Single-pass high-harmonic generationat 20.8 MHz repetition rate

Andreas Vernaleken,1,* Johannes Weitenberg,2 Thomas Sartorius,3 Peter Russbueldt,3 Waldemar Schneider,1

Sarah L. Stebbings,1 Matthias F. Kling,1 Peter Hommelhoff,1 Hans-Dieter Hoffmann,2 Reinhart Poprawe,2,3

Ferenc Krausz,1 Theodor W. Hänsch,1 and Thomas Udem1

1Max-Planck-Institut für Quantenoptik, Hans-Kopfermann-Straße 1, 85748 Garching, Germany2Lehrstuhl für Lasertechnik, RWTH Aachen University, Steinbachstraße 15, 52074 Aachen, Germany

3Fraunhofer-Institut für Lasertechnik, Steinbachstraße 15, 52074 Aachen, Germany*Corresponding author: andreas.vernaleken@mpq.mpg.de

Received June 14, 2011; accepted August 3, 2011;posted August 5, 2011 (Doc. ID 148949); published August 29, 2011

We report on single-pass high-harmonic generation (HHG) with amplified driving laser pulses at a repetition rate of20:8MHz. An Yb:YAG Innoslab amplifier system provides 35 fs pulses with 20W average power at 1030nm afterexternal pulse compression. Following tight focusing into a xenon gas jet, we observe the generation of high-har-monic radiation of up to the seventeenth order. Our results show that state-of-the-art amplifier systems have becomea promising alternative to cavity-assisted HHG for applications that require high repetition rates, such as frequencycomb spectroscopy in the extreme UV. © 2011 Optical Society of AmericaOCIS codes: 140.7090, 140.7240, 190.4160, 190.7110, 320.5520, 320.7110.

Among the different experimental techniques to createspatially and temporally coherent extreme UV (EUV) ra-diation, high-harmonic generation (HHG) [1] in gaseousmedia with an intense driving laser is highly attractivedue to the comparably simple, table-top setup. Giventhe technical challenge of generating the high peak inten-sities of more than 1013 W=cm2 required to drive the HHGprocess directly with the output of a femtosecond oscilla-tor, the most commonly used systems for HHG to datehave been Ti:sapphire-based chirped pulse amplificationsystems [2]. Keeping the average power roughly constant,these increase pulse energy at the expense of repetitionrate and, therefore, typically operate in the kilohertz re-gime. Only recently, rapid progress in fiber amplifier tech-nology has enabled HHG at 500 kHz [3] and 1MHz [4].However, a number of applications require, or at leasthighly benefit, from EUV sources with repetition ratesin the multimegahertz (multi-MHz) regime, such as fre-quency comb spectroscopy in the EUV [5].The most successful technique to provide laser-

generated multi-MHz EUV radiation so far has beencavity-assisted HHG [6–9], where a passive broadbandenhancement resonator boosts the available averagepower at the full repetition rate of the oscillator, and highharmonics are generated in an intracavity gas target. Sig-nificant progress has been made since the first proof-of-principle experiments [6,7], and a cavity-assisted systemproducing ∼100 μW in a single harmonic order in the EUVfor the first time has recently been reported [10]. At thispower level, applications such as frequency metrology inthe EUV come into reach [5]. However, further progressbecomes increasingly difficult with average intracavitypowers in the kilowatt regime, where damage and non-linear effects of the mirrors and intracavity ionization dy-namics limit the cavity performance [10]. The only otherdemonstrated multi-MHz high-harmonic source to daterelies on plasmonic field enhancement in nanostructures[11], but cannot compete yet with cavity-assisted sourcesin terms of conversion efficiency and EUV beam qualityand has yet to prove its scalability and long-term stability.

In this Letter, we show in a proof-of-principle experi-ment that, owing to the rapid development of high-powersolid-state amplifier systems for ultrashort pulses, it hasnow become possible to generate high harmonics at tensof megahertz repetition rate in a comparably simple sin-gle-pass geometry. Apart from practical convenience, theomission of an enhancement cavity has the following im-portant advantages concerning both the versatility andthe scalability of the EUV source. (i) Single-pass HHGis, by far, less sensitive to nonlinear phase shifts due tothe mirrors, the gas target, or the output coupler, which isrequired to extract the EUV radiation from the enhance-ment cavity. (ii) The efficiency of HHG can be improvedby using shorter driving pulses and specially designedgas targets together with quasi-phase-matching techni-ques. Whereas the former is more challenging in a reso-nator because of the limited bandwidth of availableprecisely dispersion-compensated cavity mirrors, the lat-ter is not yet possible in a cavity at all. (iii) The wave-length range of high harmonics available for anexperiment is not limited by the output coupling method.(iv) The same single-pass HHG setup works at differentrepetition rates without major modifications. (v) Both therepetition rate and the carrier-envelope offset frequencyare independently adjustable.

A schematic of the experimental setup is shown inFig. 1. A passively mode-locked Yb:KGW oscillator(HighQ femtoTrain) delivering 300 fs pulses with an aver-age power of 2:2Wat a center wavelength of 1030 nm and

Fig. 1. Schematic of experimental setup (VA, variable attenua-tor; OAP mirror, off-axis parabolic mirror). See text for details.

3428 OPTICS LETTERS / Vol. 36, No. 17 / September 1, 2011

0146-9592/11/173428-03$15.00/0 © 2011 Optical Society of America

a repetition rate of 20:8MHz is used to seed an Yb:YAGInnoslab power amplifier, which has been described indetail elsewhere [12,13]. In brief, the amplifier is capableof providing almost transform- and diffraction-limitedpulses with output powers of more than 600W. However,the pulse duration of ∼680 fs at the amplifier output,which is limited by the gain bandwidth of the Yb:YAGcrystal, would be unfavorably long for efficient HHGin a rare gas due to ionization of the medium [14]. There-fore, the amplified pulses are spectrally broadened andtemporally compressed. The input power to the compres-sion stage is currently limited to ∼40W by the damagethreshold of the large-mode-area photonic crystal fiber(LMA PCF, NKT Photonics LMA-35) that is employedfor spectral broadening. Self-phase modulation in anLMA fiber of 7:1 cm length broadens the amplifier outputspectrum to Δλ≃ 80 nm, where Δλ is determined at the1=e points of the outermost peaks of the spectrum (seeFig. 2). The pulses are then compressed using custom-made chirped mirrors in a multipass configuration with13 bounces introducing a group delay dispersion of ap-proximately −450 fs2 per bounce. Both the measuredautocorrelation trace (red solid curve) and the autocor-relation trace calculated from the spectrum assuming aflat spectral phase (blue dashed curve) are shown inFig. 3. A FWHM pulse duration of τp ¼ 35 fs can be re-trieved from the measured autocorrelation. The moreprominent sidelobes in the measured autocorrelationtrace compared to the calculated one probably indicatethat higher-order dispersion is not efficiently compen-sated for by the mirrors used. At a compression factorof ∼20, the overall transmission of the pulse compressionunit is about 60%, resulting in an average power of up to23W after compression. Thus, the pulse energy is limitedto ∼1 μJ, and tight focusing is required to reach a peak

intensity sufficient for HHG. We thus magnify the IRbeam with a 1∶2 telescope to a spot size (1=e2 intensityradius) ofw ¼ 2:4mm and then use a gold-coated 90° off-axis parabolic mirror with an effective focal length of20mm to focus the pulses into a noble gas target insidea vacuum chamber. With opened chamber, the beam pro-file can be analyzed and optimized by imaging the focuswith a CCD-based beam profiler. From the measuredwaist radius of w0 ¼ 4:6 μm, a peak intensity of 4 ×1013 W=cm2 at the focus can be estimated. Plasma gen-eration in air is observed and can be exploited for furtheroptimization of the alignment. For HHG, the interactionchamber is evacuated, and a continuous Xe gas jet isproduced in the vicinity of the focus by a Lavalnozzle (orifice L≃ 150 μm), which is operated at a back-ing pressure of 1–2 bars. Motorized translation stages al-low for fine adjustment of the nozzle position. Thegenerated harmonic radiation propagates into a differen-tially pumped grazing incidence EUV monochromator(McPherson Model 248/310), which is separated from theinteraction chamber by a 2mm diameter pinhole. Themonochromator is equipped with a 133:6 grooves=mmgrating and a solar blind channel electron multiplier tospectrally resolve and detect the harmonic radiation withvirtually no background. The signal from the detector dur-ing the wavelength scan is acquired with an oscilloscopeas a trace of counts versus time. The recorded count rate isaveraged with a moving Gaussian window function andprocessed into a spectrum using the scanning parametersof the monochromator.

The resulting spectrum is shown in Fig. 4. The reddashed lines indicate odd harmonics H11 to H19 of the fun-damental. High-harmonic radiation ranging from the 11th(93:6nm) up to the 17th order (60:6 nm) can be identified.Note that the spectral resolution is limited to about 3 nmby the data acquisition technique. The strongest observedpeak corresponds to the fifteenth harmonic (68:7nm).The count rate of this harmonic is 1 order of magnitudehigher than for the other harmonics. However, the relativeharmonic intensities are of limited significance in our casedue to, for example, the unknown spectral diffractionefficiency of the monochromator grating.

As a consequence of the tight focusing geometry, thegenerated EUV radiation is strongly divergent, thusmaking a determination of absolute photon numbers dif-ficult. However, taking into account the quantum effi-ciency of the detector and the limited transmissionimposed by the geometry of the monochromator, and

980 1000 1020 1040 1060 10800.0

0.2

0.4

0.6

0.8

1.0

Wavelength nm

Inte

nsity

arb.

un.

Fig. 2. Infrared spectrum after broadening due to self-phasemodulation in a 7:1 cm long piece of LMA-35 PCF. The widthof the broadened spectrum is Δλ≃ 80nm.

150 100 50 0 50 100 1500.0

0.2

0.4

0.6

0.8

1.0

Delay fs

Inte

nsity

arb.

un.

Fig. 3. (Color online) Measured (red solid curve) and calcu-lated (blue dashed curve) autocorrelation trace. The retrievedFWHM pulse duration is τp ¼ 35 fs.

40 50 60 70 80 90 1000

500

1000

1500

2000

2500

Wavelength nm

Avg

.cou

ntra

te1

s

40 50 60 70 80 90 100

510

50100

5001000

Wavelength nm

Avg

.cou

ntra

te1

s

Fig. 4. Measured high-harmonic spectrum (inset: logarithmicscale). The red dashed lines indicate odd harmonics of the fun-damental. The green line marks the background. The strongpeak near 70nm corresponds to the 15th harmonic.

September 1, 2011 / Vol. 36, No. 17 / OPTICS LETTERS 3429

assuming a typical diffraction efficiency of < 10% for themonochromator grating, we can give a conservative es-timate of ∼1 nW for the total generated power in the 15thharmonic order.The system used in this proof-of-principle demonstra-

tion is currently limited in performance mainly by thespectral broadening unit, which does not reliably with-stand pulse energies higher than 2 μJ. Consequently, tightfocusing is required for HHG, which leads to a low overallharmonic yield because of the small interaction volume(here, b≃ 130 μm, where b denotes the confocal param-eter of the laser). Moreover, the geometrical phase laginduced by focusing accumulates to π for a harmonic or-der q after a distance Lcoh

q ¼ πb=ð2ðq − 1ÞÞ, so that the ef-ficiency of HHG in a gas target of length L is additionallylimited by the reduced effective source volume due tophase matching when Lcoh

q ≲ L [15], which applies forall observed harmonic orders under our experimentalconditions (e.g., Lcoh

15 ≃ 15 μm ≪ L≃ 150 μm). Whenusing a more robust spectral broadening scheme (e.g.,the one introduced in [16]), at least 1 order of magnitudemore pulse energy should be available for HHG so thatthe size of the focus can be increased to exploit the b3

scaling of the harmonic yield [14]. In addition, differentgas target designs and quasi-phase-matching techniquescan be employed to further improve the conversion effi-ciency for specific harmonic orders. Thus, power levelsof the order of several microwatts seem feasible forfuture multi-MHz single-pass EUV sources.In conclusion, we have demonstrated single-pass HHG

in xenon with amplified laser pulses at an, to the best ofour knowledge, unprecedented repetition rate of20:8MHz. Using a state-of-the-art solid-state amplifiersystem, we generated and observed high harmonics upto the 17th order (60:6nm) at power levels of up to nano-watts. While cavity-assisted multi-MHz HHG sources arestill the benchmark in terms of generated EUV power, webelieve that, following the rapid progress in amplifier de-velopment, single-pass HHG will quickly mature into acompetitive, versatile, and easy to operate table-top high-repetition-rate EUV source with a wide range of applica-tions and improved scalability.

We gratefully acknowledge HighQ Laser InnovationGmbH for providing the seed oscillator, and L. Lötscher

for providing the chirped mirrors and for her help. Wethank C. Gohle, A. Ozawa, and M. Herrmann for helpfuldiscussions. S. L. Stebbings gratefully acknowledges sup-port via a postdoctoral fellowship from the Alexandervon Humboldt Foundation. This work was supportedby the Fraunhofer-Max-Planck cooperation projectKORONA.

References

1. J. G. Eden, Prog. Quantum Electron. 28, 197 (2004).2. D. Strickland and G. Mourou, Opt. Commun. 56, 219 (1985).3. S. Hädrich, J. Rothhardt, M. Krebs, F. Tavella, A. Willner, J.

Limpert, and A. Tünnermann, Opt. Express 18, 20242(2010).

4. J. Boullet, Y. Zaouter, J. Limpert, S. Petit, Y. Mairesse, B.Fabre, J. Higuet, E. Mével, E. Constant, and E. Cormier,Opt. Lett. 34, 1489 (2009).

5. M. Herrmann, M. Haas, U. D. Jentschura, F. Kottmann, D.Leibfried, G. Saathoff, C. Gohle, A. Ozawa, V. Batteiger, S.Knünz, N. Kolachevsky, H. A. Schüssler, T. W. Hänsch, andTh. Udem, Phys. Rev. A 79, 052505 (2009).

6. C. Gohle, Th. Udem, M. Herrmann, J. Rauschenberger, R.Holzwarth, H. A. Schuessler, F. Krausz, and T. W. Hänsch,Nature 436, 234 (2005).

7. R. J. Jones, K. D. Moll, M. J. Thorpe, and J. Ye, Phys. Rev.Lett. 94, 193201 (2005).

8. A. Ozawa, J. Rauschenberger, C. Gohle, M. Herrmann, D. R.Walker, V. Pervak, A. Fernandez, R. Graf, A. Apolonski, R.Holzwarth, F. Krausz, T. W. Hänsch, and Th. Udem, Phys.Rev. Lett. 100, 253901 (2008).

9. D. C. Yost, T. R. Schibli, and J. Ye, Opt. Lett. 33, 1099 (2008).10. J. Jones, in High Intensity Lasers and High Field Phe-

nomena, OSA Technical Digest (CD) (Optical Society ofAmerica, 2011), paper HFB5.

11. S. Kim, J. Jin, Y.-J. Kim, I.-Y. Park, Y. Kim, and S.-W. Kim,Nature 453, 757 (2008).

12. P. Russbueldt, T. Mans, G. Rotarius, J. Weitenberg, H. D.Hoffmann, and R. Poprawe, Opt. Express 17, 12230 (2009).

13. P. Russbueldt, T. Mans, J. Weitenberg, H. D. Hoffmann, andR. Poprawe, Opt. Lett. 35, 4169 (2010).

14. A. L’Huillier, K. J. Schafer, and K. C. Kulander, J. Phys. B 24,3315 (1991).

15. A. L’Huillier, M. Lewenstein, P. Salières, P. Balcou, M. Y.Ivanov, J. Larsson, and C. G. Wahlström, Phys. Rev. A48, R3433 (1993).

16. M. Nisoli, S. D. Silvestri, and O. Svelto, Appl. Phys. Lett. 68,2793 (1996).

3430 OPTICS LETTERS / Vol. 36, No. 17 / September 1, 2011