ULTRAFAST OPTICAL SYSTEMS

Steve Kane, Jeff Squier, Ralph Jimenez,Lyuba Kuznetsova, Frank Wise, HenryKapteyn, and Bruno Touzet

Reflection grismscompensate dispersionin ultrafast systemsA new class of pulse compressor based on "grisms"—diffraction gratings coupledto prisms—can compensate for dispersion from hundreds of meters of glass path,allowing for fiber-delivery applications.

henever an ultrafastlaser pulse is propa-gated through ma-terial—an opticalfiber, a focusing ob-

jective in a microscopy system, or even the gainmedium in the laser that generated or ampli-fied the pulse in the first place—the wavelength-dependent refractive index of the material dis-torts the temporal shape of the pulse and reducesits peak power. Compensation of this materialdispersion is a fundamental aspect of nearly allullrafast optical systems, from the production ofpulses to the delivery of pulses to experimentsand applications. Broadening of the pulse by ma-terial dispersion is sometimes intentional anddesirable, and sometimes it is an unwelcome con-sequence of the optical system. However, the ulti-mate goal is always the same: to eventually prop-agate the pulse through a compensating opticalsystem^a pulse compressor—whose wavelength-depen-dent path length is opposite that of the dispersive materi-al, resulting in a fully compressed, high-peak-power pulseat the pulse's final destination.

Traditional pulse compressors are based on pairs ofdiffraction gratings or prisms, and both types of com-pressors have limited value for compensating materialdispersion. Grating-pair compressors can imperfectlycompensate for large amounts of material dispersion, andprism pairs can nearly perfectly compensate for only tiny

STEVE KANE and BRUNO TOUZET are with the Diffraction Gratings Divi-sion of Horiba Jobin Yvon, 3880 Park Ave, Edison, NJ 08820; e-maii: steve.kane@jobinyvon.com; www.lobinyvon.com/grisms. JEFF SQUIER is withthe Colorado School of Mines, Golden, CO 80401; RALPH JIMINEZ is withthe National Institute of Standards and Technology, 325 Broadway, Boul-der, CO 80305, and the University of Colorado, Boulder. CO 80309; LYUBAKUZNETSOVA and FRANK WISE are with Comell University, Ithaca, NY14853; and HENRY KAPTEYN is at the University of Colorado. Boulder.

FIGURE 1. A reflectiongrism is used for dispersioncompensation and pulsecompression. Grisms canbe cemented or air spaced,depending on the nature of theapplication, {Courtesy HoribaJobin Yvon)

amounts of material disper-sion. Until recently there havebeen no practical solutionsto the problem of completelycompensating modest to large{1 to 500 m) amounts of dis-persive material Over thepast year we have developedand demonstrated a new classof pulse compressors hasedon reflection grisms—metal-coated diffraction gratingscoupled to prisms—that fully

compensate material dispersion (see Fig. 1). ITiese grismshave been demonstrated in Ti:sapphire laser systems at800 nm and at 1030 nm in ytterbium (Yb)-doped fiber sys-tems with equal success, producing transform-limited puls-es after propagation through meters of dispersive material.

Dispersion in materialsand pulse compressorsAny dispersive system—cither a material with a wavelength-dependent index or a pulse compressor with a wavelength-dependent optical path—imparts a delay to a pulse that canbe expressed as a function of wavelength (or more conve-niently, frequency). A delay that is linear with frequency iscalled group-delay dispersion (GDD); a delay that is qua-dratic with frequency is termed third-order dispersion(TOD). When a pulse experiences GDD. it is symmetricallystretched in time (for example, a short Gaussian pulse will

Laser Focus World www.laserfocusworld.cam May 2007 3 5

Increase BeamIntensity WithHard Clad FibersFibergujde Industries, Inc.introduces its new line oflarge core, step-index, andmultimode optical fibers:A Hard Clad Silica instandard and low OHfeaturing a steady stateNumerical Aperture(N.A,) ot 0.39, and HardCoated Silica/Silicia instandard and low OHwith a steady stateN A of 0.22 andattenuation of<10dB/km@850nm.

Features & Benefits:

• Higti concentricityand core-to-cladratio for excellentconnection alignment

• Hard polymercoating for increasedfiber strength andreduced static fatigue inhumid environments

• High core-to-clad ratioto prevent pistoningand provide morestable terminations

• Siocompatibility

Product Applications:

• Laser deiiverysystems

•Medical for laser andsensor technologies

•Shortto mediumdistance datacommunications

• Factory automation•NearHR

spectroscopy• Optical pyrometry• Automotive cabling

Contact FlberguMe today and askfor our now CD-ROM. It's freo!

fiberguide*^ industries

THE FIBER OPTIC SOLUTIONS COMPANY

1 Bay StreetStirling, NJ 07980Telephone: 908-647-6601Fax: 908-647-8464info@fiberguide.com

Cavity-dumpedoscillator

Grism stretcher

10-pass,cryo-cooled

amplifier

20 cm SFIBglass(compressor)

20 cm SF18 glass(compressor)

Intensity (a.u.)

1.0

0.2

0.0-100 100

FIGURE 2. In a DPA laser system a grismstretcher and compression through 120 cm ofglass (top) produces a 35 fs pulse (bottom).

simply become a longer Gaussian pulse),and if it sees TOD, the pulse is asymmet-rically distorted with a series of prepuls-es or postpulses. Group-delay dispersionand TOD can be either positive or nega-tive in sign; for example, a system withpositive GDD {like glass) has a longeroptical path length for blue wavelengthsthan red wavelengths. After a pulse trav-els through a positive-sign GDD system,the only way to recover the original shortpulse is to propagate through anothersystem with GDD of identical magnitudebut negative sign (having a longer pathlength for red wavelengths than bluewavelengths). Sending a pulse through aseries of opposite-signed dispersive sys-tems is the essence of femtosecond pulsecompression.

For ultrashort pulses, especially thosethat have suffered a great deal of disper-sion, it is necessary not only to compen-sate for the GDD, but aiso the TOD. If leftuncompensated, leftover TOD will pre-vent the pulse from being compressed toits ideal, transform-limited duration.

The magnitude and sign of GDD andTOD of materials is determined by thewavelength dependence of the refrac-tive index and the amount of materialthrough which the pulse propagates.For all optical materials in the 700-to-1100 nm wavelength range, GDD isalways positive in sign, and TOD is al-

ways positive in sign. There are noexceptions to this rule among optical glasses and common crystal-line materials (Ti:sapphire, TeO ,LiSAF, YAG, and so on.); when-ever a pulse propagates throughany optical material, it alwaysaccumulates positive GDD andpositive TOD.

To compensate for material dis-persion, it is therefore necessaryto find a pulse compressor withnegative GDD and negative TODin the same proportions. Can thisbe done? The most common pulsecompressor is the Treacy gratingpair, which was shown in 1969 tohave negative GDD. The exact dis-persion of the grating-pair com-pressor is determined by severalfactors (grating line density, inci-dence angle, grating separation),but it is straightforward to provethat, for any configuration of aTreacy compressor, the GDD is al-

ways negative and the TOD is alwayspositive. There is no way to arrange agrating pair to provide negative GDDand negative TOD, which would benecessary to compensate for materialdispersion.

Another commonly used pulse com-pressor is a sequence of prisms, whichprovide negative GDD in a variety ofapplications. In addition, a prism paircan be arranged (by choosing the prop-er prism materials and usage geometry)to have negative GDD and negativeTOD, ideal for compensating material.For example, the well-known 6 fs pulsesproduced at Bell Labs in the 1980s werecompressed using a prism sequence toremove excess TOD. However, the use-fulness of prisms is limited: in general,compensation of a single centimeterof material requires a prism spacingof about 10 cm. As the material pathlengths in the system become long (asin the case of a fiber-delivery systemor a regenerative amplifier) a prism se-quence becomes entirely impractical formaterial-dispersion compensation.

In the early 1990s P. Tournois demon-strated a method for compensating tensof meters of material, which was furtherrefined by two of us (Steve Kane andJeff Squier).'' ^ Tournois showed that aTreacy-style pulse compressor using tra-ditional grisms—transmission gratings

stretcher:lOmSMF

V_ Fiber laser60 MHz

1000 1040

Wavelength (nm)

engraved directly ontothe surface of pr i sms-would yield a dispersivesystem with zero TOD.In 1995 we demonstratedexperimentally that trans-mission grisms could bedesigned and fabricatedto give large amounts ofnegative GDD and nega-tive TOD in the properproportions for compen-sating material dispersion;we used these grisms tocompress 800 nm, 130 fspulses after propagationthrough lOfl m of fiber.

Tliis early demonstra-tion of material-disper-sion compensation usinggrisms was interesting,but not of great practical significance. Thetransmission grisms were very inefficient{25% or less transmission per pass), andit was easy to show that the required ge-ometry for material-dispersion compen-sation was totally incompatible with theefficiency constraints of diffraction grat-ings. 71ie hope of using traditional grismsfor dispersion compensation and pulsecompression quickly faded.

Grisms revisited1 he need tor a material-dispersion com-pensator continued to increase, however,especially with the growth of ultrafastfiber lasers. Ten years after the originalgrism experiments, we began to inves-tigate additional design parameters thatmight overcome the shortcomings of theoriginal grisms and allow us to produce apulse compressor with the proper GDD,TOD, and diffraction efficiency for prac-tical use. 'l"he result was a grism consist-ing of a reflection grating mounted to aprism, with a critical design differenceas compared to traditional grisms. Afterdiffracting from the grating, the pulsepropagates through the prism material,and then refracts out of the prism at anair-glass interface that is tilted by a largeangle with respect to the grating. Thisnew parameter—the angle of the prismwith respect to the diffracted beam—al-lows us to design reflection grisms atnear-Littrow usage, which is the idealcondition for high diffraction efficiency.

By carefully optimizing the grism pa-rameters, we have been able to produce

Yb-dopedLMA PCF

Grism 2Pumplaser18W

CompressorMirror

Grism 1

T F W H M = 1 2 O I S

-600 0 600

Delay (fs)

-600 0 600

Delay (ts)

FIGURE 3. An experimental setup with a Yb-doped large-mode-area photonic-crystal fiber (LMA PCF) includesa 10 m length of single-mode fiber (SfvlF) as a stretcher{top left). The output passes through a grism compressor(top right), resulting in amplified femtosecond pulses(spectrum, bottom left).Pulse compression using grism-paircompressors (bottom right) is more effective than whenusing a grating (bottom center).

pulse-compression grisms with provengroove densities {600, 1200, 1480 lines/mm) and standard prism geometriesand materials. A well-designed grismcan exhibit 90% efficiency, which iscompetitive with the best pulse com-pression-grating specifications on themarket. A patent application has beenfiled for the use of these new grisms ascompensators for material dispersion.*

Applications at 800 nmA novel technique tor producing veryhigh-average-power amplified pulses isdownchirped pulse amplification (DPA),introduced in 2004. In DPA, pulses arestretched by a grating-based delay line,amplified in Tirsapphire, and then com-pressed by propagating through 1 to 2 mof dispersive glass.' Contrary to con-ventional chirped-pulse amplification, aDPA system relies on the negative GDDof gratings to stretch the pulse, and thepositive GDD of material to compress it.For relatively modest pulse energies, thistechnique yields extremely high through-put from the compressor, reduces spatialchirp, and is exceedingly simple to align.To produce ultrashort amplified pulses,however, it is necessary to have a stretch-er whose GDD and TOD are perfectlymatched to the dispersive material in thecompressor. Reflection grisms are a natu-ral fit for this application.

Using 600 lines/mm gratings andoff-the-shelf BK7 prisms as stretchergrisms, pulses as short as 35 fs wereobtained af^er compression through

OPTICAL POWER METERS' Interchangable

Photodiode or

Thermal Sensors

' Digital & Analog

Displays

' Computer

Interface

OPTOMECHANICS

• Lens Holders

• Mirror Mounts

• Cage Systems

• Rail Systems

• Stainless Steel

Posts

LIGHT SOURCES• Tunable Lasers

• Laser Diodes

• Rber Coupled tLasers

• FemtosecondRber Lasers

POLARIZATIONTest & Measurement

Systems

• Deterministic

Polarization

Controllers

• PMD

Measurements

NANOPOSITtONING

• Ptezo ElectricControl

• 6-AxisNanopositioning

hungry for yourthoughts...

Share your product ideaswww.iiiuiiau:>.uuiii

120 cm of SF18 glass (see Fig. 2). Lossesin the compressor are limited by surfacereflections, which can be reduced byantireflection coatings to levels an orderof magnitude more favorable than con-ventional compressors, Reflection-grismpairs are extremely compact—the two-grism system was less than 7 cm longand compensated nearly 200,000 fs ofGDD and 120,000 fs of TOD.

Fiber applications at 1030 nmIn Yb-doped liber-laser systems, an ultra-short pulse propagates through a verylarge amount of dispersive fiber duringall stages of the generation and amplifica-tion process. Recently we demonstrated

a chirped-pulse-amplification system us-ing a 10 m fiber stretcher and a Yb-dopedfiber amplifier, with reflection grismsas the pulse compressor.^ These grismswere specifically designed to compensatethe large amounts of TOD present at1030 nm, and were able to fully compen-sate for 10 m of fiber (with a grism spac-ing of only 1.2 cm). This system was gain-narrowing-limited to 120 fs pulses, andthe extremely clean interferometric auto-correlations indicate full GDD and TODcompensation by the grisnis (see Fig. 3).These initial results have encouraged usto build a high power CPA system with a400 m fiber stretcher; these experimentsare planned for spring 2007.

In addition to fiber-laser applications,reflection grisms will be very useful inapplications (such as multiphoton mi-croscopy) that would benefit from fiberdelivery of femtosecond pulses. LI

REFERENCES1. p. Tournois, Electronics Lett. 29. 1414 (1993).2- S, Kane and J. Squier, IEEE i. Quantum Elec-

tron, 31, 2052 (1995).3. S. Kane, "Grating with angled output prism

face for providing wavelength dependentgroup delay," US Patent Application (2006).

4. H.C. Kapteyn and SJ. Backus, "Downchirpedpulse amplification," U.S. Patent 7,072,101,July 4, 2006.

5. D. Caudiosi et al., Optics Express 14, 9277{2006).

6. L. Kuzneisovaeta\..Advaticed Solid-statePhotonics 2007. Vancouver, Canada, paperTuB3 (2007).

Beam Steering / Active Optics

PIFOC^ Objective Scanners• Millisecond Step Time• Travel to 460 pm• Resolution <1 nm

Prezo Steering Mirrors• NanoRadian Resolution• Microsecond Response• Coplanar Axes

Image Dithering Stages• XY Piezo Scanners• Sub-Millisecond Response• Compact OEM Designs

1I\Pie ro N a n o Positioning

PI is the Global Leaderin lUanopositioning.ISO 9001 since 19S4.30+ Years Experience.Custom Designs.Global Expert Support.Get the 500 page PI catalog'

PI [Physik Instrumente] L.RTel: (508) 832-3456www.pi.wB/lfwbs ^

Moving the isianoWorldjwww.pi.ws See US at Nanotech, Booth #601

- tunableall new series of BBO-OPOs

205 nm - >2500 nm available

fit most Nd:YAG lasers

many options and configurations

GWU-LasertechnikFon +49 . (0) 22 35 . 9 55 22-0Fax +AQ . (0)22 35.7 56 32info@gwu-group.de

www-gwu-group.de/laser