89_RSIjag

Gain-switched,all-acousto-optic,femtosecond pulse ampli?er Julie A.Gruetzmacher,a),b)Matthew A.Horn,b),c)Bret N.Flanders,d)Xiaoming Shang,e)

and Norbert F.Scherer f)

Department of Chemistry,James Franck Institute,and Institute for Biophysical Dynamics,

University of Chicago,5735S.Ellis Avenue,Chicago,Illinois60637

?Received12May2003;accepted19June2003?

The design and performance of a gain-switched,all acousto-optic?AO?Ti:Sapphire regenerative laser ampli?er is presented.An AO Bragg cell is used to send pulses into and out of the ampli?er cavity,and an AO modulator serves as an active isolation device.Pumping the high-Q ampli?er with

a short duration??40ns?532nm pulse allows gain-switched operation in which the seed pulse

dominates the ampli?ed spontaneous emission;no Q switch is required.The ampli?ed pulse energy is?110?J at a4kHz repetition rate,and the compressed pulse duration isр50fs.Detailed measurements are reported demonstrating that this design facilitates low-noise operation.?2003 American Institute of Physics.?DOI:10.1063/1.1619581?

Titanium-doped sapphire has been the workhorse laser and ampli?er medium of ultrafast spectroscopy since the rec-ognition of its suitability in the1990’s.The latest generation of ampli?ed Ti:Sapphire laser systems produces terawatt peak powers,generating sub-20fs pulses with energies in the millijoule range1–3at multikilohertz repetition rates.4The de-velopment of such systems has been driven by ultrafast high-harmonic/x-ray generation,high-?eld measurements,and now attosecond science.5While many can?t on a‘‘table-top,’’most are multistage systems with sophisticated pulse compression schemes.Conversely,the technological require-ments for ultrafast electronic spectroscopy6,7are decidedly more modest:the laser system should be as simple and as ?exible as possible,generate suf?cient energy to perform basic frequency conversion processes,maintain short pulse durations,and provide low-noise operation.

This note presents a unique Ti:Sapphire regenerative am-pli?er that attains these characteristics.The design is mod-eled after a cavity-dumped oscillator,8combining the sim-plicity of that cavity design with the?exibility of regenerative ampli?cation.Gain-switched operation and use of acousto-optic?AO?devices minimize the number of opti-cal elements inside the ampli?er cavity and actively isolate the ampli?er from the seeding laser.The ampli?er is pumped with a low-noise diode-pumped,frequency-doubled Nd:YAG laser.The system provides modest per pulse energies??100?J?and operates at a4kHz repetition rate so that signal averaging techniques,such as lock-in ampli?er detection, can be utilized in experiments.The compressed pulses have energies suf?cient for white-light continuum generation and frequency doubling,permitting pump–probe studies of samples absorbing from the near-UV to the near-IR with sub-50fs time resolution.

Two AO devices are employed.First,an AO Bragg cell replaces the commonly used Pockels cell/thin-?lm polarizer combination to inject the seed pulse into and eject the am-pli?ed pulse from the cavity.This con?guration presents sev-eral advantages.9,10Re?ection losses in the ampli?er are minimized since the Bragg cell is used at Brewster’s angle. Furthermore,dispersion is reduced since a typical Bragg cell has a shorter material path length and is made of fused silica, a low refractive index material.Second,an AO modulator supplants the conventional passive isolation scheme—a Far-aday rotator and polarizers—for directional rather than polarization-based discrimination of incoming and outgoing beams.The AO modulator is composed of a?int glass?SF8?that has lower and more well-determined dispersion than the terbium gallium garnet used in Faraday rotators?see EPAPS material11for a table comparing material dispersion in differ-ent ampli?er designs?.

The use of a diode-pumped,frequency-doubled, Nd:YAG laser as a pump laser further optimizes the ampli-?er.Shot-to-shot noise speci?cations of?1%can be achieved by diode-pumped systems.In addition,the short duration pulses??40ns?from diode-pumped lasers create rapid gain buildup that facilitates gain-switched,rather than Q-switched,ampli?er operation.While a Q switch sup-presses ampli?ed spontaneous emission?ASE?,it also intro-duces re?ection losses inside the ampli?er,thereby reducing the ef?ciency.Gain-switched operation also affords a further reduction in material dispersion;this design leaves only two transmissive optics in the cavity.

A schematic of the laser system is shown in Fig.1.The seed pulse for the ampli?er is provided by a four-mirror Ti:Sapphire oscillator?KM Labs TS-Laser?pumped by a frequency-doubled Nd:YVO4laser?Spectra Physics Millen-nia?.The oscillator produces a91MHz train of22fs pulses with spectra centered at800nm.The pulses?rst pass through an all-re?ective,single-grating stretcher with gold-coated optics.The gratings in the stretcher and compressor

a?Current address:Argonne National Laboratory,Argonne,IL60439.

b?These authors contributed equally to this work.

c?Current address:Utah Valley State College,Orem,UT84058.

d?Current address:Oklahoma State University,Stillwater,OK74078.

e?Current address:Columbia University,New York,NY10027.

f?Author to whom correspondence should be addressed;electronic mail:

nfschere@https://www.sodocs.net/doc/c82934010.html,

REVIEW OF SCIENTIFIC INSTRUMENTS VOLUME74,NUMBER11NOVEMBER2003

4961

0034-6748/2003/74(11)/4961/3/$20.00?2003American Institute of Physics

have groove densities of 600mm ?1and are blazed at 13.9°?Richardson Gratings ?.The stretcher and compressor designs are similar to those previously reported by Wynne et al.,12except that paraboloidal,rather than spherical mirrors,are used to eliminate spherical aberrations,and both are operated in an off-Littrow angle con?guration to reduce clipping of the beam.A 90°roof re?ector ?PLX Inc.?is employed to reduce vertical spatial chirp by minimizing the deviation of the re?ection angle from the grating normal.The stretched pulse length is ?90ps.

After passing through the stretcher,the beam is gently focused into the AO modulator ?Intra Action AOM 125?.The diffraction ef?ciency of the AO modulator is 70%at full rf power.The gating window is ?75ns long and is limited by the acoustic transit time across the diameter of the focused laser spot.The diffracted beam is collimated with a lens,identical to the focusing lens that mode-matches the size of the seed beam to that of the ampli?er.

The regenerative ampli?er cavity design is similar to that of a cavity-dumped oscillator,8but lacks a prism sequence.9,10All mirrors in the ampli?er are broadband,low-dispersion dielectrics ?CVI TLM2-800?.The Bragg cell is placed at the focus of the M4–M5mirror pair terminating the long arm of the cavity.ABCD matrix calculations show that this position allows a tight focus in the Bragg cell,yield-ing ef?cient diffraction and small pre-and post-pulses.8The ampli?er is pumped by a Q -switched,diode-pumped,frequency-doubled Nd:YAG laser ?Lightwave Electronics 210G ?that produces pulses ?40ns in duration.The pump pulse energy at 4kHz is ?525?J.

The timing sequence of the rf and laser pulses is de-picted in Fig.2.The master frequency needed for synchro-nization of the ampli?er electronics is established by the os-cillator;light leaked through the high re?ector end mirror is monitored by a fast photodiode.The repetition rate of the ampli?er is set by a timing system ?Camac TS 2004?that independently triggers the pump laser,AO modulator,and Bragg cell driver ?Camac CD 5000?with rf power ampli?er ?Camac PB 1800?.Independent phase control of the injection and ejection rf pulses sent to the Bragg cell is important for

obtaining maximum ampli?ed power and is achieved with a home-built circuit.In addition,properly timing the arrival of the pump laser pulse relative to seed pulse injection allows suppression of ASE.The pulse is ejected from the ampli?er cavity after the maximum gain is achieved—in this case,after ten round trips.It is spatially separated from the incom-ing seed pulses as it passes through the AO modulator;on this pass,the modulator is off and the beam is therefore not de?ected,allowing 100%of the ampli?ed pulse energy to proceed to the

compressor.

FIG.1.Schematic of all-AO regenerative ampli?er.M1—?at mirror.M2–M4—20cm concave radius of curvature ?ROC ?mirror.M5—10cm con-cave ROC mirror.TS—Ti:Sapphire crystal.B—Bragg cell.L —50cm focal length ?f.l.?lens.PD—photodiode.AOM—AO modulator.G—grating.P—171/2in.f.l.parabolic mirror.RR—90°roof re?ector.LP—Nd:YAG pump laser.Nd:YVO 4—Nd:YVO 4pump laser.OSC—Ti:Sapphire oscilla-

tor.

FIG.2.Relative timing sequence for ampli?er

operation.

FIG.3.?A ?Ampli?ed pulse pro?le in the ampli?er cavity.?B ?Pro?le of ampli?ed pulse with contrast ratio of 37:1?main to post-pulse ?.Data is normalized to a maximum of one.

4962Rev.Sci.Instrum.,Vol.74,No.11,November 2003Gruetzmacher et al.

Light leaked through the dog-leg fold mirror is moni-tored by a fast photodiode;the intracavity pulse train is shown in Fig.3?A ?.An ejection ef?ciency of 65%is ob-served in this case;ef?ciencies as high as 70%have been obtained.The contrast ratio between the main pulse and pre-/post-pulse is typically у35:1,as shown in the photodiode trace of an ampli?ed pulse in Fig.3?B ?.The ejected pulse energy of more than 110?J represents ?30%gain extraction of the absorbed pump pulse energy and a gain factor of ?3?104.

Ampli?ed pulses then pass through a pair of BK7prisms after the compressor for readily tunable compensation of group velocity dispersion and third-order dispersion ?TOD ?imparted by the stretcher and materials in the ampli?er.This strategy is feasible due to the reduction of material disper-sion by the all-AO design.For more dispersive ampli?er designs,the required prism separation would be several meters long,necessitating the use of grating angle adjust-ments in the compressor for TOD minimization.The ?nal pulse energy after recompression exceeds 30?J.The ampli-?ed pulses were characterized by self-diffraction frequency-resolved optical gating,13yielding durations of ?50fs.

To characterize the intensity stability of the ampli?er,the pulse intensity distribution was measured with a fast photo-diode and a gated photon counter with a discriminator ?Stan-ford Research Systems SR400?.8The total number of pulses counted at each discriminator voltage yielded a distribution of pulse intensities;distributions for both the Lightwave and ampli?ed pulses were measured.The counts are ?t to the

integral of a normal distribution,as shown in Fig.4.The deviation of the counts from the mean is ?t to the function:

F ?x ??N erfc

?x ?a

b

?

,?1?

where erfc is the complementary error function,N is the maximum number of counts,a is the mean,and b is the standard deviation.The standard deviation is 1.0%for the Lightwave pump laser and 0.5%for ampli?ed pulses.That the ampli?er shows less pulse intensity noise than the pump source suggests that saturation is achieved.The noise ?gure of the ampli?er exceeds the stability of commercially avail-able Ti:Sapphire regenerative ampli?ers,which typically specify root-mean-square energy stabilities of ?1%.14The authors thank Dr.Andreas-Neil Unterreiner ?Univer-sita

¨t Karlsruhe ?and Dr.Yish-Hann Liau ?Industrial Technol-ogy Research Institute ?for assistance.This work was sup-ported by grants from the NIH ?Grant No.RO1-GM57768?and NSF ?Grant Nos.CHE93-57424and CHE03-17009?.

1

E.Zeek,R.Bartels,M.M.Murnane,H.C.Kapteyn,S.Backus,and G.Vdovin,Opt.Lett.25,587?2000?.2

K.Yamakawa and C.P.J.Barty,IEEE J.Sel.Top.Quantum Electron.6,658?2000?.3

Z.Cheng,F.Krausz,and C.Spielmann,https://www.sodocs.net/doc/c82934010.html,mun.201,145?2002?.4

S.Backus,R.Bartels,S.Thompson,S.Christensen,H.Kapteyn,and M.Murnane,in Ultrafast Phenomena XIII ,edited by M.Murnane,N.F.Scherer,https://www.sodocs.net/doc/c82934010.html,ler,and A.M.Weiner ?Springer,Berlin,2003?,p.128.5

M.Drescher,M.Hentschel,R.Kienberger,M.Uiberacker,V .Yakovlev,A.Scrinzi,T.Westerwalbesloh,U.Kleineberg,U.Heinzmann,and F.Krausz,Nature ?London ?419,803?2002?.6

L.D.Book,A.E.Osta?n,N.Ponomarenko,J.R.Norris,and N.F.Scherer,J.Phys.Chem.B 104,8295?2000?.7

M.A.Horn,J.A.Gruetzmacher,J.Kim,S.-E.Choi,S.M.Anderson,K.Moffat,and N.F.Scherer,in Femtochemistry IV ,edited by F.deShryver,S.De Feyter,and G.Schweizter ?Wiley-VCH,Weinheim,2001?,p.381.8

Y .-H.Liau,A.N.Unterreiner,D.C.Amett,and N.F.Scherer,Appl.Opt.38,7386?1999?.9

A.J.Ruggerio,N.F.Scherer,G.M.Mitchell,G.R.Fleming,and J.N.Hogan,J.Opt.Soc.Am.B 8,2061?1991?.10

T.B.Norris,Opt.Lett.17,1009?1992?.11

See EPAPS Document No.E-RSINAK-74-035311for a table comparing material dispersion in electro-and AO ampli?ers.A direct link to this document may be found in the HTML reference section of the online article.The document may also be reached via the EPAPS homepage ?https://www.sodocs.net/doc/c82934010.html,/pubservs/epaps.html ?or from https://www.sodocs.net/doc/c82934010.html, in the directory/epaps.See the EPAPS homepage for more information.12

K.Wynne,G.D.Reid,and R.M.Hochstrasser,Opt.Lett.19,895?1994?.13

K.W.DeLong,R.Trebino,and D.J.Kane,J.Opt.Soc.Am.B 11,1595?1994?.14

For example,https://www.sodocs.net/doc/c82934010.html,/products/pdfs/spit?re50fs.pdf ?Spectra-Physics ?,https://www.sodocs.net/doc/c82934010.html,/downloads/RegA ?9000?9500?DS.pdf ?Coherent ?,http://64.227.154.50/Scienti?c/CPA2001.htm ?Clark-MXR ?

.

FIG.4.Noise characterization of ampli?ed and Lightwave pump pulses.Circles:Lightwave counts.Triangles:ampli?er counts.Lines are ?ts of the deviation from the mean ?as given by Eq.?1??.Solid:Fit to ampli?ed pulse data.Dashed:fit to Lightwave pulse data.

4963

Rev.Sci.Instrum.,Vol.74,No.11,November 2003Notes