Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
Dispersive and nonlinear effects in high-speed reconfigurable WDM optical fiber communication systems
(USC Thesis Other)
Dispersive and nonlinear effects in high-speed reconfigurable WDM optical fiber communication systems
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
NOTE TO USERS Page(s) missing in number only; text follows. Page(s) were scanned as received. 115 This reproduction is the best copy available. ® UMI Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. DISPERSIVE AND NONLINEAR EFFECTS IN HIGH SPEED RECONFIGURABLE WDM OPTICAL FIBER COMMUNICATION SYSTEMS by Changyuan Yu A Dissertation Presented to the FACULTY OF THE GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulfillment of the Requirements for the Degree DOCTOR OF PHILOSOPHY (ELECTRICAL ENGINEERING) August 2005 Copyright 2005 Changyuan Yu Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. UMI Number: 3196919 Copyright 2005 by Yu, Changyuan All rights reserved. INFORMATION TO USERS The quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleed-through, substandard margins, and improper alignment can adversely affect reproduction. In the unlikely event that the author did not send a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion. ® UMI UMI Microform 3196919 Copyright 2006 by ProQuest Information and Learning Company. All rights reserved. This microform edition is protected against unauthorized copying under Title 17, United States Code. ProQuest Information and Learning Company 300 North Zeeb Road P.O. Box 1346 Ann Arbor, Ml 48106-1346 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Dedication To my parents, brother and wife, fo r their everlasting love and support. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Acknowledgements I would like to thank my academic advisor and dissertation committee chairman, Dr. Alan Eli Willner, for his support, guidance, and mentorship throughout my graduate studies. I would also like to extend my great appreciation to the other dissertation and qualifying examination committee members, Dr. William H. Steier, Dr. Edward Goo, Dr. John O ’Brien, and Dr. Stephan Haas. I would like to pay my heartiest thanks to my colleagues for their greatest support and invaluable discussions throughout my graduate career. They are Dr. Zhongqi Pan, Dr. Qian Yu, Dr. Yong-Won Song, Dr. A saf Sahin, Dr. Deniz Gurkan, Dr. Paniz Ebrahimi, Dr. Reza Motaghian, Dr. Lianshan Yan, M ichelle Hauer, Yan Wang, Ting Luo, John McGeehan, Poorya Sahari, Saurabh Kumar, Lou Christen, Bo Zhang, Reza Gholizadeh, Lin Zhang, and Irfan Fazal. I would also like to express my deep appreciation to my family for their everlasting love and support. In addition, I would like to acknowledge Dr. Daniel Nolan from Corning Inc. for providing the highly nonlinear fiber. iii Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Last but not least, I would like to thank those people who have contributed to the success of my academic endeavors. They are Milly Montenegro, Mayurni Thrasher, Gerrielyn Ramos, Diane Demetras, Ramona Gordon, and Kristan Venegas. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table of Contents Dedication.......................................................................................................ii Acknowledgements.......................................................................................iii List of Figures................................................................................................ x Abstract......................................................................................................... xv Chapter 1 .........................................................................................................1 Introduction.................................................................................................... I 1.1 Progress in Optical Fiber Communication S ystem s.............................................. 1 1.2 Dispersive Effects in Optical Fiber Communication System s.............................3 1.2.1 Chromatic D ispersion.................................................................................................3 1.2.2 Polarization Mode Dispersion (PM D ).................................................................... 8 1.3 Fiber Nonlinearities............................................................................................................14 1.3.1 Self Phase M odulation (S P M )................................................................................ 16 1.3.2 Cross-Phase Modulation (X PM ).............................................................................17 1.3.3 Four-Wave Mixing (F W M ).....................................................................................18 1.3.4 Stimulated Scattering................................................................................................20 1.4 Sum m ary..............................................................................................................................21 Chapter 2 ...................................................................................................... 22 V Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Optical Compensation of the PMD-Induced RF Power Fading for Single-Sideband Subcarrier-Multiplexed Systems.............................. 22 2.1 PM D-Induced RF Power Fading.................................................................................... 22 2.2 Concept of Optical Compensation of PMD-Induced RF Power Fading for SSB SCM System s..............................................................................................................................25 2.3 Experimental Setup........................................................................................................... 26 2.4 Results and D iscussions................................................................................................... 27 2.5 Sum m ary............................................................................................................................... 33 Chapter 3 ...................................................................................................... 34 Chromatic-Dispersion-Insensitive PMD Monitoring for NRZ Data Based on Clock Power Measurement.....................................................34 3.1 Chromatic-Dispersion-Insensitive PMD M onitoring.............................................. 34 3.2 Clock Power Recovered Using Notch Filter for NRZ D ata.....................................36 3.3 PMD Compensation Using the Recovered Clock Power as a Feedback Signal.41 3.4 Sum m ary..............................................................................................................................42 Chapter 4 ...................................................................................................... 43 40-GHz RZ and CS-RZ Pulse Generation Using a Phase Modulator and PM Fiber...............................................................................................43 4.1 Generation o f High-Speed RZ and CS-RZ Optical Pulse T rain .............................43 4.2 Concept o f Pulse Generation Using a Phase M odulator and PM F iber................45 4.3 Experimental Setup........................................................................................................... 48 4.4 Results and Discussion......................................................................................................49 vi Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 4.5 Sum m ary............................................................................................................................. 51 Chapter 5 ...................................................................................................... 52 160-GHz Pulse Generator Using a 40-GHz Phase Modulator and PM Fiber....................................................................................................... 52 5.1 Generation of High-Speed Optical Pulse Train beyond 40 G H z............................52 5.2 Concept of Pulse Generation with a Four Fold Repetition R ate.............................54 5.3 Experimental Setup........................................................................................................... 59 5.4 Results and Discussion......................................................................................................60 5.5 Sum m ary..............................................................................................................................63 Chapter 6 ...................................................................................................... 64 Polarization-Insensitive All-Optical Wavelength Conversion Using Dispersion-Shifted Fiber with a Fiber Bragg Grating and a Faraday Rotator Mirror............................................................................................ 64 6.1 Four-W ave M ixing Wavelength Conversion and Polarization Insensitive O peration......................................................................................................................................64 6.2 Concept o f Polarization-Insensitive Technique Using an FBG and an FR M 66 6.3 Experimental Setup........................................................................................................... 67 6.4 Results and Discussion......................................................................................................68 6.5 Sum m ary..............................................................................................................................75 Chapter 7 ...................................................................................................... 76 Width-Tunable Optical RZ Pulse Train Generation Based on Four- Wave Mixing in Highly-Nonlinear Fiber.............................................. 76 vii Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 7.1 W idth-Tunable Optical Pulse G eneration.................................................................... 76 7.2 Concept o f W idth-Tunable Pulse Generation Based on FWM in H N L F .............78 7.3 Experimental Setup........................................................................................................... 79 7.4 Results and Discussion......................................................................................................81 7.5 Sum m ary..............................................................................................................................88 Chapter 8 ...................................................................................................... 90 All-Optical XOR Gate Using Polarization Rotation in Single Highly- Nonlinear Fiber........................................................................................... 90 8.1 All-Optical XOR G a te ......................................................................................................90 8.2 Concept o f All-Optical XOR Gate Using Polarization Rotation in Single Highly— Nonlinear F ib er.......................................................................................................................... 92 8.3 Experimental Setup........................................................................................................... 94 8.4 Results and Discussion......................................................................................................96 8.5 Sum m ary............................................................................................................................101 Chapter 9 .................................................................................................... 102 3R Regeneration of a 40-Gbit/s Optical Signal by Optical Parametric Amplification in a Highly-Nonlinear Fiber...................102 9.1 3R Regeneration of Fligh-Speed Optical S ignal.......................................................102 9.2 Concept o f 3R Regeneration by Optical Parametric Amplification in Fiber With a Clock-Modulated P u m p ......................................................................................................104 9.3 Experimental Setup......................................................................................................... 105 9.4 Results and Discussion....................................................................................................107 9.5 Sum m ary............................................................................................................................I l l viii Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Chapter 10.................................................................................................. 112 Conclusion.................................................................................................. 112 References...................................................................................................116 Bibliography.............................................................................................. 126 ix Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. List of Figures Figure 1.1 Dispersion parameter D versus wavelength of SMF and D SF................... 5 Figure 1.2 Chromatic dispersion induces optical pulse broadening, proportional to the transmission distance and with the data rate...............................................7 Figure 1.3 Origin of PM D.........................................................................................................9 Figure 1.4 Probability distribution o f DGD in a typical fiber........................................ 11 Figure 1.5 Graphical representation o f the effect o f PMD on an optical pulse 12 Figure 2.1 Explanation o f PM D-induced RF power fading in an SSB SCM system in the optical dom ain................................................................................................. 26 Figure 2.2 Experimental setup for the optical compensation technique...................... 27 Figure 2.3 Optical Spectrum o f (a) the light at the PMD emulator output; (b) the optical carrier after FBG filtering; (c) the SSB after passing through the FBG; (d) after recombination of the optical carrier and the SSB using an optical coupler....................................................................................................... 29 Figure 2.4 Power fading due to first-order PMD and the compensation results. Solid line shows the theoretical RF fading curve.................................................... 30 Figure 2.5 Power fading distribution after the PMD emulator with and without optical compensation: (a) average DGD=42 ps, subcarrier frequency=20 GFIz; (b) average DGD=42 ps; subcarrier frequency=18 GHz; (c) average DGD=31 ps, subcarrier frequency=20 GHz; (d) average DGD=31 ps, subcarrier frequency=18 G H z .......................................................................... 31 x Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 2.6 M easured BER vs. received optical power for 155 M bit/s BPSK signal at 20 GHz. The compensated signal exhibited 14 dB o f RF fading before com pensation......................................................................................................... 32 Figure 3.1 Concept of PMD monitoring for NRZ data based on the recovered clock using a narrowband FBG notch fdter...............................................................37 Figure 3.2 Experimental setup............................................................................................. 38 Figure 3.3 Electrical spectrum: (a) w/o and (b) w/ fdter.................................................38 Figure 3.4 Optical spectrum: (a) w/o and (b) w/ fdter.......................................................39 Figure 3.5 Relative clock power as a function o f DGD for different C D ...................40 Figure 3.6 Relative clock power distribution for CD = 0 and CD = 600 ps/nm 41 Figure 3.7 Power Penalty distribution: (a) without and (b) with com pensation 42 Figure 4.1 Concept o f RZ/CS-RZ pulse generation using a phase modulator and PM fiber...........................................................................................................................45 Figure 4.2. Simulation results for waveforms of pulse trains: (a) RZ and (b) CS- R Z .............................................................................................................................43 Figure 4.3 Q-penalty vs. dispersion through SM F............................................................ 48 Figure 4.4 Experimental setup for 40G RZ/CS-RZ pulse train generation................49 Figure 4.5. Optical spectrum o f pulse trains: (a) RZ and (b) CS-RZ............................ 50 Figure 4.6. Waveforms of pulse trains: (a) RZ and (b) CS-RZ.......................................50 Figure 5.1 Concept o f CS-RZ pulse generation with a four fold repetition rate using a phase modulator and PM fiber............................................................................55 Figure 5.2 Details o f (a) the first and (b) the second stage o f PM fiber in the setup. 56 Figure 5.3 Simulation results for waveforms of pulse trains: (a) 80-GHz 33% RZ and (b) 160-GHz 50% CS-RZ.................................................................................... 59 xi Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 5.4 Experimental setup for 160G CR-RZ pulse train generation................... 60 Figure 5.5 Generated 40G CS-RZ pulse trains:(a) optical spectrum and (b) waveform ................................................................................................................ 61 Figure 5.6 Generated 80G RZ pulse trains at the first stage:(a) optical spectrum and (b) autocorrelator measurement........................................................................ 62 Figure 5.6 Generated 160G CS-RZ pulse trains at the second stage: (a) optical spectrum and (b) autocorrelator measurement.............................................63 Figure 6.1 Concept for polarization-insensitive FWM wavelength conversion 67 Figure 6.2 Experimental setup for polarization-insensitive FWM wavelength conversion................................................................................................................68 Figure 6.3. Polarization sensitivity o f the dual pass system vs. the loss o f the pump while passing forward through the D SF...........................................................70 Figure 6.4. (a) Power o f the converted wave vs. pump wave when power of signal wave=6 dBm. (b) Power o f the converted wave vs. signal wave when power o f pump wave=8 dBm..............................................................................71 Figure 6.5. Conversion efficiency vs. conversion distance............................................. 72 Figure 6.6. Experimental polarization sensitivity for up conversion (top) and down conversion (bottom), both single pass and dual pass.................................... 73 Figure 6.7. Power penalty induced by wavelength conversion at BER = 10'9........... 74 Figure 7.1 Concept o f width-tunable pulse generation based on four-wave mixing in FINLF: the pulse-width o f the generated pulse at G can be tuned continuously by tuning x ......................................................................................79 Figure 7.2 Experimental setup for width-tunable pulse generation based on four- wave mixing in FINLF..........................................................................................81 Xll Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 7.3 The FWHM o f generated pulse train vs. the delay x at different repetition rates........................................................................................................................... 82 Figure 7.4 Waveforms of the generated pulse train: (a) 5G; (b) 10G............................ 84 Figure 7.5 Optical spectrum (a) at the output o f FINLF; (b) o f 10G generated 25-ps pulse train.................................................................................................................85 Figure 7.6 Power Penalty vs. transmission distance through SMF w/o compensation. ....................................................................................................................................87 Figure 7.7 Power Penalty vs. transmission distance through SMF w/o compensation. Figure 8.1 Concept for all-optical XOR gate based on Kerr effect in a single highly— nonlinear fiber........................................................................................................ 93 Figure 8.2 Simulation results o f the output pattern at G with the input patterns at X\ and ^ 2 at 10 Gbit/s................................................................................................. 94 Figure 8.3 Experimental setup for all-optical XOR gate based on Kerr effect in a single highly-nonlinear fiber.............................................................................. 95 Figure 8.4 The output vs. intput optical power of the XOR gate...................................97 Figure 8.5 The output optical spectrum o f the XOR gate when inputs A -i or k2 are continuous waves: (a) /C off and X ,2 off; (b) Xi off and X ,2 on; (c) A ,i on and off; (d) X] on and A .2 on.................................................................................. 98 Figure 8.6 The input patterns at Xi and X2 and the resulting output pattern at X} for 10 Gbit/s data..........................................................................................................99 Figure 8.7 Eye diagram o f output of XOR gate for 10 Gbit/s PRBS data................. 100 Figure 9.1 Concept of 3R regeneration using OPA in FINLF with a clock-modulated pum p........................................................................................................................104 xm Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 9.2 (a) Experimental setup for 3R regeneration using OPA in HNLF with a clock-modulated pump, (b) Details of 3R regenerator............................... 106 Figure 9.3 (a) Recovered 20-GHz clock and (b) waveform o f the 40-GHz CS-RZ pump pulse train..................................................................................................107 Figure 9.4 Optical spectrum at the output o f the H NLF..................................................108 Figure 9.5 Detailed optical spectrum of the signal at (a) the input o f the HNLF; (b) the output of the HNF; (c) after the optical fdter........................................ 109 Figure 9.6 (a) Eye before the 3R generator; (b) eye after 3R generator................... 110 Figure 9.7 BER measurement for 40 Gbit/s D ata............................................................111 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Abstract Chromatic dispersion, polarization mode dispersion (PMD) and nonlinear effects are important issues on the physical layer of high-speed reconfigurable WDM optical fiber communication systems. For >10 Gbit/s optical fiber transmission system, it is essential that chromatic dispersion and PMD be well managed by dispersion monitoring and compensation. One the other hand, dispersive and nonlinear effects in optical fiber systems can also be beneficial and has applications on pulse management, all-optical signal processing and network function, which will be essential for high bite-rate optical networks and replacing the expensive optical-electrical-optical (O/E/O) conversion. In this Ph.D. dissertation, we present a detailed research on dispersive and nonlinear effects in high-speed optical communication systems. We have demonstrated: (i) A novel technique for optically compensating the PM D-induced RF power fading that occurs in single-sideband (SSB) subcarrier-multiplexed systems. By aligning the polarization states of the optical carrier and the SSB, RF power fading due to all orders o f PMD can be completely compensated, (ii) Chromatic-dispersion-insensitive PMD monitoring by using a narrowband FBG notch filter to recover the RF clock power for lOGb/s NRZ data, and apply it as a xv Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. control signal for PMD compensation, (iii) Chirp-free high-speed optical pulse generation with a repetition rate o f 160 GHz (which is four times o f the frequency of the electrical clock) using a phase modulator and polarization maintaining (PM) fiber, (iv) Polarization-insensitive all-optical wavelength conversion based on four- wave mixing in dispersion-shifted fiber (DSF) with a fiber Bragg grating and a Faraday rotator mirror, (v) W idth-tunable optical RZ pulse train generation based on four-wave mixing in highly-nonlinear fiber. By electrically tuning the delay between two pump pulse trains, the pulse-width o f a generated pulse train is continuously tuned, (vi) A high-speed all-optical XOR gate based on polarization rotation induced by Kerr effect in a single highly-nonlinear fiber, (vii) Wavelength- shift-free 3R-regeneration of 40-Gbit/s optical RZ signal by OPA with a clock- modulated pump in highly-nonlinear fiber. These techniques will play key roles in future high-speed dynamic WDM optical fiber communication systems and reconfigurable networks. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Chapter 1 Introduction This chapter provides a brief perspective of the progress in the field o f optical communication systems. Fiber properties such as chromatic dispersion, polarization mode dispersion (PMD), and various nonlinear effects which are considered as important effects in the high speed fiber communication systems are discussed briefly. 1.1 Progress in Optical Fiber Communication Systems Optical fiber communication systems use high carrier frequencies (-100 THz) in the visible or near-infrared region of the electromagnetic spectrum and employ optical fibers for information transmission. Such systems have been deployed worldwide since 1980 and revolutionized the technology of telecommunications [1]. The optical communication technology, together with microelectronics, brings the advent o f the “information age”. The major break through in optical fiber transmission came after invention o f Erbium-Doped Fiber Amplifier (EDFA). Due to the wide gain bandwidth of the EDFA, the wavelength- division multiplexing (WDM) channels can be simultaneously amplified and transmitted over long distances. The bit rates have reached 3.73-Tb/s (373 x 10- 1 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Gb/s, C+L band) over 11,000 km [2], 1.28-Tb/s (32 x 40-Gb/s, C band) over 4500 km [3], and 10.2-Tb/s (256 x 42.7-Gb/s, C+L band) over 100 km for WDM systems [4]. The demand for network bandwidth is outpacing even the astounding advances of recent years. The ever-increasing fiber optic base and the acceptance of WDM as an established technology are waiting to fulfill the enormous future potential of next-generation Internet services. The proliferation o f online services and network access providers coupled with low cost computers result in exponentially increasing numbers o f customers, with increasing bandwidth demands to support multimedia and other revolutionary applications. Faster processors fuel this demand, as today’s computers are outdated tomorrow. People spend more and more time online to perform everyday tasks. Because o f its high capacity and performance, optical fiber communications have already replaced many conventional communication systems in point-to-point transmission and networks and also have been considered as a good candidate for wireless backbone. To fully utilize the bandwidth of the fiber and achieve a high performance, the physical effects on the physical layer of optical fiber communication systems must be detailed studied. Optical signals suffer from many physical effects in the fiber, including chromatic dispersion, polarization mode dispersion (PMD) and fiber nonlinearities. For high-speed reconfigurable WDM optical communication 2 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. systems, it is essential that chromatic dispersion and PMD be well managed by using some type o f dispersion compensation. Furthermore, for automated timely tunable dispersion compensation, some methods o f dispersion monitoring have to be implemented. By managing fiber chromatic dispersion, PMD and nonlinearities, the capacity of optical systems has been greatly expanded over the past few years [1]. There are still a lot of issues in this area need to be addressed and better solutions need to implement to achieve ultra-high capacity and performance systems. On the other hand, nonlinear effects in optical fiber transmission system can also be beneficial and has applications on pulse management, all-optical signal processing and network function [5], which will be essential for high bite-rate optical networks and replacing the expensive optical-electrical-optical (O/E/O) conversion. 1.2 Dispersive Effects in Optical Fiber Communication Systems Fiber dispersive effects include chromatic dispersion and polarization-mode dispersion (PMD). Both effects degrade the performance o f high-speed optical fiber communication systems. 1.2.1 Chromatic Dispersion When an electromagnetic wave propagates through fiber, the medium response depends upon optical frequency co. This property, referred to as chromatic dispersion, manifests through the frequency dependence o f the refractive index n(co) 3 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. [1,5-7]. Fiber dispersion plays a critical role in propagation o f optical pulses since different spectral components associated with the pulse travel at different speeds given by c/n{co). Consequently, the optical pulse at the output o f the fiber will be distorted. Dispersion effect can be considered by expanding the mode-propagation constant P in a Taylor series around the center frequency ®o [5] The pulse envelope travels at the group velocity (vg = 1/Pi), while the parameter p2 is responsible for pulse broadening to the first order. p2 also depends on the wavelength (i.e., frequency) o f the optical signal. The wavelength for which p2 = 0 is often referred to as dispersion-zero wavelength (A .o); X o = 1.3 pm for standard single mode fiber (SMF). More commonly used system parameter is the dispersion parameter D; the quantity D is related to p2 by the equation D = i K =_lnc dA A2 2 The unit o f D is ps/(nm-km). D for SMF is about +17 ps/(nm-km) at 1.5 pm. Tailoring the waveguide profile can change dispersion parameters. In dispersion shifted fibers (DSF), Xo is in the neighborhood o f 1.5 pm and D usually between -2.5 and +2.5 ps/nm-km at 1.5 pm. The dispersion parameter, D as a 4 ( 1-1) Where ) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. function o f wavelength for both SMF and DSF is shown in Figure 1.1. Negative D values are referred normal dispersion (P 2 is positive), and positive D values are referred anomalous dispersion (P2 is negative). The wavelength dependency of D is usually considered through dispersion slope which is dD/dX = 0.08 ps/nm -km (for both SMF and for DSF around 1.5 pm). 15 10 SMF DSF 5 0 X — 1.31 ja m 10 15 Wavelength (urn) Figure 1.1 Dispersion parameter D versus wavelength o f SMF and DSF [3]. Both negative and positive dispersion cause pulse broadening at the output of the fiber. The broadening increases with the fiber length, imposing a limit on the maximum distance and/or data rate without regeneration. Therefore, chromatic dispersion must be mitigated for high-speed or long-distance systems. Even though it is possible to manufacture fiber with zero dispersion, it is not practical to use such fiber for WDM transmission, because o f large four wave mixing (FWM) induced penalties. Therefore, that chromatic dispersion must be managed, rather than eliminated in optical fiber transmission systems. There are several important aspects 5 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. of optical systems and networks that make tunable dispersion compensation solutions attractive, especially in high-speed optical networks. As shown in Figure 1.2, because o f the non-zero spectral with of modulated data (optical pulse stream), dispersion leads to pulse broadening, proportional to the distance and with the data rate, thus imposing a limit on the maximum distance transmission without regeneration. Dispersion-limited distance can be approximated by determining the transmission distance at which a pulse is broadened by one bit interval. The estimated dispersion limited distance L d for a signal having non-return to zero (NRZ) intensity modulation can be obtained by D Ld = T = — = > Ld = —----- (1.3) c D R D A2D R2 where R is the data rate, T (= 1/R) is the bit time, c is speed o f light, and 1 is the wavelength o f the optical signal. The dispersion limited distance decreases as square of bit rate. One criterion for detecting this limit for an externally modulated NRZ signal is [8]: B 2 DL <\04,000{G b / sf-ps/nm which corresponds to a dispersion-induced power penalty o f IdB. For single-mode fibers with D=17 ps/nm/km, the maximum distance is approximately 1000 km for a bit rate o f B = 2.5-Gb/s, but decreases to about 60 km for B = 10-Gb/s and 5 km for B = 40-Gb/s in an externally modulated system. Some method o f dispersion Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. compensation must be employed for a system to operate beyond these distance limits. Information Bandwidth of Data 0 Fourier^ 0 transform Time v = velocity _ l ^ Optical ^ C a rrie r Freq. ps Temporal Pulse Spreading —> j [distance, bit rate] nm*km 000 . Time Figure 1.2 Chromatic dispersion induces optical pulse broadening, proportional to the transmission distance and with the data rate. For >10-Gb/s data rates that are transmitted over >100 km, it is essential that chromatic dispersion be well managed by using some type of dispersion compensation. In theory, compensation o f chromatic dispersion for high-speed or long-distance systems can be fixed in value. However, static, fixed dispersion compensation is inadequate when system conditions can change in the following scenarios: (i) reconfigurable optical networks for which a given channel's accumulated dispersion will change when the network routing path is reconfigured, and (ii) >40-Gb/s long-distance links for which chromatic dispersion and signal degradation may change substantially due to normal changes in temperature [9]. The required accuracy in dispersion compensation increases dramatically with the data 7 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. rate. While the amount o f residual dispersion that is tolerable at 10-Gb/s is large, of the order of 1000 ps/nm, in 40-Gb/s systems this margin shrinks to only 60 ps/nm. Thus the use of tunable modules is the only way o f managing accumulated dispersion. 1.2.2 Polarization Mode Dispersion (PMD) Single-mode fibers actually support two perpendicular polarizations o f the original transmitted signal (fundamental mode). In an ideal fiber (perfect) these two modes are indistinguishable, and have the same propagation constants owing to the cylindrical symmetry o f the waveguide. However, practical fibers are not perfect and, as a result, the two perpendicular polarizations may travel at different speeds and consequently arrive at the end o f the fiber at different times. This phenomenon is called polarization mode dispersion (PMD). As shown in Figure 1.3, the major cause of PMD is the asymmetry of the fiber-optic strand. Asymmetry is simply the fact that the fiber core is slightly out-of round, or oval. Fiber asymmetry may be inherent in the fiber from the manufacturing process, or it may be a result of mechanical stress on the deployed fiber. The inherent asymmetries of the fiber are fairly constant over time, while the mechanical stress due to movement of the fiber can vary, resulting in a dynamic aspect to PMD. 8 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The mechanical stress on the optical fiber can originate from a variety of sources. One source that is very difficult to control is the day/night and seasonal heating and cooling of the optical fiber. Although much fiber is deployed in the ground and often within conduits, it is still subject to temperature variations and corresponding mechanical stress. Another source o f mechanical stress can originate from nearby sources of vibration. For example, much fiber is deployed alongside railroad tracks because of the ease of right-of-way and construction. However, vibration from passing trains can contribute to stress on the optical fiber. And wind can cause swaying o f the fiber cable and can contribute to PMD o f fiber deployed aerially. Geometric Internal Stress External Stress Bend Twist Figure 1.3 Origin o f PMD. 9 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The difference in propagation constants (differential phase velocity) of these two modes is responsible for PMD in the fiber, and can be related to the difference in refractive indices between the two orthogonal polarization axes as p0 - pc = wn0 /c - cone/c = coAneff/c (1.5) where n0 and ne are the effective refractive indices o f two orthogonal axes, and Aneff is the differential index of refraction. The differential index o f refraction is a measure of birefringence in the fiber, and is usually between 1 C T 7 and 10’5. Since PMD is caused by the different transmission speeds o f the signal's two states-of-polarization (SOPs) as they propagate along a fiber having a small birefringence, and the birefringence o f a fiber changes randomly along a fiber link, PMD is a statistically random quantity. PMD is characterized by differential group delay (DGD) between two principal states of polarization (SOPs) after a given length of fiber. Because of random variations in the perturbations along a fiber span, PMD in long fiber spans accumulates in a random-walk-like process that leads to a square root o f transmission-length dependence [10]. Therefore, PMD does not have a single value for a given span o f fiber. Rather, it is described in terms of average DGD, and a fiber has a distribution of DGD values over time. The probability o f the DGD of a fiber section being a certain value at any particular time follows a Maxwellian distribution (see Figure 1.4). The probability o f D G D = A t is given by: 10 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 32A x2 4 Ax2 PDF (Ax) = j f exp( -------"-----t ) (1.6) 7 1 < Ax > 7 1 < Ax >*" ’ 1/2 with mean value < A r > . PMD is usually expressed in ps/km in long fiber spans, and the typical value of < A r > is 0.1 to 10 ps/km l/2 [7,11,12]. Q O Q o o O h Average (M ean) DGD Figure 1.4 Probability distribution of DGD in a typical fiber. Additionally, in a cascaded fiber link, there may be many discrete components (i.e., isolators, couplers, wavelength multiplexers), which are polarization dependent due to molecular asymmetry (anisotropy) o f the waveguide material. Although PMD caused by polarization dependence o f a single component may be negligible, cascaded components may add significant PMD in a long link. The combined PMD-induced broadening in a long link may be up to a few tens of ps, which can degrade systems operating at >10 Gbit/s. In systems operating at >40 Gbit/s, PMD has been proved to be deleterious. In order to enable ultra-fast 11 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. TDM/WDM communications over long distances of optical fiber, the remaining critical issue o f PMD must be considered. In addition to the time variance of DGD, PMD also varies over wavelengths, known as higher-order PMD. This variance results in an optical dispersion that is a function o f both the channel bandwidth and the value of DGD over that bandwidth [7], Figure 1.5 is a graphical representation o f the effect o f PMD (both first- and higher-order) on an optical pulse. The optical pulse and its constituent photons travel from the source, or transmitter, at distance = 0, along the single-mode optical fiber. At some distance after PMD has affected the pulse, the polarized energy is separated by some time (i.e., DGD). If DGD is severe, the receiver at some distance L cannot accurately decode the optical pulse, and bit errors can result. If the bit errors caused by PMD are too numerous, then the transmitted information is too corrupt to recover and the transmission link should be considered out o f service. at distance = 0 at distance = L higher-order PM D 1st order PM D Figure 1.5 Graphical representation of the effect o f PMD on an optical pulse. The quantity o f bit errors encountered at the receiver is directly influenced by the amount of PMD in a fiber optic transmission span. DGD o f this magnitude, in a 12 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 10-Gb/s transmission system, can be expected to result in a bit-error rate that is severe enough to cause service problems. Some general rules on limitations of distances caused by PMD are given in Table 1.1. PMD Limited Maximum Optical Transmission Distance (km) Data Speed (Gbit/s) Fiber PMD (ps/km 1 /2 ) 1.0 (Old fiber) 0.5 (New fiber) 0.1 (Future) 10 100 km 400 km 10000 km 40 6 km 25 km 600 km Table 1.1 Limitations of transmission distances caused by PMD PMD induced problems can be reduced simply by regeneration, i.e., shortening the optical transmission distance. However, from a network point o f view, a regeneration site is an inefficient and costly optical-electronic conversion site. Adding to the expense and inefficiency o f a regeneration site is the fact that most long-haul transmission systems are now multiwavelength, dense wavelength division multiplexing (DWDM) systems. In this application, the transmission link must first be demultiplexed, then regenerated, then multiplexed again. This is a very costly operation compared to the preferred alternative o f a multiwavelength amplifier. From a network and cost perspective, a more efficient method o f addressing the PMD problem is to fix the effects o f PMD while the transmission is in an optical 13 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. state, before a receiver tries to decode the bits. A PMD compensator (PMDC), deployed at the destination of the transmission system, can reduce the effects of the PMD in the fiber and ensure that the optical bits are correctly decoded by the receiver before they are to be routed and switched. The most reliable and efficient PMDC technology is the use o f adaptive optics to realign and correct the pulses of dispersed optical bits. 1.3 Fiber Nonlinearities There are two categories of fundamental optical nonlinear effects that can cause degradation o f the transmitted signal. They are refractive-index effects and stimulated scattering effects. Refractive index effects are associated with modulation of the refractive index due to changes in the light intensity. Stimulated scattering effects arise from parametric interactions between light and acoustic or optical phonons (due to lattice or molecular vibrations) in the fiber. The refractive index n of silica is not a constant but increases with power (or light intensity) according to the relationship: P h{co, P) = nQ (co) + n2I = n{ ) (a>) + n2 (1-7) Aeff where n„(co) is the linear refractive index of silica, is the intensity-dependent refractive index coefficient, and I=P/Ae f is the effective intensity in the medium. The typical value of m is 2.6 x 1(T2 0 m2 /W. This number takes into account the 14 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. averaging o f the polarization states o f the light as it travels in the fiber. The intensity dependence o f the refractive index gives rise to three major effects [5,13]: (i) self phase modulation (SPM), (ii) cross-phase modulation (XPM), and (iii) four wave mixing (FWM). All these three nonlinear effects can significantly degrade the performance of a WDM lightwave system [14,15]. XPM and FWM are more severe in multi-channel WDM systems, while SPM can occur in both single channel and WDM systems. The relevant power-times-distance products for amplified transmission systems can be so large as to make fiber nonlinear effects the dominant factor in determining the design o f long-distance systems. System specifications such as the non-regenerated span length L, amplifier spacing l\, number o f WDM channels N, channel frequency spacing Af, and power per channel Po are all affected. Understanding how system performance is degraded by fiber dispersion in the presence of fiber nonlinearities is crucial for designing amplified transmission systems. One the other hand, nonlinear effects can also be beneficial: refractive- index effects has applications on pulse compression and logic gates for all-optical signal processing and network function; and stimulated scattering effects can be used for signal amplification [5]. 15 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1.3.1 Self Phase Modulation (SPM) Self-phase modulation (SPM) is the phenomenon where any modulation on the signal power gives rise to modulation of the signal phase and spectral broadening. The nonlinear contribution of the index o f refraction due to optical power P results in a phase change O nl, which for light propagating in a fiber given in which AC ff is the effective mode area of the fiber, and a is the fiber attenuation loss. Lerris the effective nonlinear length of the fiber that accounts for the fiber loss, and y is the nonlinear coefficient measured in rad/(km-W). Although nonlinear coefficient is small, the lengths and powers that have been made possible by the use of the optical amplifiers (EDFAs) can cause the nonlinear phase large enough to play a significant role in the state-of-the-art lightwave systems [5,7]. When an intensity-modulated signal travels through an optical fiber, the peak of the pulse accumulates phase more quickly than the wings due to nonlinear refractive index. This results in a nonlinear chirping o f the signal. The SPM induced chirp may interact with dispersion induced chirp and can cause a totally different behavior depending upon positive or negative dispersion values [5,7], In the normal by [5] = yPLeff (1.8) where the quantities y and Leff are defined as 2 ^ 2 and L = ------ a 16 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. dispersion regime (D < 0), the SPM induced nonlinear-chirp will add to the dispersion-induced linear-chirp, thereby causing not only the enhanced pulse broadening but also distorting the shape o f the pulse. In the anomalous dispersion regime (D > 0), the SPM induced nonlinear-chirp will tend to partially negate the dispersion induced linear-chirp, thereby slightly reducing the pulse broadening, but still will distort the pulse shape. Therefore, SPM induced chirp can impose a limitation on bit rate and transmission distance in lightwave systems. The SPM induced chirp is dependent upon the power and the shape o f the optical pulse. Therefore, if the power and the shape of the pulse is right, the SPM induced chirp and the dispersion-induced chirp can completely negate each other in anomalous dispersion regime (D > 0) [6]. The pulse with the right shape and power is called soliton. 1.3.2 Cross-Phase Modulation (XPM) Cross-phase modulation (XPM) is the phenomenon in which intensity fluctuations in one channel propagating in the fiber modulate the phase o f all the other channels or alternatively all the WDM channels (at different wavelengths) in the fiber modulate the phase of any one channel [5,16], In a multi-channel system, the excess bandwidth generated by this XPM effect is given by <‘-9> 17 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Note that the XPM induced chirp is twice as much as that o f the SPM induced chirp. This factor o f 2 arises from counting of terms in the expansion o f the nonlinear polarization inside the fiber [5,16]. Therefore, it appears that XPM can impose more severe limitation than SPM for WDM systems because effect is twice as large for each interfering channel, and there can be a lot o f interfering channels. However, fiber dispersion plays a significant role in the system impact o f XPM [5], Due to dispersion, pulses at different wavelengths travel with different speeds inside the fiber because o f group velocity mismatch. In normal dispersion regime (D < 0), a longer wavelength travels faster while the opposite occurs in the anomolous- dispersion regime (D > 0). This feature leads to a w alk-off effect that tends to reduce XPM effect. 1.3.3 Four-wave mixing (FWM) Like SPM and XPM, four-wave mixing (FWM) is also generated by the intensity-dependence of refractive index of silica. However, impact on performance of WDM system is completely different. In FWM, the beating between two channels of a WDM system at their difference frequency, modulates the phase o f one of the channels at that frequency, generating new tones as sidebands [17], When three waves o f frequencies f, fj, and interact through fiber nonlinearity, they generate a wave o f frequency fuk = f ^ f j - f k d-10) 18 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Therefore, three waves give rise to nine new optical waves by FWM. In WDM system with equally spaced channels, most of the product terms generated by FWM fall at the channel frequencies, giving rise to crosstalk. The center channels are more vulnerable to this cross talk since the number o f FWM products, which fall on center channels, is higher than those, which fall on end channels [7,17], The efficiency of FWM depends on the channel spacing and the fiber dispersion. Increasing channel spacing or fiber dispersion will reduce mixing efficiency. High-speed WDM systems require simultaneously high launched power and low dispersion values. This greatly enhances the efficiency o f FWM, making FWM the dominant nonlinear effect in WDM lightwave systems. FWM can impose severe limitation on bit rate/channel, transmission distance, and number o f WDM channels [’ 7 1 - Dispersion limits the maximum transmission distance and the bit rate. But, the effects o f XPM and FWM are reduced by dispersion because dispersion destroys the phase matching conditions. In order to achieve good system performance, it is important to consider the chromatic dispersion and the nonlinear effects of the transmission fiber together. Dispersion management is a solution for this dilemma: use two different types o f fibers having opposite dispersions periodically. The total accumulated dispersion is zero after some distance, but the absolute dispersion is non-zero at all points along the link. The result o f this dispersion management 19 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. scheme is that the total effect o f dispersion is negligible for all channels, and non zero dispersion causes phase mismatch between channels thereby destroying efficient nonlinear interactions. 1.3.4 Stimulated Scattering The nonlinear effects described above are governed by the power dependence of refractive index, and are elastic in the sense that no energy is exchanged between the electromagnetic field and the dielectric medium. A second class o f nonlinear effects results from stimulated inelastic scattering in which the optical field transfers part o f its energy to the nonlinear medium. Two important nonlinear effects fall in this category [5]: (i) stimulated Raman scattering (SRS), and (ii) stimulated Brillouin scattering (SBS). The main difference between the two is that optical phonons participate in SRS, while acoustic phonons participate in SBS. In a simple quantum- mechanical picture applicable to both SRS and SBS, a photon o f the incident field is annihilated to create a photon at a downshifted frequency. The new photon is propagated along the original signal in the same direction in SRS, while the newly generated photon propagates in the backward direction in SBS. Furthermore, the downshifted frequency range where new photons can be generated is ~30 THz in SRS and only -3 0 MHz in SBS. Therefore, SBS does not impose any significant limitations in high-speed (Gb/s systems) digital lightwave systems. However, SRS can impose some limitations on WDM systems because the effect o f SRS is to deplete the energy o f some channels (higher frequency channels) on behalf o f the 20 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. other channels (low frequency channels). The effect of SRS is not very significant unless the number of channels are more than 100 [7], On the other hand, SRS can be used for signal amplification in a fiber (so called Raman amplifier). Raman amplifier is becoming more and more cost effective now and is extensively developed in recent years because of some its unique features. Indeed, unlike EDFA, Raman amplification can virtually occur at any wavelength by properly choosing the pump wavelength and a large bandwidth can be achieved by combing several pump wavelength. 1.5 Summary In this chapter, we introduced the most important dispersive and nonlinear effects in optical fiber transmission systems. Chromatic dispersion, PMD, and different fiber nonlinearities have been discussed briefly. 21 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Chapter 2 Optical Compensation of the PMD-induced RF Power Fading for Single-Sideband Subcarrier-Multiplexed Systems In this chapter, we demonstrate a novel technique for optically compensating the PM D-induced RF power fading that occurs in single-sideband (SSB) subcarrier- multiplexed systems. By aligning the polarization states o f the optical carrier and the SSB, RF power fading due to all orders o f PMD can be completely compensated. 2.1 PMD-induced RF Power Fading Subcarrier multiplexing (SCM) has several important applications in optical systems including: cable television, antenna remoting, microwave photonics, and for control and routing information in optically-switched networks. However, deleterious RF power fading has been reported in the transmission o f analog and digital SCM signals over fiber due to polarization mode dispersion (PMD) [18,19], As discussed in Chapter 1, PMD is caused primarily by the asymmetry of the optical fiber core that causes a birefringence such that light polarized along one axis will travel faster than light polarized along the orthogonal axis. A key feature of 22 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. PMD is its statistical behavior, since the relative orientation between the state-of- polarization (SOP) o f the input signal and the principal-states-of-polarization (PSPs) of the fiber varies randomly with time [19]. The differential group delay (DGD) between the fast and slow PSPs, i.e. first-order PMD, is a random process with a Maxwellian probability distribution, such that network outages can occur due to rare events in the tail of the distribution. The deleterious PMD-induced power-fading effect in SCM can be described in the time domain as follows. The light can be decomposed along two orthogonal PSPs, with one axis traveling faster than the other. The time delay between the faster and slower axes causes a phase difference in the corresponding received subcarrier signals. This phase difference induces destructive interference and may lead to serious RF power fading that is a function of subcarrier frequency and accumulated DGD [20], Another explanation involves the polarization state in the optical frequency domain. PMD-induced RF power fading occurs when the polarization state of the optical carrier wave is not aligned with the polarization state of the subcarrier, since PMD will cause the polarization state o f the carrier and subcarrier to wander at different rates. In general, double-sideband (DSB) transmission will be affected by chromatic dispersion because o f the relative time (phase) shift that develops between the upper and lower sidebands due to frequency-dependent velocities in the fiber. To 23 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. avoid this, many SCM systems employ single-sideband (SSB) subcarriers and are therefore relatively immune to chromatic dispersion [21], However, PMD-induced RF power fading remains a problem even for SSB signals [19], since the relative polarization state o f the carrier to the SSB changes through the transmission fiber. For example, in a 40-GHz optical SSB SCM system, the RF power is completely faded with 12.5-ps instantaneous DGD. Furthermore, we note that higher-order PMD can cause additional power fading [22,23]. Therefore, robust transmission o f an SCM signal may necessitate the use of some technique to compensate or mitigate the power fading effects o f PMD. Published work includes: (i) using a first-order PMD compensator [19], and (ii) employing polarization diversity, which requires two optical detectors [24], We propose and demonstrate a novel technique for optical compensation of the PM D-induced RF power fading that occurs in SSB SCM transmission systems. Given that the PM D-induced power fading in the optical domain is caused by the difference between the polarization states o f the SSB signal and the optical carrier, we can overcome this problem by: (i) splitting the optical carrier and the SSB signal at the receiver using a narrowband optical grating fdter, (ii) realigning their polarization states relative to each other, and (iii) recombining them. In this way, first-order and all-higher-order PM D-induced RF power fading can be completely compensated. We show that RF power fading of the 5% distribution tail can be 24 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. decreased from 13.5 dB to <1.5 dB for both 300 experimental samples as well as for a 10,000-sample simulated system. The technique is independent o f the DGD o f the optical fiber link and the subcarrier frequency. 2.2 Concept of optical compensation of PMD-induced RF power fading for SSB SCM systems Figure 2.1 shows the concept o f PM D-induced RF power fading in SSB SCM systems. At the transmitter, the optical carrier and the SSB have the same SOP. After propagating through the optical fiber link, first-order and higher-order PMD cause the SOPs o f the optical carrier and the SSB to vary by different amounts. To properly recover the modulated data, the SSB must beat with the carrier in the receiver. However, only the portions o f the signals that have the same polarization will effectively interfere, so a relative SOP difference will cause the received RF power to fade. Complete fading will occur when the SSB is orthogonal to the optical carrier. However, if the SOPs can be realigned such that they are the same for both the optical carrier and the SSB, the PM D-induced RF power fading can be completely removed. It is important to note that it is not sufficient to recover the faded RF signal by using only first-order PMD compensation [19]. On the other hand, no matter changes of the SOP are induced by first-order or higher-order PMD, since our technique is based on realigning the SOP of the optical carrier with the 25 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. SSB, the PM D-induced RF fading will be compensated. And it is independent of the DGD o f the fiber link and the subcarrier frequency. SSB Subcarrier Transmitter Carrier n SSB Optical Spectrum State of polarization (SOP) changes while propagating through the fiber Fiber Output Fiber link w/ PMD Change of SOPC a rrier * Change of SOPS S B •SOPC a r r ie r I I SOPS S B . No Fading X *SOPC a r r j C r k SOPS S B . Partial Fading C a r rie r T S O P s S B : X -sop -* Total Fading! Figure 2.1 Explanation of PM D-induced RF power fading in an SSB SCM system in the optical domain. 2.3 Experimental Setup Figure 2.2 shows the experimental setup. We first generate an 18~20 GHz double sideband signal by externally modulating the 1550 nm optical carrier. An SSB signal is obtained by using a fiber Bragg grating (FBG) to filter out the lower sideband. The light then propagates through a PMD emulator. Generally, the optical carrier and the SSB will have different SOPs at the output o f the PMD emulator, since the PMD will induce different changes in the SOP for the optical carrier and the SSB. At the receiver, a circulator and a narrowband FBG filter is used to separate the optical carrier and the SSB. The FBG has a reflection of 99.7% at the optical carrier wavelength o f 1550 nm, with a bandwidth o f 0.1 nm. The 26 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. reflected optical carrier passes through a polarization controller (PC) that is used to align its SOP with the SSB. The optical carrier and SSB are then recombined with a coupler and sent to the receiver. By adjusting the PC to maximize the received RF power, the faded RF signal can be completely recovered after detection. We adjust the PC manually during the experiment, and it is possible to make the PC adjustment automatically by using an electrically tunable PC. PMD Compensator PMD Emulator SSB Transmitter SSB FBG _ /L > EDFA Circulator FBG Rx Laser E-O PC Coupler Carrier 4 , SSB RF Signal 18-20 GHz Carrier Optical Spectrum L Figure 2.2 Experimental setup for the optical compensation technique. 2.4 Results and Discussions Several experiments were performed using different PMD emulator configurations. Initially, the emulator simply consisted o f a PC followed by a single section of polarization maintaining (PM) fiber to generate only first-order PMD (DGD). Different fiber lengths were used to obtain different DGD values. By tuning the PC, the optical input was launched into the PM fiber with 50/50 splitting between the two PSPs. The subcarrier frequency was set to 20 GFlz. The optical spectaim of the light is shown for different points in the setup in Figure 3.3. The spectrum of the 27 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. signal exiting the emulator is shown in Figure 2.3 (a). The spectra o f the carrier and the sideband after being separated by the FBG filter are shown in Figure 2.3 (b) and (c), respectively. We can see that the filter reduces the power o f the optical carrier in the spectrum of the SSB by about 25 dB. After the SOP o f the optical carrier is aligned with the SSB, the optical carrier and the SSB are recombined and the spectrum is shown in Figure 2.3 (d). When an equal amount of optical power is distributed between the fast and slow PSPs, RF power fading at the receiver can be calculated using the following equation [24]: F = cos (7rf*Ax) (2.i) where F is the RF power fading factor (F = 0 for complete fading), f is the frequency of the subcarrier, and At is the DGD value. Figure 2.4 shows the measured RF power fading due to first-order PMD with and without compensation. The theoretical curve is plotted as a solid line. W ithout compensation, the RF power fading is > 3 dB when the DGD value is >12.5 ps and increases rapidly when the DGD value is close to 25 ps (corresponding to A the period o f the subcarrier). However, after compensation, the RF power fading is reduced to less than 1 dB. 28 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. After Emulator -20 C 3 " 3 0 O o -40 -50 1549 1549 1550 1550 1551 W avelength (nm ) (a) Carrier -30 -40 - 50'-------- 1549 1551 1549 1550 1550 W avelength (nm ) (b) SSB -1 0 -20 £ O ^ '3 0 13 o -50-------- 1549 1549 1550 1550 1551 W avelength (nm ) (c) P Q £ o O h _ o + 3 a O Recombined 1549 1550 1550 1551 W avelength (nm ) (d) Figure 2.3 Optical Spectrum o f (a) the light at the PMD emulator output; (b) the optical carrier after FBG filtering; (c) the SSB after passing through the FBG; (d) after recombination of the optical carrier and the SSB using an optical coupler. 29 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 0 f e t) .3 -10 -o C S t. -15 £ o ^ -20 0 10 20 30 40 50 DGD (ps) Figure 2.4 Power fading due to first-order PMD and the compensation results. Solid line shows the theoretical RF fading curve. Compensated fW/O Compensation To investigate the higher-order PMD compensation ability of this new technique, we used a PMD emulator with 15 sections o f PM fiber separated by polarization controllers that are randomly varied between each experimental sample. This emulator will generate both Maxwellian distributed DGD values as well as higher-order PMD effects [25], Furthermore, the emulator could be configured to yield an average DGD o f either 31 or 42 ps. The amount RF fading was measured for 300 independent samples for both o f these average DGD values as well as two subcarrier frequencies, 18 and 20 GHz, as shown in Figure 2.5 (a - d). Additionally, a computer simulation was used to generate the “theoretical distribution” curve from a set of 10,000 random samples. 30 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 99.999 99.99 99.9 99 95 90 : 80 i 70 ; 50 30 20 10 5 • a i £ - < D C D > -42 ps f=20 G H z 300 ex perim ental sam ples w/o com pensation T h eoretical D istribution w/o com pensation .1 .01 .001 300 ex perim ental sam ples w/ com pensation -35 -30 -25 -20 -15 -10 -5 Power Fading (dB) c 4 ) (J s- a- 99.999 99.99 99.9 ; 99 I 95 j 90 80 ! 70 : 50 | 30 ■ 20 10 5 1 i .1 .01 .001 <DG D >=42 ps 1=18 G H z 300 experim ental sam ples w/o com pensation T heoretical D istribution w /o com pensation 300 experim ental sam ples w / com pensation -35 -30 -25 -20 -15 -10 Power Fading (dB) (a) (b) 4 > O U C m 99.999 99.99 99.9 99 . 95 90 80 ; 70 50 30 20 1 0 5 1 <DCD >=31 ps 1-20 G H z 300 experim ental sam ple w/o com pensation T h eoretical D istribution w/o com pensation .1 .01 .001 300 ex perim ental sam ples w /co m p en satio n -35 -30 -25 -20 -15 -10 -5 Power Fading (dB) 99.999 99.99 99.9 99 S3 4 > O 95 90 80 70 50 30 20 10 ; 5 ■ ' <DGD>=31 ps f=18 G H z 300 experim ental sam ples j w/o com pensation j^i T heoretical D istribution w/o com pensation .1 .01 .001 300 experim ental sam ples vv/ com pensation -35 -30 -25 -20 -15 -10 -5 Power Fading (dB) (c) (d) Figure 2.5 Power fading distribution after the PMD emulator with and without optical compensation: (a) average DGD=42 ps, subcarrier frequency=20 GHz; (b) average DGD=42 ps; subcarrier frequency=18 GHz; (c) average DGD=31 ps, subcarrier frequency=20 GHz; (d) average DGD=31 ps, subcarrier frequency=18 GHz . 31 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Ideally the distributed RF power fading can be totally compensated to 0 dB using our technique. The results show that without compensation, 5% o f the samples exhibit more than 13.5 dB o f RF power fading. The experimental results differ a little from the simulation results because only 300 samples were taken during the experiments. Using our technique, all o f the compensated data exhibit less than 1.5 dB RF power fading, which may be caused by a small residual misalignment left between the SOPs o f the optical carrier and the SSB. Therefore, the 5% power fading tail is improved by more than 12 dB using our compensation technique. By extrapolation, the improvement will be even greater for lower probability areas of the tail. The performance of this technique is independent o f the average DGD of the optical fiber link and the subcarrier frequency. Figure 2.6 shows the measured bit error rate (BER) versus the received optical power for a 155 Mbit/s binary-phase-shift-keyed (BPSK) signal modulated onto the 20 GHz subcarrier. The back-to-back BER is measured without the PMD emulator. The average DGD o f the PMD emulator was 31 ps. We choose an arbitrary sample for which the RF fading was approximately 14 dB at the emulator output. Using this sample, we could not get a measurable BER directly after the PMD emulator, since the data is almost entirely lost without compensation. On the other hand, the power penalty is around 0.4 dB with compensation. This penalty is consistent with the calculation result o f the power penalty induced by the optical interference effect when the optical carrier and the SSB branches are recombined. 32 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Since the power o f the optical carrier in these two branches differs by 25 dB, the effect is small. B ack-tb back C om pensated -6.5 -8.5 R eceived optical Pow er (dB m ) -7.5 Figure 2.6 Measured BER vs. received optical power for 155 Mbit/s BPSK signal at 20 GHz. The compensated signal exhibited 14 dB o f RF fading before compensation. 2.5 Summary In this chapter, we propose and demonstrate a novel technique for optically compensating the PMD-induced RF power fading that occurs in single-sideband (SSB) subcarrier-multiplexed systems. By aligning the polarization states of the optical carrier and the SSB, RF power fading due to all orders o f PMD can be completely compensated. The 5% RF power fading tail is improved from 13.5 dB to <1.5 dB, as verified from both experimental measurement (300 samples) and computer simulation (10,000 samples). This technique is independent o f the DGD of the optical fiber link and the subcarrier frequency. 33 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Chapter 3 Chromatic-Dispersion-Insensitive PMD Monitoring for NRZ Data Based on Clock Power Measurement In this chapter, we demonstrate chromatic-dispersion-insensitive PMD monitoring by using a narrowband FBG notch filter to recover the RF clock power for lOGb/s NRZ data, and apply it as a control signal for PMD compensation. 3.1 Chromatic-Dispersion-Insensitive PMD Monitoring As we mentioned in Chapter 1, high-bit-rate transmission systems (>10 Gb/s/channel) are highly susceptible to deleterious optical-fiber-based effects, such as chromatic dispersion (CD), polarization-mode dispersion (PMD), and nonlinearities. In particular, PMD accumulates due to either high-PM D legacy fiber or PMD o f many in-line components. Deleterious PMD effects are stochastic, time varying, and temperature dependent. M oreover, the instantaneous first-order PMD (i.e., differential group delay (DGD)) follows a Maxwellian probability distribution, always with some finite possibility of network outage. 34 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. A key challenge for systems deployment is that these effects are not static but change with time, including: (i) temperature changes, (ii) reconfigurable optical networking, (iii) wavelength drifts, and (iv) periodic repair and maintenance. These degradations may require the monitoring of signal quality in order to either dynamically tune a compensator or simply to determine the network location that must be diagnosed and repaired. One straightforward method o f monitoring optical signal quality is to electronically determine either the Q o f the eye diagram or actually measure the bit- error-rate. Unfortunately, this approach cannot distinguish between the various effects that may cause signal degradation. Several types o f PMD monitors have been reported, including: (a) adding a subcarrier tone [26], (b) measuring the signal’s degree-of-polarization (DOP) [27], (c) spectral analysis o f the detected signal [28], and (d) measuring the chromatic-dispersion-generated clock tone [29]. However, each o f these techniques suffers from one or more of the following disadvantages: (i) high cost and complexity, (ii) necessitating transmitter modification, (iii) low sensitivity, and (iv) sensitive to CD. In this chapter, we propose and demonstrate chromatic-dispersion-insensitive PMD monitoring o f non-return-to-zero (NRZ) data based on clock power measurement using a narrowband fiber Bragg grating (FBG) notch filter. The clock tone does not appear at the receiver for NRZ data. Using a narrowband FBG notch 35 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. filter to filter off one o f the optical clock sidebands, the RF clock tone can be recovered from the beating between the carrier and the remaining optical clock sideband. The recovered RF clock power depends on the relative polarization state of the carrier to the optical clock sideband, which is determined by PMD of the transmission link. CD only affects the phase o f the recovered RF clock tone but not the amplitude. Therefore, the recovered RF clock power can be used as a PMD monitoring signal, and is insensitive to CD. Using a FBG notch filter with a 10-dB bandwidth o f 15 GHz at the receiver, we measured the recovered clock power to monitor PMD in a lOGb/s NRZ system. The variation of the detected clock power is within 1.5 dB when the accumulated dispersion increases from 0 to 600 ps/nm. For 300 independent samples using the 10 GHz recovered clock as a feedback control signal to a PMD compensator, the 5% worst case value o f the power penalty is reduced from 6.0 dB to 1.5 dB. 3.2 Clock Power Recovered Using Notch Filter for NRZ Data Figure 3.1 shows the concept for using a narrowband FBG notch filter to recover the RF clock power and use it for PMD monitoring. Ideally the clock tone does not appear at the receiver for NRZ data. However, using a narrowband FBG notch filter to filter off one o f the optical clock sidebands, the RF clock tone can be recovered from the beating between the carrier and the remaining optical clock sideband. The recovered RF clock power depends on the relative state o f polarization 36 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. (SOP) o f the carrier to the optical clock sideband. At the transmitter, the optical carrier and the optical clock sideband have the same SOP. After propagating through the optical fiber link, PMD cause the SOPs of the optical carrier and the optical clock sideband to vary by different amounts. Only the portion o f the clock sideband that has the same polarization as the carrier can effectively interfere, so the resulted RF clock power is determined by PMD of the transmission fiber link. On the other hand, CD can only affect the phase o f the recovered RF clock tone but not the amplitude. Therefore, the recovered RF clock power can be used as a PMD monitoring signal, and it is insensitive to CD. Receiver F B G N otch F ilter Transmitter Optical Spectrum ( ( 0 f w/o PMD F B G N otch Filter * f Electrical Spectrum V nr\ w/ PMD \ / i i X Figure 3.1 Concept o f PMD monitoring for NRZ data based on the recovered clock using a narrowband FBG notch filter. Figure 3.2 shows the experimental setup. After propagating through the transmission link, which consists o f a single mode fiber (SMF) and a PMD emulator, 37 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1% o f the 10 Gb/s NRZ signal is tapped off for PMD monitoring. A FBG notch filter with a 10-dB bandwidth of 15 GFIz is placed in front o f a photodetector to filter off the upper optical clock sideband in order to recover the RF clock power. When both CD and DGD are zero, as shown in Figure 3.3, there is no clock power without filter, but the clock is regenerated by 28 dB with the fdter in place. Figure 4.4 shows the corresponding optical spectrum. Fiber Link Tx d PMD Emulator PMD Comp. Data Rx FBG Notch Filter — tt+tt— RF Analyzer PMD Monitoring i i i Feedback i i Figure 3.2 Experimental setup. -55 w/o Filter -60 « TJ -80 -85 -55 w/ Filter r-v " 60 P Q w -65 U ^ -70 O to t o -75 -80 10 G oo n -85 (b) (a) Figure 3.3 Electrical spectrum: (a) w/o and (b) w/ fdter. 38 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. I ' 5 s- I -2 5 "ft a -45 O w 'o F ilter 0.1 n in / \i iV-N 1 „ I ‘ 5 a > I ’2 5 f t 73 (J ft -45 O \\7 Filter C arrier +10 G .. ■ 1 . I W (a) (b) Figure 3.4 Optical spectrum: (a) w/o and (b) w/ filter. We perform both first-order and all-order PMD monitoring. For first-order PMD (DGD), we launch the input signal to a first-order PMD emulator, which consists o f a polarization controller (PC) and a piece o f polarization maintaining (PM) fiber. The power splitting ratio between the two principle-polarization-states is 0.5. Figure 3.5 shows the simulation and experimental results o f the relative recovered RF clock power changing with different DGD value. When we change the length of the SMF from 0 to 35 km. (i.e. the corresponding CD increases from 0 to 600 ps/nm.) The variation o f the RF clock is < 1.5 dB. 39 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. p q o w -4 U a> £ -8 O Ph -12 g -16 ^ -20 0 > -24 a 13 -28 p2 0 10 20 30 40 50 DGD (ps) Figure 3.5 Relative clock power as a function of DGD for different CD. To investigate all-order PMD monitoring o f this technique, we then used a PMD emulator which has 30 sections o f PM fiber with polarization controllers distributed between the sections to realize different polarization coupling and therefore emulate both M axwellian distribution of DGD and higher-order PMD effects [30]. The average DGD o f the PMD emulator is 42 ps. We take 300 samples with CD at 0 and 600 ps/nm. As shown in Figure 3.6, the distribution o f the relative recovered clock power is essentially unchanged as the CD varies. And the experimental distributions are close to the 10000 samples simulation results. 40 P CD = 0 • CD = 600 Lines: Simulation results I ® . Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 5 1 .1 .01 .001 ■ CD = 600 ps/nm Line: Simulation results _l 1 .1 , I, I I I I I I— 1 I ■ I . I I . -1 L. -30 -25 -20 -15 -10 -5 0 Relative Clock Power (dB) Figure 3.6 Relative clock power distribution for CD = 0 and CD = 600 ps/nm. 3.3 PMD Compensation Using the Recovered Clock Power as a Feedback Signal Using the recovered clock power as a feedback signal, we compensate PMD in the transmission link (with an average DGD o f 42 ps) using a PC and a piece of 50-ps PM fiber. The PC is adjusted to maximize the recovered clock. As shown in Figure 3.7, for 300 experimental samples, the 5 % worst case o f the power penalty distribution tail is reduced from 6 dB to 1.5 dB using this monitoring technique. 41 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 5% W orst Case: 6 dB 5% W orst Case 1.5 dB Power Penalty (dB) Power Penalty (dB) (a) (b) Figure 3.7 Power Penalty distribution: (a) without and (b) with compensation. 3.4 Summary In this chapter, we propose and demonstrate chromatic-dispersion-insensitive PMD monitoring by using a narrowband FBG notch fdter to recover the RF clock power for lOGb/s NRZ data. Using a FBG notch fdter with a 10-dB bandwidth o f 15 GHz at the receiver, we measured the recovered clock power to monitor PMD in a lOGb/s NRZ system. The variation o f the detected clock power is within 1.5 dB when the accumulated dispersion increases from 0 to 600 ps/nm. Applying the recovered clock power as a control signal for PMD compensation, for 300 independent samples, the 5% worst case value o f the power penalty is reduced from 6.0 dB to 1.5 dB. 42 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Chapter 4 40-GHz RZ and CS-RZ Pulse Generation Using a Phase Modulator and PM Fiber In this chapter, we demonstrate chirp-free RZ and CS-RZ pulse generation with a repetition rate of 40 GHz using a phase modulator driven by a 20 GHz clock and a single piece o f polarization maintaining (PM) fiber. 4.1 Generation of High-Speed RZ and CS-RZ Optical Pulse Train The generation o f a high-speed optical pulse train is critical for many applications, including transmission and signal processing. Specifically, retum-to- zero (RZ) modulation formats can be created from pulse trains and have been shown to be robust to fiber-based degrading effects for many high-speed, long-distance systems. This applies to conventional RZ as well as carrier-suppressed RZ (CS-RZ). The most common method for generating a high-speed optical pulse train for ultimately producing RZ modulated data is to use a M ach-Zehnder intensity modulator that is driven by an electrical clock at the desired data rate (e.g., 40 GHz for a 40-Gbit/s data channel) [31]; note that CS-RZ is produced using an intensity modulator driven at half the bit rate with a different bias [32], This technique 43 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. requires the use o f a high-speed modulator and driver for RZ generation and tends to produce pulses that are chirped. Recently, two methods have been reported for generating a high-speed optical pulse train [33,34], which employs a phase modulator that is driven by a clock at only half the bit rate. The first method [33] used an optical filter after the phase modulator to convert 20-GHz phase modulation into 40-GHz amplitude modulation, but it induces chirp in the optical pulses. The second method [34] used a M ach-Zehnder interferometer after the phase modulator, also translating phase modulation into amplitude modulation, but without any induced chirp. However, the employment o f a planar waveguide interferometer adds complexity and cost to the implementation. We demonstrate a simple technique that generates an RZ and a CS-RZ 40- GHz optical pulse train from a 20-GHz electrical drive clock. We use a phase modulator that is followed by a single piece o f polarization-maintaining (PM) fiber. After 20-GHz sinusoidal phase modulation, the light is split equally into the two principal-states-of-polarization (PSPs) of the PM fiber. Differential group delay (DGD) provides a one-bit time shift (25 ps) between the two polarization components o f the light. At the output of the PM fiber, due to beating o f the two replicas o f the light, one polarization, aligned 45° with respect to the PSPs, generates an RZ pulse train, whereas the orthogonal polarization generates a CS-RZ pulse train 44 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. with a repetition rate of 40 GHz. Based on fundamental principles, the generated pulse trains are chirp free. This technique may be cost-effective and seamlessly integrated with optical fiber systems. 4.2 Concept of Pulse Generation Using a Phase Modulator and PM Fiber Figure 4.1 shows a conceptual diagram of RZ and CSRZ pulse generation using a phase modulator and a single piece o f PM fiber. Phase Mod. Light beating Intensity ^ A A A A A PM Fiber Polarizer Pulse Train Figure 4.1 Concept o f RZ/CS-RZ pulse generation using a phase modulator and PM fiber. For 40 GHz pulse generation, f= 4 0 GHz and T= l/f=25 ps. After phase modulation by a 20-GHz (f/2) sinusoidal clock tone with a modulation depth (peak to peak) o f 7i, the light is aligned at 45° relative to the PSPs o f the PM fiber, so that the power splitting ratio between the two principle-polarization-states is 0.5. The light 45 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. splits equally into the two PSPs o f the PM fiber with a DGD o f 25 ps. The DGD makes a one-bit time shift (T=25 ps) between the two polarization components of the light. The optical fields o f these two replicas can be described by the following equations: A ,71 . T C t . £, = — prcost—sin----1 - cot) 1 V 2 2 T v A E 7 = — j= c o s 4 2 7 1 . 7 l { t - T ) — sin------------i- cot 2 T A 7 1 . 7 lt -cos( sin vcot) 4 2 2 T At the output o f the PM fiber, after a polarizer aligned at 45° relative to the PSPs, the two replicas o f the light beat together, producing the following optical field: E , + E 2 ,71 . 7tt. , . £3 = — c x A cos(“ sin “ ) cos(ty?) (4.2) The resulting field has the characteristics of a 40-GHz chirp-free 33% RZ pulse train. With the polarizer aligned to the orthogonal polarization direction (-45° relative to the PSPs), the output optical field can be expressed as: E — E T C T C t £ 4 = 1 2 o c A sin(“ sin ~ ) sin(&tf) ( 4 3 ) This field has the characteristics o f a 40-GHz chirp-free 67% CS-RZ pulse train. Figure 2 shows the simulation results for waveforms of the generated RZ and CS-RZ pulse trains. 46 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1.2 3 1 'S ' 0.8 0.6 (/) 0.4 fl 0.2 o > 0 3 -0.2 0 20 40 60 80 100 Time (ps) « 0.8 20 40 60 80 100 Time (ps) (a) (b) Figure 4.2. Simulation results for waveforms o f pulse trains: (a) RZ and (b) CS- RZ. To further prove that the generated pulse trains are chirp free, we simulated 40- Gb/s data transmission using the generated RZ and CS-RZ pulse trains as the 40- GHz pulse train sources. We modulate 40 Gb/s (223-l) PRBS data and transmit it through different lengths of single mode fiber (SMF) with a dispersion o f 16 ps/nm/km. Figure 4.3 shows the Q-penalty as a function o f dispersion value. For comparison, we also simulated the transmission o f 33% theoretical RZ data with a rise time of 'A bit time. We can see that the generated 33% RZ pulse has quite similar performance to the purely theoretical 33% RZ. The small difference is caused by pulse shape differences between the generated RZ pulse and the theoretical 33% RZ pulse. 47 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 8 Theoretical 33% RZ Generated 33% RZ Generated 67% CS-RZ M T5 Oil 0 10 20 30 40 50 Dispersion (ps/nm) Figure 4.3 Q-penalty vs. dispersion through SMF. 4.3 Experimental Setup The experimental setup is shown in Figure 4.4. The phase modulator is driven by a 20 GHz clock tone. This is followed by a polarization controller (PC) used to align the light to be along the direction of 45° with respect to the PSPs of the PM fiber, which has a DGD of 25 ps. Another PC is used to align the polarization beam splitter (PBS) to also be 45° to the PSPs of the PM fiber. The generated RZ pulse train is obtained at one output port of the PBS and the generated CS-RZ is obtained at the other output port. Note that the PBS is only used here as a polarizer in order to observe the generated pulse trains in the experiment. In a real application, a LiNbCfi intensity modulator, following the pulse generator, can perform this 48 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. function while simultaneously modulating data onto the pulse train. And both of the PCs could be eliminated by 45° splicing the PM fiber. 20 G Clock PC 25 ps PM fiber PC PC PBS Tunable non Phase non v J non / 40G RZ pulse Train Laser Mod. r 40G CS-RZ pulse Train Figure 4.4 Experimental setup for 40G RZ/CS-RZ pulse train generation 4.4 Results and discussion Figure 4.5 shows the optical spectrum of the generated RZ and CS-RZ pulse trains. We can see that the unwanted 20 GFlz tone in the spectrum o f the RZ pulse train is suppressed by more than 20 dB, while the optical carrier is suppressed by more than 30 dB in the spectrum of the CS-RZ pulse train. The residual 20 GHz clock components may be caused by misalignment o f the PM fiber. 49 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. (dBm) 0 -10 -20 -30 -40 (dBm) 0.32 nm ■4------- ► (40 (ill/.) 0.2 nm 1550 nm 0.32 nm (40 (ill/) -10 0.2 nm -20 -40 1550 nm (a) (b) Figure 4.5. Optical spectrum of pulse trains: (a) RZ and (b) CS-RZ. Figure 4.6 shows the waveforms of the generated RZ and CS-RZ pulse trains observed by a 40G photodiode. The observed waveforms are not identical to the simulation results in Figure 2 due to the bandwidth limitation o f the photodiode. 10 ps '"G A ■ ; .a ’ GG a '■ ■ ■ /S GG \ H V i i l ¥ / Vw' ‘ v n ; I G 10 ps G _ G v i c a n y G’ G / / G v >/ G v: G G (a) (b) Figure 4.6. Waveforms of pulse trains: (a) RZ and (b) CS-RZ. 50 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Note that a similar scheme with PM fiber can replace one-bit-delay interferometer at the receiver as a decoder for an optical DPSK transmission system. 4.5 Summary In this chapter, we demonstrate a simple technique that generates an chirp-free RZ and a CS-RZ 40-GHz optical pulse train from a 20-GHz electrical drive clock. We use a phase modulator that is followed by a single piece o f polarization- maintaining (PM) fiber as a one-bit delay interferometer, the unwanted 20 GHz tone in the spectrum o f the pulse train is suppressed by more than 20 dB. 51 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Chapter 5 160-GHz Pulse Generator Using a 40-GHz Phase Modulator and PM Fiber In this chapter, we extend our work in the last chapter to demonstrate chirp- free CS-RZ pulse generation with a repetition rate o f 160 GHz, which is four-fold of the frequency o f the electrical clock. 5.1 Generation of High-Speed Optical Pulse Train beyond 40 GHz An ever-growing fraction of the research-and-development community in optical communications is performing experiments at speeds higher than 40-Gbit/s. An invaluable element for any ultra-high-speed system is a pulse-train generator at the data clock speed or bit rate. Such a generator can be used as an optical clock, for optical sampling, or to imprint optical data bits. Presently, the ready availability o f a low-cost high-speed pulse train has been elusive. There are various methods that can generate high-speed pulse trains. The most common method is to use an actively-mode-locked laser, either by mode- locking at a very high rate (i.e., 140 GHz) [35] or by passively splitting/delaying/multiplexing a very-short-pulse train [36], In general, mode- 52 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. locked lasers are not considered low-cost elements. Other published methods include: tunable-rate pulse generation using a specially designed dual-wavelength DFB laser diodes [37]; the rate multiplication process through the temporal fractional Talbot effect o f chirped pulses[38]; and spectral selection by arrayed waveguide gratings [39], where each burst o f pulses lies at different wavelength; using spectral selection by a Fabry-Perot (FP) optical filter to multiply 10 GHz pulse train to 40 GHz, which needs either high finesse or additional components such as semiconductor optical amplifiers [40]. In last chapter, we have demonstrated the technique that generates chirp-free return-to-zero (RZ) and carrier-suppressed RZ (CS-RZ) optical pulse trains at a repetition rate that is double the frequency o f the electrical clock. In that method, a phase modulator and polarization-maintaining (PM) fiber are used in a single-stage subsystem to double 20-GHz phase modulation to 40-GHz pulse trains. In this chapter, we significantly extend our previous work to demonstrate 160-GHz chirp-free pulse generation at a repetition rate which is four-fold of the frequency o f the electrical clock. To achieve this, we use a two-stage subsystem that uses a 40-GHz phase modulator plus two low-cost PM fibers and two polarizers. The mechanism for the four-fold increase in the two stages is as follows: an 80-GFIz optical pulse train is generated by constructive beating of a 40-GHz phase modulated light at the first stage, then it is split equally into the two principal-states-of- 53 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. polarization (PSPs) of the PM fiber at the second stage. Differential group delay (DGD) provides a one-bit-time shift (6.25 ps) between the two polarization components o f the light. At the output o f the PM fiber, due to the destructive beating of the two replicas o f the light, one polarization, aligned -45° with respect to the PSPs, generates a 50% CS-RZ pulse train with a repetition rate o f 160 GHz. The unwanted low frequency tones are suppressed by more than 15 dB, and the measured pulse width is 3.3 ps. Again, we emphasize that our method has the potential to be a cost-effective source of high-speed optical pulses, which can be seamlessly integrated with optical fiber systems. 5.2 Concept of Pulse Generation with a Four Fold Repetition Rate Figure 5.1 shows a conceptual diagram o f an optical CS-RZ pulse generation at a repetition rate which is four-times o f the frequency o f the electrical clock using a phase modulator and two PM fibers and two polarizers. 54 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. //4 Phase Mod. f= \ 60 GHz Phase T=l/f=6.25 ps JJ4 Clock //2 RZ Amplitude / CSRZ Intensity i 4 T : Laser I 1- *; i - Phase 2T PM Fiber Mod. 1st Stage \45° -45°/ 2T PM Fiber 2n d Stage Figure 5.1 Concept o f CS-RZ pulse generation with a four fold repetition rate using a phase modulator and PM fiber. Figure 5.2 shows details o f the first and second stage o f PM fiber in the setup. 55 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. //4 Phase Mod. Phase Phase Mod. Light 2T pep / i / phase f] 2 RZ PM Fiber t \PSP t 4T: beating Amplitude A A A A 1st Stage phase Polarizer 2T (a) f/2 RZ Amplitude Amplitude beating * A A A 45« 2T PM Fiber 2n d Stage 2T Polarizer / CSRZ Intensity |\ (b) Figure 5.2 Details o f (a) the first and (b) the second stage o f PM fiber in the setup. For 160-GHz pulse generation, / = 160 GHz and T - l/f- 6 .2 5 ps. After phase modulation by a 40-GFIz (f/4) sinusoidal clock tone with a modulation depth (peak to peak) o f 7i, at the first stage, the light is aligned at 45° relative to the PSPs o f the first PM fiber, so that the power splitting ratio between the two principle-polarization- states is 0.5. The light splits equally into the two PSPs o f the PM fiber with a DGD of 2T (2T=12.5 ps). The DGD makes a two-bit time (27) shift between the two 56 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 71 polarization components o f the light. Similar to Chapter 4, the optical fields of these two replicas can be described by the following equations: A 7 1 . 7 Z t E, = — p=cos(— sin------f - cat); V2 2 2T _A_ cos n . n { t - 2T) —sin--------------+ cot 2 2 T A , n . 7ti = — p=cos(— sin------h cot) 4 l 2 2T (5.1) At the output of the PM fiber, after a polarizer aligned at 45° relative to the PSPs, the two replicas of the light beat together, producing the following optical field: Ei “t E ■ y tz . jti . , . if, = — _ oc A cos(—sin — )cos(m t) -v 2 2 2T (5.2) The resulting field has the characteristics o f an 80-GHz chirp-free 33% RZ pulse train. Note that if the polarizer aligned to the orthogonal polarization direction (-45° relative to the PSPs), the output optical is an 80-GHz chirp-free 67% CS-RZ pulse train. In the second stage, the 80-GHz RZ pulse train is again aligned at 45° relative to the PSPs of the second PM fiber, so that split equally into the two PSPs of the PM fiber with a DGD o f T (T=6.25 ps). The DGD makes a one-bit time (7) shift between the two polarization components o f the light. At the output o f the PM fiber, 57 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. after a polarizer aligned at -45° relative to the PSPs, the two replicas o f the light destructively interfere with each other, producing the following optical field: E\ = f{t) oc A , 7 1 . T C t ,71 T C t . cost —sin— ) - cos(—cos— ) 2 I T 2 2 T cos{oot) (5.3) W hich is a 160-GHz 50% CS-RZ pulse train with a period o f T (T=6.25 ps). And it is chirp free as we can see from the equation. Figure 5.3 shows the waveform simulation results o f the generated 80-GHZ 33% RZ and 160-GHz CS-RZ pulse trains. 58 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 s a O 50 1 0 0 O S 3 & o o Tim e (ps) (a) " l l ..I1 , i 1 . U .j ... I ' . 50 Tim e (ps) 100 (b) Figure 5.3 Simulation results for waveforms o f pulse trains: (a) 80-GHz 33% RZ and (b) 160-GHz 50% CS-RZ. 5.3 Experimental Setup The experimental setup is shown in Figure 5.4. The phase modulator is driven by a clock tone with a frequency o ff/4 . This is followed by a polarization controller (PC) used to align the light to be along the direction of 45° with respect to the PSPs of the PM fiber, which has a DGD o f 2 T. Another PC is used to align the 59 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. polarization beam splitter (PBS) to also be 45° to the PSPs o f the PM fiber. The generated RZ pulse train with a repetition rate off/2 is obtained at one output port of the PBS. After the first stage, another PC is used to align the generated f/2 optical RZ pulse train to be 45° to the PSPs of the PM fiber, and another PC is used to align the polarizer to be -45° to the PSPs of the second PM fiber. The generated CSRZ pulse train with a repetition rate o f/ is obtained after the polarizer. Note that in a real application, all the PCs could be eliminated by 45° splicing the PM fiber. 40 G C lo c k 8 0 G R Z P u ls e T ra in 1 6 0 G C S R Z P u ls e T ra in 6 .25 ps P M fib e r 12.5 ps P M fib e r P C n n P h a se M o d . L a s e r c Stage 1 Stage 2 Figure 5.4 Experimental setup for 160G CR-RZ pulse train generation 5.4 Results and discussion As proof o f four-fold frequency multiplication effect in our technique, first we demonstrate 40-GHz CS-RZ (f-4 0 GHz) pulse generation by providing a 10- GFIz clock to the phase modulator and using DGD value o f 50 ps and 25 ps for the two PM fibers respectively. Figure 5.5(a) shows the optical spectrum of the generated 40-GHz RZ pulse trains. We can see that the unwanted lower tones (10 & 20 GHz) in the spectrum are suppressed by about 23 dB, Figure 5.5(b) shows the waveform of the generated pulse trains observed by a 40G photodiode. 60 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. J(K 5 t ' S R / Pulse 0 .4 II til (L>J ii in (4(K. ) 10 ps LA. 1 54'J.X inn (a) (b) Figure 5.5 Generated 40G CS-RZ pulse trains:(a) Optical spectrum and (b) waveform Then we demonstrate 160 GHz (f=160 GHz) pulse generation by modulating the phase using a 40-GHz clock and replacing the two PM fibers in the setup with DGD values o f 12.5 ps (27) and 6.25 ps (7) respectively. Figure 5.6 (a) shows the optical spectrum of the 80-GHz 33% RZ pulse train at the output of the first stage, and we can see that the unwanted 40GHz tone in the spectrum o f the RZ pulse train is suppressed by about 17 dB. We also measure its pulse width by using an autocorrelator with a scale factor o f 7.41 ps/ms, and the result is shown in Figure 5.5(b). By using the pulse shape factor o f 0.71, we get measured pulse width of 4.5 ps, which is close to theoretical value o f 4.16 ps. 61 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. I im 0.4 mu (a) (b) Figure 5.6 Generated 80G RZ pulse trains at the first stage:(a) optical spectrum and (b) autocorrelator measurement. Then we measure the properties o f the 160-GHz 50% CSRZ pulse train at the output o f second stage. Figure 6 (a) shows its optical spectrum: the unwanted lower tones (40 & 80 GHz) are suppressed by about 15 dB. The residual lower clock components may be caused by misalignment of the PM fibers. Figure 6 (b) shows the pulse width measurement result obtained using the autocorrelator. By using the pulse shape factor of 0.75, we get measured pulse width o f 3.3 ps, which is also close to theoretical value of 3.12 ps. 62 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. - i n tfLLut -SO i (a) (b) Figure 5.6 Generated 160G CS-RZ pulse trains at the second stage: (a) optical spectrum and (b) autocorrelator measurement. 5.5 Summary In this chapter, we demonstrate chirp-free CS-RZ pulse generation with a repetition rate o f 160 GFIz using a phase modulator driven by a 40 GFIz clock and two low-cost polarization-maintaining fibers. The unwanted low frequency tones are suppressed by more than 15 dB. The measured pulse width is 3.3 ps. 63 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Chapter 6 Polarization-Insensitive All-Optical Wavelength Conversion Using Dispersion-Shifted Fiber with a Fiber Bragg Grating and a Faraday Rotator Mirror In this chapter, we demonstrate a simple technique for polarization- insensitive all-optical wavelength conversion based on four-wave mixing in dispersion-shifted fiber (DSF) with a fiber Bragg grating and a Faraday rotator mirror. 6.1 Four-Wave Mixing Wavelength Conversion and Polarization Insensitive Operation In high-speed wavelength-division multiplexed (WDM) optical networks, wavelength conversion is an important function for translating data carried on one wavelength to another, to reduce wavelength blocking and provide more flexibility in network management [41]. All-optical wavelength conversion based on four wave mixing (FWM) in optical fibers has the following potential advantages: (i) it eliminates optical-electrical-optical (O/E/O) conversion and thus enables transparent all-optical networks, (ii) it is ultra-fast and transparent to both modulation format and 64 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. bit rate, (iii) it induces negligible signal degradation since there is little chirp or added noise, and (iv) the optical fiber itself is low cost, low loss, and seamlessly compatible with the transmission fiber. However, FWM in fiber strongly depends (i.e., >20 dB) on the relative state-of-polarization (SOP) o f the signal relative to the pump. Therefore, polarization-insensitive operation is essential for any future application o f fiber FW M -based wavelength conversion since there is generally no control o f the signal’s polarization in real optical networks. There are several techniques reported for polarization-insensitive FWM wavelength conversion in fiber: (i) a polarization diversity method using a fiber loop with a beam splitter which requires an in-line tunable polarization controller [42], (ii) a scheme using non-degenerate FWM operation that requires two orthogonal pumps of different wavelengths [43], and (iii) a method using a pump composed o f cross polarized high frequency pulses in which the pump is modulated at speeds much higher than the data rate [44], There are also reports using Faraday rotator mirrors (FRM) for polarization-insensitive phase conjugation in fiber [45] and wavelength conversion in periodically poled lithium niobate (PPLN) waveguide [46]. However, a high-speed wavelength conversion using fiber with an FRM has yet to be studied. We demonstrate a simple technique to minimize polarization sensitivity in a fiber-based FWM wavelength converter using a fiber Bragg grating (FBG) and an FRM. The FBG is set to reflect only the pump wavelength without changing its 65 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. polarization state and pass the signal wavelength to the FRM, which rotates the signal’s polarization state by 90°. Both the pump and the signal make a dual pass through dispersion-shifted fiber (DSF). However, the rotation o f the signal polarization on the return pass guarantees that both orthogonal polarization components o f the incoming signal wave will efficiently mix with the pump to produce a polarization-insensitive wavelength-converted output. We experimentally demonstrate that the residual polarization sensitivity is reduced from >22 dB to 2 dB with 4 km o f DSF. The power penalty incurred in wavelength conversion is less than 1 dB for a 20-nm conversion distance, and the difference in power penalty for different polarization states is less than 0.3 dB. Theoretically, this technique will have negligible polarization sensitivity when the propagation loss o f the pump through the DSF approaches 0. 6.2 Concept of Polarization-Insensitive Technique Using an FBG and an FRM Figure 6.1 shows the concept o f the technique for polarization-insensitive FWM wavelength conversion using an FBG and an FRM. The FBG is designed to only reflect the pump wavelength without changing its polarization state; the signal wavelength is outside the reflection band of the FBG and passes through to the FRM, where the signal is reflected and the w ave’s polarization state is rotated by 90°. Both the pump and the signal make a dual pass through the DSF. Only the polarization 66 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. component o f the signal wave that is the same as the polarization state o f the pump contributes to wavelength conversion during a single-pass through the DSF. Flowever, the rotation o f the signal polarization on the return pass guarantees that both orthogonal polarization components o f the incoming signal will efficiently mix with the pump during either the forward pass or backward pass through the DSF. Thus the wavelength-converted output will be polarization-insensitive. Note that this method can extend to multi-channel wavelength conversions, since it does not require controlling the polarization state of each input signal channel. Signal F o r w a r d . N DSF Pump w/o Rotation f b G Faraday Rotator Mirror B a c k w a r d Signal (Rotated 90°) Figure 6.1 Concept for polarization-insensitive FWM wavelength conversion. 6.3 Experimental Setup The experimental setup is shown in Figure 6.2. The polarization controller (PC) aligns the pump light to be linearly polarized when incident on the FBG, so that 67 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. the polarization sensitivity of the wavelength converter is minimum. The FBG has a reflectivity of 99.5% at the pump wavelength with a bandwidth o f 0.5 nm. Both the signal at X§ and the pump at A .p are coupled into the circulator. After a dual pass through the DSF, the converted wave, Xc, is produced at a wavelength X c (1/A,C =2/X,P - 1 /As). After the output o f the circulator, an optical filter separates the converted wave, Xc, from the other wavelengths. The polarization dependent loss (PDL) of the DSF, FBG, FRM and circulator is negligible. The polarization mode dispersion (PMD) o f the DSF is less than 1 ps. The nonlinear coefficient y of the DSF is 2.6/w/km. The zero dispersion wavelength o f the DSF is 1553.8 nm, and the dispersion slope is 0.07 ps/nm /km. 10 Gb/s Data Tunable E-O Laser MOD EDFA V Coupler Tunable Laser EDFA p c D Q Q L p Wavelength Converter DSF Circulator / FBG Faraday Rotator Mirror Optical Filter Xs, A ,p, Xc Converted Output ► Figure 6.2 Experimental setup for polarization-insensitive FW M wavelength conversion. 6.4 Results and discussion For a conventional single-pass scheme (without the FBG and FRM), the conversion efficiency is highest when the polarization state o f the signal is the same 68 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. as the polarization state o f the pump. However, when the polarization state of the signal is orthogonal to the polarization state of the pump, there is no converted wave, resulting in large (>20 dB) polarization sensitivity. For our dual pass scheme with the FBG and FRM, the theoretical polarization sensitivity equals twice the loss o f the pump while passing forward through the DSF. Figure 6.3 shows both our experimental and simulation results for the residual polarization sensitivity o f the dual pass system. The simulation results come from the calculation o f a MATLAB program under phase- matched conditions, without considering Rayleigh, Raman and Brillouin scattering effects. We use 2 km, 4 km and then 10 km DSF in our setup. The losses of the pump while passing forward through the DSF are 0.6 dB, 1.0 dB, and 2.3 dB respectively, and the corresponding measured polarization sensitivities are 1.2 dB, 2.0 dB, and 4.7 dB respectively, which are consistent with the simulation results. Theoretically, our technique will have negligible polarization sensitivity when the propagation loss o f the pump through the DSF is low. For example, for 100 m o f highly nonlinear DSF [47], the propagation loss of the pump through the fiber is 0.05 dB, so the corresponding polarization sensitivity would be 0.1 dB. 69 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. C Q 3 Vi a <u c v* c s tN 'S "o & h 6 5 0 Experim ent 4 Sim ulation 3 2 1 0 0 0.5 1 1.5 2 2.5 3 Loss o f pum p w hile passing forw ard through D SF (dB) Figure 6.3. Polarization sensitivity o f the dual pass system vs. the loss of the pump while passing forward through the DSF. In our experiment for the 4 km DSF, the pump is tuned to the zero dispersion wavelength o f the DSF (1553.8 nm). Figure 6.4 shows the power o f the converted wave at the output of the circulator as a function o f the power o f the input pump and signal wave, for both up and down conversion, when the polarization state of the input signal is aligned to the polarization state o f the pump. The insert picture is the spectrum o f wavelength conversion at the output o f the circulator. We can see that the power of the converted wave is proportional to the square o f the power of the pump wave, and proportional to the power o f the signal wave, which is consistent with the theory. 70 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 0 1 5 5 0 to 1557.6 nm ■ 1557.6 to 1550 nm -20 ■ a -25 -30 -35 0 2 4 6 8 -15 O 1550 to 1557.6 nm ■ 1557.6 to 1550 nm 04 > T t * -o < 3 4 L . 04 > fl O U 0 4 JZ -w o 04 £ © O h -20 -25 4 2 4 ■ 2 0 6 8 Power of the Pump Wave (dBm) Power of the Signal Wave (dBm) Figure 6.4. (a) Power of the converted wave vs. pump wave when power o f signal wave= 6 dBm. (b) Power o f the converted wave vs. signal wave when power of pump wave= 8 dBm. Figure 6.5 shows the dependence o f conversion efficiency (the power o f the converted wave at the output o f the circulator divided by the power of the input signal wave) on the wavelength conversion distance (the distance between the wavelength of the converted wave and the signal wave), when the pump power is 8 dBm. Conversion efficiency is about -2 2 dB for up to a 20-nm conversion distance and is symmetric for up conversion and down conversion, for both continuous wave and 10 Gb/s data signals. 71 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. ^ - o - 10 Gb/s Data Modulated P Q -20 Continuous Wave > g -30 U -30 -20 -10 0 10 20 30 Conversion Distance (nm) Figure 6.5. Conversion efficiency vs. conversion distance. Figure 6 . 6 shows both simulation and experimental results o f polarization sensitivity o f the 4 km DSF, for both single and dual pass configurations. By placing a half-wave plate right after the input signal and rotating it from 0° to 90°, we varied the angle between the polarization state of the signal and the pump from 0 ° to 180°. The polarization sensitivity for a single pass was larger than 22 dB. For dual pass, the conversion efficiency changed by 2 dB when the angle between the polarization state of the signal and the pump changed from 0° to 90°, then returned to the original value when the angle changed from 90° to 180°, in a manner consistent with the simulation results. The 2 dB residual polarization sensitivity in our experimental setup was due to the 1 dB pump propagation loss through the DSF. 72 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. > > -15 5 -20 1550 nm to 1557.6 nm Sim ulation -25 s s 0 20 40 60 80 100 120 140 160 180 -25 0 20 40 60 80 100 120 140 160 180 Angle between pump and signal polarization states (degrees) Figure 6 .6 . Experimental polarization sensitivity for up conversion (top) and down conversion (bottom), both single pass and dual pass. To determine the power penalty induced by wavelength conversion, 10 Gb/s (223-l) PRBS non-retum-to-zero (NRZ) data was modulated onto the signal wave. BER measurements were performed for both the original signal wave and the converted wave. The power penalty for the converted signal compared to the original signal is measured at 10' 9 BER. Figure 6.7 shows that, within a 20-nm wavelength conversion distance, for both up conversion and down conversion, the power penalty induced by wavelength conversion is less than 1 dB. For the best polarization state (when the polarization state o f the input signal is the same as the 73 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. pump) and the worst polarization state (the polarization state o f the input signal is orthogonal to that o f the pump), the power penalty difference is less than 0.3 dB. Power Penalty (dB) A A A A A ■ ■ A A A ■ ■ ■ ■ 1 0.8 0.61 0.4 0.21 ■ Best Polarization States A Worst Polarization States ▲ ▲ A A A A A -25 -20 -15 -10 -5 0 10 15 20 25 Conversion Distance (nm) Figure 6.7. Power penalty induced by wavelength conversion at BER = 10"9. A key disadvantage of fiber-based wavelength converters is that the pump must be located close to the zero dispersion wavelength. In conventional fiber, which has a large dispersion slope, the pump wavelength can only be tuned by < 1 nm, thereby making this a fixed wavelength converter. However, in order to make a tunable wavelength converter, non-conventional fiber with a reduced dispersion slope could be used to create a reasonably-wide tuning range ( > 8 nm) [48]. Our polarization-insensitive technique using a widely tunable FBG [49] could be applied to such a scheme. 74 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 6.5 Summary In this chapter, we demonstrate a simple technique for polarization- insensitive fiber four-wave mixing wavelength conversion using a fiber Bragg grating and a Faraday rotator mirror. We experimentally demonstrate that the residual polarization sensitivity is reduced from more than 22 dB to 2dB with 4 km of DSF. For 10 Gbit/s NRZ data, the power penalty incurred in wavelength conversion is less than 1 dB for a 20-nm conversion distance, and the difference in power penalty for different polarization states is less than 0.3 dB. 75 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Chapter 7 Width-Tunable Optical RZ Pulse Train Generation Based on Four-Wave Mixing in Highly-Nonlinear Fiber In this chapter, we demonstrate a simple technique for width-tunable optical RZ pulse train generation based on four-wave mixing in highly-nonlinear fiber. By electrically tuning the delay between two pump pulse trains, the pulse-width of a generated pulse train is continuously tuned. 7.1 Width-Tunable Optical Pulse Generation Narrow optical retum-to-zero (RZ) pulse trains have many applications in optical communications including: RZ data transmission, soliton systems, optical packet switching network signaling, all-optical switching, and various optical-signal- processing techniques [50-55], Achieving maximum performance in nearly all optical systems can depend critically on matching the optical pulse width to the optimal overall system parameters. A laudable goal is the ability at the transmitter to tune the pulse width in order to optimize system performance. 76 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Perhaps the best illustration o f the importance o f tunable pulse-width generation is the use o f RZ pulses for long-distance transmission. There are many scenarios for which RZ pulses with different pulse-widths are more robust to various fiber-based degradations, such as nonlinearity and polarization-mode-dispersion (PMD). Previous reports have shown that the performance o f a link can vary significantly depending on the pulse width, even for small changes in fiber characteristics and small difference of pulse width [50-54]. For example, when changing a pulse width from 50 to 35 ps in a 10-Gbit/s system, the achievable transmission distance could change from 600 km to 2000 km [54], Other applications for a pulse-width tunable pulse train include optical time division multiplexing (OTDM), all-optical 3R and optical sampling [55], In previous published work [54], width-tunable pulse trains have been accomplished by adjusting both the chirp applied to the pulse and the dispersion value o f a tunable dispersive element. This method needs to overdrive the phase modulator to provide sufficient chirp, which is difficult at higher (~ 40 Gb/s) bit- rates. Furthermore, any non-ideal property o f the tunable dispersion element, such as group velocity dispersion ripple and non-uniform transmission spectrum, will affect the quality of the tunable pulses, limiting the system performance. We demonstrate width-tunable optical pulse generation based on four-wave mixing (FWM) in highly-nonlinear fiber (HNLF). The technique involves: (i) 77 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. generating two parallel optical pulse trains on different wavelengths, (ii) combining them in a fiber to generate FWM, and (iii) varying the delay between the two pulse trains. Due to FWM, a product term will be generated that is a stream of optical pulses whose width is determined by the overlap time between the two original pulses. Moreover, the fiber nonlinearity compresses the pulse-width in the generated pulse train [56,57], In our experiment, by electrically tuning the delay, the full width of half maximum (FWHM) o f a 5G pulse train is tuned continuously from 85 ps to 25 ps, and the FWHM of a 10G pulse train is tuned continuously from 33 ps to 18 ps. And the simulation results show that the FWHM o f a 40G pulse train can be tuned continuously from 10.6 ps to 2.9 ps. A transmission experiment is performed to examine the quality o f the generated tunable pulses. Negligible power penalty is observed after transmission through 59-km o f single mode fiber (SMF) and 11.4-km dispersion compensation fiber (DCF) for different pulse widths at 10-Gb/s. 7.2 Concept of Width-Tunable Pulse Generation Based on FWM in HNLF A conceptual diagram o f our technique for width-tunable pulse generation based on FWM in HNLF is shown in Figure 7.1. Two RZ pulse trains with fixed pulse-width of T at wavelengths M and M are launched into an HNLF. The delay between the two pulse trains is x. Due to FWM in HNLF, a new pulse train is generated at a wavelength of M (1/ M = 2 / M - 1 / M ) . Since FWM is an ultra-fast 78 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. process, and the new pulse at X3 can be generated only when the pulses at Li and X2 are both present (overlapping each other in the time domain), the pulse-width of the generated pulse is approximately T-x. Therefore, because x can be tuned continuously, the pulse-width o f the generated pulse at X3 can also be tuned continuously. Input Pulse Train @ ^ 1 ____ Input Pulse Train r -L . z I FWM W idth-Tunable Output Pulse Train (a), A,, Figure 7.1 Concept o f width-tunable pulse generation based on four-wave mixing in HNLF: the pulse-width o f the generated pulse at X3 can be tuned continuously by tuning x. 7.3 Experimental Setup Figure 7.2 shows the experimental setup. The wavelength o f laser 1 (X1) is 1548 nm, and the wavelength o f laser 2 (X2) is set at 1552 nm, which is the zero- 79 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. dispersion wavelength of the HNLF. We first generate chirp-free pulse trains at both X] and Xz by externally modulating an RF clock using LiNbCfi modulators to both of the wavelengths. The delay between the two pulse trains can be tuned continuously by tuning the electrical delay x between the RF clock sent to modulator 1 and the clock sent to modulator 2. Both A ,] and X2 are amplified by an EDFA followed by a bandpass filter with a 3-dB bandwidth o f 0.7 nm, then coupled into 1km of dispersion-shifted HNLF, with a non-linear coefficient of 9.1 W ’km ' 1 and a fiber loss of 0.45 dB/km. A polarization controller (PC) is used to align the polarization states o f Li to X2 to achieve the highest FWM efficiency. After propagation through the HNLF, a new pulse train is generated at the wavelength L3 =1556 nm. An optical filter with a bandwidth o f 1 nm separates X3 from the residual signals. A 40G photo detector and a 50G oscilloscope are used after the optical filter to capture the waveforms o f the generated pulse train. By electrically tuning the delay, we continuously tune the pulse-width o f the generated pulse train at X3. This experiment was demonstrated at both 5 and 10 GHz clock rates. 80 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. EDFA r u i \ W idth-T unable P u lse G en erator R F C lo c k lkm H N L F E D F A O p tic a l F ilte r C o u p le r A ,, , X 3 G e n e ra te d W id th - T u n a b le P u ls e T ra in P D - O s c illo s c o p e L a s e r 2 - L a s e r 1 D elay ! E -O M o d . 1 Figure 7.2 Experimental setup for width-tunable pulse generation based on four- wave mixing in FINLF. 7.4 Results and discussion The peak powers o f X\ and Xi launched into the HNLF are both approximately 10 dBm. The optical peak power o f the output o f the HNLF at A 3 is approximately -5 dBm. W hen the frequency of the RF clock is 5 GHz, the FWHM of the 5G input pulse trains at both ^ 1 and X2 is 98 ps, with a rise time o f 27 ps. As shown in Figure 7.3, by electrically tuning the delay x from 0 to 80 ps, the FWHM of the generated 5G pulse train at the output X3 is continuously tuned from 85 ps to 25 ps. When the frequency o f the RF clock is 10 GHz, the FW HM o f the 10G input pulse trains at both A ,| and X2 is 44 ps, with a rise time of 18 ps. By tuning the delay x from 0 to 40 ps, the FWHM of the generated 10G pulse train at the output X3 is continuously tuned from 33 ps to 18 ps. The experimental results are consistent with the simulation results for both 5G and 10G data rates. We also simulate our system 81 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. at 40G data rates by setting the FWHM and the rise time for the input 40G pulse train to V 2 and 1/8 o f the bit time, respectively. Simulation results show that the FWHM o f the generated 40G pulse train can be continuously tuned from 10.6 ps to 2.9 ps by tuning the delay, x from 0 to 9 ps. 0 Experiment — Simulation (/) CL C D j/> “ ■ 40 5G H — o 10G X 5 LL 40G Delay t (ps) Figure 7.3 The FWHM of generated pulse train vs. the delay x at different repetition rates. Figure 7.4(a) shows three examples of waveforms generated from a 5G pulse train with FWHMs o f 30 ps, 50 ps and 80 ps. Figure 7.4(b) shows three examples 82 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. of waveforms generated from a 10G pulse train with FWHMs o f 20 ps, 25 ps and 30 ps. Figure 7.5(a) shows the optical spectrum at the output o f the HNLF and Figure 7.5(b) shows the details of the optical spectrum o f the 10G generated 25-ps pulse train, with a resolution o f 0 . 0 1 nm. 83 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. FWHM = 30 ps 50 ps < — ► FWHM = 50 ps FWHM 80 ps (a) FWHM = 20 ps T ....Ti" ^ 1 1 50 ps . FWHM „ . _ = 25 ps FWHM = 30 ps (b) Figure 7.4 Waveforms of the generated pulse train: (a) 5G; (b) 10G. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 10 m o 2 -20 a O -30 1548 1552 1556 W avelength (nm) (a) s P Q 3 u < u £ o p- w • — '+ ■ > a O 1 -11 -21 -31 -41 1555 1556 1557 Wavelength (nm) (b) Figure 7.5 Optical spectrum (a) at the output o f HNLF; (b) o f 10G generated 25- pulse train. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. From both the experimental and simulation results, we conclude that the maximum FWHM o f the generated pulse train at X3 is about 80% o f the FWHM of the input pulse train. This can be attributed to the nonlinear compression effect of the pulse shape during FWM. Note that if the rise and fall times are faster, the FWHM o f the generated pulses can be greater than 80% o f the input pulses. The minimum FWHM of the generated pulse train is limited by the rise time o f the input pulses. As stated previously, the pulse-width between these maximum and minimum FWHMs can be changed linearly by electrically adjusting the delay, t. Using this width-tunable pulse generation as a 10G pulse train source, we modulate 10G (223-l) PRBS data and transmit it through transmission links with a 0- dBm launched power. The receiver is thermal-noise limited. First we compare the performance of 25-ps pulse signal with that o f a regular pure 50% RZ signal through different lengths o f single mode fiber (SMF) without dispersion compensation. As shown in Figure 7.6, the induced power penalties follow the same trend, which means the generated pulse train has negligible chirp. For the same transmission distance, the power penalty difference between 25-ps and 50-ps RZ signals is mainly caused by the differences of pulse widths and profiles. 86 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. CQ w "cS f l Pl h £ o C L h 4 25-ps Pulse 2 1 R egular 50% RZ 50-ps pulse 0 SMF Length (km) Figure 7.6 Power Penalty vs. transmission distance through SMF w/o compensation. We also transmit different pulse-width signals through 59 km of SMF and 11.4 km of DCF, with a zero residual chromatic dispersion. Figure 7.7 shows the BER measurement results for 25-ps pulse-width signal. We can see that there is a negligible power penalty after transmission. Similarly, the power penalties for other pulse-width signals are also negligible. Therefore, the generated pulse train is suitable for high speed transmission. 87 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. O ' LU DO D ) O 3 A Back-to-Back O After Trans. 4 25 ps Pulse 5 6 7 8 9 10 -21 -20 -24 -23 -22 Received Optical Power (dBm) Figure 7.7 Power Penalty vs. transmission distance through SMF w/o compensation. 7.5 Summary In this chapter, we demonstrate a simple technique for width-tunable optical RZ pulse train generation based on four-wave mixing in highly-nonlinear fiber. By electrically tuning the delay between two pump pulse trains, the pulse-width o f a generated pulse train is continuously tuned. In our experiment, the FWHM o f a 5G pulse train is tuned from 85 ps to 25 ps, and the FWHM o f a 10G pulse train is tuned from 33 ps to 18 ps. And the simulation results show that the FWHM o f a 40G pulse train can be tuned continuously from 10.6 ps to 2.9 ps. Negligible power penalty is Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. observed after 59-km SMF and 11.4-km DCF transmission for different pulse widths at 10-Gb/s. 89 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Chapter 8 All-Optical XOR Gate Using Polarization Rotation in Single Highly-Nonlinear Fiber In this chapter, we demonstrate a high-speed all-optical XOR gate based on polarization rotation induced by Kerr effect in a single highly-nonlinear fiber. 8.1 All-Optical XOR Gate Future high-speed optical communication networks, especially packet- switched networks, may rely on the processing o f signals completely in the optical domain. Such processing could occur either: (i) in the data plane, such as for header replacement, so that no optical-to-electronic conversion is necessary, or (ii) in the control plane, such as for determining correlation, destination and contention resolution, so that high-speed resolution can be achieved. A key building block in many areas of optical signal processing is the XOR logic gate. The XOR logic gate is a commonly used device in half adders, pattern recognition circuits, ultrahigh speed pattern generation, data encoding and encryption circuits [58-61], Previous reported techniques for achieving the all-optical XOR function include: (i) fiber based interferometers [60,61], which may be limited by instability 90 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. unless special design is applied; (ii) utilization of integrated M ach-Zehnder interferometers based on semiconductor optical amplifiers (SOAs) [62-64], and using cross-polarization modulation in SOAs [65] which results in different bits on different wavelengths at the output. Due to the carrier dynamic in the SOAs, SOA based techniques may suffer from additional noise and speed limitations (unless special high-speed SOAs are used [63]), and some o f the techniques require more than one SOA to achieve the XOR function. In this chapter we demonstrate a 10 Gbit/s all-optical XOR gate using polarization rotation induced by the Kerr effect in a single 2-km highly-nonlinear fiber (HNLF). Due to the Kerr effect, two different input light waves at A -i and Xj induce birefringence in the HNLF, thereby rotating the polarization state of a third light wave at X3 . The resulting amount o f induced birefringence is determined by the on/off state o f the two input waves. The resulting output at X3 after a polarizer represents the XOR operation o f the two input signals. We are able to obtain an on- off extinction ratio at the output o f our all-optical XOR gate o f 25 dB using a 2 1cm spool of HNLF with a non-linear coefficient of 9.1 W 'k n T 1 . The results can be further improved through the use o f fiber with a higher non-linear coefficient. Since nonlinear effects in fiber are ultra-fast, the ultimate speed limitation o f our proposed XOR gate is above 100 Gbit/s. 91 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 8.2 Concept of All-Optical XOR Gate Using Polarization Rotation in Single Highly-Nonlinear Fiber A conceptual diagram of our XOR generation technique is shown in Figure 8.1. At the input of the HNLF the polarization states of the input waves at X\ and X2 are orthogonal to each other. Both o f these inputs are aligned 45° with respect to the third "dummy" continuous wave (CW) signal at X2. A polarizer is placed at the output o f the HNLF, aligned orthogonal to the original polarization state o f X3. When both X\ and X2 are off, there is no output at X2 after the polarizer. When only the input signal at X\ is present, the Kerr effect creates a difference in optical index between the polarization direction aligned with Xi and the direction orthogonal to X\ (i.e., the polarization direction of X2). In this condition, the polarization state of X2 will rotate due to the birefringence induced by the presence o f X\. Since the output at X2 is no longer orthogonal to the polarizer, a portion of the signal will pass through and an output at X .3 will be present after the polarizer. Similarly, when only the input signal at X2 is present, X2 will experience a polarization rotation and an output at X2 will be present after the polarizer. However, when both input signals at X\ and X2 are present, the birefringence induced by Li and X2 will cancel, resulting in a net zero rotation o f the polarization state of X2, In this situation, the polarization state of X2 at the output o f the fiber will be orthogonal to the polarizer and no signal at X2 will be present at the output of the polarizer. 92 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Based on the four above-mentioned situations, the polarizer output at X3 is the XOR operation o f the input signals at X\ and X2 (see logic table in Figure 8.1). It is worth noting that the Kerr effect is insensitive to the wavelengths chosen for X\, X2 and X3, as long as the three wavelengths are located within a small dispersion range. Furthermore, w alk-off effects can be controlled by limiting the length o f the HNLF. J T \ o J T \ Input, @ A ,, Output @ X3 Dummy CW X3 joJTT \_ HNLF 1 1 \0_ ( Xx) XOR (A,) Input, Polarizer *1 x 2 X3 0 0 0 0 1 1 1 0 1 1 1 0 Figure 8 .1 Concept for all-optical XOR gate based on Kerr effect in a single highly- nonlinear fiber. Figure 8.2 shows simulation results obtained using the VPI software modeling tool. The 2 km HNLF used in the simulation has a non-linear coefficient of 9.1 w 'k m '1 , with a zero dispersion wavelength at 1552 nm. The optical power of both inputs is approximately 16 dBm, with Li and X2 having wavelengths o f 1548 and 1550 nm, respectively. The input power o f X3 is approximately 3 dBm at a wavelength o f 1554 nm. Input data patterns at 10 Gbit/s are shown (10100100 at X\ and 1 0 0 0 0 1 1 1 at X2), along with the resulting output pattern at X3 after the polarizer 93 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. (00100011). The resulting output represents the XOR operation o f the two input data streams. Similar simulation results were also obtained at 40 Gbit/s date rate. Figure 8.2 Simulation results of the output pattern at A 3 with the input patterns at a i and X2 at 10 Gbit/s. 8.3 Experimental Setup Figure 8.3 shows the experimental setup o f our all-optical XOR Gate. The wavelengths of laser 1 (X ) and laser 2 (iG) are 1548 nm and 1550 nm, respectively and the dummy wavelength X3 is 1554 nm. The input signals X\ and X2 are combined through a PBS, the output of which is coupled with X3 by a 90:10 coupler into 2 km of FINLF. The HNLF has a non-linear coefficient o f 9.1 W 'k n fi1 and a fiber loss of rutfi_j ^_____Input2 1 [ f \ ° utput n 0 200 400 600 800 Time (ps) 94 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 0.45 dB/km, with a zero dispersion wavelength at 1552 nm. The polarization-mode dispersion (PMD) o f the HNLF is less than 1 ps. Through tuning o f PC3, the polarization state o f X3 is aligned perpendicular to the direction o f the polarization state of X2 by minimizing their resulting four-wave mixing product term. The polarization state o f L3 is then rotated 45° through a rotation o f the X/2 plate by 22.5°. PC4 is used to adjust the polarization state o f X3 at the output o f the HNLF to be orthogonal to the polarizer when both X\ and X2 are off. The distinction ratio of the polarizer is approximately 35 dB and sets a limit on the resulting XOR extinction ratio. An optical fdter with a bandwidth o f 0.5 nm is used after the polarizer to separate the desired signal at X3 from the residual wavelengths. E D F A All-Optical XOR Gate 1 0 G D a ta E D F A P B S H N L F P C 4 O p tical F ilter X O R O utput Polarizer Coupler D u m m y X - Oo Laser 2 L aser 3 L aser 1 M Z M o d . M Z M o d . Figure 8.3 Experimental setup for all-optical XOR gate based on Kerr effect in a single highly-nonlinear fiber. 95 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 8.4 Results and discussion In Figure 8.4 we illustrate the output optical power o f A .3 as a function of the input power when only X\ or X2 is on. The input power o f X3 into the HNLF is approximately 3 dBm. As shown in Fig. 8.3, as the optical power o f X\ or X2 at the input of the XOR gate increases from 0 to 0.11 W, the output power o f changes continuously from about -4 0 dBm to -15 dBm. It can be observed that there is negligible difference between the results when either or X2 is on. The highest output power o f X3 is 18 dB less than the input power o f X3 in our experiment. By using a HNLF with a larger non-linear coefficient, a larger rotation o f X3 can be obtained. The output signal at X3 will therefore be more aligned with the polarizer and a larger output level will be observed. 96 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. -15 S os ■ d -20 a > < 2 -25 C 3 'ZJ ♦ Input @ X , O Input @ X2 3 a -35 O 0 0.02 0.04 0.06 0.08 0.1 0.12 Input Optical Power (W) Figure 8.4 The output vs. intput optical power o f the XOR gate. In Figure 8.5, we show the optical spectrum of the output o f the XOR gate for CW inputs at X| and Xz. The input optical power of both and Xz is 0.11 W. Due to loss in components, the corresponding power at the input o f the FINLF is 16 dBm (just below the stimulated Brillouin scattering threshold). When A -i and Xz are both off or on, the output power o f the XOR gate at Xz is ^40 dBm and increases to - 15 dBm when only one o f X . i or Xz is on. The on-off extinction ratio o f the XOR gate is therefore 25 dB. According to Figure 8.4, the on-off extinction ratio increases as the input optical power of X\ and Xz increases. Note that the input optical power of X\ and Xz should be lower than the stimulated Brillouin scattering threshold, which 97 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. can be increased by using shorter fiber with higher nonlinearity. After the HNLF, Li, L2, and the other wavelengths generated by four wave mixing are filtered away by an optical bandpass filter. Input 1 1 u|>ut2 Input 1 Inf ut2 20 OFF OFF -20; ojji ; C N 1 30 O U T Pl 1 -30 s............j-|........ O U TPU T -40 1547 1551 (a) -40dBm 1555 nm (c) -40 -40dBm 1547 1551 (b) 1555 nm Inputl Input2 -15dBm Inputl Im -20 _ . C N . OFF 1 -20 .OFF ,. .6 i o ! i -30 OU TPU T .................. F -40 1 I ' : ’ • iii -40 i 1 1 / 1547 1551 1555 nm 1547 1551 -15dBm ...OUTPUT 1555 nm (d) Figure 8.5 The output optical spectrum o f the XOR gate when inputs L| or L2 are continuous waves: (a) L, off and L2 off; (b) Li off and L2 on; (c) L| on and L2 off; (d) Li on and L2 on. In Figure 8.6 we illustrate the input and output waveforms for 10 Gbit/s NRZ data. The input patterns are at Li and L2 and the resulting XOR output pattern is at L3. For the input data pattern 0110001001 at Li and input data pattern 0011101100 98 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. at X2, the resulting data pattern at X2 is 0101100101 functionality. 100 ps 4 —►' Input2 Output Figure 8.6 The input patterns at X\ and X2 and the resulting output pattern at X3 for 10 Gbit/s data. Figure 8.7 shows the eye diagram of the output o f the XOR gate when using 10 Gbit/s (223-l) PRBS NRZ data. The signal-to-noise ratio (SNR) o f the input data at A ,] is 18.7 dB and at X2 is 19.6 dB, while the SNR of the output data is 15.7 dB. The corresponding SNR penalty at the output compared to the input data at X2 is 3 dB, which could be caused by polarization fluctuation/instability o f the polarization states in the system. Since the Kerr effect in fiber is ultra-fast and the delay of output signal response due to input signal change is less than 10 ps, the data rate of our proposed XOR gate can be above 100 Gbit/s in principle. 99 This demonstrates XOR Inputt Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 8.7 Eye diagram of output of XOR gate for 10 Gbit/s PRBS data. Since the Kerr effect is proportional to the fiber nonlinearity, by using HNLF with higher nonlinearity [66] or using HNL holey fiber [67], both the output power and extinction ratio of the XOR gate can be improved considerably and the required input optical power can be decreased. For example, using a HNLF with a non-linear coefficient o f 20 W 'k m ’1 [66] in our scheme, with the same input optical power levels, both the output power and extinction ratio of the XOR gate are expected to increase by approximately 10 dB. We can also achieve the same induced birefringence by using shorter fiber with higher nonlinearity, thereby reducing polarization instability and increasing the stimulated Brillouin scattering threshold. Note that we can simultaneously obtain the XNOR gate operation on the orthogonal polarization state o f Xj. 100 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 8.5 Summary In this chapter, we demonstrate a 10 Gbit/s all-optical XOR gate based on polarization rotation induced by Kerr effect in a single highly-nonlinear fiber. Using 2-km o f highly-nonlinear fiber with a non-linear coefficient o f 9.1 W 'k n f 1 , we obtain a 25-dB extinction ratio at the XOR output. 101 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Chapter 9 3R Regeneration of a 40-Gbit/s Optical Signal by Optical Parametric Amplification in a Highly-Nonlinear Fiber In this chapter, we demonstrate wavelength-shift-free 3R-regeneration of 40- Gbit/s optical RZ signal by OPA with a clock-modulated pump in highly-nonlinear fiber. 9.1 3R Regeneration of High-Speed Optical Signal In order to achieve high throughput and efficiency, future transparent optical networks may require an optical data signal to traverse long-distances and many switching nodes all-optically before reaching its destination. O f course, the data signal will become degraded in both time and amplitude by many possible effects, including fiber-based dispersion and nonlinearities as well as non-idealities of optical devices inside a switching node [68], Therefore, there has been much interest in a high-speed 3R (retiming, reshaping, re-amplification) regenerator as a network element [69], 102 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. There have been several published results on optical 3R regenerators. In planar optoelectronic devices, 3R was achieved using a wavelength shifter in the form of: (i) a semiconductor optical amplifier (SOA) that employed cross-gain- or cross-phase-modulation [70], and (ii) an electro-absorption (EA) modulator that employed cross-absorption-modulation [71]; these methods generally induced wavelength shift o f the signal. Since fiber-based nonlinear effects have the potential of operating at a much higher speed than optoelectronic devices, there have also been reports on using a wavelength shifter in the form o f highly-nonlinear fiber (HNLF), which the regenerated signal are also at a different wavelength.[72,57]. An additional wavelength converter has been used to demonstrate wavelength-shift free 3R-regenerator [36], Previously an optical parametric amplification (OPA) with a clock-modulated pump has been used to make a continuous wave (CW) becoming a high-quality pulse train source [56], Here we demonstrate 3R regeneration of a 40-Gbit/s optical signal by OPA with a clock-modulated pump in a highly-nonlinear fiber (HNLF). We use a recovered 20-GHz clock to drive an OPA pump laser and produce a clean carrier- suppressed RZ (CS-RZ) pulse optical pulse train. This pump pulse train mixes with the degraded 40-Gbit/s data signal in the OPA to achieve retiming/reshaping/re - amplification. After the 3R regenerator, the power penalty of the signal is improved by 2.6 dB at 10"9, and the signal optical power is amplified by 7 dB. We emphasize that the regenerated data rem ains on its original wavelength, such that no 103 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. wavelength shifting is required to achieve our results and thereby potentially making network control and management somewhat less complex. 9.2 Concept of 3R Regeneration by Optical Parametric Amplification in Fiber With a Clock-Modulated Pump Figure 9.1 shows a conceptual diagram of 3R signal regeneration using OPA in FINLF with a clock-modulated pump. Pump Pump Signal Amplified Re-amplification Signal OPA Reshaping Retiming HNLF Degraded Signal A i A: A; K Optical Filter O o A A Clock-Modulated Pump Reshaping Retiming 3R Regenerated Signal Figure 9.1 Concept of 3R regeneration using OPA in HNLF with a clock-modulated pump. First we recover the clock at the half bit rate o f the degraded data via a clock recovery module, and use it to modulate the pump light to become a clean CS-RZ pulse train at bit rate. Then both the degraded signal and the pump pulse train are 104 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. combined together into a HNLF. The signal is amplified by OPA based on four wave mixing (FWM). Since the OPA efficiency is proportional to the square of the pump power, in the time domain, a pump pulse train makes a CW signal wave into a pump train with a narrower pulse width due to nonlinear effects. When the signal wave carrying degraded data, the data will be retimed and reshaped. Due to the inherent amplification of OPA, the data will be amplified at the same wavelength. By filtering out the signal wave, the 3R regenerated data is obtained without a wavelength shift. 9.3 Experimental Setup The experimental setup is shown in Figure 9.2. The wavelength o f the signal (As) is 1546.8 nm, and the wavelength o f the pump (Zp) is set at 1551.8 nm. First we generate a 40-Gbit/s optical RZ signal by modulating a 40-GHz clock and 40 Git/s (2n-l) PRBS data to the signal wave. Then the 40 Gbit/s RZ data is degraded by passing through 2.2 km o f single mode fiber with a dispersion value o f 39 ps/nm, and attenuated by an attenuator to -25 dBm, then amplified by EDFA back to 0 dBm to add ASE noise. 1% signal power is tapped to a 40G photodiode and a clock recovery which recovers clock at 20 GHz. Then the CW pump wave is modulated by the 20 GHz clock to become a 40-GHz CS-RZ pulse train. In order to further reduce stimulated Brillouin smattering (SBS), the pump wave is then randomly phase modulated by a 10 Gbit/s PRBS signal, which is synthesized to the recovered 105 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. clock, so as to minimize the power penalty caused by phase modulation. Both the signal and pump wave are coupled into 1-km of dispersion-shifted HNLF, with a non-linear coefficient o f O .lW 'k m '1 and a fiber loss o f 0.45 dB/km. The zero- dispersion wavelength o f the HNLF is 1552 nm. After the OPA process in the HNLF, an optical filter with a 3-dB bandwidth o f 0.5 nm separates 3R regenerated signal wave from the residual signals. 2.2 km SMF EDFA Degraded 40G RZ Data Regenerated Data Att. MZ Mod. MZ Mod. Tunable Laser 3R Regenerator 40G 40G Clock PRBS Data (a) Degraded 40G RZ Data PD Tunable Laser EDFA PC X EL X Clock Recovery EDFA MZ Random Mod. Phase Mod. 3R Regenerator n 1 km HNLF (y=9.1/w/km) n 0.7-nm Optical Filter Regenerated Data (b) Figure 9.2 (a) Experimental setup for 3R regeneration using OPA in HNLF with a clock-modulated pump, (b) Details o f 3R regenerator. 106 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 9.4 Results and discussion Figure 9.3(a) shows the recovered 20-GHz clock. The pump light is modulated by this 20-GHz clock and becomes a 40-GFlz CS-RZ optical pulse train, whose waveform is shown in Figure 9.3(b). 20 ps (a) (b) Figure 9.3 (a) Recovered 20-GHz clock and (b) waveform o f the 40-GHz CS-RZ pump pulse train. The optical power o f the signal wave and the pump wave at the input of HNLF are -9 dBm and 20 dBm respectively. Figure 9.4 shows the optical spectrum at the output o f the HNLF. 107 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 10 0 -10 -20 -30 Figure 9.4 Optical spectrum at the output o f the HNLF. Figure 9.5 (a), (b), (c) shows the detailed optical spectrum o f the signal at the input o f the HNLF, at the output of the HNLF, and after the optical fdter respectively. We can see that the optical spectrum o f the signal is broadened after OPA, which means the pulse width o f the RZ signal is narrowed by OPA. After the 0.5-nm filter, the optical spectrum o f the signal is changed back to be similar to the original optical spectrum. And the measured optical power o f the signal wave after the filter is -2 dBm, which means the signal is amplified by 7 dB after 3R regeneration. < m m U _ A L _ J _ _ _ _ _ _ _ _ _ _ i_ _ _ _ _ _ _ _ _ _ ■ V ' l . 1551.8 nm 108 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. -5 lilini -15 -25 -35 -45 I 0.4 nm M il i i . . . 1... [IJ : M P : 1 \ # i l l i , J 1. . . 1 J ... :... h 1 I I : 1546.8 nm (a) 0 dBm 0.4 nm f 1546.8 nm i I t 0.4 nm 1546.8 nm (c) Figure 9.5 Detailed optical spectrum o f the signal at (a) the input o f the HNLF; (b) the output o f the HNF; (c) after the optical filter. 109 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The eye diagram of the signal before and after the 3R regeneration is shown in Figure 9.6 (a) and (b). We can see the quality of the 40 Gbit/s RZ signal is improved after the 3R regeneration. (a) (b) Figure 9.6 (a) Eye before the 3R generator; (b) eye after 3R generator; Figure 9.7 shows the BER measurement results. We can see the power penalty at BER o f 10'9 is improved by 2.6 dB through the 3R regenerator. The power penalty compared to back to back is 0.2 dB. 110 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Degraded Data O 7 Back to Back Regenerated s J D a t a 13 -12 -11 -10 -9 -8 -7 -6 -5 Received Optical Power (dBm) Figure 9.7 BER measurement for 40 Gbit/s Data. 9.5 Summary In this chapter, we demonstrate wavelength-shift-free 3R-regeneration of 40- Gbit/s optical RZ signal by OPA with a clock-modulated pump in highly-nonlinear fiber. The power penalty is improved by 2.6 dB, and the signal power is amplified by 7 dB. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Chapter 10 Conclusion Management of fiber physical effects is essential in high-speed reconfigurable WDM optical fiber communication systems and networks. The physical effects in optical fiber include chromatic dispersion, polarization mode dispersion (PMD) and nonlinear effects.. For >10 Gbit/s optical fiber transmission system, it is critical that chromatic dispersion and PMD be well monitored and compensated by using some type of dispersion monitoring and compensation. One the other hand, dispersive and nonlinear effects in optical fiber systems can also be beneficial and has applications on pulse management, all-optical signal processing and network function, which will be essential for high bite-rate optical networks and replacing the expensive optical-electrical-optical (O/E/O) conversion. In this Ph.D. dissertation, we present a detailed study on dispersive and nonlinear effects in high-speed optical communication systems. We have demonstrated: (i) A novel technique for optically compensating the PM D-induced RF power fading that occurs in single-sideband (SSB) subcarrier-multiplexed systems. By 112 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. aligning the polarization states of the optical carrier and the SSB, RF power fading due to all orders o f PMD can be completely compensated. The 5% RF power fading tail is improved from 13.5 dB to <1.5 dB, as verified from both experimental measurement (300 samples) and computer simulation (10,000 samples). This technique is independent o f the DGD of the optical fiber link and the subcarrier frequency. (ii) Chromatic-dispersion-insensitive PMD monitoring by using a narrowband FBG notch filter to recover the RF clock power for lOGb/s NRZ data. Using a FBG notch filter with a 10-dB bandwidth o f 15 GHz at the receiver, we measured the recovered clock power to monitor PMD in a lOGb/s NRZ system. The variation o f the detected clock power is within 1.5 dB when the accumulated dispersion increases from 0 to 600 ps/nm. (iii) Chirp-free high-speed optical pulse generation with a repetition rate of 160 GHz (which is four times of the frequency of the electrical clock) using a phase modulator and polarization maintaining (PM) fiber. The unwanted low frequency tones are suppressed by more than 15 dB, and the measured pulse width is 3.3 ps. (iv) Polarization-insensitive all-optical wavelength conversion based on four- wave mixing in dispersion-shifted fiber (DSF) with a fiber Bragg grating and a 113 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Faraday rotator mirror. We experimentally demonstrate that the residual polarization sensitivity is reduced from more than 22 dB to 2dB with 4 km o f DSF. (v) W idth-tunable optical RZ pulse train generation based on four-wave mixing in highly-nonlinear fiber. By electrically tuning the delay between two pump pulse trains, the pulse-width o f a generated pulse train is continuously tuned. In our experiment, the FWHM o f a 5G pulse train is tuned from 85 ps to 25 ps, and the FWHM o f a 10G pulse train is tuned from 33 ps to 18 ps. And the simulation results show that the FWFIM of a 40G pulse train can be tuned continuously from 10.6 ps to 2.9 ps. (vi) A high-speed all-optical XOR gate based on polarization rotation induced by Kerr effect in a single highly-nonlinear fiber. Using 2-km of highly- nonlinear fiber with a non-linear coefficient o f 9.1 W 'k m '1 , we obtain a 25-dB extinction ratio at the XOR output. (vii) W avelength-shift-free 3R-regeneration of 40-Gbit/s optical RZ signal by OPA with a clock-modulated pump in highly-nonlinear fiber. The power penalty is improved by 2.6 dB, and the signal power is amplified by 7 dB. 114 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. References 1 G. P. Agrawal, Fiber-Optics Communication Systems, John W iley & Sons, 1997 2 J.-X. Cai, D.G. Foursa, C.R. Davidson, Y. Cai, G. Domagala, H. Li, L. Liu, W W . Patterson, A.N. Pilipetskii, M. Nissov, and, N.S. Bergano, “A DWDM Demonstration of 3.73 Tb/s over 11,000 km using 373 RZ-DPSK Channels at lOGb/s”, Optical Fiber Communication Conference(OFC) 2003, Paper: PD22, 2003 3 J. -X. Cai, M. Nissov, A. N. Pilipetskii, C. R. Davidson, R. -M . Mu, M. A. Mills, L.Xu, D. Foursa, R. Menges, P. C. Corbett, D. Sutton, and Neal S. Bergano, “ 1.28 Tbit/s (32 x 40 Gb/s) Transmission over 4,500 km”, 27th European Conference on Optical Communication (ECOC) 2001, Paper PD5, 2001 4 S. Bigo, Y. Frignac, G. Charlet, W. Idler, S. Borne, H. Gross, R. Dischler, W. Poehlmann, P. Tran, C. Simonneau, D. Bayart, G. Veith, A. Jourdan, and J.- P. Hamaide, “ 10.2Tbit/s (256x42.7Gbit/s PDM/W DM) transmission over 100km TeraLight fiber with 1.28bit/s/Hz spectral efficiency”, Optical Fiber Communication Conference (OFC) 2001, Paper PD25, 2001 5 G.P. Agrawal, Nonlinear Fiber Optics, New York: Academic Press, 2001 6 D. Marcuse, Light Transmission Optics, Van Nostrand Reinhold, New York, 1982 7 I.P. Kaminow and T.L. Koch, Optical Fiber Telecommunication. Ill, Academic Press, 1997 116 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 8 Tingye Li, “The Impact of optical amplifiers on long-distance light telecommunications”, Proceedings o f the IEEE, vol. 81, pp. 1568-1579, 1993 9 T. Kato, Y. Koyano, and M. Nishimura, “Temperature dependence of chromatic dispersion in various types of optical fibers,” Optical Fiber Communication Conference (OFC) 2000, Vol. 1, 2000, pp. 104 -106 vol.l 10 C.D. Poole, “Statistical treatment of polarization dispersion in single-mode fiber”, Opt. Lett. Vol. 13, pp. 687-689, 1988 11 C.D. Poole, “M easurement of polarization-mode dispersion in single-mode fibers with random mode coupling”, Opt. Lett. Vol. 14, pp. 523-525, 1989 12 Y. Namihira, and H. Wakabayashi, “Fiber length dependence of polarization mode dispersion measurement in long-length optical fibers and installed optical submarine cables”, J. Opt. Commun. Vol. 12, pp. 1-8, 1991 13 Y.R. Shen, Principles o f Nonlinear Optics, Willey, New York, 1984 14 A.R. Chraplyvy, “Limitation on lightwave Communication imposed by optical-fiber nonlinearities”, IEEE J. Lightwave Technol., vol. 8, pp 1548- 1557,1990 15 D. Marcuse, A.R. Chraplyvy, and R.W. Tkach, “Effect of fiber nonlinearities on long distance transmission”, IEEE J. Lightwave Technol., vol. 9, no. 1, 1991 16 D. Marcuse, A.R. Chraplyvy, and R.W. Tkach, “Dependence o f cross-phase modulation on channel number in fiber WDM systems”, IEEE J. Lightwave Technol., vol. 12, pp 885-890, 1994 17 R.W. Tkach, A.R. Chraplyvy, F. Forghieri, A.H. Gnauck, and R.M. Derosier,” Four-Photon M ixing and High-Speed WDM Systems”, IEEE J. Lightwave Technol., vol. 13, pp. 841-849, 1995 117 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 18 C.D. Poole and T.E. Darcie, “Distortion related to polarization-mode - dispersion in analog lightwave systems,” J. Lightwave Technol., vol. 11, pp. 1749-1759, 1993 19 O. H. Adamczyk, A. B. Sahin, Q. Yu, S. Lee, and A.E. Willner, “Statistics of PM D-induced power fading for intensity-modulated double-sideband and single-sideband microwave and millimeter-wave signals,” IEEE Trans. M icrowave Theory Technol., vol. 49, pp. 1962-1967, 2001 20 R. Hofstetter, H. Schmuck, and R. Heidemann, “Dispersion effects in optical millimeter-wave systems using self-heterodyne method for transport and generation,” IEEE Trans. Microwave Theory Technol., vol. 43, pp. 2263- 2269, 1995 21 G. H. smith, d. Novak and Z. Ahmed, “Technique for optical SSB generation to overcome dispersion penalties in fibre-radio systems,” Electron. Lett., vol. 33, pp. 74-75, 1997 22 R. M. Jopson, L.E. Nelson, H. Kogelnik, and G.J. Foschini, “Statistics of second-order PMD depolarization,” J. Lightwave Technol., vol. 19, pp. 1882- 1886, 2001 23 P. Ciprut, B. Gisin, N, Gisin, R. Passy, J. P. Von der Weid, F. Prieto, and C.W. Zimmer, “Second-order polarization mode dispersion: Impact on analog and digital transmissions,” J. Lightwave Technol., vol. 16, pp. 757- 771, 1998 24 R. FTui, C. Allen and K. Demarest, “PM D-insensitive SCM optical receiver using polarization diversity,” IEEE Photon. Technol. Lett., vol. 14, pp. 1632- 1634, 2002 25 R. Khosravani, I.T. Lima, P. Ebrahimi, E. Ibragimov, A.E. Willner, and C.R. Menyuk, “Time and frequency domain characteristics o f polarization-mode 118 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. dispersion emulators,” IEEE Photon. Technol. Lett., vol. 13, pp. 127-129, 2001 26 S.M.R. Motaghian Nezam, Y. Wang, M.C. Hauer, S. Lee, and A.E. Willner, “Simultaneous PMD monitoring of several WDM channels using subcarrier tones,” Conference on Lasers and Electro-Optics (CLEO) 2004, Paper CFE1, 2004 27 M.W. Chbat, J.-P. Soignb, T. Fuerst, J.T. Anthony, S. Lanne, H. Fevrier, B.M. Desthieux, A.H. Bush, and D. Penninckx, “Long term field demonstration of optical PMD compensation on an installed OC-192 link,” Optical Fiber Communication Conference (OFC) 1999, Paper PD12, 1999 28 G. Ishikawa and H. Ooi, “Polarization-mode dispersion sensitivity and monitoring in 40-Gbit/s OTDM and 10-Gbit/s NRZ transmission experiments,” Optical Fiber Communication Conference (OFC) 1998, paper WC5, 1998 29 S.M.R. Motaghian Nezam, Y.W. Song, Z. Pan, and A.E. Willner, “PMD monitoring in WDM systems for NRZ data using a chromatic-dispersion- regenerated clock,” Optical Fiber Communication Conference (OFC) 2002, Paper WE5, 2002 30 M.C. Hauer, Q. Yu, A.E. Willner, E.R. Lyons, C.H. Lin, A.A. Au, and H.P. Lee, “Compact, all-fiber PMD emulator using an integrated series of thin- film micro-heaters,” Optical Fiber Communication Conference (OFC) 2002, Paper ThA3, 2002 31 J.J. Veselka and S.K. Korotky, “Compact, all-fiber PMD emulator using an integrated series of thin-film micro-heaters,” IEEE J. Sel. Top. Quantum Electron., vol. 2, pp. 330-310, 1996. 119 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 32 Y. Miyamoto, A. Hirano, K. Yonenaga, A. Sano, H. Toba, K. Murata, and O. Mitomi, “320 Gbit/s (8x40 Gbit/s) WDM transmission over 367 km with 120 km repeater spacing using carrier-suppressed return-to-zero format,” IEE Electron. Lett., Vol. 35, pp. 2041-2042, 1999. 33 R. Ohhira, D. Ogasahara, and t. Ono, “Novel RZ signal format with alternate- chirp for suppression of nonlinear degradation in 40 Gb/s based W DM,” Optical Fiber Communication Conference (OFC) 2001, Paper WM2, 2001 34 X. Wei, J. Leuthold, and Z. Liming, “Delay-interferometer-based optical pulse generator,” Optical Fiber Communication Conference (OFC) 2004, Paper WL6, 2004 35 Y. J. Wen, D. Novak, H. F. Liu, A. Nirmalathas, “Generation o f 140GHz optical pulses with suppressed amplitude modulation by subharmonic synchronous modelocking of Fabry-Perot semiconductor laser.” IEE Electron. Lett., Vol. 37, pp. 581-582, 2001 36 S. Watanabe, F. Futami, R. Okabe, Y. Takita, S. Ferber, R. Ludwig, C. Schubert, C. S chm idt, H.G. Weber, “ 160Gbit/s Optical 3R-Regenerator in a Fiber Transmission Experiment,” Optical Fiber Communication Conference (OFC) 2003, pp. PD16-1 - PD16-3, 2003 37 R. Flui, B. Zhu, K. Demarest, C. Allen, J. Hong, “Generation of ultrahigh- speed tunable-rate optical pulses using strongly gain-coupled dual wavelength DFB Laser diodes” IEEE Photonics Technology Letters, vol. 11, pp. 518 - 520, 1999 38 I. Shake, H. Takara, S. Kawanishi, M. Saruwatari, “High-repetition-rate optical pulse generation by using chirped optical pulses,” IEE Electron. Letters, Vol. 34, pp. 792-793, 1998 120 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 39 D.E. Leaird, S. Shen, A.M. Weiner, A. Sugita, S. Kamei, M. Ishii, K. Okamoto, “Generation o f High-Repetition-Rate WDM Pulse Trains from an Arrayed-W aveguide Grating,” IEEE Photonics Tech. Letters, vol. 13, pp. 221-223, 2001 40 K. Yiannopoulos, K. Vyrsokinos, D. Tsiokos, E. Kehayas, N. Pieros G. Theophilopoulos, T. Eloubavlis, G. Guekos H. Avramopoulos , “Pulse Repetition Frequency M ultiplication With Spectral Selection in Fabry-Perot Filters,” IEEE J. Q. Elec., vol. 40, pp. 157-165, 2004 41 B. Ramamurthy and B. Mukherjee, “Wavelength conversion in WDM networking,” IEEE Selected Areas in Communications, Vol. 16, pp. 1061 - 1073, 1998 42 T. Flasegawa, K. Inoue and K. Oda, “Polarization independent frequency conversion by fiber four-wave mixing with a polarization diversity technique,” IEEE Photon. Tech. Lett., vol. 5, pp. 947-949, 1993 43 K. Inoue, “Polarization independent wavelength conversion using fiber four- wave mixing with two orthogonal pump lights o f different frequencies,” J. Lightwave Tech., Vol. 12, pp. 1916-1920, 1994 44 K. Tsuji, H. Yokota and M. Saruwatari, “Polarization-independent wavelength conversion using four-wave mixing in single-mode fibers pumped with cross-polarized high frequency pulses,” IEICE Trans. Electron., vol. E84-C, pp. 564-570, 2001 45 C. R. Giles, V. Mizrahi, and T. Erdogan, “ Polarization-independent phase conjugation in a reflective optical mixer,” IEEE Photon. Tech. Lett., vol. 7, pp. 126-128, 1995 46 1. Cristiani, V. Degiorgio, L. Socci, F. Carbone, M. Romagnoli, “Polarization-insensitive wavelength conversion in a lithium niobate 121 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. waveguide by the cascading technique,” IEEE Photon. Tech. Lett., vol. 14, pp. 669-671,2002 47 M. Onishi, T. Okuno, T. Kashiwada, S. Ishikawa, N, Akasaka and M. Nishimura, “Highly nonlinear dispersion fiber and its application to broadband wavelength converter,” European Conference on Optical Communication (ECOC) 1997, pp. 115-118, 1997 48 J. Hiroishi, N. Kumano, K. Mukasa, R. Sugizaki, R. Miyabe, S. Matsushita, H. Tobioka, S. Namiki and T.Yagi, “Dispersion slope controlled HNL-DSF with high y o f 25 W ’km '1 and band conversion experiment using this fiber,” European Conference on Optical Communication (ECOC) 2002, Paper PD1.5, 2002 49 Y. W. Song, Z. Pan, D. Starodubov, E. Salik, S. A. Havstad, Y. Xie, A. E. Willner, and J. Feinberg, “All-fiber WDM optical crossconnect using ultrastrong widely tunable FBGs,” IEEE Photon. Tech. Lett. , Vol. 13, pp. 1103 -1105,2001 50 L.J. Richardson, V.K. Mezentsev, and S.K. Turitsyn, “Limitations o f 40 Gbit/s based dispersion managed WDM transmission: solitons versus quasi- linear propagation regime,” Optical Fiber Communication Conference (OFC) 2001, Vol. 1, pp. MF5/1-3, 2001 51 D. Kovsh, E.A. Golovchenko and A.N. Pilipetskii, “Enhancement in performance of long-haul DWDM systems via optimization of the transmission format,” Optical Fiber Communication Conference (OFC) 2002, pp. 361 - 362, 2002 52 G.C. Gupta, K. Fukuchi and T. Ogata, “Highly efficient 1 Tb/s (50 chx20 Gb/s)-2000 km RZ transmission experiment by suppressing XPM with 122 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. optimized pulsewidth,” Optical Fiber Communication Conference (OFC) 2000, Vol. l,pp. 152 - 154, 2000 53 A. Sano, Y. Miyamoto, T. Kataoka and K. Hagimoto, “Long-span repeaterless transmission systems with optical amplifiers using pulse width management,” /. Lightwave Technol., vol. 16, pp. 977 - 985, 1998 54 L.-S. Yan, S. M. R. Motaghian Nezam, A. B. Sahin, J. E. McGeehan, T. Luo, Q. Yu, and A. E. Willner, “Enhanced Robustness o f RZ WDM Systems Using Tunable Pulse-W idth Management at the Transmitter,” European Conference on Optical Communication (ECOC) 2002, Paper 10.6.2, 2002 55 K. Suzuki, K. Iwatsuki, S. Nishi and M. Saruwatari, “Error-free demultiplexing of 460 Gbit/s pulse signal using optical loop mirror including semiconductor laser amplifier,” IEE Electron. Lett., vol. 30, pp. 1501 - 1503, 1994 56 J. Hansryd, and P.A. Andrekson, “Wavelength tunable 40GHz pulse source based on fiber optical parametric amplifier,” IEE Electron. Lett., vol. 37, pp. 584-585,2001 57 S. Radic, C.J. McKinstrie, R.M. Jopson, J.C. Centanni, and A.R. Chraplyvy, “All-optical regeneration in one- and two-pump parametric amplifiers using highly nonlinear optical fiber,” IEEE Photonics Technol. Lett., vol. 15, pp. 957 - 959, 2003 58 J.H. Kim, Y.T. Byun, Y.M. Jhon, S. Lee, D.H. Woo and S.H. Kim, “All- optical half adder using semiconductor optical amplifier based devices,” Optics Comm. Vol. 218, pp. 345-349, 2003 59 A.J. Poustie, K.J. Blow, R.J. Manning, and A.E. Kelly, “All-optical pseudorandom number generator,” Optics Comm. Vol. 159, pp. 208-214, 1999 123 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 60 C. Bintjas, M. Kalyvas, G. Theophilopoulos, T. Stathopoulos, H. Avramopoulos, L. Occhi, L. Schares, G. Guekos, S. Hansmann, and R. Dall'Ara, “20 Gb/s all-optical XOR with UNI gate,” IEEE Photon. Technol. Lett., vol. 12, pp. 834 - 836, 2000 61 B.S. Robinson, S.A. Hamilton, S.J. Savage and E.P. Ippen, “40 Gbit/s all- optical XOR using a fiber-based folded ultrafast nonlinear interferometer,” Optical Fiber Communication Conference (OFC) 2002, pp. 561 - 563, 2002 62 T. Fjelde, D. Wolfson, A. Kloch, B. Dagens, A. Coquelin, I. Guillemot, F. Gaborit, F. Poingt, and M. Renaud, “Demonstration o f 20 Gbit/s all-optical logic XOR in integrated SOA-based interferometric wavelength converter,” Electron. Lett., vol. 36, pp. 1863 - 1864, 2000 63 R.P. Webb, R.J. Manning, G.D. Maxwell, and A.J. Poustie, “40 Gbit/s all- optical XOR gate based on hybrid-integrated M ach-Zehnder interferometer,” IEE Electron. Lett., vol. 39, pp. 79 - 81, 2003 64 J.H. Kim, Y.M. Jhon, Y.T. Byun, S. Lee, D.H. Woo and S.H. Kim, “All- optical XOR gate using semiconductor optical amplifiers without additional input beam ,” IEEE Photon. Technol. Lett., vol. 14, pp. 1436-1438,2002 65 H. Soto, D. Erasme and G. Guekos, “5-Gb/s XOR optical gate based on cross-polarization modulation in semiconductor optical amplifiers,” LEEE Photon. Technol. Lett., vol. 13, pp. 335 - 337, 2001 66 M.C. Ho, F.S. Yang, M.E. Marhic, and L.G. Kazovsky, “Optical parametric amplification in a high-nonlinearity fiber,” Conference on Lasers and Electro-Optics (CLEO) 1998, paper 496 - 497, 1998 67 T.M. Monro, K.M. Kiang, J.H. Lee, K. Frampton, Z. Yusoff, R. Moore, J. Tucknott, D.W. Hewak, H.N. Rutt and D J . Richardson, “High nonlinearity 124 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. extruded single-mode holey optical fibers,” Optical Fiber Communication Conference (OFC) 2002, pp. FA1/1 - FA1/3, 2002 68 O. Leclerc, B. Lavigne, D. Chiaroni, Optical Fiber Telecommunications IV A, Chapter 15, pp. 732 - 783, 2002 69 O. Leclerc, B. Lavigne, E. Balmefrezol, P. Brindel, L. Pierre, D. Rouvillain, F. Seguineau, “Optical Regeneration at 40Gb/s and Beyond,” IEEE Journal o f Lightwave Technology, vol. 21, pp. 2779-2790, 2003 70 B. Lavigne, P. Guerber, C. Janz, A. Jourdan, M. Renaud, “Full validation of an optical 3R regenerator at 20Gbit/s,” Optical Fiber Communication Conference (OFC) 2000, Vol. 3, pp. 93-95, 2000 71 E. S. Awad, P. S. Cho, C. Richardson, N. Moulton, J. Goldhar, “Optical 3R Regeneration Using a Single EAM for All-Optical Timing Extraction With Simultaneous Reshaping and W avelength Conversion,” IEEE Photon. Technol. Lett., vol. 14, pp. 1378-1380, 2002 72 C. Schubert, R. Ludwig, S. Watanabe, F. Futami, C. Schmidt, J. Berger, C. Boemer, S. Ferber, H.G. Weber, “ 160 Gbit/s wavelength converter with 3R- regenerating capability,” IEE Electron. Letters, vol. 38, pp. 903-904, 2002 125 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Bibliography O. H. Adamczyk, A.B. Sahin, Q. Yu, S. Lee, and A.E. Willner, “Statistics of PM D-induced power fading for intensity-modulated double-sideband and single sideband microwave and millimeter-wave signals,” IEEE Trans. Microwave Theory Technol., vol. 49, pp. 1962-1967, 2001 G.P. Agrawal, Fiber-Optics Communication Systems, John Wiley & Sons, 1997 G.P. Agrawal, Nonlinear Fiber Optics, New York: Academic Press, 2001 E.S. Awad, P.S. Cho, C. Richardson, N. Moulton, J. Goldhar, “Optical 3R Regeneration Using a Single EAM for All-Optical Timing Extraction With Simultaneous Reshaping and W avelength Conversion,” IEEE Photon. Technol. Lett., vol. 14, pp. 1378-1380, 2002 S. Bigo, Y. Frignac, G. Charlet, W. Idler, S. Borne, H. Gross, R. Dischler, W. Poehlmann, P. Tran, C. Simonneau, D. Bayart, G. Veith, A. Jourdan, and J.-P. Flamaide, “ 10.2Tbit/s (256x42.7Gbit/s PDM/WDM) transmission over 100km TeraLight fiber with 1.28bit/s/Hz spectral efficiency,” Optical Fiber Communication Conference (OFC) 2001, Paper PD25, 2001 C. Bintjas, M. Kalyvas, G. Theophilopoulos, T. Stathopoulos, H. Avramopoulos, L. Occhi, L. Schares, G. Guekos, S. Hansmann, and R. Dall'Ara, “20 Gb/s all- optical XOR with UNI gate,” IEEE Photon. Technol. Lett., vol. 12, pp. 834 - 836, 2000 J.-X. Cai, D.G. Foursa, C.R. Davidson, Y. Cai, G. Domagala, FI. Li, L. Liu, W.W. Patterson, A.N. Pilipetskii, M. Nissov, and, N.S. Bergano, “A DWDM Demonstration o f 3.73 Tb/s over 11,000 km using 373 RZ-DPSK Channels at 126 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. lOGb/s, ” Optical Fiber Communication Conference(OFC) 2003, Paper: PD22, 2003 J. -X. Cai, M. Nissov, A. N. Pilipetskii, C. R. Davidson, R. -M . Mu, M. A. Mills, L.Xu, D. Foursa, R. Menges, P. C. Corbett, D. Sutton, and Neal S. Bergano, “ 1.28 Tbit/s (32 x 40 Gb/s) Transmission over 4,500 km ,” 27th European Conference on Optical Communication (ECOC) 2001, Paper PD5, 2001 M.W. Chbat, J.-P. Soignb, T. Fuerst, J.T. Anthony, S. Lanne, H. Fevrier, B.M. Desthieux, A.H. Bush, and D. Penninckx, “Long term field demonstration of optical PMD compensation on an installed OC-192 link,” Optical Fiber Communication Conference (OFC) 1999, Paper PD12, 1999 A.R. Chraplyvy, “Limitation on lightwave communication imposed by optical- fiber nonlinearities,” IEEE J. Lightwave Technol., vol. 8, pp 1548-1557,1990 P. Ciprut, B. Gisin, N, Gisin, R. Passy, J.P. Von der Weid, F. Prieto, and C.W. Zimmer, “Second-order polarization mode dispersion: Impact on analog and digital transmissions,” / Lightwave Technol., vol. 16, pp. 757-771, 1998 I. Cristiani, V. Degiorgio, L. Socci, F. Carbone, M. Romagnoli, “Polarization- insensitive wavelength conversion in a lithium niobate waveguide by the cascading technique,” IEEE Photon. Tech. Lett., vol. 14, pp. 669-671, 2002 T. Fjelde, D. Wolfson, A. Kloch, B. Dagens, A. Coquelin, I. Guillemot, F. Gaborit, F. Poingt, and M. Renaud, “Demonstration o f 20 Gbit/s all-optical logic XOR in integrated SOA-based interferometric wavelength converter,” IEE Electron. Lett., vol. 36, pp. 1863 - 1864, 2000 C.R. Giles, V. Mizrahi, and T. Erdogan, “ Polarization-independent phase conjugation in a reflective optical mixer,” IEEE Photon. Tech. Lett., vol. 7, pp. 126-128, 1995 127 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. G.C. Gupta, K. Fukuchi and T. Ogata, “Highly efficient 1 Tb/s (50 chx20 Gb/s)- 2000 km RZ transmission experiment by suppressing XPM with optimized pulsewidth,” Optical Fiber Communication Conference (OFC) 2000, Vol. 1, pp. 152 - 154, 2000 J. Hansryd, and P.A. Andrekson, “W avelength tunable 40GHz pulse source based on fiber optical parametric amplifier,” IEE Electron. Lett., vol. 37, pp. 584-585, 2001 J. Hansryd, and P.A. Andrekson, “Wavelength tunable 40GHz pulse source based on fiber optical parametric amplifier,” IEE Electron. Letters, vol. 37, pp. 584-585,2001 M.C. Hauer, Q. Yu, A.E. Willner, E.R. Lyons, C.H. Lin, A.A. Au, and H.P. Lee, “Compact, all-fiber PMD emulator using an integrated series o f thin-film micro heaters,” Optical Fiber Communication Conference (OFC) 2002, Paper ThA3, 2002 T. Hasegawa, K. Inoue and K. Oda, “Polarization independent frequency conversion by fiber four-wave mixing with a polarization diversity technique,” IEEE Photon. Tech. L e tt, vol. 5, pp. 947-949, 1993 J. Hiroishi, N. Kumano, K. Mukasa, R. Sugizaki, R. Miyabe, S. Matsushita, H. Tobioka, S. Namiki and T.Yagi, “Dispersion slope controlled HNL-DSF with high y o f 25 W 'k m '1 and band conversion experiment using this fiber,” European Conference on Optical Communication (ECOC) 2002, Paper PD1.5, 2002 M.C. Ho, F.S. Yang, M.E. Marhic, and L.G. Kazovsky, “Optical parametric amplification in a high-nonlinearity fiber,” Conference on Lasers and Electro- Optics (CLEO) 1998, paper 496 - 497, 1998 128 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. R. Hui, C. Allen and K. Demarest, “PMD-insensitive SCM optical receiver using polarization diversity,” IEEE Photon. Technol. Lett., vol. 14, pp. 1632-1634, 2002 R. Hui, B. Zhu, K. Demarest, C. Allen, J. Hong, “Generation o f ultrahigh-speed tunabl e-rate optical pulses using strongly gain-coupled dual-wavelength DFB Laser diodes,” IEEE Photonics Technology Letters, vol. 11, pp. 518 - 520, 1999 R. Hofstetter, H. Schmuck, and R. Heidemann, “Dispersion effects in optical millimeter-wave systems using self-heterodyne method for transport and generation,” IEEE Trans. Microwave Theory Technol., vol. 43, pp. 2263-2269, 1995 K. Inoue, “Polarization independent wavelength conversion using fiber four- wave mixing with two orthogonal pump lights of different frequencies,” J. Lightwave Tech., Vol. 12, pp. 1916-1920, 1994 G. Ishikawa and H. Ooi, “Polarization-mode dispersion sensitivity and monitoring in 40-Gbit/s OTDM and 10-Gbit/s NRZ transmission experiments,” Optical Fiber Communication Conference (OFC) 1998, paper WC5, 1998 R. M. Jopson, L.E. Nelson, H. Kogelnik, and G.J. Foschini, “Statistics of second- order PMD depolarization,” J. Lightwave Technol., vol. 19, pp. 1882-1886, 2001 I.P. Kaminow and T.L. Koch, Optical Fiber Telecommunication. I ll, Academic Press, 1997 T. Kato, Y. Koyano, and M. Nishimura, “Temperature dependence o f chromatic dispersion in various types o f optical fibers,” Optical Fiber Communication Conference (OFC) 2000, Vol. 1, 2000, pp. 104 -106 vol.l 129 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. R. Khosravani, I.T. Lima, P. Ebrahimi, E. Ibragimov, A.E. Willner, and C.R. Menyuk, “Time and frequency domain characteristics o f polarization-mode dispersion emulators,” IEEE Photon. Technol. Lett., vol. 13, pp. 127-129, 2001 J.H. Kim, Y.M. Jhon, Y.T. Byun, S. Lee, D.H. Woo and S.H. Kim, “All-optical XOR gate using semiconductor optical amplifiers without additional input beam,” IEEE Photon. Technol. Lett., vol. 14, pp. 1436-1438, 2002 J.H. Kim, Y.T. Byun, Y.M. Jhon, S. Lee, D.H. Woo and S.H. Kim, “All-optical half adder using semiconductor optical amplifier based devices,” Optics Comm. Vol. 218, pp. 345-349, 2003 D. Kovsh, E.A. Golovchenko and A.N. Pilipetskii, “Enhancement in performance of long-haul DWDM systems via optimization o f the transmission format,” Optical Fiber Communication Conference (OFC) 2002, pp. 361 - 362, 2002 B. Lavigne, P. Guerber, C. Janz, A. Jourdan, M. Renaud, “Full validation of an optical 3R regenerator at 20Gbit/s,” Optical Fiber Communication Conference (OFC) 2000, Vol. 3, pp. 93-95, 2000 D.E. Leaird, S. Shen, A.M. Weiner, A. Sugita, S. Kamei, M. Ishii, K. Okamoto, “Generation of High-Repetition-Rate WDM Pulse Trains from an Arrayed- Waveguide Grating f IEEE Photonics Tech. Letters, vol. 13, pp. 221-223, 2001 O. Leclerc, B. Lavigne, E. Balmefrezol, P. Brindel, L. Pierre, D. Rouvillain, F. Seguineau, “Optical Regeneration at 40Gb/s and Beyond,” IEEE Journal o f Lightwave Technology, vol. 21, pp- 2779-2790, 2003 O. Leclerc, B. Lavigne, D. Chiaroni, Optical Fiber Telecommunications IV A, Chapter 15, pp. 732 - 783, 2002 130 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Tingye Li, “The Impact o f optical amplifiers on long-distance light telecommunications,” Proceedings o f the IEEE, vol. 81, pp. 1568-1579, 1993 D. Marcuse, Light Transmission Optics, Van Nostrand Reinhold, New York, 1982 D. Marcuse, A.R. Chraplyvy, and R.W. Tkach, “Effect o f fiber nonlinearities on long distance transmission,” IEEE J. Lightwave Technol., vol. 9, no. 1, 1991 D. Marcuse, A.R. Chraplyvy, and R.W. Tkach, “Dependence of cross-phase modulation on channel number in fiber WDM systems,” IEEE J. Lightwave Technol., vol. 12, pp 885-890, 1994 Y. Miyamoto, A. Hirano, K. Yonenaga, A. Sano, H. Toba, K. Murata, and O. Mitomi, “320 Gbit/s (8X 40 Gbit/s) WDM transmission over 367 km with 120 km repeater spacing using carrier-suppressed retum-to-zero format,” IEE Electron. Lett., Vol. 35, pp. 2041-2042, 1999 T.M. Monro, K.M. Kiang, J.H. Lee, K. Frampton, Z. Yusoff, R. Moore, J. Tucknott, D.W. Hewak, H.N. Rutt and D.J. Richardson, “High nonlinearity extruded single-mode holey optical fibers,” Optical Fiber Communication Conference (OFC) 2002, pp. FA1/1 - FA1/3, 2002 S.M.R. Motaghian Nezam, Y. Wang, M.C. Hauer, S. Lee, and A.E. Willner, “Simultaneous PMD monitoring o f several WDM channels using subcarrier tones,” Conference on Lasers and Electro-Optics (CLEO) 2004, Paper CFE1, 2004 S.M.R. Motaghian Nezam, Y.W. Song, Z. Pan, and A.E. Willner, “PMD monitoring in WDM systems for NRZ data using a chromatic-dispersion- regenerated clock,” Optical Fiber Communication Conference (OFC) 2002, Paper WE5, 2002 131 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Y. Namihira, and H. Wakabayashi, “Fiber length dependence of polarization mode dispersion measurement in long-length optical fibers and installed optical submarine cables,” /. Opt. Commun. Vol. 12, pp. 1-8, 1991 R. Ohhira, D. Ogasahara, and t. Ono, “Novel RZ signal format with alternate- chirp for suppression of nonlinear degradation in 40 Gb/s based W DM ,” Optical Fiber Communication Conference (OFC) 2001, Paper WM2, 2001 M. Onishi, T. Okuno, T. Kashiwada, S. Ishikawa, N, Akasaka and M. Nishimura, “Highly nonlinear dispersion fiber and its application to broadband wavelength converter,” European Conference on Optical Communication (ECOC) 1997, pp. 115-118, 1997 C.D. Poole, “Statistical treatment o f polarization dispersion in single-mode fiber,” Opt. Lett. Vol. 13, pp. 687-689, 1988 C.D. Poole, “Measurement o f polarization-mode dispersion in single-mode fibers with random mode coupling,” Opt. Lett. Vol. 14, pp. 523-525, 1989 C.D. Poole and T.E. Darcie, “Distortion related to polarization-mode-dispersion in analog lightwave systems,” J. Lightwave Technol., vol. 11, pp. 1749-1759, 1993 A.J. Poustie, K.J. Blow, R.J. Manning, and A.E. Kelly, “All-optical pseudorandom number generator,” Optics Comm. Vol. 159, pp. 208-214, 1999 S. Radic, C.J. McKinstrie, R.M. Jopson, J.C. Centanni, and A.R. Chraplyvy, “All-optical regeneration in one- and two-pump parametric amplifiers using highly nonlinear optical fiber,” IEEE Photonics Technol. Lett., vol. 15, pp. 957 - 959,2003 132 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. S. Radic, C.J. McKinstrie, R.M. Jopson, J.C. Centanni, A.R. Chraplyvy,” All- optical regeneration in one- and two-pump parametric amplifiers using highly nonlinear optical fiber,” IEEE Photon. Technol. Lett., pp. 957 - 959, 2003 B. Ramamurthy and B. Mukherjee, “Wavelength conversion in WDM networking,” IEEE Selected Areas in Communications, Vol. 16, pp. 1061 -1073, 1998 L.J. Richardson, V.K. Mezentsev, and S.K. Turitsyn, “Limitations o f 40 Gbit/s based dispersion managed WDM transmission: solitons versus quasi-linear propagation regime,” Optical Fiber Communication Conference (OFC) 2001, Vol. l,pp. MF5/1-3, 2001 B.S. Robinson, S.A. Hamilton, S.J. Savage and E.P. Ippen, “40 Gbit/s all-optical XOR using a fiber-based folded ultrafast nonlinear interferometer,” Optical Fiber Communication Conference (OFC) 2002, pp. 561 - 563, 2002 A. Sano, Y. Miyamoto, T. Kataoka and K. Hagimoto, “Long-span repeaterless transmission systems with optical amplifiers using pulse width management,” J. Lightwave Technol., vol. 16, pp. 977 - 985, 1998 I. Shake, H. Takara, S. Kawanishi, M. Saruwatari, “High-repetition-rate optical pulse generation by using chirped optical pulses,” IEE Electron. Letters, Vol. 34, pp. 792-793, 1998 Y.R. Shen, Principles o f Nonlinear Optics, Willey, New York, 1984 G. H. Smith, d. Novak and Z. Ahmed, “Technique for optical SSB generation to overcome dispersion penalties in fibre-radio systems,” IEE Electron. Lett., vol. 33, pp. 74-75, 1997 133 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Y.W. Song, Z. Pan, D. Starodubov, E. Salik, S. A. Havstad, Y. Xie, A.E. Willner, and J. Feinberg, “All-fiber WDM optical crossconnect using ultrastrong widely tunable FBGs,” IEEE Photon. Tech. Lett. , Vol. 13, pp. 1103 -1105, 2001 H. Soto, D. Erasme and G. Guekos, “5-Gb/s XOR optical gate based on cross polarization modulation in semiconductor optical amplifiers,” IEEE Photon. Technol. Lett., vol. 13, pp. 335 - 337, 2001 C. Schubert, R. Ludwig, S. Watanabe, F. Futami, C. Schmidt, J. Berger, C. Boemer, S. Ferber, H.G. Weber, “ 160 Gbit/s wavelength converter with 3R- regenerating capability,” IEE Electron. Letters, vol. 38, pp. 903-904, 2002 K. Suzuki, K. Iwatsuki, S. Nishi and M. Saruwatari, “Error-free demultiplexing o f 460 Gbit/s pulse signal using optical loop mirror including semiconductor laser amplifier,” IEE Electron. Lett., vol. 30, pp. 1501 - 1503, 1994 R.W. Tkach, A.R. Chraplyvy, F. Forghieri, A.H. Gnauck, and R.M. Derosier, “Four-photon mixing and high-speed WDM systems,” IEEE J. Lightwave Technol., vol. 13, pp. 841-849, 1995 K. Tsuji, H. Yokota and M. Saruwatari, “Polarization-independent wavelength conversion using four-wave mixing in single-mode fibers pumped with cross polarized high frequency pulses,” IEICE Trans. Electron., vol. E84-C, pp. 564- 570, 2001 J.J. Veselka and S.K. Korotky, “Compact, all-fiber PMD emulator using an integrated series of thin-film micro-heaters,” IEEE J. Sel. Top. Quantum Electron., vol. 2, pp. 330-310, 1996. S. Watanabe, F. Futami, R. Okabe, Y. Takita, S. Ferber, R. Ludwig, C. Schubert, C. Schmidt , and Fl.G. Weber, “ 160Gbit/s Optical 3R-Regenerator in a Fiber Transmission Experiment,” Optical Fiber Communication Conference (OFC) 2003, pp. PD16-1 - PD16-3, 2003 134 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. S. Watanabe, F. Futami, R. Olcabe, Y. Takita, S. Ferber, R. Ludwig, C. Schubert, C. Schmidt , and H.G. Weber, “ 160Gbit/s Optical 3R-Regenerator in a Fiber Transmission Experiment,” Optical Fiber Communication Conference (OFC) 2003, pp. PD16-1 - PD16-3, 2003 R.P. Webb, R.J. Manning, G.D. Maxwell, and A.J. Poustie, “40 Gbit/s all-optical XOR gate based on hybrid-integrated M ach-Zehnder interferometer,” IEE Electron. Lett., vol. 39, pp. 79 - 81, 2003 X. Wei, J. Leuthold, and Z. Liming, “Delay-interferometer-based optical pulse generator,” Optical Fiber Communication Conference (OFC) 2004, Paper WL6, 2004 Y. J. Wen, D. Novak, H. F. Liu, A. Nirmalathas, “Generation o f 140GHz optical pulses with suppressed amplitude modulation by subharmonic synchronous modelocking o f Fabry-Perot semiconductor laser.” IE E Electron. Lett., Vol. 37, pp. 581-582,2001 L.-S. Yan, S. M. R. Motaghian Nezam, A. B. Sahin, J. E. McGeehan, T. Luo, Q. Yu, and A. E. Willner, “Enhanced Robustness of RZ W DM Systems Using Tunable Pulse-W idth Management at the Transmitter,” European Conference on Optical Communication (ECOC) 2002, Paper 10.6.2, 2002 K. Yiannopoulos, K. Vyrsokinos, D. Tsiokos, E. Kehayas, N. Pieros G. Theophilopoulos, T. Houbavlis, G. Guekos H. Avramopoulos , “Pulse Repetition Frequency M ultiplication With Spectral Selection in Fabry-Perot Filters,” IEEE J. Q. Elec., vol. 40, pp. 157-165, 2004 135 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
Investigation of degrading effects and performance optimization in long -haul WDM transmission systems and reconfigurable networks
PDF
Experimental demonstration of optical router and signal processing functions in dynamically reconfigurable wavelength-division-multiplexed fiber -optic networks
PDF
Chromatic and polarization mode dispersion monitoring for equalization in optical fiber communication systems
PDF
Electro-optic and thermo -optic polymer micro-ring resonators and their applications
PDF
High performance components of free -space optical and fiber -optic communications systems
PDF
Dispersive properties in photonic crystals applications to beam-steering and wavelength demultiplexing
PDF
Experimental demonstrations of all -optical networking functions for WDM optical networks
PDF
All-optical devices based on carrier nonlinearities for optical filtering and spectral equalization
PDF
Experimental demonstration of techniques to improve system performance in fiber -optic communication systems using subcarrier -multiplexed and digital baseband signals
PDF
Experimental demonstrations of optical techniques to implement networking functions in dynamically reconfigurable wavelength-division-multiplexed fiber -optic communication networks
PDF
Experimental demonstration of techniques to improve system performance in non-static microwave frequency analog and digital signal transmission over fiber -optic communication systems
PDF
Generation, filtering, and application of subcarriers in optical communication systems
PDF
Active microdisk resonant devices and semiconductor optical equalizers as building blocks for future photonic circuitry
PDF
Carrier transport in photorefractive multiple -quantum -well spatial light modulators
PDF
Finite-difference time-domain modeling of photonic crystal microcavity lasers
PDF
Design, fabrication, and integration of a 3-D hybrid electronic/photonic smart camera
PDF
Management of dispersion, nonlinearity and polarization-dependent effects in high-speed reconfigurable WDM fiber optic communication systems
PDF
Joint data detection and parameter estimation: Fundamental limits and applications to optical fiber communications
PDF
Design, synthesis and characterization of novel nonlinear optical chromophores for electro-optical applications
PDF
Energy -efficient information processing and routing in wireless sensor networks: Cross -layer optimization and tradeoffs
Asset Metadata
Creator
Yu, Changyuan
(author)
Core Title
Dispersive and nonlinear effects in high-speed reconfigurable WDM optical fiber communication systems
School
Graduate School
Degree
Doctor of Philosophy
Degree Program
Electrical Engineering
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
engineering, electronics and electrical,OAI-PMH Harvest,physics, optics
Language
English
Contributor
Digitized by ProQuest
(provenance)
Advisor
Willner, Alan E. (
committee chair
), Goo, Edward (
committee member
), Hass, Stephen (
committee member
), O'Brien, John (
committee member
), Steier, William H. (
committee member
)
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-c16-462115
Unique identifier
UC11341085
Identifier
3196919.pdf (filename),usctheses-c16-462115 (legacy record id)
Legacy Identifier
3196919.pdf
Dmrecord
462115
Document Type
Dissertation
Rights
Yu, Changyuan
Type
texts
Source
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Access Conditions
The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...
Repository Name
University of Southern California Digital Library
Repository Location
USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA
Tags
engineering, electronics and electrical
physics, optics