We report on the developed front-end/pump system for optical parametric chirped pulse amplifiers. The system is based on a dual output fiber oscillator/power amplifier which seeds and assures all-optical synchronization of femtosecond Yb and picosecond Nd laser amplifiers operating at a central wavelength of 1030 nm and 1064 nm, respectively. At the central wavelength of 1030 nm, the fiber oscillator generates partially stretched 4 ps pulses with the spectrum supporting a <120 fs pulse duration and pulse energy of 0.45 nJ. The energy of generated 1064 nm pulses is 0.15 nJ, which is sufficient for the efficient seeding of high-contrast Nd:YVO chirped pulse regenerative amplifier/post amplifier systems generating 9 mJ pulses compressible to 16 ps duration. The power amplification stages, based on Nd:YAG crystals, provide 62 mJ pulses compressible to 20 ps pulse duration at a repetition rate of 1 kHz. Further energy scaling currently is prevented by limited dimensions of the diffraction gratings, which, because of the fast progress in MLD grating manufacturing technologies is only a temporary obstacle.
© 2016 Optical Society of America
Optical parametric chirped pulse amplification (OPCPA)  technology enables generation of ultrashort high energy pulses in broad spectral range which extends from the visible to mid-infrared. Pulses as short as 5 fs (1.7 optical cycles at 880 nm) can be generated by using OPCPA technique . High peak (16 TW)  and average (22 W)  powers could also be achieved.
High energy OPCPA systems rely on high energy pump sources. While designing/selecting a pump source for an OPCPA system one should consider such parameters as pump pulse energy, duration, wavelength, and repetition rate. A synchronization of seed and pump pulses at different stages of OPCPA system is also an important aspect. The most convenient approach is to use passive optical synchronization: when all pump lasers of an OPCPA system are seeded by the same oscillator. Furthermore, for an ultrashort pulse OPCPA or OPA, a generation of stable coherent seed is an important issue, which usually is achieved through self-phase modulation (SPM) based spectral broadening of pump pulses. Laser pulses with the duration considerably below 1 ps are desired for stable generation of such white-light (WL) seed in transparent dielectrics . This brings a demand on combination and synchronization of low energy femtosecond WL seed generating lasers and high-energy picoseconds/nanosecond pump lasers.
Duration of the pump pulses depends on their energy which is dictated by the nonlinear phase accumulation (also referred to as a break-up or B-integral). In the case of Joule-level systems typically nanosecond pump pulses are implemented [5, 6], while millijoule-level OPCPA systems can be pumped by 50-100 ps [7, 8] or even 1 ps pulses . When pump pulse shortens, peak power of the pulse increases which allows the use of shorter nonlinear optical crystals and extension of the amplification bandwidth leading to the generation of shorter pulses. The negative aspect of using short pump pulses is increased sensitivity to the synchronization accuracy of the seed and pump pulses. Fluctuations of optical path lengths in an OPCPA setup due to thermal effects and mechanical instabilities could be relatively large compared to the pulse duration of the pump source. Ytterbium-doped active media (e.g. Yb:YAG) based chirped pulse amplification (CPA) systems usually generate short ~1 ps pump pulses [10–12] although in comparison to the neodymium-doped active media, which are capable of producing ~10-20 ps pulses (8 ps pulse duration was achieved in ), gain cross-sections of Yb-doped crystals are typically lower. Since the durations of seed and pump pulses need to be matched, the duration of pump pulse also determines dispersion management in an OPCPA system: in the case of (sub-) picosecond pump pulses bulk or chirped-mirror compression of generated OPCPA output is usually used while in the case of few-tens of picoseconds/nanosecond pump diffraction grating based compressors are employed.
Repetition rate of pump lasers also depends on anticipated energy, which is dictated by the thermal management in the pump chamber. Higher repetition rates are desired because of more effective data collection, especially in the experiments with coherent and incoherent X-rays [14, 15]. During the last decades flash-lamp pumping of Nd-doped laser crystals was increasingly replaced by laser diode pump modules which led to the development of diode- pumped solid state laser (DPSSL) technology. Due to their significantly higher pump efficiency DPSSL systems are capable of operating at much higher repetition rates as compared to flash-lamp pumped analogues.
Picosecond pump pulses as compared to nanosecond ones allow achieving higher parametric conversion efficiencies and shorter generated parametric pulses however the energy of a pump pulse is mainly limited by the accumulation of nonlinear phase (B-integral) in optical components of the amplifier (Laser crystals, Pockels cells, polarizers etc.). One way to overcome the problem would be to apply CPA technology for Nd based amplifiers. However since the gain in Nd-doped laser crystals is rather narrowband, stretching and compression of the pulses is challenging because of the large dimensions of stretcher and compressor units. Recently hyper-stretcher/hyper compressor approach was introduced  where the use of a pair of gratings instead of a single grating allows scaling down the dimensions of the stretcher/compressor units. One of the obstacles of the hyper-compression technology is diffraction efficiency of the gratings. In the case of typical 90% diffraction efficiency, after 8 reflections (in case of hyper-compressor) one can achieve only ~43% throughput. It must be noted that the problem is partially solved by recently developed diffraction gratings with 98-99% diffraction efficiency . Another challenge of the hyper-compressor is related to the complexity of the alignment.
We present concepts of the seed and pump sources for a hybrid multi-millijoule femtosecond OPA/picosecond OPCPA system operating at 1 kHz repetition rate based on an earlier reported 20 Hz mid-IR OPCPA system [Fig. 1] . Kilohertz repetition rates in the case of high-energy Nd-amplifiers became reachable through recently developed DPSSL technology. In our laser system we implemented the chirped pulse amplification (CPA) technology for Nd:YAG/Nd:YVO active medium based amplifier using custom diffraction gratings designed and manufactured by Fraunhoffer IOF. Higher gain of Nd doped active media (as compared to Yb doped) leads to a less complex amplifier design without a need of a large number of passes and/or cryogenic cooling. The paper is organized as follows: after describing the design and performance of dual wavelength seed source, we characterize performance of the amplification chain consisting of chirped pulse regenerative amplifier, post-amplifier and power amplifier; before concluding we discuss the issue of energy scalability of the developed amplification system.
2. Seed source
To realize all optical synchronization of femtosecond OPA and picosecond OPCPA stages [Fig. 1] while providing enough seed for the generation of high contrast pump pulses, a dual-output fiber frontend was designed. The front-end consisted of a passively mode-locked fiber oscillator, a double stage fiber amplifier, and spectral and temporal pulse formation chain (see Fig. 2). Polarization maintaining single mode fiber was used in all parts of the system.
A chirped fiber Bragg grating (CFBG1) and a semiconductor saturable absorber mirror (SESAM) were employed as end mirrors of the resonator of the oscillator. CFBG1 with ~60% reflectivity at 1047 nm, 25 nm spectral bandwidth and anomalous dispersion of 4.5 ps/nm was implemented for the dispersion management of the resonator. A free space long pass filter (LPF) placed in the collimated beam before the SESAM provided wavelength tuning of the oscillator. The central wavelength of the oscillator was set at 1047 nm by adjusting the angle of incidence on the LPF. The SESAM, placed on a mirror mount for angular adjustment and a translation stage for optimization of position along the beam, enabled self-starting passive mode-locking of the oscillator. Laser radiation was focused on the SESAM by a 12 mm focal length lens to a spot of 7 µm in diameter. The oscillator was pumped through a wavelength division multiplexer (WDM) and CFBG1 by a 20% fraction (about 80 mW) of an output of a laser diode (LD1). LD1 was protected from back returning pulsed radiation by a band-pass filter (PF1). Oscillator at a repetition rate of 60.3 MHz generated 2.9 ps pulses, which were close to transform limit. Average output power after the CFBG1, which served as an output coupler, was in the order of 3.5 mW. 10% of the output power reflected by a beam splitter BS1 (10/90) was used for synchronization of the system to the clock of the master oscillator.
Pulses from the oscillator were directed through the WDMs and isolators to a two-stage amplifier. The first amplification stage was pumped by the remaining 80% output (320 mW) of LD1, while the second amplification stage was pumped by the high power (900 mW) single mode laser diode LD2. After two stages pulses were amplified to 6.7nJ pulse energy which corresponded to 435 mW of average power. The spectrum of amplified pulses was broadened due to self-phase modulation (SPM) in an 11 m long passive fiber. Resulting spectrum spanning from 1025 nm to 1067 nm, presented in Fig. 3, covered gain spectra of both Yb-and Nd-doped laser media. Duration of the spectrally broadened pulses was measured to be around 15 ps.
Spectrally broadened pulses were then divided by a 30/70 ratio beam splitter (BS3) (we used a beam splitter instead of a broadband circulator because the later was not available in the laboratory) with 30% fraction being dumped and 70% fraction being passed to a narrowband FBG2, which back-reflected 0.4 nm spectral bandwidth around the central wavelength of 1064 nm for further amplification. The spectral fraction centered around 1030 nm passed FBG2 for seeding femtosecond 1030 nm amplifier. A free space bandpass filter (BPF) was used to reject spectral components, which would not be amplified in Yb-amplifier. Output power after the BPF was measured to be 27 mW which corresponded to 450 pJ pulse energy. Spectral bandwidth of about 10 nm, and pulse duration of 4 ps indicated that the 1030 nm pulses were partially chirped. The 10 nm bandwidth achieved through the SPM in fiber supported <300 fs pulse duration.
Narrowband pulses at 1064 nm were back reflected by FBG2 and coupled through 30% branch of the beam-splitter BS3 to WDM4. An average power (~0.1 mW) and energy (~1.6 pJ) was too low for efficient seeding of a RA based on Nd-doped laser crystal. In order to boost pulse energy an additional fiber amplification stage, in which 1064 nm pulses were amplified by a factor of 100 to 150 pJ pulse energy, was installed. The resulting energy of 150 pJ was sufficient for high contrast operation of the solid-state regenerative amplifier . The duration of the pulses was measured to be 8.5 ps with ~0.4 nm spectral bandwidth.
At the output of the fiber oscillator chirped 1064 nm pulses were further stretched using a chirped fiber Bragg grating (CFBG2) with 0.3 nm spectral bandwidth (at FWHM) and −2000 ps/nm dispersion parameter. The duration of the resulting pulses was ~520 ps.
3. Regenerative amplifier
Stretched output pulses of the oscillator were amplified in a RA [Fig. 4] based on end-pumped Nd:YVO4 active medium of 0.3% Nd concentration and having dimensions of 3x3x8 mm (width x height x length). The RA was pumped by 15 mJ, 100 µs (10% duty cycle) pulses generated by a 15 W average power GaAs laser diode operating at the central wavelength of 808 nm (pump diameter ~1.2 mm). The Nd:YVO4 active medium was chosen due to its broader than Nd:YAG amplification bandwidth and relatively high single pass gain. Broader amplification bandwidth allowed preserving wider spectrum as for chirped pulses this corresponds to longer pulses. In addition it allowed straightforward further amplification of pulses generated in Nd:YVO4 RA with Nd:YAG power amplifier since no precise wavelength tuning was required. The Nd:YVO4 RA generated ~3 mJ pulses at a repetition rate of 1 kHz. After the RA pulse duration was shortened to ~300 ps due to the spectral narrowing during the amplification (note that in case of stretched pulses spectral narrowing or filtering leads to pulse shortening as spectral components of the pulse are spread in time).
After the RA pulses were further amplified in a Nd:YVO4 post-amplifier [Fig. 5] based on 0.3%-doped Nd:YVO4 crystal with dimensions of 5x5x8 mm3). The crystal was end-pumped by two fiber coupled laser diodes each providing 100 µs, 24 mJ pump pulses. In the post-amplifier a single pass arrangement was chosen as a double-pass configuration appeared to be not efficient enough to justify more complex layout and additional thermal lens aberrations experienced by the amplified beam. After a single pass 2.9 mJ input pulses were amplified to 9 mJ energy while 12 mJ pulses were obtained in a double-pass configuration. The B integral after the post-amplifier was estimated to be about 1.1 with the contributions from the RA and post-amplifier being ~1.0 and ~0.1 respectively.
As it can be seen from the Fig. 6 where the spectra recorded after the RA and after the post-amplifier are presented, during the amplification in the post amplifier spectrum broadened by a nearly factor of 2 (from ~0.12 nm FWHM to 0.22 nm FWHM).
We believe this happened because the central wavelength of the RA output was red shifted with respect to the central wavelength of the amplification spectrum of post-amplifier, so the blue wing was amplified more efficiently. This partially compensated the shift of the spectrum that positively chirped pulses typically undergo during amplification, i.e. red spectral components arrive to the laser crystal before the blue ones and experience stronger amplification, which is valid for the case of high-level amplification with the depletion of population inversion. The scenario is supported by the spectral transformations of the RA output during amplification in the post-amplifier [Fig. 6]: the peak of the spectrum during amplification shifted to the red side, while blue components, being closer to the maximum of the gain spectrum, were amplified more efficiently, which lead to the broadening of the spectrum.
5. Power amplifier
9 mJ pulses originating from the post amplifier were further amplified to the energy of 62 mJ in a double-pass amplification stage [Fig. 5] with a Nd:YAG laser module produced by Northrop Grumman Cutting Edge Optronics (CEO) and having active medium of 5 mm in diameter. Relay imaging and Faraday polarization rotator were used to compensate the thermally induced birefringence . In order to improve beam profile spatial filtering was employed between the post amplifier and the power amplifier stages and at the output of the power amplifier. This resulted in a quasi-Gaussian output beam profile [Fig. 7] featuring quite good symmetry of far field intensity distribution.
In order to examine compressibility of the amplified pulses we designed a compressor based on a pair of custom made 1818 gr/mm dielectric gratings placed at a 3 degrees with respect to the Littrow angle. With the distance between the gratings of 3 m the compressor provided group delay dispersion of approximately 2000 ps/nm. By using a folded design [Fig. 8], the geometrical dimensions of the compressor were reduced to merely ~0.9 x 0.25 meters. The output and input of the compressor were vertically separated by ~1 cm by a roof mirror. Diffraction efficiency of the gratings of 96% corresponded to a maximum achievable transmission of the compressor of ~85%.
As one can see from the Fig. 9 the beam at the output of the compressor became slightly elliptical, which was caused by geometrical transformations during propagation. Due to the large angle of diffraction from the first diffraction grating GR1 the horizontal dimension of the diffracted beam was reduced, while the vertical dimension stayed the same. Since the smaller dimension experienced stronger divergence, after some distance the beam became circular again. However in our layout this distance coincided with the position of the second diffraction grating GR2. During diffraction from the second grating GR2 the horizontal dimension of the beam was increased. In fact the beam in our layout was circular before it was diffracted by the first diffraction grating GR1 for the second time. After diffraction the horizontal dimension was increased, leading to an elliptical intensity distribution at the output of the compressor. Possible solutions to this problem are either to increase the input beam diameter or to use diffraction gratings at a smaller angle of incidence. An increase of the input beam diameter would mean that larger dimensions of the gratings and other optical components of the compressor are required. In order to use gratings at a smaller incidence angle one would need either to increase the distance between the gratings or to change the grove density of the gratings as the reduction of the angle reduces the group delay dispersion of the compressor.
62 mJ pulses after the power amplifier were compressed to a duration of about 18 ps [Fig. 10]. The pulse energy after the compressor was ~38 mJ, meaning that only ~60% of energy was transmitted through the compressor. Additional losses are probably caused by a slight clipping of the beam on optical components of the compressor (including the diffraction gratings themselves). The clipping could also be a reason for slightly non-Gaussian shape of the autocorrelation function shown in Fig. 10., and for the duration of compressed pulse being longer than the time bandwidth limited value of <11 ps.
7. Energy scaling
In order to increase the efficiency of the amplifier, we installed a spatially variable retardation plate-based beam shaper , which was designed to convert a Gaussian beam into a beam with a 6th order super-Gaussian intensity distribution. Unfortunately it appeared that after the post-amplifier the beam profile was not strictly Gaussian anymore, so the beam shaper didn’t function properly. Therefore, in order to extract more energy from the system, we placed the beam shaper at the output of the RA and used an optical layout without the post amplifier, instead employing two consequent double-pass stages based on two CEO laser modules. In this configuration we were able to achieve ~130 mJ output pulse energy. The problem of this approach was that the amplified beam possessed strong thermal aberrations and broke up after propagating a distance that was much smaller than the optical path length of the compressor. Consequently, compression of the generated 130 mJ pulses could not be demonstrated. We believe that the problem could be overcome in a few ways, namely: i) by increasing the diameter of the beam at the output of the amplifier (however this would lead to larger dimensions of diffraction gratings and other optical components in the compressor; the size of the diffraction gratings currently limits achievable pulse energy); ii) by reducing the distance between the gratings of the compressor (this would mean less stretching of the pulses); or iii) by switching to a hyper-compressor configuration .
In conclusion, we demonstrate a novel picosecond driver for high repetition rate OPCPA system. The driver is based on a dual wavelength fiber seed source and a hybrid Nd:YVO/Nd:YAG chirped pulse amplification system. The amplifier at a 1 kHz repetition rate currently produces 18 ps, 38 mJ pulses, which are centered at 1064 nm. We believe that achieved 18-ps pump pulse duration will allow generation of broadband mid-IR pulses with rather high parametric conversion efficiency. Further scaling of the output energy above 100 mJ level is possible with the compressor being designed by using optical components having larger clear apertures.
Eurostars project E! 6 655 CHEOPS. Austrian Science Fund (FWF) SFB NextLite F4903-N23.
References and links
1. A. Dubietis, G. Jonušauskas, and A. Piskarskas, “Powerful femtosecond pulse generation by chirped and stretched pulse parametric amplification in BBO crystal,” Opt. Commun. 88(4-6), 437–440 (1992). [CrossRef]
2. J. Rothhardt, S. Demmler, S. Hädrich, J. Limpert, and A. Tünnermann, “Octave-spanning OPCPA system delivering CEP-stable few-cycle pulses and 22 W of average power at 1 MHz repetition rate,” Opt. Express 20(10), 10870–10878 (2012). [CrossRef] [PubMed]
3. D. Herrmann, L. Veisz, R. Tautz, F. Tavella, K. Schmid, V. Pervak, and F. Krausz, “Generation of sub-three-cycle, 16 TW light pulses by using noncollinear optical parametric chirped-pulse amplification,” Opt. Lett. 34(16), 2459–2461 (2009). [CrossRef] [PubMed]
4. M. Bradler, P. Baum, and E. Riedle, “Femtosecond continuum generation in bulk laser host materials with sub-μJ pump pulses,” Appl. Phys. B 97(3), 561–574 (2009). [CrossRef]
5. H. Yoshida, E. Ishii, R. Kodama, H. Fujita, Y. Kitagawa, Y. Izawa, and T. Yamanaka, “High-power and high-contrast optical parametric chirped pulse amplification in β-BaB2O4 crystal,” Opt. Lett. 28(4), 257–259 (2003). [CrossRef] [PubMed]
8. S. Witte, R. T. Zinkstok, A. L. Wolf, W. Hogervorst, W. Ubachs, and K. S. Eikema, “A source of 2 terawatt, 2.7 cycle laser pulses based on noncollinear optical parametric chirped pulse amplification,” Opt. Express 14(18), 8168–8177 (2006). [CrossRef] [PubMed]
9. Y. Deng, A. Schwarz, H. Fattahi, M. Ueffing, X. Gu, M. Ossiander, T. Metzger, V. Pervak, H. Ishizuki, T. Taira, T. Kobayashi, G. Marcus, F. Krausz, R. Kienberger, and N. Karpowicz, “Carrier-envelope-phase-stable, 1.2 mJ, 1.5 cycle laser pulses at 2.1 μm,” Opt. Lett. 37(23), 4973–4975 (2012). [CrossRef] [PubMed]
10. S. Klingebiel, C. Wandt, C. Skrobol, I. Ahmad, S. A. Trushin, Z. Major, F. Krausz, and S. Karsch, “High energy picosecond Yb:YAG CPA system at 10 Hz repetition rate for pumping optical parametric amplifiers,” Opt. Express 19(6), 5357–5363 (2011). [CrossRef] [PubMed]
11. M. Schultze, T. Binhammer, G. Palmer, M. Emons, T. Lang, and U. Morgner, “Multi-μJ, CEP-stabilized, two-cycle pulses from an OPCPA system with up to 500 kHz repetition rate,” Opt. Express 18(26), 27291–27297 (2010). [CrossRef] [PubMed]
12. S. Hädrich, S. Demmler, J. Rothhardt, C. Jocher, J. Limpert, and A. Tünnermann, “High-repetition-rate sub-5-fs pulses with 12 GW peak power from fiber-amplifier-pumped optical parametric chirped-pulse amplification,” Opt. Lett. 36(3), 313–315 (2011). [CrossRef] [PubMed]
13. M. Y. Shverdin, F. Albert, S. G. Anderson, S. M. Betts, D. J. Gibson, M. J. Messerly, F. V. Hartemann, C. W. Siders, and C. P. Barty, “Chirped-pulse amplification with narrowband pulses,” Opt. Lett. 35(14), 2478–2480 (2010). [CrossRef] [PubMed]
14. T. Popmintchev, M.-C. Chen, D. Popmintchev, P. Arpin, S. Brown, S. Ališauskas, G. Andriukaitis, T. Balčiunas, O. D. Mücke, A. Pugzlys, A. Baltuška, B. Shim, S. E. Schrauth, A. Gaeta, C. Hernández-García, L. Plaja, A. Becker, A. Jaron-Becker, M. M. Murnane, and H. C. Kapteyn, “Bright coherent ultrahigh harmonics in the keV x-ray regime from mid-infrared femtosecond lasers,” Science 336(6086), 1287–1291 (2012). [CrossRef] [PubMed]
15. J. Weisshaupt, V. Juvé, M. Holtz, S. Ku, M. Woerner, T. Elsaesser, S. Alisauskas, A. Pugzlys, and A. Baltuska, “High-brightness table-top hard X-ray source driven by sub-100-femtosecond mid-infrared pulses,” Nat. Photonics 8(12), 927–930 (2014). [CrossRef]
16. K. Hehl, J. Bischoff, U. Mohaupt, M. Palme, B. Schnabel, L. Wenke, R. Bödefeld, W. Theobald, E. Welsch, R. Sauerbrey, and H. Heyer, “High-efficiency dielectric reflection gratings: design, fabrication, and analysis,” Appl. Opt. 38(30), 6257–6271 (1999). [CrossRef] [PubMed]
17. G. Andriukaitis, T. Balčiūnas, S. Ališauskas, A. Pugžlys, A. Baltuška, T. Popmintchev, M.-C. Chen, M. M. Murnane, and H. C. Kapteyn, “90 GW peak power few-cycle mid-infrared pulses from an optical parametric amplifier,” Opt. Lett. 36(15), 2755–2757 (2011). [CrossRef] [PubMed]
18. J. Adamonis, R. Antipenkov, J. Kolenda, A. Michailovas, A. P. Piskarskas, and A. Varanavicius, “High-energy Nd: YAG-amplification system for OPCPA pumping,” Quantum Electron. 42(7), 567–574 (2012). [CrossRef]
19. K. Michailovas, V. Smilgevičius, and A. Michailovas, “High average power effective pump source at 1 kHz repetition rate for OPCPA system,” Lithuanian J. Phys. 54(3), 150-154 (2014).
20. A. Michailovas, J. Adamonis, A. Aleknavicius, S. Balickas, T. Gertus, A. Zaukevičius, K. Michailovas, and V. Petrauskiene, “A new beam shaping technique implemented in 150 W 1kHz repetition rate picosecond pulse amplifier,” in Conference on Lasers and Electro-Optics, OSA Technical Digest (online) (Optical Society of America, 2016), paper JTu5A.40. [CrossRef]