Open Access
Add to BibSonomy
Abstract
Performance of the novel high repetition rate HF-PW laser system of ELI ALPS is presented in its first operation phase at 400 TW and 700 TW levels. Long-term operation was tested at 2.5 and 10 Hz repetition rates, where an exceptional 0.66% and 1.08% shot-to-shot energy stability was demonstrated, respectively. Thorough spatio-spectral and temporal measurements confirmed high quality output pulses with a Strehl ratio of >0.9 after compression at both repetition rates. Amplified pulses with an unprecedentedly high 240 W average power were reached for the first time from a PW-class amplifier chain by using novel pseudo-active mirror disk amplification-based pump lasers.
© 2023 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
The development of petawatt lasers has recently moved towards increased robustness, more stable pulse parameters, higher repetition rates and decreased pulse duration to obtain ultrahigh peak intensity on target. BELLA, the first rep-rated petawatt laser was implemented in Berkeley Lab, which after upgrades is able to produce 40 J pulses with 30 fs duration at 1 Hz [1]. Similar performance was achieved in HZDR, where the DRACO PW laser can produce 20 J with 30 fs pulse duration at 1 Hz repetition rate on target [2]. The same 1 Hz repetition rate is reached in case of the 1 PW level compression of the HPLS laser system installed recently in ELI Nuclear Physics, where the 10 PW peak power is also achieved at a 1 shot/min operation [3]. The average power available from a petawatt-class system was increased to 85 W in the Colorado State University, where 30 fs pulses with 26.3 J were reached at 3.3 Hz repetition rate [4]. The HAPLS laser system in ELI Beamlines aims to achieve similar output performance, where 13.3 J pulses are produced currently at 3.3 Hz with >27 fs duration corresponding to 490 TW [5]. All the above mentioned systems are seeded by standard Ti:Sa oscillator based chirped pulse amplification (CPA) systems, mixed with picosecond optical parametric chirped pulse amplification (OPCPA) stages only in case of the 10 PW laser in ELI Nuclear Physics. Seed pulse generation by a highly stable, white light seeded OPCPA system for a 3 J 30 fs laser has been just recently implemented as part of the KALDERA project in DESY, which is planned to be used for laser plasma acceleration at 1 kHz repetition rate [6].
The High Field PetaWatt (HF-PW) laser of ELI ALPS, designed by Amplitude in collaboration with ELI ALPS and implemented by the former, features all the above-mentioned trends of laser development in the PW regime. Consequently, the commissioning of this system represents a new milestone in the world of user-ready petawatt-class lasers [7]. The major aim of the implemented laser infrastructure in ELI ALPS is to provide reproducible and stable sources of ultrashort optical pulses for driving secondary sources of high harmonics, accelerated particles, X-ray radiation, and for performing versatile experiments with unprecedented parameter spaces and control. The HF-PW laser is designed to produce high contrast 34 J energy pulses with 17 fs temporal duration at a repetition rate as high as 10 Hz. The realization of these parameters requires a different approach in terms of design and implementation compared to other petawatt-class laser systems worldwide. To this end, the laser architecture implements an industrial Yb-laser pumped OPCPA-based frontend, several stages of nonlinear frequency conversion, and multipass Ti:Sa amplifiers with spectral corrections. Operation at an unprecedentedly high average power is enabled by special laser heads in the power amplifiers and by using next generation high energy and high average power pump lasers for the last amplifier stage.
2. Design and performance of subsystems
The HF-PW laser is a dual arm system composed of a high energy HF-2PW arm, running at 10 Hz, and a secondary, millijoule level HF-100 arm with carrier-envelope phase (CEP) stabilization running at 100 Hz. Both arms have a common high contrast frontend based on OPCPA technology. The main HF-2PW arm is designed to amplify the stretched seed pulses up to 50 J at 10 Hz repetition rate by using multipass Ti:Sa amplifier stages. After amplification, a beam transport system (HFBTS) transfers the pulses to the highly shielded target area (HTA), where the vacuum compressor is located. This way, amplified pulses are compressed to 17 fs directly before the laser-matter interaction stations generating attosecond pulses and accelerated particles with an output energy of 34 J, resulting in a peak power of 2 PW (Fig. 1). The secondary arm is composed of a mJ level Ti:Sa-based CPA system followed by a spatio-temporal filter utilizing cross-polarized wave (XPW) generation [8]. This allows to eliminate amplified spontaneous emission (ASE) and pre/post pulses generated in the Ti:Sa amplifiers. Running at 100 Hz repetition rate, the laser produces 6 mJ pulses with 23 fs duration. After temporal filtering and recompression with chirped mirrors, the HF-100 arm generates 1 mJ pulses with 10 fs duration and an extremely clean temporal intensity contrast. Detailed description of the system will be given in the following subsections.
Fig. 1. Schematic layout of the HF-2PW dual arm laser system. MPA refers to multipass, RGA to regenerative amplifier, while DM for deformable mirror.
The OPCPA frontend (see Fig. 1) is followed by the stretcher, a 3 J amplifier frontend (amplifiers MPA0 – MPA4) and two power amplifiers (MPA5 and MPA6). Right up to the end of the 3 J amplifier frontend, the laser system always runs at 10 Hz, while the repetition rate of the power amplifiers can be changed between 10 Hz and 2.5 Hz.
At full power, the MPA5 and MPA6 stages of the PW arm are pumped by a set of eight TITAN HE and two PremiumLite Nd:YAG pump lasers delivering a total energy of 32 J and 100 J at 532 nm, respectively.
In the current configuration, PremiumLite pump lasers and the MPA6 stage are not yet fully commissioned. As a consequence, the output energy after compression is limited to a maximum of 10 J with a more relaxed 24 fs pulse duration. Furthermore, the development of the secondary sources requires to decrease the repetition rate of the amplifiers located after MPA4 to 2.5 Hz, which allows for a safer operation of the system in the first phase. Additionally, it also keeps the average power in the grating compressor at a moderate level of 35 W. Within the framework of in-house development, a 120 mm clear aperture, fast shutter with single-shot capability was implemented before the HFBTS (Fig. 1, Fast shutter). Most of the power supplies and chiller units are located on mezzanines above the laser laboratory to decouple the mechanical vibrations from the floor.
2.1 High contrast seeder
One of the major design aspects was to provide a seed source, which offers robust day-to-day operation within a short startup time and without the necessity of daily maintenance. In most high intensity lasers Kerr-lens mode-locked Ti:Sa oscillators are used [1–4], which can generate a spectral bandwidth spanning from 620 to 1100 nm at the 10−2 intensity level [9] with few nJ pulse energy at 70–80 MHz repetition rate and a temporal intensity contrast of typically >109 [10]. These femtosecond oscillators are highly sensitive to environmental changes, and typically require fine tuning every day to obtain the same performance. An additional issue, is that they provide pulses with few-nJ energy, which through amplification in several stages suffer from significant temporal contrast degradation. In contrast, our system starts with a femtosecond Yb-doped fiber oscillator (T-pulse, Amplitude), which produces pulses with 14 nm full width at half maximum (FWHM) spectrum centered at 1029 nm. It runs at 40 MHz repetition rate, with 30 mW average power (Fig. 1, top left area). Seed pulses are fiber-coupled to an Yb:CaF2 thin disk CPA system (S-Pulse HP2, Amplitude), which then generates 500 fs pulses with 1.4 mJ energy at 4 kHz repetition rate and a shot-to-shot energy stability of <0.25% rms. This thin disk laser is an ideal pumping source for an OPCPA, because it provides high stability and about 5.5 W of average power.
In the OPCPA, a 4 µJ energy portion of the thin disk laser pulses is focused into a YAG crystal, which is enclosed in an air-tight compartment to avoid air movement.
This allows us to generate a stable supercontinuum (WLG in Fig. 2), from which the spectral region of 1400-1800nm is selected by a spectral filter. The white-light pulses are compressed by bulk silicon, then overlapped with the 1030 nm pump pulses in a periodically poled lithium niobate (PPLN) crystal to generate difference frequency (Fig. 2, OPA1 stage). The passively CEP-stable idler pulses at 3.2 µm are frequency doubled in an AGS crystal (Fig. 2, SHG1 stage), and are stretched by another silicon crystal. The 1.6 µm pulses are then sent through an acousto-optic programmable dispersive filter (DAZZLER, Fastlite), which provides fine spectral phase tuning and also active CEP control. Energy of the stretched pulses is increased in two consecutive OPA stages (Fig. 2, OPA2 and OPA3) in PPLN crystals, pumped by 400 µJ and 600 µJ energy 1030 nm pulses, respectively. All parametric stages are realized in a collinear geometry, which allows for the highest spatio-temporal quality of the amplified pulses. Finally, the amplified 1.6 µm pulses with >60 µJ energy are compressed by propagation through ZnSe (Fig. 2, ZnSe), and are then frequency doubled in a BiBO crystal, resulting in >10 µJ pulses at 4 kHz repetition rate centered at 800 nm wavelength. The output spectrum is limited to >20 fs Fourier transform limited (FTL) duration due to the collinear geometry of the OPA stages.
Fig. 2. Schematic layout of the OPCPA seeder. WLG refers to the white-light generation stage, DFG to the difference frequency generation stage, SHG1 and SHG2 to the second harmonic generation stages, OPA2 and OPA3 to the optical parametric amplification stages, CMC to the chirped mirror compressor, and SPM to the self-phase modulation-based spectral broadening stage.
A small portion with about 1 µJ energy after SHG2 is sent to the HF-100 arm. In the meantime, pulses with 9 µJ are compressed by chirped mirrors, and spectrally broadened by self-phase modulation in a fused silica plate (positioned at Brewster angle) to obtain a slightly enhanced bandwidth supporting <17 fs FTL. These pulses serve as the seed for the HF-PW arm of the laser system. The key output parameters of the OPCPA are visualized in Fig. 3. Due to the Gaussian near field (NF) profile of the thin disk pump laser, the output NF of the OPCPA also has a Gaussian spatial distribution (Fig. 3(a-c)). The spectrum directly at the output of the OPCPA is shown in Fig. 3(d). The long-term stability of the OPCPA was tested by measuring the output average power for 14 hours, which resulted in a rms stability of 3% for the complete measurement (Fig. 3(e)). It needs to be mentioned, that this long-term stability was reached with only slow drift compensation of the pump beam at the input of the OPCPA, and no internal stabilization loops are implemented at the moment. The temporal intensity contrast was measured with a SEQUOIA HD (Amplitude) device. As a consequence of the high dynamic range required for the measurement, output pulses of the OPCPA required further amplification. For this reason, in a test setup the OPCPA pulses were stretched, amplified in two multipass Ti:Sa stages, and compressed to 1 mJ energy [11]. The measured contrast is inherently ultra-high due to the femtosecond OPA stages, which is further enhanced by frequency doubling two times (Fig. 3(f-g)). The features at –16 ps and –39 ps in Fig. 3(f) are measurement artifacts, consequently they are not real pre-pulses.
Fig. 3. Near-field (a) and far-field (b) spatial intensity distribution of the S-Pulse pump laser, near-field profile of the OPCPA (c), output spectrum of the OPCPA (d), long-term power stability of the OPCPA (e), short range (f) and long range (g) temporal intensity contrast.
2.2 3 J frontend
Seed pulses of the HF-PW arm after the SPM stage are coupled into the double-pass Öffner-type stretcher, where they are expanded temporally to 1 ns. The spectral transmission window of the PW stretcher was designed to be 725–875 nm. It accommodates a bandwidth sufficient to reach 17 fs pulses at the end of the laser at the 2 PW level. To suppress gain narrowing in the first two Ti:Sa amplifiers, stretched pulses with a temporal duration of 1 ns are then shaped (Fig. 4, SPF1) with a spectral filtering mirror [12].
Fig. 4. Spectral and phase shaping configuration of the 3 J frontend. SPF1 and SPF2 are spectral shaping mirrors.
The spectrally shaped seed pulses with an energy of about 1 µJ are then amplified in MPA0, which is a 4-pass amplifier pumped by 10 mJ from an InLite II (Continuum) Nd:YAG laser providing a shot-to-shot stability of <0.5% at 10 Hz repetition rate. After MPA0, the amplified pulses with 200 µJ energy are shaped with a DAZZLER in both spectral phase and amplitude. MPA1 further amplifies the seed pulses in 5 passes to 1 mJ with applying a pump energy of 30 mJ delivered by the same InLite II as in case of MPA0. 25 mJ energy is reached in the MPA2 stage (Fig. 4), which is driven by a second InLite II with 110 mJ energy pump pulses. The efficiency of the MPA0-MPA2 amplifiers is kept low by design to minimize gain shifting and narrowing. Laser pulses are then temporally cleaned by a Pockels cell, which also cuts the remaining part of the 4 kHz pulse train from the OPCPA seeder. A pair of spectral filtering mirrors [12] is used to pre-shape again the long wavelength side of the pulse spectrum with a transmission of 16%, accounting for a transmitted pulse energy of 4 mJ. Main parameters of the first five amplifiers are summarized in Table 1.
Table 1. Main parameters of the MPA0–MPA4 stages.
More efficient amplification is reached in MPA3 with a total gain of 100. This stage is pumped by a PowerLite DLS Plus (Continuum) Nd:YAG laser providing 2 J pulses at 532 nm and 10 Hz repetition rate. The amplified pulse energy is increased to 400 mJ in five passes, after which the polarization and temporal contrast are cleaned with thin film polarizers and a fast Pockels cell. Finally, high energy stability is reached in the 100 TW-class MPA4 stage, which is pumped by four PowerLite DLS Plus lasers, providing a total pump energy of 8 J at 10 Hz repetition rate. After four passes, the amplified pulses reach an energy level of 3 J, which accounts for a gain of 8 in this stage.
Amplifier crystals in MPA3 and MPA4 are water-cooled and mounted in standard edge cooled laser heads. The near-field profile of the amplified pulses is transformed to Super-Gaussian mainly in MPA4 (Fig. 5(b)), which still results in a high quality far-field intensity distribution (Fig. 5(d)). The input pulses of both MPA3 and MPA4 stages are stabilized for long term thermal drifts by using motorized mirror mounts in both near- and far-fields. The far-field profile, i.e., the wavefront of the amplified pulses stabilizes in 10 s and 30 s in MPA3 and MPA4, respectively (Fig. 5(c) and (d)). As a result of the spectral amplitude shaping obtained with the DAZZLER and the filtering mirrors after MPA2, the output spectrum of MPA4 is tuned to the short wavelength side (Fig. 5(e)). Such a spectral intensity distribution leads to a much more balanced spectral evolution in the following power amplifier stages, where deep saturation is reached, and gain shifting will occur [13]. A 55 mm aperture Pockels cell is installed at the output of MPA4 in combination with thin film polarizers, which are able to decrease the repetition rate down including the single shot mode, if required. This setup also serves as the backscattering protection from the laser-matter interaction area.
Fig. 5. Output near-field profiles of MPA3 (a) and MPA4 (b), output far-field profiles of MPA3 (c) and MPA4 (d). Typical output spectrum of MPA4 is shown in (e).
2.3 Power amplifiers
2.3.1 MPA5
The first 3-pass power amplifier stage, MPA5, is pumped with eight TITAN HE type (Amplitude) Nd:YAG lasers, which are based on a dual head architecture. As a standard, both laser heads run at 5 Hz repetition rate, and each of them provides pulses with 5 J of energy. In case of 10 Hz operation of MPA5, the TITAN heads run in the temporal multiplexing mode, when their two 5 Hz pulse trains are combined to create the 10 Hz pump chain.
On the other hand, when ≤5 Hz operation of MPA5 is required, half of the laser heads can be turned off to slow down the degradation of laser components and reduce power consumption. To reach a truly homogeneous spatial distribution of pump intensity on the Ti:Sa crystal, diffractive optical elements are utilized for all eight TITAN lasers, which produce a smooth spatial profile in the image plane, i.e., at the entrance plane of the 120 mm diameter Ti:Sa crystal of MPA5. Seed pulses propagate through the laser head (Fig. 6(a-b)) three times, where the beam size is set to be 55 mm. As a result, the energy at the output is increased to 14 J, while its shot-to-shot stability for 100 seconds was found to be 1.08% rms, measured at 10 Hz repetition rate by using a QE95LP-H-MB-QED-D0 (Gentec) type detector (Fig. 6(d)). The high thermal load originating from the high repetition rate operation is handled by a special laser head solution: the diameter to thickness ratio of the Ti:Sa crystal is 4:1, while cooling is performed from the optical surfaces on both sides (Fig. 6(a-b)). To control the parasitic lasing in the laser head, diiodomethane is applied at the edge of the Ti:Sa crystal. The required output spectrum of MPA5 (Fig. 6(c)) was reached by a combination of spectral filters installed in amplifiers and tuning the spectral pre-shaping with the DAZZLER so that MPA5 would serve as the last amplifier. As a sign of deep saturation, the long wavelength side of the spectrum shown in Fig. 6(c) is significantly elevated compared to the output of MPA4 visualized in Fig. 5(e). Spatial quality of the output beam is demonstrated in Fig. 6(e) and (f), where the output near field and far field profiles are visualized for the 10 Hz operation, respectively.
Fig. 6. Photographic picture of the MPA5 laser head (a), and schematic visualization of its central cross-section (b), where FS denotes fused silica windows. Typical output spectrum of MPA5 is shown in (c), while the energy stability at 10 Hz repetition rate for 100 s is presented in (d). Near and far field spatial profiles of the MPA5 output are shown in (e) and (f), respectively.
2.3.2 MPA6
The last amplification stage, MPA6, is currently under implementation. It is pumped by two prototype PremiumLite (Amplitude) pump lasers, which are able to provide 50 J energy at 532 nm each, at a maximum repetition rate of 10 Hz. They are based on pseudo active mirror laser heads with Nd:YAG ceramics [14], which are pumped by high energy flashlamps. The seed pulses with an energy of 2 J at 10 Hz repetition rate, generated by a PowerLite DLS (Continuum) laser, are first spatially shaped to get an optimal input for the amplifier heads (Fig. 7(c)).
Fig. 7. Schematic view of an amplifier module’s horizontal cross-section (a) and a photo of the same module during simmering (b). TLM refers to material for transverse parasitic lasing mitigation. Schematic arrangement of a PremiumLite laser is shown in (c), where diffraction patterns in the near field profile are caused by contaminations on the neutral density filters in front of the CCD camera.
Amplification is achieved in six disk amplifier modules (DAM), where the first two are used in a double pass mode, while the remaining four in a single pass mode (Fig. 7(a) and (b)). By applying vacuum spatial filtering with an imaging system in between the DAMs, and by implementing a deformable mirror in a double pass configuration, high spatial quality is achieved at the exit of the amplifier system (Fig. 7, near field profile). Finally, the 1064 nm wavelength pulses with 72 J are frequency doubled in a large aperture LBO crystal with a conversion efficiency of 70%. To handle the high thermal load in the laser heads, a 25 l/min flow rate cooling is applied on each of them, where the coolant flows along both optical surfaces of the Nd:YAG ceramic slabs (Fig. 7(a)). The coolant was chosen to be heavy water (D2O), by which absorption of the fundamental wavelength is decreased to about 30% of the value for simple deionized water. Performance at 532 nm was measured to be 51 J at 10 Hz repetition rate in both PremiumLites with a short-term energy stability of 0.9% rms and a pulse duration of 7.2 ns.
The same type of face-cooled laser head, as in MPA5, will be utilized in MPA6 for the thermal management of the 1 kW average power pumped gain medium. Output pulses of MPA5 are first increased in beam size to 80 mm by a Galilei telescope, then amplified in two passes through the 120 mm diameter Ti:Sa crystal (Table 2). Pump pulses with 100 J in total from the PremiumLite lasers’ output are imaged onto the Ti:Sa crystal surface. Wavefront distortions of the amplifier system are corrected by a 150 mm diameter deformable mirror (ILAO STAR, Imagine Optic), set as the first mirror of MPA6. Leakage of a turning mirror at the output of MPA6 is sampled for focal spot and wavefront characterization by a wavefront sensor (HASO4, Imagine Optic). A tunable afocal telescope consisting of two singlet lenses is used after MPA6 for fine control of collimation of the amplified beam. After MPA6 for alignment of secondary sources a Fresnel reflection-based attenuator consisting of four uncoated wedges can be installed in the laser chain. This reduces the pulse energy to <5 mJ with a maximum attenuation of 4 orders of magnitude. The attenuator is mounted on a movable frame, by which it can be moved in, or out of the beam path keeping the beam position and direction unchanged.
Table 2. Main parameters of the power amplifier stages.
2.4 Beam transport and compressor area
At the output of the laser amplifier, the beam is magnified from >80 mm to >110 mm before entering the air section of the beam transport system, located in the HF-PW laser laboratory. A fast shutter enabling single-shot operation even at 10 Hz repetition rate was internally developed. The shutter is based on a linear translation stage and reflective optical elements, and it was installed in the first periscope box (PRC1 on the left side of Fig. 8). The pulses blocked by the shutter are sent to a water-cooled beam dump with the capability to handle the full 50 J energy at 10 Hz repetition rate. Due to radiation protection requirements, two chicanes with periscope assemblies mounted behind the 2 m thick concrete wall at the entrance of the highly radiation shielded target area (HTA) were added to the propagation chain. After the beam is transported to the HTA in air, pulses enter the vacuum section of the beam transport system, where an imaging telescope resizes the beam to the original >80 mm diameter (Fig. 8).
The amplified pulses are magnified to their final beam size of 22 cm just before the vacuum compressor, by using an aberration corrected off-axis reflective telescope. When the pulse enters the vacuum chamber with a vacuum level of 2·10−7 mbar, it is compressed by using a four-diffraction-grating setup combined with a retro-reflector mirror pair, which shifts the beam in the vertical direction. The diffraction gratings have a line spacing of 1,480 lines/mm, where the smaller aperture ones with a size of 485 × 300 mm are manufactured by Plymouth Grating Laboratory, while the larger ones with 650 × 300 mm are produced by Horiba. Compressed pulses then propagate to a turning chamber, where the last turning mirror of the laser system is used to deliver the beam to the entrance of the experimental beamlines. Distance, rotation angle, and the turning mirror tip/tilt movements are motorized inside the vacuum chambers for the fine alignment of the optical arrangement. In the current configuration, the compressor gratings are not cooled.
The output of the system is used in two modes of operation. In mode 1 (M1), the turning mirror in the turning chamber (TC1) is in the beam, and the pulses are delivered to the secondary sources, either with attenuated or with full energy pulses. This way, the calibrated transmission of the turning mirror propagates to the metrology bench (Fig. 9). In mode 2 (M2), the pulses are attenuated after the power amplifier to the few-mJ level, and the turning mirror is removed from the beam path. Consequently, the pulses propagate directly to the metrology bench. To provide a beam with acceptable size for the diagnostics, an off-axis telescope is installed before the metrology bench with a demagnification factor of 12:1, which decreases the beam size to 18 mm. Further demagnification is provided by different telescopes for each diagnostic device.
Fig. 9. Schematic layout of the compressor area. TC1 and TC2 refer to turning chambers, and PW-BD to petawatt beam dump.
The in-house developed PW beam dump (Fig. 9, PW-BD) is the final element of the HF-PW system. It is capable to accept full aperture 2 PW pulses for the case of beam diagnostics and testing the alignment of the high aperture beam attenuator installed behind the final amplification stage of the laser.
2.5 HF-100 frontend
The secondary arm of the HF-PW laser, the HF-100 frontend is seeded by 1 µJ pulses directly from the OPCPA. The seed pulses are then stretched by a standard Öffner-type stretcher, after which the repetition rate is reduced by a Pockels cell to 1 kHz, and the pulses pass through a DAZZLER for compensation of the high order spectral phase. The stretched pulses are injected into a Ti:Sa regenerative amplifier, in which an acousto-optic programmable gain filter (MAZZLER, Fastlite) shapes the amplified spectrum in every pass of the cavity. The regenerative amplifier is pumped by 8 mJ pulses from a Nd:YLF based pump laser (Terra, Continuum) at 1 kHz and 527 nm with an energy stability of <0.5% rms in the short term. After polarization cleaning and repetition rate reduction with a Pockels cell to 100 Hz, the 0.5 mJ pulses are sent to a multipass amplifier, which increases the energy to 9 mJ. The multipass amplifier is pumped with the remaining portion of the pulse energy from the first Terra laser, where the repetition rate is reduced to 100 Hz with a Pockels cell, and with another 100 Hz Terra laser, accounting for a total pump energy of 30 mJ. After polarization cleaning, 8 mJ pulses are coupled into a Treacy-type reflective diffraction gratings - based compressor positioned in air. The key output parameters of the HF-100 arm are presented in Fig. 10.
Fig. 10. Output near and far-field profiles of the MPA (a and b), output spectrum of the MPA (c, blue) and XPW stage (c, orange). Energy (d) and FWHM pulse duration (e) stabilities are shown for 1- and 2-minute time windows, respectively. The temporal contrast after the XPW stage is shown for −50 to 50 ps (f) and -500 to 150 ps (g) time windows of the same measurement.
Before temporal ‘cleaning’ by XPW, the spatial mode of the beam is improved. The compressed pulses with 6 mJ energy and <25 fs duration at 100 Hz repetition rate are focused into a vacuum tube, where they are focused in a metallic pinhole. Then, the pulses are temporally cleaned and spectrally broadened in an XPW stage using BaF2, after which the beam is collimated in air, and sent through a tunable wedge pair and a chirped mirror compressor for dispersion compensation. At the end of the system, 1 mJ pulses with 10 fs duration are obtained with a very clean temporal contrast. In Fig. 10(f), the intensity pulse shape shows 3 pre-pulses in the temporal range of –21 to –9 ps. The pre-pulse at −21 ps is the frequency doubled replica of the corresponding post-pulse created by SEQUOIA. The ones at −9 ps and −13 ps are real pre-pulses. Removing the latter pre-pulses will require further work to be carried out on the removal of their post-pulse origin. Based on the performance presented in Fig. 10, this frontend is an ideal source of seed pulses for further amplification in subsequent amplifiers at 100 Hz repetition rate.
3. Performance at 400 TW level
3.1 Spatio-spectral and temporal characterization at 10 Hz
To confirm, that the spectral content of the pulse is homogeneous in the whole cross-section of the beam, we measured the 2D spectrum in both horizontal and vertical directions by using a MISS-L-B imaging spectrometer (Femto Easy) positioned in the image plane of the compressor output (Fig. 11(a)). We found a good agreement between the on-axis and integrated spectrum. It must be noted, that this measurement was performed after the down-collimation of the attenuated beam with the factor of 24:1, and after imaging the output of the compressor to diagnostics. This means that spatial distortions are much more pronounced than in the real beam. Nevertheless, spectral homogeneity [15] over the aperture of the beam is >0.87 in the central 80% of the full diameter (Fig. 11(b)). At the edges of the beam, the spatio-spectral manifestation of the gain shifting was captured, which is characteristic for high energy power amplification stages [16,17].
Fig. 11. Spatially resolved spectrum in the horizontal cross-section of the beam, measured in an image plane of the compressor (a). The spectrally integrated beam profile (black) and the spectral homogeneity (orange) is shown in (b), while spatial phase measured by the wavefront sensor is visualized in (c).
Spatial quality beyond the beam profile and spatio-spectral distribution measurements was further investigated by measuring the output wavefront of the complete laser system in the metrology bench. The measurement was performed by using an interferometric type wavefront sensor (SID4, Phasics) in a dedicated image plane, which resulted in a Strehl ratio of 0.9 at 1/e2 intensity level of the beam (Fig. 11(c)). In addition to energy stability, experiments also require the beam to be stable in the focal plane. For this reason, we investigated the pointing stability in the focal plane of an 0.7 m focal length lens after the 12:1 demagnifying telescope, accounting for an effective focal length of 8.4 m. By using a Gentec Beamage device, beam pointing was recorded for 100 shots, and the rms value was found to be 8.5% of the beam divergence, which is an excellent value for such a large-scale laser system.
Temporal characterization of PW-class lasers after the vacuum compressor over the full aperture is far from straightforward due to the large demagnification factor required to obtain a beam of acceptable size for any diagnostic device. Typically, the applied telescope is of Galilei type to avoid nonlinear propagation in air by using an intermediate focus, hence adding a large amount of virtual propagation to the recollimated pulses. This results in significant spatial degradation even after centimeters of propagation. Thus, one needs to image the output plane of the compressor on any type of spatial diagnostic measurements. In the meantime, in temporal diagnostic devices there is propagation in the device itself, which makes it nearly impossible to provide a high-quality beam with homogeneous intensity distribution for such diagnostic apparatus. Reliable measurement of the pulse duration was achieved by coupling out the compressed pulses before the high demagnification factor telescope, and by applying spatial filtering in both near and far-fields, where a total energy of <15 µJ was propagated through the focusing arrangement. This mitigated the emergence of significant nonlinear phase accumulation, while it provided enough pulse energy with a nice Gaussian spatial distribution to drive a single-shot D-scan device (D-shot R, Sphere Ultrafast Photonics). Typical pulse characterization results performed on the center of the compressed beam are shown in Fig. 12.
Fig. 12. Results of the D-shot measurement: measured D-shot trace (a), retrieved trace (b) with 1.8% rms retrieval error, measured and reconstructed pulse spectrum with retrieved spectral phase (c), FTL and retrieved temporal profile (d).
As a comparison, compressed pulses were also measured by using a standard Wizzler (Fastlite) device. Results of this measurement are in good agreement with the D-shot values.
Temporal intensity contrast of the compressed pulses was measured in the metrology bench by using a SEQUOIA (Amplitude) third order cross-correlator, where pulses with an energy of 300 µJ were sent through the input iris of the device. After confirmation of pulse compression, and signal maximization, the full available temporal range of the SEQUOIA was scanned through (Fig. 13). As it can be observed in Fig. 13, the picosecond pedestal reaches the 10−10 intensity at –30 ps, which already shows the advantage of using the OPCPA system as a seeder. The origin of the measured pre-pulses was identified, however to define their physical nature we will conduct a nanosecond range scanning measurement as soon as a suitable device becomes available. It is worth to mention that the short pre-pulses will be to much extent mitigated by the plasma mirror which is the first element of the beam transport system to secondary sources. Since it does not belong to the PW laser, we do not consider it here.
Fig. 13. Temporal intensity contrast of the complete laser system at 14 J energy output of MPA5, and the identified origin of the pre-pulses: −50 to 50 ps range (a) and full temporal range (b) of the same measurement. NA refers to not known.
3.2 2.5 Hz operation
As it was mentioned in the Design section, MPA5 can be configured to run either at 10 Hz or at 2.5 Hz repetition rate. The latter operation mode is important for the development of secondary sources, and it is also favorable for preserving the laser status for a longer time. It reduces energy consumption also. Parameters like the spectrum, pulse duration, spatio-spectral distribution, and the temporal contrast were measured to be basically identical in the 10 Hz and 2.5 Hz operation modes. As mentioned in the section related to the power amplifier, the laser system implements a tunable collimator and deformable mirror, which allow to compensate for the difference of the heat flow in amplifiers for different repetition rates. Thus, the wavefront at the end of the laser was also found to be similar at 10 Hz and 2.5 Hz, while the pointing stability was slightly enhanced in the latter case.
Energy stability at 2.5 Hz was measured after MPA5 for a 100-minute time period, where only a slight degradation of the detector sensitivity due to its heating was found with cross-calibration of the NF measurement. Results of the energy logging of MPA5 are presented in Fig. 14. Pumping with eight lasers had a very positive effect on the energy stability of MPA5, which was actually further improved during operation at 2.5 Hz, resulting in a 100-minute shot-to-shot stability of only 0.66% rms. Spectral stability was also analyzed at the output of MPA5 by continuous logging of the output amplified spectrum. For this measurement, we used a wavelength and amplitude calibrated Qmini VIS (Broadcom) fiber coupled spectrometer.
Fig. 14. Directly measured output energy (a) and spectrum (b) of MPA5 with its probability on a 100-minute timescale at 2.5 Hz repetition rate.
The spectrum was logged with a 1 Hz frame rate. The results of the measurements are presented in Fig. 14(b). Based on the probability plot of the spectral intensity in Fig. 14, spectral stability is rather homogeneous in the whole wavelength range covered by the amplified pulses.
4. Operation at 700 TW level at 10 Hz
As a part of the commissioning phase of MPA6, PremiumLite lasers were tested at moderate energy for pumping the 120 mm diameter Ti:Sa crystal in the two pass amplifier stage. Due to limitations on the pulse energy of the beam transport between the pump lasers and the Ti:Sa medium, only 12 J energy was delivered to the amplifier crystal from each PremiumLite. At the time of this first operation of MPA6, we used a standard water-cooled laser head, where the edge of the crystal is cooled. Performance of MPA6 at 10 Hz repetition rate is summarized in Fig. 15. The long wavelength side of the spectrum develops further during amplification in MPA6 (Fig. 15(a)), which still lacks any high frequency modulations. Energy stability at 24 J was measured for 30 minutes of operation, and it was found to be 0.68% rms shot-to-shot. Output near field profile was showing high homogeneity (Fig. 15(e)), while thermalization of the laser head took about 30 minutes, after which a far field profile presented in Fig. 15(f) was reached.
Fig. 15. Amplified pulse spectrum at the output of MPA6 at 24 J (a), energy stability during 30 minutes of continuous operation with 0.68% rms stability, reconstructed spectrum, phase (c) and temporal intensity distribution (d) after compression with attenuated pulses measured by Wizzler. Output near field profile of MPA6 is shown in (e), while the far field is visualized in (f). A photograph of the operating MPA6 stage is presented in (g).
By optimizing the wavefront with the deformable mirror at the entrance of MPA6, a Strehl ratio of 0.91 was obtained before the HFTBS. It needs to be noted, that the face-cooled laser head in MPA5 pumped by 32 J takes less than 5 minutes to thermalize, which shows the superiority of the new type of head in terms of thermal performance. Increasing the pumped area on the other hand was already shown in [18] to improve the cooling performance in face-cooled geometries. For this reason, we believe, that using this type of laser head in MPA6 will significantly improve the thermal performance, especially at full energy pumping with 90 J in the near future. Since the high average power of 240 W at 10 Hz injected to the compressor in a long time period can lead to heating of diffraction gratings and the consequent degradation of beam and pulse properties, compression of the 24 J pulse was tested by using attenuation of the beam after amplification. An energy level of 100 mJ was sent to the vacuum compressor during the demonstration of compression. We measured <23 fs pulse duration by using a standard Wizzler device, where the FTL was 22.9 fs, thus we reached complete compression. By compressing the full energy peak power of the output pulses would reach 740 TW, where the average power of the compressed pulses would be as high as 170 W. Despite of this issue it is worth to mention that, the fast shutter installed at the exit of the laser amplifier chain allows to sample laser pulses at 10 Hz and form a burst mode of 10 Hz pulses, if required.
5. Conclusions and outlook
In conclusion, we presented the design of the HF-PW laser system of ELI ALPS, the performance of larger subsystems, and the stability of output pulse parameters. As of mid-2023, the laser was tested at 10 Hz and has already been operating for 6 months at the 400 TW level at 2.5 Hz repetition rate and with 24 fs pulse duration.
In the next step, the MPA6 amplification stage will be commissioned by driving it with the novel PremiumLite pump lasers at 2.5 Hz repetition rate. After the first tests at the 700 TW level, the next step is to ramp up the amplified pulse energy to 35 J in MPA6, and hence to reach a compressed pulse energy of 25 J with a peak power of >1 PW after the compressor. Here we must take into account that the targetry for PW-class laser pulses requires significant development to adapt even to the 2.5 Hz repetition rate, not to mention the final 10 Hz value.
Furthermore, we plan to upgrade our OPCPA frontend for a next generation arrangement. More specifically, the bandwidth directly available from the OPCPA will be increased to support 10 fs FTL pulses, while the energy will be increased to >0.5 mJ at a repetition rate of 1 kHz. By injecting 50 times higher energy to the Öffner stretcher, we expect significant improvement in the leading front pedestal of the main pulse, resulting in a greatly increased temporal intensity contrast. This will help us to decrease the number of Ti:Sa amplifiers in the 3 J frontend, which will also have a positive impact on the spectral management, and on the available shortest pulse duration at the end of the system.
Funding
European Regional Development Fund (GINOP-2.3.6-15-2015-00001).
Acknowledgments
We greatly acknowledge the contribution of Károly Osvay, Vladimir Chvykov, Huabao Cao, Nikita Khodakovskiy, Julien Houzet, Mathieu Paurisse, Alexandre Thai, Catalin Neacsu, Djamel-Dean Braham, Vincent Trubert, Pierrick Leroy, Ivan Delgado, Hugo Caherec, Abdollah Malakzadeh, Árpád Mohácsi, Arnold Farkas, Frederic Estable, David Armier, and Gilles Riboulet to this project.
Disclosures
The authors declare no conflicts of interest.
Data availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
References
1. K. Nakamura, H. -S. Mao, A. J. Gonsalves, et al., “Diagnostics, Control and Performance Parameters for the BELLA High Repetitiosn Rate Petawatt Class Laser,” IEEE J. Quantum Electron. 53(4), 1–21 (2017). [CrossRef]
2. U. Schramm, M. Bussmann, A. Irman, et al., “First results with the novel petawatt laser acceleration facility in Dresden,” J. Phys.: Conf. Ser. 874, 012028 (2017). [CrossRef]
3. F. Lureau, G. Matras, O. Chalus, et al., “High-energy hybrid femtosecond laser system demonstrating 2 × 10 PW capability,” High Power Laser Sci. Eng. 8, e43 (2020). [CrossRef]
4. Y. Wang, S. Wang, A. Rockwood, et al., “0.85 PW laser operation at 3.3 Hz and high-contrast ultrahigh-intensity λ = 400 nm second-harmonic beamline,” Opt. Lett. 42(19), 3828–3831 (2017). [CrossRef]
5. https://www.eli-beams.eu/facility/lasers/laser-3-hapls-1-pw-30-j-10-hz/
6. T. Eichner, T. Hülsenbusch, J. Dirkwinkel, et al., “Spatio-spectral couplings in saturated collinear OPCPA,” Opt. Express 30(3), 3404–3415 (2022). [CrossRef]
7. N. C. Danson, C. Haefner, T. Bromage, et al., “Petawatt and exawatt class lasers worldwide,” High Power Laser Sci. Eng. 7, e54 (2019). [CrossRef]
8. A. Jullien, O. Albert, F. Burgy, et al., “1010 temporal contrast for femtosecond ultraintense lasers by cross-polarized wave generation,” Opt. Lett. 30(8), 920–922 (2005). [CrossRef]
9. https://novantaphotonics.com/product/venteon-ultrafast-femtosecond-lasers/
10. G. Cheriaux, G. Rey, S. Ferre, et al., CLEO Conference (IEEE, 2002), pp. 51–52.
11. A. Thai, E. Gontier, C. Ferchaud, et al., “High contrast CEP-stable OPCPA front-end for PW-class Ti:Sapphire system,” in International Conference on Ultrafast Phenomena (Optica Publishing Group, 2016), paper UTh4A.25.
12. F. Giambruno, C. Radier, G. Rey, et al., “Design of a 10 PW (150 J/15 fs) peak power laser system with Ti:sapphire medium through spectral control,” Appl. Opt. 50(17), 2617–2621 (2011). [CrossRef]
13. C. Le Blanc, P. Curley, and F. Salin, “Gain-narrowing and gain-shifting of ultra-short pulses in Ti: sapphire amplifiers,” Opt. Commun. 131(4-6), 391–398 (1996). [CrossRef]
14. F. Falcoz, “220W average power reached on a PW TiSa laser,” Proc. SPIE PC11980, PC119800A (2022). [CrossRef]
15. J. Weitenberg, A. Vernaleken, J. Schulte, et al., “Multi-pass-cell-based nonlinear pulse compression to 115 fs at 7.5 µJ pulse energy and 300 W average power,” Opt. Express 25(17), 20502 (2017). [CrossRef]
16. A. Jeandet, A. Borot, K. Nakamura, et al., “Spatio-temporal structure of a petawatt femtosecond laser beam,” J. Phys. Photonics 1(3), 035001 (2019). [CrossRef]
17. A. Jeandet, S. W. Jolly, A. Borot, et al., “Survey of spatio-temporal couplings throughout high-power ultrashort lasers,” Opt. Express 30(3), 3262–3288 (2022). [CrossRef]
18. R. S. Nagymihaly, H. Cao, D. Papp, et al., “Liquid-cooled Ti:Sapphire thin disk amplifiers for high average power 100-TW systems,” Opt. Express 25(6), 6664–6677 (2017). [CrossRef]
