THE SOCIETY FOR PHOTONICS · fibers, Prof. Phillip Russell of Max-Planck Research Group for Optics Information and Photonics and the University of Erlangen-Nuremberg has contributed

October 2007Vol. 21, No. 5www.i-LEOS.org IEEE

THE SOCIETY FOR PHOTONICS

NEWS

Photonic Crystal Fibers:A Historical Account

Photonic Crystal Fibers:A Historical Account

Margin Measurements

in Optical Amplifier Systems

Margin Measurements

in Optical Amplifier Systems

21leos05.qxd 10/25/07 10:28 AM Page cov1

21leos05.qxd 10/25/07 10:28 AM Page cov2

October 2007 IEEE LEOS NEWSLETTER 1

IEEE

THE SOCIETY FOR PHOTONICS

NEWS

Page 16, Figure 1:

Color photograph of two-dimensionalPhCs on a silicon-on-insulator substratefabricated by x-ray lithography. October 2007 Volume 21, Number 5

COLUMNS

Editor’s Column………………2 President’s Column………………….3

21

DEPARTMENTS

News . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27• LEOS 2007 Graduate Student Fellowship Winners• LEOS 2007 Chapter Award Winners

Careers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28• 2007 IEEE Daniel E. Noble Award recipients:

Stephen R. Forrest, Richard H. Friend, and Ching Tang

Membership . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29• Benefits of IEEE Senior Membership• New Senior Members• LEOS Goes to Washington, DC• Chapter Highlight: New LEOS Chapter formed in Western Australia

Conferences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32• 19th Annual Workshop on Interconnections Within High Speed

Digital Systems• 2007 CLEO/Pac Rim

Publications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33• Call for Papers:

IEEE Journal of Selected Topics in Quantum Electronics (JSTQE)IEEE/OSA Journal of Display Technology on Medical Displays (JDT)

17

12

FEATURES

Special 30th Anniversary Feature: . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4• Most Cited Article from IEEE PTL “Margin Measurements in Optical Amplifier Systems,”

by Neal S. Bergano et al• “Margin Measurements in Optical Amplifier Systems,” Commentary by Neal S. Bergano• “Photonic Crystal Fibers: A Historical Account” by Philip Russell

Industry Research Highlights: . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16• “Adiabatic Tuning of Optical Microsystems,” by Masaya Notomi

University Research Highlights: . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22• “High Intensity Fiber Lasers: Emerging New Applications and New Fiber Technologies,” by K. C. Hou et al

Column by LEOS Leaders: . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10• “Dream Comes True”, by LEOS Past President - Tetsuhiko Ikegami

21leos05.qxd 10/25/07 10:29 AM Page 1

Welcome to the October Newsletter!This month, we conclude our special 30th anniversary

series of reprints of the most cited articles in LEOS jour-nals. Neal Bergano of Tyco Telecommunications haswritten a nice commentary about his 1993 article onMargin Measurements in Optical Amplifier Systems,which is the most cited Photonics Technology Letter(228 citations). In keeping with the theme of opticalfibers, Prof. Phillip Russell of Max-Planck ResearchGroup for Optics Information and Photonics and theUniversity of Erlangen-Nuremberg has contributed areview of photonic crystal fibers, which he was instru-mental in developing. Prof. Almantas Galvanauskaspresents a nice overview of ongoing work in fibers andother areas in his group at the University of Michigan.Dr. Masaya Notomi of NTT Basic Research Laboratoriesin Japan presents an Industry Research Highlights arti-cle on Adiabatic Tuning of Optical Microsystems.Finally, we continue our series of columns by LEOSLeaders with a commentary by 1994 LEOS President Dr.Tetsu Ikegami.

The LEOS Annual Meeting takes place this month inLake Buena Vista, Florida, USA, where our communityshares the latest results in photonics. Watch for high-lights from the meeting in the December Newsletter. Iplan to attend, and hope to meet many of you there.

As always, please feel free to send any comments andsuggestions to [email protected]. I have yet tohear from many of you, and would love to get morefeedback!

Krishnan Parameswaran

Editor’sColumnKRISHNAN PARAMESWARAN

PresidentAlan WillnerUniversity of Southern CaliforniaDept. of EE-Systems/ Rm EEB 538Los Angeles, CA 90089-2565Tel: +1 213 740 4664Fax: +1 213 740 8729Email: [email protected]

President-ElectJohn H. MarshIntense Photonics, Ltd.4 Stanley BoulevardHamilton International Tech ParkHigh Blantyre G72 0UX Scotland, UKTel: +44 1698 827 000Fax: +44 1698 827 262Email: [email protected]

Secretary-TreasurerFilbert BartoliLehigh University19 West Memorial DrivePackard Lab 302Bethlehem, PA 18015Tel: +1 610 758 4069Fax: +1 610 758 6279Email: [email protected];[email protected]

Past PresidentH. Scott HintonUtah State UniversityDean of Engineering4100 Old Main HillLogan, UT 84322-4100Tel: +1 435 797 2776Fax: +1 435 797 2769Email: [email protected]

Executive DirectorRichard LinkeIEEE/LEOS445 Hoes LanePiscataway, NJ 08855-1331Tel: +1 732 562 3891Fax: +1 732 562 8434Email: [email protected]

Board of GovernorsM. Amann K. HotateF. Bartoli D. HuffakerK. Choquette H. KuwaharaC. Doerr C. MenoniS. Donati D. PlantC. Gmachl A. Seeds

Vice PresidentsConferences – E. GolovchenkoFinance & Administration – S. NewtonMembership & Regional ActivitiesAmericas – S. UnluAsia & Pacific – C. JagadishEurope, Mid-East, Africa – J. BuusPublications – C. MenoniTechnical Affairs – N. Jokerst

Newsletter Staff

Executive EditorKrishnan R. Parameswaran Physical Sciences, Inc.20 New England Business CenterAndover, MA 01810Tel: +1 978 738 8187Email: [email protected]

Associate Editor of Asia & PacificHon TsangDept. of Electronic EngineeringThe Chinese University of Hong KongShatin, Hong KongTel: +852 260 98254Fax: +852 260 35558Email: [email protected]

Associate Editor of CanadaAmr HelmyThe Edward S. Rogers, Sr. Department of Electrical and Computer EngineeringUniversity of Toronto10 King’s College RoadToronto, Ontario, Canada M5S 3G4Tel: +1 416-946-0199Fax: +1 416-971-3020Email: [email protected]

Associate Editor of Europe/MidEast/AfricaKevin A. WilliamsEindhoven University of TechnologyInter-University Research Institute COBRA on Communication TechnologyDepartment of Electrical EngineeringPO Box 5135600 MB Eindhoven, The NetherlandsEmail: [email protected]

Staff EditorGiselle BlandinIEEE/LEOS445 Hoes LanePiscataway, NJ 08855-1331Tel: +1 732 981 3405Fax: +1 732 562 8434Email: [email protected]

IEEE Lasers and Electro-Optics Society

LEOS Newsletter is published bimonthly by the Lasers and Electro-Optics Society of the Institute of Electrical and Electronics Engineers,Inc., Corporate Office: 3 Park Avenue, 17th Floor, New York, NY10017-2394. Printed in the USA. One dollar per member per year isincluded in the Society fee for each member of the Lasers andElectro-Optics Society. Periodicals postage paid at New York, NYand at additional mailing offices. Postmaster: Send addresschanges to LEOS Newsletter, IEEE, 445 Hoes Lane, Piscataway, NJ08854.

Copyright © 2007 by IEEE: Permission to copy without fee all or partof any material without a copyright notice is granted provided thatthe copies are not made or distributed for direct commercialadvantage, and the title of the publication and its date appear oneach copy. To copy material with a copyright notice requires spe-cific permission. Please direct all inquiries or requests to IEEECopyrights Office.

21leos05.qxd 10/25/07 10:29 AM Page 2


President’sColumnALAN E. WILLNER

“LEOS 2017”“It is not the strongest of the species that survives, nor themost intelligent, but the one most responsive to change.”Charles Darwin.

I recently asked my students as to what topic they would wantto read about in my LEOS Column. One bright individual wascurious as to my vision concerning what he could expect LEOS tobe like in the future, say in the year 2017.

“Change is inevitable - except from a vending machine.”Robert C. Gallagher, Author.

Of course, it is fanciful that I could predict anything that mightoccur in 10 years. Therefore, I’ll try to mix together some of the fol-lowing ingredients:

(i) aspects that I think might be different due to cultural andtechnological changes.

(ii) aspects that I think might remain the same given thatthey relate to our core value proposition.

(iii) aspects that I “hope” will be with us in 2017.

Please read below with a skeptical mind,and please do NOT show this article to me in10 years.

Aspects That Might be Different“He who rejects change is the architect ofdecay. The only human institutionwhich rejects progress is the cemetery.”Harold Wilson, former Prime Ministerof the U.K.

Change brings new and potentially excitingopportunities to nearly all aspects of LEOS.Moreover, as a warning from Gen. Eric Shinsekiof the U.S. Army, “If you don't like change,you're going to like irrelevance even less.”

(1) Publications: The Internet will cer-tainly have an impact on our publications.Will we have true “open access,” in whicharticles will be available on-line at no costto the reader? Perhaps. Will print stillexist? Probably in a more limited form.However, there is unmistakable valuefrom LEOS-enabled peer-review and edi-tor-monitored revisions that adds signifi-cant quality to the publications process.Therefore, our business model will adaptso that LEOS publications flourish in thenew “virtual” reality. In fact, we probably

have a great opportunity to use the Internet for creating a pub-lications model that is of maximum value to our stakeholders.Will the reader be able to electronically customize their mul-timedia reading experience and provide feedback to authors? Ihope so.

(2) Committee Meetings: Many committees, especially con-ference committees, will convene by Internet-enabled videoconferencing. This will save an enormous amount of time andmoney, and it eliminates the problem of obtaining visas fortravel. Moreover, this will effectively remove any barriers forpeople to accept committee memberships from anywhere onthe planet. As generally pointed out by Thomas Friedman,the LEOS world will truly be “flat.”

(3) Employment: Since our field is relatively young and the com-mercial adoption of our technologies is fairly recent, the employ-ment scenario for our industry has experienced dramatic and rapidchanges. We have emerged from one of the worst elimination ofjobs in modern history, and yet the vast majority of the uprootedpeople are now gainfully employed. In 10 years, the basic marketvalue of our industry will become more fully entrenched, and the

(continued on page 26)

21leos05.qxd 10/25/07 10:29 AM Page 3

AbstractThe margin, or the difference between the received signal-to-noiseratio (SNR) and the SNR required to maintain a given bit errorratio (BER), is important to the design and operation of opticalamplifier transmission systems. A new technique is described forestimating the SNR at the receiver’s decision circuit when the BERis too low to be measured in a reasonable time. The SNR is deter-mined from the behavior of the BER as a function of the decisionthreshold setting in the region where the BER is measurable. Weobtain good agreement between the BER predicted using themeasured SNR value and the actual measured BER.

IntroductionThe bit error ratio (BER) in an optical amplifier transmission sys-tem is set by the electrical signal-to-noise ratio (SNR) of the datasignal at the decision circuit. Therefore, SNR is a natural figure ofmerit for transmission performance and system health. The marginin an optical amplifier system is the decibel difference between thereceived SNR and the SNR required to maintain the system errorrate specification. The received SNR, and therefore the margin, isset by optical noise and waveform degradations accumulating overthe entire length of the system. In this letter, we propose a methodof measuring margin in an optical amplifier system, by estimatingthe SNR at the receiver’s decision circuit, when the BER is toosmall to be directly measured. We observe good correlationbetween measured and calculated BER for a 5 Gb/s NRZ signalthrough a 4500 km optical amplifier system.

SNR in Optical Amplifier SystemsIn an optical amplifier system, the SNR at the decision circuit isdegraded by optical noise, fiber chromatic dispersion, polarizationmode dispersion, and fiber nonlin-earities. While the exact proba-bility density function for optical noise is not exactly Gaussian [1],a Gaussian approximation can lead to close BER estimates [2]. TheQ factor [3] is the signal-to-noise ratio at the decision circuit involtage or current units, and is typically expressed by

Q = |μ1 − μ0|σ1 + σ0

(1)

where μ1,0 is the mean value of the marks/spaces rail, and σ1,0 isthe standard deviation. A voltage histogram down the center of theeye can be measured with a digital sampling oscilloscope to esti-mate Q. This technique fails, however, to give good correlationbetween the measurement of Q and the BER, since the variationseen around each rail represents a mix of pattern effects, such as

intersymbol interference (ISI) and noise. Such effects in turn artifi-cially broaden the estimates of σ1,0, thus giving erroneous results.In addition, this method operates on a limited set of bits (i.e., thedata arrives at 5 Gb/s, while the oscilloscope’s analog-to-digital con-verter samples at 10–100 kHz).

Alternatively, the voltage histograms can be made at a specificpoint in the pattern as opposed to the data eye. This eliminates thepattern effects from the measurements of σ1,0, but yields a poten-tially inaccurate measure of μ1,0. Also, this approach has the draw-back of recording even fewer bits than the measurement in the eye,and it is not practical in a real transmission system, where the databits arc random. Our new technique avoids these problems byusing the decision circuit to probe the rails of the eye, thus usingevery bit in the data stream. Moreover, it includes the ISI present inthe regenerator’s linear channel as well as that generated in the sys-tem from dispersion and fiber nonlinearity.

Measurement TechniqueThe Q factor is measured by recording the BER versus decisionlevel down the center of the eye (i.e., a fixed timing phase). Theequivalent mean and sigma of the marks and spaces are determinedby fitting this data to a Gaussian characteristic, given by [4]

BER(D) = 1

2

{erfc

( |μ1 − D|σ1

)+ erfc

( |μ1 − D|σ1

)}

(2)

where μ1,0 and σ1,0 arc the mean and standard deviation of themark and space data rails, D, is the decision level, and erfc (x) is aform of the complementary error function given by:

erfc(x) = 1√2π

∫ ∞

xe−β2/2dβ ≈ 1

x√

2πe−x2/2 (3)

where the approximation is nearly exact for x > 3. Here the μ1,0

and σ1,0 are not the physical values in the eye, rather they are equiv-alent values used to fit the data for the purpose of estimating Q.

The Q factor is calculated as follows: using the μ and σ of eachrail in a fashion similar to (1), the data are divided into two sets thathave the measured BER dominated by the marks rail and spacesrail. The raw data arc separated at the point of minimum error ratefor measurable BER’s, or at any value of D that yields error-free per-formance, for cases where the SNR is high. Each data set is fitted toan ideal curve, assuming Gaussian noise statistics, to obtain anequivalent mean and sigma for the positive and negative rail.Equation (2) naturally separates into errors dominated by markerrors and space errors. Once separated, the BER is a simple expres-sion given by a single 1

2 erfc (·) function. Each set of BER data ispassed through an inverse error function, and then a linear regres-sion is performed with the decision levels Di. The equivalent σ1,0

4 IEEE LEOS NEWSLETTER October 2007

Margin Measurements in Optical Amplifier SystemsReprint of most cited article from IEEE PTL Vol. 5, pg 304, 1993

Neal S. Bergano, F. W. Kerfoot, and C. R. Davidson

Special 30th Anniversary

MANUSCRIPT RECEIVED NOVEMBER 30, 1992; REVISED DECEMBER 16,1992. THE AUTHORS ARE WITH AT&T BELL LABORATORIES, HOLMDEL.

NJ 07733. IEEE LOG NUMBER 9207607.

21leos05.qxd 10/25/07 10:29 AM Page 4

A CODEC CHIP THAT ACHIEVED 50% COMPRESSION IN DEVELOPMENT TIME.

Accelerating the pace of engineering and science

THAT’S MODEL-BASED DESIGN.

To be first to market with a next-generation mixed-signalchip, the IC design team at Realtek used system models tocontinuously verify their design.The result: 50% less time to completion and a 50% marketshare. To learn more and torequest the IBS study on SimulinkROI in Electronic Design, visit mathworks.com/mbd

©2007 The MathWorks, Inc.

21leos05.qxd 10/25/07 10:29 AM Page 5


and μ1,0, are given by the slope and intercept of the linear regres-sions. For ease of computation, the inverse 1

2 erfc (·) function is per-formed by first taking the logarithm of the BER. Log (1

2 erfc (·)) isa smooth one-one function that can be inverted by many numericalmethods, or more simply by using a polynomial fit:

{log

(1

2erfc(·)

)}−1

(x) ≈ 1.192 − 0.6681x − 0.0162x2

(4)

where x is log (BER), and (4) is accurate to ±0.2% over the rangeof BER’s from 10−5 to 10−10. The optimum decision level Dopt isdetermined from μ1,0 and σ1,0 as the cross point for the twoGaussian probability density functions.1 The calculated BER isgiven by (2) evaluated at Dopt. and to a good approximation isgiven by erfc(Q).

ResultsThe BER was measured at 5 Gb/s using an NRZ transmitter regen-erator pair similar to that used in previous experiments [5]. Thetransmitter consists of a CW DFB laser source and a Mach-Zehnderintensity modulator, with a rise time of 50 ps. The regenerator usesan optical amplifier automatic gain control amplifier (AGO front-end, a 2 nm bandpass filter, a p-i-n diode O/E converter, phase-locked loop timing recovery, and a decision circuit with a variabledecision level. The 10−9 sensitivity of the transmitter/regeneratorpair is −36.4 dBm at 5 Gb/s using a 215 − 1 word. Measurementswere made in the back-to-back configuration and through a 4500km amplifier chain [5].

Fig. 1 shows typical measured data for the logarithm of the BERversus the decision threshold in the decision circuit, and the solidline shows the best fit of (2). In this measurement, the optical powerinto the receiver was about 3.5 dB over the 10−9 sensitivity point.The BER was measured over one second intervals and was consid-ered valid if at least five errors were recorded; thus, the minimummeasured BER was 10−9. The maximum BER measured wasabout 10−5. The Q factor was determined to be 8.5 in linear ratio.The down arrow shows the predicted optimum decision threshold,and the two vertical segments show the equivalent μi, for themarks and spaces rail.

The Q factor and BER were measured for different opticalpowers into the receiver. Changing the optical power into thereceiver in turn changed the SNR at the input to the optical-to-electrical converter (O/E), since the receiver had an opticalamplifier AGC front-end that maintained constant total powerinto the O/E. The circles in Fig. 2 show the measured BER ver-sus Q factor for the back-to-back operation, and the broken lineshows the ideal characteristic. Since the Q factor is estimatedfrom the BER as measured in the range of 10−5 to 10−9, thedata necessarily matches for a BER greater than 10−9. Theinteresting points are those with a BER less than HP9, wherethe data show a good match to the prediction down to BER’sas low as 10−13. The good match between measurements andcalculations with the 4500 km amplifier chain gives us confi-dence that the technique works in the presence of realistic noiseand waveform distortion. The BER’s were measured at theoptimum decision threshold predicted from the fitting algo-rithm. The good fit to the data in Fig. 2 also shows that theGaussian approximation predicts the proper decision thresholdfor measurable BER’s.

This measurement technique allows us to make accurate pre-dictions of the margin available in an optical amplifier system.For example, Fig. 3 shows the measured Q factor versus thereceived optical power (ROP) into the receiver and includes thedata shown in Fig. 2. Similar to electrical SNR, the Q factor inthe figure is expressed in decibels as Q (dB) = 20 log (Q). Foran ROP greater than 2 dB over the sensitivity point, it is imprac-tical to measure the BER; however, Q is easily measured and

Figure 1: Bit error ratio versus decision threshold for back-to-backoperation.

−2

−4

−6

−8

−10−0.4 −0.2 0.0 0.2 0.4

Decision Level (Volts)

Q = 8.5

μ1 μ0

Opt

imum

Dec

isio

nP

oint

Log(

BE

R)

Figure 2: Measured bit error ratio versus Q factor for back-to-back,and 4500 km operation.

Extrapolated

Ideal BERBack-to-Back4,500 km

−6

−8

−10

−12

−145.0 5.5 6.0 6.5 7.0 7.5 8.0

Measured Q Factor (Linear Ratio)

Mea

sure

d Lo

g(B

ER

)

Figure 3: Measured Q factor versus received optical power, back-to-back and 4500 km.

Back-to-Back

4,500 km

35

30

25

20

15

10−40

−5

−9

−15

−30 −20 −10 0Received Optical Power (dBm)

Log(

BE

R)

20*

Log

(Q)

18.5 dB

5.5 dB

21leos05.qxd 10/25/07 10:29 AM Page 6

21leos05.qxd 10/25/07 10:29 AM Page 7


shows continued increases with an ROP for the back-to-backcase, until the maximum input power of −6.7 dBm is reached.The margin or the difference in Q value for 10−9 BER operationand the maximum input power was 18.5 dB. If more transmit-ter power were available, the measured Q value would eventual-ly reach a constant value corresponding to the SNR of the trans-mitted signal. With the 4500 km amplifier chain, the Q factorreached a steady state value of 21.2 dB, for a margin of 5.5 dB.

ConclusionsWe have described a technique for measuring the signal-to-noise ratio at the decision circuit (or Q factor) of an opticalamplifier transmission system. The measured system bit errorratio is accurately predicted from the SNR measurement. Thetechnique can be used to adjust the decision point of theregenerator in the terminal adaptivcly, since the measurementpredicts the optimum threshold, and the results can beextended to select the optimum phase. This measurementshould apply equally well to loop experiments using NRZ orsoliton transmission.

AcknowledgmentWe thank J. M. Siprcss, P. K. Rungc, and W. O. Schlosser for theircontinued support and encouragement.

References[1] D. Marcuse, “Derivation of analytical expressions for the bit-

error probability in lightwave systems with optical ampli-fiers,” J. Lightwave Technol., vol. 8, Dec. 1990.

[2] P.A. Humblet and M. Azizoglu, “On the bit error rate oflightwave systems with optical amplifiers,” J. LightwaveTechnol., vol. 9, Nov. 1991.

[3] S.D. Personick, “Receiver design for digital fiber optic com-munications systems, I,” Bell Syst. Tech. J., vol. 52, no. 6,July–Aug. 1973.

[4] J.M. Wozencraft and I.M. Jacobs, Principles of CommunicationEngineering. New York: Wiley, 1965, pp. 77–84.

[5] N.S. Bergano et al, “9000 km, 5 Gb/s NRZ transmissionexperiment using 274 erbium-doped fiber-amplifiers,”Topical Meeting on Optical Amplifiers and TheirApplications, Sante Fe, NM, June 24–26, 1992. Post-Deadline paper PD11.

In the summer of 1989, we were given a golden opportu-nity to start working on Erbium-doped optical amplifiersas a technology replacement for electro-optic regeneratorsin our undersea cable transmission systems. Of course, intoday’s logic EDFAs in undersea cable systems makes per-fect sense; however, back then it was not obvious that thetechnical challenges associated with stringing hundredsof optical amplifiers together could be surmounted. Ourfirst priority was to have a feasibility demonstration (orexperimental sanity check) of a transoceanic length sys-tem with at least a single 2.5 Gb/s channel. Our solutionfor this experimental challenge was the circulating looptransmission technique, where a relatively short “amplifi-er chain” was used in combination with optical switchingto emulate long-distance propagation. In those first trans-mission demonstrations, we simply measured the bit errorrate of our data pattern as a function of transmission dis-tance. In 1991, our error free distance for a 2.5 Gb/s datasignal was about 20,000 km, which was interesting andcertainly demonstrated feasibility but did not answer thefundamental question of how much margin was reallyavailable at say 9,000 km (the transpacific distance).

What we needed was a methodology of describingtransmission performance and an experimental way toestimate it. Two points were obvious to us at that time;we wanted a measurement technique that was based on biterror rate and a methodology capturing an “SNR-like”definition. We wanted a measurement technique based onBER because system performance has always been speci-fied in terms of BER and we knew that it could be accu-

rately measured over a wide range. Forward ErrorCorrection (FEC) was not in common use in optical sys-tems at the time and customers wanted error-free or near-ly error-free performance (in the range of 10-15 BER), soBER was not directly measurable in the range where mar-gin was desired. We started with the concept of using thereceiver’s decision circuit to “probe” the received data bychanging the decision point to cause measurable BER.Our experience with electro-optic regenerators pointed usto work by Stuart Personick on optical receivers. Hedescribed error rates in terms of the “…number of noisestandard deviations between signal and threshold at thereceiver’s input…” and simply labeled this parameter“Q”. Not wanting to pick a different letter, we simplyused the letter “Q” and then named it “Q-Factor”. The Q-factor describes the performance of the entire transmis-sion path from the optical transmitter, through the opti-cal path, and into the receiver. The Q-factor is expressedas a dimensionless ratio denoted by a lower case “q”.Often it is more convenient to express Q-factor in dBs as20log(q). We used the factor of 20 (or 10log(q2)) to main-tain consistency with the linear noise accumulationmodel. For example, a 3 dB increase in the average launchpower in all of the spans results in a 3 dB increase in Q-factor (an ideal simplification, but useful nonetheless).

By the summer of 1992, we were up and running withour decision circuit method of estimating margin on along optical amplifier testbed, and were confident thatthe Q-factor methodology was correct. At this point, wedebated the merits of keeping the measurement technique

Commentary by Neal S. Bergano

Margin Measurements in Optical Amplifier Systems

21leos05.qxd 10/25/07 10:29 AM Page 8

From Imagination to Market

IEEE Information Driving Innovation

Up-to-date, Relevant Information Driving the Bottom Line

Fueling Imagination

“Saving time in any way you can is critical. Accessto IEEE articles and papers is key in this regard.”

– Jon Candelaria, Project Manager, Motorola

Free Trial!Experience IEEE – request a trial for your company.

www.ieee.org/innovate

Today’s top companies rely on IEEE information to drive new research and generate cutting-edge patents.

Periodicals and conference proceedings that defi ne the future of innovation

Over 1.5 million documents in the IEEE Xplore®

digital library Top cited journals in the fi eld

07-PIM-0069b_Generic.indd 1 6/6/07 11:42:48 AM

21leos05.qxd 10/25/07 10:29 AM Page 9


In 1994, I was appointed the first President of LEOS withoutNorth American citizenship. It was the Biggest Time for opticaltransmission technologies and LEOS was really the top profession-al society in the world. I gave a plenary talk at OFC 1993 entitled“Network Toward Tera Bits Realm”, which sounds funny nowsince FTTH carries more than 100 Mb/s and undersea cable withcapacity beyond Terabits/s is in service. At the time, the questionwas “When would the marvelous enabling technology be imple-mented”. We encouraged ourselves by saying “if we build it, theywill come” referring to the film, “Field of Dreams” with increduli-ty and with the expectation of video service. As we now know, thedream came true, but from the Internet, not Video.

The key players at that time were Bell Systems, NTT, BritishTelecom, and French Telecom, and every conference was a bat-tlefield for champion records in the bit rate race. Now these play-ers have left the stage. “Best effort by IP” has become commonservice around the world following broadband video service.Definitely, present IT service could not have been realized with-out our crazy effort, but I am homesick for the good old times.

During my presidency, “Put our Members, First” wentinto Europe and Asian professional communities, and LEOSand OFC have succeeded in making our society truly inter-national today. In fact, LEOS lost half million dollars inEurope at that time but the return on that investment hasfortunately been great. Innovation is a key word in our soci-ety and Electro-Optics is not only an Enabler but also aDriver for it.

I am now in charge of Software development and appreci-ate the realization of convenient access through the efforts of

LEOS members. Ironically, I am also in charge of Satellitedevelopment in Japan, and I bear a grudge against the out-come that is killing satellite communication service.

As I look back, I sincerely thank my old colleagues for shar-ing the Biggest Time with me and hope that young chal-lengers can have new dreams for the future.

Biography: Tetsuhiko IkegamiTetauhiko (Tetsu) Ikegami was appointed Commissioner,Space Activities Commission, Ministry of Education, Culture,Sports, Science and Technology (MEXT) in January 2007 andserves as Senior Advisor to the National Institute of AdvancedScience and Technology (AIST) under METI.

He received his Doctor of Engineering in Electronics fromTokyo Institute of Technology in 1968. He joined NipponTelegraph & Telephone Co. and was appointed SeniorExecutive Manager of Optoelectronics Labs (1991), BasicResearch Labs (1994), board member of NTT (1993), andPresident and CEO of NTT Advanced Technology Co.(1996). He has performed significant work in optical-fibercommunication technology R&D and was elected Presidentof IEEE LEOS (1994).

In 1998, he moved to university and was elected Presidentof the University of Aizu having special features in ComputerScience & Engineering (2001-2005).

He has been serving as member of Government Committeesfor implementation of Science & Technology Policy.

He is Life Fellow of IEEE and a Fellow of the OpticalSociety of America and IEICE and JSAP in Japan.

Dream Comes TrueTetsuhiko Ikegami

Column by LEOS Leaders

a company trade secret, patenting the technique, or sim-ply publishing the results hoping to get buy-in from thefiber optic community. Luckily, we chose the later andsubmitted the manuscript to PTL late in 1992. To ourdelight, the fiber optics industry embraced the techniqueand methodology, and now Q-factor is the standard lan-guage for capturing and specifying lightwave transmis-sion performance.

Biography: Neal S. BerganoNeal S. Bergano is Managing Director of System Research& Network Development at Tyco Telecommunications. In1981, he received a BS degree in Electrical Engineeringfrom the Polytechnic Institute of New York and in 1983,

he received an MS degree in Electrical Engineering andComputer Science from the Massachusetts Institute ofTechnology. In 1981, he joined the technical staff of BellLaboratories undersea systems division. In 1992, he wasnamed a distinguished member of the technical staff ofAT&T Bell Labs. In 1996, he was promoted to AT&TTechnology Consultant and was promoted to AT&TTechnology Leader the following year. He holds 27 USpatents in the area of lightwave transmission systems. Heis a Fellow of the OSA, the IEEE, and AT&T.Neal S. Bergano is the recipient of the 2002 John TyndallAward “for outstanding technical contributions to andtechnical leadership in the advancement of global under-sea fiber optic communication systems.”

21leos05.qxd 10/25/07 10:29 AM Page 10


IntroductionIn 1991, when I first proposed to develop anew kind of fiber – which I initially called“holey fiber,” defusing any anxious looks byadding that the word needed an “e” – I wasmet with a good deal of disbelief. Somefriends even questioned my sanity. Who intheir right mind could possibly want tomake a fiber with an array of microscopichollow channels (“holes”) running along itslength, and anyway would this new thing(the “photonic bandgap”) really work?Surely, the refractive index of silica glasswas too small. Anyone – who knew any-thing – knew that the literature clearlystated that photonic bandgaps only appearif the refractive index ratio is very large, say2.2:1 for a two-dimensional dielectric-airstructure [1]. And then there were thepracticalities of making it. Distressed lookswould pass over the faces of those withlonger memories, who recalled how diffi-cult it was to make “single material” fibers[2]. Proposed in the early 1970s as low-losssingle-mode waveguides for telecommuni-cations, and made entirely from pure silicaglass, these structures consisted of a tubu-lar cladding shell connected to a centralcore by two thin webs of glass. They werevery hard to make, and work on them wasabandoned when modified chemical vapordeposition (MCVD) came along (quiterecently, after almost thirty years of inactiv-ity and prompted it seems by the emer-gence of PCF, there has been a revival ofinterest in single-material fibers for nonlin-ear optics and gas sensing [3, 4]).

So why bother to tackle such a difficult– and apparently impractical – technology?I suppose there were two main reasons. Thefirst was simple curiosity – the idea ofusing a photonic band gap to trap light ina hollow core was intriguing, even if yetunproven. Secondly, conventional fiber hadbecome a sort of highly respected elderstatesman with a wonderful history. One

had to concede that, whatever step-indexfiber could do, it did it extremely well. Thetrouble was that it could not do enough.The world was full of scientists and engi-neers who wanted fibers that could carrymore power, fibers that could be used moreeasily for sensing, fibers that had multiplecores, higher nonlinearities, lower nonlin-earities, higher birefringence, more ther-mal stability and widely engineerable dis-persion landscapes … in short, fibers withmore versatility.

Two main factors contribute to this lackof flexibility. The first is the smallness ofthe core-cladding index difference (Δ is lessthan 1% in telecommunications fiber),which limits the degree to which groupvelocity dispersion and birefringence canbe manipulated, and produces bend loss(0.5 dB at 1550 nm in Corning SMF-28for one turn around a mandrel 32 mm indiameter [5]). Although higher values of Δcan be attained (modified chemical vapordeposition yields an index difference of0.00146 per mol.% GeO2, up to a maxi-mum Δ ~ 10% for 100 mol.% [6]), thecore radius for single-mode behaviorbecomes very small and the attenuationrises through increased absorption andRayleigh scattering. The second factor isthe reliance on total internal reflection(TIR), so that guidance in a hollow core isimpossible (at least at UV, visible, andtelecommunications wavelengths), howev-er useful it would be in fields requiringelimination of glass-related nonlinearitiesor enhancement of laser interactions withdilute or gaseous media.

Glass-air PCF seemed to offer for thefirst time the revolutionary opportunity ofescaping the straitjacket of total internalreflection, allowing low-loss guidance oflight – in a single-mode – in a hollow fibercore. It also offered a very large index dif-ference (~1.46:1 for silica), setting thescene for solid-core fibers with much small-er mode volumes (and hence higher effec-tive nonlinearities), higher birefringence,and more widely engineerable dispersion.Furthermore, as they were made from onlyone solid material, the thermal stability of

these fibers was likely to be much higherthan that of fibers made from two differentglasses.

However, I am getting ahead of myself.Before launching into a discussion of themany new aspects of PCF, I would like tomake a detour and talk a little about wherethe idea came from.

BackgroundIt is almost always true that inventions arethe result of a multitude of different influ-ences and ideas. In my case, I had been fas-cinated, since my PhD work, by the behav-iour of light in periodically structuredmedia. This started in 1976, when I beganwork under the supervision of LaszloSolymar at Oxford, who at that time wasdeveloping two and three dimensional cou-pled mode theories to model the propaga-tion of light in volume holograms. Createdby the interference pattern of two or morearbitrary beams, and occupying a volumeof space, these holograms bear some resem-blance to three-dimensional periodicmedia, at least locally. Because the behaviorof light in these structures was often verycomplex and therefore hard to understand,I became interested in borrowing conceptsand intuitive tools from other fields, espe-cially the dynamical theory of x-ray diffrac-tion [7] and Floquet-Bloch theory forwaves in periodic media [8,9]. This natu-rally led to thinking in terms of electro-magnetic band-structure, Bloch waves andthe curious effects that appear when groupand phase velocity point in different direc-tions, or when the group velocity is inde-pendent of the direction of the incidentwavevector. This made me very receptive tothe suggestion by Eli Yablonovitch andSajeev John in 1987 that a full electromag-netic band gap might be created by peri-odically structuring a high refractive indexmaterial to produce a “photonic band gapcrystal”. Their main interest lay in creatingan absence of photonic states in threedimensions, something that Yablonovitchwent on to demonstrate experimentally atmicrowave frequencies [10]. I quickly real-ized that one might be able to achieve low-

Photonic Crystal Fibers: A Historical AccountPhilip Russell

Special 30th Anniversary

MAX-PLANCK RESEARCH GROUP,UNIVERSITY OF ERLANGEN-NUREMBERG

D91058 ERLANGEN, GERMANYWWW.PCFIBER.COM

21leos05.qxd 10/25/07 10:29 AM Page 11


loss guidance of light in a hollow fiber core.The challenge would be to increase thescattering sufficiently so that, over a rangeof axial wavevectors, propagation is closedoff for all radial and azimuthal directions inthe transverse plane – in other words, atwo-dimensional photonic band gap(PBG) appears.

Why not annular Bragg fibers?Another kind of structure that couldpotentially create this effect had beensuggested theoretically in 1968 byMelekin [11] and then studied in moredetail by Yariv some ten years later [12].It consisted of concentric tubular layers ofalternating refractive index (Fig. 1(d)).The idea was that rays traveling at anangle to the axis encounter a cylindricalBragg stack, and are fully reflected backinto the core, where they becometrapped. The modes that would most nat-urally guide in such a structure are thosewhere the electric or magnetic fields areparallel to the boundary, i.e., the fieldpolarization turns with azimuthal angle

(these are the TE01 and TM01 modes). Infact, the TE01 mode had been used previ-ously at microwave frequencies in hollowmetallic waveguides; the field movesaway from the waveguide walls as the fre-quency increases, resulting in very lowattenuation, although the guide must bekept very straight to avoid bending-relat-ed losses caused by field penetration intothe metal.

Although it is straightforward to pro-duce solid-core versions of such Braggfibers by MCVD [13], for guidance in ahollow core one is up against the need forlow effective index (the radial stop-bandmust appear at values of axial refractiveindex nax <1) and high index contrast (forstrong confinement). As a result, the indi-vidual layers must be very thin (thickness ≈(1.462–nax)-1/2 λ/4<0.69λ, where λ isthe vacuum wavelength), which enhancesthe effects of dopant diffusion during fiberdrawing, further reducing the already weakindex contrast. Small index contrast has thedrawback that, for good confinement, alarge number of periods is needed and thestructure must be highly perfect to avoid

leakage through defect states in thecladding layers.

The ideal structure would be a series ofconcentric glass layers with air betweenthem; of course, this structure would nothold together mechanically. One couldthink of increasing the index contrast usingtwo solid materials, but here the problemsare extreme for another reason. Pairs ofdrawable glasses with compatible meltingand mechanical properties, a large refrac-tive index difference, and high opticaltransparency have not yet been found.More exotic combinations of chalcogenideand polymer overcome the mechanicalproblems, but suffer from extremely highabsorption in the polymer layers (neverthe-less, 1 dB/m loss at 10 μm wavelength hasbeen reported for the TE01 mode [14]).

In the end, it turns out that PCF, witha triangular array of hollow channels, satis-fies all of the requirements in one step:high index contrast, mechanical stability,no problems with thermal expansion, andextremely low material absorption.

The undiscovered has no mapLike Thomas Edison in his search to find asuitable material for the element of a lightbulb (apparently he tried out 3000 fila-ments before coming up with one thatworked), I had no real idea how to go aboutproducing a fiber with holes that might beas small as half a micron in diameter,spaced a few microns apart in a crystallinelattice. There was no “map”. After all, noone had tried making something quite likethis before. Lithography was good for verythin structures, but it was hard to see howit could be adapted to produce even cmlengths of PCF. More promising was workat Naval Research Laboratories inWashington, D.C., where Tonucci hadshown that multi-channel glass plates withhole diameters as small as 33 nm, in atightly packed array, could be producedusing draw-down and selective etchingtechniques [15]. The maximum channellength was limited by the etching chem-istry to ~1 mm, and though the structureswere impressively perfect, they were clear-ly not fibers.

My earliest attempt, in 1991, involveddrilling a pattern of holes into a stub of sil-ica glass, my hope being that it could bedrawn into fiber. Machining an array of 1mm holes in a stub of silica ~2.5 cm in

Figure 1: Drawings of various structures: (a) birefringent PCF; (b) ultra-small core PCF;(c) hollow core PCF; (d) hollow core Bragg fiber. The white regions represent silica, the blackregions are hollow, and the colored regions are other materials (glasses or polymers).

(a) (b) (c) (d)

Figure 2: Some PCF structures. (a) the original stub of silica (2.5 cm wide) with holes 1mm wide partly drilled through it; (b) the hexagonal tube used in the first PCF stack; (c)scanning electron micrograph (SEM) of the first working solid-core PCF; (d) SEM of thelatest low-loss hollow core PCF designed for 1550 nm transmission (BlazePhotonics Ltd.)

(a) (b) (c) (d)

21leos05.qxd 10/25/07 10:29 AM Page 12


diameter (the largest the drawing furnacewould accommodate) proved beyond thecapabilities of the ultrasonic drill that wasavailable, so this approach was abandoned(Fig. 2(a)). I then had the good fortune toraise some funding from DRA Malvern inthe UK, and was joined successively by twopost-docs: Tim Birks in 1993 and JonathanKnight in 1995.

After trying various differentapproaches, we made the first successfulsilica-air PCF structure in late 1995 bystacking 217 silica capillaries (8 layersoutside the central capillary), speciallymachined with a hexagonal outer and acircular inner cross-section (Fig, 2(b)).The diameter-to-pitch ratio d/λ of theholes in the final stack was ~0.2, whichtheory showed was too small for PBGguidance in a hollow core, so we decidedto make a PCF with a solid central coresurrounded by 216 air channels (Fig.2(c)) [16]. This led to the discovery ofendlessly single-mode PCF, which, if itguides at all, only supports the funda-mental guided mode [17].

We soon realized that simple circularcapillaries were just as good, allowingcomplex lattices to be assembled fromindividual stackable units of the correctsize and shape. Solid, empty, or dopedglass regions could easily be incorporat-ed. The stacking technique resembles atechnology first used in the 2nd centuryBC by the Egyptians to make mosaicglass [18] – and by modern-day Italiansto fashion millefiori glass jewelry [19]. Ithas proved remarkably successful, large-ly because the lattice of holes is mechan-ically stable – the surface tension forcestend to balance out, allowing formationof highly regular structures during thedrawing process. Overall collapse ratiosof ~100,000 have been realized, and con-tinuous holes as small as 25 nm in diam-eter demonstrated. I am told that thisearned us a mention in the GuinnessBook of Records for the World’s LongestHoles, though I have not seen the entrymyself (Fig. 3).

Another versatile technique is extru-sion, first used by a team of researchersat Corning Incorporated [20]. In thisprocess, molten glass is forced througha die containing a suitably designedpattern of holes. It allows fiber to bedrawn directly from bulk material,almost any structure can be produced,

and it works for many materials,including polymers [21].

Designing the first hollow core PCFBack in 1991 I suspected that workingout-of-plane (nax > 0) would enhance thechances of finding a full two-dimensionalPBG. Transfer matrix calculations for sili-ca/air multilayer stacks confirmed thatbroad stop-bands could appear in theregion nax < 1 – an essential prerequisite forhollow core guidance. Precise numericalsolutions of Maxwell’s equations, carriedout by John Roberts at DRA Malvern, con-firmed this in 1995 [22]. His calculationsshowed that the familiar stop-bands, whichexist in all periodic structures, can coalesceto block all radial and azimuthal propaga-tion over a range of axial refractive indicesΔnax in the region nax < 1. Light could thenform a guided mode in a hollow core, whileits escape is blocked by a PBG in thecladding.

It turns out that the core diameter mustbe greater than a minimum value if a guid-ed mode is to appear under these condi-

tions. To understand why this is so – andindeed to understand many other featuresof PCF – we need to introduce the conceptof transverse effective wavelength. For thei-th material, with refractive index ni, thisis defined as follows:

λieff = λ(n2

i – n2ax)–1/2 (1)

where nax is the axial refractive index.This wavelength can be many times the

Table 1: Key milestones in the development of PCF

Year Milestone Ref.1995 2D bandgaps can exist in silica/air PCF for nax < 1 [22]1996 First solid-core PCF [16]1997 Endlessly single mode concept [17]1998 Ultra-large mode area [25]1999 Dispersion-shifted ultra-small core [26]1999 Hollow core photonic band gap PCF [23]2000 Multi-core PCF [27]2000 Polarisation-maintaining [28]2000 Rare-earth doped PCF laser [29]2000 Supercontinuum generation [30]2001 Carbon-dioxide laser processing of PCF [31]2001 Four-wave mixing [32]2001 Polymer PCF [33]2001 Soliton self-frequency shift [34]2002 Laser-tweezer guidance of particles in HC-PCF [35]2002 Long-period gratings [36]2002 PCF made from Schott SF6 glass for SC generation [37]2002 Stimulated Raman scattering in hydrogen [38]2003 Phononic bandgaps [39]2003 Rocking filters in PM PCF [40]2003 Cancellation of the soliton self-frequency shift [41]2003 Tellurite glass PCF [42]2004 Four-wave mixing [43]2004 Twin-photon generation in PCF [44]2005 EIT in acetylene [45, 46]2005 High energy transmission in HC-PCF [47]2005 Low-loss transitions between different PCFs [48, 49]2005 Photonic band gaps at 1% index contrast [50]

Figure 3: If the Channel Tunnel (connect-ing England and France) extended all theway to Jupiter, it would have the samelength-to-diameter ratio as a hole 25 nmwide and 1 km long.

The World’s Longest Holes

Earth

Jupiter

21leos05.qxd 10/25/07 10:29 AM Page 13


vacuum value, tending to infinity atthe critical angle (nax = ni), and beingimaginary when nax > ni (the fields arethen evanescent in the transverseplane). It is a measure of whether or notlight is likely to be resonant within aparticular feature of the structure, e.g.a hole or a strand of glass, and definesPCF as a wavelength-scale structure.For a strong enough band gap, we canapproximately assume that the fieldreaches zero at the hollow core edge,which for the fundamental mode yieldsthe condition ρ = z01λco

eff / 2π; 0.38λco-

eff where z01 is the first zero of a BesselJ0 function. For example, at a vacuumwavelength of 1.55 μm the hollow corediameter would have to be 8.4 mm ifthe axial refractive index of the guidedmode was 0.99.

In 1999, the first hollow-core PCF wasreported [23], and the best hollow-coretransmission losses now stand at 1.1dB/km at 1550 nm, only about six timeshigher than in telecommunications fiber(Fig. 2(d)) [24].

ApplicationsOne particularly attractive feature of PCFstructures is that they are highly uniformover very long distances. This means thatlight launched in at one end has time tosort itself out into a single mode, permit-ting highly reproducible detection of verysmall effects. Essentially, unwantedcladding modes are efficiently filtered outbefore they can interfere with the measure-ments. This is in sharp contrast to mostother kinds of photonic crystals, where tak-ing reliable optical measurements is a chal-lenging and painstaking process. As aresult, new PCF structures and PCF-basedapplications can rapidly be developed (seeTable 1 for some key milestones), perhaps

the most celebrated being the report in2000 of supercontinuum generation fromun-amplified Ti:sapphire fs laser pulses in aPCF with a core small enough to give zerodispersion at 800 nm wavelength [30].

Hollow core PCF has many fascinatingapplications, including gas-Raman cells forhigh efficiency [38], low threshold color-conversion of laser light. and laser-tweezerpropulsion and guidance of small particlesalong a curved path [35]. Another relative-ly unexplored area is optical sensing, withmyriad opportunities spanning many fieldsincluding environmental detection, bio-medical sensing, and structural monitoring[51].

As PCF becomes more widely used,there is increasing demand for effectivecleaves, low-loss splices, multi-port cou-plers, intra-fiber devices, and mode-areatransformers. The air-holes provide anopportunity not available in standardfibers: the creation of dramatic morpholog-ical changes by hole collapse (under surfacetension) or inflation (under internal over-pressure) when heating to the softeningtemperature of the glass [52]. Thus, notonly can the fiber be stretched locally toreduce its cross-sectional area, but themicrostructure can itself be radicallyaltered. Components (couplers, filters,transitions, etc.) made in this way have onegreat advantage over equivalent devicesmade in conventional fiber: being formedby permanent changes in structure, theyare highly stable with temperature andover time.

Orders of magnitude into the futureAt post-deadline sessions in telecommuni-cations conferences, companies oftenannounce incremental improvements inbit-error rates on long-haul fiber systems –the sign of a mature technology. PCF, onthe other hand, has often delivered ordersof magnitude improvement over the priorart – the sign of a successful disruptivetechnology. Here is a list of some of themost striking examples:• Seven orders of magnitude improve-

ment in nonlinear gas-laser devices(e.g. Raman cells), by offering pathlengths of order 1 km and single-mode operation

• More than octave-wide supercontin-uum sources that are five orders ofmagnitude brighter than an incan-

descent lamp (pumped by compactmicrochip or fiber lasers)

• Pulses of white light ~1 ps long and10,000 times brighter than the sun

• Frequency measurement to unprece-dented accuracy (1 part in 1015) in acompact desktop apparatus using anoctave-spanning fs frequency comb(this helped Theodore Hänsch real-ize a long-standing dream – and wina Nobel Prize in the process)

• Vanishingly small index contrasts canyield photonic band gaps under cer-tain conditions in all-solid PCF [50]

• Low-loss adiabatic tapered trans-formers allow conversion betweenthe guided modes of radically differ-ent PCF structures

• Fiber nonlinearity adjustable overthree orders of magnitude (by mov-ing from ultra-small solid core to ahollow core PCF)

• Probably the world’s best mirrors inhollow core PCF (1 dB/km, 20 µmcore, 1550 nm): 2.8 million bouncesper km; 0.35 µdB/bounce (betterthan 1 part per 10 million loss perbounce); a new mirror at everyreflection; and all azimuthal angles& polarisation states are reflected.

• Birefringence that is 10 times high-er, and 100 times more stableagainst temperature fluctuations,than is possible using conventionalapproaches.In some of these cases, the penetration is

already significant. For example, white-light supercontinuum sources are rapidlybecoming an essential tool in laboratoriesworldwide, even appearing in commercialmicroscopes.

It is clear that many exciting develop-ments are emerging, and will emerge,based on PCF in its many forms, withapplications spanning many disparatefields of science [53].

References[1] J. D. Joannopoulos, R. D. Meade,

and J. N. Winn, Photonic Crystals(Princeton University Press,Princeton, NJ, 1995).

[2] P.V. Kaiser, and H. W. Astle, BellSystem Technical Journal 53, 1021(1974).

[3] N. Y. Joly et al., Optics Letters 30,2469 (2005).

[4] A. S. Webb et al., OpticalEngineering 46 (2007).

[5] www.corning.com/opticalfiber/.[6] E. M. Dianov, and V. M. Mashinksy,

Figure 4: White-light supercontinuum(10,000 times brighter than the sun) gen-erated in PCF with zero dispersion at 800nm, pumped by 150 fs Ti:sapphire pulsesof energy 1∼nJ.

21leos05.qxd 10/25/07 10:29 AM Page 14


Journal of Lightwave Technology23, 3500 (2005).

[7] Z. G. Pinsker, Dynamical Scatteringof X-Rays in Crystals (Springer-Verlag, Berlin, 1978).

[8] L. Brillouin, Wave Propagation inPeriodic Structures: Electric Filtersand Crystal Lattices (McGraw-Hill,New York, 1946).

[9] T. Tamir, H. C. Wang, and A. A.Oliver, IEEE Transactions ofMicrowave Theory and Techniques12, 323 (1964).

[10] E. Yablonovitch, Journal of theOptical Society of America B -Optical Physics 10, 283 (1993).

[11] V. N. Melekin, and A. B. Manenkov,Soviet Physics - Technical Physics13, 1698 (1968).

[12] P. Yeh, A. Yariv, and E. Marom,Journal of the Optical Society ofAmerica 68, 1196 (1978).

[13] F. Brechet et al., Electronics Letters36, 514 (2000).

[14] S. G. Johnson et al., Optics Express9, 748 (2001).

[15] R. J. Tonucci et al., Science 258, 783(1992).

[16] J. C. Knight et al., in Conference onOptical Fiber Communications(Optical Society of America, SanJose, California, 1996).

[17] T. A. Birks, J. C. Knight, and P. St.J.Russell, Optics Letters 22, 961(1997).

[18] http://www.cmog.org[19] http://www.ilmosaicoveneziano

.com.[20] D. C. Allan et al., in Photonic

Crystals and Light Localisation inthe 21st Century, edited by C. M.Soukoulis (Kluwer AcademicPublishers, 2001), pp. 305.

[21] M. C. J. Large et al., MolecularCrystals and Liquid Crystals 446,219 (2006).

[22] T. A. Birks et al., Electronics Letters31, 1941 (1995).

[23] R. F. Cregan et al., Science 285,1537 (1999).

[24] P. J. Roberts et al., Optics Express13, 236 (2005).

[25] J. C. Knight et al., ElectronicsLetters 34, 1347 (1998).

[26] T. A. Birks et al., IEEE PhotonicsTechnology Letters 11, 674 (1999).

[27] B. J. Mangan et al., ElectronicsLetters 36, 1358 (2000).

[28] A. Ortigosa-Blanch et al., OpticsLetters 25, 1325 (2000).

[29] W. J. Wadsworth et al., ElectronicsLetters 36, 1452 (2000).

[30] J. K. Ranka, R. S. Windeler, and A.J. Stentz, Optics Letters 25, 25(2000).

[31] G. Kakarantzas et al., Optics Letters26, 1137 (2001).

[32] J. E. Sharping et al., Optics Letters26, 1048 (2001).

[33] M. A. van Eijkelenborg et al., OpticsExpress 9, 319 (2001).

[34] X. Liu et al., Optics Letters 26, 358(2001).

[35] F. Benabid, J. C. Knight, and P. St.J.Russell, Optics Express 10, 1195(2002).

[36] G. Kakarantzas, T. A. Birks, and P.St.J. Russell, Optics Letters 27,

1013 (2002).[37] V. V. R. K. Kumar et al., Optics

Express 10, 1520 (2002).[38] F. Benabid et al., Science 298, 399

(2002).[39] P. St.J. Russell et al., Optics Express

11, 2555 (2003).[40] G. Kakarantzas et al., Optics Letters

28, 158 (2003).[41] D. V. Skryabin et al., Science 301,

1705 (2003).[42] V. V. R. K. Kumar et al., Optics

Express 11, 2641 (2003).[43] C. J. S. de Matos, J. R. Taylor, and K.

P. Hansen, Optics Letters 29, 983(2004).

[44] J. E. Sharping et al., Optics Express12, 3086 (2004).

[45] F. Benabid et al., Optics Express 13,5694 (2005).

[46] S. Ghosh et al., Physical ReviewLetters 94 (2005).

[47] A. H. Al-Janabi, and E. Wintner,Laser Physics Letters 2, 137 (2005).

[48] S. G. Leon-Saval et al., Optics Letters30, 1629 (2005).

[49] A. Witkowska et al., Optics Letters31, 2672 (2006).

[50] A. Argyros et al., Optics Express 13,309 (2005).

[51] T. M. Monro et al., MeasurementScience & Technology 12, 854(2001).

[52] T. A. Birks et al., Laser Focus World42, 70 (2006).

[53] P. St.J. Russell, Journal of LightwaveTechnology 24, 4729 (2006).

21leos05.qxd 10/25/07 10:29 AM Page 15


AbstractRecently, rapid progress has been made on ultrahigh-Qmicrocavities and slow-light media. These systems arenow providing novel ways of controlling light, namely bytuning optical systems within their photon dwell time.This article shows that such tuning results in interestingphenomena, such as adiabatic wavelength conversion, andoptomechanical wavelength and energy conversion. Thesephenomena are physically similar to those observed inclassical oscillators, but they have yet to be seriously con-sidered in the field of optics.

Recent Progress on High-Q Nanocavities and Slow-light MediaThis letter introduces a novel way of controlling lightusing the adiabatic tuning of optical systems. Thismethod is closely related to recent developments in high-Q microcavities and slow-light media, so let us beginwith some of the latest work on these topics.

Rapid progress has been made as regards the qualityfactor (Q) of miniature-sized cavities, such as whispering-gallery-mode (WGM) cavities and photonic crystal (PhC)cavities2,3,4,5,6. Among them, PhC cavities have beenconsidered the most advantageous in terms of Q normal-ized to mode volume (V). PhCs are artificial dielectricstructures whose refractive index is periodically modulat-ed on the scale of the light wavelength, and can stronglyconfine light by their photonic bandgaps. Figure 1 showsa color photograph of two-dimensional silicon PhCs fab-

Figure. 1. Color photograph of two-dimensional PhCs on a silicon-on-insulator substrate fabricated by x-ray lithography. Various col-ors are due to different photonic band structures for different crys-tal patterns.

NTT BASIC RESEARCH LABORATORIES, NTT CORPORATION, 3-1MORINOSATO-WAKAMIYA, ATSUGI, 243-0198, JAPAN

Masaya Notomi

Industry Research Highlights

Adiabatic Tuning of Optical Microsystems

Figure. 2. Ultrahigh-Q silicon PhC cavities. (a) Design, (b) trans-mission spectrum, (c) ring-down measurement.

1575.3401575.336

1.2 pm

Wavelength (nm)

Tra

nsm

issi

on

Q = (1.28±0.06)×106

1 2 3 410

100

τ:1.12 nsec

Time (nsec)

Inte

nsity

c1-hole c2-hole c3-hole

(b)

(c)

Q = (1.34±0.08)×106

(a)

21leos05.qxd 10/25/07 10:29 AM Page 16

ricated by state-of-the-art semiconductor nanofabricationprocesses. The colors show the establishment of the pho-tonic band structures in the structures. The Q value ofwavelength-sized PhC cavities has increased from 103 to106 in the last few years and is still increasing. Figure 2shows a result we recently obtained for ultrasmall cavitieswith an ultrahigh Q based on two-dimensional PhCslabs7. This cavity is in fact a line defect whose width islocally modulated to realize in-line light confinement.We have reported that the theoretical Q of this particularcavity is higher than 108 with a cavity mode volume of1.6 (λ/n)3. We fabricated this design in silicon PhC slabsby using state-of-the-art electron-beam lithography and adry etching technique. The results of spectral-domaintransmission and time-domain ring-down measurementsare shown in Fig. 2(b) and (c). Both results showedapproximately identical loaded Q of 1.3 million.

With small- and high-Q cavities, it has been thoughtthat various light-matter interactions (which are general-ly very weak) would be greatly enhanced. If these cavitiesare applied to optical switching devices, the thresholdpower can be reduced by Q2/V8. If the cavities are appliedto emission devices, the spontaneous emission process canbe greatly enhanced or inhibited by the Purcell effect6.The cavities are also applied to solid-state cavity QED(quantum electro-dynamics) systems to realize a strong-coupling regime9. This article introduces another applica-tion of high-Q microcavities. Figure 2(c) shows that lightcan be stored in a tiny cavity for nanoseconds, longenough to allow tuning the cavity system within its pho-ton lifetime. Subsequent sections explain the possibilitiesthat arise with such tuning. It should be noted that it isusually difficult to tune an optical system within its pho-ton dwell time because the speed of light is so high. Thistuning is practical only for small systems with a long pho-ton dwell time.

Another class of optical system should be consideredfor such applications. Recently, various slow-light mediahave been studied10,11,12, mainly with a view to realizingoptical (or quantum) memories. In well-designed slow-light media, the photon dwell time is sufficiently longand the light pulse is well localized both spatially andspectrally thus enabling the system to be tuned withinthe photon dwell time. In this article, we concentrate onmicro-cavities, but we can expect to find similar featuresfor dielectric slow-light media. Slow-light media havemany issues in common with microcavities. In fact, wehave demonstrated that a device based on an ultrahigh-QPhC nanocavity exhibits ultraslow light propagation (thegroup velocity is 2 × 10−5 c ), as shown in Fig. 3.

Adiabatic Wavelength Conversion (Tuning Light Like a Guitar)Here, we examine a very simple situation. Suppose that asmall high-Q cavity captures a light pulse, and we changeits resonance frequency within its photon lifetime. Whathappens to the light pulse? We performed a series ofnumerical experiments by undertaking extensive electro-magnetic simulation using the finite-difference time-domain method13,14. We assumed realistic silicon PhCcavities (Fig. 4(a)) that can be fabricated with currenttechnology. We tuned the cavity resonance frequency byvarying its refractive index as a function of time as shownin Fig. 4(b). Our calculations revealed apparent wave-length conversion after the tuning as shown in Fig. 4(c).The wavelength shift is precisely the same as the reso-nance frequency shift. Note that we did not assume anynonlinear susceptibility in this simulation. One mightthink that this spectral change is related to self-phasemodulation induced by the time-dependent refractiveindex. Such χ3 nonlinearity phenomena should depend onthe tuning rate. However, the resultant spectra are iden-

Figure. 3. Slow-light transmission experiment using an ultrahigh-Q silicon PhC cavity.

8.4 μm

−3 −2 −1 0 1 2 3 4 5 6 7

Reference

PhC Cavity

Time (nsec)

Inte

nsity

Measured transit speedVg=5.8 km/s= c/50,000

Transit Time =1.45 ns

Width-Modulated PhC Cavity(Q=7.4x105)

Output PulseInput Pulse

Reference

21leos05.qxd 10/25/07 10:29 AM Page 17


tical even if we change the tuning rate (in this example,we used two different tuning rates (5 fs and 10 ps for thesame index shift)). This finding clearly shows that thiswavelength conversion is fundamentally different fromconventional χ3 nonlinearity such as self-phase modula-tion. As seen in the figure, there is no noticeable peak atthe original wavelength after the tuning, which showsthat the conversion efficiency is 100% if we disregard theloss caused by the finite photon lifetime.

Now it is clear that this wavelength conversion is fun-damentally different from conventional wavelength con-version using χ2 or χ3 optical nonlinearity, where higher

order polarization plays an essential role. With conven-tional wavelength conversion, the total energy is conservedat the photon level. However, in the present situation, theentire electromagnetic energy Uem is not conserved. Wecalculated the temporal variation of the electromagneticenergy in the cavity, which clearly shows that Uem is notconserved during the tuning process (Fig. 5). However, wefound that the quantity Uem/ω is always constant, asshown in Fig. 5. Uem/ω is a well-known quantity, which isan adiabatic invariant for classical oscillator systems15.This fact tells us that the wavelength conversion in Fig. 4is physically the same as the adiabatic tuning of classicaloscillators. One of the simplest examples is the tuning ofa guitar. Suppose that after plucking a guitar string togenerate sound with a certain tone, we twist the tuningpeg before the sound dies away. We can easily change thetone of the sound in this way. In such a process, Uem/ω isconserved. This is simply the conversion of the frequencyof the sound. What we observed in Fig. 4 is essentially thesame as this phenomenon but in optics. Such simple tun-ing has not been fully considered because it is practicallydifficult to achieve with conventional optical systems.However, it is now becoming possible because of recentprogress on microcavities and slow-light systems. Notethat we can draw the same conservation rule if we startfrom the conservation of the photon number (=Uem/ηω),

Figure. 4. Adiabatic wavelength conversion via dynamic tuning ofa PhC cavity. (a) Schematic, (b) tuning conditions, (c) calculatedspectra before and after the tuning.

Δn=+0.5%

1570 1580 1590 16001E-5

1E-4

1E-3

0.01

0.1

1

Fie

ld In

tens

ity (

a. u

.)

Wavelength (nm)

Δn=0 Δn=+0.5%

ΔT=5 fs

Δλ~+6 nm (+0.36%)

Tuned Area

1.005

1.000

20 25 30 35 40

α dT=5 fsdT=10 ps

Time (ps)

(a)

(b)

(c)

Figure. 5. Electromagnetic energy during the adiabatic wave-length conversion process.

−6 −4 −2 0 2 4 6

−4

−2

0

2

4

Δλ, Δ

U, Δ

(U/ω

)(%

)

Δn (%)

Δλ

ΔU

Δ(U/ω)

ΔU, Δλ vs. Δn

Pendulum

Guitar

21leos05.qxd 10/25/07 10:29 AM Page 18


Figure. 6. Optomechanical wavelength conversion. (a) Coupled PhC cavity structure, (b) resonant wavelength and Q of the cavity as a func-tion of d, (c) tuning condition, (d) calculated spectra before and after the tuning.

0 100 200 300 4000

1

2

3

4

5

6

7

Q (

x 10

7 )

Q

λc

1300

1400

1500

1600

Wavelength (nm

)

1350 1400 1450 1500 1550 1600 1650102

103

104

105

106

107

108After Tuning(d=60 nm)

After Tuning(d=30 nm)

Fie

ld In

tens

ity (

a. u

.)

Wavelength (nm)

No Tuning(d=0 nm)

0 1 2 3 4 5 6 7 80

60

120

d → 60 nm

d → 120 nm

d(n

m)

Time (ps)

d

(a) (c)

(b) (d)

d (nm)

Figure. 7. Optomechanical energy conversion. (a) Calculated radiation force and momentum exchange normalized by the stored energy, (b)energy conversion efficiency for different cavity Q values.

0 100 200 300 400 5000.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

Δp=Fτ

d (nm)

η=Δp/(U/c)

F

−dU/dz

100

101

102

103

104

105

MomentumExchange

Radiation Force

0.01 0.1 1 10 100−23

−21

−19

−10

−8

−6

0

−4

Log[

ΔU/U

]

Reference

Q=2x105

Q=3x106

U (pJ)

>1012

Q=5x107−2

(a) (b)

F (

µN)/

U (

pJ)

Fτ/(U/c)

21leos05.qxd 10/25/07 10:29 AM Page 19


but the phenomenon itself is purely classical because thisagreement is simply because a photon is defined using adi-abatic invariant in quantum mechanics. Very recently, ourtheoretical prediction has been experimentally demon-strated in silicon micro-ring cavities16

We summarize the features of our scheme in compari-son with a conventional wavelength conversion scheme. 1)With our scheme, conversion is a purely classical and lin-ear phenomenon, and 2) it does not depend on the lightintensity. The same conversion occurs even at the singlephoton level. 3) Our scheme does not need special nonlin-ear optical materials. As is apparent from the above expla-nation, adiabaticity is important for this conversion. If thecavity mode separation is larger than the inverse of thetuning time or the tuning is spatially homogeneous, adia-baticity is always maintained. Note that this condition israther difficult to realize for larger optical systems. If thiscondition is not satisfied, several peaks appear where thepeak number and amplitude sensitively depend on thetuning rate. In such a situation, it is difficult to distin-guish the process from conventional self-phase modula-tion. This issue has been discussed in detail elsewhere14.

Optomechanical Wavelength Conversion Although we assumed index tuning in the previous exam-ple, adiabatic wavelength conversion is possible by employ-ing any tuning method for the resonance frequency orguiding modes. When we employ index tuning, the shiftamount is limited by the extent to which we can change therefractive index within the photon lifetime (normally lessthan 1%). To enlarge the shift amount, we designed thecoupled cavity structure shown in Fig. 6(a), in which weinserted an air slit in the center of an ultrahigh-Q PhC-slabcavity17. The resonance frequency of this cavity dependssensitively on the thickness of the air slit d, but the Q ofthis cavity is insensitive to d, as shown in Fig. 6(b). Thismeans that we can change the resonance frequency withinthe long photon lifetime simply by changing d. Thisprocess should lead to adiabatic optomechanical wave-length conversion. We numerically simulated such aprocess for Minkowski equations by using the finite-differ-ence time-domain method and found that it enabled us torealize a very large wavelength conversion (Δλ/λ>20%) asshown in Fig. 6(c, d). Although we assumed a relativelyfast displacement rate in this calculation, the same resultshould occur for a much slower displacement if the dis-placement time is shorter than the photon lifetime. Webelieve that such mechanical tuning should be possiblebecause the longest photon lifetime of a microcavity islonger than a nanosecond and the fastest MEMS can oper-ate in the GHz range.

Optomechanical Energy Conversiontowards Ultra-efficient Optical MEMSThis optomechanical wavelength converter has anotherinteresting feature. This particular cavity can generate asignificantly large radiation force if it holds a light pulsein it. If the system is adiabatic, the radiation force F=-

dUem/dz should be proportional to dω/dz, which is verylarge for this cavity. Figure 7(a) shows the calculatedradiation force per unit energy of the pulse. It revealsthat approximately 1 μN is generated per 1 pJ, which isa much larger value than that of conventional opticaltweezers18. However, this is not the most essential issuefor this force generator. The most important issue is thatit can produce an extremely large momentum exchangeΔp=Fτ (τ. is the photon lifetime). Since the cavity massis extremely small, this large Δp naturally results in anextremely high efficiency for conversion from opticalenergy to mechanical energy. We solved an equation ofmotion and derived the energy conversion efficiencyΔU/U as shown in Fig. 7(b). For conventional situationsincluding optical tweezers, ΔU/U is extremely small.This can be intuitively understood by imagining a casewhere one moving ball collides with another ball at restand they exchange energies. The maximum energyexchange (100%) is obtained when the two balls haveequal mass. In the current situation, light does not havemass. Thus other than for relativistic situations such as aphoton rocket, ΔU/U is negligible. However, Fig. 7shows that ΔU/U can approach 10%, which is more thantwelve orders of magnitude greater than the convention-al value. This extremely efficient optomechanical energyconversion is one of the most interesting features of thiscavity.

Where does this ΔU come from? In fact, it comes fromthe wavelength conversion in this process. That is, thisoptomechanical energy conversion process is simply thereverse of optomechanical wavelength conversion. Inother words, very efficient optomechanical wavelengthconverters are automatically very efficient optomechanicalenergy converters. We expect that such an efficient opto-mechanical energy converter will be applied to ultra-effi-cient optical MEMS in future.

Other Topics and Related PhenomenaIn the two examples described above, we have shown thatwe can change the frequency of light by employing adiabat-ic tuning processes. Conventionally, linear optics systemsonly affect the momentum (or wavevector) of light and can-not change its frequency. However, this is not true.Recently, such dynamic tuning processes have been studiedin various fields. Stopping (halting) light using electromag-netically induced transparency (EIT) or coherent populationoscillation is a well known example of dynamic tuning,where the EIT dispersion is dynamically tuned and the trav-eling light state is adiabatically connected to the haltingatomic spin state19. This idea has been applied to all-dielectric systems consisting of slow-light waveguides orcoupled micro-resonator systems20,21,22. For such aprocess, the frequency spectrum of the initial light issqueezed or expanded during tuning. The origin of the vari-ation in the frequency spectrum is the same as that foundwith adiabatic wavelength conversion. Dynamic tuningprocesses provide us with a novel way of controlling the fre-quency of light.

21leos05.qxd 10/25/07 10:29 AM Page 20


References[1] D. K. Armani, T. J. Kippenberg, S. M. Spillane, K.

J. Vahala, “Ultra-high-Q toroid microcavity on achip,” Nature 421, 925 (2003).

[2] B-S. Song, S. Noda, T. Asano, Y. Akahane,“Ultrahigh-Q photonic crystal double-heterostruc-ture nanocavity,” Nature Mat. 4, 207 (2005).

[3] E. Kuramochi, M. Notomi, S. Mitsugi, A. Shinya, T.Tanabe, T. Watanabe, “Ultrahigh-Q photonic crystalnanocavities realized by the local width modulationof a line defect,” Appl. Phys. Lett. 88, 041112(2006).

[4] T. Tanabe, M. Notomi, E. Kuramochi, A. Shinya, H.Taniyama, “Trapping and delaying photons for overone nanosecond in an ultrasmall high-Q photonic-crystal nanocavity,” Nature Photon. 1, 49 (2007).

[5] T. Tanabe, A. Shinya, E. Kuramochi, S. Kondo, H.Taniyama, M. Notomi, “Single point defect photoniccrystal nanocavity with ultrahigh quality factorachieved by using hexapole mode,” Appl. Phys. Lett.91, 021110 (2007).

[6] S. Noda, M. Fujita, T. Asano, “Spontaneous-emissioncontrol by photonic crystals and nanocavities,”Nature Photon. 1, 449 (2007).

[7] T. Tanabe, M. Notomi, E. Kuramochi, H. Taniyama,“Large pulse delay and small group velocity achievedusing ultrahigh-Q photonic crystalnanocavities,” Opt. Express. 15, 7826(2007).

[8] T. Tanabe, M. Notomi, S. Mitsugi, A.Shinya, and E. Kuramochi, "Fast bistableall-optical switch and memory on siliconphotonic crystal on-chip," Opt. Lett. 30,2575 (2005).

[9] T. Yoshie et al. “Vacuum Rabi splittingwith a single quantum dot in a photoniccrystal nanocavity,” Nature 432, 200(2004).

[10] L. V. Hau, S. E. Harris, Z. Dutton, C. H.Behroozi, “Light speed reduction to 17metres per second in an ultracold atomicgas,” Nature 397, 594 (1999).

[11] M. Notomi, K. Yamada, A. Shinya, J.Takahashi, C. Takahashi, I. Yokohama,“Extremely large group velocity disper-sion of line-defect waveguides in photon-ic crystal slabs,” Phys. Rev. Lett. 87,253902 (2001).

[12] D. Mori, T. Baba, “Wideband and lowdispersion slow light by chirped photon-ic crystal coupled waveguide,” Opt.Express 13, 9398 (2005).

[13] M. Notomi, T. Tanabe, A. Shinya, E.Kuramochi, S. Mitsugi “Dynamic non-linear control of resonator-waveguide

coupled system in photonic crystal slabs", CLEOPacific Rim 2005, Tokyo (2005).

[14] M. Notomi, S. Mitsugi, “Wavelength conversion viadynamic refractive index tuning of a cavity,” Phys.Rev. A73, 051803(R) (2006).

[15] H. Goldstein, Classical Mechanics (Addison-Wesley,Massachusetts, 1980) p.531.

[16] S. F. Preble, Q. Xu, M. Lipson, Changing the colourof light in a silicon resonator, Nature Photon. 1, 293(2007).

[17] M. Notomi, H. Taniyama, S. Mitsugi, E. Kuramochi,“Optomechanical wavelength and energy conversionin high-Q double-layer cavities of photonic crystalslabs,” Phys. Rev. Lett. 97, 023903 (2006).

[18] M.P. Sheetz ed., Laser Tweezers in Cell Biology(Academic Press, San Diego, 1988).

[19] M. Fleischhauer, M.D. Lukin, “Dark-state polaritonsin electromagnetically induced transparency,” Phys.Rev. Lett. 84, 5094 (2000).

[20] M.F. Yanik, S. Fan, “Stopping light all-optically,”Phys. Rev. Lett. 92, 083901 (2004).

[21] B.P.J. Bret, T.L. Sonnemans, T.W. Hijmans,“Capturing a light pulse in a short high-finesse cavi-ty,” Phys. Rev. A 68, 023807 (2003).

[22] Q. Xu, P. Dong, M. Lipson, “Breaking the delay-bandwidth limit in a photonic structure,” NaturePhysics 3, 406 (2007).

21leos05.qxd 10/25/07 10:29 AM Page 21


AbstractIn this paper, we review our recent work on developing high inten-sity fiber laser technology and demonstrating its new applications.Use of very-large-core fibers has enabled the generation of opticalpulses with very high peak intensities suitable for achieving hotplasma conditions in laser driven targets. This development has ledto recent demonstrations of efficient EUV and X-ray radiation gen-eration with fiber lasers. Further progress in high intensity fiberlasers is critically dependent on developing very-large-core fiberswith performance indistinguishable from that of traditional single-mode fibers. One such technology, based on chirally-coupled fiberstructures, has been demonstrated to provide effectively single-mode performance for core sizes exceeding 30 μm, and holds greatpromise for future integrated high power fiber laser systems.

IntroductionFiber lasers represent the latest “generation” of laser technology,which is distinguished from all the previous “generations” by beingmuch more practical, as well as by its potential for generatinguniquely high optical powers. This distinction between fiber lasersand all the other types of lasers originates in the fact that light infiber laser cavities is guided and controlled by integrated glassstructures (optical fibers, various types of optical fiber couplers, andfiber-connected optical devices). This feature enables a previouslyunprecedented degree of control over the light in a laser medium,which allows for the design of very efficient, robust, compact, andreliable diode-pumped lasers with a great variety of novel laserarchitectures and functionalities (for example, fiber-array combin-ing), and previously unattainable laser-light characteristics (forexample, very high laser powers or electronically-controllable puls-ing characteristics). In fact, the very “classical” notion of a laser asconsisting of a gain medium placed in free-space between two mir-rors forming an open-cavity optical resonator is becoming replacedby a notion of an all-integrated “tweak-free” laser structure – some-what akin to the revolutionary transition from vacuum tubes to sil-icon-chip electronics that occurred in mid-20th century.

As a result, a true fiber revolution is currently occurring, with adramatic increase in the power levels generated from fiber lasersinto the few kW range (sufficient to cut and weld steel and othermetals in a variety of industrial applications [1]. This revolution hasmanifested itself in rapidly expanding market share for fiber lasers,and in numerous research programs being conducted towards fur-ther development of unique fiber laser capabilities. It is anticipatedthat this development will continue towards even higher optical

powers ranging from a few kW to tens of kW, as required by a vari-ety of emerging scientific, industrial, and defense needs. Manyapplications, however, require not only high average powers butalso high peak powers and high peak intensities. In this paper, wereview results showing the generation of high peak intensities fromfiber lasers, and their use in generating X-ray and EUV radiation.We also briefly describe here demonstration of a new type of large-core fiber technology, based on chirally-coupled core structures,which enables single-mode performance of very large-core fibers.This technology is crucial for further development of integratedhigh power laser sources.

13.4-nm EUV radiation generation for future lithography systemsEfficient high-power EUV light sources are presently considered asstrong contenders for high-volume semiconductor manufacturingat the 32-nm node, scheduled to be reached in 2010–2013 [2].Laser-produced plasma EUV sources have already demonstratedefficient EUV generation using solid-state lasers [3]. High-powerpulsed fiber lasers have a significant potential as cost-effectivemulti-kW power scalable laser drivers for high power laser-pro-duced-plasma EUV lithography sources. However, the question ofwhether fiber lasers can produce the same efficiency as solid-statelasers, in which high energy per pulse is possible. We recentlydemonstrated that, despite the significantly different operationalparameter space of pulsed-fiber lasers when compared with solid-state lasers, high conversion efficiency of more 2% of in-band 13.5-nm EUV generation is achievable with Sn-doped droplet targetsusing nanosecond pulse-fiber lasers.

The laser experimental setup for EUV generation study isshown in Fig. 1. A four-stage Master Oscillator Power Amplifier(MOPA) configuration using ytterbium-doped all fiber amplifierswith an electric-pulse driven semiconductor diode seed emitting at1064 nm was developed for EUV conversion efficiency study [4].The pulse is first pre-amplified by two stages of SM pre-amplifierfollowed by two stages of power amplifiers. The first stage of poweramplifier is 30-μm-diameter 0.06-NA core with 250-μm diameter0.45 NA inner pump cladding with available pump power up to50 W. The final amplifier fiber utilized 80-μm-diameter 0.06-NAcore with 400-μm-diameter 0.46-NA inner pump cladding. Theamplifier is end-pumped using diode lasers with available pumppower up to 650 W by combining sources at 915 nm and 974 nm.Despite the very large effective mode area (2749 µm2) of this stage,the beam quality of this fiber is controlled by a combination of fiber

University Research HighlightsHigh Intensity Fiber Lasers: Emerging NewApplications and New Fiber TechnologiesK.-C. Hou1, C.-H. Liu1, K.-H. Liao1, A. Mordovanakis1, M.-Y. Cheng1, G. Chang1, J. Nees1, S. George2, K. Takenoshita2, M. Richardson2, and A. Galvanauskas1

1CENTER FOR ULTRAFAST OPTICAL SCIENCE, UNIVERSITY OF MICHIGAN, 2200 BONISTEEL BOULEVARD, ANN ARBOR, MICHIGAN 48109

2LASER PLASMA LABORATORY, TOWNES LASER INSTITUTE, COLLEGE OF OPTICS AND PHOTONICS, UNIVERSITY OF CENTRAL FLORIDA,ORLANDO, FLORIDA 32816

21leos05.qxd 10/25/07 10:29 AM Page 22


coiling and mode matched excita-tion to produce an output withM2≈1.3-1.5. An arbitrary wave-form generator was used as the seeddiode driver to enable temporalshaping for compensation of pulsetemporal deformation due to gainsaturation as well as providingcapability of studying plasmadynamics on laser temporal shape.Energy up to 6 mJ is obtained atthis stage for 6 ns pulse durationwith Megawatt peak power.

A Sn-doped water droplet tar-get from the University ofCentral Florida [5] was used as arealistic target source with theadvantages of less debris genera-tion and higher repetition ratecompatibility. The spot size onthe target in the EUV generationexperiment is measured to be ≈18µm (1/e2 diameter) on the targetwith a 60-mm focusing lens. Inthe first round of experiments, an optimization study was con-ducted at a lower repetition rate at 1.75 kHz, corresponding toaverage laser output power greater than 10 W. Scaling of theaverage power output can be easily achieved by increasing therepetition rate. Peak power up to 1.7 MW was tested by using1.7-ns pulses with 2.9 mJ, corresponding to a peak intensitylevel of 6.7 x 1011 W/cm2. Fig. 2 shows the spectra from fiber-laser-produced-plasma with Sn-doped water droplet targets. The13.4-nm spectral region required for the next generation EUVlithography sources is marked separately. The photon countsstart dropping once the intensity exceeds 8x1010 W/cm2, indi-cating that this intensity is optimum for maximum conversionefficiency. The highest conversion efficiency achieved was 2.0%(at 13.4-nm within 1% BW) at the intensity of 8.4 x 1010

W/cm2. At higher intensity, conversion efficiency drops. Thiswork is the first demonstration of efficient 13.4-nm EUV gener-ation with a fiber laser source. Achieved conversion efficiency isin the same range as that obtained using Q-switched solid-statelasers in which the energy per pulse is much higher. High beamquality and large gain bandwidth of a ytterbium doped fused-sil-ica gain medium also promise an unrivalled potential for powerscalability to tens of kW level by spectral and spatial multiplex-ing, thus offering a technological path towards practical and eco-nomical EUV lithography sources.

Fiber-laser-driven hard X-ray generationUltrafast-laser-produced plasma hard X-ray sources are invaluabletools for time resolved diffraction studies, with applications rang-ing from biology to solid-state physics. Significant further advance-ment of these techniques requires substantial increase in sourcebrightness and practicality. However, the existing solid-state ultra-fast laser technology is limited in achievable powers. Fiber lasers area new rapidly developing and highly practical technology suitablefor high average powers, including a recent demonstration of ultra-

short pulse generation in the mJ range [6]. Recently we achievedthe first demonstration of efficient hard X-ray generation using afemtosecond fiber laser driver [7].

We used a high-energy fiber CPA system based on 65-μm-coreYb-doped LMA fiber [6]. Such a large fiber core enabled pulse ener-gies in the mJ range with pulse durations of ~300-fs and a meas-ured pre-pulse contrast ratio of better than 2.5x10-5 in a diffraction-limited output beam (measured M2 < 1.06). Due to the high fideli-ty of the generated pulses and beams we were able to demonstrateefficient X-ray generation starting with pulse energies of 100 μJ(after compressor).

The experimental set-up is shown in Fig. 3(a). The 1064-nm p-polarized pulses are focused on a thick Ni target using an f=5 cmgold paraboloidal mirror. An adaptive deformable mirror was usedto optimize the wavefront to achieve a near-diffraction limit spot sizeof 2.5 μm in the focal plane. The plasma is imaged along the spec-

Figure 1. Experimental high average power pulsed fiber laser system showing all fiber MOPA config-uration amplifier with Fabry-Perot (FP) diode seed. Two stages of monolithic single-mode fiber ampli-fier amplify the seed to kW peak power. Further amplification by two stages of large mode area (LMA)power amplifiers generates more than 1 MW of peak power in nanosecond pulses with excellent beamquality close to single transverse mode.

FP LaserDiode

AO

M

2.0-m 30-μm Yb PM-LMA DC Fiber

Pump (650-W)975-nm + 915-nm

Cooled 3.3-m 80-μmYb LMA DC Fiber

SM 974-nm pump diodes

Output Beam of Pulsed-Fiber Laser

Figure 2. Spectrum showing emission from Sn-doped water droplet tar-get with different intensity levels

5,000

10,000

15,000

20,000

11 12 13 14 15 16 17 18

2.0×1010W/cm2

6.9×1010W/cm2

9.3×1010W/cm2

1.2×1011W/cm2

Wavelength (nm)

Pho

ton

Cou

nts

(a.u

.)

(b)

21leos05.qxd 10/25/07 10:29 AM Page 23


ular direction onto the entrance slit of a visible spectrometer and thetarget position is optimized to maximize the second harmonic gen-eration from the target surface at high intensities. A 9 mm2 CdTeX-ray detector with a 25-μm-thick Be window is mounted 50 cmfrom the plasma along the surface normal. The X-ray energy spec-trum is shown in Fig 3(b) with the intensity on target at 1x1015

W/cm2. Assuming an isotropic emission into the half space, theenergy conversion efficiency ηKα into the Ni Kα photons (7.48keV) was calculated to be around 4.5x10–8. At 1 kHz repetitionrate, this corresponds to 1.7x106 Kα photons per second into 2π sr.The bremsstrahlung emission is characteristic of an effective electrontemperature Te= 3.6keV and extends to photon energies higherthan 20 keV.

The demonstrated first fiber-based hard X-ray source has yieldefficiencies similar to those obtained with solid state laser drivers withcomparable pulse energies and intensities [8]. Further experiments onincreasing X-ray yield efficiency with higher pulse energies from thecurrent system are in progress. Since this fiber laser technology is scal-able to >100 W of average power, this demonstration opens a pathtowards compact high-brightness hard X-ray sources for imagingand time resolved diffraction experiments.

Novel Fiber StructuresThe fiber laser revolution of achieving diffraction-limited multi-kW average power [1] and multi-MW peak power [4] outputbeams originates in the adoption of so called Large Mode Area(LMA) fibers. LMA fibers are large-core multimode fibers (core sizesare typically 20 µm to 30 µm diameter, with 65 µm to 80 µm stateof the art demonstrations [6], [4]), which require careful mode-

management [9],[10] for achiev-ing single-modeo u t p u t .Consequently,this approach isquite lacking in atraditional tech-nological advan-tage of conven-tional single-mode fibers ofbeing suitable forintegration into amonolithic all-fiber system.Indeed, whileconventional sin-

gle-mode fibers can be monolithically connected together usingfusion splicing equipment, it is increasingly difficult to achievemode-distortion-free splicing of LMA fibers with core sizes largerthan 20 µm. This limitation is currently hindering the develop-ment of all-monolithic large-core fiber systems for high powerapplications.

Recently we demonstrated that a new type of fiber structure, so-called chirally-coupled-core (CCC) fibers, can provide robust single-mode output in a manner identical to conventional single-modefibers, but with core sizes greatly exceeding the standard 30-µm ofLMA fibers (in fact exceeding 50-µm). Large-core CCC fibers donot require any external mode management, can be spliced intocomplex systems or used for development of a variety of fiber-opticdevices using standard fiber splicing and processing techniques,and can be coiled for compact packaging purpose only.

The geometry of CCC fiber, as shown in Fig.4, contains astraight central core and at least one helical satellite core, wrappedaround and in the optical proximity of the central core. The centralcore guides modes propagating along fiber axis, while the helix-coiled side core (or cores) supports modes that propagate in a heli-cal path around the central core. This composite structure can bedesigned to: (1) provide phase-velocity matched coupling betweenLP01 mode of the central straight core and helix-side core by select-ing suitable side-helix parameters such that β(side mode) + Δβhelix =(central mod e) ,where Δβhelix =2πn/λ [

√(2π R/�)2 + 1 − 1] is the

phase-velocity of the side core component that can be controlled byhelix period ∧ and radius R; (2) provide efficient modal-symmetryselective coupling between higher-order modes (HOM) in the cen-tral straight core and the side-helix modes through quasi-phasematching of phase velocities ΔβQPM =�βQPM = �βhelix ± (2π/�)l Dβhelix ± (2π/Λ)l (l is the HOMazimuthal number); and (3) allow the selection of suitable side-coreparameters and period to provide high loss for selected helix-sidemodes. These features permit a large variety of functionality forsuch a fiber structure. For example, CCC fiber can be made effec-tively single-mode by using (2) and (3), i.e. by coupling all HOMof the central core into high-loss helix side modes and thus allow-ing only LP01 mode in the central core to propagate. As anotherexample, by inducing partial coupling between the central-coreLP01 mode and a low-loss side mode using 1), the phase velocity

Figure. 3 (a) The experimental setup used for x-ray generation. (b) X-ray energy density spectrum showing thebremsstrahlung and the Kα line at 7.43 keV. The energy conversion efficiency into the Kα line was 4.5x10-8.

60° off-AxisParaboloid

Be Window

AmptekCdTe Detector

BG-39 Filter

VisibleSpectrometer

Ni Target

DeformableMirror

300fs, 100μjp-Polarized1,064 nm Pulses

107

106

105

104

103

5 10 15 20 25 30

X-R

ay E

nerg

y D

ensi

ty [k

eV/s

/2π]

Photon Energy [keV]

Spectral Energy DensityMaxwellian Fit Te~3.6keV

ηkα=4.5.10−8

Figure 4: Chirally-Coupled Core fiber structure (left) and cross sec-tion of a fabricated CCC fiber (insert).

21leos05.qxd 10/25/07 10:29 AM Page 24


and dispersion of the LP01mode can be controlled.Indeed, both positive andnegative dispersion valuescan be achieve in arbitrarilylarge fiber cores with magni-tudes >>10 times the disper-sion of fused silica.

Several CCC fiber sam-ples with 34-µm core (NAvalues between 0.065 and0.07) and a variety ofdesign parameters for oper-ation in 1550-nm and1064-nm spectral regionshave been fabricated andexperimentally character-ized. CCC fiber geometry can be implemented using commer-cially available fiber fabrication techniques by spinning fiberpreforms during the fiber drawing process. It is critical forachieving effectively single-mode performance that all fiberHOM would be sufficiently suppressed. An approximate crite-rion is to achieve >100 dB/m loss for HOM in the central core.Fig. 5 shows calculated power transient of a truly single-mode30-µm fiber (achieved by taking core NA=0.02 with V < 2.402– red line in the figure) compared to that of an effectively sin-gle-mode CCC fiber of the same core size (but 0.06 NA) with >100 dB/m HOM loss (blue line). In both cases, the fiber core isexcited with a significantly mode-mismatched beam in whichonly 12.5% of the total power matches the LP01 mode.Consequently, the remaining power is lost into the claddingwithin a short 10-20-cm length of each fiber, i.e. only the fun-damental mode remains. Results of experimental characteriza-tion of one of the fabricated samples are shown in the right-handside of the figure. A standard SM fiber have been spliced to theCCC sample, ensuring that very poor modal match will beachieved at the excitation port of the CCC fiber, representingsituation similar to the calculated results on the left-hand sideof the figure. Experimental results clearly show that, despitesuch deliberately poor coupling and large core size, this CCCfiber acts as a single-mode fiber for fiber lengths as short as 25-cm. Further characterization confirmed that such CCC fiberscan be used as a conventional SM fiber, allowing splicing withno modal-content degradation and splice loss of ~0.1, insensi-tivity to fiber perturbations and coupling conditions, negligibleLP01 modal loss (<0.1-dB/m), etc. Such CCC fibers represent asignificant technological breakthrough, permitting mode-sizescaling of conventional telecom-type fiber technology, whichshould be of critical importance for further practical develop-ment of high-power fiber laser technology. Our modeling indi-cates that effectively SM core sizes of > 50 µm are achievable.

SummaryLarge-core fiber laser technology allowed achieving pulsed peakpowers sufficiently high for efficient hot-plasma generation of EUVand X-ray spectral range radiation. Although current demonstra-tions rely on conventional large-core LMA fibers, the emerging newtechnologies of large-core effectively single-mode fibers hold the

promise of enabling all-fiber systems with core sizes exceeding 50μm. Development of these new fiber technologies is crucial formaking such sources practical. Practicality and high-power scala-bility of fiber lasers opens a path towards compact and robust highbrightness sources of such short-wavelength radiation required fora wide range of applications.

References[1] V. Gapontsev, D. Gapontsev, N. Platonov, O.

Shkurikhin, V. Fomin, A. Mashkin, M. Abramov, S.Ferin, “2 kW CW ytterbium fiber laser with record dif-fraction-limited brightness,” Proc. Conference on Lasersand Electro-Optics/Europe, Munich, Germany, June 12-17, 2005, Page(s) 508

[2] R. H. Stulen, "13-nm Extereme Ultraviolet Lithography,”Journal of Selected Topics in Quantum Electronics 1, 970-975(1995)

[3] C-S. Koay, Simi George, K. Takenoshita, et al Proc. of SPIE,Emerging Lithographic Technologies IX 5751, 279-292(2005)

[4] A. Galvanauskas, Ming-Yuan Cheng, Kai-Chung Hou, andKai-Hsiu Liao,” High Peak Power Pulse Amplification inLarge-Core Yb-Doped Fiber Amplifiers,” IEEE JSTQE 13,559-566 (2007)

[5] F. Jin, K. Gabel, M. Richardson, M. Kado, A. F. Vasil'ev & D.Salzmann, "Mass limited laser plasma cryogenic target for 13nm point x-ray sources for lithography," Proc. SPIE 2015,151-159 (1993)

[6] Kai-Hsiu Liao et al. “Diffraction-Limited 65-µm Core Yb-doped LMA Fiber based High Energy Fiber CPA Systems,”CLEO 2006, Long Beach, Calif., May 21–26, 2006, post-deadline paper CPDB4

[7] K.-H. Liao, A. G. Mordovanakis, B. Hou, G. Chang, G.Mourou, J. Nees, and A. Galvanauskas, “An efficient Ni K_X-ray source driven by a high energy fiber CPA system,”CLEO/Europe-IQEC Conference, June 2007, Munich,Germany, postdeadline paper CP1-4-THU

[8] M. Hagedorn et al. “High-repetition-rate hard X-ray genera-tion with sub-millijoule femtosecond laser pulses,” AppliedPhysics B, 77, 49-57 (2003)

[9] M.E.Fermann, “Single-mode excitation of multimode fiberswith ultrashort pulses,” Optics Letters, 23, pp.52-54, 1998

[10] J.P.Koplow, D.A.V.Kliner, and L.Goldberg, “Single-modeoperation of a coiled multimode fiber amplifier,” Optics

Figure 5. Calculated power transients in SM and effectively SM fibers (left-hand side of the figure). Cut-back measurement of the modal quality from a fabricated CCC sample demonstrating effectively single-mode performance (right-hand side of the figure).

1.00

0.75

0.50

0.25

00 0.2 0.4 0.6

Tota

l Pow

er in

the

Fib

er (

a.u.

)

Propagation Distance (m)

0

0.5

1.0

1.5

2.0

0 25 50 75Length of CCC Fiber (cm)

M2

21leos05.qxd 10/25/07 10:29 AM Page 25


thirst for bandwidth/lumens/processing willcontinue unabated. The basic skills of ourmembers will have value among a wide vari-ety of markets (i.e., communications, datastorage, health care, lighting, manufactur-ing, sensing), and our technologies will findever-more applications in inter-disciplinaryfields. All this will enable people to migrateto different flourishing sectors, therebyreducing risks.

(4) Exhibit Halls: Exhibit halls will still beof significant value as a meeting venue forcustomers and vendors. However, theexchange of specific product information willbe relatively unimportant. Rather, it will bedriven by the need for face-to-face contact to:(i) build a collaborative relationship, (ii) dis-cuss specifications of next-generation prod-ucts, and (iii) provide a hand-on feel of therelative user-friendliness of a product.

(5) Membership: Given the longer andhealthier lifespan, the demographics of ourmembership will contain more retired “Life”members. Many of these members will stillbe interested in some of our activities, butthey will have limited financial resources toparticipate as fully as they might wish.LEOS’ service and financial models willevolve to more fully serve and engage thispopulation segment. I firmly believe thatyoung members experience tremendousinspiration when interacting with seniormembers who still value our community.

(6) Interdisciplinary Activities:Information exchange will enable techno-logical developments within one disciplineto become rapidly intertwined with tech-nologies from other disciplines. Multi-dis-

ciplinary and multi-society activities will beas common as single-society activities.

Our Immutable Core Values“What's past is prologue.” WilliamShakespeare's The Tempest, Act 2 Scene 1.

Our core values bind us together, andthey are manifest in our enduring character-istics.

Furthermore, novelist Ellen Glasgowrightly points out that,“all change is notgrowth, as all movement is not forward.”

(1) Journals and Conferences: Manuscriptdecisions for journals and conferences will bedriven, first and foremost, by quality andintellectual honesty. If this changes, wemight as well close up shop.

(2) Networking: In-person meetings andconferences will remain important.Technology can fill in gaps and make com-munication easier, but there is no substitutefor personal interactions. It will remain crit-ical for people to directly and effectivelymarket themselves to present and futureleaders.

(3) Membership: Our members will behighly educated, highly motivated scientistsand engineers. We will remain a member-driven, non-profit society, for which soundfinancial operation is meant to ensure thatwe continue to provide value to our mem-bers.

(4) Neutrality: LEOS will have no hiddenagenda, other than to provide a neutral, non-profit forum for the effective exchange ofhigh-quality technical ideas for the benefitof our community.

My Hopes for the LEOS’ Future“Things alter for the worse spontaneously,if they be not altered for the betterdesignedly.” Sir Francis Bacon.

Some of my hopes for our future include:

(1) “Flat” Rejection Ratios: Althoughmost manuscript submissions come fromoutside the U.S., the rejection ratio of non-U.S.-authored papers is higher than thosefrom U.S. authors. This is a gentrified situ-ation, one that should hopefully change in a“flat” world.

(2) Increased Member Retention:Each year, LEOS tends to lose roughly asmany members as we gain. We willhopefully become stronger in terms ofretaining members. The best way to dothis is by: (i) providing clear value toprofessionals in our technical field, and(ii) identifying emerging LEOS-relatedtechnologies and approaches to servethese new communities.

(3) Student “Buzz”: We need to be rele-vant to students, such that our activities arethe ones that they aspire to attend, con-tribute, and lead. Our activities must be theones that students discuss among them-selves. They are our future.

In summary, we all hope for positivechanges, no deviation from our core values,and a bright and secure future. Anythoughts on LEOS 2027?

Respectfully submitted,Alan E. Willner

University of Southern California

President’s Column(continued from page 3)

21leos05.qxd 10/25/07 10:29 AM Page 26


News

The IEEE Lasers & Electro-Optics Society announced the winners of the LEOS Graduate Student Fellowships for 2007.The Fellowships will be presented during the Awards Banquet at the LEOS Annual Meeting to:

Amit Agrawal - University of UtahMaria Ana Cataluna - University of St AndrewsIgnace Gatare Gahangara - Vrije Universiteit Brussel and Ecole Supérieure d’ElectrictéMaria Garcia Larrode - Eindhoven University of TechnologyZhensheng Jia - Georgia Institute of TechnologyHannah Joyce - The Australian National UniversityYannick Keith Lize - Ecole Polytechnique de MontrealCicero Martelli - University of SydneyHouxun Miao - Purdue UniversityJoris Van Campenhout - Ghent UniversityDirk van den Borne - Eindhoven University of TechnologyLin Zhu - California Institute of Technology

The LEOS Graduate Student Fellowship Program was established to provide Graduate Fellowships to outstanding LEOSstudent members pursuing graduate education within the LEOS field of interest.For more information on the program, check the Student site on the LEOS Home Page (www.i-LEOS.org)

LEOS 2007 Graduate Student Fellowship Winners Announcement

The IEEE Lasers & Electro-Optics Society announced the winners of the five (5) LEOS Chapter Awards for 2007. TheChapters will be recognized during the Awards Banquet at the LEOS Annual Meeting to:

The Chapter of the Year Award - Santa Clara Valley ChapterThe Largest Membership Increase Award - Poland ChapterThe Most Improved Chapter Award - Rochester ChapterThe Most Innovative Chapter Award - Ukraine ChapterThe LEOS Chapter Senior Member Initiative Award - Montreal Chapter

LEOS 2007 Chapter Award Winners

“Nick” Cartoon Series

21leos05.qxd 10/25/07 10:29 AM Page 27


Career Section

Education: B. A. Physics, 1972, University ofCalifornia, MSc and PhD Physics in 1974 and1979, University of Michigan. First at BellLabs, he investigated photodetectors for opti-cal communications. In 1985, Prof. Forrestjoined the Electrical Engineering andMaterials Science Departments at USC wherehe worked on optoelectronic integrated cir-cuits, and organic semiconductors. In 1992,Prof. Forrest became the James S. McDonnellDistinguished University Professor ofElectrical Engineering at PrincetonUniversity. He served as director of theNational Center for Integrated Photonic Technology and asDirector of Princeton’s Center for Photonics andOptoelectronic Materials (POEM). From 1997-2001, heserved as the Chair of the Princeton’s ElectricalEngineering Department. In 2006, he rejoined theUniversity of Michigan as Vice President for Research, andas the William Gould Dow Collegiate Professor inElectrical Engineering, Materials Science and Engineering,and Physics. A Fellow of the IEEE and OSA and a member

of the National Academy of Engineering, hereceived the IEEE/LEOS DistinguishedLecturer Award in 1996-97, and in 1998 hewas co-recipient of the IPO NationalDistinguished Inventor Award as well as theThomas Alva Edison Award for innovationsin organic LEDs. In 1999, Prof. Forrestreceived the MRS Medal for work on organicthin films. In 2001, he was awarded theIEEE/LEOS William Streifer ScientificAchievement Award for advances made onphotodetectors for optical communicationssystems. In 2006 he received the Jan

Rajchman Prize from the Society for Information Displayfor invention of phosphorescent OLEDs, and is the recipi-ent of the 2007 IEEE Daniel E. Nobel Award for innova-tions in OLEDs. Prof. Forrest has authored ~400 papers inrefereed journals, and has 155 patents. He is co-founder orfounding participant in several companies, includingSensors Unlimited, Epitaxx, Inc., Global Photonic EnergyCorp., Universal Display Corp. (NASDAQ: PANL) andASIP, Inc. (now Apogee Photonics).

Ching W. Tang recently joined the Universityof Rochester as the Doris Johns CherryProfessor of Chemical Engineering. Dr. Tangis known internationally for his pioneeringwork on organic solar cells and organic lightemitting diodes (OLEDs). His invention ofthe organic bi-layer heterojunction structureforms the basis of modern organic electronics.

Dr. Tang obtained a B.Sc. degree in chem-istry from the University of British Columbiain 1970 and a Ph.D. degree, also in chem-istry, from Cornell University in 1975. Hejoined the Kodak Research Laboratories inRochester, New York immediately after graduating fromCornell, and he left Kodak in 2006 to join the Universityof Rochester as Professor of Chemical Engineering andChemistry.

In addition to his groundbreaking research on organicsolar cells and organic light emitting diodes, Dr. Tang hasbeen credited with a number of key innovations leadingto the commercialization of a new flat-panel display tech-nology. These include the development of robust trans-

port and luminescent materials, improveddevice architectures, novel color pixilationmethods, and fabrication processes for themanufacture of passive-matrix OLED dis-plays; and the adaptation of active-matrixbackplane technology for high-definitionOLED displays.

Dr. Tang is a Fellow of the AmericanPhysical Society and a Fellow of the Societyfor Information Display. He was elected amember of the National Academy ofEngineering in 2006. He holds more than 70U.S. patents and has published 70 papers,

including three highly cited papers based on his originalwork on solar cells and OLEDs. Dr. Tang has received anumber of awards, including the Eastman InnovationAward (2000) from Eastman Kodak Company, theCarothers Awards (2001) from ACS, the Jan RajchmanPrize (2001) from the Society for Information Display,the Team Innovation Award (2003) from ACS, theHumboldt Research Award (2005), and the Daniel E.Noble Award (2007) from IEEE.

2007 IEEE Daniel E. Noble Award RecipientsStephen R. Forrest

Ching Tang

21leos05.qxd 10/25/07 10:29 AM Page 28


Career Section (cont’d)

Richard Friend has been on the Faculty in theDepartment of Physics, University ofCambridge, since 1980, where he is theCavendish Professor of Physics. ProfessorFriend has pioneered the study of organicpolymers as semiconductors, and his researchgroup has demonstrated that these materialscan be used in wide range of semiconductordevices, including light-emitting diodes andtransistors. He has been very active in theprocess of technology transfer of this researchto development for products. He co-founded

Cambridge Display Technology Ltd in 1994.Light-emitting polymer displays developedby Cambridge Display Technology are nowbeing manufactured under license and arenow used in a number of consumer products.He co-founded Plastic Logic Ltd in 2000 todevelop directly-printed polymer transistorcircuits, and these are now being developed asflexible active-matrix backplanes for e-paperdisplays. He is currently working on the useof polymer and related materials for thin-filmphotovoltaic diode applications.

Richard H. Friend

Membership Section

There are many benefits to becoming an IEEE Senior Member:• The professional recognition of your peers for technical and professional excellence• An attractive fine wood and bronze engraved Senior Member plaque to proudly display.• Up to $25 gift certificate toward one new Society membership.• A letter of commendation to your employer on the achievement of Senior member grade

(upon the request of the newly elected Senior Member.)• Announcement of elevation in Section/Society and/or local newsletters, newspapers and notices.• Eligibility to hold executive IEEE volunteer positions.• Can serve as Reference for Senior Member applicants.• Invited to be on the panel to review Senior Member applications.

The requirements to qualify for Senior Member elevation are, a candidate shall be an engineer, scientist, educator, technical execu-tive or originator in IEEE-designated fields. The candidate shall have been in professional practice for at least ten years and shall haveshown significant performance over a period of at least five of those years.”To apply, the Senior Member application form is available in 3 formats: Online, downloadable, and electronic version. For moreinformation or to apply for Senior Membership, please see the IEEE Senior Member Program website: http://www.ieee.org/organi-zations/rab/md/smprogram.html

The following individuals were elevated to Senior Membership Grade thru July - August:

Marc P. ChristensenGana N. Dash

Gregory R. KilbyZhongqi Pan

David Z. TingJunichiro Yamashita

Benefits of IEEE Senior Membership

New Senior Members

21leos05.qxd 10/25/07 10:29 AM Page 29


Membership Section (cont’d)

On May 9 and 10, 2007, LEOSvolunteers traveled to Capitol Hillto meet with members of Congressas part of a Washington “Drive-In”. Participants discussed theimportance of R&D spending andscience, math, and engineeringeducation with their legislators.They then asked Congress toincrease federal investments inbasic research and invest more ineducation at all levels.

The annual CLEO/QELS confer-ence, sponsored in part by LEOS, isheld regularly in Baltimore,Maryland, located only 45 minutesfrom Washington, D.C. This year,CLEO attendees were invited totake advantage of their proximityto our nation’s capital by spendingone day of the week-long confer-ence interacting with lawmakersand explaining their profession.

The Drive-In was organized bya coalition of science and engineer-ing groups, including IEEE-USA,the American Physical Society, andThe Optoelectronics IndustryDevelopment Association. Theevent was lead by the OpticalSociety of America.

In the end, over 30 CLEO atten-dees from 14 states made the trip,including 10 IEEE members. Theparticipants each met with both oftheir Senators and theirRepresentatives (or their staff) todiscuss innovation, R&D, and edu-cation issues. The LEOS members’meetings were arranged and facili-tated by IEEE-USA staff. All par-ticipants were given a three-hourtraining class to prepare them fortheir visits.

Studies of Congress have founda broad consensus among legisla-tors, their staff, and academics thatpersonal visits from constituentsare the single best way to influencea legislator’s vote. Legislators pay

attention when voters take thetime to come to Capitol Hill todiscuss the issues.

The approach certainly workedin this case. One of the top issuesthat participants discussed duringtheir meetings was the AmericaCompetes Act. This bill wasdesigned to improve America’sability to innovate by investingsignificant amounts of money inbasic research and education,specifically in science, math, andengineering.

The America Competes Act waspassed by both Houses of Congress onAugust 3. President Bush is expectedto sign the bill into law soon. The billrepresents the first notable increase infederal spending on science and engi-neering (outside of medical research)in the past 15 years.

The CLEO Drive-In built on thesuccess of the Science and Engineering

Congressional Visits Day (CVD), anannual event that brings over 200scientists and engineers from severaldozen associations into Washingtonto discuss R&D issues withCongress. This year, CVD occurredon May 1 and 2 and was chaired byIEEE-USA.

The messages and requests com-municated by participants in theDrive-In and CVD were the same,allowing participants to reinforceeach other. This approach provedto be a powerful way to influenceCongress.

Any IEEE member in theUnited States who is interested inmeeting with legislators shouldcontact IEEE-USA staffer RussHarrison at [email protected], Chapters, Societies, andStudent Groups that are interestedin speaking with a legislatorshould also contact Russ.

LEOS Goes To Washington

Group photograph of CLEO attendees (including several LEOS members) who attended the2007 Washington Drive In.

21leos05.qxd 10/25/07 10:29 AM Page 30


Membership Section (cont’d)

New LEOS Chapter Formed in Western AustraliaAssociate Professor Adam Osseiran, Edith Cowan University, Perth, Australia,[email protected]

Engineeringand R&DActivity Boomin the RegionWhen I came toWestern Australiafrom Europe fiveyears ago, I saw thatthe region waspoised for a promis-

ing future with numerous innovationpossibilities.

Western Australia is a young andvibrant part of the world offeringmany opportunities, and although itis more renowned for its flourishingresource industries, there are manyactivities and potential in engineer-ing research and development.

Western Australia has four univer-sities with a strong emphasis on engi-neering (Curtin University ofTechnology (CUT), Edith CowanUniversity (ECU), MurdochUniversity (MU), and the Universityof Western Australia (UWA)). Eachhas its strength in engineering topics,and collaboration between researchersfrom different universities and engi-neers from industry is common.

These new resources and activitiesare attracting thousands of new resi-dents to the region each month. Thisnew rush necessitates strengtheningthe existing basis of technologydevelopment to sustain a boomingresource-based economy.

Therefore, it made sense to set upa new WA Chapter to gather mem-bers from the region and federatethem around shared technical inter-ests while also promoting IEEE andLEOS.

How the Western AustraliaChapter was FormedI invited the IEEE DistinguishedLecturer Professor Vojin Oklobdzija

(SSCS Distinguished Lecturer) fromthe University of California at Davisto give a series of talks in Perth inNovember 2006. During these semi-nars, Prof Oklobdzija encouraged meto establish a SSCS chapter in theregion. Later, during another seminarat the University of WesternAustralia hosted by Professor LorenzoFaraone (EDS DistinguishedLecturer), we discussed the possibili-ty of forming the Chapter and theyboth encouraged me to establish ajoint LEOS/EDS/SSCS Chapter.

I contacted the Section Chair, Dr.Douglas Chai, collected therequired signatures on a petitionand forwarded it to the IEEE teamin Piscataway. Everything went veryquickly, and on 11 July, the newWestern Australia chapter was born.In time, I aim for the WesternAustralian Chapter to become asolid foundation for members to useas an innovation platform.

Western Australian Chapter MissionThe new Chapter is a joint associationbetween three Societies: Lasers andElectro-Optics (LEOS), ElectronDevices (EDS), and Solid StateCircuits (SSCS). Many activities inthis region are covered by these threeSocieties, but the objective is to even-tually split the Chapter into two orthree Society Sections as the numberof members grows and the activityincreases.

As the founder of the WesternAustralia Chapter, I will endeavour to:1. Invite distinguished researchers/

lectures to Western Australia2. Maintain a list of members and

invite new members to join3. Set up a website for the Chapter4. Organize IEEE conferences and

symposia

5. Hold meetings and seminars formembers

However, our first action item will beto ensure that IEEE plays a signifi-cant role in Western Australia so thatthe links within its engineering com-munity grow stronger.

Biography: Adam OsseiranAdam Osseiran (M’92-SM’06)obtained the B.Sc. degree in 1981and the M.Sc. in 1982 from theUniversity Joseph Fourier inGrenoble-France. He obtained thePh.D. in 1986 from the INPG,National Polytechnic Institute ofTechnology in Grenoble, France.

Adam is currently AssociateProfessor at Edith Cowan Universityin Perth, and the Director of theAustralian National NetworkedTeleTest Facility, a Major NationalResearch Facility project. Adam hasheld positions as Professor,Consultant, Sales and TechnicalDirector, Senior Research Fellowand instructor to many engineers onmicroelectronics design and designfor test, in both the academic andindustrial arenas.

He was Chair of the WorkingGroup of the IEEE Std 1149.4Mixed-Signal Test Bus 2000-2005;Vice Chair Test TechnologyTechnical Council Asia of theComputer Society; General Chair ofthe IEEE IMSTW 2002; GeneralChair of the IEEE DELTA 2004;General Chair of the VLSI-SOC2005; Vice General Chair of IEEEIDT 2006; Program Chair of sever-al Conferences and member of sev-eral Conference Steering andProgram Committees. AdamOsseiran was the invited editor forthe Microelectronics JournalSpecial Edition in 2002

Adam Osseiran

21leos05.qxd 10/25/07 10:29 AM Page 31


Conference Section

21leos05.qxd 10/25/07 10:30 AM Page 32


Publication Section

Announcing an Issue of the IEEE JOURNAL OF SELECTED TOPICS IN QUANTUM ELECTRONICS on Nonlinear-Optical Signal ProcessingSubmission Deadline – November 1st, 2007The IEEE Journal of Selected Topics in Quantum Electronicsinvites contributions of original papers in the area of non-linear-optical signal processing. The purpose of this issueof JSTQE is to document the state of the art and recentdevelopments in the field from both experimental and the-oretical perspectives. Solicitations topics include (but arenot limited to):

Wavelength conversionOptical limitingFrequency translation2R and 3R regenerationPhase conjugationOptical data format conversionPhase-sensitive amplificationAll-optical clock recoveryQuantum information processingAll-optical switchingFrequency comb generationParallel multi-wavelength processing

Call for Papers

Conference Section (cont’d)

2007 CLEO/Pac RimPhoto taken at the LEOS Member Reception at 2007 CLEO/Pac Rim in Seoul, Korea.

From left to right: Richard Linke, El Hang Lee, Yoshinori Namihira, Alan Willner and Hideo Kuwarhara

IEEE LEOS 20th Annual Meeting (LEOS 2007)Conference Dates: 21-Oct-2007 to 25-Oct-2007

Buena Vista Palace Resort & Spa, Lake Buena Vista, FL USA

Conference URL: http://www.i-LEOS.org

Conference E-mail: [email protected]

21leos05.qxd 10/25/07 10:30 AM Page 33


Publication Section (cont’d)

Nonlinear pulse shapingOptical bufferingOptical bi-stabilityOptical burst / packet switching Micro- and nano-scale nonlinear- All-optical OTDM multiplexing /optical devices demultiplexing

The Guest Editors for this issue are: Michael Vasilyev(University of Texas at Arlington, USA), and Yikai Su(Shanghai Jiao Tong University, China) and Colin J.McKinstrie (Alcatel-Lucent Bell Labs, USA),

JustificationAll-optical processing of high-speed signals, enabled by non-linear-optical devices, represents critical technology for futureoptical communications, computing, and military applica-tions. Originally demanding pulsed high-power lasers, non-linear-optical signal processing can now be achieved with low,semiconductor-laser compatible, powers and in small devicesizes, owing to the progress in quasi-phase-matched materials,semiconductor optical amplifiers, photonic-crystal and high-ly-nonlinear fibers, and silicon photonics. This, in turn, facil-itates the development of novel optical signal processingarchitectures and schemes, which bring nonlinear-optical sig-nal processing to the verge of practical deployment.

The deadline for submission of manuscripts isNovember 1, 2007; publication is scheduled for thebeginning of 2008. Please send a .pdf or Word file ofeach manuscript, including keywords, all author biogra-phies and IEEE Copyright Form to Chin Tan-yan at [email protected]. IEEE Copyright Form is available online at:http://www.ieee.org/about/documentation/copyright/cfrmlink.htm. Additional information for authorsregarding manuscript format may be found on theinside back cover of any issue of the IEEE Journal ofSelected Topics in Quantum Electronic or online athttp://www.ieee.org/organizations/pubs/transactions/information.htm All submissions will be reviewed inaccordance with the normal procedures of the Journal.

Email your manuscript and supporting documents to:JSTQE Editorial OfficeChin Tan-yanNonlinear-Optical Signal Processing IssueIEEE/LEOS45 Hoes Lane Piscataway, NJ 08854 USA

Contact [email protected] for any questions about thisissue. For all papers published in JSTQE, there are volun-

tary page charges of $110.00 per page for each page up toeight pages. Invited papers can be twelve pages andContributed papers should be 8 pages in length beforeoverlength page charges of $220.00 per page are levied.The length of each paper is estimated when it is received.Authors of papers that appear to be overlength are notifiedand given the option to shorten the paper. Additionalcharges will apply if color figures are required.

Announcing an Issue of the IEEE JOURNAL OF SELECTED TOPICS IN QUANTUM ELECTRONICS on Semiconductor Photonic MaterialsSubmission Deadline – September 1, 2007The IEEE Journal of Selected Topics in QuantumElectronics invites contributions of original papers in thearea of Semiconductor Photonic Materials. The purpose ofthis issue of JSTQE is to document the state of the art andrecent developments in the field from both experimentaland theoretical perspectives. The clear emphasis of thisspecial issue is the materials and the processing methodsfor semiconductor photonic devices, rather than devices.Contributions are, however, encouraged in which deviceresults are presented to highlight materials and processingadvances.Solicitations topics include (but are not limited to):

Epitaxial growth, e.g. MBE, MOCVD, CBEHeterostructuresQuantum wells, wires, dotsSelective Area EpitaxyQW/QD intermixingProcessing techniquesMicro and nanofabrication Wet/Dry etchingOhmic and Schottky contactsIon implantation, plasma processesOptical characterization Electrical characterisationNon-linear optical materials and devicesSilicon and Group IV photonicsWide Bandgap materials, e.g. nitrides, oxides Narrow bandgap semiconductorsMaterials for THz photonicsPhotonic CrystalsNanophotonic structuresNanowires, nanotubes

The Guest Editors for this issue are: Catrina Bryce(University of Glasgow, UK), Chennupati Jagadish(Australian National University, Australia) and JamesColeman (University of Illinois, USA)

21leos05.qxd 10/25/07 10:30 AM Page 34


Publication Section (cont’d)

JustificationSemiconductor materials played a crucial role in thedevelopment of photonics technologies. This issue isdedicated to all semiconductor materials and process-ing of relevance to photonics applications. Purpose ofthis special issue is to highlight recent developmentsin the field covering a broad range of materials,processes and techniques developed for a variety ofdevice applications.

The deadline for submission of manuscripts isDecember 1, 2007; publication is scheduled for theSummer of 2008. Please send a .pdf or Word file ofeach manuscript, including keywords, all author biog-raphies and IEEE Copyright Form to Chin Tan-yan [email protected]. IEEE Copyright Form is availableonline at:http://www.ieee.org/about/documentation/copyright/cfrmlink.htm. Additional information for authors regard-ing manuscript format may be found on the inside backcover of any issue of the IEEE Journal of Selected Topicsin Quantum Electronic or online athttp://www.ieee.org/organizations/pubs/transactions/information.htm All submissions will be reviewed inaccordance with the normal procedures of the Journal. Email your manuscript and supporting documents to:

Chin Tan-yan – Publications Coordinator JSTQE EditorialSemiconductor Photonic Materials IssueIEEE/LEOS445 Hoes LanePiscataway, NJ 08854 USA

Contact [email protected] for any questions about thisissue. For all papers published in JSTQE, there are volun-tary page charges of $110.00 per page for each page up toeight pages. Invited papers can be twelve pages andContributed papers should be 8 pages in length beforeoverlength page charges of $220.00 per page are levied.The length of each paper is estimated when it is received.Authors of papers that appear to be overlength are notifiedand given the option to shorten the paper. Additionalcharges will apply if color figures are required.

Announcing a Special Issue of the IEEE/OSA Journal of displayTechnology on Medical DisplaysSubmission Deadline – 31 December, 2007The IEEE/OSA Journal of Display Technology (JDT)invites manuscript submissions for a special issue. Thepurpose of this special issue is to document the currentstatus of Medical Displays through a collection of originalpapers. With a central theme that of medical visualization

and diagnosis, topics of interests include (but are not lim-ited to) 2D and 3D emerging grayscale and color tech-nologies, advanced multisensory interfaces for immersiveand augmented reality displays (e.g. visual haptic dis-plays), and task-based assessment of all display typesincluding stereo rendered 3D and 4D displays, and dis-plays for telemedicine and telerobotics. Articles related totechnological and perceptual bottlenecks are encouragedas well as semantics-driven technological solutions.The Primary Guest Editor for this issue is ProfessorJannick Rolland, University of Central Florida, Orlando,FL USA. Associate Guest Editors are Professor YongtianWang , Beijing Institute of Technology, P.R. China,Professor Itoh from Tsukuba University, Japan, Dr. GloriaMenegaz, University of Siena, Italy, and Dr. Aldo Badano,US Food and Drug Administration, Rockville, MD USA.

The deadline for submission of manuscripts is 31December 2007 and publication is tentatively scheduledfor the September 2008 issue. Manuscripts should con-form to requirements for regular papers (up to 8 double-column, single-spaced journal pages in length, keywords,biographies, etc.). All submissions will be reviewed inaccordance with the normal procedures of the Journal.

The IEEE Copyright Form should be submitted afteracceptance. The form will appear online in the AuthorCenter in Manuscript Central after an acceptance decisionhas been rendered.

For all papers published in JDT, there are voluntarypage charges of $110.00 per page for each page up to eightpages. Invited papers can be twelve pages in length beforeoverlength page charges of $220.00 per page are levied.The length of each paper is estimated when it is received.Authors of papers that appear to be overlength are notifiedand given the option to shorten the paper.

Authors may opt to have figures displayed in color onIEEE Xplore at no extra cost, even if they are printed inblack and white in the hardcopy edition. Additionalcharges will apply if figures appear in color in the hardcopyedition of the Journal.

Manuscripts should be submitted electronicallythrough IEEE’s Manuscript Central:

http://mc.manuscriptcentral.com/leos-ieee. Be sure toselect “2008 Medical Displays Special Issue” as theManuscript Type, rather than “Original Paper.” This willensure that your paper is directed to the special issue edi-tors. IEEE Tools for Authors are available online at:

http://www.ieee.org/organizations/pubs/transactions/information.htm

Inquiries can be directed to Lisa Jess, PublicationsAdministrative Assistant, IEEE LEOS Editorial Office,[email protected] (phone +1-732-465-6617; fax +1 732981 1138).

21leos05.qxd 10/25/07 10:30 AM Page 35


MANAGEMENTJames A. VickStaff Director, AdvertisingPhone: 212-419-7767Fax: [email protected]

Susan E. SchneidermanBusiness Development ManagerPhone: 732-562-3946Fax: [email protected]

Marion DelaneyAdvertising Sales DirectorPhone: 415-863-4717Fax: [email protected]

PRODUCT ADVERTISINGMidatlanticLisa Rinaldo Phone: 732-772-0160Fax: [email protected], NJ, PA, DE, MD, DC,KY, WV

New England/ConnecticutStan GreenfieldPhone: 203-938-2418Fax: [email protected]

New England/Eastern CanadaJody EstabrookPhone: 978-244-0192Fax: [email protected], VT, NH, MA, RICanada: Quebec, Nova Scotia,Newfoundland, Prince EdwardIsland, New Brunswick

SoutheastBill HollandPhone: 770-436-6549Fax: [email protected], NC, SC, GA, FL, AL,MS, TN

Midwest/Central Canada Dave JonesPhone: 708-442-5633Fax: 708-442-7620

[email protected] IL, IA, KS, MN, MO, NE,ND, SD, WICanada: Manitoba,Saskatchewan, Alberta

Midwest/Ontario, CanadaWill HamiltonPhone: 269-381-2156 Fax: [email protected], MI. Canada: Ontario

OhioJoe DiNardoPhone: 440-248-2456Fax: [email protected]

SouthwestSteve LoerchPhone: 847-498-4520Fax: [email protected], LA, TX, OK

So. California/Mountain StatesMarshall RubinPhone: 818-888-2407Fax: 818-888-4907 [email protected], AZ, NM, CO, UT, NV,CA 93400 & below

Northern California/Northwest/ Western CanadaLori KehoePhone: 650-458-3051Fax: [email protected], ID, MT, WY, OR, WA,CA 93401 & aboveCanada: British Columbia

Europe/Africa/Middle EastHeleen VodegelPhone: +44-1875-825-700Fax: [email protected], Africa, Middle East

Asia/Far East/Pacific RimSusan SchneidermanPhone: 732-562-3946Fax: [email protected], Far East, Pacific Rim,Australia, New Zealand

RECRUITMENT ADVERTISINGMidatlanticLisa RinaldoPhone: 732-772-0160Fax: [email protected], NJ, CT, PA, DE, MD,DC, KY, WV

New England/Eastern CanadaJohn RestchackPhone: 212-419-7578Fax: [email protected], VT, NH, MA, RICanada: Quebec, Nova Scotia,Prince Edward Island,Newfoundland, NewBrunswick

SoutheastThomas FlynnPhone: 770-645-2944Fax: [email protected], NC, SC, GA, FL, AL,MS, TN

Midwest/Texas/Central CanadaDarcy GiovingoPhone: 847-498-4520Fax: [email protected]; AR, IL, IN, IA, KS, LA, MI,MN, MO, NE, ND, SD, OH, OK, TX, WI. Canada:Ontario, Manitoba,Saskatchewan, Alberta

West Coast/Southwest/Mountain StatesTim MattesonPhone: 310-836-4064Fax: [email protected], CO, HI, NV, NM, UT,CA, AK, ID, MT, WY, OR,WA. Canada: BritishColumbia

Europe/Africa/Middle EastHeleen VodegelPhone: +44-1875-825-700Fax: [email protected], Africa, Middle East

IEEE Lasers and Electro-OpticsSociety Newsletter

Advertising Sales Offices445 Hoes Lane, Piscataway NJ 08854

www.ieee.org/ieeemediaImpact this hard-to-reach audience in their own Society

publication. For further information on product and recruitment advertising, call your local sales office.

ADVERTISER’S INDEXThe Advertiser’s Index contained in this issue iscompiled as a service to our readers and advertis-ers. The publisher is not liable for errors or omis-sions although every effort is made to ensure itsaccuracy. Be sure to let our advertisers know youfound them through the IEEE LEOS Newsletter.

Advertiser’s Index . . . . . . . . . . .Page #

R Soft . . . . . . . . . . . . . . . . . . . . . CVR2

Tempo Plastics . . . . . . . . . . . . . . . . . 3

Mathworks . . . . . . . . . . . . . . . . . . . . 5

Luna Technologies . . . . . . . . . . . . . . 7

IEEE Marketing . . . . . . . . . . . . . . . . . 9

California Scientific . . . . . . . . . . . . . 15

Third Millenium . . . . . . . . . . . . . . . 21

IEEE MDL . . . . . . . . . . . . . . . . . CVR3

General Photonics . . . . . . . . . . . CVR4

LEOS Mission StatementLEOS shall advance the interests of its mem-bers and the laser, optoelectronics, and pho-tonics professional community by:• providing opportunities for informa-

tion exchange, continuing education,and professional growth;

• publishing journals, sponsoring con-ferences, and supporting local chapterand student activities;

• formally recognizing the professionalcontributions of members;

• representing the laser, optoelectronics,and photonics community and servingas its advocate within the IEEE, thebroader scientific and technical com-munity, and society at large.

LEOS Field of InterestThe Field of Interest of the Society shall belasers, optical devices, optical fibers, andassociated lightwave technology and theirapplications in systems and subsystems inwhich quantum electronic devices are keyelements. The Society is concerned with theresearch, development, design, manufac-ture, and applications of materials, devicesand systems, and with the various scientificand technological activities which con-tribute to the useful expansion of the fieldof quantum electronics and applications.

The Society shall aid in promoting closecooperation with other IEEE groups andsocieties in the form of joint publications,sponsorship of meetings, and other forms ofinformation exchange. Appropriate cooper-ative efforts will also be undertaken withnon-IEEE societies.

21leos05.qxd 10/25/07 10:30 AM Page 36

FreedomTO INNOVATE

Subscribe by 31 December — Get your first month FREE! IEEE Member

Digital LibraryThe information you need to succeed canbe at your fingertips when you subscribeto the IEEE Member Digital Library.

■ The only way for individuals to access anyIEEE journal or conference proceeding

■ Over a million full-text documents■ The latest online research, plus a 50 year

archive for select titles■ Access to the top-cited publications you

need to make your project a success

Power up. Learn more at:www.ieee.org/ieeemdl

703-Qa MDL Freedom FP MoFree.ind1 1 6/26/06 9:35:13 AM

21leos05.qxd 10/25/07 10:30 AM Page 37

21leos05.qxd 10/25/07 10:30 AM Page 38

Documents

THE SOCIETY FOR PHOTONICS · fibers, Prof. Phillip Russell of Max-Planck Research Group for Optics Information and Photonics and the University of Erlangen-Nuremberg has contributed