Open Access
Add to BibSonomy
Abstract
In order to enhance the performance of a continuous-wave photocathode electron gun at Peking University, and to achieve electron beams with higher current and brightness, a multifunctional drive laser system named PULSE (Peking University drive Laser System for high-brightness Electron source) has been developed. This innovative system is capable of delivering an average output power of 120 W infrared laser pulse at 81.25 MHz, as well as approximately 13.8 W of green power with reliable stability. The utilization of two stages of photonic crystal fibers plays a crucial role in achieving this output. Additionally, the incorporation of two acousto-optic modulators enables the selection of macro-pulses with varying repetition frequencies and duty cycles, which is essential for effective electron beam diagnosis. Furthermore, the system employs a series of birefringent crystals for temporal pulse shaping, allowing for stacking Gaussian pulses into multiple types of distribution. Overall, the optical schematic and operating performance of PULSE are detailed in this paper.
© 2024 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
The electron gun serves as a critical component in electron accelerators, playing a pivotal role in producing high-quality electron beams. Photocathode electron guns have emerged as the preferred choice for numerous facilities, including free-electron lasers (FELs) and energy-recovery linacs (ERLs), due to the lower emittance and shorter length of the generated bunches [1,2]. The drive laser system could directly control various parameters of electron beams from the photocathode, such as bunch charge, longitudinal distribution, and repetition rate [3]. FEL facilities require higher bunch charge, leading the drive laser system to prioritize obtaining higher pulse energy [4–7]. Meanwhile, ERL facilities aim to achieve electron beams with high current [8,9]. Consequently, ERLs necessitate drive lasers with high repetition frequency and average power.
Efforts to improve the average power of the drive laser have led to the widespread use of photonic crystal fiber (PCF) and rod-type photonic crystal fiber (Rod-PCF) as the gain medium of the main amplifier. Cornell University has developed two fiber laser systems for the Cornell energy recovery linac photoinjector [10,11]. One operates at a 50 MHz repetition rate for beam emittance diagnosis, while the other at a 1.3 GHz repetition rate, generating a high average power output that yields a 75-mA beam current from the injector [9,10]. Numerous other facilities have also designed high average power laser systems with the aim of achieving high average current, including cERL at KEK [8,12], the Accelerator Test Facility (ATF) at BNL [13], and PAPS at IHEP [14]. In addition, the transverse mode and power stability are also critical indicators that require improvement, as they significantly impact the operational status of the photocathode electron gun.
The distribution of the generated electron beam in a photocathode electron gun can be adjusted by modifying the temporal distribution of the drive laser pulse. In order to suppress the growth of emittance caused by the space charge effect, the pulse temporal stacking system designed for photoinjectors typically aims for a flattop distribution [1]. To achieve this, birefringent crystals are commonly employed within the picosecond regime using two different methods. The first method involves a series of birefringent crystals with decreasing thickness, arranged in a Lyot filter structure [15]. This method is widely adopted by laboratories worldwide due to its simplicity and low energy loss [10,12,16–20]. The second method, known as coherent pulse stacking, utilizes crystals of equal length and set up like Solc filter structure [21]. Peking University has successfully demonstrated a pulse shaper capable of producing laser pulses with various temporal distributions and linear polarization [22]. The coherent pulse stacking system exhibits superior shaping performance compared to the pulse shaping method with decreasing crystal thickness, enabling the generation of pulses with arbitrary temporal structures.
Peking University has been developing a photocathode electron gun with high repetition rate, brightness, and current [23–28]. To further enhance its performance, Peking University has developed a new drive laser system, named as PULSE (Peking University drive Laser System for high-brightness Electron source). PULSE utilizes both PCF and Rod-PCF as main amplifiers to achieve high average power, while maintaining a transverse distribution that is very close to the Gaussian mode. Additionally, stable green light is generated through frequency doubling with a noncritical phase-matched LBO crystal. As a unique part, PULSE incorporates a pulse stacking module under two multi-crystal configurations. One aims at a flattop-distributed output with higher efficiency, while the other is designed to enable real-time adjustments to the longitudinal profile of electron bunches. Moreover, PULSE offers the flexibility to switch between continuous wave mode and electron beam diagnosis mode. This paper provides a detailed description of the design and performance of the PULSE system.
2. Design of PULSE
2.1 Design objectives
PULSE primarily employs an all-fiber master oscillator power amplifier (MOPA) structure to achieve high average power. In order to relieve damage to the amplifying medium, a pulse stretching and compression module is designed based on the chirped pulse amplification (CPA) method. The layout of the system, which also incorporates the second harmonic generation (SHG) and pulse stacking modules, is illustrated in Fig. 1.
Fig. 1. Structure of PULSE. (BF: bandpass filter; WDM: wavelength division multiplexing; YDF: Yb-doped fiber; AOM: acousto-optical modulator; DM: dichroic mirror; PCF: photonic crystal fiber; FM: flipping mirror)
The electron gun at Peking University adopts K2CsTe as the photocathode [28]. To attain an electron beam with 5-mA current at 81.25 MHz, the incident laser power at the photocathode needs to exceed 1.2 W, assuming a quantum efficiency (QE) of 1% [29]. Approximately 50% of the power will be lost at the spatial shaping and transportation system. The efficiencies of pulse compression, SHG, and pulse stacking are approximately 70%, 30%, and 20%, respectively. Considering the power loss at these modules and other optical elements, the target output infrared power after the main amplifier is set at 100 W.
A mode-lock femtosecond ytterbium laser from Menlo Systems Inc. is selected as the master oscillator. This laser provides free-space output 1038 nm infrared laser pulses at a repetition rate of 81.25 MHz, with a pulse duration of 300 fs. The output average power is approximately 97 mW. The spectral bandwidth at 1038 nm is 11 nm, as depicted in Fig. 2. In order to enhance frequency doubling and temporal shaping performance, we have employed two band-pass filters to control the spectrum of the seed pulses. The first filter has a 4 nm bandwidth at 1030 nm, while the second filter has a bandwidth of 2 nm at 1030 nm.
We have integrated two fiber pre-amplification modules prior to the main amplifiers, allowing the incident laser to meet the minimum seed power requirement of the main amplifier. These pre-amplifiers utilize 976 nm laser diodes (LD) as pumps and Yb-doped fiber as the gain medium. The incident seed pulses are amplified by coupling the pump energy into the gain fiber using wavelength division multiplexing (WDM). The average power of the seed laser is designed to be amplified to over 200 mW through the pre-amplifiers.
In order to achieve an output power of 100 W in the infrared (IR) range with high stability and transverse fundamental mode, we have selected two PCFs as the main amplifiers. The first main amplifier utilizes a PCF module with a 31 μm mode field diameter (aeroGAIN-BASE-1.3, NKT Photonics). It features a 1.8-meter-long Yb-doped gain medium and can reach a maximum output power of 45 W with highly stable single-mode operation. This module is designed to amplify the seed signal to 5 W, which serves as the input seed for the subsequent amplifier. The PCF amplifier is counter-pumped to minimize nonlinearity [30].
The second main amplifier module is a Rod-PCF module (aeroGAIN-ROD-2.1, NKT Photonics) with 804 mm of Yb-doped gain fiber. With a 65 μm mode field diameter and high pump absorption, the Rod-PCF is an ideal gain medium for ultrafast high-power amplifiers. The Rod-PCF amplifier is counter-pumped by a fiber-coupled diode laser with a maximum output power of 250 W and a central wavelength locked at 976 nm. The application of the wavelength-locked diode laser can significantly improve the pump absorption efficiency of the Rod-PCF amplifier module, so as to improve the output stability [31]. Any unabsorbed pump energy is directed into a beam dump. To ensure higher stability for the output high-power infrared laser, both main amplifiers and their pumps are connected to a water cooling system.
In the context of the CPA structure, the infrared pulses are coupled into a 300-meter-long fiber for pulse stretching between two filters. The pulse compression module is designed to optimize pulse temporal shaping. This module comprises a series of YVO4 crystals, each 3.3 mm thick, as the pulse width of the green light needs to be maintained at approximately 4 ps. Consequently, the compression module is built to produce infrared pulses with a full width at half maximum (FWHM) of 5.6 ps, which is approximately 1.4 times larger than the green pulse due to the frequency doubling process.
To introduce negative dispersion, a pair of transmission gratings are configured as Treacy structure. The transmission gratings have 1600 lines/mm and designed to operate near 1030 nm central wavelength at 55.5° angle of incidence (AOI). In experiments, the AOI was adjusted slightly for maximized transmittance. A retroflecting mirror is employed to allow the beam to travel backward through the gratings, effectively resolving the issue of spatial incoherence while doubling the amount of negative dispersion.
Considering the spectral response of the K2CsSb photocathode [29], it is necessary to double the frequency of the infrared pulse to produce green light. To achieve this, we employ an LBO crystal for frequency doubling, also known as SHG, due to its high efficiency and damage threshold. A 3 × 3 × 10 mm3 LBO crystal has been selected for this purpose.
The compressed infrared light spot is focused at the center of the LBO crystal using a convex lens with a focal length of 150 mm. This focusing increases the power density, thereby improving the SHG efficiency. Subsequently, another lens with a focal length of 50 mm is positioned after the LBO crystal to adjust the output beam to be nearly parallel.
2.2 Pulse picking module with two-stage AOMs
During the experiment with the photocathode electron gun, it is essential to switch the operation mode between continuous-wave mode and diagnosis mode. Additionally, the repetition frequency and duty cycle of the macro-pulses need to be adjusted based on experimental requirements.
To modulate laser pulses, PULSE utilizes a combination of two acousto-optic modulators (AOMs) to pick the macro-pulse train. Initially, the frequency of pulses from the main oscillator is reduced by AOM-I, which is fiber-coupled on both the input and output sides. For pulses at an 81.25-MHz repetition rate, the interval between two adjacent micro-pulses is 12.3 ns. The rising and falling edges of AOM-I are both below 10 ns, ensuring the complete picking of each micro-pulse, thereby maintaining the same shape and energy for each pulse.
Subsequently, the amplified macro-pulses are picked by the second AOM (AOM-II), which possesses a much higher energy tolerance and is utilized to generate macro-pulses with much lower repetition rate. AOM-II features longer rising and falling edges, approximately 150 ns. Consequently, AOM-I is designed to pick macro-pulses with the appropriate repetition rate and duty cycle, so that the interval of macro-pulses is greater than 150 ns.
2.3 Pulse stacking module with YVO4 crystals
PULSE employs two different setups to stack green pulses separately: one uses a series of birefringent crystals of decreasing thickness and the other uses crystals of equal thickness. For the birefringent crystals, YVO4 crystals are selected to divide the 515 nm green pulse into multiple pulses with approximately 1 ps time delay per millimeter length of the crystal.
A series of birefringent crystals with decreasing thickness is commonly utilized to generate a temporally flattop-distributed pulse. In the case of PULSE, three YVO4 crystals are employed with thicknesses of 16 mm, 8 mm, and 4 mm, with the optical axis oriented at 45 degrees and 0 degrees alternatively to the polarization direction of the incident pulse. The horizontally polarized green pulse passes through these crystals sequentially and is divided into eight sub-pulses with an approximate 4 ps delay between two adjacent sub-pulses. Consequently, the 4 ps incident pulse is theoretically stacked into a 32 ps flattop-distributed pulse.
In order to achieve arbitrary pulse shaping, a coherent pulse stacking module has been specifically designed for PULSE. This module utilizes eight 3.3-mm-long YVO4 crystals, along with two Glan prisms positioned before and after the crystals. The linearly polarized light pulse from the first Glan prism is divided into 28 pulses after passing through the eight birefringent crystals with nine different delays. Subsequently, the second Glan prism projects their polarizations in a specific direction for coherent addition. Each YVO4 crystal is housed in a crystal oven, allowing for precise control of its temperature to finely adjust the phase difference between ordinary ray (o-ray) and extraordinary ray (e-ray). Additionally, by rotating the crystals, the angle between the optical axis of the birefringent crystal and the direction of incident light polarization can be altered to adjust the relative magnitude of o-ray and e-ray.
Figure 3 shows a cross-correlation system which has been constructed to measure the temporal structure of the output pulses. Approximately 10% of the infrared light before frequency doubling is extracted by a beam splitter and compressed to about 1 ps to serve as the detection pulse. An optical delay line is set here to adjust the relative phase between IR light and green light. After the adjustment, the two pulses are focused on a BBO crystal together. By scanning the relative phase between the IR and green pulse with the motorized stage, the changing intensity of UV light generated by the sum frequency generation (SFG) effect in BBO reveals the temporal profile of the stacked green pulse.
3. Experiment results
3.1 Output of the main amplifier and SHG module
According to the schematic mentioned above, the main design objectives are summarized in Table 1.
Table 1. The main design objectives
The entire laser system has been constructed and tested incrementally. To suppress nonlinear effects, the incident seed power at the fiber collimator before the pulse stretching structure is reduced to less than 1 mW. Preamp I amplifies the seed laser to 113.4 mW under a 500 mA LD current. Subsequently, after passing through the fiber bandpass filter, isolator, and AOM-I, a significant portion of the power is dissipated. The incident laser is once again amplified by preamp II. Adjusting LD current to 700 mA, a 142-mW average power is measured at the output port of PCF, enough for serving as the seed signal for the main amplifier.
For the main amplification modules, 5-axis stages mounted with optical lenses are utilized to enhance the coupling of pump light into the fiber. The output infrared power demonstrates a linear increase with the rising pump power, as depicted in Fig. 4(a). The transverse distributions of the output laser, also shown in the figure, were measured by a beam profiler at the pump current of 0.6A, 1.2A, and 1.88A, revealing a relatively high quality with a Gaussian distribution. Considering actual operating conditions, the maximum output power has been tested up to 6.7 W. Notably, with a pump power of 12.9 W, the output infrared power is stably amplified to 5.27 W, as illustrated in Fig. 4(b). The relative standard deviation (RSD) is lower than 0.5%.
Fig. 4. Output power of PCF vs pump LD current and beam transverse distribution at the pump current of 0.6A, 1.2A, and 1.88A (a); stability of output infrared power (b).
In the second main amplifier, the infrared pulses are further amplified. As depicted in Fig. 5(a), the output infrared power also demonstrates a linear increase with the pump current. The beam transverse distributions, measured at the pump current of 7A, 12A, 22A, and 32A, are displayed in Fig. 5(a). We can see as the pump power increases, the output beam profile deviates from the Gaussian distribution. This could be a superposition of higher-order modes [32]. The maximum output power reaches approximately 120 W when the current is set at 38 A. The spectrum distribution at 100W output power is illustrated in Fig. 5(b), exhibiting a FWHM of 2.26 nm at 1029.5 nm. By setting the pump LD current of the Rod-PCF to 30 A, the infrared light is stably amplified to 99.3 W, with an RSD of less than 0.4% as shown in Fig. 5(c).
Fig. 5. Output power of Rod-PCF vs pump LD current and beam transverse distribution at the pump current of 7A, 12A, 22A, and 32A (a); spectrum of infrared beam (b); output power stability at 99.3 W (c).
Following this, the infrared pulses are compressed using a grating pair. The autocorrelation signals of the input and compressed infrared pulses are depicted in Fig. 6. Adjusting the distance between the two gratings to 60 mm, the pulse is compressed to 5.28 ps from 22.21 ps. The transmission of the gratings is over 91%, indicating a total efficiency of approximately 70%.
By setting the crystal temperature to 90 ℃, the SHG efficiency rises to above 35% with an increase in incident infrared light power exceeding 40 W, as illustrated in Fig. 7(a). The output green light power reaches 13.8 W with a SHG efficiency of 34.8% when the incident IR power is 39.7 W. Subsequent to power monitoring, as depicted in Fig. 7(b), the RSD of the green power is approximately 0.7%, indicating exceptional stability. Furthermore, Fig. 7(c) presents the autocorrelation signal of the output green light, with a FWHM of 5.55 ps, corresponding to a Gaussian-fitted pulse width of 3.92 ps. The transverse distribution is nearly Gaussian-like, as shown in Fig. 7(d).
Fig. 7. Green light power and SHG efficiency vs infrared light power (a); power stability of green light power (b); autocorrelation signal of green pulse (c); transverse beam distribution (d).
3.2 Performance of pulse stacking module
Figure 8 displays the cross-correlation measurement of the green pulse subsequent to pulse stacking module with crystals of decreasing thickness. The actual pulse width is around 34 ps, and the ripples in the profile may be attributed to the interference of the pulses resulting from the random phase difference between two adjacent sub-pulses.
Fig. 8. Cross-correlation measurement of the flattop green light after pulse shaping module with crystals of decreasing thickness.
The coherent pulse stacking module features each crystal mounted on a vertically-placed motorized rotation stage. These stages can be controlled online, enabling real-time control of the pulse profile. Additionally, a plate has been designed to fix the stages facing the same direction, as depicted in Fig. 9, enhancing collimation. This design consolidates the coherent pulse stacking module into an integrated unit, facilitating effortless relocation to other positions.
Figure 10 depicts the output pulses with temporal flattop distribution. In addition, the coherent pulse stacking module is capable of generating pulses with numerous other temporal distributions. As exemplified in Fig. 11, three triangle pulses with different rise and fall times are generated by the coherent pulse shaping module.
3.3 Macro-pulse diagnosis mode
To test the pulse picking performance, a photodetector connected to an oscilloscope is used to measure the waveform of output pulse trains. Waveform of the original 81.25 MHz output pulses is depicted in Fig. 12(a). Square wave pulses are used as the modulating signal to control the two AOMs. AOM-I modulates pulses under square wave pulses with 0.4 μs width at 1 MHz frequency, while AOM-II modulates with 1.6 μs width at 250 kHz pulses. The blue lines in Fig. 12(b) and Fig. 12(c) display the pulse sequence after AOM-I and AOM-II respectively. And the red lines are intentionally drawn with different rise and fall time to show the difference between two AOMs.
Fig. 12. Pulses sequence of main oscillator (a); macro pulses picked by AOM-I (b); macro pulses picked by AOM-II (c).
During the operation of the photocathode electron gun, the repetition frequency and the number of micro-pulses in each macro-pulse can be adjusted to meet the experimental requirements.
4. Conclusion
A stable multifunctional drive laser system, PULSE, has recently been constructed and is currently operational. PULSE amplifies the infrared pulse from the main oscillator through PCF and Rod-PCF, achieving an average output power of 120 W at an 81.25 MHz repetition rate. Following frequency doubling, stable green light with a power output of 13.8 W and a 3.98-ps pulse width is generated. The system's output mode can be adjusted to CW mode and macro-pulse diagnosis mode using a two-stage AOM, offering convenience for DC-SRF-II experiments. Moreover, the green light can be shaped into flattop or other temporal distributions through pulse stacking modules before being transported to the photocathode. The steady operation of PULSE could provide an excellent research environment for electron beam longitudinal manipulation.
Funding
State Key Laboratory of Nuclear Physics and Technology, Peking University (NPT2020KFY15, NPT2022ZZ01); National Natural Science Foundation of China (12175251); National Key Research and Development Program of China (2017YFA0701001).
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. F. Stephan and M. Krasilnikov, “High Brightness Photo Injectors for Brilliant Light Sources,” in Synchrotron Light Sources and Free-Electron Lasers (2014), pp. 1–38.
2. R. Xiang, A. Arnold, and J. W. Lewellen, “Superconducting radio frequency photoinjectors for CW-XFEL,” Front. Phys. 11, 1166179 (2023). [CrossRef]
3. F. Zhou, A. Brachmann, P. Emma, et al., “Impact of the spatial laser distribution on photocathode gun operation,” Phys. Rev. Spec. Top.--Accel. Beams 15(9), 090701 (2012). [CrossRef]
4. S. Schreiber, C. Grün, K. Klose, et al., “FLASH Photoinjector Laser Systems,” in FEL2019 (JACoW Publishing, 2019), WEP048.
5. L. Winkelmann, A. Choudhuri, U. Grosse-Wortmann, et al., “The European XFEL Photocathode Laser,” in FEL2019 (JACoW Publishing, 2019), WEP046.
6. S. Gilevich, S. Alverson, S. Carbajo, et al., “The LCLS-II Photo-Injector Drive Laser System,” in Conference on Lasers and Electro-Optics (Optica Publishing Group, 2020), SW3E.3.
7. C. J. Milne, T. Schietinger, M. Aiba, et al., “SwissFEL: The Swiss X-ray Free Electron Laser,” Applied Sciences 7(7), 720 (2017). [CrossRef]
8. M. Akemoto, D. Arakawa, S. Asaoka, et al., “Construction and commissioning of the compact energy-recovery linac at KEK,” Nucl. Instrum. Methods Phys. Res., Sect. A 877, 197–219 (2018). [CrossRef]
9. B. Dunham, J. Barley, A. Bartnik, et al., “Record high-average current from a high-brightness photoinjector,” Appl. Phys. Lett. 102(3), 395 (2013). [CrossRef]
10. Z. Zhao, A. Bartnik, F. W. Wise, et al., “High-power fiber lasers for photocathode electron injectors,” Phys. Rev. ST Accel. Beams 17, 053501 (2014). [CrossRef]
11. Z. Zhao, B. M. Dunham, and F. W. Wise, “Generation of 167 W infrared and 124 W green power from a 1.3-GHz, 1-ps rod fiber amplifier,” Opt. Express 22(21), 25065–25070 (2014). [CrossRef]
12. Y. Honda, “DEVELOPMENT OF A PHOTO-INJECTOR LASER SYSTEM FOR KEK ERL TEST ACCELERATOR,” in Proceedings of IPAC2012 (2012).
13. Z. Zhao, B. Sheehy, and M. Minty, “Generation of 180 W average green power from a frequency-doubled picosecond rod fiber amplifier,” Opt. Express 25(7), 8138–8143 (2017). [CrossRef]
14. H. Xu, J. Xu, X. Li, et al., “High power drive laser system for photocathode at IHEP,” Opt. Express 29(18), 29550–29556 (2021). [CrossRef]
15. J. W. Evans, “The Birefringent Filter,” J. Opt. Soc. Am. 39(3), 229–242 (1949). [CrossRef]
16. D. Wang, “UV PULSE SHAPING WITH α-BBO CRYSTALS FOR THE PHOTOCATHODE RF GUN,” in Proceedings of IPAC2016 (2016).
17. A. K. Sharma, T. Tsang, and T. Rao, “Theoretical and experimental study of passive spatiotemporal shaping of picosecond laser pulses,” Phys. Rev. Spec. Top.--Accel. Beams 12, 033501 (2009). [CrossRef]
18. B. Zhang, X. Li, Q. Liu, et al., “High repetition-rate photoinjector laser system for S3FEL,” Front. Phys. 11, 862 (2023). [CrossRef]
19. T. Schietinger, M. Pedrozzi, M. Aiba, et al., “Commissioning experience and beam physics measurements at the SwissFEL Injector Test Facility,” Phys. Rev. Accel. Beams 19, 100702 (2016). [CrossRef]
20. L. Feng, T. Wang, H. Jia, et al., “Peking University's DC-SRF-II photoinjector drive laser system,” High Power Laser and Particle Beams 34(10), 104016 (2022). [CrossRef]
21. I. Will and G. Klemz, “Generation of flat-top picosecond pulses by coherent pulse stacking in a multicrystal birefringent filter,” Opt. Express 16(19), 14922–14937 (2008). [CrossRef]
22. F. Liu, S. Huang, S. Si, et al., “Generation of picosecond pulses with variable temporal profiles and linear polarization by coherent pulse stacking in a birefringent crystal shaper,” Opt. Express 27(2), 1467–1478 (2019). [CrossRef]
23. K. Zhao, J. K. Hao, Y. Hu, et al., “Research on DC-RF superconducting photocathode injector for high average power FELs,” Nucl. Instrum. Methods Phys. Res., Sect. A 475(1-3), 564–568 (2001). [CrossRef]
24. J. Hao, X. Lu, Y. Ding, et al., “Primary beam-loading tests on DC–SC photoinjector at Peking University,” Nucl. Instrum. Methods Phys. Res., Sect. A 557(1), 138–141 (2006). [CrossRef]
25. S. W. Quan, J. K. Hao, L. Lin, et al., “Stable operation of the DC-SRF photoinjector,” Nucl. Instrum. Methods Phys. Res., Sect. A 798, 117–120 (2015). [CrossRef]
26. Y. Q. Liu, “ENGINEERING DESIGN OF LOW-EMITTANCE DC-SRF PHOTOCATHODE INJECTOR,” in FEL 2019 (2019).
27. T. Tan, H. Jia, S. Zhao, et al., “Sub-micron normalized emittance measurement for a MeV continuous-wave electron gun,” Nucl. Instrum. Methods Phys. Res., Sect. A 1045, 167552 (2023). [CrossRef]
28. S. Huang, K. Liu, K. Zhao, et al., “DC-SRF photocathode gun,” Chin. Sci. Bull. 68(9), 1036–1046 (2023). [CrossRef]
29. Y. Zhao, D. Ouyang, H. Xie, et al., “QE evolution of bialkali photocathode at cryogenic temperature,” Front. Phys. 11, 1157729 (2023). [CrossRef]
30. R. Li, H. Wu, H. Xiao, et al., “More than 5 kW counter tandem pumped fiber amplifier with near single-mode beam quality,” Opt. Laser Technol. 153(17), 108204 (2022). [CrossRef]
31. B. Köhler, T. Brand, M. Haag, et al., “Wavelength stabilized high-power diode laser modules,” Proc. SPIE 7198, 719810 (2009). [CrossRef]
32. J. Limpert, F. Stutzki, F. Jansen, et al., “Yb-doped large-pitch fibres: effective single-mode operation based on higher-order mode delocalisation,” Light: Sci. Appl. 1(4), e8 (2012). [CrossRef]
