Spectral control of high-harmonic generation via drive laser pulse shaping in a wide-diameter capillary

We experimentally investigate spectral control of high-harmonic generation in a wide-diameter (508 μm) capillary that allows using significantly lower gas pressures coupled with elevated drive laser energies to achieve higher harmonic energies. Using phase shaping to change the linear chirp of the drive laser pulses, we observe wavelength tuning of the high-harmonic output to both larger and smaller values. Comparing tuning via the gas pressure with the amount of blue shift in the transmitted drive laser spectrum, we conclude that both adiabatic and non-adiabatic effects cause pulse-shaping induced tuning of high harmonics. We obtain a fractional wavelength tuning, Δλ/λ, in the range from −0.007 to + 0.01, which is comparable to what is achieved with standard capillaries of smaller diameter and higher pressures. ©2016 Optical Society of America OCIS codes: (140.7240) UV, EUV, and X-ray lasers; (190.4160) Multiharmonic generation. References and links 1. R. L. Sandberg, A. Paul, D. A. Raymondson, S. Hädrich, D. M. Gaudiosi, J. Holtsnider, R. I. Tobey, O. Cohen, M. M. Murnane, H. C. Kapteyn, C. Song, J. Miao, Y. Liu, and F. Salmassi, “Lensless diffractive imaging using tabletop coherent high-harmonic soft-X-ray beams,” Phys. Rev. Lett. 99(9), 098103 (2007). 2. S. Witte, V. T. Tenner, D. W. E. Noom, and K. S. E. Eikema, “Lensless diffractive imaging with ultra-broadband table-top sources: from infrared to extreme-ultraviolet wavelengths,” Light Sci. Appl. 3(3), e163 (2014). 3. D. Descamps, C. Lyngå, J. Norin, A. L’huillier, C. G. Wahlström, J. F. Hergott, H. Merdji, P. Salières, M. Bellini, and T. W. Hänsch, “Extreme ultraviolet interferometry measurements with high-order harmonics,” Opt. Lett. 25(2), 135–137 (2000). 4. P. Balcou, P. Salieres, K. S. Budil, T. Ditmire, M. D. Perry, and A. L’Huillier, “Two-colour time-resolved spectroscopy of helium using high-order harmonics,” Z. Phys. D Atom. Mol. Cl. 34, 107–110 (1995). 5. J. Larsson, E. Mevel, R. Zerne, A. Lhuillier, C. G. Wahlstrom, and S. Svanberg, “2-color time-resolved spectroscopy of helium using high-order harmonics,” J. Phys. B. 28(3), L53–L58 (1995). 6. H. Niikura and P. B. Corkum, “Attosecond and angstrom science,” in Advances in Atomic Molecular and Optical Physics 54, P. R. Berman, C. C. Lin and E. Arimondo, eds. (Elsevier Inc. 2007) pp 511–548. 7. A. Föhlisch, P. Feulner, F. Hennies, A. Fink, D. Menzel, D. Sanchez-Portal, P. M. Echenique, and W. Wurth, “Direct observation of electron dynamics in the attosecond domain,” Nature 436(7049), 373–376 (2005). 8. F. Calegari, D. Ayuso, A. Trabattoni, L. Belshaw, S. De Camillis, S. Anumula, F. Frassetto, L. Poletto, A. Palacios, P. Decleva, J. B. Greenwood, F. Martín, and M. Nisoli, “Ultrafast electron dynamics in phenylalanine initiated by attosecond pulses,” Science 346(6207), 336–339 (2014). 9. M. Labat, M. Bellaveglia, M. Bougeard, B. Carré, F. Ciocci, E. Chiadroni, A. Cianchi, M. E. Couprie, L. Cultrera, M. Del Franco, G. Di Pirro, A. Drago, M. Ferrario, D. Filippetto, F. Frassetto, A. Gallo, D. Garzella, G. Gatti, L. Giannessi, G. Lambert, A. Mostacci, A. Petralia, V. Petrillo, L. Poletto, M. Quattromini, J. V. Rau, C. Ronsivalle, E. Sabia, M. Serluca, I. Spassovsky, V. Surrenti, C. Vaccarezza, and C. Vicario, “High-gain harmonic-generation free-electron laser seeded by harmonics generated in gas,” Phys. Rev. Lett. 107(22), 224801 (2011). 10. E. Takahashi, Y. Nabekawa, T. Otsuka, M. Obara, and K. Midorikawa, “Generation of highly coherent submicrojoule soft x rays by high-order harmonics,” Phys. Rev. A 66(2), 021802 (2002). #254691 Received 30 Nov 2015; revised 8 Jan 2016; accepted 11 Jan 2016; published 20 Jan 2016 (C) 2016 OSA 25 Jan 2016 | Vol. 24, No. 2 | DOI:10.1364/OE.24.001604 | OPTICS EXPRESS 1604 11. V. Tosa, H. T. Kim, I. J. Kim, and C. H. Nam, “High-order harmonic generation by chirped and self-guided femtosecond laser pulses. I. Spatial and spectral analysis,” Phys. Rev. A 71(6), 063807 (2005). 12. V. Tosa, H. T. Kim, I. J. Kim, and C. H. Nam, “High-order harmonic generation by chirped and self-guided femtosecond laser pulses. II. Time-frequency analysis,” Phys. Rev. A 71(6), 063808 (2005). 13. A. Rundquist, C. G. Durfee 3rd, Z. Chang, C. Herne, S. Backus, M. M. Murnane, and H. C. Kapteyn, “Phasematched generation of coherent soft X-rays,” Science 280(5368), 1412–1415 (1998). 14. R. A. Bartels, A. Paul, H. Green, H. C. Kapteyn, M. M. Murnane, S. Backus, I. P. Christov, Y. Liu, D. Attwood, and C. Jacobsen, “Generation of spatially coherent light at extreme ultraviolet wavelengths,” Science 297(5580), 376–378 (2002). 15. C. A. Froud, E. T. F. Rogers, D. C. Hanna, W. S. Brocklesby, M. Praeger, A. M. de Paula, J. J. Baumberg, and J. G. Frey, “Soft-x-ray wavelength shift induced by ionization effects in a capillary,” Opt. Lett. 31(3), 374–376 (2006). 16. T. J. Butcher, P. N. Anderson, R. T. Chapman, P. Horak, J. G. Frey, and W. S. Brocklesby, “Bright extremeultraviolet high-order-harmonic radiation from optimized pulse compression in short hollow waveguides,” Phys. Rev. A 87(4), 043822 (2013). 17. S. J. Goh, Y. Tao, P. J. M. van der Slot, H. J. M. Bastiaens, J. Herek, S. G. Biedron, M. B. Danailov, S. V. Milton, and K. J. Boller, “Single-shot fluctuations in waveguided high-harmonic generation,” Opt. Express 23(19), 24888–24902 (2015). 18. D. Geskus, “Design & construction of regenerative amplifier and compressor for chirped pulse amplification,” M.Sc. dissertation, University of Twente, Dept. of Science & Technology, Laser Physics & Non-Linear Optics, (2006). 19. C. Delfin, C. Altucci, F. D. Filippo, C. Lisio, M. B. Gaarde, A. L’Huillier, L. Roos, and C.-G. Wahlström, “Influence of the medium length on high-order harmonic generation,” J. Phys. At. Mol. Opt. Phys. 32(22), 5397– 5409 (1999). 20. J. Xia, C. Altucci, S. Amoruso, R. Bruzzese, R. Velotta, and X. Wang, “Generation of high energy, 30 fs pulses at 527 nm by hollow-fiber compression technique,” Opt. Express 16(6), 3527–3536 (2008). 21. S. J. Goh, H. J. M. Bastiaens, B. Vratzov, Q. Huang, F. Bijkerk, and K. J. Boller, “Fabrication and characterization of free-standing, high-line-density transmission gratings for the vacuum UV to soft X-ray range,” Opt. Express 23(4), 4421–4434 (2015). 22. P. Tournois, “Acousto-optic programmable dispersive filter for adaptive compensation of group delay time dispersion in laser systems,” Opt. Commun. 140(4-6), 245–249 (1997). 23. G. L. Wang, C. Jin, A. T. Le, and C. D. Lin, “Influence of gas pressure on high-order-harmonic generation of Ar and Ne,” Phys. Rev. A 84(5), 053404 (2011). 24. C. Altucci, T. Starczewski, E. Mevel, C.-G. Wahlström, B. Carré, and A. L’Huillier, “Influence of atomic density in high-order harmonic generation,” J. Opt. Soc. Am. B 13(1), 148–156 (1996). 25. H. J. Shin, D. G. Lee, Y. H. Cha, J. H. Kim, K. H. Hong, and C. H. Nam, “Nonadiabatic blueshift of high-order harmonics from ar and ne atoms in an intense femtosecond laser field,” Phys. Rev. A 63(5), 053407 (2001). 26. X. S. Zhang, A. L. Lytle, O. Cohen, M. M. Murnane, and H. C. Kapteyn, “Quantum-path control in high-order harmonic generation at high photon energies,” New J. Phys. 10(2), 025021 (2008). 27. C. G. Durfee, A. R. Rundquist, S. Backus, C. Herne, M. M. Murnane, and H. C. Kapteyn, “Phase matching of high-order harmonics in hollow waveguides,” Phys. Rev. Lett. 83(11), 2187–2190 (1999). 28. W. M. Wood, C. W. Siders, and M. C. Downer, “Measurement of femtosecond ionization dynamics of atmospheric density gases by spectral blueshifting,” Phys. Rev. Lett. 67(25), 3523–3526 (1991). 29. N. B. D. M. V. Ammosov and V. P. Krainov, “Tunnel ionization of complex atoms and of atomic ions in an alternating electromagnetic field,” Sov. Phys. JETP 64, 1191 (1986). 30. M. C. Chen, P. Arpin, T. Popmintchev, M. Gerrity, B. Zhang, M. Seaberg, D. Popmintchev, M. M. Murnane, and H. C. Kapteyn, “Bright, coherent, ultrafast soft X-ray harmonics spanning the water window from a tabletop light source,” Phys. Rev. Lett. 105(17), 173901 (2010). 31. J. Zhou, J. Peatross, M. M. Murnane, H. C. Kapteyn, and I. P. Christov, “Enhanced high-harmonic generation using 25 fs laser pulses,” Phys. Rev. Lett. 76(5), 752–755 (1996). 32. G. Tempea and T. Brabec, “Nonlinear source for the generation of high-energy few-cycle optical pulses,” Opt. Lett. 23(16), 1286–1288 (1998).


Introduction
High-harmonic generation (HHG) is a nonlinear optical process that provides coherent radiation in the form of ultra-short pulses, covering a broad spectrum including the extreme ultraviolet (XUV).In recent years, HHG has found numerous applications such as in imaging at the nanoscale [1,2], XUV interferometry [3], spectroscopy [4,5] and attosecond science [6][7][8] while future applications are under way, e.g, the spectral control of free-electron lasers via injection seeding [9].All these applications would clearly benefit if available elevated drive laser pulse energies could be utilized to increase the high-harmonic (HH) output [10].However, when using higher drive pulse energies it remains essential to maintain control over the spatial and spectral properties of the generated radiation [11,12].For spatial control, generation of HHG along increased interaction lengths, i.e., using long, thin capillaries, seems particularly promising due to the strong waveguiding of the drive laser pulse [13,14].Additionally, in thin capillaries, spectral control, in terms of wavelength tuning of the various harmonic orders, can be achieved via pulse shaping of the drive laser [15].However, in these standard capillaries, where the inner diameter is small (on the order of 150 µm diameter), HHG suffers from several fundamental limitations.First, the output pulse energies are limited by ionization-induced phase mismatch, created at relatively low drive laser pulse energies, typically beyond a few hundred μJ.Second, the propagation of the drive laser pulse becomes subject to complicated nonlinear propagation effects.Under certain circumstances these, through temporal self-compression, increase the output [16], but otherwise reduce the degree of control.The reduction in control shows up as multimodal drive laser propagation and nonlinear mode coupling caused by ionization-induced defocusing, and can lead to significant spatial shot-to-shot fluctuations [17].
A promising way to circumvent such limitations and drive up the output energies while still maintaining spatial and spectral control would be to generate HHG in a capillary with significantly increased cross section and lower gas pressures, such that mJ-levels of drive laser pulse energy can be applied without inducing undesired nonlinear pulse propagation dynamics.However, such experiments have not been done before, and simple scaling arguments are not straightforward.For instance, in a wide capillary where the waveguidinginduced dispersion is weaker, phase matching occurs at lower gas pressure.This might impose limitations on the spectral control via drive laser pulse shaping, as such a technique involves critical timing and sizing of the ionization-induced blue shift of the drive laser [15].It is therefore important to perform an experimental investigation of the spectral control of HHG in wide-diameter capillaries in order to demonstrate the effectiveness in this new regime of spectral control via drive laser pulse shaping.
Here we present the first investigation of spectral control of HHG through drive laser pulse shaping within a wide-diameter capillary.The results show that, in spite of rather different operational parameters in such capillaries, compared to standard thin capillaries, pulse shaping induced spectral control remains effective and attractive for applications.

Experimental setup
Figure 1 depicts the overall setup used for the experiments.To drive the HHG, we used a pulsed Ti:Sapphire laser with a 795-nm center wavelength, operated at a repetition rate of 1 kHz (Legend Elite Duo HP USP, Coherent Inc.), and with a maximum pulse energy of 6-7 mJ.Infrared (IR) pulse lengths as short as 40 fs were measured via single-shot frequency resolved optical gating (Grenouille) [18].The pulses are close to the Fourier limit (timebandwidth product of 0.5) and the beam parameter, M 2 , is close to unity (M x 2 = 1.1;M y 2 = 1.3).We used a capillary with a wide inner diameter of 508 μm to guide the laser pulse.The 67-mm long waveguide capillary, with an active length of 37 mm [19], is mounted in a vacuum chamber and is filled through two slits with Ar gas to the desired pressure.The capillary has four additional slits for pumping and the configuration is identical to the 150 μm diameter previously used [17].To optimize the injection of the drive laser pulse into the waveguide capillary we need to mode-match the drive laser beam [20] and compensate for the astigmatism of the beam.This can both be realized using a weakly focusing mirror (2.5 m radius of curvature) at a 14 deg angle of incidence.We measure a maximum pump laser throughput of around 60% under optimum alignment conditions of the capillary.We attribute the deviation of the measured maximum transmission from the theoretical maximum of 96% to the non-unity beam parameter of the driving laser pulse and to scattering from the 6 slits.In an equivalent setup, but using a thin capillary of 150 μm diameter having the same slit configuration, we observed a comparable maximum transmission of 50% [17].Using a drive laser energy of 6.9 mJ and estimating 30% losses due to incoupling and scattering at the first gas fill slit (5 mm from the entrance) we calculate a peak intensity of up to 2.1 × 10 14 W/cm 2 in the waveguide.
Following the capillary, we let the high harmonic and drive laser beams co-propagate over a large distance (9 m) allowing the drive laser intensity to be reduced by diffraction.The infrared part of the beam is then blocked by a set of two 200-nm thick Aluminum (Al) filters placed in series.The XUV radiation is then sent for spectral analysis to a transmission spectrometer equipped with home-fabricated, high-line-density gratings [21].The Al filters also act as a bandpass for the XUV radiation with a spectral transmission window from 16 nm to 80 nm.The HHG radiation is detected with an XUV CCD camera (Andor DO240).Absolute pulse energies are determined by taking into account the spectrally dependent camera sensitivity and transmission of Al filters [17].Fig. 1.Schematic of the experimental setup for spectral tuning of high-harmonic generation (HHG) via drive laser pulse shaping in a wide capillary.The drive laser system incorporates an acousto-optic programmable dispersive filter (AOPDF) for phase shaping of the drive laser pulses (HP = horizontal polarization; VP = vertical polarization).The shaped pulses are sent to a Grenouille for measurement of the pulse duration and are then passed through an Ar-filled capillary (508 µm wide) for high-harmonic generation.Spectral tuning of HHG is observed with a transmission XUV spectrometer.
We investigated spectral control through two methods.The first via a variation of the drive laser chirp and the second through a variation of the gas pressure.Drive laser pulse shaping is performed with an acousto-optic programmable dispersive filter (AOPDF, DAZZLER HR-800/T1, Fastlite) [22].The filter is positioned following the laser oscillator in order to prevent optical damage.The polarization of the shaped pulse leaving the AOPDF, having been rotated by 90 deg, is rotated back with a half-wave plate.The pulse is then stretched and injected into the regenerative amplifier and subsequently followed by a singlepass amplifier.After amplification, the pulse is sent to the compressor.Calibration of the phase settings applied by the AOPDF was performed by measuring with the Grenouille the pulse duration of the drive laser after compressor.The experimental error of the pulse measurement using the Grenouille is estimated to be ± 2.5 fs.Although a large variety of different pulse shapes can be provided with the programmable filter, here we restrict ourselves to phase-only shaping that avoids significant loss of pulse energy.In order to concentrate on the most basic effects only, we restrict ourselves to shaping with second-orderparabolic phase spectra.This corresponds to generating drive laser pulses with various degrees of linear frequency chirp.
The spectrum of the transmitted IR laser pulse is measured following exit from the capillary.This allows us to directly compare the wavelength shift of the high-harmonic output with a possible wavelength shift of the drive laser acquired during propagation through the gas-filled capillary under a variety of drive laser chirp and gas pressure conditions.For this purpose, a flip mirror is inserted 2 m after the capillary.This directs the drive laser beam into an IR spectrometer (Ocean Optics 2000).We also analyzed the transmitted IR pulse profile at this location and found no significant features in the mode profile that would indicate the presence of higher-order modes.This means that the driving laser pulse is exciting to good approximation only the fundamental capillary mode.To demonstrate that the high harmonic output can be increased noticeably by using higher drive laser pulse energies in combination with a wider-diameter capillary and lower gas pressures, we measure the harmonic output pulse energy obtained with the 508 µm diameter capillary and compare it with our previous measurement [17] that had used a thin (150 µm diameter) standard capillary.For these measurements, the transmission grating is removed from the XUV beam path.Figure 2 shows the highest attainable total (spectrally integrated from 16 to 80 nm) pulse energy in the harmonic beam averaged over 100 single shots versus pressure for both the small and wide capillary, which requires IR pulses energies of 1.0 mJ and 6 mJ (maximum available), respectively.It can be seen that the highest output of 4.1 nJ at 6 mbar is obtained from the wide capillary as compared to 1.2 nJ at 53 mbar from the standard capillary.This corresponds to a factor of 3.3 increase in the harmonic output from the wide capillary which is why we are interested to investigate also the degree of spectral control that can be obtained at such elevated output.

Experimental results
In Fig. 3(a) we show a series of high-harmonic spectra recorded following generation in the wide capillary and for various different Ar gas pressures.The drive laser was set for the shortest pulse length (40 fs), i.e., drive laser pulses with zero chirp and the maximum available IR pulse energy of 6.9 mJ was used.At this energy the optimum pressure producing maximum XUV energy was found to be also 6 mbar.The spectra are obtained by integrating the HH output over 20 pulses.Harmonic orders, q, between 17 and 29, can be seen as spectral peaks.Each spectrum is normalized to the highest peak and for clarity the different spectra are offset with regard to each other by the same vertical distance.As the pressure increases from 2 mbar to 10 mbar, we observe that the order of the strongest harmonic increases from q = 17 to 27.This observation is consistent with what was reported in [23] by Wang et al., and can be understood as due to the pressure dependence of phase matching and reabsorption of harmonics.A second effect can be identified in the spectra, a wavelength blue-shift (tuning) of all harmonic orders with increasing pressure [24].In the figure we have used the center frequency (determined from a Gaussian least-square fit to each harmonic) of the harmonics at lowest pressure (2 mbar) as reference, and these are indicated as vertical dashed lines.Figure 3(b) plots the corresponding spectra of the drive laser radiation as recorded after interaction with the Ar gas (transmitted through the capillary).The center wavelength of the drive laser spectra, as indicated by the short vertical line, is obtained again with a Gaussian least-square fit.The vertical dashed line indicates the center wavelength of the drive laser at the lowest pressure of 2 mbar.It can be seen that, similar to the harmonic orders, the drive laser central wavelength shows an increasing blueshift as the gas pressure increases from 2 to 10 mbar.
In Fig. 4(a), we show a series of high-harmonic spectra recorded at a fixed gas pressure of 6 mbar but with various different drive laser pulse durations (obtained using the previously determined chirp settings of the programmable filter) and with a constant drive laser pulse energy of 6.9 mJ.The gas pressure at 6 mbar corresponds to the pressure that yielded the highest total pulse energy of the high-harmonic beam.The dashed lines corresponds to the center wavelength of each harmonic as found for a driving laser pulse with zero chirp with a gas pressure of 2 mbar and are at the same wavelengths as in Fig. 3(a).The sign given with the pulse duration (FWHM in fs) corresponds to the sign of the applied laser chirp.The driving pulse energy is kept constant during the measurement and hence its intensity becomes lower as the pulse duration becomes longer.The corresponding drive laser spectra measured following the capillary are given in Fig. 4 center wavelength of the drive laser pulse as measured following an empty capillary, i.e., without interaction with Ar gas.This wavelength is obtained again with a Gaussian leastsquare fit.In Fig. 4(a) we observe harmonics ranging from q = 17 to 29 with the strongest harmonic being the 23rd for most cases.This indicates that, unlike in the measurements for Fig. 3, the phase matching conditions are approximately unchanged during the pulse shaping experiments.We observe a second effect in the spectra and that is a wavelength shift (tuning) of all harmonic orders as a function of the pulse duration.Here, the blue-shift of the harmonic peaks is barely measureable when positive chirp ( + 95 and + 57 fs) is applied to the drive pulses.At zero chirp (40 fs), the blue-shift is clearly visible.However, for negative chirp, the measurements of the harmonic peaks recorded for the −48 fs and −68 fs conditions show an even larger blue-shift than that for zero chirp (40 fs).This is in contrast to the spectrum of the drive laser, Fig. 4(b), which displays the highest blue-shift at minimum (zero) chirp (40 fs).And finally, at a chirp setting of −98 fs, the harmonic peaks show a large red-shift while the drive laser is clearly seen to be blue-shifted.In the following we will discuss the implication of these observations.From Figs. 3 and 4, we clearly observe a shift of the high-harmonic output wavelengths as a function of pressure and drive laser chirp.In comparison with the observations of others [15,25], although based on different experimental parameters (in [15], Froud et al. used a drive laser peak intensity of I p = 1.2 × 10 14 W/cm 2 and an Ar pressure of P = 80 mbar, while in [25], Shin et al. used I p = 9 × 10 15 W/cm 2 and P = 40 mbar), the blue-shift may result from three mechanisms.One of the mechanisms is an ionization induced blue-shift of the drive laser.The ultrafast rise of the free-electron density during the drive laser pulse causes a rapid decrease of the refractive index, which generates a temporally increasing instantaneous frequency, ω L (t) [15].In this case the frequencies of the high-harmonic orders, ω q (t) = qω L (t), become blue-shifted as well.In this sense, the high-harmonic output adiabatically follows the frequency shift of the drive laser.If we denote the blue-shift of the drive laser wavelength as δλ, the corresponding blue-shift of the q th harmonic order is expected to be δλ/q.
A second mechanism, called the non-adiabatic shift [19], occurs due to a non-adiabatic response of the driven medium to the rapidly increasing field strength of the ultrashort drive laser pulse.To understand this mechanism, consider the following where only the short electron trajectories are taken into account as these contribute the most to HHG in a capillary [26].Assume an intensity-dependent phase of the generated high-harmonic field, where the harmonic undergoes a blue-shift during the leading edge of the drive laser pulse and a redshift at the trailing edge.When the drive laser peak intensity is higher than the so-called saturation intensity, high harmonics are generated only in the leading edge.This saturation intensity coarsely corresponds to an intensity when the so-called critical fractional ionization, η cr , is reached, beyond which high harmonics can no longer be phase matched [27].In this case the resultant harmonics will be blue-shifted with the degree of blue-shift dependent on the harmonic order.For short electron trajectories, the degree of blue-shift increases with harmonic order [25].
The third mechanism is related to the degree and sign of the linear chirp of the drive laser pulse.For instance, if the drive laser pulse is positively chirped, the instantaneous frequency grows with time, i.e., the carrier frequency at the leading edge is lower than at the trailing edge.In this case, as discussed before, when the peak intensity is higher than the saturation intensity, high harmonics are generated only at the leading edge of the pulse.Due to the positive chirp the leading edge of the drive pulse is redshifted and so the high-harmonic generation is driven by the longer-wavelength part of the drive laser spectrum.This gives rise to harmonics that are red-shifted.Similarly, with a negatively chirped drive laser pulse we expect a blue-shift of the generated harmonics.
To investigate which of the discussed mechanisms may be responsible for the observed blue-shift in our measurements, we carried out a qualitative analysis via the relative (or fractional) wavelength shifts that the various different harmonics and the drive laser undergo.
Here we define the fractional wavelength shift as δλ q /λ q , where λ q is the reference center wavelength of the q th harmonic defined as the center (peak) wavelengths found via a multipeak Gaussian least-square fit to the measured HH spectrum for a pressure of 2 mbar and a zero-chirped IR drive laser pulse, and these wavelengths are indicated by the dashed lines in Figs.3(a) and 4(a), and δλ q is the difference between the measured center wavelength and the reference wavelength for harmonic q.This definition is chosen for simple comparison with the case that the HHG adiabatically follows the drive laser frequency, ω q = qω L .If the drive laser frequency is shifted by δω L , the harmonics are shifted accordingly, i.e., ω q + δω q = (ω L + δω L ).In terms of wavelength, λ = 2πc/ω, this reveals that δλ q /λ q = -δω L /ω L , independent of the considered harmonic order.
Figure 5(a) shows the measured relative shift of three selected harmonic orders, i.e., the value of δλ q /λ q plotted as a function of gas pressure as extracted from Fig. 3.For the plot we selected the most representative orders, i.e., the lowest, the central and the highest order that were observed (q = 17, 21, and 25) and we also plot the measured shift of the drive laser for which q = 1.A weakly curved, almost linear increase as indicated by the dashed line (linear least-square fit) of the blue-shift with pressure can be seen for the drive laser at q = 1.This observation can be explained by a linearly increasing pressure leading to a linearly increasing electron density which should be approximately proportional to the ionization induced blueshift of the drive laser (see Fig. 3(b)).It should be noted that, at higher pressures, more complicated changes are found in the transmitted IR spectrum [28].As was described before, if the wavelength shift of the high harmonics were just caused by conversion of the frequency-shifted drive laser, then the relative wavelength shift, δλ q /λ q , would be the same for all orders.When looking at Fig. 5(a), however, it can be seen that the relative wavelength shift for q = 17, 21, and 25 is larger than for q = 1, and that the relative shift generally increases with q.This implies that other mechanisms are contributing as well to the overall harmonic blue-shift.In order to check for the possibility that a non-adiabatic blue-shift may have occurred, we use the Ammosov-Delone-Krainov (ADK) model [29].Using this model we calculate, for the peak intensity applied in our experiments (I p = 2.1 × 10 14 W/cm 2 for the unchirped driving laser pulse having an energy of 6.9 mJ), a fractional ionization of η ≈7%.The maximum fractional ionization that still allows phase matching, also called the critical fractional ionization, was calculated to be η cr = 5% for a pressure of 6 mbar.This was determined using the following expression [30], Here, r e is the classical electron radius, N atm is the atomic number density at ambient pressure, and Δn is the difference in the indices of refraction of the gas at the fundamental and high harmonic wavelengths.The condition for observing a non-adiabatic blue shift occurs when η cr < η [19].This is indeed the case in our measurements and is consistent with our observation that the relative shift increases with the order.This agrees with the earlier results from [25], where a non-adiabatic blue-shift is reported as well.It can be concluded that the total observed harmonic blue-shift obtained is due to both ionization (adiabatic) and non-adiabatic effects.
As for the third mechanism discussed before, a positive (negative) linear chirp of the drive laser pulse would, given that the harmonics are generated only at the leading edge of the drive laser, generate a red-shift (blue-shift) in the harmonics.If this mechanism were significant, it would allow for spectral control of the high harmonics via the drive laser chirp.To investigate whether this is the case, we plot the relative wavelength shift (Fig. 5(b)), δλ q /λ q , for the same harmonic orders (q = 17, 21, 25) and the drive laser (q = 1), but now as a function of the drive laser chirp.These data are extracted from Fig. 4 where δλ q is again the difference between the measured center wavelength and the reference wavelength (center wavelength of harmonics at a pressure of 2 mbar and driven by an unchirped IR laser pulse).The shortest pulse duration of 40 fs corresponds to the zero chirp setting of the programmable dispersive filter.
Settings with positive or negative chirp are indicated with a positive or negative sign respectively in front of the measured pulse duration.The pulse energy is constant and therefore we expect that by applying a chirp, whether positive or negative, the drive laser intensity drops and so it should experience less ionization and less blue-shift due to the lower peak intensity.Furthermore the blue-shift from the non-adiabatic response should decrease.Therefore, if only these two mechanisms are contributing to the harmonic shift, we expect that the blue-shift of the harmonics should independent of the sign of the chirp.
From Fig. 5(b), we see that the blue-shift of the harmonics decreases relative to the zero chirp setting (40 fs) when the pulse duration is increased from 40 fs to + 95 fs.However, for an increasingly negative chirped drive laser, from 40 fs to −68 fs, the blue-shift of harmonic becomes stronger.These observations match well the experiments described in [31] by Zhou at al., where the same Ar gas pressure (6.7 mbar) was employed.We conclude that in our case the harmonic blue-shift at negative chirp (> −1000 fs 2 ) can be understood as being generated at the leading edge, and thereby a shorter wavelength is generated here as compared to positive chirp.
As we apply an even larger negative chirp to the drive laser (data points at −98 fs corresponding to a chirp of −1500 fs 2 ), we observe that the harmonics are no longer blueshifted.Instead, the harmonics show a significant amount of red-shift compared to a pulse with similar duration but with the opposite chirp ( + 95 fs).We also observe that the harmonics start to broaden, which is indicated by the spectral tails, specifically at q = 17 and 19, and a side peak shows up at q = 19 (see Fig. 4(a)).The sharp peak in Fig. 2 indicates that the plasma dispersion is important for phase matching.When the pulse intensity is high, the optimal electron density is achieved in the first half of the pulse.However, the peak intensity of a strongly chirped laser pulse is so small that, possibly, the optimal electron density is only reached after the peak of the pulse.In a negatively chirped pulse, the trailing edge is redshifted.So, if most harmonics are generated at the trailing edge (due to phase matching), then the harmonic spectrum produced by a weak negatively chirped laser pulse must be redshifted.This is what is observed.Fig. 6.Relative wavelength shift of the 17th, 21st and 25th harmonic orders as a function of pulse duration with respect to the harmonic peak wavelengths at zero chirp (Ar gas pressure of 6 mbar).The dashed line indicates the relative wavelength shift of the drive laser, the dashed dotted line is reproduced from [15].The sign of the pulse duration corresponds to the sign of the applied drive laser chirp.
To compare the harmonic wavelength shift that we obtained via chirping of the drive laser pulses to that observed with a standard capillary of small diameter (150 μm), we plot the result for harmonic q = 23 reported by others [15] as the dashed dotted line in Fig. 6.Other harmonics and the fundamental follow a similar pattern.We note that in [15], Froud et al. define δλ q as the shift relative to the harmonic center wavelength with zero chirp setting at a gas pressure of 80 mbar.To enable a direct comparison with our results, we plot our measurements for δλ q /λ q (for q = 17, 21, 25 and 1) as a function of the drive laser chirp in Fig. 6, now with δλ q given as the shift relative to the harmonic center wavelength at zero chirp for a gas pressure of 6 mbar.
Compared to our results, it can be seen that the degree of shift obtained in [15] at a chirp of + 1000 fs 2 is higher, which we address to their using much higher Ar gas pressure.However, with 80 mbar of Ar, which is a factor of 13 more than in our case, their measured wavelength shift was only 2.4-times (0.012/0.005) more than what we measured.On the other hand, over the whole range of chirps we investigated, the wider capillary at lower pressure shows a slightly larger range of wavelength shifts.This observation indicates that in our experiment the second and the third mechanisms, which are based on the single-atom response, are larger than the contribution of the first mechanism.This is also supported by experiments in the thin capillary, such as apparent in the data of [15], where the relative wavelength shift, δλ q /λ q , for all harmonic orders is the same as for the drive laser implying that there the main mechanism there is ionization induced blue-shift.This is confirmed qualitatively because, regardless of the sign of the chirp, the harmonics in [15] (the dasheddotted curve in Fig. 6) show only a reduced blue shift (maximum value of + 0.018) relative to the maximum shift at zero chirp setting.For the large diameter capillary, we find that with respect to the blue shift at zero chirp the harmonics are shifted in both directions, red and blue shifted (from −0.0025 to + 0.015) following the sign of the chirp.When the chirp is decreased from −1000 fs 2 tp −1500 fs 2 , the blue shift is changed into a strong red shift with the largest tuning rate observed.

Summary and conclusions
We have shown that waveguided high-harmonic generation (HHG) at elevated drive laser energy propagated through a wide capillary can be wavelength tuned, and we compare the tuning with that of using a thin, standard capillary.By applying an elevated drive laser energy to the three-times wider capillary, we have increased the harmonic output approximately three-fold over that achieve in the standard tine capillary.With experimental parameters much different from a standard capillary, specifically, a much lower gas pressure at which phase matching occurs, we have investigated the spectral control of HHG as a function of gas pressure, and as a function of the drive laser pulse duration adjusted via a linear chirp with a programmable dispersive filter.The relative shift of the high harmonics, δλ q /λ q , are found to range from 0 to −0.0075 via increasing the gas pressure, and from −0.0075 to + 0.010 following the sign of the chirp.We found that the wavelength shift is based not only on the ionization induced blue-shift as reported for a standard capillary, but that other non-adiabatic mechanisms provide significant contributions as well.The results show that in spite of rather different operational parameters, the wavelength shifting of high harmonics in such capillaries can be as effective as with a standard capillary.Furthermore, our scheme appears largely free of ionization-induced nonlinear propagation effects in contrast to what is found in [32] by G. Tempea et al.The increased output pulse energy and the significantly lower electron densities (due to much lower gas pressure) appear attractive for applications that require spatially and spectrally controlled XUV radiation.

Fig. 2 .
Fig. 2. Total (spectrally integrated from 16 to 80 nm) pulse energy of the high-harmonic beam measured as a function of Ar gas pressure for two different capillary diameters, for d = 150 μm (triangular symbols, standard capillary) with a drive energy of E = 1.0 mJ, and for d = 508 μm (square symbols, wide capillary) with E = 6.0 mJ.The shown pulse energies are obtained by measuring the individual pulse energies of 100 single shots and taking the average.The error bars show the pulse-to-pulse root mean square deviation.

Fig. 3 .
Fig. 3. (a) High-harmonic and (b) drive laser spectra measured following the capillary filled with different Ar gas pressures, for a drive laser energy of E = 6.9 mJ, and with zero chirp setting of the programmable filter.The spectra are obtained by integrating over 20 pulses.The dashed lines indicate the center wavelengths of the harmonic orders and of the transmitted drive laser recorded at the lowest Ar gas pressure of 2 mbar.The short solid lines indicate the center (peak) wavelength of the transmitted IR pulse.

Fig. 4 .
Fig. 4. (a) High-harmonic and (b) drive laser spectra measured following the capillary filled with 6 mbar of Ar using a drive laser energy of E = 6.9 mJ and with various different chirps applied.The sign of the pulse duration corresponds to the sign of the applied chirp.The total XUV energy (integrated from 16 to 80 nm) is (0.37, 1.38, 3.36, 5.71, 2.28 1.99) nJ for the IR pulse durations (−98, −68, −48, 40, + 57, + 95) fs, respectively.The vertical dashed line in (b)shows the center wavelength of the drive laser transmitted through an empty capillary (no Ar gas), whereas the short solid lines show the measured drive laser wavelength in the presence of gas.The vertical dashed lines in (a) are the reference wavelengths for the harmonics as given in Fig.3a.

Fig. 5 .
Fig. 5. Measured relative wavelength shift of the 17th, 21st and 25th harmonic orders as a function of (a) gas pressure and zero chirp of the drive laser pulse, and (b) as a function of the linear chirp of the drive laser pulse for a pressure of 6 mbar.The latter is expressed also as various pulse durations in the upper horizontal scale, where the sign given with the pulse duration corresponds to the sign of the applied drive laser chirp.The shortest pulse duration of 40 fs is achieved with a zero chirp setting (0 fs 2 ).Negative values of the relative wavelength shift on the vertical axes represent a blue-shift.The dashed traces indicate the relative blue shift of the drive laser.