Generating energetic, few-cycle laser pulses with stabilized carrier-envelope phase at a high-repetition rate constitutes a first step to access the ultra-fast dynamics underlying the interaction of matter with intense, ultrashort pulses in attosecond science or high-field physics. We present here a Ti:Sa-based 1 kHz TW-class laser delivering 17.8 fs pulses with 350 mrad shot-to-shot CEP noise based on an original 10 kHz front-end design. In parallel to this short pulse duration operation mode, it is possible to tune the output wavelength of the front end within a 90 nm range around 800 nm.
© 2019 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
High energy, few cycle laser pulses represent the main tool for accessing the strong-field interactions ruling attosecond temporal dynamics. When a high-intensity (above 1014 W/cm2), ultrashort driving pulse interacts with a medium, a train of XUV attosecond pulses will be generated through a strong non-linear effect like high harmonic generation [1,2]. Many applications  need a single isolated attosecond pulse rather than a train of attosecond pulses, like pump-probe experiments , nano-plasmonics , molecular vibrational wave packet mapping , atomic correlation investigations  or attosecond lighthouse . The easiest way to obtain such isolated attosecond pulses is to confine the XUV radiation emission to one half of the optical cycle near the peak of the driving pulse. This temporal confinement has been obtained by post-compression of spectrally enlarged pulses through Self Phase Modulation (SPM) [9,10] or polarization gating [11,12] of high energy, sub-20 fs pulses. Accessing such short pulse duration is quite challenging because of the experimental difficulty to produce a sufficiently large spectral bandwidth with a controllable and compensable spectral phase variation.
Besides those non-linear techniques, the possibility to use the technique of synthesizing laser pulses covering different spectral regions [13,14] for generating energetic sub-cycle laser pulses has been demonstrated recently. This technique also provides the possibility to tailor specific sub-cycle waveforms for applications such as arbitrary waveform generation  or quantum control .
A classical commercial OPA system pumped typically with 20 mJ pulses, allows to produce pulses in the mJ range with a spectrum extending from 1000 nm to 2200 nm when associating Signal and Idler spectra. The spectral range can be extended from 500 to 1100 nm when combining a SPM spectrum [13,17]. A direct wavelength tunability of amplified pulses at the output of the laser within a smaller spectral range between 750 and 850 nm can also be very useful for finely tuning wave mixing, for above-described light field synthesis, for direct coherent control in atomic or molecular physics experiments , and even for driving photo-injector electron guns on some Free Electron Lasers (FEL) facilities .
Nevertheless, besides single cycle pulse duration, another key ingredient for achieving isolated attosecond pulses is the control of relative Carrier-Envelope Phase (CEP). It is crucial to ensure a faithful replicability of the produced electrical field through laser–matter interaction. Numbers of applications will also benefit from a driving source exhibiting simultaneously high repetition rate and high pulse energy. Laser systems associating those parameters with a shot to shot CEP stabilization better than 300 mrad have been technically challenging. And so far, the pulse energy is mostly restricted to a few mJ at 1 kHz or even less at 10 kHz for Ti:Sapphire (Ti:Sa) based systems, while synthesized pulses are limited to μJ level .
Grating-based CPA system has been demonstrated to be, until now, the most suitable choice for achieving high-energy pulses, keeping a low B-integral value, for Ti:Sa based amplifier . On the other hand, compressors and stretchers based on reflection gratings are notably more sensitive to mechanical vibration: small vibrations and movements of the gratings could induce a significant CEP shift [21–23]. Distance and mutual angular stability of the gratings in stretcher and compressor become crucial parameters to keep a good CEP stability. To date, only a few CEP-stable lasers exhibiting tens of mJ have been reported with shot to shot CEP stability lower than 350 mrad [10,17,24]. It should be mentioned that OPCPA technology starts to become mature and it can be an interesting alternative to get short pulse duration at high repetition rate, with mJ energy level and low CEP residual noise  even though those systems remain quite complicated and expensive.
We present here a Ti:Sa based laser system with dual simultaneous output delivering 16 mJ, 17.8 fs pulses (0.9 TW) at 1 kHz and 1.6 mJ sub-20 fs pulses at 10 kHz with long-term CEP stability. This laser, based on Chirped Pulse Amplification (CPA) technique , has been developed by Amplitude Technologies in collaboration with CEA Saclay within the joint laboratory Impulse.
2. Laser system specific front-end description
We will first introduce an original full water-cooled 10 kHz front-end used to seed different amplification stages. The so-called front-end relays on a full water-cooled Ti:Sa regenerative amplifier with a specific design allowing efficient thermal management. Its output is split in two parts, seeding simultaneously two amplifiers operating respectively at 1 and 10 kHz. We will then also present the possibility to use this front end in an output wavelength tuning mode and discuss the amplified results in the 1 kHz arm. Finally, we will present the broadband amplified mode reaching 23 mJ energy per pulse at 1 kHz after two ‘bow tie’ multi-pass amplifier stages, leading to 16 mJ, 17.8 fs after a grating-based compressor. The 10 kHz arm delivers close to 1.6 mJ, 18 fs laser pulses at the final output.
2.1. 10 kHz single crystal cavity
A standard regenerative cavity (Fig. 2(a)) operated generally between 10 Hz and 1 kHz repetition rate allows efficient amplification with more than 6 orders of magnitude gain while keeping a very good spatial beam profile thanks to its spatial filtering effect. The main drawback is gain narrowing effect during the amplification . An intracavity programmable filter (AOPGCF or Mazzler)  can be used in order to counteract this effect. Spectral losses with an optimal shape to ensure a homogeneous amplification over a large spectrum are created where the spectral gain is higher. This leads to an amplification enhancement at the edges of the spectrum with respect to the center of the spectrum, thus ensuring a broadband amplification allowing sub-25 fs amplified pulse compression.
Such a cavity can be pumped with a pump power up to 15 or 20 W at 1 kHz repetition rate when using a non-programmable intracavity filtering device like a passive spectral filter. Unfortunately, when using higher repetition (10 kHz) pump lasers with same pump power (around 15 W), the output beam exhibits strong deformations at high pump power, as shown in Fig. 2(b) right. Pumping with high average power at this repetition rate indeed causes thermal heat accumulation in the crystal. The cavity becomes then unstable leading to a very distorted spatial beam shape and limiting the output power as well as the efficient use of spectral filters counteracting gain narrowing. Pump power has to be divided by a factor of 2 in order to recover a good spatial beam shape as shown in Fig. 2(b) left. This limits also the efficient output power of the standard cavity at 10 kHz to 1 W with a spectral width at FWHM of 80 nm.
2.2. Thermal focal length measurement
Heat load in the crystal leads to two main phenomena : refractive index gradient and thermo-mechanic deformation of the crystal. Both of those phenomena contribute to increase the thermal lens effect inside the crystal. Its focal length is described by the following formula [30,31]:Eq. (1) [30–33] which permits to calculate the thermal focal length variation as a function of pump power in our experimental conditions. The results are shown in full red line in Fig. 3; the thermal focal length varies from 40 cm at 8 W pump power down to less than 20 cm at 14 W pump power considering our experimental water cooling system. This very short induced thermal focal length has to be taken into account for predicting the cavity stability parameter and to properly design the W-like double crystal regenerative cavity presented in the next section. One can then estimate the shortest value of thermal focal length that can be supported by the cavity while keeping its stability.
To confirm the validity of our calculations based on the theoretical formula of Eq. (1) we performed an experimental measurement of the induced thermal lens length. The experimental principle is relatively simple : a probe constituted by a He-Ne beam passes through the amplifier crystal. The probe beam is imaged using a 4f-line in the plane of the crystal output surface and its waist is adapted to the waist of the IR mode. A magnification telescope can be used in order to image this waist on a wave front detector. This waist-to-waist conjugation makes it possible to obtain a plane wave front, equivalent to an infinite focal length, when the crystal is not pumped. It can be demonstrated that the measured wavefront curvature radius is:Fig. 3, without extraction of the IR from the cavity. The experimental data are well fitted by the theoretical calculations, especially at high pump power, leading to a 20 cm thermal focal length when pumped at 14 W. At lower pump power the agreement between theoretical and experimental data is weaker. We suppose that this is due to a low sensitivity of our detector when the wavefront curvature is less pronounced. The measured thermal focal length can be 6 cm longer when the cavity is seeded and the produced power extracted. In this configuration the radiative stimulated emission is encouraged: most of the electronic population is relaxed to the ground level with radiative emission, reducing the probability of non-radiative emission, which causes heat accumulation. Thus in order to operate in different alignment and optimization conditions it is important that the cavity remains stable over a large range of thermal lens focal length values. The good agreement between experimental measurements and calculations confirms the huge importance of considering this short thermal lens in the cavities or other high power amplifiers designs.
Even though proper dimensioning of the crystal and cooling systems help reducing the heat load, the thermal effect is critically influenced by the pump power density. As the pump power increases, the focal length of the thermal lens becomes shorter, leading to instability of the cavity, aberration and thus poor beam profile as shown in previous section.
2.3. Two-crystal cavity design
In order to reduce the detrimental effect of the thermal load and allow higher pump power in the regenerative cavity, a new design using two crystals has been proposed as shown in Fig. 4(a). A detailed description can be found in  and we will recall here only the main features and results. The thermal load is there distributed over two amplifying media, thus limiting the detrimental effect of the induced thermal lens. Figure 4(b) shows that the output power increases as a function of the pumping power. In the case of one crystal configuration (blue points) it is possible to achieve close to 3 W output power at 21 W pump power while the AOPGCF is off but the beam shape becomes distorted because of the strong induced thermal lens. Pump power has to be limited to 12 W, limiting the cavity output power to 1.5 W in order to keep a very good spatial beam shape (spectrum shows a Gaussian shape and 35 nm at FWHM). On the other hand, the double crystals cavity supports pump power up to 28 W while keeping a very good IR output beam spatial profile as it can be seen on Fig. 4(b) (upper inset). One can notice that the double crystal cavity remains stable over a very broad pumping power range, from 14 W to 28 W (in total) thanks to its specific design taking into account the measured thermal focal lens length. The output power reaches 5.6 W at maximal pump power which is more than 3 times more efficient than the single crystal configuration. We should specify that the pump power was here limited by the IR power density threshold of the AOPGCF placed in the cavity for broad band amplification.
2.4. Different operating modes of the two-crystal regenerative cavity
This new W-like designed cavity with a better thermal load management enables efficient operation in three different modes: narrow band (NB), broad band (BB) or tunable narrow band (TNB) mode.
As described previously, in NB mode, when the intracavity spectral shaping device is switched off, the output spectrum width is around 35 nm at FWHM. In addition to the intracavity filter, a Dazzler module (AOPDF) is placed at the output of the stretcher  which can provide at the same time a dispersion compensation for pulse duration optimization as well as spectral shaping and CEP slow drift correction. In BB mode, when turning the AOPGCF on, different spectral bandwidth optimization loops allow to increase the spectral width of the out coming pulse as shown on Fig. 5. Finally, it is possible to enlarge the amplified pulse spectrum up to 110 nm at FWHM with a super- Gaussian shape. Since losses are introduced, the output power is then lowered to 2.7 W. Increasing the pump power allows generally to partially compensate this energy loss but at 10 kHz it is less efficient because of the strong thermal lensing. Coupling this cavity to an additional one-pass two crystal amplifier, again to limit thermal effects, pumped by 15 W green light allows additional amplification up to 10 W and 7W respectively without and with gain narrowing compensation loop.
When associated to a grating based compressor, the front-end delivers 500 µJ, 17 fs pulses at 10 kHz repetition rate, CEP stabilized with an active analogical correction loop leading to a remaining shot-to-shot CEP noise of 200 mrad over 3 hours and 170 mrad over 30 seconds as described in .
Alternatively, in TNB mode, pulses with tunable spectral width and central wavelength within ± 40 nm around 800 nm can be produced. This spectral tunability is achieved by a coupled use of the two acousto-optic filters. In order to be able to tune the output wavelength within 80 nm bandwidth, the final amplification spectral gain width has to be at least 100 nm large like in the BB mode. Thus, the first step is to operate the cavity in BB mode. The second step consists in using the first programmable spectral shaper (Dazzler) to select a super-gaussian profile within the large spectrum of the pulses that seed the cavity. We have to precise here that we use a CEP-stable oscillator (Rainbow CEP4, Femtolasers) delivering octave spanning pulses with a theoretical residual CEP noise below 60 mrad  that are temporally stretched by a grating-based Öffner stretcher with 150 nm spectral bandwidth throughput. It is thus quite easy to select a part of the spectrum of the seed pulses setting a spectral width between 35 and 50 nm FWHM and a central wavelength between 760 and 850 nm. Those spectrally shaped pulses can then be seeded into the regenerative double-crystals cavity operating in BB mode. Since the spectral gain profile of the regenerative amplifier is larger than the seeded pulse spectrum, the amplified spontaneous emission (ASE) is higher than that in the BB mode. We show in Fig. 6 the case of a 30 nm FWHM spectrally super-gaussian shaped seed pulse centered at 800 nm amplified in the regenerative cavity (blue line) compared to the BB mode spectrum (black line). An ASE signal is arising around 750 nm even though these spectral components are not injected in the amplifier. The amplification process induces strong red-shift that produces an unsymmetrical spectrum as well as some spectral narrowing (25 nm at FWHM). Those effects can be compensated by changing the shape of the intra-cavity losses introduced by the AOPGCF in order to get rid-off this unwanted ASE as well as preserving the width and shaping of the seeded pulse as shown in red in Fig. 6. An enhancement of the spectral brightness of the pulse can then be seen while comparing the spectra before and after spectral re-shaping with the applied acousto-optical wave on the AOPGCF device, resulting in a more efficient amplification.
Those optimization routines using Dazzler and Mazzler filters are implemented to adjust the spectral width and shape of the out-coming amplified pulse. The spectral FWHM of the amplified pulses can thus be adjusted between 17 and 66 nm with different spectral shapes as shown on Fig. 7(a). This technique can also be used to produce pulses with a varying central wavelength whereas FWHM and spectral shape are fixed. Such an example of wavelength tunability is shown in Fig. 7(b). For the sake of clarity, we decided to show only 5 selected spectra with central wavelength set at 760, 780, 800, 820, and 850 nm. It is worth to mention that the spectral tunability precision can be as precise as 1 nm. The spectral width of the emitted spectra could be varied between 20 nm and 40 nm FWHM. Those tunability results are obtained for the first time to our knowledge at a 10 kHz repetition rate in a Ti:Sa regenerative cavity output. When compressing temporally those pulses, their duration can thus be varied from 25 to 100 fs with Fourier limited pulse duration without any temporal chirp.
3. High amplification results in BB and TNB modes
After this detailed description of the innovative front-end, we will describe the two amplification arms of the FAB 1/10 laser system.
3.1. BB and TNB amplified spectra and durations
The 10 kHz front-end output, consisting of the double-crystal cavity followed by a single pass double crystal amplifier, is injected into two high energy amplifiers. A 30% fraction of the front-end output energy is sent through a Pockels cell to reduce the pulse repetition rate from 10 kHz to 1 kHz, resulting in 180 µJ in BB mode. The 1 kHz line will be described later.
The remaining 70% of the output energy from the 10 kHz front-end, corresponding to 510 µJ in BB mode, is sent to the 10 kHz cryo-cooled main amplifier. To counteract thermal lensing, the Brewster cut Ti:Sa crystal of the last amplifier stage is cooled down to −150°C under vacuum with a vibration-free cryogenic cooler. The amplifier does not reach full saturation condition and it is thus more sensitive to the injected power leading to 28 W and 23 W output powers when seeded respectively with a 80 nm or 110 nm FWHM spectrum and pumped at 10 kHz with 100 W of green pump laser light. After an expansion telescope, the amplified laser beam is send to a CEP-compatible reflection grating-based compressor. To prevent our system from being critically sensitive to mechanical vibrations, and also considering the extended beam dimension, special tuning-free CEP-compatible grating mounts have been designed for the compressor to ensure a solid and stable support for the optics and less sensitivity to vibrations. Its transmission efficiency of about 70% leads to 1.6 mJ pulses with a duration of 18.7 fs at 10 kHz after compression. The measured broadband spectrum and spectral phase together with temporal profile of the 10 kHz line pulses are shown respectively in Figs. 8(a) and 8(b).
The 1 kHz train of pulses are firstly amplified to 4.3 mJ in a water-cooled 5-pass pre-amplifier, and further boosted up to 23 mJ with 70 W of pump power in the 4-pass cryo-cooled main amplification stage. This last amplifier is identical to the 10 kHz one but with enhanced saturation effect. After an expansion telescope, the train of pulses is sent to a CEP-compatible reflection grating compressor, with tuning-free optical mount, already described for the 10 kHz line. An overall transmission efficiency of ~70% leads to 16 mJ output at 1 kHz repetition rate with the broadest achieved spectral bandwidth. The 110 nm spectral width of the front-end used in BB mode is maintained after the different amplification stages. Some red shift and spectral asymmetry occur in the high energy amplifiers; these effects can be corrected performing the spectral enlarging shaping loops on the amplified spectrum for final optimization. This allows to keep an ultra-broad band spectrum with a super-gaussian shape after the cryogenic amplifiers as shown in Fig. 9(a) which also displays the variation of the optimized spectral phase. Optimal pulse compression by Wizzler-Dazzler correction loops leads to a 17.8 fs pulse duration at the 1 kHz output, very close to the 17.7 fs Fourier limited pulse duration (Fig. 9(b)), when we obtained simultaneously 18.7 fs on the 10 kHz line. This pulse duration difference is due to the remaining spectral phase difference between the 1 and 10 kHz pulses due to different amplification stages that are not compensated simultaneously on both arms. As can be seen on Fig. 8(b), the spectral phase of the 10 kHz pulses is not as well corrected as the one of the 1 kHz resulting in a slightly longer pulse duration. Indeed, we have to choose one of the two arms to perform the temporal optimization loop and flatten the corresponding spectral phase variation. This has been done on the 1 kHz line in order to get as close as possible to the TW-class level.
As described in previous section the front-end can be used in TNB mode and seed the high power amplifiers. Our stretcher exhibits a 5 ps/nm stretching factor: the pulse duration is 500 ps for a 100 nm spectrum while it becomes 100 ps for a 20 nm FWHM spectrum, leading to an increase of the whole B-integral. The calculated B-integral values for a 100 nm FWHM spectrum are 0.6 rad and 1.5 rad respectively at the front-end and 1kHz amplifiers output. In order to check the most extreme conditions and still run the amplifiers safely, we reduced the TNB mode mean spectral width to 27 nm for the feasibility experiments on the 1 kHz arm. To avoid any damage on the FAB 1/10 facility, we limited the amplification to 10W instead of 23 W thus limiting the final B-integral value to 3.4 rad and working in a safer amplification regime. Figure 10(a) shows the effect of high power amplification on the central wavelength and FWHM bandwidth of different selected spectra. The spectral bandwidth is reduced by 10% and the central wavelengths are globally shifted towards 800 nm by approximately 3 nm. . This can be due to the higher gain around 800 nm in the 1 kHz amplifiers, since its spectral gain bandwidth is not optimized with a Mazzler loop. Here, the wavelength tunability loops are calculated using the regenerative cavity output spectra. According to the spectrometer resolution and to the uncertainty in the central wavelength determination some more precise measurement should be performed to be sure if those values are significant. Spectral shift and small gain narrowing limits the tunable bandwidth of the amplified pulses within a range of 70 nm instead of 80 nm at the front-end output. Figure 10(b) compares the output power of the front end with the10 mJ, 1 kHz amplifiers output in TNB mode. The energy of the front end pulses varies as a function of the wavelength between 220 µJ and 380 µJ, with a minimum at 800 nm and highest energy on the edges. One can assume an average output value of 300 ± 80 µJ and thus an energy variation of 25%. This is mainly due to the shape of the losses introduced by the intra-cavity spectral filter allowing broadband spectral gain and ASE attenuation. The output after amplification at 1 kHz shows less variation for different central wavelengths thanks to the strong saturation regime. The pulse energy is now about 10.0 ± 0.2 mJ limiting the peak to peak energy variation to 2%.
The experimental results and B-integral calculations show that we will have to limit to 40 nm the spectral bandwidth reduction of the pulses we want to amplify when using the TNB mode at full amplification. B-integral values should reach respectively 1.5 rad and 3.8 rad at front-end and 1 kHz amplifier outputs. The duration of those pulses after compression can still be tuned between 17.5 fs and 50 fs with Fourier limited pulse duration without any temporal chirp.
3.2. CEP stabilization of the amplified system
We have also measured the residual CEP noise after the two amplification arms in BB mode. As mentioned before and detailed in  shot to shot CEP residual noise is as low as 200 mrad over 3 hours using the front end output compressed at 17.5 fs. We use the same CEP stabilization technique based on a complete analog acquisition system that we have developed coupled to a fast actuator . Simultaneously digital and analog detection have been performed. A home-made f-to-2f interferometer is used to deliver a beam encoding the CEP variation information that is split into two parts. One fraction is sent to a digital fast spectrometer for long time measurements limited to 1 kHz shot to shot acquisition rate. The interference fringes produced in the f-2f interferometer are spectrally resolved and their displacement in the spectral range is directly related to the CEP variations. The other part is the analog acquisition running at the full laser repetition rate and coupled with the correction loop. Here the shot-to-shot CEP noise is now described by a spatial displacement of the interference fringe pattern. The displacement of one selected fringe can be correlated to the CEP drift and the system generates a corresponding error signal. The latter is sent to a PID controller which generates a correction signal according to the proportional P and integral I values of the correction loop algorithm. The produced feedback signal is applied to the CEP control device which is here the Dazzler. A fast correction closed-loop can thus be achieved without any mechanical displacement allowing for high frequency correction range.
Shot to shot remaining CEP noise of ~350 mrad RMS (Fig. 11(a)) has been measured for the 1 kHz beamline, while it is as low as 260 mrad RMS for the 10 kHz beamline (Fig. 11(b)) over 30 minutes. The 10 kHz acquisition rate gives access to a larger noise frequency range and allows to correct high frequency noise more efficiently than in the 1 kHz case. This larger correction bandwidth and the lower number of amplification stages explain that the 10 kHz CEP residual noise can be quite lower than for the 1 kHz arm. We have to stress that the compressors are installed in the experimental rooms which are separated from the laser room and placed more than 5 meters away from the power amplifiers making those values quite remarkable.
In TNB operation mode the CEP noise should remain the same than in BB mode since the IR energy stability during amplification remains the same and acousto-optic filters are working at a set point close to the BB mode. A difference could appear in the measurement since f-2f interferometry is based on white light generation which can be more or less stable . Pulse duration can then definitely play a role on the CEP measurement and increase arbitrarily the measured CEP noise.
In conclusion, we demonstrated a Ti:Sa based, CEP-stabilized, sub-20 fs, 1 kHz / 10 kHz hybrid laser system with an original front-end design, delivering 0.9 TW peak power pulses at 1 kHz and close to 20 W average power at 10 kHz. We have also demonstrated that the front end can be used to produce pulses with a tunable central wavelength and variable spectral width between 760 and 850 nm. Amplification up to 10 mJ at 1 kHz is experimentally demonstrated conserving the wavelength tunability possibility with spectral widths as narrow as 25 nm. To our knowledge, this is one of the best results reported for kHz TW-class Ti:Sa laser based on grating stretcher/compressor with large stretching ratio, in terms of CEP stability, short pulse duration and wavelength tunability. The FAB 1/10 dual output laser with such quite unique characteristics is already installed and daily running in the French ATTOLab facility. Moreover, the innovative regenerative cavity design in this system shows one of the most promising configurations for future Ti:Sa high intensity CEP stabilized lasers.
H2020 Marie Skłodowska-Curie Actions (ITN-641789-MEDEA); Agence Nationale de la Recherche (ANR11-EQPX0005-ATTOLAB); Conseil Régional Ile-de-France (SESAME2012-ATTOLITE).
1. M. Chini, K. Zhao, and Z. Chang, “The generation, characterization and applications of broadband isolated attosecond pulses,” Nat. Photonics 8(3), 178–186 (2014). [CrossRef]
2. A. Scrinzi, M. Y. Ivanov, R. Kienberger, and D. M. Villeneuve, “Attosecond Physics,” J. Phys. B 39(1), R1–R37 (2006). [CrossRef]
3. M. Nisoli and G. Sansone, “New frontiers in attosecond science,” Prog. Quantum Electron. 33(1), 17–59 (2009). [CrossRef]
4. M. Uiberacker, T. Uphues, M. Schultze, A. J. Verhoef, V. Yakovlev, M. F. Kling, J. Rauschenberger, N. M. Kabachnik, H. Schröder, M. Lezius, K. L. Kompa, H. G. Muller, M. J. J. Vrakking, S. Hendel, U. Kleineberg, U. Heinzmann, M. Drescher, and F. Krausz, “Attosecond real-time observation of electron tunnelling in atoms,” Nature 446(7136), 627–632 (2007). [CrossRef] [PubMed]
5. M. I. Stockman, M. F. Kling, U. Kleineberg, and F. Krausz, “Attosecond nanoplasmonic-field microscope,” Nat. Photonics 1(9), 539–544 (2007). [CrossRef]
6. C.-D. Lin, X.-M. Tong, and T. Morishita, “Direct experimental visualization of atomic and electron dynamics with attosecond pulses,” J. Phys. At. Mol. Opt. Phys. 39(13), S419–S426 (2006). [CrossRef]
7. S. Laulan and H. Bachau, “Correlation effects in two-photon single and double ionization of helium,” Phys. Rev. A 68(1), 013409 (2003). [CrossRef]
8. H. Vincenti and F. Quéré, “Attosecond Lighthouses: How to Use spatiotemporally coupled Light Fields to Generate Isolated Attosecond Pulses,” Phys. Rev. Lett. 108(11), 113904 (2012). [CrossRef] [PubMed]
9. C. F. Dutin, A. Dubrouil, S. Petit, E. Mével, E. Constant, and D. Descamps, “Post-compression of high-energy femtosecond pulses using gas ionization,” Opt. Lett. 35(2), 253–255 (2010). [CrossRef] [PubMed]
10. F. Böhle, M. Kretschmar, A. Jullien, M. Kovacs, M. Miranda, R. Romero, H. Crespo, U. Morgner, P. Simon, R. Lopez-Martens, and T. Nagy, “Compression of CEP-stable multi-mJ laser pulses down to 4fs in long hollow fibers,” Laser Phys. Lett. 11(9), 095401 (2014). [CrossRef]
11. G. Sansone, E. Benedetti, F. Calegari, C. Vozzi, L. Avaldi, R. Flammini, L. Poletto, P. Villoresi, C. Altucci, R. Velotta, S. Stagira, S. De Silvestri, and M. Nisoli, “Isolated Single-Cycle Attosecond Pulses,” Science 314(5798), 443–446 (2006). [CrossRef] [PubMed]
12. H. Mashiko, S. Gilbertson, C. Li, S. D. Khan, M. M. Shakya, E. Moon, and Z. Chang, “Double optical gating of high-order harmonic generation with carrier-envelope phase stabilized lasers,” Phys. Rev. Lett. 100(10), 103906 (2008). [CrossRef] [PubMed]
13. A. Wirth, M. T. Hassan, I. Grguraš, J. Gagnon, A. Moulet, T. T. Luu, S. Pabst, R. Santra, Z. A. Alahmed, A. M. Azzeer, V. S. Yakovlev, V. Pervak, F. Krausz, and E. Goulielmakis, “Synthesized light transients,” Science 334(6053), 195–200 (2011). [CrossRef] [PubMed]
14. S.-W. Huang, G. Cirmi, J. Moses, K.-H. Hong, S. Bhardwaj, J. R. Birge, L.-J. Chen, E. Li, B. J. Eggleton, G. Cerullo, and F. X. Kärtner, “High-energy pulse synthesis with sub-cycle waveform control for strong-field physics,” Nat. Photonics 5(8), 475–479 (2011). [CrossRef]
15. M. T. Hassan, A. Wirth, I. Grguraš, A. Moulet, T. T. Luu, J. Gagnon, V. Pervak, and E. Goulielmakis, “Invited article: attosecond photonics: synthesis and control of light transients,” Rev. Sci. Instrum. 83(11), 111301 (2012). [CrossRef] [PubMed]
17. B. Langdon, J. Garlick, X. Ren, D. J. Wilson, A. M. Summers, S. Zigo, M. F. Kling, S. Lei, C. G. Elles, E. Wells, E. D. Poliakoff, K. D. Carnes, V. Kumarappan, I. Ben-Itzhak, and C. A. Trallero-Herrero, “Carrier-envelope-phase stabilized terawatt class laser at 1 kHz with a wavelength tunable option,” Opt. Express 23(4), 4563–4572 (2015). [CrossRef] [PubMed]
18. D. Busto, L. Barreau, M. Isinger, M. Turconi, C. Alexandridi, A. Harth, S. Zhong, R. J. Squibb, D. Kroon, S. Plogmaker, M. Miranda, Á. Jiménez-Galán, L. Argenti, C. L. Arnold, R. Feifel, F. Martín, M. Gisselbrecht, A. L’Huillier, and P. Salières, “Time–frequency representation of autoionization dynamics in helium,” J. Phys. At. Mol. Opt. Phys. 51(4), 044002 (2018). [CrossRef]
19. A. Zholents, “Next generation X-ray free electron lasers,” IEEE J. Sel. Top. Quantum Electron. 99, 1–10 (2011).
20. S. Backus, C. G. Durfee III, M. M. Murnane, and H. C. Kapteyn, “High power ultrafast lasers,” Rev. Sci. Instrum. 69(3), 1207–1223 (1998). [CrossRef]
23. X. Chen, L. Canova, A. Malvache, A. Jullien, R. Lopez-Martens, C. Durfee, D. Papadopoulos, and F. Druon, “1-mJ sub-5-fs carrier-envelope phase-locked pulses,” Appl. Phys. B 99(1-2), 149–157 (2010). [CrossRef]
24. J.-F. Hergott, O. Tcherbakoff, P.-M. Paul, P. Demengeot, M. Perdrix, F. Lepetit, D. Garzella, D. Guillaumet, M. Comte, P. D. Oliveira, and O. Gobert, “Carrier-Envelope Phase stabilization of a 20 W, grating based, chirped-pulse amplified laser, using electro-optic effect in a LiNbO3 crystal,” Opt. Express 19(21), 19935–19941 (2011). [CrossRef] [PubMed]
25. R. Budriūnas, T. Stanislauskas, J. Adamonis, A. Aleknavičius, G. Veitas, D. Gadonas, S. Balickas, A. Michailovas, and A. Varanavičius, “53 W average power CEP-stabilized OPCPA System delivering 5.5 TW few cycle pulses at 1 kHz repetition rate,” Opt. Express 25(5), 5797–5806 (2017). [CrossRef] [PubMed]
26. D. Strickland and G. Mourou, “Compression of amplified chirped optical pulses,” Opt. Commun. 56(3), 219–221 (1985). [CrossRef]
28. D. F. Hotz, “Gain Narrowing in a Laser Amplifier,” Appl. Opt. 4(5), 527–530 (1965). [CrossRef]
29. T. Oksenhendler, D. Kaplan, P. Tournois, G. M. Greetham, and F. Estable, “Intracavity acousto-optic programmable gain control for ultra-wide-band regenerative amplifiers,” Appl. Phys. B 83(4), 491–494 (2006). [CrossRef]
30. M. Zavelani-Rossi, F. Lindner, C. Le Blanc, G. Chériaux, and J. P. Chambaret, “Control of thermal effects of high-intensity Ti:sapphire laser chains,” Appl. Phys. B 70(S1), S193–S196 (2000). [CrossRef]
31. W. Koechner, Solid-state laser engineering (Springer, 1999) 5th ed.
33. S. Ferre, Caractérisation expérimentale et simulation des effets thermiques d’une chaîne laser ultra-intense à base de saphir dope au titane, PhD thesis, Ecole Polyechnique X, Palaiseau, 2002.
34. S. Chénais, F. Druon, F. Balembois, G. Lucas-Leclin, Y. Fichot, P. Georges, R. Gaume, B. Viana, G. P. Aka, and D. Vivien, “Thermal lensing measurements in diode-pumped ybdoped gdcob, ycob, yso, yag and kgw,” Opt. Mater. 22(2), 129–137 (2003). [CrossRef]
35. A. Golinelli, X. Chen, E. Gontier, B. Bussière, O. Tcherbakoff, M. Natile, P. d’Oliveira, P.-M. Paul, and J.-F. Hergott, “Original Ti:Sa 10 kHz front-end design delivering 17 fs, 170 mrad CEP stabilized pulses up to 5 W,” Opt. Lett. 42(12), 2326–2329 (2017). [CrossRef] [PubMed]
36. F. Verluise, V. Laude, Z. Cheng, C. Spielmann, and P. Tournois, “Amplitude and phase control of ultrashort pulses by use of an acousto-optic programmable dispersive filter: pulse compression and shaping,” Opt. Lett. 25(8), 575–577 (2000). [CrossRef] [PubMed]
37. F. Lücking, A. Assion, A. Apolonski, F. Krausz, and G. Steinmeyer, “Long-term carrier-envelope-phase-stable few-cycle pulses by use of the feed-forward method,” Opt. Lett. 37(11), 2076–2078 (2012). [CrossRef] [PubMed]
38. C. Feng, J.-F. Hergott, P.-M. Paul, X. Chen, O. Tcherbakoff, M. Comte, O. Gobert, M. Reduzzi, F. Calegari, C. Manzoni, M. Nisoli, and G. Sansone, “Complete analog control of the carrier-envelope-phase of a high-power laser amplifier,” Opt. Express 21(21), 25248–25256 (2013). [CrossRef] [PubMed]
39. M. Bradler, C. Homann, and E. Riedle, “Investigation and optimization of continuum generation in crystals - white-light beyond sapphire,” in European Conference on Lasers and Electro-Optics and the European Quantum Electronics Conference (Optical Society of America, 2009). [CrossRef]