Open Access
Add to BibSonomy
Abstract
We compare multiple temporal pulse characterization techniques in three different pulse duration regimes from 15 fs to sub-5 fs, as there are no available standards yet for measuring such ultrashort pulses. To accomplish this, a versatile post-compression platform was developed, where the 100 fs near infrared pulses were post-compressed to the sub-two-cycle regime in a hybrid, three-stage configuration. After each stage, the duration of the compressed pulse was measured with the d-scan, TIPTOE and SRSI techniques and the retrieved temporal intensity profiles, spectrum and spectral phases were compared. Spectral homogeneity was also measured with an imaging spectrometer to understand the input coupling conditions of the temporal measurements. Our findings suggest that the different devices give similar results in terms of temporal intensity profile, however they are extremely sensitive to alignment and to beam quality, especially in the case of the shortest pulses. We address specific steps of measurement procedures, which paves the way towards the standardization of pulse characterization in the near future.
© 2024 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
Applications in attosecond science, including high harmonic generation [1–3] or ultrafast spectroscopy [4], are based on ultrashort laser pulses generated in Chirped Pulse Amplification (CPA) systems. Although state-of-the-art laser systems are able to produce few-cycle laser pulses [5,6], the output spectrum, and consequently the pulse duration, are still limited by the bandwidth of the gain medium and the stretcher-compressor arrangement, especially in Ti:Sa systems, where the typical pulse duration is around 25 fs. Isolated attosecond pulse generation, on the other hand, requires few-to-single cycle pulses [7]. Currently, post-compression setups such as hollow-core fibers (HCF) [8–10], multiple thin plate configuration [11–13] or multipass cells [14–16] are widely used at the output of high energy lasers to increase the spectral bandwidth and thus reduce pulse duration after dispersion compensation. As these techniques are based on nonlinear effects [17], mostly on self-phase modulation (SPM), they inevitably lead to strong spectral modulations and complex high-order phase contributions, which strongly affect the temporal quality, especially at high compression factors [18]. In addition, the final spectrum and compressibility are determined by many parameters ranging from the initial pulse duration and central wavelength through nonlinear processes.
For applications, the first and most important step is the precise temporal characterization of the driver pulses. This is still a challenging task, albeit many different techniques have been developed in the last decades, e.g. attosecond streaking [19] and its air compatible variants [20,21], frequency resolved optical gating (FROG) [22], spectral phase interferometry for direct electric-field reconstruction (SPIDER) [23], self-referenced spectral interferometry (SRSI) [24], dispersion-scan (d-scan) [25,26] or tunneling ionization with a perturbation for the time-domain observation of an electric field (TIPTOE) [27]. Most of these methods are based on the response of a nonlinear medium, utilizing for example, second-harmonic (SHG), sum-frequency (SFG) or cross-polarized wave generation (XPW) signals. Since these instruments operate either in the spectral or the temporal domain, the optical processes and the reconstruction algorithms also differ. Although the bandwidth is usually limited, the most advanced devices can cover octave-spanning spectra and can thus measure single-cycle pulses. Typically, users have only one piece of equipment at their disposal to measure the above-mentioned complex pulse shapes and phases. As a compressor cannot compensate any arbitrary phase, it is crucial to determine not just the pulse duration but also the residual higher-order dispersion as precisely as possible. This allows us to better control the experimental conditions or further optimize the pulse compression with additional techniques, for example with an acousto-optic programmable dispersive filter (Dazzler, Fastlite) after feedback.
Here we specifically focus on techniques which are able to diagnose few-cycle pulses with complex spectral amplitude and phase profiles, and are becoming available in most laser laboratories working with high intensity pulses. The d-scan technique is suitable for the characterization of pulses with complex power spectra and phase. While compressing ultrashort pulses with chirped mirrors and a pair of glass wedges, fine tuning of the wedges helps to reach maximum compression. Around optimum glass insertion, the pulses can be fully characterized by measuring the fundamental and the second-harmonic spectra with a numerical iterative algorithm [25,26]. The main advantages of this technique are the simplicity of the optical setup and therefore ease of alignment, as well as high sensitivity, since this technique uses directly all the pulse energy (there is no need for pulse replicas). The main drawback can be phase matching limitation during SHG generation. Due to the stringent phase-matching requirements of SHG, the d-scan technique can work over a few hundred nm in the near-IR, to a lesser extent in the visible and to an even lesser extent in the UV. Commercial d-scan devices are not available for the Mid-infrared (Mid-IR) and long-wave infrared (LWIR). SHG-based devices also require angle tuning, when the wavelength is tuned. Lack of appropriate SHG crystals limits the minimum operating wavelength of d-scan to about 400 nm. Another issue occurs during the measurement of octave-spanning spectra, where there is an overlap between the fundamental and the second harmonic spectra. In this case, the measured signal must be filtered. The d-scan technique can be realized in a multi-shot or a single-shot apparatus; the latter is especially useful for low repetition rate laser systems.
In SRSI the reference pulse can be self-created e.g. by the XPW process, where the second order spectral phase is reduced by a factor of nine. For higher order phase terms this is even more efficient. For a nearly transform limited input pulse, the pulse duration will shorten and the spectrum will broaden, therefore the XPW generated signal can be a reference pulse with a broad spectrum and a nearly flat phase [24]. The spectral phase difference between the laser pulse and the reference is measured by Fourier-transform spectral interferometry (FTSI). Both pulses are directed into a spectrometer with a certain optical delay. The resulting spectral interference pattern allows for the direct retrieval of the spectral phase and amplitude. The advantage of this technique is single-shot operation. The main drawback is that XPW generation broadens the spectrum only if the input pulse has no significant chirp. Commercially available devices (e.g. Wizzler (Fastlite) exploiting the above-described technique are not accessible for Mid-IR, LWIR and wavelengths below ∼400 nm.
The principle of TIPTOE is based on the direct sampling of the electric field of a weak pulse (signal) that perturbs the ionization introduced by a strong pulse (fundamental). The technique exploits the extreme nonlinearity of ionization in gaseous media, hence it can be applied in ambient air or in other gases [27]. The signal laser pulse is sampled from the fundamental pulse and continuously varying time delay is applied between the two pulses. Both beams are focused in the middle of two metallic plates, that are connected to a current measuring device. The fundamental pulse ionizes the gaseous media while the signal pulse is too weak to cause ionization. However, the interference of the beams modulates the ionization yield that has information on the temporal shape. The benefits of this technique are the absence of damage limitations in the nonlinear media, as well as the ease of alignment due to the collinear setup. Furthermore, since tunneling ionization occurs at any wavelength, this technique (using suitably coated optics) can work in a spectral range spanning from 200 nm to 20 µm. While the d-scan and SRSI techniques operate in the spectral domain, TIPTOE works in the temporal domain. Due to sampling criteria, the spectral reconstruction in TIPTOE lacks the accuracy present in the other two techniques. Good compression of the input pulses is usually required for the precise measurement for TIPTOE too. The above-described techniques are utilized for single-cycle, few-cycle and multi-cycle pulse measurements.
Here we report on the systematic comparison between temporal characterization of few-cycle post-compressed pulses with three different techniques. The 100 fs pulses from a Ti:Sa laser system were compressed down to 1.7-cycle duration in three stages in a hybrid configuration, utilizing a stretched flexible hollow core fiber (SF-HCF) and multiple thin plates. The diversity of the parameters such as the nonlinear material, accumulated B-integral or spectral bandwidth gave us the opportunity to compare not only three measurement methods at the same time but also to test them under various conditions by implementing them after each stage. We demonstrate, that the different techniques can give comparable results in a certain range, but we also reveal their limitations.
2. Experimental setup
Here, we briefly present the experimental setup used for our measurement campaign, while a more detailed description can be found in the Supplement 1. The seed pulses for post-compression were generated in a Ti:Sa CPA system (Arco, Amplitude). The central wavelength, pulse duration, energy and repetition rate were 782 nm, 100 fs, 4 mJ and 1000 Hz, respectively (Fig. 1).
Fig. 1. Scheme of the experimental setup. HWP: half-wave plate, TFP: thin film polarizer, SF-HCF: stretched flexible hollow core fiber, M1-M5: concave mirrors, BS: removable beam sampler, FS: fused silica thin plates, CMC1-3: chirped mirror compressors, I1-I2: Iris.
In the first stage, the seed pulses with 1 mJ energy were coupled into a 3 m long SF-HCF (Laser-Laboratorium Göttingen e.V.) [28]. The SF-HCF was differentially pumped [29] using argon as a nonlinear medium. The pressure was set to 1170 mbar and <2 mbar (i.e. the measurement limit of our pressure gauge) at the output and input, respectively. After being propagated in the SF-HCF, the pulses were compressed in a chirped mirror compressor (CMC1) consisting of ten 1” diameter chirped mirrors (GSM030, Layertec) arranged in a double pass configuration. At this position (Diagnostics 1 in Fig. 1), the pulse duration was measured with Wizzler USP4 (Fastlite), d-cycle XR (Sphere Ultrafast Photonics) and TIPTOE (SourceLAB) devices.
For the second post-compression stage, we found that the optimal energy is 70 µJ to avoid a significant amount of SPM or any ionization in air. The beam was focused into two 200 µm thick FS plates. Then the collimated beam was sent to the second compressor (CMC2) built of four pairs of double-angle chirped mirrors (PC1332, Ultrafast Innovations). From this point, an s-shot B (Sphere Ultrafast Photonics) was used instead of the d-cycle XR to cover the much broader spectrum spanning from 600 nm to 950 nm, resulting in 7 fs pulse duration.
Due to the losses on the plates, on the chirped mirrors and on the spatial filter, 60 µJ energy pulses were sent to the last stage. The beam was focused onto two FS plates with a thickness of 200 µm, and one 150 µm thick plate. The pulses were compressed with two pairs of chirped mirrors, identical to those used in CMC2.
After each compression stage, we characterized the spectrum and the spatio-spectral homogeneity with an imaging spectrograph (MISS-Large-Broadband, Femto Easy) and recorded the near-field beam profile with a beam profiler camera (Beamage-4 M, Gentech). As these devices require very low energy, the beam was attenuated by reflection from a glass wedge and using additional neutral density filters in front of the camera when it became necessary.
3. Results
3.1 SF-HCF
Good pulse compressibility and proper temporal characterization are only possible when fairly homogenous spectral broadening is obtained with a high-quality beam profile. We observed that the latter one is crucial for the temporal diagnostic devices, hence we first recorded the spatial and the spatio-spectral profiles. Figure 2 shows the spatially resolved intensity spectrum measured with the wavelength and amplitude calibrated imaging spectrograph across the horizontal (Fig. 2(a)) and vertical directions (Fig. 2(b)). The 2D patterns show a slightly tilted structure along the y axis, the origin of which is not known exactly, but any further analysis would be beyond the scope of this paper. Spatio-spectral homogeneity [30] (Fig. 2(c) and (d)) was also calculated for both cross-sections, which indicates uniform spectral broadening across the beam profile in the SF-HCF.
Fig. 2. Spatially resolved spectrum on logarithmic scale (a) in the horizontal, and (b) in the vertical cross-section of the beam. The related (c, d), spectrally integrated spatial profile (grey) and spectral homogeneity (orange) after CMC1. Signal drops along the spatial axis both in (c) and (d) are due to contaminations on the entrance slit of the imaging spectrograph.
The near-field beam profile after CMC1 (Diagnostics 1, Fig. 1) can be seen in the inset of Fig. 3. As expected, at the fiber output we obtained a circular beam profile preserving the fundamental EH11 mode. The rings around the main beam come from diffraction on the spatial filter and their energy content is estimated to be below 5%. All temporal diagnostic tools were placed at a distance of 50 cm from the output of CMC1. Before each measurement, special care was taken to align the collimated beam parallel to the optical table and perpendicular to the entrance plane of the measurement device. This alignment was checked every time when we exchanged the diagnostic tools. With the help of the HWP and TFP pair we set the optimal pulse energy for all devices without changing any temporal or spectral feature of the beam. Figure 3 shows the compressed reconstructed temporal shapes after CMC1. The retrieved pulse durations are very close to each other: d-cycle XR, TIPTOE and Wizzler gave 13.5, 13.6 and 15.8 fs, respectively. It is clearly visible that the compression is not perfect, probably due to the residual higher order phase and each device revealed these temporal features next to the main peak. Here, we present the TIPTOE data obtained without using the measured spectrum for reconstruction. In Section 3.4 we will compare separately the slight differences in the temporal reconstruction when we used a spectrum recorded previously by another device.
The validity of the reconstructed temporal shape must be checked with the spectrum and spectral phase together. In Fig. 4 it is visible that in each case the spectral bandwidth is very similar to the measured one and the d-cycle XR could resolve the spectral features very well. It seems that the reconstructed spectrum of Wizzler is cut at 850 nm, which could lead the slightly longer pulse duration. In each case, the phase is near zero at the central wavelength but Wizzler indicates negatively chirped pulses contradictory to the other two devices. We experienced that during these measurements only the d-scan technique could consistently predict the sign of the residual group delay dispersion (GDD). In case of Wizzler USP4, the settings of the retrieval algorithm highly affected the reproducibility of the measurements, i.e. temporal filtering has to be well aligned to get accurate results. It must also be noted that typically the spectral resolution is limited by the retrieval algorithm, by which sharp features of the spectrum usually obtained from post-compression setups are not reconstructed well. This typically leads also to an underestimated pulse duration and an overestimation on the femtosecond contrast of the pulse. TIPTOE works in the time domain, hence the spectral resolution is poor compared to the other techniques. The full measurement results are shown in Fig S1-S3.
Fig. 4. (a) Spatially integrated spectrum measured with a 2D imaging spectrometer after CMC1. Reconstructed spectrum and spectral phase by (b) d-cycle XR, (c) TIPTOE and (d) Wizzler USP4.
3.2 Multiple thin plates I
In the next stages, the positions of the plates were optimized to achieve smooth spectral broadening while preserving the beam quality with low conical emission and avoiding significant ionization in the plates. To this end, the beam profile and the spectrum were monitored in parallel. The 2D spectra and the spatio-spectral homogeneity demonstrate smooth spectral broadening, although in the vertical direction (Fig. 5(b)) the tilted profile is still visible like in the case of the SF-HCF. The spectrum spanned from 600 to 950 nm at –20 dB level.
Fig. 5. Spatially resolved spectrum on logarithmic scale (a) in the horizontal, and (b) in the vertical cross-section of the beam. The related (c, d), spectrally integrated spatial profile (grey) and spectral homogeneity (orange) after CMC2.
First, we measured the compressed pulses after CMC2 by using our s-shot B device, which has a working range of 3 to 9 fs in terms of Fourier transform limit (FTL) pulse duration, and it is a single-shot apparatus employing an imaging spectrometer inside. The s-shot B has a rectangular input aperture of 10 mm × 10 mm, which must be fully illuminated with a well-collimated beam. As a consequence of the imaging conditions inside the device, the spatial homogeneity of the input beam must be rather high to obtain accurate results from the measurement. It can be seen in Fig. 6 that the spatial intensity distribution was also high quality after CMC2 with a slightly elliptical shape. In order to match the size of the input aperture, we magnified the beam after CMC2 with a telescope built with f = 1000 mm and f = –300 mm silver coated mirrors. The s-shot B was placed at a distance of 25 cm from the telescope output. The intensity of the beam was sufficient to produce second harmonic signals at an appropriate level for the s-shot B measurement, and no attenuation was necessary. We obtained reproducible measurement results, where the retrieved pulse duration was found to be 7.5 fs with a retrieval error of 2.5% rms (more details in Fig. S4). Using the same, magnified beam as in the case of the s-shot B, the intensity was sufficient to obtain proper ion yield without saturation. The input beam diameter and therefore the modulation depth were adjusted with the help of the built-in iris, where the beam had to be well centred to obtain reliable results. Reconstruction gave 7.4 fs pulse duration reproducible over multiple scans, and was very close to the result of the s-shot B measurement. For the Wizzler measurements, the mirrors in the telescope were changed to flat silver ones as the beam size was too large. The energy was attenuated by the built-in iris and half wave plate at the entrance of the device. We measured a pulse duration of 9.3 fs. The pulse shape suggests stronger negative third-order dispersion (TOD) leading to 2 fs longer FWHM than the value we got with the other techniques.
In case of s-shot B the retrieved phase was close to zero, some residual TOD could be also present, yielding negative phase values on the blue side. The difference between the measured and retrieved spectra is attributed to the sampling of the fundamental and SHG beams inside the device (for more details see Fig. S4). The reconstructed spectrum and spectral bandwidth of the TIPTOE (Fig. 7(c)) were also similar to the measured one (Fig. 7(a)) despite the low resolution. In the case of Wizzler USP4, the spectral phase indicates good compression, and shows similarities with the s-shot B data regarding the more negative phase values in the shorter wavelength range. Together with the poor reconstruction of the red side of the spectrum, this could cause the difference in the pulse duration compared to the others. The parameters for the reconstruction algorithm had to be chosen carefully to have reproducible measurements; however, this resulted in low temporal resolution in this case.
Fig. 7. (a) Spatially integrated spectrum measured with a 2D imaging spectrometer after CMC2. Reconstructed spectrum and spectral phase by (b) s-shot B, (c) TIPTOE and (d) Wizzler USP4.
3.3 Multiple thin plates II
The position of the thin plates was optimized in a similar manner as described in the previous section; the beam profile and the spectrum were again monitored simultaneously. The spatially resolved intensity spectra in Fig. 8 show a smooth spectral broadening after CMC3. The calculated spatio-spectral homogeneity indicates a still rather uniform broadening across the beam profile, however compared to the first multiple thin plate stage it is slightly deteriorated. Interestingly, in the vertical direction the 2D profile was no longer tilted (in contrast with CMC1 and CMC2). The spectrum spanned from 560 nm to 950 nm at the –20 dB level.
Fig. 8. Spatially resolved spectrum on logarithmic scale (a) in the horizontal, and (b) in the vertical cross-section of the beam. The related (c, d), spectrally integrated spatial profile (grey) and spectral homogeneity (orange) after CMC3.
It can be seen in Fig. 9 that we could still preserve a close to Gaussian intensity distribution with the main energy content in the central part.
For temporal diagnostics, the beam was again expanded with a telescope consisting of f = 1000 mm and f = –300 mm silver mirrors. The s-shot B was placed at a distance of 35 cm from the telescope output, where we could measure compressed, 4.4 fs short pulses, i.e. 1.7 optical cycles (Fig. 9) with a retrieval error of 2.8% rms (see also Fig. S7). The relative peak intensity of the compressed pulse is around 70%, but considering the complexity of the setup and the overall high compression factor, it can be considered as a well-compressed pulse. We observed that the measurements were extremely sensitive to beam collimation and alignment; especially the angle played an important role in obtaining the proper trace, but with precise alignment we could get reproducible measurement results on a daily basis. In case of TIPTOE, no magnification was necessary at this time. The beam propagated about 50 cm inside the device, which meant a lot of additional positive GDD for such a short pulse, and the effect was clearly observable in the ion yield signal. Therefore, the diagnostic tool was placed closer to the output of CMC3. A pulse duration of 4.1 fs was reproducibly measured. This value is very close to the result of the s-shot B measurement, and similar side lobes are visible next to the main peak, around ±40 fs. With the Wizzler USP4 device we were not able to measure the pulse duration after CMC3, as the measurement results were not reproducible, and the shot-to-shot variation was unacceptably high.
The spectral phase of the s-shot B is basically zero along the whole spectrum, and the retrieved spectrum covers the same bandwidth as the measured one (Fig. 10(b)) which also indicates good compression. TIPTOE shows very similar features regarding the spectrum and spectral phase but on the blue side it retrieved negative phase values contradictory to the d-scan measurements.
Fig. 10. (a) Spatially integrated spectrum measured with a 2D imaging spectrometer after CMC3. Reconstructed spectrum and spectral phase by (b) s-shot B and (c) TIPTOE.
3.4 Comparison and accuracy of TIPTOE measurements
The previous sections presented the intensity spectrum results of the TIPTOE measurements as retrieved data. The retrieval algorithm allows for loading the measured spectrum recorded by an external device (s-shot B and d-cycle XR in our case) to improve the reconstruction of both spectral phase and pulse intensity. Figure 11 shows the reconstructed temporal profiles in comparison to the reconstructed and measured spectrum. Basically, both calculation methods give similar FWHM pulse durations and pulse shapes. However, the amplitude of the side lobes, especially in Fig. 11(a) and (c), are smaller in those cases when the measured spectrum was loaded, and therefore these results are closer to the ones obtained with other devices.
Fig. 11. Comparison of reconstructed pulse intensities at different stages: (a) SF-HCF, (b) multiple thin plates, stage I and (c) stage II, when reconstruction was done without (grey) or with measured (orange) power spectra.
4. Summary
We compared three different pulse duration measurement techniques, and our results show that all devices gave similar results regarding pulse duration and shape, except that in our case the Wizzler was not able to fully recover the blue side of the broadest spectrum. Table 1 summarizes the main specifications of the devices and the retrieved, averaged pulse durations for three measurements.
Table 1. Summary of the devices and averaged pulse durations (for three measurements) with the standard deviation error
The retrieved spectral phase can be misleading and contradictory in some cases, especially when high TOD is present. We could perform reasonable measurements only when the compressed pulses were close to the Fourier limit, typically in cases where the pulse duration is <1.5 x FTL. The TIPTOE retrieval algorithm gives more reliable results in terms of pulse shape, when a measured spectrum is used.
The alignment was optimized by tuning the last mirror in front of the device. During this procedure, the beam position on the entrance iris must be always checked, and corrected if necessary. In case of d-cycle XR, insertion was set to zero. The signal-to-noise ratio of the fundamental spectrum is less crucial than the SHG signal to have a good retrieval with low RMS error. We observed that the device is less sensitive to the spatial intensity distribution than the single shot version of d-scan (e.g. s-shot B), which utilizes imaging onto the detector. For the TIPTOE, it is crucial to have the correct intensity ratio between the inner and outer parts of the beam cross-section and to reach the proper modulation depth of the measured ion yield, hence the diameter and intensity distribution of the beam are key parameters for a proper measurement. We also observed that even a minor angular misalignment from the optical axis could result in a false uncompressed pulse. SRSI was found to be operational only when the spectral phase is close to zero, thus the pulse to be measured is very close to compression. The higher order phase terms make it difficult to obtain reliable and reproducible results. For the same reason, the retrieval algorithm is highly sensitive to the chosen parameters in the temporal filtering of the Fourier transformed signal, which typically leads to low temporal resolution in the end, and underestimation of the pulse complexity.
Our experience shows that the tools are extremely sensitive to beam quality and alignment, especially in the last stage: a well-collimated and very accurately aligned beam with high spatial homogeneity is essential. Here we also confirmed by using an imaging spectrometer, that spatio-spectral homogeneity was ideal for these measurement devices in all pulse duration regimes. It needs to be noted that in case of significant spatio-spectral inhomogeneities, measurement with any of these devices can give rather different results depending on the input coupling conditions, induced spatio-temporal couplings, and dependence on spatial intensity distribution.
Funding
ELI ALPS project Economic Development and Innovation Operational Programme (GINOP-2.3.6-15-2015-00001); supported by the European Union and co-financed by the European Regional Development Fund and by the Integrated Management and reliable oPerations for User-based Laser Scientific Excellence project (IMPULSE 871161).
Disclosures
Authors declare no conflict 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.
Supplemental document
See Supplement 1 for supporting content.
References
1. O. Jahn, V. E. Leshchenko, P. Tzallas, et al., “Towards intense isolated attosecond pulses from relativistic surface high harmonics,” Optica 6(3), 280–287 (2019). [CrossRef]
2. Z. Chang, P. B. Corkum, and S. R. Leone, “Attosecond optics and technology: progress to date and future prospects,” J. Opt. Soc. Am. B 33(6), 1081–1097 (2016). [CrossRef]
3. J. Li, J. Lu, A. Chew, et al., “Attosecond science based on high harmonic generation from gases and solids,” Nat. Commun. 11(1), 2748 (2020). [CrossRef]
4. J. Liu, K. Okamura, Y. Kida, et al., “Clean sub-8-fs pulses at 400 nm generated by a hollow fiber compressor for ultraviolet ultrafast pump-probe spectroscopy,” Opt. Express 18(20), 20645–20650 (2010). [CrossRef]
5. S. Z. Tóth, T. Stanislauskas, I. Balciunas, et al., “SYLOS lasers – the frontier of few-cycle, multi-TW, kHz lasers for ultrafast applications at extreme light infrastructure attosecond light pulse source,” J. Phys. Photonics 2(4), 045003 (2020). [CrossRef]
6. S. Hädrich, E. Shestaev, M. Tschernajew, et al., “Carrier-envelope phase stable few-cycle laser system delivering more than 100 W, 1 mJ, sub-2-cycle pulses,” Opt. Lett. 47(6), 1537–1540 (2022). [CrossRef]
7. S. Mondal, M. Shirozan, M. N. Ahmed, et al., “Surface plasma attosource beamlines at ELI-ALPS,” J. Opt. Soc. Am. B 35(5), A93–A102 (2018). [CrossRef]
8. T. Nagy, M. Kretschmar, M. J. J. Vrakking, et al., “Generation of above terawatt 1.5 cycle visible pulses at 1 kHz by post-compression in a hollow fiber,” Opt. Lett. 45(12), 3313–3316 (2020). [CrossRef]
9. J.-F. Hergott, H. J. B. Marroux, R. Lopez-Martens, et al., “Near-single-cycle pulses generated through post-compression on FAB1 laser at ATTOLAB-ORME facility,” EPJ Web Conf. 255, 11006 (2021). [CrossRef]
10. T. Nagy, H. Steffen, P. Simon, et al., “Generation of three-cycle multi-millijoule laser pulses at 318 W average power,” Optica 6(11), 1423–1424 (2019). [CrossRef]
11. C.-H. Lu, Y.-J. Tsou, H.-Y. Chen, et al., “Generation of intense supercontinuum in condensed media,” Optica 1(6), 400–440 (2014). [CrossRef]
12. M. Stanfield, F. Beier, S. Hakimi, et al., “Millijoule few-cycle pulses from staged compression for strong and high field science,” Opt. Express 29(6), 9123–9136 (2021). [CrossRef]
13. S. Z. Tóth, R. S. Nagymihály, I. Seres, et al., “Single thin-plate compression of multi-TW laser pulses to 3.9 fs,” Opt. Lett. 48(1), 57–60 (2023). [CrossRef]
14. J. Schulte, T. Sartorius, J. Weitenberg, et al., “Nonlinear pulse compression in a multi-pass cell,” Opt. Lett. 41(19), 4511–4514 (2016). [CrossRef]
15. P. Rueda, T. Witting, G. A. Torchia, et al., “8 fs laser pulses from a compact gas-filled multi-pass cell,” Opt. Express 29(17), 27004–27013 (2021). [CrossRef]
16. C. Grebing, M. Müller, J. Buldt, et al., “Kilowatt-average-power compression of millijoule pulses in a gas-filled multi-pass cell,” Opt. Lett. 45(22), 6250–6253 (2020). [CrossRef]
17. T. Nagy, P. Simon, and L. Veisz, “High-energy few-cycle pulses: post-compression techniques,” Adv. Phys: X 6(1), 1845795 (2021).
18. E. Escoto, A.-L. Viotti, S. Alisauskas, et al., “Temporal quality of post-compressed pulses at large compression factors,” J. Opt. Soc. Am. B 39(7), 1694–1702 (2022). [CrossRef]
19. E. Goulielmakis, M. Uiberacker, R. Kienberger, et al., “Direct measurement of light waves,” Science. 305(5688), 1267–1269 (2004). [CrossRef]
20. A. Korobenko, K. Johnston, M. Kubullek, et al., “Femtosecond streaking in air,” Optica 7(10), 1372–1376 (2020). [CrossRef]
21. D. Zimin, M. Weidman, J. Schötz, et al., “Petahertz-scale nonlinear photoconductive sampling in air,” Optica 8(5), 586–590 (2021). [CrossRef]
22. D. J. Kane and R. Trebino, “Characterization of arbitrary femtosecond pulses using frequency-resolved optical gating,” IEEE J. Quantum Electron. 29(2), 571–579 (1993). [CrossRef]
23. C. Iaconus and I. A. Walmsley, “Spectral phase interferometry for direct electric-field reconstruction of ultrashort optical pulses,” Opt. Lett. 23(10), 792–794 (1998). [CrossRef]
24. T. Oksenhendler, S. Coudreau, N. Forget, et al., “Self-referenced spectral interferometry,” Appl. Phys. B 99(1-2), 7–12 (2010). [CrossRef]
25. M. Miranda, C. L. Arnold, T. Fordell, et al., “Characterization of broadband few-cycle laser pulses with d-scan technique,” Opt. Express 20(17), 18732–18743 (2012). [CrossRef]
26. I. Sytcevich, C. Guo, S. Mikaelsson, et al., “Characterizing ultrashort laser pulses with second harmonic dispersion scans,” J. Opt. Soc. Am. B 38(5), 1546–1555 (2021). [CrossRef]
27. W. Cho, S. I. Hwang, C. H. Nam, et al., “Temporal characterization of femtosecond laser pulses using tunnelling ionization in the UV, visible and mid-IR ranges,” Sci. Rep. 9(1), 16067 (2019). [CrossRef]
28. T. Nagy, M. Forster, and P. Simon, “Flexible hollow fiber for pulse compressors,” Appl. Opt. 47(18), 3264–3268 (2008). [CrossRef]
29. A. Suda, M. Hatayama, K. Nagasaka, et al., “Generation of sub-10-fs, 5-mJ-optical pulses using a hollow fiber with a pressure gradient,” Appl. Phys. Lett. 86(11), 111116 (2005). [CrossRef]
30. 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), 20502–20510 (2017). [CrossRef]
