the refractive index modulation is due to a periodic variation of the concentrationof one or several dopants,
(2) the diffusion of dopants out of the core (core diameter ~5-10 µm) is negligiblecompared to the diffusion of the individual grating fringes (fringe width~0.25µm), and
(3) the refractive index is linearly dependent on dopant concentration.
The following equation describes the concentration profile of periodic and paralleldiffusing planes [9],
( ) ( ),
Dt2x1/2merf
Dt2xmerf
m Dt2xmerf
Dt2x1/2merf
20C
C��
���−Λ⋅−−−Λ⋅−�
∞
−∞= ��
���
��
����
����
� −Λ⋅−���
����
� −Λ⋅+⋅=��
�
�
��
�
�
��
�
�
�
��
�
�
�
Eq. (5-1)
where C is the dopant concentration, C0 is the initial peak-to-peak concentration, x isthe axial position along the grating, t is time and Λ is the geometrical grating period.D is the temperature dependent diffusion coefficient given by D=D0exp(-Ea/RT)where Ea is the activation energy, D0 is a constant, R is the gas constant and T thetemperature in Kelvin. At time t=0, equation 5-1 describes a square wave pattern. Anearly sinusoidal dopant profile is expected from the UV exposure followed bychemical reactions. In order to use equation 5-1 to describe the effect of heating, thesquare wave pattern must be allowed to evolve first into a nearly sinusoidaldistribution. This introduces a delay time after which equation 5-1 can be used. Afterthis delay time (that is not critical), a change in peak to peak amplitude of equation 5-1 is then proportional to the change in refractive index modulation.
Kirchhof et al [10] found that adding phosphorous to fluorine doped silica glasssignificantly increased the diffusion rate of fluorine. As phosphorous is commonlyused in fiber manufacturing, three different situations need to be considered whenmodeling the decay of CCGs.
(1) When the fiber contains no phosphorous, the diffusion parameters for fluorine(DF) can be used.
(2) When the core and cladding are co-doped with phosphorous, the diffusionparameters for fluorine co-doped with phosphorous (DF(P)) should be used.
50
(3) When the inner cladding surrounding the core is co-doped with phosphorouswhile the core is not. In this case the change in refractive index or concentration(C) is obtained by superposition of two gratings with different diffusioncoefficients C=C(DF) + C(DF(P)). The reason is that the grating may extend intothe cladding and due to inter-diffusion of dopants between the core and claddingduring preform and fiber manufacturing.
5.4.2. Modeled and experimental decay of CCGs
Two single mode silica fibers were manufactured using MCVD technology in order todetermine diffusion coefficients for CCG decay and to assess the effects ofphosphorous on the grating stability. Fiber (I) was manufactured with a germanium-fluorine doped core and inner cladding co-doped only with fluorine, as in situation 1.Fiber (II) had a germanium-fluorine doped core and inner cladding co-doped withphosphorous and fluorine, corresponding to the third situation described above. Thecore diameter of both fibers was ~8.5 µm with core refractive index steps of∆nI~4.8·10-3 and ∆nII~5.2·10-3 respectively. The fibers were hydrogen loaded at 10MPa and 50 ºC for 5 days prior UV exposure. The periodic UV exposure wasperformed using a frequency doubled Argon-ion laser at 244 nm wavelength, in atwo-beam interferometric setup. The gratings were 15 mm long with a Gaussianapodized profile at a Bragg wavelength of ~1540 nm. The CCGs were developed asdescribed in section 5.2 (figure 5.2). When studying the decay of the developedgratings the starting time (t´=0) was set to zero when the CCG reflectivity was at amaximum as shown by the dashed line in figure 5.2.
The decay of the CCGs was then studied in the temperature range 1000-1200 oC and adiffusion constant could be obtained for each temperature by fitting the model to theexperimental decay curves, as plotted in a lin-log scale in figure 5.7 for fiber (I). Thecorresponding result for the decay in fiber (II) is shown in a lin-lin scale in figure 5.8.Two diffusion coefficients were required to fit the model to the decay in fiber (II),corresponding to the fast initial decay for t < 25 minutes and the following slowdecay. The determined diffusion coefficients for fiber (I) and (II) are plotted in figure5.9, where the two data points for fiber (II) are from the same grating. From a linearfit, the derived activation energy for CCG decay (without phosphorous co-doping)was Ea=312 (±20) kJmol-1 with a pre-exponential constant D0=3.69·10-3 cm2s-1.Besides the linear fit to experimental data shown as a dashed line, figure 5.9 includesthe and reported values for fluorine diffusion given for the temperature range 1000 oCto 2000 oC, and for fluorine diffusion when co-doped with phosphorous given for thetemperature range 1600 oC to 2000 oC, both represented by solid lines [10]. Tocompare with the experimental results, the value for fluorine diffusion when co-dopedwith phosphorous is extrapolated down to 1000 oC (dotted line).
The values of diffusion coefficients attained from the model are in good agreementwith reported values for fluorine diffusion as well as for fluorine diffusion whencodoped with phosphorous.
51
Figure 5.7 Decay characteristics of CCGs in germanium-fluorine doped fiberscontaining no phosphorous. Symbols are simulated values.
Figure 5.8 Decay characteristics of CCG in germanium-fluorine doped fiberscontaining phosphorous in the cladding surrounding the core. Symbols are simulated
values.
52
Figure 5.9 Arrhenius plots for diffusion coefficients showing the experimental resultsfor fiber-I (�) and fiber-II (□) and the linear fit for the slow decay (─ ─). Additional
data are reported diffusion values for fluorine (lower solid line) and fluorine co-doped with phosphorous (upper solid line) [10]. Values for fluorine co-doped withphosphorous have been extrapolated for comparison (•••). The two data points forfiber II correspond to the fast initial decay followed and the following slow decay.
5.5. Dynamics of CCG formation
In order to increase the efficiency of the development process attempts were made toremove the temperature ramping and developing the grating directly at 1000 oC usingminiature oven described in the following section (Chapter 5.6 and Paper IV). Byusing a miniature oven the length of heat-treated fiber could be decreasedsignificantly resulting in an overall increased mechanical strength of the final gratingas well as simplified handling. The initial results however showed a large variation inthe final reflectivity of the CCGs. The refractive index modulation was significantlylower than of CCGs developed using the tube furnace and in some cases, no CCGcould be detected at all. This observation indicated that the thermal history and thethermal ramping procedure are crucial for the final refractive index modulation of theCCG. To study the effects of thermal annealing three different sets of experimentswere performed.
1. In the initial experiment, the effect of annealing temperature and duration wasstudied. The gratings were annealed and developed within one week after thegratings had been written while being stored at room temperature. As the timerequired for out-diffusion of hydrogen from the fiber is approximately two weeksat room temperature, the fibers were partially hydrogen loaded.
53
2. In the second experiment, two gratings were developed in order to determine theeffect of remaining hydrogen on the final outcome of the CCG. To remove theremaining hydrogen one grating was heated to 100 oC for ~20 hrs prior annealing,while the second grating was simply stored at room-temperature forapproximately 1 week prior annealing. Both gratings were then annealed at 700 oCfor 24 minutes and subsequently developed at 1000 oC. At 100 oC thermallyinitiated reactions of hydrogen are assumed to be minimal.
3. In the third experiment the annealing time was held constant at 24 minutes whilevarying the annealing temperature from 500 to 800 oC in steps of 100 oC. Thesefibers had been previously treated at 100 oC for 20 hrs to remove any remaininghydrogen.
In the three experiments, gratings were developed at 1000 oC for ~20 min. The resultsof the experiments are shown in figure 5.10, 5.11, and 5.12 respectively. Conclusionsdrawn from the three experiments can be summarized as follows.
1. The resulting refractive index modulation of the CCG is highly dependent on thethermal treatment prior development.
2. The effect of low-temperature treatment to remove remaining hydrogen is clearlyseen as a significantly increased grating reflectivity.
3. The results when the annealing time is constant at 24 minutes and the temperatureis changed shows that there is clearly an optimum annealing temperature near600-700 oC.
Figure 5.10 Final refractive index modulation of developed CCGs as a function ofannealing time at annealing temperatures of 600, 700, and 800 oC.
54
Figure 5.11 Growth of reflectivity during the development at 1000 oC for the fiberwithout hydrogen (top) and the fiber partially hydrogen loaded (lower).
Figure 5.12 Final refractive index modulation of CCGs annealed for 24 minutes atdifferent annealing temperatures prior developing (the line is drawn to guide the eye)showing that for a constant annealing time of 24 minutes there is clearly an optimum
annealing temperature near 600-700 oC.
55
Following fringed UV exposure, the underlying mechanism of CCG formation isattributed to redistribution of fluorine as a result of HF formation according to thereaction pathways shown given by equations (5-2) through (5-4).
2≡X-OH ↔ ≡X-O-X≡ + H2O Eq. (5-2)
≡X-OH +F-X≡ ↔≡X-O-X≡ + HF Eq. (5-3)
≡X-F + H2O ↔ ≡X-OH + HF Eq. (5-4)
where X is either Si or Ge. A plausible explanation for the results shown in figure5.12 is as follows. At low temperatures, the hydroxyl groups can be consideredimmobile [11]. A consequence will be that there is no interaction between fluorineand hydroxyl groups or between hydroxyl groups. i.e. the formation of HF ormolecular water is negligible. As the temperature increases the rate of HF formationshould increase, either as a result of interaction of fluorine with hydroxyl groups (Eq.5-3) or fluorine with molecular water (Eq. 5-4). The increase in the rate of HFformation would result in an increase in refractive index modulation. As thetemperature is increased further, the rapid diffusion of hydroxyl and molecular waterwill result in a decrease in the OH-visibility, i.e. the amplitude modulation of thehydroxyl concentration. If the decrease in OH-visibility is faster than the rate of HFformation, the final refractive index modulation will be lower although the averagechange in refractive index may be the same. In addition, if equation (5-4) is the mainpathway for HF formation, the temperature dependence of equation (5.2), i.e.temperature dependence of the concentration of molecular water [11,12] and thediffusion-reaction behavior [13] will have a large effect on the outcome of the CCG.Another possibility for the decrease in refractive index modulation is a depletion ofhydroxyl groups by formation and diffusion of H2 molecules (see e.g. [13]).
The results from the second experiment indicate that thermally induced reactionsoccur with the remaining hydrogen, reducing the achievable refractive indexmodulation. A possible explanation is that thermally induced hydroxyl groups areprimarily formed in non-UV-exposed areas, which should have a larger amount ofreactive sites available (see chapter 2). The consequence of such a process would be adecrease in OH-visibility, resulting in a lower refractive index modulation of theCCG. The effect is also seen when comparing the results for 600 oC and 700 oCannealing shown in figure 5.10 with those in figure 5.12.
5.6. Ultra-stable chemical composition gratings
Similar high-temperature decay experiments, as described in section 5.4.2, were alsoperformed on CCGs formed from initially strong gratings (prior development) withrefractive index modulation greater than ∆n>1•10-3. In this case part of the gratingreflectivity remained stable after an initial decay following a behavior similar to thatin figure 5.7. For the remaining grating no significant decay was observed at 1106 ºC,measured with a spectrum analyzer, during 60 minutes as shown in figure 5.13. In thedifferent chemical pathways proposed, an additional source of change in the chemicalcomposition arise from the formation of molecular water, which diffuses within the
56
glass network resulting in a periodic reduction of oxygen as described by equation 5-2. For strong gratings, longer exposure times results in a higher concentration ofhydroxyl groups subsequently increasing the concentration of molecular water.Although the effect of oxygen concentration on the refractive index in germaniumdoped glass has not been treated in detail in the literature, the process should result ina structural and compositional modification of the glass and a subsequent change inthe refractive index after structural relaxation towards thermal equilibrium.
Figure 5.13 Stability of remaining superimposed grating at 1106 oC during 60 minafter an initial decay corresponding to diffusion of fluorine.
5.7. Miniature oven for processing chemical composition gratings
The CCGs described previously were manufactured using a large tube furnace or aminiature oven (Paper IV) during the development procedure, i.e. the thermal heattreatment to 1000 oC used to create the CCG. The miniature oven was developed inorder to increase the efficiency of the development process and to decrease the lengthof heat-treated section for overall increased mechanical strength of the fiber. In otherapplications, one may also wish to rapidly heat the fiber in matter of seconds or lesswhile monitoring transmission properties, a task that becomes very impractical withlarger ovens.
The basic design of the oven is shown in figure 5.14. The electric resistance oven usesa high-temperature alloy as heating elements, made from sections of ribbon formedinto U-shape. The ends of the heating elements were spot-welded to iron rods whichwere used for electric contact. The oven was built by using five heating elementsinserted through holes in a stainless steel block. To isolate the heating element fromthe holding block high-temperature ceramic tubing was fitted into the holes.
57
A
B
CDEF
G
HI
(a) (b)
Figure 5.14 Individual heating element (a)-left and a series of heating elementsmaking up the length of the oven (a)-right and (b) a schematic diagram of oven
showing: A) stainless steel holding block, B) inlet for gases into oven, C) ceramictubing for electric insulation, D) stainless steel heat reflector, E) iron rod for
connectors, F) pure silica half cylinder covering oven elements, G) alloy heatingelements, H) pure silica protective sleeve and I) the fiber showing the direction of
insertion.
Figure 5.15 Photograph of the miniature oven operating at 1000 oC.
58
All heating elements are connected in series, the three inner ones bypassed in parallelby a variable resistor. The variable resistor was used to tune the temperature profile. Ahole through the holding block was made to enable gas-flow into the oven behind theheating elements to enable thermal treatment in an inert atmosphere. Between theheating elements and the gas outlet, a U-shaped stainless steel sheet was placed tofunction as a heat reflector and to avoid direct gas-flow onto the heating elements.Inside the heating elements, a U-shaped pure silica glass plate was inserted. This wasused to protect the fiber from being contaminated by the surrounding material. Theheating elements were encompassed by a half-cylinder of pure silica glass with sealedtop, in order to minimize temperature fluctuation due to the environment, and tofunction as a gas container when using externally supplied gases. A slit was cut in thepure silica glass half-cylinder to enable the insertion of the fiber into the oven. Apicture of the oven operating at 1000 oC is shown in figure 5.15. The oven wascalibrated by scanning a 2 mm long CCG through the oven. The results are shown infigure 5.16. By proper adjustment of the variable resistor the temperature variationcould be controlled to be less than 2% over 16 mm and less than 10 % over 20 mm.
Figure 5.16 Measured temperature profile of oven by using a CCG as a miniaturetemperature sensor.
59
5.8 References
1. A. K. Varshneya, Fundamentals of inorganic glasses, Academic press, 1994.2. H. Hosono, J. Nishii, “High photosensitivity and nanometer-scale phase
separation in GeO2-SiO2 glass thin films,” Opt. Lett., 24, 1352-1354 (1999).3. G. Brambilla, P. Hua, “Phase separation in highly-photosensitive tin-doped and
codoped silica optical fibers and preforms,” in Bragg gratings, Photosensitivity,and Poling in Glass waveguides, Paper BThA1, OSA, Italy (2001).
4. J. Kirchhof, S. Unger, H-J. Pißler, and B. Knappe, ”Hydrogen-induced hydroxylprofiles in doped silica layers”, OFC’95, Vol. 8, 1995 OSA Tech. Dig. Series,paper WP9.
5. V. I. Karpov, M. V. Grekov, E. M. Dianov, K. M. Golant, R. R. Khrapko, “Ultra-thermostable long-period gratings written in nitrogen-doped silica fibers”, Mat.Res. Soc. Symp. Proc., 531, 391-396 (1998).
6. ToF-SIMS: Surface Analysis by Mass Spectrometry, Edited by John C Vickermanand David Briggs, IM Publications and Surface, Spectra Limited (2001).
7. A. Benninghoven, "Chemical Analysis of Inorganic and Organic and Thin Filmsby Static Time-of-Flight Secondary Ion Mass Spectrometry (ToF-SIMS),"Angew. Chem. Int. Ed. Engl., 33, 1023-1043 (1994).
8. W. Jost, “The fundamental laws of diffusion”, in Diffusion, (Acad. Press, NewYork, pp. 32-46, (1952).
9. J. Kirchhof, S. Unger, K-F. Klein, and B. Knappe, “Diffusion behavior of fluorinein silica glass,” J. Non-Cryst. Solids, 181, 266-273, 1995.
10. P. Urquhart, Optoelectronics IEEE Proceedings, 135, 385-407 (1988).11. R. H. Doremus, "Diffusion of water in silica glass," J. Mater. Res., 10, 2379
(1995).12. M. Tomozawa, J. Am. Ceram., "Concentration dependence of the diffusion
coefficient of water in SiO2 glass," J. Am. Ceram. Soc., 68, C-251 (1985)13. P. B. McGinnis, J. E. Shelby, "Diffusion of water in vitreous silica," J. Non-Cryst.
Solids, 179, 185-193 (1994).
60
61
6. Extended functionality of optical fibers
6.1. Background
As telecommunication systems become more complex, new optical components arerequired to perform specific tasks. Traditionally optical fibers have mainly been usedfor long-haul transmission by exploiting their low losses and low sensitivity toexternal perturbations. With the development of the erbium-doped fiber amplifier newfunctions were added to the fiber [1]. Today specialty fibers are increasingly beingused for developing fiber components with different optical functions. A fewexamples of components are reflection and loss filters, switches, modulators andvariable attenuators. One of the main advantages of using optical fibers is thepotentially low cost, small size and the ability to splice for easy integration into fiberoptic networks. For active components one needs to overcome the optically passivemedium, which is characteristic for glass. This can be achieved e.g. by interaction ofthe evanescent field of the waveguide with materials, such as polymers or thin films[2,3], by using active/passive dopants in the core, such as erbium or ytterbium [1], oraffecting propagation by thermal, acoustic or mechanical means, or with externalfields [4,5]. Combining these techniques with special refractive index structures ormultiple cores open up new possibilities for active fiber based components. Followingare a few examples of techniques that can be used to manufacture fibers withextended functionality:
1. Special fiber geometry- Micro-structured fibers such as holey fibers and photonic bandgap fibers [6]- Multiple core fiber [7]
2. Special processing techniques- UV exposure for FBG manufacturing [8].- OH flooding [Paper 6]
3. Micro-mechanical designs- Pressing, bending, stretching [9].
In the following sections, two examples of single fiber Mach-Zehnder interferometer(MZI) are presented relating to refractive index modifications, multiple core fibersand applied electric field.
6.2. Mach-Zehnder interferometers for filtering and switching
To efficiently transform phase information to amplitude modulation some type ofinterferometric setup needs to be used. Typically, amplitude modulators used intelecommunications are either a MZI structure or based on fast absorption modulators.A typical lab set-up of a fiber based MZI is shown in figure 6.1. The interferometer istypically made by splicing together two 50/50 fiber couplers and the transmitted lightcan be changed between the two output fibers by changing the phase (φ) of one orpossibly both arms. The problems facing this set-up is the size and the separation of
62
the two arms of the interferometer resulting in a component with high sensitivity toambient fluctuations, relatively complex and expensive.
Figure 6.1 Typical lab setup fiber based Mach-Zehnder interferometer made withtwo couplers and changing the phase of one of the arms.
In the following sections, two methods are presented for manufacturing integratedsingle-fiber MZI. In the first MZI structure the two arms of the MZI, as shown infigure 6.1, are replaced by two propagating modes in a single bi-modal core. In thesecond MZI the two arms are replaced with a single fiber containing two cores. Thebi-modal MZI structure (thesis paper VIII) is passive and acts as a wavelengthdependent filter. The two-core MZI is an active component made by incorporatingtwo metal electrodes in proximity of one of the cores (thesis paper IX). The low non-linearity of the fiber is compensated with increased length of the component, but isintended to be used in combination with poling.
6.3. Single core bimodal Mach-Zehnder interferometer using OHflooding for refractive index control
When heating hydrogen loaded fibers, OH flooding (Thesis paper VI), both thephotosensitivity and the refractive index increases. In figure 6.2 and 6.3, the dynamicsof hydroxyl formation during OH flooding and the relation between induced hydroxylconcentration and the change in refractive index is shown. To illustrate the potentialof this technique an in-fiber MZI was designed in a single piece of standardtelecommunications fiber (Corning SMF-28). The idea was to make part of the fiberbi-modal by controlled increase of the refractive index in the core. To render the fiberbi-modal a refractive index change of ∆n~1.7·10-3 was required. From the graphs infigure 6.2 and 6.3 a heating time of ~2.6 seconds was determined to reach sufficientrefractive index change. The MZI was manufactured by pulling a hydrogen loadedstandard telecommunications fiber through a miniature oven (Paper IV) at a speed toeffectively heat each portion for approximately 2.6 seconds. To excite both the LP01and LP11 mode in the heat treated section it was necessary to induce microbends at theinput and output of the interferometer [10]. The experimental setup is shown in figure6.4. A transmission spectrum of the MZI is shown in figure 6.5, which also includes atrace of the spectrum after exposing part of the interferometer to UV light, a processthat can be used to tune the output spectrum of the MZI.
50/50 50/50
δφ
63
Figure 6.2 Dynamics of hydroxyl formation during OH flooding at 1000 oC (Thesispaper I and VI).
Figure 6.3 Refractive index changes due to OH flooding as a function of OHabsorption (Thesis paper VI).
Figure 6.4 Experimental setup for the single-core bi-modal MZI (Thesis paper VIII).
64
Figure 6.5 Transmission spectrum of the MZI. The dashed line is the spectrum afterUV exposure (Thesis paper VIII).
6.4. Dual core Mach-Zehnder interferometer for electroopticswitching
In the following section, a different and externally controlled MZI was constructedbased on the combination of multiple cores and applied electric fields. By placing thecores close to each other in a two-core fiber, coupling of light can occur between thecores. To actively control the light, two electrodes are placed near one core in order tocreate a phase difference between the two cores. The low non-linearity of glass can beovercome by making long devices utilizing the weak Kerr effect [11] or by poling inorder to increase the non-linearity [4,5].
Holes
Cores
Figure 6.6 Example of a fiber structure for an integrated MZI with internalelectrodes.
65
6.4.1. Electrode insertion technique
There has been intense research into poling of glass and glass waveguides in order todemonstrate low-cost optical modulators. Poling is achieved by applying a highelectric field across the glass sample while at elevated temperature. The procedure hasalso been performed in fibers, similar to that shown in figure 6.6 containing only onehole. The electrodes are typically thin wires inserted from the cleaved end or by sidepolishing [12] the fiber to access the hole. This technique is time-consuming and thelength of inserting the electrodes is limited to typically 10-20 cm. Using the wireinsertion method requires the wire to be much smaller than the hole, which results invariations of the position of the wire in the holes. This effect will also be differentfrom device to device. This leads to uncertainty in the performance before and afterpoling (varying electric field and impedance along the fiber).
To overcome the problems with wire insertion a different approach was taken. Herethe electrodes are inserted in molten form by the use of low-melting temperaturealloys. The metal is pressed into the holes by placing the fiber in a pressure chamber,which is placed in an oven. The setup is shown schematically in figure 6.7.
High pressure
Fiber with holes
Liquid metal
High temperature
Figure 6.7 Schematic of the setup for inserting metal electrodes into fiber (Paper IX).
The electrode fabrication technique consists of heating an alloy with relatively lowmelting point in an oven, so that a liquid is available. One extreme of the fibers withholes is inserted into the molten metal and the other extreme is kept free in theatmosphere. The liquid metal is contained in small crucible in a sealed cell that ispressurized. In this way, the liquid metal is forced up into the fiber. Normally, liquidsare sucked into the holes using vacuum. Using the presented pressure cell, thedifference in pressure can greatly exceed one bar that is obtained with vacuum,enabling fast filling and manufacturing of long devices. The technique of pressurizingor sucking liquids into silica fibers and capillaries has been previously used in variousfields and to make fiber components (see for example [13]).
The described method provides smooth, continuous electrodes, as shown in figure 6.8and 6.9. Figure 6.8 shows two metal electrodes extruding from a previously filledfiber which was cleaved and then pulled apart. As can be seen the wires fill the whole
66
volume of the hole and have a smooth surface. In figure 6.9 a fiber with ~4.2 metersof continuous electrodes is shown. The picture is taken from part of the spool onwhich the fiber was wound up. Fibers with continuous electrode lengths of more than22 meters have been produced this way although care has to be taken to avoiddiscontinuities in the metal. These discontinuities occur due to the difference inthermal expansion of the metal and glass.
Figure 6.8 Cleaved fiber with two extruding metal electrode.
Figure 6.9 Fiber containing 420 cm of two continuous metal electrodes. The fiber iswound on a small spool from which the picture is taken.
67
To enable splicing the beginning and end of the fiber is made without metal. Theoutput end is free from metal by placing this part outside of the heating chamberwhile the input end is freed from metal by removing the metal from the pressurechamber and pressing the metal farther into the fiber. To access the metal electrode aside-polishing technique was used. Once the metal is accessible, it is contacted eitherby a thin wire or by the use of e.g. conductive epoxy. Figure 6.10 shows a fiber thathas been side-polished in order to have access to the electrodes.
Figure 6.10 Fiber with exposed electrode by side polishing.
6.4.2. Evaluation of dual core electrode filled Mach-Zehnderinterferometer
To evaluate the concept as described in the introduction of section 6.4, a simplifiedexperiment was carried out. A two-hole two-core fiber, similar to that shown in figure6.6, was used. A ~1.5 m long fiber piece was filled along 1 m with molten metalwhich was allowed to solidify at room temperature. Contacts were established byside-polishing the fiber and connected by thin wires. The electrodes did not extend tothe end of the fiber, to prevent electrical breakdown. In the tests, light did not couplebetween cores. Rather, an unfocussed 633-nm wavelength HeNe laser beamilluminated one fiber end, so that both cores were excited. Light propagated as twobeams guided by the twin-cores to the output side, and exiting the fiber, diffracting.After weak collimation, the far-field pattern consisted of interference vertical lines, ifat the fiber output the two cores were disposed along a horizontal plane. A singlefringe was detected with a photomultiplier through a 0.5-mm wide slit. Theexperimental setup is shown in figure 6.11.
68
1 m MZIFiber+
-Det.
HeNe laser
Figure 6.11 Experimental setup with single fiber as the two arms in a MZI(Paper IX).
The fiber was not poled. Rather, the weak intrinsic Kerr nonlinearity was exploited.Figure 6.12 (a) shows a plot of the optical signal measured as a function of voltage,and figure 6.12 (b) illustrates the signal detected as a function of the square of theapplied voltage. The periodicity seen in figure 6.12 (b) confirms the expectedquadratic dependence, and the required voltage for a π-shift is measured to be 2 kV2,which implies Vπ ~1.4 kV.
0.0 0.5 1.0 1.5 2.0 2.50
1
2
3
4
Opt
ical
sig
nal (a
.u.)
Voltage (kV)
0 1 2 3 4 5 60
1
2
3
4
5
Opt
ical
sig
nal (a
.u.)
Voltage squared (kV2)
(a)
(b)
Figure 6.12 Experimental results from the two-core electrode MZI.
69
6.5. References
1. P. C. Becker, N. A. Olsson, J. R. Simpson, Erbium-Doped Fiber Amplifiers-Fundamentals and Technology, Optics and Photonics Series, Acad. Press, 1999.
2. C. Kerbage, A. Hale, A. Yablon, R. S. Windeler, and B. J. Eggleton, “Integratedall-fiber variable attenuator based on hybrid microstructure fiber,” App. Phys.Lett., 79, 3191-3193, (2001)
3. H.G. Limberger, N. Hong Ky, D. M. Costantini, R. P. Salathe,C. A. P. Muller,G.R. Fox, “Efficient miniature fiber-optic tunable filter based on intracore Bragggrating and electrically resistive coating,” Photon. Technol. Lett., 10, 361-363(1998).
4. X. C. Long, R. A. Myers and S. R. J. Brueck, " Measurement of linear electro-optic effect in temperature/electric-field poled optical fibres," Electr. Lett., 30,2162-2163 (1994).
5. P. G. Kazansky, L. Dong and P. St. J. Russel, "High second-order nonlinearities inpoled silicate fibers," Opt. Lett., 19, 701-703 (1994).
6. J. C. Knight, T. A. Birks, P. St. J. Russell, and D. M. Atkin, “All-silica single-mode optical fiber with photonic crystal cladding ,” Opt. Lett., 21, 1547-1549(1996)
7. D. Drolet, R. Vallee, “Dual-core fiber as a tunable directional coupler,” Opt. Lett.,18, 408-410 (1993).
8. Raman Kashyap, Fiber Bragg Gratings, Academic Press (1999).9. A. Iocco, H. G. Limberger,R. P. Salathe, “Bragg grating fast tunable filter,”
Electron. Lett., 33, 2147 -2148 (1997).10. J. Canning,L. G. Carter, “Modal interferometer for in situ measurements of
induced core index change in optical fibres”, Opt. Let., 22, 561-563, (1997).11. Govind P. Agrawal, “Nonlinear fiber optics,” Academic Press (1995).12. D. Wong, W. Xu, S. Fleming, M. Janos and K.-M. Lo, " Frozen-in Electrical Field
in Thermally Poled Fibers," Opt. Fib. Tech., 5, 235-241 (1999).13. B. J. Eggleton, C. Kerbage, P. Westbrook, R. S. Windeler, A. Hale,
"Microstructured optical fiber devices," Opt. Exp., 9, 698-713 (2001).
70
71
7. Description of original work and authorcontribution
Paper I
Formation of thermally stable chemical composition gratings in optical fibers
M. FokineJOSA B, 19, 1759-1765 (2002).
Experimental results and a discussion on possible chemical pathways in the formationof thermally stable chemical composition gratings in optical fibers have beenpresented. Gratings are formed through high temperature treatment of UV exposedhydrogen-loaded fibers. The final refractive index modulation is ascribed to variationsin fluorine concentration attained by periodically increased diffusion of fluorine. Themechanism behind this increase is the formation of mobile hydrogen fluoride fromchemical reactions of fluorine and UV induced hydroxyl, which occur with the spatialperiodicity of the UV pattern. Hydroxyl-assisted increase in fluorine diffusion wasverified using time-of-flight secondary-ion-mass spectroscopy. Ultra stable gratingformation by periodic variation of oxygen concentration through diffusion ofmolecular water is also discussed.
Paper II
Thermal stability of chemical composition gratings in fluorine-germanium-doped silica fibers
M. FokineOpt. Lett., 27, 1016-1018 (2002).
A model based on diffusion of dopants in a periodic structure has been applied todescribe thermal decay of chemical composition gratings in fluorine-germaniumdoped silica fibers. The good agreement between previously reported values and thediffusion coefficients derived here from experiments and model in the temperaturerange 1000 oC and 1200 oC indicate that fluorine diffusion is the main mechanism ofgrating decay. Experimental results also indicate that the presence of phosphoroussignificantly increases the decay rate of chemical composition gratings.
72
Paper III
Growth dynamics of chemical composition gratings in fluorine-doped silicaoptical fibers
M. FokineOptics Letters, 27, 1974-1976 (2002).
The refractive index modulation of chemical composition gratings in fluorine-germanium doped silica fibers has been studied during manufacturing as a function ofthermal treatment. The final grating strength was found to depend strongly on anintermediate annealing step, with an optimum temperature near 600-700 oC, priordevelopment at a fixed temperature of 1000 oC. Low temperature annealing, aimed atremoving remaining hydrogen from the fiber, performed at 100 oC for 20 hrs prior theannealing step significantly increases the final refractive index modulation.
Paper IV
High temperature miniature oven with low thermal gradient for processing fiberBragg gratings
M. FokineReview of Scientific Instruments, 72, 3458-3461 (2001)
An oven for locally heating optical fibers in excess of 1000 oC was designed andevaluated. The dimensions and the design of the oven allow rapid insertion andremoval of the fiber during heating while allowing simultaneous transmissionmeasurements to be performed. The 22 mm long oven has a flat zone at 1000 0C of 16mm with a temperature variation of less than 2 % and 20 mm with a temperaturevariation of 10% and long term stability (hrs) of ±20 oC measured at the center of theoven. Rapid heating of optical fibers to 1000 oC shows a delay of approximately 400ms before the core reaches thermal equilibrium.
Paper V
ToF-SIMS imaging of dopant diffusion in optical fibers
M. Hellsing, M. Fokine, Å. Claesson, L.-E. Nilsson, W. MargulisApplied Surface Science, Article in press (2002).
Applications of optical fibers in telecommunication and sensing are rapidly emergingwhere the fiber properties are related to the controlled addition of dopants such asgermanium, phosphorous, fluorine and erbium. The modern ToF-SIMS instrument,with its high sensitivity and high lateral resolution, has shown to be an excellent toolto directly analyze cross-sections of as-manufactured fibers. The present workdescribes ToF-SIMS imaging of the dopant distribution in fluorine, germanium and
73
rare-earth doped fibers where dopants are confined to a few µm in the core. Theincreased fluorine diffusion in the fluorine doped fibers due to chemical reactionswith hydroxyl groups was examined. This process is utilized in the manufacture ofthermally stable chemical composition fiber Bragg gratings. We were able to produceToF-SIMS elemental images with a lateral resolution around 0.5 µm showing thedetailed distribution of the dopants.
Contribution by the author: Sample preparation and processing. Participated inexperimental planning and experiments, in discussions and in writing the paper.
Paper VI
Large increase in photosensitivity through massive hydroxyl formation
M. Fokine, W. MargulisOpt. Lett., 25, 302-304 (2000).
In this paper we deliberately create a massive amount of hydroxyl radicals (a fewmol%) by heating hydrogen-loaded fibers to 1000 oC for a short time period(seconds), using a high temperature furnace or a CO2-laser. Bragg gratings inscribedin heated sections of fiber show a large increase in photosensitivity at 242 nmcompared with untreated hydrogen loaded fiber. The loss induced by the OH-creationwas measured in a 50 cm long piece of SMF28 fiber to be ~0.02 dB/cm-mol% OHabove1.55 µm, while at 1.30 µm the absorption increase is ~0.03 dB/cm-mol% OH.Such a loss is not prohibitive in many applications of short length Bragg gratings infibers.
Contribution of the author: Background studies of hydrogen interactions with silicabased fibers. Fiber preparation, fiber processing, grating writing and characterization.Participated in discussions and writing of the paper.
Paper VII
Grating formation in pure silica-core fibers
J. Albert, M. Fokine, W. Margulis, Opt. Lett., 27, 809-811 (2002).
Strong grating formation in pure silica core fibers using 193 nm ArF-laser radiation isreported. Unsaturated refractive index changes of ∆n~0.3·10-3 were observed in non-treated fiber and ∆n~0.5·10-3 in fibers with high hydroxyl concentration. Possiblemechanisms of photosensitivity in pure silica core fibers are discussed.
Contribution of the author: Background studies of hydrogen interactions with silicabased fibers. Fiber preparation, fiber processing. Participated in discussions andwriting of the paper.
74
Paper VIII
Mach-Zehnder interferometer using standard telecommunication optical fibre
F. C. Garcia, M. Fokine, W. Margulis, R. KashyapElectronics Letters, 37, 1440-1442 (2001).
It is demonstrated that a single, short-length, standard telecommunications opticalfiber can function as an interferometer when subjected to OH flooding. Theinterferometer is based on the mode beating effect arising from the induced-refractiveindex change causing the fiber to become bi-modal.
Contribution of the author: Participated in experimental planning, in experimentsregarding OH flooding and in component manufacturing.
Paper IX
Integrated fiber Mach-Zehnder interferometer for electro-optic switching
M. Fokine, L.-E. Nilsson, Å. Claesson, D. Berlamont, L. Kjellberg, L.Krummenacher, W. MargulisOpt. Lett., 27, 1643-1545 (2002).
Molten alloys under high pressure are used to provide fibers with long internalelectrodes that are solid at room temperature. An integrated Mach-Zehnderinterferometer was constructed from a twin-core twin-hole fiber that allows applyingan electric field preferentially to one of the cores. Good stability and a switchingvoltage of 1.4 kV was measured with a 1-m long fiber device with a quadratic voltagedependence.
Contribution of the author: Background studies of liquid filling of hole-fibers.Participated in developing the metal filling technique and in fiber polishing forcomponent manufacturing. Participated in component characterization, in fiberdesigns and in discussions regarding the paper.
75
8. Conclusions and outlookThe main topic of my research has been related to chemical composition gratings(CCGs). The research has resulted in significantly better understanding of theunderlying mechanism of the refractive index change, which has been attributed tochanges in the fluorine concentration in the fiber core. The mechanisms regardingthermal stability/decay of CCGs at high temperatures have also been studied. Thedecay of CCGs is governed by diffusion and a model was developed to determinegrating diffusion parameters. These experimentally determined diffusion parameterscan subsequently be used to predict the thermal stability under different operatingconditions. Studies on the dynamics of CCG formation have resulted in a proposedmodel for some of the mechanisms that occur during grating fabrication, and a“recipe” that can be used to manufacture these gratings in a reproducible manner hasbeen provided.
Although research results presented in this thesis explain several properties of CCGs,there are still several issues that need to be addressed e.g. improving the maximumattainable refractive index modulation. A typical, non-optimized value of therefractive index modulation of manufactured CCGs is ∆n~1·10-4. For example,relating the concentration of UV induced hydroxyl groups and the final refractiveindex of the CCG should be valuable to maximize the refractive index change.
The above conclusions are related to CCGs where the modulated dopant is fluorine.Further studies need to be performed on the ultra-stable gratings, which have beenattributed to oxygen removal due to water diffusion. Exposing these gratings totemperatures as high as 1100 oC do not result in any measurable decay of therefractive index. Several questions need to be answered regarding these gratings.What is the underlying mechanism? If these gratings are CCGs, it should be possibleto determine corresponding diffusion parameters. If oxygen is the modulated dopant,is it possible to erase the grating by heating the grating in an oxygen atmosphere?What is the limiting operating temperature for sensor applications? What is thelimiting operating temperature of the fiber itself? Better understanding of the differentprocesses that occur during the development procedure should be sought after, e.g. thestructural relaxation dynamics of the core and cladding, chemical reactions and defectformation and annihilation. These issues are significant, not only from amanufacturers viewpoint, but from a scientific perspective, providing additionalinsight into glass physics and glass chemistry.
The CCGs also opens doors for new applications and devices. If it is possible to writeCCGs in pure silica core fibers with fluorine depressed cladding, these gratings mightfind applications as sensors for nuclear power plants or space applications, where thistype of fiber is generally used. It might also be possible to write gratings with Braggwavelengths in the in UV region. This could open the possibility for producing short-wavelength (UV) fibers lasers. Due to their high-temperature stability, the interest inCCGs has naturally been focused on high-temperature sensors. The advantage overother types of gratings, which may survive similar temperatures, is that the decaybehavior of CCGs is well characterized, therefore increasing the reliability of thegrating sensor. One crucial area, which will require a large research effort, is toidentify or develop coatings that can survive high operating temperatures, in order toprovide mechanical protection to the fiber.
76
CCGs has played a central role in several projects, including a European projectregarding health monitoring of industrial structures. Field trials have also been madewith the involvement of the oil-industry and regarding process control duringaluminum casting (Paper A).
Besides CCGs, a method was studied to make low-photosensitivity fibersphotosensitive. The method referred to as OH flooding has been applied to standardtelecommunications fiber resulting in fibers with extreme photosensitivity. Themethod was also applied to pure-silica core fibers, resulting in the first report ofstrong grating formation in undoped silica core fibers. Experiments also showed thatgrating formation was possible even without hydrogen treatment. Due to the extremephotosensitivity caused by OH flooding when applied to standard telecommunicationfibers, this method may not be suitable for the production of fiber Bragg grating withcomplex structures, as the growth in refractive index could be difficult control.However, for fibers with extremely low photosensitivity OH flooding could be anoption. In addition, further research needs to be performed regarding the thermalstability in gratings manufactured using OH flooding. The results of the gratingformation in pure-silica fibers might have bigger implications. For example, studies ofthe mechanisms of photosensitivity in pure silica-core fibers can be performedsystematically without the interference of other dopants as their possible role inphotosensitivity can be excluded. The literature review on material aspects onphotosensitivity presented in this thesis, and the fact that strong gratings were writtenin pure-silica core fibers demonstrates the difficulty of evaluating the contribution thata specific dopant has on photosensitivity. The possibility of writing gratings in pure-silica also opens up new possibilities for novel fiber components e.g. gratings writtenin holey fibers or photonic bandgap fibers made only of pure silica glass or formanufacturing short wavelength lasers.
Additional research was focused on manufacturing fiber component resulting in twotypes of single-fiber Mach-Zehnder interferometers (MZI's). One MZI was madeusing OH flooding as a technique to design a large and controllable refractive indexchange of the core, causing a section of a standard telecommunications fiber tobecome bi-modal. The second MZI was based on a two-core fiber integrated withinternal electrodes to enable active control of the fiber output. As the two arms of theMach-Zehnder interferometers are integrated into the same piece of fiber, thesensitivity to external perturbations is decreased significantly. The Kerr effect wasutilized for switching in the MZI with internal electrodes, but to further reduce thenecessary switching voltage, thermal poling should be utilized.
Although several questions have been answered during the course of this thesis, morequestions have been raised than answers found. To finalize the conclusion, thefollowing speculative questions are presented. Can OH flooding with a large amountof dipolar species be used for thermal poling of glass? How is the hydroxyl-assisteddiffusion of fluorine affected by an applied electric field? Will a diffusing HFmolecule react in a preferred direction with the glass matrix if under an appliedelectric field? If so, will this break the inversion symmetry of glass?
77
Reproduction of Papers I through IX
