Open Collections

UBC Undergraduate Research

Wavelength tunable femtosecond fiber laser : results and recommendations Pa, Jack; Young, Robert Apr 4, 2010

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
52966-Pa_Jack_et_al_APSC_459_2011.pdf [ 1.5MB ]
Metadata
JSON: 52966-1.0074457.json
JSON-LD: 52966-1.0074457-ld.json
RDF/XML (Pretty): 52966-1.0074457-rdf.xml
RDF/JSON: 52966-1.0074457-rdf.json
Turtle: 52966-1.0074457-turtle.txt
N-Triples: 52966-1.0074457-rdf-ntriples.txt
Original Record: 52966-1.0074457-source.json
Full Text
52966-1.0074457-fulltext.txt
Citation
52966-1.0074457.ris

Full Text

Wavelength Tunable Femtosecond Fiber Laser Results and Recommendations  Jack Pa Robert Young  Project Sponsor Dr. David Jones  Applied Science 459 Engineering Physics The University of British Columbia April 4, 2010  Group 1111  P a g e | ii  Executive Summary The research objective of this project is to construct and characterize a wavelength tunable fiber laser. The set up takes advantage of a grating compressor to shorten the width of the generated pulses, and a photonic crystal fiber to shift the output wavelength through the soliton self-frequency shift (SSFS) effect. The quantitative objective is to construct a system that outputs roughly 5mW pulses with wavelengths between 1030 and 1500 nm with pulse widths around 200fs and a pulse rate of 100MHz. To obtain these desired specifications, pulse shaping optics were used to alter the output parameters detailed below. First, a pre amplifier will be used to amplify incoming pulses to the required pulse power. However, since the diode for the pre-amplifier malfunctioned, a laser oscillator with similar parameters was used instead as a temporary replacement. As a result the pre-amplifier could not be fully validated for performance. A grating compressor made up of two transmission gratings will be used to compress the incoming pulses to reach the 200fs range. Finally, the output is then coupled to a photonic crystal fiber (PCF) where the wavelength of the beam is shifted to achieve wavelength tunability. As of the original intended deadline of this report, two of the three major components mentioned in the above paragraph has been completed and fully characterized. The grating compressor has been aligned and adjusted to obtain a confirmed pulse width of approximately 200fs (± 2.6fs) using the autocorrelator device. A minor wavelength peak shift was observed in the two tested fiber ranging from approximately1030 to 1067nm.The pre-amplifier has been completed and currently waiting for a replacement pump diode before having it revalidated. Based on the current status and results, while the pulse shaping optics and fiber coupling apparatus has been completed (some not yet fully characterized), the required tunable wavelength range was not achieved. It is recommended that for future work on the system, the pre-amplifier be completed and used to further increase the incident intensity into the PCF, the core size and length of the PCF to be made smaller and longer respectively to increase the nonlinear effects, and the system be further optimized through alignment to minimize avoidable power losses. The pulse width could also be decreased further to increase the nonlinear effect.  P a g e | iii  Contents Executive Summary......................................................................................................... ii Figures ............................................................................................................................ v Tables ........................................................................................................................... vii 1  2  Introduction............................................................................................................... 1 1.1  History and Background .................................................................................... 1  1.2  Project Scope and Limitations ........................................................................... 2  Discussion ................................................................................................................ 4 2.1  Theory ............................................................................................................... 4  2.1.1  Fiber Pre-Amplifier ..................................................................................... 4  2.1.2  Grating Pulse Compressor ......................................................................... 6  2.1.3  Nonlinear Frequency Shift and Soliton Characteristics ............................... 8  2.2  Methods .......................................................................................................... 11  2.2.1  Input Polarization Dependence................................................................. 12  2.2.2  Input Optical Mode Dependence .............................................................. 12  2.2.3  Fiber Pre-amplifier .................................................................................... 14  2.2.4  Grating Pulse Compressor and Optical Attenuator ................................... 16  2.2.5  Photonic Crystal Fiber Coupling ............................................................... 18  2.2.6  Frequency Shift Characterization at PCF Output ...................................... 20  2.3  Experimental Equipment ................................................................................. 21  2.4  Issues and Current Workarounds .................................................................... 22  2.5  Results ............................................................................................................ 24  2.5.1  Fiber Pre-amplifier .................................................................................... 24  2.5.2  Grating Compressor Performance ............................................................ 24  2.5.3  Coupling into and Characterization of Photonic Crystal Fiber ................... 25  2.6  Results Discussion .......................................................................................... 28  3  Conclusions ............................................................................................................ 29  4  Project Deliverables ................................................................................................ 30 4.1  List of Deliverables .......................................................................................... 31  4.2  Financial Summary.......................................................................................... 31  P a g e | iv 5  Recommendations .................................................................................................. 32  Appendix A – Pre Amplifier Schematic and Losses ....................................................... 33 Appendix B – Enclosure for Grating Compressor .......................................................... 35 Appendix C – Diagram of Fiber Clamp Adapter Mount .................................................. 36 Appendix D – Bill of Materials ........................................................................................ 37 References .................................................................................................................... 39  Page |v  Figures Figure 1 – Energy state diagram showing a three level type dopant such as ytterbium. After electrons are pumped to the E3 state, there is spontaneous emission to the E2 state. The wanted amplified emission is from E2 to E1 (Agrawal 2001). ......................... 4 Figure 2 - Schematic of fiber preamplifier. The coupling lenses (CL) interface the fiber with free space. Wavelength-division multiplexers (WDM) isolate the pump laser diode (LD) from back reflections and emission from the Yb-doped fiber (YDF). ....................... 5 Figure 3 – Basic schematic of a grating pair, showing the light’s path length lx, incident and refracted angles θi and θr, and grating separation d0. (Agrawal, 2001). .................... 6 Figure 4 – A typical reflective diffraction grating compressor implementation showing how the different frequency components are diffracted at different angles. This also shows how a single pass through the two gratings does not maintain spatial dimensions of the pulse (Lam, 2010). ................................................................................................ 8 Figure 5 - A scanning electron microscope image of the fine microstructure of a solid core PCF, showing the overall structure with the small solid core surrounded by holes (left) and a close up of the air holes (right). ..................................................................... 9 Figure 6 – Two different pulse chirps caused by different nonlinear effects. When the effects balance out, the pulse remains undistorted. .......................................................10 Figure 7 – Overview system schematic showing all major components, including the laser oscillator, the pre-amplifier, the grating compressor, the optical attenuator, and the photonic crystal fiber......................................................................................................11 Figure 8 – Incoming beam meeting a quarter and half wave plate as well as a polarizing beam splitter intending to polarize the incoming beam. .................................................12 Figure 9 - Cylindrical waveguide TEM mode patterns. Notice the difference between the 01 and the 00 mode (Wikipedia). ...................................................................................13 Figure 10 – Coloured beam profile from the profiling camera showing a 00 mode at the output of the photonic crystal fiber. ................................................................................14 Figure 11 – Ericsson fiber cleaving (left) and fiber splicing (right) machines. .................15 Figure 12 - Finished pre-amplifier with fiber and loops placed in a foam block for portability and protection................................................................................................16 Figure 13 – Overhead view of the Grating Compressor System with Plexiglas enclosure off ..................................................................................................................................17  P a g e | vi Figure 14 - Overview of the input polarization configuration optics, the grating compressor, and output optics including the attenuator (the pre-amplifier was left out of the image as it was not being used due to malfunctions). ..............................................18 Figure 15 – PCF coupling apparatus using a very short focal length aspherical lens mounted onto a flexure XYZ adjustment stage in order to steer the beam into the micron sized core diameter of the PCF. ....................................................................................19 Figure 16 - Femtochrome autocorrelator showing the input side (input hole at upper center of device) and various adjustment knobs for alignment. ......................................21 Figure 17 – Simple photodiode based optical power meter (detector is under the black aluminum foil). ...............................................................................................................22 Figure 18 - The failed pump laser diode attached to its heat sink and driver. .................23 Figure 19 - Spectral graph of the SC-5.0-1040 PCF coupled using a 2.0 mm focal length. Three output intensities are shown to illustrate the wavelength shifting of the smaller peak on the right side. ...................................................................................................26 Figure 20 - Spectral graph of the NL-1050-ZERO-2 PCF coupled using a 2.0 mm focal length. Three output intensities are shown to illustrate the wavelength shifting of the smaller peak on the right side ........................................................................................27 Figure 21 – SolidWorks schematic of the pre-amplifier fiber routing based on dimensions collected from a picture taken of the assembled pre-amplifier. ......................................33 Figure 22 – The assembled pre-amplfiier, showing the fiber routings in the foam block. 33 Figure 23 – The designed Plexiglas enclosure for the grating compressor, showing components inside.........................................................................................................35 Figure 24 – The resulting product of the design in the figure above. The enclosure protects the gratings from dust accumulation and damage. ...........................................35 Figure 25 – SolidWorks dimensioned and toleranced drawing of the adapter block. ......36 Figure 26 – Resulting part in use in the PCF coupling apparatus...................................36  P a g e | vii  Tables Table 1 – Focal length and depth of field calculations for aspheric lens selection for each PCF ...............................................................................................................................20 Table 2 - Grating compressor measurements showing some measurements taken to determine the compressed pulse width as a function of grating separation distance .....24 Table 3 - Deliverables and current status of completion at the end of the project timeline ......................................................................................................................................31 Table 4 - Estimated costs for parts fabricated during the project. These parts were all fabricated by the Berhnard Zender of the project lab. ....................................................31 Table 5 – Table of fiber losses at the connections between the different fiber components in the pre-amplfiier.....................................................................................34 Table 6 - Approximate bill of materials for all systems implemented ..............................37  Page |1  1 Introduction This project, sponsored by Dr. David Jones of the Physics & Astronomy Dept at the University of British Columbia, will be to setup a wavelength tunable Ytterbium (Yb)doped femtosecond fiber laser system. The purpose of such a project, and of such a device, is to ultimately use the tunable femtosecond fiber laser system in conjunction with a fixed wavelength femtosecond fiber laser to enable difference frequency generation (DFG) to generate electromagnetic waves in the mid-infrared (2um) region to study the energy absorption of different vibrational modes of molecules, essentially establishing a ‘fingerprint’ for those molecules.  1.1 History and Background Wavelength tunable femtosecond fiber lasers are not entirely new devices as they have been built and studied extensively in the past. Two recent notable papers, one by Takayanagi et al. in 2006 and one other by Sidorov-Biryukov et al. in 2010, detail tunable Ytterbium-doped laser systems very similar to the requirements of the proposed laser system. Both have used a mode locked femtosecond Ytterbium-doped fiber laser and coupled the output to a photonic crystal fiber to produce a tunable range of 1050nm to 1690nm (Takayanagi et al.) and 1060 to 1400nm (Sidorov-Biryukov et al.). More specifically, Takayanagi et al. produced a system outputting 66fs pulses at a repetition rate of 41.3MHz with a pulse energy of 2nJ and a peak power of 20kW while SidorovBiryukov et al. managed to output 100fs pulses at 70MHz with pulse energies of 7nJ. Both papers show a detailed correlation between the incident power to the PCF and the peak output wavelength with very similar results. Since the output is a pulse, a superposition of many frequency components, the spectral measurement is commonly based on the peak, most prominent, wavelength. However, being laboratory research tools, their systems are unique and not commercialized. Commercialized femtosecond laser systems do exist, such as Toptica’s FemtoFiber pro UCP: Ultra Compressed Pulse Femtosecond Fiber Laser with a wavelength tuning range from 980nm to 1400nm and 30mW pulses less than 25fs wide and repeated at 80MHz as described on their product page. This device is one of few available that have such a wide tunable range. The range that is suggested in this project, 1030nm to 1500nm, is not possible with any currently existing commercial products. More specifically, widely tunable systems have been used in many instances in the past for difference frequency generation, and a simple search shows few of these implementations were used for mid-infrared spectroscopy. For example, in 2006  Page |2 Weidong Chen et al. published their results in developing a widely tunable continuous wave laser for trace gas spectroscopy via difference frequency generation. In terms of pulsed laser systems, Hisanao Hazama et al. in 2007 published on their development of a tabletop tunable mid-infrared wavelength pulsed laser via difference frequency generation as a smaller and less expensive alternative to a tunable mid-infrared free electron laser. Again, these are laboratory systems and are not available commercially. Since many femtosecond fiber lasers were built before by the project sponsor and his graduate students, this is merely an extension of that by coupling these high power pulses into a highly nonlinear medium (PCF) to cause SSFS and thus wavelength tunable. Therefore, much of the theoretical work developed to help define the proposed solution has been fashioned by the project sponsor into a ‘recipe’. Furthermore, literature research suggests that highly nonlinear photonic crystal fiber is the primary method in enabling wavelength tuning of femtosecond lasers. This project will also incorporate some design ideas for pulse refinement mechanisms from Matt Lam in his M.Sc thesis titled Chirped pulse amplifier and passive enhancement cavity for generation of extreme ultraviolet light.  1.2 Project Scope and Limitations The original scope called for the design and/or implementation of the oscillator cavity, the pulse shaping optics, and the coupling and characterization of the wavelength shift in the photonic crystal fiber. However, due to the project timeframe, the scope of this project is reduced to the implementation of the pulse shaping optics – the grating compressor and pre amplifier, as well as the coupling and characterization of wavelength tunable range through the photonic crystal fiber. The project sponsor will be responsible for the laser cavity and will not be discussed in this report. The scope also does not consider anything after the completion and characterization of the setup such as difference frequency generation and molecular spectroscopy. The key areas addressed in the scope of this project are the characteristics of the pulses at the output of the laser. The requirements for the output pulses are 200fs wide pulses at roughly 5mW intensity. A pulse repetition rate of approximately 100MHz is expected to be generated from the laser oscillator and should remain unchanged at the output. Finally, the last issue addressed will be to enable the laser to have a tunable wavelength from 1030nm to 1500nm. The pulse width is modulated with a grating compressor while the pulse intensity is modulated with a simple Ytterbium-doped fiber pre-amplifier. The crucial wavelength tunability will be realized by the soliton self-frequency shift (SSFS) effect due to the stimulated Raman scattering (SRS) effect in highly nonlinear photonic crystal fiber (PCF). The amount of wavelength shift will depend on the incident intensity modulated by an optical attenuator. To meet the requirements for the system, the acquired range of tunable wavelengths, pulse widths and pulse intensities must be  Page |3 confirmed quantitatively to be within the specification described via a spectrum analyzer, interferometric autocorrelator, and optical power meter. This report will first present some background information regarding the physical theory of the pulse shaping optics and how frequency shift can be obtained through highly nonlinear fiber such as photonic crystal fiber. Then, this report will discuss in detail the implementation methods for each set of optics and for the coupling into the photonic crystal fiber. Finally, this report will present the findings and results from measurements, leading to the proposed recommendations and future work of the project. While this report will discuss the theory and methods in implementing the pre-amplifier, this report will not cover the performance results and characterization of the pre-amplifier due to ongoing issues with the diode laser used to pump the pre-amplifier’s ytterbiumdoped fiber. This limits the maximum possible power of the pulses going into the remainder of the system. Also, due to issues in implementing the oscillator cavity intended to supply this system, the system has been implementing using a commercially available laser with approximately 200mW intensity and approximately 130fs. Due to this, the pulse compressor will be implemented with different parameters, and the incident intensity will change when the pre-amplifier is coupled to the system. The nonlinear frequency shift effect will also change as the input intensity is changed.  Page |4  2 Discussion This section will focus on the development of the wavelength tunable femtosecond fiber laser system, focusing on the pulse shaping optics and the photonic crystal fiber behaviour as mentioned in the introduction. More specifically, this section will begin with the technical theory for each component, and then will follow with the implementation details. Here, methods and equipment used for testing and validation of the optics will be discussed. Finally, results obtained for each pulse shaping system will be presented and analyzed, including current sources of error.  2.1 Theory This section and its subsections will detail the technical concepts into the operation and design of the fiber pre-amplifier, the grating pulse compressor, and the principles in inducing pulse frequency shift due to the nonlinear properties of the photonic crystal fiber. The theory and design in this section is primarily adapted from principles discussed in Gavind P. Agrawal’s Nonlinear Fiber Optics as well as Matthew Lam’s 2010 master’s thesis. 2.1.1 Fiber Pre-Amplifier The basic operating principles of a fiber pre-amplifier is to increase the intensity of the incident light signal through stimulated emission of a lasing medium such as rare earth metal doped fibers (Agrawal, 2001). This principle is similar to that of a typical laser where a population inversion is created to cause stimulated emission, but using doped fiber instead of crystals or gas for the lasing medium (Agrawal, 2001). The gain is based on the energy levels of the medium (Figure 1) and is mathematically given as proportional to the atomic densities of the two energy states E2 and E1. Different rare earth metal dopants provide different gain characteristics, operating wavelengths (Agrawal, 2001).  Figure 1 – Energy state diagram showing a three level type dopant such as ytterbium. After electrons are pumped to the E3 state, there is spontaneous emission to the E2 state. The wanted amplified emission is from E2 to E1 (Agrawal 2001).  Page |5  While pumping the doped fiber produces amplified stimulated emission (ASE), it also produces amplified spontaneous emission in the same process where electrons of some ytterbium ions simply decay spontaneously from the higher to lower state in an incoherent fashion in both forward and backward directions in the amplifier. This increases the signal noise and also limits the maximum gain that can be achieved with the medium (Agrawal, 2001). This can be minimized with good alignment of the signal source into the pre-amplifier to better seed the population inversion. The requirement of a pre-amplifier in this system is to increase the amplitude of the input pulse signal from the linear cavity (the primary oscillator) to compensate for losses in the grating compressor, attenuator, the pre amplifier itself, as well as any coupling losses from fiber to free space and vice versa. In addition, it will be explained in a later section that the increased pulse intensity is beneficial for the nonlinear frequency shifting effect in the photonic crystal fiber. The implemented system uses a 980nm, 500mW diode laser (JDSU S30 series) to pump approximately 50cm of Ytterbium doped fiber in order to create the population inversion necessary for the emission of amplified pulses at 1030nm. Since the amplification requirements for this application is similar compared to a design Matthew Lam used in his laser system, the same design was adopted (Figure 2).  Figure 2 - Schematic of fiber preamplifier. The coupling lenses (CL) interface the fiber with free space. Wavelength-division multiplexers (WDM) isolate the pump laser diode (LD) from back reflections and emission from the Yb-doped fiber (YDF).  As shown in Figure 1, the signal pulse train is coupled with the pump through two multiplexers (the first intended to prevent back reflections from damaging the pump diode while the second actually couples the two sources together) into the Ytterbium doped fiber. Here the output is coupled into free space using a fiber collimator.  Page |6 2.1.2 Grating Pulse Compressor The grating compressor (Figure 3) takes advantages of two parallel gratings that act as a dispersive delay line and ultimately induces anomalous dispersion and compresses the incoming pulses as it separates and recombines the different frequency components (Agrawal, 2001).  Figure 3 – Basic schematic of a grating pair, showing the light’s path length lx, incident and refracted angles θi and θr, and grating separation d0. (Agrawal, 2001).  The grating pair can be described geometrically via 2πc sin θr = − sin θi ωΛ  Equation 1  where the diffracted angle θr is dependent on the frequency (ω), the grating period (line spacing of grating) Λ, and c is the speed of light. The time delay due to the different path lengths of the different frequency components of incident light can be determined geometrically from 𝜕𝜙𝑐 𝑙(𝜔) = 𝑡𝑑 (𝜔) = 𝑐 𝜕𝜔  Equation 2  where ϕc is the phase shift of frequency component ω and l(ω) is the total path length given by 𝑑0 cos(𝜃𝑟 − 𝜃𝑖 ) 𝑙(𝜔) = 𝑙1 + 𝑙2 = cos(𝜃𝑟 )  Equation 3  where d0 is the parallel grating separation as shown in Figure 3. The phase change ϕc(ω) can be Taylor expanded around ω0 because the phase shift of a certain ω is to be determined, resulting in  Page |7 𝜙𝑐 (𝜔) = 𝜙0 + 𝑡𝑐 (𝜔 − 𝜔0 ) − 𝑎𝑐 (𝜔 − 𝜔0 )2 + 𝑏𝑐 (𝜔 − 𝜔0 )3 +. . .  Equation 4  where tc is a constant delay, and ac and bc are coefficients to take account of the GVD effects of the gratings (Agarwal, 2001). The coefficients ac and bc can be obtained by Equation 1-Equation 4 and results in 4𝜋 2 𝑐𝑏0 𝑎𝑐 = 3 2 𝜔0 Λ cos2 (𝜃𝑟0 )  Equation 5  𝑏𝑐 =  4𝜋 2 𝑐𝑏0 (1 + 𝑠𝑖𝑛𝜃𝑖 + 𝑠𝑖𝑛𝜃𝑟0 ) 𝜔04 Λ cos4(𝜃𝑟0 )  Equation 6  where θr0 is the result of setting ω=ω0 in Equation 1 and b0 is the center-to-center spacing between the gratings. b0 is related to the grating to grating distance by 𝑏0 = 𝑑0 𝑠𝑒𝑐𝜃𝑟0  Equation 7  If the spectral width Δω≪ω0, then the cubic and higher order terms in Equation 4 can be removed, and if the constant and linear terms are ignored, the phase shift ϕc results to the remaining term ac(ω –ω0)2. Since ac is positive from Equation 5Equation 4, then ϕc(ω) is negative, corresponding to negative phase change, or anomalous GVD. As described in the theory of grating compression in the methods section, anomalous GVD would compress the pulse as opposed to normal GVD which broadens the pulse. The effective GVD parameter can be determined via 2𝑎𝑐 𝑒𝑓𝑓 𝛽2 = − 𝑏0  Equation 8  To consider the effects of all these relations, the overall relationship between the amount of dispersion and the separation distance 𝜆 𝜆 2 𝑏 � � � 2 � 𝛽2 = − 2 2𝜋𝑐 𝑑 cos (𝜃𝑑 )  Equation 9  where b is the separation distance between the two gratings, θd is the angle of incidence, d is the grating period, and c is the speed of light. Then the dispersion required can be approximated using this relation from a Newport Application Note 𝛽2 2 𝑡𝑐 = �1 + (16 ln(2))2 � � 𝑡𝑜 𝑡𝑜  Equation 10  Page |8 where tc is the compressed pulse width at full width half max and to is the output pulse width, also at full width half max (Newport, 2006). While passing the pulse train through the two gratings is successful in compressing the pulses in the temporal domain, the spatial domain becomes stretched in the direction of dispersion and dimensionally different than the input pulse (Figure 4). To compensate for this a simple solution mentioned in Agrawal (2001) to reflect and redirect the pulse train back through the grating system to recollimate the beam back into its original spatial dimensions. This method also has the advantage of doubling the amount of anomalous dispersion, which benefits to reducing the grating separation by 50 percent.  Figure 4 – A typical reflective diffraction grating compressor implementation showing how the different frequency components are diffracted at different angles. This also shows how a single pass through the two gratings does not maintain spatial dimensions of the pulse (Lam, 2010).  2.1.3 Nonlinear Frequency Shift and Soliton Characteristics The basis of this project is to obtain a pulse train with a tunable peak wavelength from 1030nm to 1500nm. This feature can be accomplished using nonlinear optics such as highly nonlinear photonic crystal fiber. The purpose of this section is to provide sufficient background information to understand the physical concepts at work to produce this wavelength shifting effect. This laser system will use the nonlinearity in photonic crystal fibers to cause the wavelength shifting effect. Figure 5 shows a cross section of a typical solid core silica based photonic crystal fiber (PCF). Due to the smaller effective core area of a PCF compared to normal fibers, the nonlinear effect in the fiber will be enhanced. The silica in the fiber will act as the Raman active medium for the stimulated Raman scattering of the input pulses. When light scatters from an atom/molecule, most of it is elastic where the scattered photons and incident photons are of the same energy and thus  Page |9 wavelength. Raman scattering is defined as the small portion of the scattered photons is of inelastic which means the scattered photons are of different energy thus different wavelength.  Figure 5 - A scanning electron microscope image of the fine microstructure of a solid core PCF, showing the overall structure with the small solid core surrounded by holes (left) and a close up of the air holes (right).  The Raman scattering effect is highly dependent on the incident intensity of the pulse train. By taking advantage of the enhanced nonlinearity of the PCF, the wavelength shift effect can be more prominent, where the longer wavelength portions of the pulse will experience amplification at the expense of shorter wavelength portions (Agrawal, 2001). This shifting effect of the wavelength of soliton pulses is called the soliton self frequency shift. The characterization of this gain with respect to the wavelength can be used obtain the Raman-Gain Spectrum. However, for the purposes of this project, the spectral output as the pulse frequencies shift is of greater importance. The amount of wavelength shift is dependent on the input pulse intensity as well as the length of the fiber, thus it is advantageous to use PCF with smaller core sizes, longer PCFs, or higher input intensities (Agrawal, 2001). The amount of shift can then be adjusted or modulated with a simple optical attenuator. As mentioned in the previous paragraphs, the Gaussian pulse train from the laser cavity and through the pulse shaping optics changes into soliton pulses when in the photonic crystal fiber. Usually, the spectral shape of ultra short pulses propagating through a medium is distorted due to the Kerr effect as well as dispersion (GVD). However, under the right conditions, the effects of Kerr non linearity and dispersion can cancel each other out (Figure 6). This allows the pulse to propagate through long paths without being distorted. This is the characteristic of a soliton pulse.  P a g e | 10  Figure 6 – Two different pulse chirps caused by different nonlinear effects. When the effects balance out, the pulse remains undistorted.  The Kerr effect can be characterized with the change of the refractive index in proportion to I (optical intensity) and n2 (nonlinear coefficient) as shown by Δ𝑛 = 𝑛2 𝐼  Equation 11  As for GVD, the details have been discussed conceptually in methods and theory for the grating compressors, and to greater detail in appendix B where the technical details of the grating compressors are discussed. Further details on GVD and dispersion in general can be found in Govind P. Agrawal’s Nonlinear Fiber Optics 2nd Edition. There are a few conditions that need to be met in order for pulses to be classified as solitons. With a positive n2 for the Kerr effect, which is most of the time, the chromatic dispersion needs to be anomalous. The temporal shape of the pulse also needs to be of an unchirped (chirp referring to a signal where the frequency increases or decreases with time) sech2 shaped pulse assuming that only the second order dispersions apply. This assumption is valid for pulses greater than 30fs long. This is described mathematically via 𝑃𝑝 𝑡 𝑃(𝑡) = 𝑃𝑝 sech2 � � = 𝑡 𝜏 cosh2 � � 𝜏  Equation 12  Finally, the pulse energy Ep and soliton pulse duration τ (tau) needs to meet the following equation 𝐸𝑝 =  2|𝛽2 | |𝛾|𝜏  P a g e | 11 Equation 13  where β2 is the second order dispersion GVD parameter and γ is the SPM coefficient due to the Kerr effect. To summarize, the high nonlinearity of photonic crystal fibres due to the micron sized core created by designing a hollow microstructure around the core allows for a higher overall peak intensity in the fiber. This correlates to wavelength shifting of the pulses due to the stimulated Raman scattering effect. Due to the characteristics of the PCF, Gaussian pulses at the input are converted to soliton pulses in the PCF.  2.2 Methods In the proposed wavelength tunable femtosecond fiber laser system, the laser oscillator outputs short pulses which will be slightly amplified by the pre-amplifier and then compressed by the grating compressor. There will be an optical attenuator between the grating compressor and the PCF to adjust the pulse intensity. The compressed pulse that enters the Photonic Crystal Fiber will act as a pump for Soliton Self Frequency Shift which will create pulses with varying wavelengths. The wavelength of the output pulses will depend on the pulse intensity of the compressed pulse entering the PCF. The relationship between the wavelength and input pulse intensity will be characterized. A diagram of the overall system can be seen in Figure 7.  Figure 7 – Overview system schematic showing all major components, including the laser oscillator, the preamplifier, the grating compressor, the optical attenuator, and the photonic crystal fiber.  P a g e | 12  2.2.1 Input Polarization Dependence In the proposal, the importance of the input polarization was not discussed and was not regarded as an important issue until the implementation of the grating compressor system was initiated. Due to the elliptically polarized characteristics of the beam from the oscillator cavity and through the pre-amplifier, the pulse train must then be linearly polarized before it enters the grating compressor system. The gratings are specified to perform at a certain linear polarization, P-polarized, and other polarizations would translate to losses and performance issues in the grating compressor system. Figure 8 shows the optics implemented to obtain linear polarization. This includes a quarter wave plate, half wave plate, and polarizing beam splitter (PBS); the first two allows for the rotation and linearization of the polarization while the latter removes the unwanted polarization from the system. Since the optical attenuator system later in the system uses the linear polarization to obtain a maximum range of pulse intensities, this linear polarization step will enable the attenuator to achieve zero to maximum attenuation. Quarter Wave Plate  Beam Splitter Half Wave Plate  Figure 8 – Incoming beam meeting a quarter and half wave plate as well as a polarizing beam splitter intending to polarize the incoming beam.  To achieve the correct polarization, the quarter wave plate and half wave plate are rotated to rotate and linearize the pulse train from the pre-amplifier. Since the PBS picks off the unwanted polarization, the two wave plates must be rotated to obtain maximum power at the through side of the PBS (where the grating compressor is located). The output power from the PBS can be measured using a simple power meter.  2.2.2 Input Optical Mode Dependence Optical modes correspond to the patterns of power density at the output of a waveguide produced by the dominant standing wave that forms inside the waveguide (Figure 9).  P a g e | 13 Optical fibers can be specified to support single or multiple modes, meaning certain modes can propagate while other standing waves will simply decay. The optical mode of the input pulse train is also of great importance since the photonic crystal fibers proposed for this system thus far only support a single mode, the 00 mode. Due to the current apparatus where the temporary source pulse train from a commercial laser is propagated through a multiple mode fiber to this system, the resulting mode at the input of the system can be different depending on how the pulse train is coupled into the long fiber.  Figure 9 - Cylindrical waveguide TEM mode patterns. Notice the difference between the 01 and the 00 mode (Wikipedia).  Initial measurements showed a 01 mode at the input to the PCF. This division in power density caused difficulty in coupling to the PCF’s micron sized core. Two mirrors at the input to the long fiber were used to translate and pitch the pulse train to achieve a 00 mode. Measurements to confirm a 00 mode (Figure 10) was done using a beam profiling camera connected to a computer using the BeamScan software by Coherent. The software can also measure various parameters of the beam profile such as effective diameter and average power density. To obtain meaningful profiles, the camera must not be saturated, thus only power on the order of microwatts should be incident to the camera. The half wave plate at the input described in the previous section was used in conjunction with another half wave plate at the pulse train source to attenuate the beam to sufficient power levels.  P a g e | 14  Figure 10 – Coloured beam profile from the profiling camera showing a 00 mode at the output of the photonic crystal fiber.  2.2.3 Fiber Pre-amplifier The fiber pre-amplifier, as described in the theory section, uses fiber and a diode laser to pump the medium to obtain an amplified output of the system. More specifically, the designed system based on the pre-amplifier created by Matthew Lam (2010), uses approximately 50cm of Ytterbium doped silica fiber was used as the gain medium, a 500 mW laser diode from JDSU Uniphase (S30 Series) was used to act as the pump, and two 980/1030 nm wavelength division multiplexers were used to couple the signal and pump sources together prior to the ytterbium doped fiber (see Figure 2 for a schematic of the overall pre-amplifier system). The pump diode is fiber connected and is mounted on a modified thermoelectric heat sink and current driver. The multiplexers, two wavelength division multiplexers are situated as seen in the diagram below. One of the wavelength division multiplexers is to combine the 1030nm signal source with the 980nm pumping laser diode to the gain fiber. The other multiplexer is to isolate the 1030 nm back reflection from going into the pump laser diode and damaging the device. Splicing of the fibers together requires that first the fiber ends are stripped of the acrylate protective cladding and then cleaned using methanol and optical wipes to ensure that no debris is left on the bare fiber itself. Then the fiber ends need to be cleaved to ensure a perfect flat end for effective fusing later. An Ericsson branded fiber cleaver (Figure 11)  P a g e | 15 holds down the fiber at two places and provides tension while an ultrasonic blade advances and creates a micro crack on one side of the fiber. Due to the tension, the crack propagates through the brittle material and the result is a flat ended cleave ideal for fusing. The fusing of the fiber ends was done using an Ericsson branded arc fuser (Figure 11). This device accepts two fiber ends and automates the entire process of aligning and fusing the two fiber ends together. The machine also approximates the losses through that splice and provides an image for quality checking of the splice. The losses for each splice in this system can be seen in Appendix A – Pre Amplifier Schematic and Losses.  Figure 11 – Ericsson fiber cleaving (left) and fiber splicing (right) machines.  The spliced preamplifier results in loops of fiber, thus to organize the system for debugging purposes and portability, a layout based on the desired footprint and length of fiber was carved into a Styrofoam block and the fiber was laid in the channels. This completed pre-amplifier result can be seen in Figure 12. Testing to verify the correct amplification and spectral profile of the pre-amplifier output was done using a power meter for the intensity and a fiber spectrum analyser for the profile. More thorough testing procedure included mapping the output intensity as a function of the driving current of the diode. This will allow for the calculation of the average gain of the system.  P a g e | 16  Figure 12 - Finished pre-amplifier with fiber and loops placed in a foam block for portability and protection.  2.2.4 Grating Pulse Compressor and Optical Attenuator The implementation for the grating pulse compressor in this system is very similar to the device discussed in the theory section. The double pass method is still used, but instead of reflective diffraction gratings, transmission gratings were used instead to provide higher overall optical efficiency. Also, the transmission gratings used here were significantly cheaper than gratings previously purchased by the project sponsor. The gratings selected for this system are from LightSmyth (LSFSG-1000-3212-HP) with a line density of 1000 lines/mm, an optimal incidence angle of 31°, and a diffraction efficiency of greater than 94 percent. In addition, a roof prism was used instead of a simple retroreflecting mirror as it is much easier to align the gratings to the incident and returning beam. The roof prism separates the incident and returning beam by a small height, allowing for the returning beam to be adjusted independent of the incident beam. Using a roof prism also allowed for easier alignment to the D-shaped mirror. If a retroreflecting mirror was used, the D-shaped mirror would also have to be adjusted to correct for the angle of the returning beam. A picture labelling the different components of the grating compressor can be seen in Figure 13. Prior to the assembly of the grating compressor components, Equation 10 was used to provide an estimate of the required anomalous dispersion to achieve the required pulse compression where t0 is the input pulse width and tc is the compressed pulse width. An input pulse width of approximately 1ps can be assumed based on the approximate dispersion introduced by air and the fiber of the pre-amplifier. From the dispersion coefficient β2, Equation 9 can be used to determine the required gratings separation b given the gratings period d and incident angle θd.  P a g e | 17 To align the gratings for its desired function, the angle of the first grating was configured to the optimal angle of incidence, and then adjusted about this angle to obtain the maximum transmitted power. The pitch of the gratings was also adjusted via the ThorLabs mounts to optimize the amount of transmitted power. Then, the second grating was attached to a translating stage directed along the beam path to allow for modulation of the induced dispersion, thus the compression amount. The critical alignment was to ensure that both gratings are parallel with respect to each other to ensure that the frequency components are recombined correctly. The roof prism is mounted on the translation stage as well since its incident axis is parallel with the incident beam axis and not the grating axis. Thus, mounting them together ensures that the incident beam will always hit the prism.  Roof Prism  Input Adjustable Stage (1 Axis)  D-Shaped Mirror and Output  Transmission Gratings  Figure 13 – Overhead view of the Grating Compressor System with Plexiglas enclosure off  Optimal alignment for the grating compressor was accomplished using a combinanation of an IR reactive card and an IR viewer to observe where the beam was propagating and reflecting. This included adjusting the roof prism position to match the horizontal position of the beam spot on both gratings and to ensure that the incident and returning beam paths are separated and not being clipped by non optical surfaces. The returning beam (the output) from the compressor was then reflected onto another path using a D-shaped mirror and is then directed to an attenuator before it is coupled to the PCF. The attenuator consists of a half wave plate and a polarizing beam splitter. The  P a g e | 18 rotation of the half wave plate changes the linear polarization, causing more or less of the beam being transmitted at the PBS. Two mirrors at the output of the grating compressor were also used to direct the beam to the appropriate location. This can be seen in the overview image in Figure 14.  Half Wave Plate  PBS Mirrors Beam Direction  Figure 14 - Overview of the input polarization configuration optics, the grating compressor, and output optics including the attenuator (the pre-amplifier was left out of the image as it was not being used due to malfunctions).  To validate that the output of the grating compressor is providing the required pulse width, the output was guided into an autocorrelator that will measure the pulse width as the separation distance of the gratings is changed, and outputs the signal onto an oscilloscope for pulse width measurement. An enclosure was also designed and fabricated out of Plexiglas to cover the gratings and prevent dust accumulation on the grating surfaces. Due to the microstructure of the gratings, any dust accumulation will translate into performance losses of the system. The design and fabrication of the enclosure is discussed in Appendix B – Enclosure for Grating Compressor.  2.2.5 Photonic Crystal Fiber Coupling To couple the pulse train from free space (from the grating compressor) and into the micron sized core of the photonic crystal fiber, an aspheric lens mounted on a precision three axis translational stage to allow for small adjustments in aiming the focal point of  P a g e | 19 the lens onto the core surface of the PCF. The PCF is mounted in a Newport fiber clamp (model numbers 561-FH and 561-UM), and then mounted onto a set of posts to the optical table. A simple adapter block was fabricated to interface the two hole mounting pattern of the fiber clamp to the single hole mount of the post system. This can be seen in Figure 15.  PCF Z adjust  Lens  Fiber Clamp  Y adjust  X adjust Figure 15 – PCF coupling apparatus using a very short focal length aspherical lens mounted onto a flexure XYZ adjustment stage in order to steer the beam into the micron sized core diameter of the PCF.  To focus the incident spot size down to the photonic crystal fiber’s mode field diameter, the aspheric lens with an appropriate focal length must be used. The approximate focal length can be determined via 𝜆𝑓 𝑤𝑓 = 𝜋𝑤𝑜  Equation 14  where w0 is the input beam radius (approximately 0.55mm and was later measured to be 0.64mm) while wf is the focused beam waist radius. By setting this to each PCF’s mode field diameter, the focal length (f) for each PCF was determined and is shown in Table 1. Table 1 also shows the depth of field allowable for each PCF, given by an equation on a Newport tutorial 𝐷𝑂𝐹 =  8𝜆 𝑓 2 � � 𝜋 2𝑤𝑜  Equation 15  P a g e | 20 Table 1 – Focal length and depth of field calculations for aspheric lens selection for each PCF  PCF Model Number NL-1050-ZERO-2 SC-5.0-1040 NL-3.7-980  Mode Field Diameter 2.2 +/- 0.5 µm 4.0 +/- 0.2 µm 3.1 µm  Calculated Focal Length 1.85 mm 3.40 mm 2.60 mm  Depth of Field 7.42 µm 25.1 µm 14.65 µm  Selected Focal Length 2.0mm 3.3mm 2.5mm  In addition to selecting the appropriate focal length lens, the fiber itself must be prepared for efficient coupling. This requires that the protective acrylate layer be removed, and the fiber end be spliced for a clean flat surface. Since the PCF is more delicate than typical fiber, a razor blade was used to carefully scrape off the acrylate layer. Typical fiber strippers will also work, but this method was safer. In addition, the fiber cleaver also needed to be set at a lower tension (approximately 30% lower than for typical solid fibers) as the shock wave propagation from the initial crack is disrupted in PCF by the hollow microstructure (NKT Photonics, 2009). Lastly, the NKT application note also recommended that cleaning with solvents be avoided as the microstructure acts as capillary tubes and can wick solvent into the fiber, causing undesirable effects (2009). The output power from the photonic crystal fiber was monitored with a photodiode power meter. This region was shielded from stray light using black aluminum foil. By adjusting the three axis stage and observing the power output, the free space pulse train can be coupled into the PCF. More specifically, this is done by first adjusting the X and Y axis to find the fiber, and then adjusting the Z axis to better focus the beam into the PCF. These steps were repeated many times as minute adjustments can cause misalignment. The NKT Photnics application note describes a method of coupling where the fiber was mounted on the XYZ stage instead of the lens. However, the project sponsor suggested that the opposite configuration is more efficient. The expected coupling efficiency is in the 50 percent region.  2.2.6 Frequency Shift Characterization at PCF Output Once the pulse train was sufficiently coupled into the PCF with approximately 30 to 40 percent in coupling efficiency, the profile at the output of the PCF can be measured as a function of the incident intensity into the fiber. First, the power at the output of the PCF was measured as a function of the half wave plate angle. Recording these angles, the power meter was removed and replaced with a fiber spectrometer, the Ocean Optics USB2000. Placement of this detector was at a distance away from the fiber output such that the detector does not saturate. The half wave place was then readjusted to these recorded positions (corresponding to different input intensities) and the displayed spectral profile was saved using the proprietary software. This process was repeated for each PCF type.  P a g e | 21  2.3 Experimental Equipment A variety of equipment was used throughout the implementation of this wavelength tunable laser system. While some of the equipment has been mentioned in previous sections such as the fiber splicer and cleaver, this section will provide some detail into the other equipment including the autocorrelator, the spectrum analyzer, the optical power meters, and the fiber spectrometer used in characterizing the PCF output. The autocorrelator was used to quantitatively measure the pulse widths at the output of the grating compressor. The device is capable of measuring pulses in the femtosecond order of magnitude. The autocorrelator uses a set of optics that separates the input into two beams, one with a direct path into the photomultiplier tube while another obtains a variable delay prior to the photomultiplier tube. The basis of operation is to cross both normal and delayed beam at a second harmonic generating crystal. The second harmonics then enter the photomultiplier tube and the signal is output to an oscilloscope. The width of the second harmonic signal is related to the width of the pulse by the Gaussian factor 0.707. Δ𝑇𝑠𝑐𝑜𝑝𝑒 ∗ 0.707 = Δ𝑇𝑝𝑢𝑙𝑠𝑒  Equation 16  The manual for the device can be referred to for more details regarding the device’s operation, and alignment and measurement procedures.  Figure 16 - Femtochrome autocorrelator showing the input side (input hole at upper center of device) and various adjustment knobs for alignment.  P a g e | 22 A spectrum analyzer with a fiber input was used for a brief moment during the characterization of the pre-amplifier performance prior to the pump laser diode failure. This device allowed for an analysis of the spectral profile of the output of the preamplifier, and to ensure that only the signal is amplified. An optical power meter was used throughout the implementation to measure the power of the beam at various locations to check alignment, correct polarization, and performance issues. The power meter was also used to characterize the output power dependence on the rotation angle of the half wave plate at the attenuator. Its basis of operation is a photodiode probe and a meter that contains calibration values for different wavelengths. Figure 17 shows one of the power meters used during the system’s implementation.  Figure 17 – Simple photodiode based optical power meter (detector is under the black aluminum foil).  The fiber spectrometer from Ocean Optics, the USB2000, was used to observe the spectral profile at the output of the PCF. This device processes the measurements and outputs the spectral profile to a computer using proprietary software. This allows for easy recording of spectral data for later analysis.  2.4 Issues and Current Workarounds Seldom projects are without ongoing issues and temporary workarounds during the implementation process. Currently, there are two primary ongoing issues with the project, but have been mitigated with temporary workarounds that may be removed in the future. These issues include the non completion of the intended ytterbium doped  P a g e | 23 fiber laser oscillator to be used in this system, and the failure of the pre-amplifier pump diode. While these issues don’t prevent the implementation of other components, pulse compression and power amplification characteristics of the system have changed. As mentioned in some previous sections, since the intended laser cavity was not complete, a commercial laser was used for the source and was coupled into this system through a long fiber. The commercial laser provides approximately 200mW pulses with widths 130fs wide. Depending on the dispersion provided in the long fiber, the pulse width incident to the compressor may not be the same as the intended laser cavity. Therefore, the separation distance of the two gratings will need to be changed once the actual oscillator cavity is complete. Also mentioned in some of the previous sections, the pump diode for the pre-amplifier (Figure 18) has failed due to optical damage probably from back reflections. The preamplifier was purposed to amplify the input pulses of the system and provide sufficient intensity when coupling into the PCF. Since higher intensity correlates to higher magnitudes of nonlinear effect, more wavelength shifting could have been seen if the pre-amplifier was available. Currently, the system is utilizing the full 200mW of the commercial laser, but losses at the free space to fiber couplings has reduced the input intensity at the PCF significantly. Currently 40mW makes it out of the long fiber and is considered the incident power of the system. After the grating compressor, approximately 25mW becomes incident to the PCF.  Figure 18 - The failed pump laser diode attached to its heat sink and driver.  P a g e | 24  2.5 Results This section will detail the results obtained in quantitatively validating the performance of the pre-amplifier, the grating compressor, the coupling efficiency into the PCF, and the final spectral characterization of each PCF as a function if input intensity. With the current configuration, an approximately 4.5mW output with 200fs pulse width was obtained for the SC-5.0-1040 model PCF. The grating separation was set to be approximately 1.7cm to achieve this pulse width at the output. The shift of wavelength was observed in a couple of configurations the desired range of 1030 to 1500nm was not achieved.  2.5.1 Fiber Pre-amplifier Unfortunately, the diode laser for the pump failed during the process of testing when using the spectrum analyzer. The damage was believed to be optical in nature, and so an optical isolated was proposed to be implemented to further prevent back reflections from damaging the pump diode (see Figure 2). As a result, formal testing could not be completed to verify the power output and spectral profile. Initial qualitative observations prior to failure showed that there was amplification of the signal 1030nm pulse train. However a spectral measurement of the output showed that the output had a very broad spectrum, much broader than expected. Further validation should be done once the replacement parts arrive.  2.5.2 Grating Compressor Performance The alignment of the grating compressor greatly affects both the nonlinear frequency shift characterization stage as well as the overall application of the system. The requirements called for 200fs pulses, and the following measurements shown in Table 2 were obtained. Table 2 - Grating compressor measurements showing some measurements taken to determine the compressed pulse width as a function of grating separation distance  Approximate Grating Separation Distance 1.5 cm 1.6 cm 1.7 cm  Pulse Width (± 2.6fs) 239fs 212fs 196fs  Power transmission through the grating compressor system is fairly high but can be optimized slightly more. Currently approximately 40mW enters the grating compressor and approximately 30mW is detected at the output. This 25% loss, approximately 10%  P a g e | 25 can be accounted for by the grating pair while another 5% to 10% can be accounted for by the mirrors and roof prism. Therefore there is approximately 10% of loss that can be minimized to increase the power transmission through the grating compressor.  2.5.3 Coupling into and Characterization of Photonic Crystal Fiber As mentioned previously, coupling efficiency was expected to be within the 50% region. At best, the coupling efficiency achieved with the time allotted was only 30% to 40%. This reduces the intensity of the input pulse into the PCF, thus reducing magnitude of nonlinear effect. Further fine adjustment to the couple apparatus can be made to increase this efficiency. In the remainder of this section, two different spectral graphs of the wavelength shift using two different PCFs are shown. The first PCF that was used is the SC-5.0-1040 with a core diameter of 4.8 um and the second one that was used is the NL-1050-ZERO2 with a core diameter of 2.3um. A wavelength shift can be observed in both of these graphs as the input pulse intensity increases. However, the range achieved does not reach the required specifications of 1030 to 1500nm. The spectral graph of the SC-5.0-1040 PCF coupled using a 2.0 mm focal length lens can be seen in Figure 19 below. It can be seen that the peak wavelengths shift as the input power (and thus output power as well) increases.  Arbitrary Units  P a g e | 26  Figure 19 - Spectral graph of the SC-5.0-1040 PCF coupled using a 2.0 mm focal length. Three output intensities are shown to illustrate the wavelength shifting of the smaller peak on the right side.  An expected peak at 1030.6nm with a shifted wavelength peak at 1047.8 nm with a 3.33mW output is observed. After increasing the input pulse intensity, a change of the shifted peak to 1060.7nm with an output power of 7.04 to 7.40mW was observed. The spectral graph of the NL-1050-ZERO-2 PCF coupled using a 2.0 mm focal lens can be seen in Figure 20 below. Similarly to the other PCF characterized, the peak wavelengths shifted as the input pulse power is increased.  Arbitrary Units  P a g e | 27  Figure 20 - Spectral graph of the NL-1050-ZERO-2 PCF coupled using a 2.0 mm focal length. Three output intensities are shown to illustrate the wavelength shifting of the smaller peak on the right side  This NL-1050-ZERO-2 PCF has a smaller core diameter so the nonlinear effects should be stronger compared to the previous SC-5.0-1040 PCF. This increase is nonlinearilty causes the soliton self frequency shift effect to increase as well. Wavelength peaks can be observed at 1027.9nm, 1044.38nm and 1058.9nm with an output power of 1.78mW. As the input power is increased, the shifted wavelength peak was observed to increased to 1060.5nm with a lower intensity. However, as the maximum input power is reached, multiple peaks at neighbouring wavelengths are seen. Wavelength peaks at 1029.2nm, 1037.9nm and 1067.0nm with an output power of 5.56mW can be seen along with peaks at lower wavelengths of unknown origin. Though the range of the shifted peak wavelength seems to have increased in comparison to the previous PCF, the intensity for the shifted peak has also shrunk. Because of this property it is unclear whether this peak is still the product of the soliton self frequency shift.  P a g e | 28  2.6 Results Discussion From the results above, it is evident that that peak wavelength shifts more as incident pulse power is increased. Additionally, by decreasing the core diameter of the PCF, similar peak wavelength shifting characteristics are seen. These two properties both contribute to increasing the nonlinear effects, thus increasing the SSFS in the PCF. Even though this shift of wavelength was observed, the peaks are not very high and a significant peak at1030nm can be seen in most of the graphs. Considering these facts, it can be concluded that the nonlinear effects are not strong enough in the PCF to achieve the desirable results. In order to obtain the specified results, the magnitude of the nonlinear effects must be increased. Frequency shifts in both of these PCFs were observed and measured. A slightly stronger frequency shift in the NL-1050-ZERO-2 PCF was observed, most likely due to the smaller core diameter increasing the SSFS (soliton self frequency shift) effects. It is important to note that both of these PCFs were coupled using the 2.00mm lens to minimize the loss due to the coupling. A shift to the 1060nm wavelength in the SC-5.01040 PCF and 1067nm in the NL-1050-ZERO-2 PCF was observed. Due to the limited range of pulse powers that is achievable in this set-up, it was not possible to obtain spectral data of the wavelength shift for higher ranges of incident pulse power to the PCF. Another limitation that was present was that the spectrum analyzer only reads up to the 1200nm spectrum. However, it is important to note when increasing the incident pulse power, there was no evident shifting to beyond the 1200nm spectrum. As an alternative, a longer PCF with a smaller core diameter can be used to maximize the nonlinear effects and thus obtain a greater increase in the wavelength shift. Additionally, the pulse width of the incoming pulse could have been decreased even further to obtain even stronger effects in the wavelength shift. Once the pre amplifier becomes ready to use again, similar tests can be run again with the increased incident pulse intensity. The spectrum analyzer probe is most likely the largest source of error. The probe of the spectrum analyzer is very sensitive and the distance had to be adjusted accordingly in order not to oversaturate the signal. The output signal of the probe also changes significantly as the position is adjusted.  P a g e | 29  3 Conclusions The necessary components for the scope of this project have been successfully constructed. The grating compressor has been fully characterized and tested and the desired pulse width of approximately 200fs was achieved. The pre amplifier set up has been constructed but not tested due to the diode pump failing. Finally, two out of three different types of PCFs were couple to the system and characterized using the spectral data from above. The spectral data of the tested PCFs was of significant interest and shifting of the peak wavelength at the output of the PCF was observed in two PCFs that were coupled successfully. A shift to the 1060nm wavelength in the SC-5.0-1040 PCF and 1067nm in the NL-1050-ZERO-2 PCF was observed. It can be clearly seen that the wavelength shift is more prominent with a smaller core diameter fiber which corresponds to the larger nonlinearity effects. It can be also observed that the wavelength shift effect is stronger as the incident pulse power is increased, which is within the expectations. The intensity of the shifted wavelength, however, is not high enough to be viable for the desired purpose of tunability. From the results, it can be concluded that even though the desired ~200fs pulse width and 4.5mW of pulse power was obtained, the required tunability specification of 1030 to 1500 nm was not achieved. This is most likely due to the weaker nonlinear effects that caused the shift of wavelength to be less evident. The weak nonlinear effects could be contributed by a few factors such as the length of the fiber, the pulse power of the input pulse, the pulse width of the incident pulse as well as the core diameter of the fiber.  P a g e | 30  4 Project Deliverables From the beginning of the project, and described in the proposal prior to the start date, the original expected objectives of this project were 1. Characterize the output pulse width, average power, and peak wavelength of the Yb doped femtosecond fiber laser built by Dr. David Jones and Matt Lam using an autocorrelator, power meter, and optical spectrum analyzer. 2. Design and implement a simple fiber amplifier using a length of Yb-doped fiber to amplify the average power of the pulse measured in Objective 1 to a value such that after accounting for the intensity losses through the grating compressor will result in roughly 5mW input into the PCF. 3. Design and implement a grating compressor system that will take the input pulse with a width measured in Objective 1 and output a pulse 200fs long with spatial dimensions equal to the input pulse (measurable with an IR detector card and ruler as beam is collimated and enlarged through the coupling lenses). 4. Couple laser system to a length of PCF using an XYZ translation stage. The point at which coupling is achieved is when the transmission is maximized as measured via a power meter. An optical attenuator will also be necessary to vary the power input to the PCF. 5. Characterize the tunable range of the 3 types of PCFs available in the lab and determine which adequately allows for wavelengths from 1030nm to 1500nm without driving the pump diode past recommended operating limits (see specifications). 6. Deliver completed and tested system to project sponsor by Thursday March 3, 2011. However, due to changes in the project scope and issues in completing certain aspects of these objectives as described in previous sections. For example, objective 1 regarding the characterization of the proposed ytterbium fiber laser oscillator did not occur as the oscillator cavity was never finished. The next section will describe the revised list of deliverables at the end of the project.  P a g e | 31  4.1 List of Deliverables Considering the list of deliverables described in the original proposal, the actual physical deliverables remain unchanged. The deliverables still include the pulse shaping optics; the pre-amplifier and the grating pulse compressor, and the coupling apparatus and spectral characterization of the three varieties photonic crystal fibers. However, due to time constraints and system issues described in previous sections, Table 3 shows the current deliverables and its completion status at the end of the project. Table 3 - Deliverables and current status of completion at the end of the project timeline  Name Pre Amplifier Gratings Compressor PCF Coupling Wavelength Shift Characterization based on incident pulse power  Status Completed but not tested Completed and characterized 2 out of 3 fibers 2 out of 3 fibers  4.2 Financial Summary This section provides a financial summary of any parts purchased excluding those purchased by the project sponsor as the details of those components were not disclosed. These components were not part of the initial design of the system by the project sponsors and thus are not mentioned or considered in the proposal and were added on during the course of the project on an as needed basis for sensitive component protection, and component size adjustments for alignment purposes. The details are shown in Table 4. Table 4 - Estimated costs for parts fabricated during the project. These parts were all fabricated by the Berhnard Zender of the project lab.  Name Enclosure for Gratings Compressor Fiber clamp to 1in mounting post adapter Fiber Collimator Mount Sizing Adjustments Standoff plate for the 3 axis stage  Estimated Cost ~$50 ~$10 ~$10 ~$10  P a g e | 32  5 Recommendations This section will list and detail the possible recommendations for future work of the tunable femtosecond fiber laser system. The recommendations focus primarily on attempting to achieve the specified tunable range of 1030nm to 1500nm. 1. Increase Nonlinear Effects: Since the desired range of wavelength shifting was not obtained with the current configuration of the system, it suggested that the magnitude of the nonlinear effects should be increased by either changing to a longer length PCF, change to a PCF with smaller core diameter or use a higher input power to the system and finally, the pulse width could be decreased as a shorter pulse width corresponds to a higher possible peak intensity. 2. Further Characterization of Current and Different Varieties of PCFs: Since only two photonic crystal fibers were characterized due to time constraints and system problems such as the coupling mode issue, further tweaking of the coupling into the PCF and recharacterization is recommended to obtain a better representative of the frequency shift that each PCF can provide. Alternatively, more varieties of PCFs can be tested, as described in recommendation 1. 3. Install Pre-Amplifier and Revalidate: Since a commercial laser was used as a temporary replacement for the pre-amplifier and original ytterbium doped laser oscillator setup, the complete set up will inevitably show different results once the actual parts are coupled to the system. First, it is recommended that an isolator should be added to the pre-amplifier to prevent the pump diode from failing a second time (see Figure 2). After the pre-amplifier is coupled to the rest of the system, the gratings distance and attenuation should be altered to meet the desired requirements. 4. Optimize Transmission Gratings: Additionally, the power losses in some of the system components can be optimized further to reduce these losses and to achieve higher intensity in the transmission to the PCF input. Since the maximum angle for diffraction for the transmission gratings was not systematically optimized, it is recommended that some tweaking be explored here to achieve higher compression efficiency.  P a g e | 33  Appendix A – Pre Amplifier Schematic and Losses This pre-amplifier schematic was created in SolidWorks based on initial pictures and measurements of the system spliced together. This layout was used as a template to create a foam base with channels carved for the fiber to run through. The foam base acts as an organized and portable vessel for the pre-amplifier fiber, and also protects the system from damage. Figure 21 shows the schematic and Figure 22 shows the finished product with the fiber routed through.  Figure 21 – SolidWorks schematic of the pre-amplifier fiber routing based on dimensions collected from a picture taken of the assembled pre-amplifier.  Figure 22 – The assembled pre-amplfiier, showing the fiber routings in the foam block.  P a g e | 34  A table of the optical losses from the splicing the individual fiber components for the pre amplifier is seen below in Table 5. The description notes the location of each said splice and its corresponding loss values. Generally, a loss greater than 0.05 dB should be redone as it has a higher change of breaking. Loss values below 0.05 dB would be acceptable. Losses in extreme cases can mean that portions of the fiber were badly cleaved and the entire surface of the fiber end did not fuse properly.  Table 5 – Table of fiber losses at the connections between the different fiber components in the pre-amplfiier  Loss 0.02 dB 0.01 dB 0.04 dB 0.02 dB 0.02 dB  Description of Splice output collimator and YDF YDF and WDM Common WDM 980 and WDM Common Input collimator and WDM 1030 Pump fiber and WDM 980  P a g e | 35  Appendix B – Enclosure for Grating Compressor This enclosure was designed and fabricated to house the grating compressor assembly and to protect the gratings themselves from dust accumulation. Since the grating surface is a fine microstructure of minute channels, dust accumulation on these surfaces will reduce the efficiency of the device. Figure 23 shows the conceptual design in SolidWorks, and Figure 24 shows the final product. The panels were cut out of 0.25in Plexiglas via the laser cutter and assembled with methylene chloride solvent.  Figure 23 – The designed Plexiglas enclosure for the grating compressor, showing components inside.  Figure 24 – The resulting product of the design in the figure above. The enclosure protects the gratings from dust accumulation and damage.  P a g e | 36  Appendix C – Diagram of Fiber Clamp Adapter Mount Figure 25 and Figure 26 illustrate the design and final product of an adapter block that enables the mounting of the fiber clamp rail from Newport to the 1 inch post system.  Figure 25 – SolidWorks dimensioned and toleranced drawing of the adapter block.  Fiber Clamp Rail  Adapter Block 1in Post System  Figure 26 – Resulting part in use in the PCF coupling apparatus.  P a g e | 37  Appendix D – Bill of Materials The following table below is intended to be an approximate bill of materials for the system completed as of the writing of this report. Table 6 - Approximate bill of materials for all systems implemented  System  Description  Quantity  Pre-amplifier  Fiber to air coupling lens  3  Wavelength division multiplexer  2  Yb doped fiber  Length determined based on laser output power and grating losses  Diode laser pump  1  Heat sink for diode laser pump -modified with reverse voltage protection  1  Retroreflecting mirror  1  Fused-silica transmission gratings with a pitch of 1250 l/mm (Ibsen Photonics A/S, Denmark)  2  D-shaped mirror  1  Single axis linear adjustment stage  1  Optics mount with pitch/roll adjustments  4  Linear polarizer  2  Half wave plate  1  Optics mount with yaw adjustment (rotation about optical axis)  3  Grating Compressor  Optical Attenuator  P a g e | 38  Wavelength tuning system  Alternative: off the shelf variable optical attenuator  1  Aspherical lens to couple with the PCF  2  Photonic crystal fibers (PCFs)  3 types, see spec sheets to be attached in the final recommendation report  XYZ linear adjustment stages  2  Optics mount with pitch/roll adjustments  2  P a g e | 39  References Agrawal, G. P. (1995). Nonlinear Fiber Optics (2 ed.). San Diego: Academic Press. Agrawal, G. P. (2001). Applications of Nonlinear Fiber Optics. San Diego: Academic Press. Chen, W., Gao, X., Cousin, J., Poullet, E., Boucher, D., Sigrist, M., et al. (2006). Laser Difference-Frequency Generation in the Mid-Infrared and Applications to HighResolution Molecular Spectroscopy and Trace Gas Detection . Infrared Millimeter Waves and 14th International Conference on Teraherz Electronics, 18, 583. Hazama, H., Takatani, Y., & Awazu, K. (2007). Integrated ultraviolet and tunable midinfrared laser source for analyses of proteins. Nonlinear Frequency Generation and Conversion: Materials, Devices, and Applications VI, 6455, 645507.1645507.6. Lam, Matthew Hon C. (2010). Chirped-pulse laser amplifier and passive enhancement cavity for generation of extreme ultraviolet light. Newport Corporation | Gaussian Beam Optics Tutorial. (n.d.). Newport Corporation. Retrieved April 22, 2011, from http://www.newport.com/servicesupport/Tutorials/default.aspx?id=112 Newport Corporation. (2006). Prism Compressor for Ultrashort Laser Pulses. Application Note, 29, 1-9. NKTPhotonics. (2009). Fiber Handling, Stripping, Cleaving and Coupling. NKT Photonics Application Note, 1, 1-7. Sidorov-Biryukov, D., Kudinov, K., Podshivalov, A., & Zheltikov, A. (2009). Widely tunable 70-MHz near-infrared source of ultrashort pulses based on a modelocked ytterbium laser and a photonic-crystal fiber. Laser Physics Letters, 7(5), 355-358. Takayanagi, J., Sugiura, T., Yoshida, M., & Nishizawa, N. (2006). 1.0-1.7µm Wavelength-Tunable Ultrashort-Pulse Generation Using Femtosecond YbDoped Fiber Laser and Photonic Crystal Fiber. IEEE Photonics Technology Letters , 18(21), 2284-2286. Toptica: Ultra Compressed Pulse Femtosecond Fiber Laser. (n.d.). Toptica: Home. Retrieved April 22, 2011, from http://www.toptica.com/products/ultrafast_fiber_lasers/femtofiber_pro/femtofiber _pro_ucp.html  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.52966.1-0074457/manifest

Comment

Related Items