Edited by
and
Springer Science+Business Media, LLC
The Library of Congress cataloged the first volume of this title as follows:
International Conference on Applications of Photonic Technology, Sens ing, Signal Processing, and Communications (1994: Toronto, Ont.)
Applications of photonic technology 1 edited by George A. Lampropoulos, Jacek Chrostowski, and Raymond M. Measures. p. cm.
lncludes bibliographical references and index.
1. Photonics-Congresses. 1. Lampropoulos. George A. II. Chrostowski, Jacek. III. Measures, Raymond M. IV. Title. TAI520.f56 1995
621.36-dc20 95-32967
ISBN 978-1-4757-9252-2 ISBN 978-1-4757-9250-8 (eBook) DOI 10.1007/978-1-4757-9250-8
© 1997 Springer Science+Business Media New York
Originally published by Plenum Press, New York in 1997.
Softcover reprint ofthe hardcover lst edition 1997
http://www.plenum.com
10987654321
Ali rights reserved
No part of this book may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, rnicrofilming, recording, or otherwise, without written permission from the Publisher
PREFACE
This book presents a current review ofphotonic technologies and their applications. The papers published in this book are extended versions of the papers presented at the Inter­ national Conference on Applications ofPhotonic Technology (ICAPT'96) held in Montreal, Canada, on July 29 to August 1, 1996. The theme of this event was "Closing the Gap Between Theory, Developments and Applications."
The term photonics covers both optics and optical engineering areas of growing sci­ entific and commercial importance throughout the world. It is estimated that photonic tech­ nology-related applications to increase exponentially over the next few years and will play a significant role in the global economy by reaching a quarter of a trillion of US dollars by the year 2000. The global interest and advancements of this technology are represented in this book, where leading scientists of twenty-two countries with advanced technology in photon­ ics present their latest results.
The papers selected herein are grouped to address six distinct areas ofphotonic tech­ nology. The reader will find throughout the book a combination of invited and contributed papers which reflect the state of the art today and provide some insight about the future of this technology.
The first two papers are invited. They discuss business aspects ofphotonic engineer­ ing. One examines if chip-to-chip interconnections by means of optical technology are a good economic choice, while the other discusses the photonic technology from entre­ preneurial viewpoint.
Papers related to materials and considered for photonic applications, e.g., lasers and laser technology and applications using optics, have been assembled under the section "Opti­ cal Materials, Lasers and Applied Optics."
In the following section the reader will find papers on optical transmission and ampli­ fication of signals under the title "Optical Communications." Optical Communications cur­ rently dominate the way oftransmission through optical fibers, optical switches, and gigabit networks. Furthermore, photonics play an important role in the evolution oflnternet and in­ formation superhighways.
A large section is devoted to "Photonic Devices" where the reader will find papers on optical interconnects, new optoelectronic components, waveguide devices, and photon de­ tection devices and techniques.
"Optical Computing, Signal and Image Processing" also represents a very important area in photonic technology. Invited papers present optical neural networks used as a security devices and for image classification. This section also includes papers related to remote sens­ ing and infrared technology. With the increasing availability of information through multi­ spectra and hyper-spectra multi-sensor distributed systems, advanced signal and im­ age-processing algorithms, devices with large storage (i.e., optical devices) and high-speed computing (i.e., optical computing) are required.
v
The last section puts together papers related to "Optical Measurements." Invited and contributed papers describe sensors and measurements techniques from theory to applica­ tions for mechanical, structural, medical, industrial, chemical, and environmental purposes.
All the contributions in this book were selected due to their great impact and potential in applications of photonics. We feel confident that this book will be a good reference for those working in photonics and a good starting point for those entering the field.
G.A. Lampropoulos R.A. Lessard
Optical Chip Interconnections: Economically Viable? ......................... . Maarten Kuijk, Kamel Ayadi, Roger Vounckx, GustaafBorghs, Gerhard Bickel
and Paul Heremans
Photon Capital: A Subsidy for Photonic Entrepreneurs Carl W. Nelson
Part II OPTICAL MATERIALS, LASERS AND APPLIED OPTICS
Materials
7
Resonant Excitation of Guided Modes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 G. Blau, G. Vitrant, P-A. Chollet, P. Raimond, and F. Kajzar
Graded Effective Index Planar Polymer Waveguides with Application to Erbium-Doped Waveguide Amplifiers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
J. N. McMullin, D. W. Boertjes, M. Krishnaswamy, and B. P. Keyworth
Nonlinear Polarization Shaping in a Nematic Liquid Crystal . . . . . . . . . . . . . . . . . . . . . 29 Gary Galstian and Tigran Galstyan
Refractive Index Gratings in CA3MN2GEp 12 • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • 33 R.A. Rupp, B. Sugg, and S.L. Gnatchenko
Ti:LiNb03 Single-Mode Waveguide Fabrication Parameters . . . . . . . . . . . . . . . . . . . . . 35 S. Fernandez Fernandez, J. Rodriguez Garcia, S. L. Palacios Diaz, R. Diaz Crespo,
J.M. Fernandez Diaz, and A. Guinea Rueda
Optical Parametric Process in Crystalline Molecular Layered Confinement Structures: Second Harmonic Generation in Microcavities . . . . . . . . . . . . . . . . 41
Serge Gauvin and Joseph Zyss
PZT Thin Films by Radio Frequency Magnetron Sputtering . . . . . . . . . . . . . . . . . . . . . 51 D.X. Lu, E.M.W. Wong, E.Y.B. Pun, P.S. Chung, and G.C. Jia
vii
Ferroelectric and Second Order Nonlinear Optical Effects in Surface Crystallized 0.50 Te02-0.50 LiNb03 Glasses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
M.V. Shankar and K.B.R. Varma
Lasers
Solitons in Femtosecond Lasers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61 Michel Piche, Jean-Francois Cormier, Allaoua Belahlou, Isabelle Richard, and
Xiaonong Zhu, and Simon Debluis
Fast Frequency Shift of Laser Diode by Optical Short Pulse Irradiation . . . . . . . . . . . . 69 Ryoji Ohba, Sasono Rahardjo, Seiichi Kakuma, Norihiko Takahashi,
Toshiyuki Suzuki, Takushi Tanaka, and Yuichiro Nariyoshi
Threshold Condition of Single-Mode Gain-Coupled DFB Lasers . . . . . . . . . . . . . . . . . 75 Jianyao Chen, Alain Champagne, Roman Maciejko, and Toshihiko Makino
Demonstration of a Cold Start Procedure for a Laser Source Frequency-Locked to Molecular Absorption Lines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
Jean Fran~ois Cliche, Michel Tetu, Christine Latrasse, Claude Gamache, Normand Cyr, and Bernard Villeneuve
Holographic Mirrors for Laser Mode Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87 Reza Massudi, Jean-Fran~ois Lepage, Genevieve Anctil, Sebastien Gilbert,
Damien Stryckman, Nathalie McCarthy, and Michel Piche
Phase-Noise Measurement of a Stabilized Single-Longitudinal-Mode Er3+-Doped Fiber Laser . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
Veronique Fran~ois, Gregory W. Schinn, Serguei Tchouragoulov, and Marc Levesque
Theoretical Modeling ofEr-doped Waveguide Laser Using Time-Domain Algorithm 99 T.I. Yuk, Y.M. Ma, S.F. Yu, and J.C. Palais
Applied Optics
Airborne Obstacle Avoidance Lidar Using Q-Switched Erbium-Doped Fiber Lasers .. 105 Fran~ois Seguin, Eric S. Johnstone, and Franz Blaha
Diffractive Lens Dispersion Characteristic as Wavemeter For Optical Communications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
Simon Thibault
High Efficiency Diffractive Optical Elements for Beam Coupling and Array Generation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
Yunlong Sheng, Dazeng Feng, Francois Galmiche, and Peter Kung
New Model for the Mutual Enhancement of Nonlinear Optical Phenomena in Composite Media ................................................. 125
G.R. Flynn, L. Malley, C.R. Schwarze, D.A. Pommet, and M.A. Fiddy
Second-Harmonic Generation in the Nonlinear Relief Grating Coupler Mitsuhiro Yokota
viii
131
Rainer Buhleier, Jean Louis Iehl, Jacques Henri Collet, Veronique Bardinal, Chantal Fontaine, Martin Hubner, and Jurgen Kuhl
Opto-Mechantronics: A New Concept of System Design ........................ 143 Kexing Liu
Proposal for a New Actuator for Optical Disk ................................ 149 Miguel Angel Partida Tapia, Francisco Flavio Cordova Quiroz,
Eduardo Rodriguez Escobar, Oscar Camacho Nieto, Adriano D'Luca, and Jose F. Delgado Frias
Relaxation of Holographic Record in the System with Annihilating Centers ........ 157 Michael G. Kucherenko
Part III OPTICAL COMMUNICATIONS
Suppression of the Soliton Frequency Shifts in a Two Channel WDM System ....... 167 A. Selvarajan and C.S. Aparna
Optimization of 10 Gbit/s Linear Transmission over Standard Single Mode Fiber Using Distributed Feedback Lasers ................................... 173
Paulo A.S. Goulart and Adolfo V.T. Cartaxo
Development of Optical Transmitter and Receiver for 10GBIT/s Lightwave System .. 179 M.S. Park, T.W.Yoo, J.H. Han, T.Y.Yun, B.S. Kwark, J.H. Song, J.S. Kim,
W. J. Kang, and M.S. Lee
Secure Coherence Modulation Telecommunication System ...................... 183 B. Wacogne, H. Porte, P. Mollier, and D.A. Jackson
Low Latency Optical Bus for Multiprocessor Architecture ...................... 189 Laurent Fesquet and Jacques Henri Collet
A Novel Multiple-Grating Fiber Device for Optical Communications Applications .. 195 Lawrence Chen, Seldon D. Benjamin, Peter W.E. Smith, John E. Sipe, and
Salim Juma
A Novel Single Fiber Self-Healing Structure for CATV Super Trunk .............. 201 Zan Shi, Quing Li, and Peida Ye
Constraints on the Design of Ultrahigh-Speed Optical Soliton Communication ...... 205 Chunyan Zhang and Peida Ye
The Timing Jitter Caused by Fluctuation of Gain in Soliton Communication Systems 211 Hong Wang and Peida Ye
Enlarged Shuffienet Architecture for Optical Networks ......................... 217 Andrea Borella and Franco Chiaraluce
ix
Indoor Wireless Infrared Link with a Holographic Multiple-Spot Diffuser .......... 223 Eli Simova, Ming Tai, and Mohsen Kavehrad
A High Speed Self-Routing WDM-ATM Network for Wireless Personnel Communication Services ........................................... 229
Shyh-Lin Tsao
A Phase-Noise Insensitive Modulation Scheme for CDMA Optical Local Area Networks ........................................................ 235
Ennio Gambi and Franco Chiaraluce
Harmonic Distortion Characteristics of Directly-Modulated Semiconductor Laser followed by an FM-to-IM Interferometer .............................. 243
G.Yabre and J. Le Bihan
Ultra-Short Soliton Stability in Distributed Fiber Amplifiers with Different Pumping Configurations ................................................... 249
Mario F.S. Ferreira
The Distributed Erbium-Doped Fiber Amplifier with Gradually Changing Doped Concentration .................................................... 255
Jinlong Yu, Xiaohong Ma, Jufeng Dai, Xiaoyi Dong, and Enze Yang
Part IV PHOTONIC DEVICES
Charles C. Zhou, I. Joseph, and Ray T. Chen
Bi-Directional Optical Backplane Bus With Multiple Bus Lines for High Performance Bus Systems .......................................... 269
Chunhe Zhao and Ray T. Chen
A Signal-Transparent Distributed Optical Interconnect Based 1 Ox 10 Optoelectronic Switch .......................................................... 275
Rohit Sharma and R. Ian MacDonald
A Comparison Study between 2D VLSI Circuits and 3D Circuits Based on Multifunctional Smart Pixels ........................................ 281
DietmarFey
Performance Analysis of a WDM-SDM Hybrid Optical Cross-Connect System ..... 287 Yongdong Jin and Mohsen Kavehrad
New Optoelectronic Components
Characteristics of Grating Assisted Couplers Used as Narrow-Band Optical Filters .. 293 Otto Schwelb
X
Study of SAW on New GaAs-CUTS used in Signal Processing Monolithic Nanostructured Devices ............................................ 299
J. E. Lefebvre, V. Zhang, J. Gazalet, arid T. Gryba
A New Modeling Method of Narrow Pulse Transmission in Fiber-Bragg Gratings ... 305 Xingguo Wang, Laurent Meuleman, and Michel Blonde!
Focal Plane Imaging Arrays Based on GaAs/AlGaAs Quantum Well Infrared Photodetectors .................................................... 311
H.C. Liu, M. Buchanan, Jianmeng Li, Z.R.Wasilewski, P. H. Wilson, P.A. Marshall, R. A. Barber, P. Chow-Chong, J.W. Fraser, and J. Stapledon
Limits of Acousto-Electro-Optic GaAs/GaAlAs Multiple Quantum Well Modulators 319 J. Gazalet, S. Bahlak, J.E. Lefebvre, and T. Gryba
Optical-to-Electrical Power Conversion and Data Transmission Module ........... 325 Pentti Karioja, Risto Jurva, Kimmo Keranen, Kari Tukkiniemi,
Jouko Lammasniemi, Ari Tervonen, and Marja Englund
8x8 GaAsP/GaP LED Arrays Fully Integrated with 64 Channel Si-Driver Circuits ... 333 R. Bub, M. GroB, T. Alder, W. Brockherde, and D. Jager
Polariser Formed by Tapered Metallic Fiber .................................. 339 Bing Zhu, Zhiyu Liu, Weiwei Hu and Jingren Qian
Validation of Small-Signal Analysis for Nonlinear Dispersive Fiber Systems Using Numerical Simulation .............................................. 343
Adolfo V.T. Cartaxo and Paulo A.S. Goulart
Bragg Grating Filter Synthesis with a Piezo-electric Transducer .................. 349 A.T. Alavie, M.M. Ohn, and R. Maaskant
Active Wavelength Measurement Systems Based on Quantum Well Electroabsorption Devices ......................................................... 355
T. Coroy and R.M. Measures
Low-Cost Pigtailing of Vertical Cavity Surface Emitting Laser Arrays ............. 361 B.P. Keyworth, D.J. Corazza, and J.N. McMullin
Optical WDM Components
High-Finesse Periodic Coupler as a Wavelength Selective Device for WDM ........ 367 Vincent Delisle, Udo Trutschel, Hugues Tremblay, Michel A. Duguay, and
Falk Lederer
An Integrated Optical Spectrometer for WDM ................................ 373 H.J. Hnatiuk, K.A. McGreer, and J. N. Broughton
Experiment on Optical Code-Division-Mutiple-Access Switch System Using Spectral Amplitude Encoding of Light-Emitting Diodes ......................... 379
Lucie Adam, Eli Simova, and Mohsen Kavehrad
xi
A Robust, Adaptive Multi wavelength WDM Transceiver for Optical Fiber Transmission Systems ............................................. 385
H. Willebrand, J. Sauer, M. Sprenger, P. Kalra, A. Jayasumana, and H. Temkin
Unequally Spaced Channels for Upgrading WDM System from 3-Channel to 22-Channel Preserving No FWM Crosstalk ............................ 391
Xin Miao
All-Optical SDH/SONET Self-Healing Ring with WDM ....................... 399 Xiao-Ping Yan and Peida Ye
The Availability Analysis of Many WDM Channels through a Cascade ofEDFAs without Gain Equalization .......................................... 405
Xiao Shen and Peida Ye
Photon Detection Devices and Techniques
Development of a Miniaturized Digital High Angular Resolution Laser Irradiation Detector (HARLID) and its Integration in a Laser Warning Receiver ........ 411
A. Cantin, J. Dubois, P. Webb, D. Luisi, and D. Pomerleau
Three Concepts of High Angular Resolution Laser Irradiation Detectors (HARLID) .. 417 J. Dubois, A. Cantin, P. Webb, and D. Luisi
Optimized TCM Transmission Formats in a Photon Counting Channel ............ 423 G. Cancellieri, F. Chiaraluce, and M. Mazzone
Picosecond Optical Pulse Autocorrelator Using Second-Harmonic Generation In AlGaAs Waveguides ............................................... 429
Y. Beaulieu, B. K. Garside, A. Delage, S. Janz, M.P. VanderMeer, and R. Normandin
Fast Uncooled BiPbSrCaCuO Detectors ..................................... 437 L. Ngo Phong and I. Shih
Waveguide Devices
Compact Planar Waveguide Couplers for Broadband Dual-Channel Wavelength (De)Multiplexing ................................................. 445
M.R. Paiam, C.F. Janz, R.I. MacDonald, B.P. Keyworth, and J.N. Broughton
Millimeter Wave Coplanar Structures on InP for Electro-Optic Modulation Applications ..................................................... 451
Hamid R. Khazaei, R. James, E. Berolo, and F. Ghannouchi
Ion-Exchanged Waveguides in Erbium-Doped Phosphate Glasses ................ 457 G. C. Righini, S. Capecchi, S. Pelli, A. Verciani, Y. Yan, A. J. Fabre, and
H. DerWaal
Integrated Optical Channel-Guide Vertical Couplers in Glass for Edge Coupling to Hybrid Photodetectors ............................................. 463
P.C. Noutsios and G.L. Yip
xii
Traveling Wave Electrode Design for High Speed Mach-Zehnder LiNb03 Intensity Modulators ...................................................... 469
F.Y. Gan and G. L. Yip
Fully-Packaged, Self-Calibrated, Absolute Optical Frequency Controller Based on a Surface-Emitting Nonlinear Semiconductor Waveguide: Applications to Multifrequency Optical Communication Systems ........................ 477
Martin Guy, Michel H\tu, and Steeve Boissinot
Part V OPTICAL COMPUTING, SIGNAL AND IMAGE PROCESSING
Optical Computing and Processing
Bahram Javidi
Optical Neutral Network for Rotation Invariant and Parallel Classification of 2-D Images Using Optical Image Compression ............................. 487
Henri H. Arsenault and Philippe Gagne
Optically Interconnected Analog Focal Plane Image Processor Using Hybrid-Seed Technology ...................................................... 493
Haibo Liu and Yunlong Sheng
40-Channel !-Nanosecond Digital Fiber Optic Correlator ....................... 499 Adrian Gh. Podoleanu, Ryan K. Harding, and David A. Jackson
Photo-Electric Crossbar Switching Network for Multiprocessor Systems ........... 505 A. Iwata, T. Doi, M. Nagata, S. Yokoyama, and M. Hirose
Broadband Adaptive Optoelectronic Discrete Time Signal Processor .............. 511 Emannuelle Perrenoud, Rohit Sharma, R. Ian MacDonald, and D. Lam
Ray-Tracing Approach to Computer-Generated Holography for Precison 3D Beam Patterns ......................................................... 519
S. Chang, D.A. Pommet, M.A. Fiddy, and F.C. Lin
Error-Diffusion Binarization for Neural Networks ............................. 527 Andre Granger, Tigran Galstyan, and Roger A. Lessard
Optical Position Converter for Target Tracking and Neural Network .............. 537 Alain Bergeron, Henri H. Arsenault, Michel Doucet, and Denis Gingras
A 3-D Optical Database Machine .......................................... 543 Selim Akyokus and P. Bruce Berra
Optism: A Tool for the Design of Optical Parallel Computer Architectures ......... 553 N. Langloh, V. Christopoulos, S. Van Langendonck, J. Camelis, R. Vounckx,
A. Kirk, R. Buczynski, H. Thienpont, and I. Veretennicoff
xiii
Shape Reconstruction of Metallic Objects by Artificial Vision ................... 563 C. Coulot, S. Kohler-Hemmerlin, C. Dumont, and B. Lamalle
Automatic Inspection System for Strip ofPreweathered Zinc .................... 571 J. Caron, L. Duvieubourg, J.J. Orteu, and Ph. Revolte
2-D Phase Unwrapping of Noise Contaminated Interferograms .................. 577 Ching-Wei Liao and M.A. Fiddy
Application of Two Dimensional Wavelet Transform to the Visualization of Glass Fibers in a Turbulent Flame ......................................... 583
S. Belald, D. Lebrun, and C. Ozkul
Optoelectronic Signal Processor for SAR Image Formation and Correlation ........ 591 William R. Franklin and Robert R. Kallman
Towards a Better Compression of Self-Similar Images .......................... 599 Erwin Hocevar and Walter G. Kropatcsch
Remote Sensing and Infrared Technology
Modelling of Natural Background Terrain in the Infrared and Its Effect on the Perfonnance of a High Resolution FLIR System ........................ 605
R.N. Singh and K. Rajmohan
Infrared Imaging System Using FM/TDM Hybrid Reticle ....................... 611 J.K. Bae, Y.H. Doh, D.S. Noh, and S.J. Kim
A Windows Based Facility for SIDP SBS Applications ......................... 619 A. Kong, A. Halet, G. A. Lampropoulos, J. F. Boulter, and M. Rey
OSIRIS - An Applications of Tomography for Absorbed Emissions in Remote Sensing ......................................................... 627
E.J. Llewellyn, D.A. Degenstein, I.C. McDade, R.L. Gattinger, R. King, R. Buckingham, E.H. Richardson, D.P. Murtagh, W.F.J. Evans, B.H. Solheim, K. Strong, and J.C. McConnell
Wildfire Detection with a Microsatellite ..................................... 633 Paul J. Thomas, Charles Hersom, Saad AI Kenany, and Doug Staley
3D Data Compression Systems Based on Vector Quantization for Reducing the Data Rate of Hyperspectral Imagery ...................................... 641
Shen-En Qian, Allan B. Hollinger, Dan Williams, and Davinder Manak
Calibration of the CASI Airborne Imaging Spectrometer and Application to Generating Reflectance Imagery ..................................... 655
L. Gray, J. Harron, C. Hersom, J. Freemantle, P. Shepherd, and J. Miller
Image and Signal Processing
New Optimal Tracking Algorithms for Reference with RandomTextures ........... 661 P. Refregier, F. Goudail, T. Gaidon, and M. Guillaume
xiv
D. Romare, M. Sabry-Rizk, and K.T.V. Grattan
Hyper-Spectra Space-Based Infrared Image Restoration and Composition .......... 675 G.A. Lampropoulos and J.F. Boulter
Target Classification by a New Class of Linear Discriminants .................... 685 A. Halet, G. A. Lampropoulos, and Tile Huynh
An AVME Clutter Map Design Approach for CFAR Adaptive Threshold Selection of SAR Images ..................................................... 695
V. Anastassopoulos, G. A. Lampropoulos, A. Drosopoulos, and M. Rey
Comparison of Compression Methods for SAR Imagery: Preliminary Investigation . . 70 I A. Drosopoulos and A. Damini
Radar Image Modelling and Detection Using Neural Networks .................. 71 I G. Hennessey, H. Leung, and A. Drosopoulos
Part VI OPTICAL MEASUREMENTS
Development and Performance of Fiber-Optic Sensor Systems for Absolute and Quasistatic Measurements in Civil Engineering Applications . . . . . . . . . . . . . . 723
W. J. Bock, W. Urbanczyk, T. A. Eftimov, and J. Chen
A Critical Review of Fiber-Optic Based Smart Structures ....................... 731 Kexing Liu
Semiconductor Integrated Photonic Transducer Chip for High-Resolution Displacement Measurement ......................................... 739
Hans P. Zappe and Daniel Hofstetter
Fiber-Optic Vibration and Acoustic Sensor Systems for Airport Ground Traffic Monitoring ...................................................... 745
N. Fiirstenau, T. Beyer, A. Werner, W. Schmidt, and W. Goetze
A High Resolution Reflectometer For Measuring Dynamic Strain in a Single Mode Optical Fibre ..................................................... 751
Veijo Lyori, Kari MiHitUi, Seppo Nissilil, Harri Kopola, Marja Englund, and Asko Mitrunen
Spatial-and Wavelength-Division-Mulitplexed In-Fibre Bragg Grating Sensor Network for Smart Structures Applications ............................. 757
Y.J Rao, D.A. Jackson, L. Zhang, and I. Bennion
Interferometric Distance Meter Using a Fequency-Modulated Laser Diode ......... 765 U. Minoni, G. Scotti, and F. Docchio
XV
Fibre Optic Remote Displacement Sensor for Seismic Events at High Temperature .. 771 W. Ecke, K. H. Jackel, P. Pfeifer, J. Schauer, and R. Willsch
Vibration Measurement Using a Novel Fibre-Optic Electronically-Scanned White-Light Interferometer ......................................... 777
R.H. Marshall, Y.N. Ning, A.W. Palmer, and K.T.V. Grattan
Experiences from Embedded Optical Fiber Based Cure and Stress Sensing in Composite Structures .............................................. 783
Harri Kopola, Pekka Suopajarvi, Veijo Lyori, Seppo Nissilii, and Reijo Johansson
Scientific
Measurement of the Nonlinear Ratio (n/ Aeff) in Optical Fibers Using Self-Phase Modulation Effect ................................................. 789
Ahmad K. Atieh, Piotr Myslinski, Jacek Chrostowski, and Peter Galko
Dispersion Compensating Fibre in an Optical Access Network: A Complete Component and System Characterisation .............................. 793
F. Ravet, L. Meuleman, M. Vandroogenbroek, D. Daniaux, and M. Blonde!
Time Division Multiplexing oflnterferometric Sensors at 1300nm Using a Litium Niobate Amplitude/Phase Modulator .................................. 799
C. McGarrity and D.A Jackson
Bragg Mirror Reflectance in a VCSEL Scanning Optical Microscope ............. 807 R.R. Burton, M.S. Stem, P.C. Kendall, P.N. Robson, and T.M. Benson
Aspects ofthe Use of Optical Feedback for Frequency Stabilization of Laser Diodes .. 813 R.C. Addy, A.W. Palmer, and K.T.V. Grattan
A Wavelength Monitor Based on Electroabsorption in Quantum Well Waveguide Photodetectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 819
X. Wu, D.M. Bruce, P.E. Jessop, B.J. Robinson, and D.A. Thompson
Optical Measurements Using Bessel Beams .................................. 825 Brigitte Belanger, Tigran Galstian, Xiaonong Zhu, and Michel Piche
A Silica Based Integrated Optic Microwave Power Sensor ...................... 831 A.J.P. Hnatiw, R.I. MacDonald, P.S. Apte, and W.O. MacDonald
Dimensional Measurements Analysis of a Silicon HalfLadeer Microvertex Detector .. 837 G. Ambrosi, T. Angelescu, R. Battiston, M. Biasini, D. Dascalu, C. Gingu,
M. Ionica, Al. Mihul, M. Pauluzzi, and V. Postolache
Medical and Industrial
Vision Control System for the Laser Cladding Process ......................... 843 Fabrice Meriaudeau, Eric Renier, Christophe Dumont, and Frederic Truchetet
Fiber-Link Vibration Immunity for an Extrinsic Faraday Current Sensor ........... 855 Norman Fisher and David A. Jackson
xvi
Applications of Low Coherence Interferometry to Dynamic Oil Film Thickness Measurement ..................................................... 863
S.R. Taplin, A. Gh. Podoleanu, D.J. Webb, D.A. Jackson, and S.R. Nattrass
Electroabsorption Modulators for Broadband Fiber Electro-Optic Field Sensors ..... 871 A. StOhr, R. Heinzelmann, T. Alder, R. BuB, and D. Jager
Optical Measurements of Psychomotor Interventions on Dysfunctional Knee Cycle: A Case Study .................................................... 877
J.H. Alvoeiro, A.N. Pinto, J.L. Pinto, H.M. Oliveira, M.A. Santos, and L.F. Semblano
Chemical and Environmental
Sensor for High Air Humidity Measurement ................................. 883 Vojko Matko and Dali Donlagic
Dynamic Measurements with Polymer Lightguides for a Selective Organic Vapor Detection ........................................................ 889
T. Ehrenreich, J. Wolters, R. Podgorsek, and H. Franke
Photolytic Spectroscopic Detection of Herbicides and BTEX in Water ............. 897 Mark Johnson and Paul Melbourne
Brillouin Loss Based Distributed Temperature Sensor Using a Single Source ....... 903 N.A. Heron, X. Bao, D.J. Webb, and D.A. Jackson
Integrated Optical Sensors For Pesticide Analysis ............................. 909 B.J. Luff, R.D. Harris, J.S. Wilkinson, J. Piehler, A. Brecht, G. Gauglitz,
R.A. Abuknesha, and U. Hollenbach
Holographic Gratings in Photopolymers as Optical Gas Monitors ................. 915 El Mehdi El Joudi and Hilmar Franke
Index ................................................................. 923
Maarten Kuijk, Kamel Ayadi, and Roger Vounckx
Electronics Department I Lab. of Microelectronics & Technology ETRO/LAMI University of Brussels, Pleinlaan 2, 1050 Brussels, Belgium
Gustaaf Borghs, Gerhard Bickel, and Paul Heremans
Department MAP /MBE IMEC Kapeldreef 75, 3001 Leuven, Belgium
INTRODUCTION
The idea of optical interconnections between integrated circuits or even within these circuits has recently gained support based on the ability of optics to cope with a much larger numbers of parallel paths. Yet, electronics designers have been very reluctant to turn to optics. The main reason is the inferior performance of opto-electronic devices compared to electronics in a number of significant ways. Due to diffraction limits opto-electronic devices are large and consume a lot of energy. They also have a significant impact on the production efficiency of the produced integrated circuits. The integration of III-V semiconductors with silicon is currently considered to have a major negative impact on productivity and yield. Optical pathways have been difficult to realize and remain costly. In this paper we .indicate possible ways to improve the performance of optical interconnects in all those areas where optics are considered inferior. We illustrate this through some of our own results in this area.
IMPROVING THE OPTICAL CHIP INTERCONNECT
In the optical chip interconnect proposals, one of the main yield killers is the integration problem of the Si-VLSI chip with III-V (GaAs) semiconductor chip. We provide two original ways to attack this problem. At the transmission side we propose to use capacitive coupling between the Si chip and the GaAs chip, and at the reception side, we propose to combine sense-amplifiers with integrated photo-diodes.
Applications of Photonic Technology 2, Edited by G.A. Lampropoulos and R.A. Lessard, Plenum Press, New York, 1997
Integrated Receiver: Sense Amplifier with integrated CMOS photo-diodes
For optical chip-interconnections, the location where the information is to arrive finally, is somewhere on a Si-VLSI chip. If the receiving photo-diodes can be part of the standard CMOS of this Si-VLSI-chip, hybrid solutions like flip-chipped solder-bumped GaAs on Si (which substantially lower the total yield) can be avoided. The light is received and converted on the place where it is used.
What we propose is to combine the well-known sense-amplifier with integrated photo-diodes. Sense amplfiers are extensively used in all types of memories, to convert the weak signal on the column lines into clear digital signals. They operate with thousands in parallel. The first to report about sense-amplifiers, be it with external photo-diodes was K. Miyake et all . The photo-diodes are formed here by extending the drains of the NMOS transistors, or the drains of the PMOS transistors of the coupled inverter from the sense-amplifier.
a 4-------t
Figure 1. The combination of a sense-amplifier with integrated photo-diodes. Differential light input is converted into a digital signal (0 - SV) in the VLSI system itself. At the right the demonstrator circuit of the first sense-amplifier receiver with integrated photo-diodes in 0,7 J.im CMOS. Two light-spots are incident on the two photo-diodes of 15 J.im x 15 J.iffi. The effective circuit area is about 40 J.im x 40 )lm.
The extra advantage of the sense-amplifier is that h does not consume any De­ power. Typical light receivers are based on trans-impedance converters, dissipating continuously. In the system described here, light is converted into charges which are accumulated on the photo-diode capacitance, and at the moment the elk-signal is pulsed to high a competition (by a positive feedback) system renders a digital output signal for use on the same chip. The dynamic operation is very analogous to the very sensitive differential pair of optical thyristors2, with which even systems demonstrators were built3.
Figure 1 shows the first sense-amplifier with integrated photo-diodes, and a photograph of the chip during test. One of the digital outputs is shown in
2
Figure 2. The other output gives exactly the opposite signal (both Return-To­ Zero). After a conversion, when the clock is back low, the reset-signal is to be pulsed shortly to a high level, to reset the system. We kept two clock signals separately addressable, to leave room for interesting experiments. However the clock and the reset signals can be combined to one clock signal by a small circuit consisting of a few extra transistors.
~ 5
0 20 40 60 80 100 time (ll'l)
Figure 2. The first circuit of this type already operates at 180 MHz, with an optical sensitivity of 170 fJ. Shown is one digital output Q leaving the chip. The technology is not intended for this high operational frequency. More recent technology will increase performance.
The receiver as shown in Figure 1 operated at a considerable speed, allowing a bit rate per channel of 180 Mbit/second. At this speed, a good reception is achieved with light input pulses of 170 fJ. With respect to the optical thyristor pair -where we obtained 2.6 fJ2- this sensitivity is relative poor. However we are confident that (since the operation is quite similar to the operation in the thyristor-pair system) we can easily enhance the sensitivity by at least one order of magnitude.
Hybridisation of Light Sources and Si-VLSI: Capacitive Coupling
A second place where we would like to avoid hybridisation of GaAs with Si-VLSI chips, is at the light emission side. Unfortunately, standard CMOS Si circuits cannot generate light, and hybridisation remains necessary.
The state-of-the-art connection between GaAs and Si is the flip-chip solder bump connection. Several problems related to this connection render this system less likely to be used in optical chip interconnects. To start with, plating is required, at both sides: at the silicon and the GaAs side. To achieve a high number of contacts and hence a high resolution, extra masks and associated lithography are required. A heat step is necessary for melting the solder. Detaching the GaAs -in case of poor operation of the light emitters- is difficult. And finally, the expansion coefficients are different for GaAs ad Si. This induces unavoidably shear stress (unless the substrate is removed), with possibly delamination when having smaller contacts. Moreover, laser operation is not benefiting from a stress-full environment.
3
What we propose here for the first time is a low stress, and detachable system by capacitive coupling4 . We demonstrate by simulations its useability for connection of light sources to Si-VLSI.
air, vacuum or glue or liquid with high E
Figure 3. Capacitive coupling as the high-yield alternative for conductive coupling systems like flip-chip solder bump technology. Instead of connecting the electrodes with difficult metallurgical compounds, a dielectric is sufficient to couple high speed information and energy from one chip to the other chip.
The principle of capacitive connection is shown in Figure 3. On both chips capacitive plates are fabricated. These can be the usual pads for bonding, or smaller. Only little or no extra device processing is required for this. A layer of air, glue or a liquid fills the gap between the two chips. When a glue is used, the light sources have to be tested before the glue is cured. With a liquid intermediate, it will be possible to exchange the chip with light emitters until the construction is sealed. This sealant should be somewhat flexible, to allow the GaAs die to expand differently than the Si chip to which it is attached. The surface of at least one of the chips should be relatively flat to support some sliding. In this way, shear stress is avoided.
SI·VLSI 111-V technology
~ VCSEL (luor) Dr LEO
Figure 4. A set-up to connect Si circuitry with a light emitter in e.g. GaAs by capacitive coupling . The three-stage oscillator at the left gives can be allowed to oscillate by the enable input. When oscillating, the AC-current is rectified in the GaAs by fast Schottky diodes. The Schottky diodes are only a small overhead.
Some overhead is added at the GaAs side, where it is necessary to have rectifying elements. In Fig. 4 the following system is shown: a VCO (voltage controlled oscillator), controlled by the enable input, modulates the high oscillation frequency, which in its turn determines the mean value of the current through the LED or Laser. In the GaAs, a rectifier turns the AC into DC. The oscillation frequency is a few GHz with current Si technology, and the allowed modulation frequency is near to half the oscillation frequency. The current through the light emitter (not shown) is gently pulsing. When this is of concern, a three phase rectifier can be used instead. A DC bias current for lasers (to stay in standby) can in this way also be generated.
4
Besides the aforementioned nice expected advantages, one would expect a low coupling efficiency. First, this depends on the obtained value of the coupling capacitance, and secondly, if stated in terms of quantum efficiency, (current through the light emitter/ current drained by the oscillator) this can be relatively high (Fig 5). This is because the system of Fig. 4 is also a step-down converter. The maximum quantum efficiency is 20.0 %, and not 100% if the voltage drop across the LED/Laser is smaller than Vcc/2. The simulated result of Fig. 5 expects a voltage drop of 1,5 V across the LED/Laser, 2 x 0.5 V across conducting schottky diodes (with 5 fF capacitance), and a Vee of SV. To obtain enough capacitance, e.g. 50 fF, 30 Jlm x 30 Jlm plates with a filling of Ethylene Glycol at a distance of 2 Jlm would suffice (if the oscillator operates at 3 Ghz).
100
#' 80
..1' ....... 40 , "' .:r
0 200 400 600 800 1000 1200
Figure 5. The curves are linked to a given simulator set-up. The achieved coupling capacitance gives the final current through the light emitter (over which a 1,5 V voltage drop is assumed) and the current efficiency of the operation. Isupply is the current drawn by the oscillator.
eo Water
EO 0
Frequency (Hz)
Figure 6. The relative permittivity versus frequency for various liquids. To achieve sufficient coupling capacitance, a material in-between the plates of the capacitance can help. Higher stated oscillator operates above 1 GHz.
5
The permittivity versus frequency of various liquids is shown in Fig. 6, and demonstrates that at high frequency the permittivity lowers. Not only the permittivity determines the final choice of he filling material, also mechanical and chemical aspects have to be considered. Water, e.g. will pose problems of freezing, evaporation and corrosion. For easy handling, a tixotropic material might be the better solution.
CONCLUSION
We showed some of the weak performances of optical systems for interchip interconnections can be improved. Capacitive coupling between the silicon chip and the optical emittors are an interesting alternative to flip chip bonding. It offers a number of advantages as far as yield, productivity and testability are concerned. Although very sensitive GaAs based detectors do exist2, there is hope that silicon detectors in conjunction with sense amplifiers, fully integrated in standard CMOS technology, can perform well enough to avoid hybridized systems on the reception side.
References
1 Use of separate sense-amplifiers and detectors first demonstrated by: K. Miyake et all.:"Fabrication and evaluation of... " Jpn. J. Appl. Phys, vol 34, 1246 (1995).
2 M. Kuijk, B. Kniipfer, P. Heremans, R. Vounckx and G. Borghs: 'Down­ scaling differential pairs of depleted optical thyristors,' IEEE Phot. Techn. Lett. vol. 7, 646, 1995.
3 A. Kirk, H. Thienpont, A. Goulet, P. Heremans, G. Borghs, M. Kuijk, R. Vounckx and I. Veretennicoff: 'Demonstration of Optoelectronic Logic Operations with Differential Pairs of Optical Thyristors., IEEE Photonics Technology Letters, vol 8, 467, (1996).
4 D. Salzman & T. Knight: IEEE on compn. pack, manufact. techn., vol18, 277 (1995)
5 A. von Hippel "Dielectric Materials and Applications", Artech House, ISBN 0-89006-805-4
6
STARTING SMALL
Carl W Nelson
Carl Nelson Consulting, Inc* PO Box 18371, Washington, DC 20036 e-mail: cwnelson@ids2.idsonline.com Web: http://www. idsonline. comlbusinesslcwnelson
Small entrepreneurial companies, with their clear agility advantage over the giants, created America's information technology industry. They started their comer of the industry with little help from government which favored the giants' mainframe mindset. Microsoft, Intel, and Apple became common names after Xerox developed (and let go) the seeds of new industries: laser printer, desk-top, mouse. A few giants (Motorola, Hewlett-Packard, AT&T) then furnished the new industry with an infrastructure that could only come from giant investments that the revolutionaries could neither raise nor implement. Some of the gnats rose into that class, like Intel which has invested billions in chip fabrication facilities.
Most of those entrepreneurs started their companies with sweat and savings, without committees of vice-presidents figuring ROI to three significant figures for markets that did not yet exist. Now comes another information wave, photonics, as the transfer medium for gazigabits per second. This wave, though, has some subsidies that once went to the giants. One subsidy in particular helped the gnats- Small Business Innovation Research (SBIR)- but only when imposed by Congress on government's R&D spending agencies.
One SBIR program is investing a large share in photonics and in companies that could threaten revolution. The Ballistic Missile Defense Organization (aka Star Wars) has economic results from its SBIR to prove that a small investment (a political euphemism for spending) in infant technology can attract private capital to continue the development of new technology. BMDO departs from the government R&D investment model in that it uses the private capital commitment as a market signal and does not depend solely on the scientific opinion of government technology experts.
SBIR, ESPECIALLY AT BMDO
SBffi funds US small firm R&D in two phases: Phase 1 under $100,000 to study a new concept, and Phase 2 normally under $1M(illion) to prototype it. After SBIR the government retains only a right to royalty-free use for government purposes; no repayment, no equity. It's free money.
SBIR assumes a market failure, where a public good will not happen because individual economic actors have no incentive to invest1• Although legitimate market failure happens frequently (like basic research in the physics of light waves), politics tends to sully it with subsidies (corporate welfare) for the politically active, like ethanol producers.
In 1982 Congress again found market failure in US R&D's ignoring small business's advantages2:
more innovation per employee, more new jobs created, and more new products. So Congress enacted SBIR with a double purpose: get small business innovations into federal R&D and then into the marketplace. By
Applications of Photonic Technology 2. Edited by G.A. Lampropoulos and R.A. Lessard, Plenum Press, New York, 1997 7
1992 more studies3, including by the General Accounting Office4, led Congress to double SBIR with more commercialization.
BMDO needs new technology created and commercialized so that private capital will create products and improve the technology which a healthy industry can later supply to anti-missile defenses. But BMDO, a typical mission agency, would not have made dual-use investments without a mandate. The two large mission agencies, DOD and NASA, mostly treat dual-use as merely a windfall. BMDO SBIR reduces the technical risk- the chance that the idea will not work; it does not attack business risk - the chance that it will not sell. BMDO typically gives the firm a financial incentive: $300,000 to reach some technical performance goal followed by a matching of private sector investment for another $400,000. If all goes well, BMDO may offer more funding with an rising match ratio which makes BMDO an oxymoronic government capitalist.
Other agencies (the other 95% of SBIR spending) have relatively rigid formulae for SBIR awards consistent with a government tendency to uniform mistreatment for politically weak actors and little thought of private investment.
BMDO ECONOMIC SUCCESS
Although BMDO's investment made the expected technological advance from Defense R&D, the bonus is economic. BMDO's SBIR showed an auditable economic return to its firms, and thus to the nation, for BMDO's $300M investment in 453 Phase 2s by 223 firms through March 1996.
Eleven firms who sold public stock by Initial Public Offering (IPO) had BMDO SBIR total funding of $16M. Seven won their first Phase 2 from BMDO at a median size of eight employees. Their total market cap(italization) has swung between $300M and $800M, normal gyrations for young stocks. In addition, two more firms, with small SBIR contributions, raised another $60M by IPO. The 13 companies together had a spring-1996 market cap of $1500M riding the froth of technology stocks. Ten more firms say they will go public for about $200M "when market conditions are right".
Private investment, a predictor of future market action, is pouring into BMDO SBIR projects - over $80M spent or planned for $75M of Phase 2 SBIR starts since 1992. Money for the matching comes from many sources, most usefully from firms already in a market. Typical matches are: multi-level secure imagery with a software company, diamond-like field emitters with a video company, and a laser transmitter with a dental products company.
Six firms each had spring-1996 market cap over $100M5• Two of interest to photonics are:
Company A, which had raised 20% of its initial capital in 1988 with two BMDO SBIRs, raised $11M by IPO while it was the sole world supplier of a blue LED. It has a $200M market cap in summer-1996 in the world blue LED competition.
Company B, which started with two SBIRs and got its first from BMDO in 1987 when it had six employees, went public in late 1993. Two years later it had already entered into and cashed out of a two-year old joint venture with an estimated $20M (including future royalties) from a large filter maker, completed a secondary offering for $18M, bought three companies, and reported first quarter 1996 profits near $0.5M on revenues of $10M. Its July 1966 market cap is $150M.
PHOTONICS COMPANIES
Company C, a spinoff from a large defense laboratory, developed a semiconductor laser that produced record power levels in a circularly symmetric diffraction-limited beam with no external beam shaping optics needed. After attracting $3 of private investment for every BMDO SBIR dollar, it expects large scale production in 1997 for at least two market segments.
Company D's founder, an escapee from an American giant, started and left a firm that grew to 80 employees, and then started another (this time profitable) company that is growing rapidly with $3M of BMDO SBIR in optical storage and displays, especially exploiting polarization. It now has 45 employees and
8
has spun out a subsidiary to mass-market a $69 3-D stereo display product for TV sets.
Company E started as four people in 1990, got a BMDO Phase 2 for a VCSEL technology, raised several millions including $11M in late 1995 (from a consortium of venture capital, an electronics firm, a university, and a musician), for a growing matching ratio, and now claims to sell the only commercially available VCSEL product.
Companies F and G, alongside $4M BMDO SBIR, arranged independent manufacturing companies funded by venture capital of $1OM (and rising). Company F has a variety of holographic products. Company G does waveguides and switches.
Company H was started by an entrepreneurial escapee from a top-ten SBIR winner. Its three BMDO Phase 2s (plus two Phase 2s from other agencies) and private funds are moving it to stability if it can get the price of sophisticated optical devices down to marketable levels. It expects to market a line of photonic products in 1997.
New Phase 2 BMDO wagers on future photonics include fiber-to-fiber connection, wavelength division multiplexing, holographic filters, VCSELs, asynchronous logic, optical correlator, tunable diode lasers, E-0 switch, blue laser, MEMS, fiber lasers, inter-chip connection, gallium nitride films, wavelength tolerant laser receiver, transparent transistors, anti-reflection coatings, conducting polymer films, optical clock signals, and multilayer mass storage.
WHAT HATH SBIR WROUGHT?
These unique BMDO successes help the particular companies but the uniqueness hurts Congress's ever getting the wanted commercial punch. Has SBIR even improved the lot of the implied customer, the high-tech small business community? No easy answer, as GAO found in its 1991 review. As in all subsidy programs, the beneficiaries and their representatives claim great good in the absence of those who would have won under different rules (or no rules). For example, companies with 500-1000 employees who cannot compete with Rockwell found their expected pool of R&D money had shrunk. When Congress found in 1992 that the percentage of federal R&D going to the client small business group had changed little in nine years of SBIR, its remedy was another eight years of the same medicine and a doubled dose.
On balance, SBIR has supplied a pleasant fiction but neither changed much nor did any great harm to the markets. The best companies would have succeeded anyway; the rest merely did government contract R&D. BMDO SBIR, though, did what BMDO would not have done without SBIR. In 1992, most agencies, especially DOD, were proud to have funded the same R&D as before. Government managers, after all, have little incentive for small firm success, and are under pressure to show "relevance" to current programs. In practice, the long term R&D of these mission agencies has a shorter focus than perhaps their managers realize.
A revealing measure is the market value of the top SBIR users government-wide (the 19 with over $10M each of SBIR). After together receiving over $300M6, only two have a public market value; one declined by half after $28M of SBIR and the other, Company B, went from no market cap to $150M. Of the 800 firms with over $1M of SBIR, not one showed up on Business Week's list of best three-year growth companies (of which half are information technologyf. Only one of the 800 made Inc's 100 list8 - the one that settled a government fraud prosecution for $3M. The public markets put little value on SBIR. For all but B, if and when the subsidy ends, the jobs will die (which in politics assures that the subsidy will not end). Those good-for-government companies attract few investors because the ROI is uncompetitively low (which is one reason they depend on low margin government R&D contracts).
If there is actually commerce spinning off, in quantities enough to justify the government's investment, the private companies have kept their secrets (as is their right). Even most company commercialization reports that I saw feebly recited hopes and assumptions ("assuming we can capture 5% of that market"). In my years at BMDO deciding 1100 Phase 2 proposals I saw little commercial success from the top government-wide SBIR recipients. In fact, they practically stopped proposing when they realized that BMDO sought commercialization. One firm which had submitted 26 proposals one year now submits none or one per year to BMDO.
9
Wait, don't blame the companies. They only proposed what they did best and wanted to do more of (for whatever reason). The government picked the winners. If any VC or mutual fund manager picked the same "winners" as government has, the fund would soon have a new manager. BMDO's experience, though, says that government can make relatively efficient investments (at least by government standards). If Congress wants dual-use and commercialization, it will have to redraw the SBIR law. The federal agencies will not act uncompelled, an attitude stiffened by its experiences with the Technology Reinvestment and Advanced Technology Programs.
THE SMALL PROTONIC FUTURE
Unpredictable! As Priestly said in 1927, "Solemn prophecy .. is obviously a futile proceeding, except in so far as it makes our descendants laugh." But Marr's two biggest political events since the Cold War - free markets and information technologies9 - combine to guarantee photonics a lively future.
Unpredictable. For the information technology industry, where photonics finds its home, IPO predictor The Red Herring magazine in spring-1994 missed the 1995 explosions of both Netscape and UUNet. New technologies open new doors to unfamiliar and inestimable worlds. The digital versatile disk (DVD), having now generally agreed formats, may well incentivize another generation of photonics products to keep up the doubling every three years of CD players sold 10• Optical correlators have started a march on silicon11 • The "mediaprocessor" from firms like MicroUnity 12 will speed video into mainstream information at speeds as different from today's as the Intel chips are different from the Hollerith card. Photonics are needed to fill the vacuum of the unused 90% of submarine fiber-optic cable capacityY
Nevertheless, predictions:
Investment will boom until the bubble bursts again, "a money bomb has hit our industry" 14 • With the information technology market in pell-mell advance, capital will eagerly seek opportunities to exploit the new Telecommunications Act and the exploding demand for both software and hardware with ample share for photonics. The links among video, photonics, software, and Internet will be giant and bound to expand beyond any estimates made today. The investment evidence is compelling:
1) Returns. Annual returns on VC investments were 51% in 199515 and four of the top 21 venture firms with 1993 Information Technology IPOs had post-IPO performance over 100% 16 and IPOs in the Standard and Poor's New Issues Index nearly doubled in 199517; in the last four years 2300 IPOs have come to market18 and in 1995, five IPOs doubled in price the first day whereas only five did so in the entire decade before 199519 ; small stocks outperforming the Standard & Poor's index by 26% since 199!2°; small-company mutual funds had the highest five-year return and in a virtual tie for ten-years21 , the lottery psychology also operates in the stock markets as in Oracle that went from a market cap of $188M at IPO to $8200M; and "you pick the right start-up, the return beats any other form of legal investment and probably most illegal investments "22;
2) Volume. VC firms invested $2200M in private companies in the first quarter of 1996, double the amount of first quarter 199523 , and 163 companies sold $8200M in first quarter of 1996 IPOs24 and the amount of VC in the average five-year old venture-backed firm rose 65% to $11.6M in 1994 up from $7M in 198525 ; technology is 10% of the US Gross Domestic Product and predicted to grow to 20%26
(which implies a growth of $600,000M); many mutual funds for small companies are so oversubscribed that they have closed their gates to new investors; one billion dollar fund advertises itself as post-VC; in first quarter 1996, $71 billion of new money flowed into mutual funds27 (oh, those baby boomers); and complaints that "there are so many transactions that it's killing the bankers"28 ;
3) Information Technology Growth. Of the $7200M in private equity financing by venture backed companies 47% went to information technology29 ; of Inc's fastest growing 100 firms (1991-1995) half are information technology; the most profitable IPO ever was a company with a powerful new switch for networks of PCs and workstations30•
The government invests too. SBIR wiii pass out a billion dollars a year to the best investment opportunities. In principle. The practice wiii depart because the government's definition of investment and value don't match the private sector's . .The National Science Foundation and the Air Force espouse the same criteria despite seeing far different worlds. Eventually the government wiii buy photonics for such Air Force
10
uses as fly-by-light31 but probably only from large firms which can both resonate with large agencies and influence procurement appropriations.
The mission agencies (who would use the technology themselves) will still buy for government. The non-mission agencies will continue to focus on minimizing criticism of the decision makers. Since they have a higher discount rate than they will admit, they will forego a future gain to feed an ongoing effort, especially if the short term results are predictable. And since they have no future profit stream, they can make no present worth nor ROI calculation. Such behavior is no way evil, only a natural consequence of laws from Congress that wrote the commercialization mandate.
Obtusely, SBIR-funded start-ups may actually lose out to the non-SBIRs because SBIR ignores two aspects that drive market competition: time-to-market and cost. Mission agencies don't think of markets. They want performance at whatever cost they have to pay and whenever they can get it. When whole generations of multi-media products see product rollout windows of 12-18 months32 the plodding government-fed will be leapfrogged by the next generation. When that happens, the free-money won't seem such a bargain. Photonics, like other competitive technologies, cannot afford the luxury of academic thoroughness and maximum performance.
SBIR will continue by law until at least 2000. Then the small business beneficiaries will argue for more subsidy while promoting national principles of individual enterprise and self-reliance. The Congressional Small Business Committees will take credit for any and all success in an election year - a replay of 1992. The committees will study the successes (and ignore the failures) and declare success without any control group - a clear violation of Campbell's dictum against one-shot case studies33• It's just a continuation of a corporate welfare approach where politics overwhelms economic calculus34 • Large institutions (government agencies, non-profits, and DOD prime contractors), out of whose pie the subsidy is sliced, will remain silent as long as the size of the subsidy in still smaller than the general small business percentage goals (into which they can conveniently divert SBIR).
Government could invest intelligently in photonics companies but it is unlikely to have any consistent approach since each agency's SBIR autonomy impedesjoint action. Agency inflexibility and continual shifting of "priorities" will overfund some poor market prospects and underfund some potential market winners. Even the agencies seeking late-stage development are unwilling to invest more than a standard uniform "cookie­ cutter" amount in any project. BMDO did range widely in its funding when private capital showed which projects might breathe unaided when the government subsidy respirator was removed.
While stock markets and technology investment will fluctuate (a law of nature), market-failure will still rationalize subsidy. The subsidy advocates rely on politically useful images like linear advance, science is progress, and the rugged pioneer. The subsidy opponents, who find government inept at choosing whom to subsidizel5, must inconveniently argue that nothing is better than something. Unfortunately, in a political free market (where politicians buy votes with public funds36) economics will lose in politics where short­ term benefits lead to long term (the next election) costs as politicians use economics like the drunk uses the lamp post, more for support than for light.
Photonics has many futures. One was confirmed when Hewlett-Packard pulled the plug on its mainframe computer while it boosted its dividend37• Guru Gilder predicts "Unknown entrepreneurs will invent new technologies to solve problems that hex Internet commerce. The Internet will multiply by a factor of millions the power of one person at a computer. "38 Dumb networks, with giant bandwidth coupled to smart terminals, open a gaping portal to the photonic entrepreneur. Government may even do more good than harm with its SBIR for those entrepreneurs.
Unpredictable!
*Author's Note: Author managed the BMDO SBIR program from 1987 to early 1996. Statements about present BMDO policies presume a continuation of the author's policies.
REFERENCES
1. See for example RR Nelson, PM Romer, "Science, Economic Growth, and Public Policy", Challenge, Vol 39, March 1966,
11
2. Fineman S and Fuentevilla W, Indicators of International Trends in Technological Innovation, Gelman Research Associates, Report to the NSF, 1976; and Armington C, Harris C, Olds M, Formation and Growth in High Technology Businesses: A Regional Assessment, The Brookings Institution, Report NSF/ISI-3-83016, September 30, 1983; and Cooper AC, "R&D Is More Efficient in Small Companies", Harvard Business Review, p75-83, May/June1964
3. Romeo AA and Rapoport J, Social Versus Private Returns to the Innovations by Small Firms Compared to Large Firms, University of Connecticut study for SBA (NTIS PB85-196996), July 1984; and Dearden J, Ickes BW, Samuelson L, "To Innovate or Not to Innovate: Incentives and Innovation in Hierarchies", The American Economic Review, Dec 1990; and Achs ZJ and Audretsch AB, "Innovation in Large and Small Firms: An Empirical Analysis", The American Economic Review, 78, p678-690, 1988; and Markusen AR, Hall P, Glasmeier A, High Tech America: The What, How, Where, and Why of the Sunrise Industries, Allen and Unwin, Boston, 1983
4. US General Accounting Office, Federal Research: Small Business Innovation Research Shows Success but Can Be Strengthened, Report GAO/RCED-92-37, 1992
5. see descriptions in CW Nelson, "Wavelength Division Mini-Money: Small Subsidies for Small Business", paper at SPIE International Symposium, Photonics West, Jan 1996, to be published by SPIE.
6. Data from Community of Science, on the Web at http:l!cos.gdb.orglbestlfedfundlsbir!sbir-intro.html
7. "1996 Hot Growth Companies", Business Week, May 27, 1996
8. "Show Time", Inc, May 1996
9. A Marr, Ruling Britannia: The Failure and Future of British Democracy, Michael Joseph, London, 1995
10. AE Bell, "Next-Generation Compact Discs", Scientific American, July 1966, pp42-46
11. "Correlation is now causation", The Economist, May 4, 1996
12. RD Hof, "Silicon Dreamers vs. The PC", Business Week, May 13, 1996, p78-80
13. "The death of distance: A Survey of telecommunications", The Economist, Sept 30, 1995
14. TC Draper, "The Return of Can-Do Entrepreneuralism", Upside, February 1996, on the Web at http :I /www. upside. comlresourcelprint/9602/vv9602. html
15. Source: Venture Economics, quoted by M Selz, Wall Street Journal, May 14,1996, pB2
16. Source: VentureOne Corp, quoted by ZA Her lick, "Venture Capital Roundup", The Red Herring, Parch 94, Issue 9, on the Web at http://www.herring.com/mag/issue09/capital.html
17. D Lohse, "Small Issues' Big Gains May Hold for Long Haul", Wall Street Journal, Apri129, 1996, pCI
18. "No Tech, No Takers", Inc, May 1996, p45
19. M Shah, quoted in The Red Herring Guide to Technology Finance, Spring 1996, pl7
20. C Gould, "Why Small Stocks Are Leading the Parade", The New York Times, May 26, 1996
21. Source: Lipper Analytic Services, quoted Wall Street Journal, July 3, 1996
22. M Moritz, quoted by S Kaufman, "Good ideas get early help from venture incubators", San Jose Mercury News, Feb 19, 1996, on Web at http://www.sjmercury.comlbusinesslventure!vent219.htm
12
23. "Venture Capitalists Pour over $2.2 Billion into Firms", Wall Street Journal, May 2, 1996, quoting study by Coopers and Lybrand, pB2
24. Source: Securities Data Corp, quoted in Fortune, May 13, 1996, p187
25. Source: National Venture Capital Association, quoted by M Selz, Wall Street Journal, May 14, 1996, pB2
26. D Lohse, "Investors Try to Tally Value of Tech Issues", quoting B Lupatkin, Wall Street Journal, May 6, 1996
27. J Wyatt, "America's amazing IPO bonanza", Fortune, May 27, 1996, p76-80
28. C Morgan, quoted by J Burke, "Getting to an IPO", The Red Herring, January 1996
29. DT Gleba, "No End in Sight for the Info Technology Boom", Upside, May 1966, p82
30. S Tully, "How to make $400,000,000 in just one minute ... ", Fortune, May 27, 1996, p85-92
31. R Braham, "The Air Force R&D Balancing Act", IEEE Spectrum, March 1996
32. HL Poppell, M Toole, "The Bleeding Edge of Information Technology", The Red Herring, August 1995, on the Web at http://www.herring.com/mag/issue22/bleeding.html
33. DT Campbell, JC Stanley, Experimental and Quasi-Experimental Design for Research, Rand­ McNally, Chicago, 1963. [When Campbell died in May 1996 at age 79, The New York Times obit misdated his classic work]
34. S Moore, D Stansel, "How Corporate Welfare Won: Clinton and Congress Retreat from Cutting Business Subsidies", CATO Policy Analysis No 254, May 15, 1996, available from The CATO Institute, 1000 Massachusetts Ave, Washington, DC 20001 or on the Web at http://www.cato.org
35. McDonald S, "Theoretically sound: practically useless? Government grants for industrial R&D in Australia", Research Policy, 15, p269-283, 1986
36. S Brittan, Capitalism with a Human Face, Edward Elgar, London, 1993
37. D Takahashi, "But shed no tears: Dividend boosted", San Jose Mercury News, May 18, 1996
38. G Gilder, quoted by P Bronson, "George Gilder", Wired, March 1966, p122
13
FUNCTIONALIZED POLYMERS FOR ELECTRO-OPTIC MODULATION THROUGH GRATING-INDUCED RESONANT EXCITATION OF GUIDED MODES
G. Blau,l G. Vitrant,l P-A. Chollet,2 P. Raimond,2 and F. Kajzar2
ILEMO/ENSERG BP 257 38016 Grenoble Cedex (France) 2CEA (LET!-Technologies Avancees) DEIN/SPE, 91191 GifCedex (France)
INTRODUCTION
Most optical modulators are built in the channel waveguide configuration. This is so for the Mach-Zehnder interferometer and of the directionnal couplers where the outcoming guided wave can be monitored by applying an external electric field. All these devices imply sophisticated technologies. On another hand it is also interesting to build devices controling laser beam propagation in the free space by means of a corrugated grating (Reinisch 1985). The intensities diffracted in the various orders by a grating made with an electro-optical polymer can be modulated with an applied electric field (Shi 1994). Such devices can find applications in computers for board to board information transmission. We report in this paper a modulator where both the transmitted and reflected intensities of a laser beam can be modulated. The electro-optic material is paranitroaniline (PNA) grafted on a PMMA polymer backbone. A laser beam is coupled into the planar film by means of a grating. When an electric field is applied, the effective index of the guided mode varies which results in a change of the coupling conditions and subsequently in the intensities of the .transmitted and reflected beams. The modulation yield is 40% when a 40 V external voltage is applied.
ELECTRO-OPTIC POLYMER
The functionalized nitroaniline was synthesized by substitution of 4-fluoro-1- nitrobenzene with N-(hydroxyethyl) ethanolamine in dry DMSO in presence of potassium carbonate. The product was purified by column chromatography and then esterified by methacrylic acid with DCCI/DMAP in methylene chloride. The ester was purified by column chromatography.
Applications of Photonic Technalogy 2, Edited by G.A. Lampropoulos and R.A. Lessard, Plenwn Press, New York, 1997 15
HO HO
t.
HO
n N fj ' NO ---- _) - 1 AIBN
OMF t. 1o•c
N-0-' NO _) - l
Figure 1. Functionalization of the NLO chromophore and grafting to the PMMA chain
The polymers were synthesized by copolymerisation under vacuum in DMF solution of freshly distilled methyl methacrylate and the methacrylate of the dye. The initiator was AIBN.
BASIC PRINCIPLES AND DEVICE MODELING
The modulator presented here is based on the dependence of the coupling angle with the electric field. The resonance condition implies the equality of the in-plane component of the incident wave vector and that of the guided mode which is related to the effective index of the mode (figure 2):
TM 27t · k0Re(neff) +I d = k0 sm0 (1)
ko is the wave vector of the incident beam, n~lf the effective index of the guided mode (TM-polarized, see below), d the spatial period of the grating and I an integer. The grating adds a mutiple of its reciprocal vector to the in plane component of k. In fact a complete analysis of the transmitted or reflected intensities can be rigorously calculated within the assumption of a plane incident beam (Vincent 1980). A cross section of the device is shown in figure 1. The electro-optical film is sandwiched between two buffer layers (one corrugated by the grating) and two transparent electrodes. The polarisation is choosen in the incident plane (TM) in order to take advantage of the highest component r33 of the Pockels tensor (the normal direction to the film plane is labeled 3).
16
y
modulated transmitted beam
Figure 2. Structure of the modulator with the corrugated grating, buffer layers, transparent electrodes and nonlinear polymer.
When Eqn(1) is fullfilled, there is a dip in the transmitted intensity and an oscillation of the reflected intensity. The narrower the resonance, the larger the device sensitivity, which is the case when the imaginary parts of the refractive indices are minimized. The rigorous model enables the optimization of the parameters that we control. Some parameters cannot be varied because of experimental requirements. This is the case of the refractive indices of the different layers, which are listed in Table 1.
Table 1. Thickness and refractive index of the different layers
thickness material refractive index@ 633 nm
top electrode 50nm ITO 2.022 + 0.055 i top buffer layer d (variable) MgF2 1.36 guiding film 900nm Polymer no= 1.610 bottom buffer layer d (variable) MgF2 ne = 1.655 bottom electrode 50nm ITO 2.022 + 0.055 i substrate lmm Si02 1.457
In Table 2 we report the results concerning the resonance (contrast and width) for different thickness d of the buffer layers, using the other parameters listed in Table 1. The thickness of the buffer layers must be large enough in order that the electromagnetic wave vanishes at the metal interface (otherwise the resonance is widened). One can see that the device response reaches a maximum for a buffer layer thickness equal to 300 nm. Larger values will be at the detriment of the applied field in the polymer film. The electrode material has a great importance in the device performances. There conductivity should as large as possible (to increase the bandwidth of the device) and there optical absorption as low as possible. Even though they are isolated from the guided wave by they are isolated from the guided wave by the buffer layers, if they absorb at the laser frequency, they contribute to increase the resonance width since the transmitted and reflected beams propagate through them. The modelization enables the determination of the optimum grating structure (period: 700 nm, grove depth: 70 nm, width of the groves: 350 nm).
17
Table 2. Modulator performances calculated with the parameters listed in Table 1.
Buffer layer Effective index of the angular reso- transmission reflection thickness d (nm) guided mode nanceFWHM contrast contrast
150 1.628 + 7.6x 1 0-5i 0.0128° 33% 13% 200 1.628 + 5.0x10-5i 0.0084° 48% 18% 300 1.627 + 3.9x IQ-5i 0.0065° 81% 74% 300 * 1.628 + 3.6xl0-5i 0.0061° 91% 93%
•assuming no optical losses in the electrodes
The optical response of the device is connected to the Pockets coefficient through the extraordinary refractive index variation :
A = r33n] E o.n, 2 3 (2)
The value of r 33 at 632.8 nm (12 prnN) have been determined previously in corona poled PNAIPMMA films (Gadret, 1993) and yields to a modulation of 81% when a 10 V voltage is applied (11ne = 4.5x10-4).
DEVICE FABRICATION
The first step is to etch the grating on the silica substrate. This is achieved by the following procedure: a poly(methyl-methacrylate) film (PMMA) is spin-coated on the silica substrate, then electron beam patterned and the silica is etched by neutral ion beam. The size of the grating is 5mmx5mm. The grove profile is etched according to the model, that is quasi-rectangular with period 700 nm and depth 70 nm. The grove width is 350 nm (half the period). The buffer layers could be a spin coated polymeric film; but such materials have two disadvantages: firstly in almost all the cases, they act as a planarization layer, decreasing the diffraction yield of the grating and secondly the solvent may dissolve the electro-optical film. This is the reason why we have choosen chiolite (NasAl3F 14) which is an extremely porous material with low refractive index (1.33) and which can be deposited by cold thermal sublimation.
As stated before the electrical and optical properties of the electrodes have a strong influence on the device performances. Tin doped indium oxyde (ITO) possesses the necessary requirements, but its deposition heats the sample (150°C) resulting in a disorientation of the chromophores and consequently the disapearance of the electro­ optical properties of the polymer film. This is why we have prefered Sn02 electrodes deposited by low temperature reactive sputtering (Goodchild 1985) and having an electric conductivity cr = 1.2x102 n·1cm·1 and an optical transmission T = 80% for a 80 nm thickness.
The electro-optical polymer film was spin coated in a clean room in order to minimize the risk of short-circuit when applying a DC voltage to the electrodes. The chromophore orientation was achieved by electrode poling: the temperature was raised to TP = 93°C which is 5°C below the glass transition temperature T g = 98°C.
18
4
- greti.ng W5mm) - transparent SnOx electrodcs
12 11 10
Figure 3. Pateming of the electrodes for increasing the bandwidth
A 250 V voltage was applied to the electrodes and maintains during cooling. It corresponds to a poling field equal to 2.1 MV /em if we assume it is homogeneous within all the sample thickness. Then the upper buffer layer and electrode were deposited. In order to increase the electrical cutoff frequency of the device, the time constant RC (R being the electrode resistance and C the device capacitance) should be reduced. In that purpose we have patterned the electrodes as shown in Figure 3. Owing to the fact that the Sn02
electrode cannot be etched by acids or bases the patterning process was achieved by lift off.
EXPERIMENTAL RESULTS
Electric field response
The device performances were measured in transmission and reflection at a wave­ length of 632.8 nm wavelength. The sample was mounted on a goniometer in order to determine precisely the angular resonance. The incident polarized laser beam was weakly focused (f= 120 em) on the grating. The transmitted and reflected beams were measured by CCD cameras. The results are shown in Figure 4. The angular resonance width is O.l2°and the contrast 40%.
An applied voltage amplitude of 74 V (-37 to 37 V) results in an angular shift equal to 0.23° which corresponds to a Pockels coefficient equal to 4.3 pm/V, which is 2.8 times smaller than the one determined previously on the same kind of films (Gadret 1992). We have assumed that the electric field had the same value in the organic film and chiolithe buffer layers. In addition to account for the value of the contrast, we have to introduce an imaginary component (0.08) to the refractive index of the electrodes. There are two possible explanations for the low value of the Pockels coefficient:
i. During the deposition of the upper buffer layers and electrodes there is an increase of the temperature of the polymer film, with the consequence of a disorientation of the chromophores.
19
-- re~ectlvlty (+37V) ~ re~ectivity (-37V) -o-- transmlssMty (-37V) ~- transmlssMty (+37V)
o~--~--~---~---~------~---J 82.0 82.5 83.0
incident angle [degrees]
Figure 4. Shift of the transmitted and reflected beam resonance with the appllied voltage. Upper curves correspond to the transmitted beam and lower curve~ to the reflected beam
ii. The electric conductivity is larger in the polymer layer than in the buffer layers, and consequently most of the voltage is applied to the buffer layers and not to the electro­ optic film.
Frequency response
As pointed out previously, there is a bandwidth limitation of device response, due to the time constant introduced by the conjugated action of electrode resistance and modulator capacitance. It is desirable to reduce the latter by reducing the surface to the laser beam cross section. This is why we have patterned the electrodes as shown in Figure 3. This reduction of the surface results in an increase of the SdB bandwidth from 100 Hz for a 5mmx5mm modulator to 80 kHz for a 1 mmxO.Smm modulator. The frequency response of these two devices is shown in Figure 5.
CONCLUSION
We were able to demonstrate an electro-optic modulator based on polymer materials using a diffraction grating. We measured a light modulation of 40% with a change in the applied voltage of AV = 60 V and an electro-optic coefficient ofr33 = 4.3 pm!V. This modulator was based on paranitroaniline grafted on a PMMA matrix as non linear waveguiding material. We obtained after numerical simulation an optimized grating depth of 70nm corresponding to a coupling length of lc = 0.6lmm. As a result, the performance of the modulator is not affected by linear losses smaller or equal to 35dB/cm. The requirements concerning wave guide homogeneity are less stringent than in the case of a waveguide with an interaction length which is usually~ 1 em. An additional advantage is given by the fact, that temporal stability of the polymer does not affect the operating capabilities of the device. A slight change in the value of the refractive indices can be easily compensated changing the angle of incidence Bi.
20
~ -10
Figure 5. Frequency response of two electro-optic modulators with different electrode dimensions.
REFERENCES
Gadret 0.,1993, Proprietes optiques non lineaires des polymeres polarisees, PhD Thesis, Universite de PARIS XI (ORSA Y).
Goodchild, R.G., Webb, J.B., Williams, D.F., 1985, Electrical properties of highly conducting and transparent thin films of magnetron sputtered Sn02, J. Appl. Phys. 57:2308.
Reinisch, R., Vincent, M., Neviere, M., Pic, E., 1985, Fast Pockels light modulator using guided wave resonance, Appl. Opt. 24:2001.
Shi, Y., Bechtel, J.H., Kalluri, S., Steier, W.H., Xu, C., Wu, B., Dalton, L.R., 1994, Measurement of the electro-optic coefficient dispersion in poled polymer materials, Proc. SPIE 2285:131.
Vincent P., 1980 in: Electromagnetic theory of gratings, R. Petit, ed., Springer Verlag, Berlin.
21
ABSTRACT
J. N. McMullin, D. W. Boertjes, M. Krishnaswamy, and B. P. Keyworth
Telecommunications Research Laboratories (TRLabs) 800 Park Plaza, 10611-98 Avenue, Edmonton, AB, T5K 2P7 http://www. trlabs.ca
A polymer waveguide of parabolic cross-section is shown to produce a lateral graded effective index for each of the discrete vertical modes of the polymer layer due to lateral variation of the waveguide height. Beam propagation simulations are in excellent agreement with recorded images of oscillating guided beams in a waveguide fabricated on glass. Results are also shown for a waveguide fabricated on an erbium/ytterbium co-doped bulk glass substrate in which a transmitted 1550 nm signal showed enhancement of up to 1. 7 dB with 210 mW of975 nm pump light in a 1.0 em long guide.
INTRODUCTION
Currently, much attention is being given to polymer waveguides for the purposes of optical interconnects and optoelectronic circuits. 1 Many methods may be used to fabricate these guides including photolithography,2 laser writing,3 and selective exposure of photorefractive materials.4 The waveguides discussed here were fabricated by direct dispensing of a UV curable optical adhesive. 5
Another area of intense research is in planar waveguide optical amplifiers. Such amplifiers are expected to play an important role in the development of optical, and hybrid optoelectronic integrated circuits. Erbium doped waveguide amplifiers operating in the 1.55 j.t.m wavelength region are of particular interest to the field of telecommunications. 6 Recent results have shown net gain in amplifiers using Er-doped glass as the core of the
Applications of Photonic Technology 2, Edited by G.A. Lampropoulos and R.A. Lessard, Plenum Press, New York, 1997 23
waveguide? This is perhaps the optimal configuration; however, it requires that the Er­ doped glass be processed and in a way that maintains the stoichiometry crucial to the device performance.8 Another configuration involves depositing the waveguide core material on bulk Er-doped glass. The gain is then produced in the evanescent tails of the propagating mode.9
In this paper we examine the properties of polymer waveguides with parabolic cross­ sections. The shape of the guide produces a graded index behaviour, which is explained in terms of the vertical slab modes' effective indices. A computer simulation is used to show that the effect is well understood. The application of these guides to an evanescently pumped waveguide amplifier is also demonstrated.
FABRICATION AND MEASUREMENT
The waveguides were fabricated by dispensing a UV curable optical adhesive (Norland NOA81) through a syringe onto a substrate under computer control. The surface of the liquid forms a circular arc, which, in the case of this material, has a low height to width ratio. The low aspect ratio means that the surface fits well to a parabola, which can be used for the following analysis. Fig. 1 shows the measured surface profile of a typical guide superimposed with a parabolic curve fit.
e ~ .i: Cl 'iii J: Ql "tl 3 (!)
4
3
2
\ 300 400
Figure 1. Measured cross-section of a typical parabolic waveguide (dots) and a parabolic fit to the data (line).
Figure 2. Experimental setup. A three axis positioner is used to align the fibre to the input facet of the guide. A CCD camera was used to capture an image of the light scattered out of the top of the guide.
24
a
b
The experiment was carried out by launching 672 nm light at the input facet of the guide and using a camera to capture an image of the scattered light as shown in Fig. 2. Fig. 3(a) shows an image of light oscillating horizontally in a waveguide 340 jlm wide and 3.6 jlm high. The light was launched off-axis by approximately 110 jlm. The polymer waveguide core bad a refractive index np=l.56 and the glass substrate had an index n,=1.45. Since the waveguide is multimoded in the vertical direction, the light splits into different vertical modes. Each vertical mode has a different effective index profile, thus each oscillates with a different period. The lower order modes have longer oscillation periods. The result of a beam propagation simulation with no free parameters is shown in Fig 3(b) and is in excellent agreement with the experiment.
\ .. , .................... ,.,. ,........... .... . - ... -· ... ·· · t->;; 4""' .. . ·- ·-· ... ,...._....,... . ...~..,. . . . . .... ... . .. ...:· ~~ ·: ...... .. :· .... , ...... . :. . . -. . · ..... .. ,..... ............. ._.,-~. . . ····-- ·- ,~ .......... ,., .
~•~----1mm----,•~
Figure 3. (a) Image taken of the experiment. The light was launched approximately 110 Jlm off axis. (b) Computer simulation output shown on the same scale. This shows the dependence of the oscillation length on the mode number.
To test the evanescent waveguide amplifier configuration, the polymer was dispensed on a commercially available erbium/ytterbium co-doped glass (K.igre QX!Er) having a refractive index of 1.53. The waveguide had a height of 8.0 jlm and a width of 180 jlm. The glass was then scribed and cleaved to produce input and output facets resulting in a guide 1.0 em in length. Using the analysis to be presented in the following section, the mode shape in the guide can be calculated for both wavelengths of interest, i.e. 1550 nm (signal) and 980 nm (pump). From this, the percentage of the signal and pump modes in the substrate was calculated to be 0.74% and 0.21% respectively. The overlap between these modes in the substrate was found to be very high (97%) since the form of both modes in this region is a decaying exponential in the y-direction. The light was launched into the waveguide from a fibre which was single moded at both 1550 and 980 nm. The coupling Joss, determined by the overlap of the waveguide and fibre modes, was estimated to be 3 dB.
The device was pumped using a tunable Ti:Sapphire laser. A multimode fibre was used to collect the light at the waveguide output and direct it through a monochrometer before detection. The spontaneous emission spectrum with 210 mW of pump light is shown in Fig. 4. The fluorescence lifetime was measured to be 8 ms under the same conditions.
In order to separate the absorption loss due to the erbium from the insertion and propagation losses, the fibre to fibre loss was measured at three wavelengths. The losses were found to be 5.8, 4.2, and 4.5 dB at 1550, 1300, and 980 nm respectively. Neither the erbium nor the ytterbium has an absorption near 1300 nm, therefore 4.2 dB of loss can be attributed to coupling and propagation losses in the polymer waveguide. This leaves 1.6 and 0.3 dB of erbium induced Joss at 1550 and 980 nm respectively.
A 1550 nm laser diode was modulated at 2 kHz and used as a signal to study the
25
stimulated emission characteristics of the waveguide. Fig. 5 shows the relationship between the transmitted signal enhancement and the residual pump power. The attenuation of the signal was decreased by up to 1.7 dB when pumped with an estimated 210 mW of light launched from