1 Introduction
High-peak-power femtosecond lasers based on chirped pulse amplification (CPA) and optical parametric chirped pulse amplification (OPCPA) technology have developed rapidly in the last few decades[ Reference Wang, Wang, Rockwood, Luther, Hollinger, Curtis, Calvi, Menoni and Rocca 1 – Reference Sung, Lee, Yoo, Yoon, Lee, Yang, Son, Jang, Lee and Nam 5 ]. Laser systems with the peak power of 10 petawatt (PW, 1 PW = 1015 W) have also been reported in recent years[ Reference Li, Gan, Yu, Wang, Liu, Guo, Xu, Xu, Hang, Xu, Wang, Huang, Cao, Yao, Zhang, Chen, Tang, Li, Liu, Li, He, Yin, Liang, Leng, Li and Xu 6 ], which have become powerful tools for the investigation of laser–matter interactions. During amplification, unwanted amplified spontaneous emission (ASE) and pre-pulses can be generated and amplified, which deteriorate the temporal contrast of the high-peak-power femtosecond pulse. When the intensity of ASE or pre-pulses reaches 1013 W/cm2, they can ionize the target material and generate pre-plasma before the main pulse reaches the target, which would seriously affect the laser–matter interactions.
To eliminate negative impact of the undesired pre-pulses and ASE, several temporal pulse cleaning technologies have been proposed and demonstrated, such as cross-polarized wave generation (XPW)[ Reference Ramirez, Papadopoulos, Pellegrina, Georges, Druon, Monot, Ricci, Jullien, Chen, Rousseau and Lopez-Martens 7 , Reference Petrov, Albert, Etchepare and Saltiel 8 ], optical parametric amplification (OPA)[ Reference Liu, Wang, Li, Zhang, Han, Teng and Wei 9 , Reference Hu, Wang, Yang, Bai, Wu, Zhang, Chen, Yang, Qian, Gui, Li, Liu, Lu, Xu and Leng 10 ], self-diffraction (SD)[ Reference Wang, Wang, Wang, Zhang, Song and Wei 11 , Reference Liu, Okamura, Kida and Kobayashi 12 ], nonlinear elliptical polarization rotation (NER)[ Reference Homoelle, Gaeta, Yanovsky and Mourou 13 – Reference Kalashnikov, Risse, Schonnagel and Sandner 15 ], plasma mirrors[ Reference Lévy, Ceccotti, D'Oliveira, Réau, Perdrix, Quéré, Monot, Bougeard, Lagadec and Martin 16 – Reference Rödel, Heyer, Behmke, Kübel, Jäckel, Ziegler, Ehrt, Kaluza and Paulus 18 ], and nonlinear Fourier filtering[ Reference Gilicze, Dajka, Foldes and Szatmari 19 , Reference Szatmari, Dajka, Barna, Gilicze and Foldes 20 ]. Among the temporal contrast improvement technologies mentioned above, XPW based on the BaF2 crystal is widely implemented in ultra-intense femtosecond laser facilities worldwide[ Reference Sung, Lee, Yoo, Yoon, Lee, Yang, Son, Jang, Lee and Nam 5 , Reference Yu, Xu, Liu, Li, Li, Liu, Li, Wu, Yang, Yang, Wang, Lu, Leng, Li and Xu 21 ], as it improves the temporal contrast by several orders of magnitude and shortens the pulse by a factor of approximately 1.73 at the same time. Therefore, XPW is good for obtaining higher peak power after amplification. Nevertheless, the total efficiency of XPW is typically only about 20%, and the focused intensity of the femtosecond pulse on the BaF2 crystal is at the 1012 W/cm2 level, which implies that the damage of the BaF2 crystal is unavoidable after long-term operation. So, it is necessary to shift the position of the BaF2 crystal periodically. As for NER, it also features temporal cleaning and spectral broadening simultaneously. Moreover, the spectral broadening capability of NER exceeds that of XPW. Based on NER, the generation of high temporal contrast pulses with few-cycle pulse duration has been reported. In 2019, Smijesh et al. [ Reference Smijesh, Zhang, Fischer, Muschet, Salh, Tajalli, Morgner and Veisz 22 ] demonstrated the generation of high temporal contrast sub-4 fs pulses by integrating NER into a spectral broadening device of argon filled hollow-core fiber (HCF). In 2020, Khodakovskiy et al. [ Reference Khodakovskiy, Kalashnikov, Pajer, Blumenstein, Simon, Toktamis, Lozano, Mercier, Cheng, Nagy and Lopez-Martens 23 ] demonstrated the generation of high temporal contrast 5 fs pulses by NER in an HCF compressor. However, the HCF suffers from low transmittance and more complex alignment requirements. Another spectral broadening device, called the multi-pass cell (MPC), has been widely investigated in recent years[ Reference Schulte, Sartorius, Weitenberg, Vernaleken and Russbueldt 24 – Reference Song, Wang, Lv, Wang, Teng, Zhu and Wei 33 ]. It consists of two concave mirrors and nonlinear elements between them. The pulse roundtrips inside MPCs can induce dramatic spectral broadening. Compared with HCF, the MPC features high efficiency, excellent energy and power scalability and a low requirement for pointing stability of the input laser. In 2021, Pajer and Kalashnikov[ Reference Pajer and Kalashnikov 34 ] simulated the simultaneous nonlinear spectral broadening and temporal cleaning of femtosecond pulses by NER in an MPC device. In their simulations, 103 temporal contrast enhancement with 50% internal efficiency is realized. Recently, Pfaff et al. [ Reference Pfaff, Forster, Barbiero, Rampp, Klingebiel, Brons, Teisset, Wang, Jung, Jaksic, Woldegeorgis, Saraceno and Metzger 35 ] demonstrated temporal contrast improvement by NER in a gas filled MPC device, and the temporal contrast is enhanced by more than a factor of 50 with an optical efficiency of 56%.
In this paper, the temporal cleaning and compression are experimentally achieved simultaneously by integrating NER into a solid plate-based MPC device. The pedestal of the input pulse is cleaned effectively with a total efficiency higher than 50%, while the pulse duration is shortened from 181 to 36 fs, corresponding to a compression factor of 5. The beam quality after temporal filter is excellent, with M 2 values reaching 1.18 × 1.16.
2 Experimental setup
The layout of the experimental setup is depicted in Figure 1. An ytterbium (Yb)-doped femtosecond amplifier is employed as the driving laser, which delivers 2 mJ pulse energy and 181 fs pulse duration at a 1 kHz repetition rate. The central wavelength and the spectral full width at half maximum (FWHM) of the driving laser are 1036.8 and 9.3 nm, respectively. A small portion of pulse energy is used for our proof of principle experiment. The mode matching between the Yb-doped driving laser and eigenmode of the MPC is realized by utilizing three lenses (L1–L3). A Glan prism (GL1) is used to improve the linear polarization degree. The pulse energy after GL1 is 29 μJ. Then, the polarization of the input laser is converted to elliptical through a quarter wave plate (QWP1). The MPC cavity consists of two concave mirrors with 50.8 mm diameter and 300 mm radius of curvature. The MPC mirrors are low group delay dispersion (GDD) coated in the spectral range of 950–1110 nm. The distance between the MPC mirrors is 525 mm, so the eigenmode diameter of the MPC is 2w 1 = 1.023 mm on the MPC mirrors and 2w 0 = 0.36 mm in the middle of the MPC. A 5 mm thick fused silica plate coated with high transmittance at 1 μm is used as the Kerr medium. As material dispersion of the fused silica plate is not compensated inside the MPC, the pulse duration broadens after multiple round trips. The single-pass nonlinear phase shift will decrease gradually. Therefore, a small number of round trips (13) are proposed, corresponding to laser pulse propagation of 0.13 m in fused silica and 13.5 m in air. It is obvious that the propagation length is much shorter than the respective dispersion lengths of fused silica L D,FS = 0.64 m and air L D,air = 725 m. The accumulated total material dispersion inside the MPC is calculated to be 2630 fs2, which just broadens the input pulse from 181 to 185 fs. As a result, the nonlinear phase shift is barely affected by material dispersion. The pulse coupling in and out of the MPC is performed with a rectangle mirror with 3 mm width. The output beam is collimated with lens L4. Then output pulses are delivered through QWP2 to eliminate the retardation generated by QWP1. The pulse temporally filtered by NER is discriminated by a high extinction ratio (>2 × 105) GL2.
Figure 1 Layout of the experimental setup. HWP, half wave plate; TFP1 and TFP2, thin film polarizers; L1–L4, lenses; GL1 and GL2, Glan prisms; QWP1 and QWP2, quarter wave plates; CM1 and CM2, concave mirrors; FS, fused silica plate.
3 Experimental results
The output power measured before GL2 is 27.5 mW, corresponding to a transmission efficiency of 95% in the MPC. QWP1, whose optical axis is rotated by an angle γ with respect to the polarization of the input pulse, is employed to set the original ellipticity, while QWP2 is rotated with an angle of 90° + γ. The fused silica plate is positioned at 14, 16, 18 and 20 cm away from CM2, respectively, corresponding to the eigenmode diameter at the position of the plate reducing from 574 to 427 μm, while ignoring the Kerr lens effect. Therefore, the peak intensity on the plate is estimated to be at the 1011 W/cm2 level. The total efficiency of the NER filter is investigated as a function of angle γ, and results are described in Figure 2(a). The total efficiency is given by ηt = P 2/P 1, where P 2 denotes the power after GL2 and P 1 denotes the power after GL1. As the angle γ increases from 0° to 44°, the total efficiency first increases and then decreases. When the fused silica plate is 14 and 16 cm away from CM2, the angle γ corresponding to maximum efficiency is 14°. Angle γ for maximum efficiency shifts to 12° for the plate at 18 and 20 cm away from CM2. Figure 2(b) summarizes the optimal total efficiency when the plate is placed at different positions. The highest total efficiency of 53% is obtained when the plate is placed 18 cm away from CM2.
Figure 2 (a) Total efficiency of the NER in the MPC as a function of elliptical angle. (b) Optimum total efficiency when the fused silica plate is placed at different positions.
By placing the plate at different positions, the spectra after GL2 are measured at the maximum total efficiency, as presented in Figure 3(a). As the plate moves away from CM2, stronger spectral broadening occurs because the nonlinear phase shift increases gradually. The spectral bandwidth at the intensity of –20 dB is broadened from 94.1 to 131.6 nm, as summarized in Figure 3(b) (red line), and the corresponding Fourier transform limit (FTL) pulse duration is shortened from 44.2 to 31 fs, as shown in Figure 3(b) (blue line). At the strongest spectral broadening condition, the total efficiency of the system is 52%.
Figure 3 (a) Broadened spectra when the plate is placed at different positions. (b) Spectral bandwidth at the intensity of –20 dB (red line) and the corresponding FTL (blue line).
After being ejected from the MPC, the positive chirp of the pulse accumulated with propagation through the MPC is compensated by utilizing chirped mirrors. The pulse duration is characterized by using a homebuilt second harmonic frequency-resolved optical gating (SHG-FROG) device. A pulse duration of 36 fs is obtained when the compensated GDD is –3200 fs2, corresponding to a compression factor of 5. The measured and retrieved traces are shown in Figures 4(a) and 4(b). The spectral intensity and phase are shown in Figure 4(c). The input pulse, compressed pulse and calculated FTL pulse durations are presented in Figure 4(d).
Figure 4 SHG-FROG characterization of the filtered pulse. (a) Measured and (b) retrieved SHG-FROG traces (0.45% FROG error on a 512 × 512 grid). (c) Measured spectrum (red line) and spectral phase (blue line). (d) Input pulse duration (blue dashed line), retrieved pulse duration (red line) and calculated FTL pulse duration (black line).
The beam quality (M 2) of the temporally cleaned pulse is measured to be 1.18 × 1.16 by employing a commercial M-squared factor meter (BSQ-SP300, Ophir Spiricon), which is shown in Figure 5. Compared with the M 2 of the input laser, the beam quality is almost maintained.
Figure 5 Beam quality after the MPC device.
The energy of the temporally cleaned pulse is no more than 20 μJ, which is insufficient to perform temporal contrast measurement with a commercial third-order cross-correlator device. So, a Yb:KGW-based chirped pulse amplifier was built to boost the filtered pulse energy. The schematic of the amplifier is depicted in Figure 6. The stretcher and compressor share one transmission grating. One Ng-cut Yb:KGW crystal with a size of 3 mm × 3 mm × 5 mm is employed as a gain medium in the regenerative amplifier. The dumped pulse energy of the regenerative cavity is 800 μJ. Limited by the emission spectral bandwidth of the Yb:KGW crystal and gain narrowing, the spectral bandwidth of the amplified pulse is narrowed to 4.8 nm, as shown in Figure 7(a). Assuming a sech2 shaped pulse, the compressed pulse duration is 457 fs, as shown in Figure 7(b).
Figure 6 Schematic of the Yb:KGW chirped pulse amplifier.
Figure 7 (a) Spectrum and (b) pulse duration of the Yb:KGW amplifier.
Figure 8 shows the measured temporal contrast of the amplified pulse (red line), which is 109. As a comparison, the temporal contrast is 107 when the amplifier is seeded by a semiconductor saturable absorption mirror (SESAM) mode-locked fiber oscillator (black line). It can be seen that the fluctuation of the ASE intensity increases dramatically when employing the temporally cleaned pulse as the seed. This is because the distance between the Yb-doped driving laser and the homemade Yb:KGW regenerative amplifier is around 10 m, which deteriorates the energy stability of the amplified pulse.
Figure 8 Temporal contrast of the homemade Yb:KGW amplifier with different seed injections.
4 Conclusion
In conclusion, the temporal contrast of a Yb-doped femtosecond laser is enhanced to nearly 109 by integrating NER into a solid-state MPC device. The pulse duration is compressed from 181 to 36 fs, corresponding to a compression factor of 5. The beam quality after the MPC is excellent, with the M 2 value of 1.18 × 1.16. The total efficiency of the MPC-based temporal contrast enhancement device is higher than 50%. Benefitting from excellent power and energy scalability of the MPC device, we believe that this scheme can be applied to high-energy pulse cleaning system.
Acknowledgments
This work was supported by the National Key R&D Program of China (No. 2017YFE0123700), the National Natural Science Foundation of China (Nos. 61925507 and 62075227), the Shanghai Rising-Star Program (No. 21QA1410200), and the Youth Innovation Promotion Association CAS (No. 2020248).
Open access
