Open Access
Add to BibSonomyAbstract
We report on a high-power third-order dispersion managed amplification system that delivers 33-fs pulses of 93.5 W at a repetition rate of 55 MHz. A pair of grisms are used as the pre-chirper for optimizing the third order dispersion (TOD) to group velocity dispersion (GVD) ratio. Detail experiments show that the use of a grsim pre-chirper significantly enhances the quality of the compressed pulses. We demonstrate that the third order dispersion of both the amplifier and the compressor can be compensated for by the grisms. Furthermore, the nonlinear phase shift introduced by spectral asymmetry during amplification can be restrained. This type of scheme, applied in our experiment, can be used for further development of a high power laser with ultrashort pulse and wide spectrum.
© 2016 Optical Society of America
1. Introduction
High power, ultrafast fiber laser sources are required in scientific and industrial applications, such as frequency comb [1], nonlinear optics exploration [2], and micro- and nano-machining of materials [3], where their high peak powers and short temporal durations are exploited. In some applications, such as generation of filamentary plasma gratings [4], high harmonic generation [5], and resolution enhancement of coherent anti-Stokes Raman scattering [6], the precise temporal chirp and duration of the pulses significantly affect the behavior of the laser-matter interaction. Consequently, compact fiber laser sources with high powers and ultrashort pulses are demanded for leveraging ultrafast laser technology. A number of fiber amplifier architectures such as fiber chirped pulse amplification [7], cubicon amplification [8], and similariton amplification [9], indeed allow for significant power scaling of the fiber laser to tens Watts with sub 200-fs pulses.
Further refining the performance of the Yb-doped fiber laser to shorter pulses at a high output power remains a challenging task. For chirped pulse amplification, the residual third-order dispersion (TOD) typically degrades the recompressed pulse quality, which is hard to eliminate by managing dispersion match between stretcher and compressor since the gain medium produce unavoidable TOD and the chirped pulses experience gain-mediated spectral distortion during amplification [10]. Even though a special fiber with negative TOD was successfully used as the temporal stretcher to match the TOD and group-velocity dispersion (GVD) of the overall laser system, the shortest pulses obtained from the fiber chirped pulse amplifier were limited to 100 fs by gain narrowing [11]. In contrast with fiber chirped pulse amplification, employing self-similar amplification enabled the Yb-doped fiber laser source delivering 48-fs pulses with 18-W average power [12]. Unfortunately, further power scaling of the fiber amplifier with direct amplification schemes is limited by additional restrictions, such as finite gain bandwidth, TOD and higher order dispersion, and simulated Raman scattering [13]. Precise control of nonlinear spectral phase has been used to compensate for TOD, resulting in 63-fs pulses with 4.1 MW peak power [14]. However sufficiently long gain fiber with low gain is required to support self-similar evolution, which induces stimulated Raman scattering. Recently, we demonstrated that pre-chirping management played a key role on the self-similar amplification in a short gain fiber. Optimizing pre-chirping GVD leads to 38-fs pulses with 80-W average power [15]. Lately, experiments demonstrated that pre-chirping management was an ideal technology for the generation of 60-fs pulses with 22-MW peak power [16]. However, the attainable shortest pulses were limited by TOD originated from asymmetric spectral phase [17]. This nonlinear phase consequently distorts the temporal chirp during the amplification and eventually, degrades the temporal shape of the recompressed pulses [18].
In this letter, we demonstrate an approach to suppress detrimental effects from TOD in the high-power self-similar fiber amplifier by using a grism-based pre-chirper to optimize the pulse chirp prior to the self-similar amplification. Our experimental results confirmed the existence of an optimum TOD to GVD ratio (TGR) that could lead to high-quality compressed pulses. As a result, the generation of 33-fs pulses with 93.5-W average power and 51-MW peak power in an Yb-doped fiber amplifier is achieved. To our knowledge, these are the shortest pulse duration and highest peak power ever reported for an Yb-doped fiber amplifier based self-similar amplification. This TOD managed self-similar amplification is a competitive scheme to chirped pulse amplification for further development of femtosecond fiber laser source in scientific and industrial applications.
2. Experimental setup
A schematic diagram of the laser system is shown in Fig. 1. The master oscillator is an Yb-doped fiber ring mode-locked laser consisting of both fiber and free space elements. This oscillator emits up to 20 mW average power at a repetition rate of 55 MHz and a center wavelength of 1040 nm. Net normal dispersion of 0.001 ps2 supports a train of stretched pulses with a spectral bandwidth of 45 nm. Additional chirp is introduced by a section of single-mode fiber (β2≈23 ps2/km, β3≈0.044 ps3/km). Linear amplification occurs in a two-stage pre-amplifier, where the first stage consists of a single-mode Yb-doped fiber (β2≈23 ps2/km, β3≈0.044 ps3/km) pumped by a pigtailed single-mode laser diode. The output beam is then coupled into the second pre-amplifier which includes a single-mode Yb-doped double-cladding photonics crystal fiber (Yb-PCF). This stage is reverse-pumped by a multimode laser diode centered at 975 nm. Both second pre-amplifier and main amplifier are consisted of same gain fiber. The grism-based compressor is composed of separated grating and prism pairs. As shown in Fig. 1, two anti-parallel isosceles SF10 prisms are placed between two parallel transmission Bragg gratings (1000 l/mm). The prisms with an apex angle of 59°, are anti-reflection coated for S polarization. The choice of the prisms is based on group delay calculations of the grism compressor with different prism materials and apex angles [19, 20]. This combination provides a TGR ranging from 1.9 fs to 2.7 fs, which covers the TGR ratio of the single-mode fiber used in the fiber pre-amplifier and PCF-based main amplifier. The TOD managed self-similar amplification is obtained in a 1.8-m-long single-mode Yb-doped PCF (DC-200/40-PZ-Yb, NKT, β2≈23 ps2/km, β3≈0.06 ps3/km), with a doped core diameter of 40 μm (NA = 0.03) and a pump cladding of 200 μm (NA = 0.6). Due to the high pump absorption (10 dB/m at 976 nm), a relatively short active fiber is needed in order to obtain high optical-to-optical conversion efficiency and avoid stimulated Raman scattering. To suppress the environmental disturbance and parasitic lasing, both ends of the fiber are sealed and cleaved at 8 degrees. The maximum available seed power of 4.8 W is coupled into the fiber amplifier. A fiber-coupled multimode laser diode with central wavelength of 975 nm is employed as pump source. We pump the Yb-PCF from the back through a dichroic mirror with maximum pump power up to 202 W.
Fig. 1 Setup of high-power Yb-doped fiber amplifier system. ISO, isolator; LD, laser diode; WDM, wavelength division multiplexer; YDF, Yb-doped fiber; PCF, photonics crystal fiber; MM LD, multi-mode laser diode; DM, dichroitic, mirror.
3. Experimental results
To characterize the pulse properties after the grisms, a commercial autocorrelator (PulseCheck, APE) and an optical spectrum analyzer (AQ6370, Yokogawa) were used. The measured spectrum of the compressed pulses delivered from the two-stage pre-amplifier is depicted in Fig. 2(a). The spectrum of the laser is centered at a wavelength of 1038 nm and covers a full width at half-maximum bandwidth of 10.4 nm. Figure 2(b) shows the second-harmonic autocorrelation traces of the measured pulse (blue solid) and the corresponding transform-limited pulse (red dash). The measured autocorrelation trace of the compressed pulse has a temporal duration of 158 fs, which is slightly longer than the Fourier transform limit of 151 fs. The good overlap between the two autocorrelation traces indicates that the compressed pulse has 98% pulse energy in the main peak. The temporal profile confirms the high quality of grism-compressed pulse after pre-amplification, which corresponds to time-bandwidth product (TBP) of 0.46 and the Strehl ratio of 0.93 (defined as the ratio of the peak power of the actual pulse to the peak power of its transform-limit). The comparison of two traces shows that residual higher order contributions limit the compressed pulse quality only at the temporal wings. The minimal temporal pedestal evidences successful compensation of TOD with the grisms. Both the spectral and temporal profiles have near-Gaussian shapes with no modulation, indicating absence of observable nonlinear effects. Moreover, the grism-based pre-chirper could further optimize the finally compressed pulses after the main amplification with respect to adjusting the TGR.
Fig. 2 The measured (a) spectral intensity and (b) the temporal intensity (red solid) of the compressed pulse from pre-amplifier, TBP: time-bandwidth product. (c) Measured output power of the self-similar amplifier. The maximum output power of 124 W corresponds to optical to optical conversion efficiency of 61%. (d) Recompressed pulse duration and spectral bandwidth as a function of the compressed output power.
Figure 2(c) shows the average output power of the self-similar amplifier as a function of the incident pump power. The maximum average output power is 124 W at pump power of 202 W, corresponding to an amplification factor of 14 dB. At the maximum output power, the beam profile are measured by a CCD (SP300, Spiricon), as shown in Fig. 2(c). The highest optical-to-optical efficiency of 61% coincides with the maximum output power. No gain saturation is observed at the maximum pump power, indicating that even more power could be achieved from this system with a higher pump power. The double-pass compression stage consists of two 1000 l/mm transmission gratings. The total transmission through the four-pass compressor is measured to be 75.4%, resulting in 93.5-W average power after the compression. Figure 2(d) illustrates the compressed pulse duration and spectral bandwidth versus different compressed output power. The input pulse (blue curve) is shortened from ~66 fs to 33 fs, and the spectral bandwidth (red curve) is increased from 29 nm to 86 nm with the increase of the compressed output power. As the output power increases beyond 60 W, the pulse shortening slows down in the presence of finite gain bandwidth.
In self-similar amplification, the TOD affects the ideal self-similar evolution and finally changes the spectral and temporal shape of the pulses [21]. In Fig. 3, we show the simulation results of the impact of different TODs on the spectral shape evolution along the fiber with a fixed pre-chirping GVD of −1 × 10−4 fs2 and varied TOD of −3 × 10−4, 0, and −3 × 10−4 fs3, leading to TGR of 3, 0, and −3fs, respectively. The parameters of the simulations were taken as the experiments described below. The pulses with negative chirp experience initial spectral and temporal narrowing before the linear broadening. In the stretched pulse regime, the spectral peak shifts towards the spectral edges of the pulse, resulting in an asymmetric spectral shape. And the direction of the shift depends on the sign of the TOD. The temporal shape of the pulses experience similar propagation process. For a broad spectrum, the accumulated nonlinear spectral phase is approximately proportional to the spectral intensity [22]. Since the spectral shape depends on the TOD, the pre-chirping TOD could be used to modify the spectral profile, and this leads to the compensation of the nonlinear spectral phase.
Fig. 3 Results from numerical simulations showing temporal and spectral evolution along the fiber amplifier with different TGRs: (a) and (d) 3 fs, (b) and (e) 0 fs, (c) and (f) −3 fs.
In order to quantitatively investigate the optimal pre-chirping TGR ratio for compensating the TOD originated from nonlinear amplification, we fix the compressed pulse at the direct output power of 90 W, which corresponds to a compressed power of 68 W, and tune the TGR of the compressor. According to [23], the optimal value of the TGR for an ideal self-similar amplifier is depended on the parameter of the gain fiber. For a fixed input power of 4.8 W and fiber parameters (i.e., L = 1.8 m, β2≈23 ps2/km, β3≈0.06 ps3/km) involved in our self-similar amplifier, this calculated ratio equals to 2.2 fs at the direct output power of 90 W. With the grisms used in the experiment, the pre-chirping TGR ratio could be tuned from −0.3 fs to 0.5 fs by changing the distance between the prism apices. The TGR introduced by the self-similar amplification has the same sign to the grism pair pre-chirper. This means that the TOD of both the self-similar amplification and the compressor could be compensated by the grism. Furthermore, the temporal and spectral asymmetry introduced by the TOD during self-similar amplification could be restrained.
To verify the above assumption, the compressed pulses with different pre-chirping TGR ratios were measured with second-harmonic autocorrelator. The resulting temporal intensity profiles are shown in Figs. 4(a)-4(d). The measured pulses are compared to the calculated traces using commercial split-step Fourier nonlinear propagation software. From Fig. 4(a), one can observe that tuning pre-chirped ratio to a value around −0.3 fs, the pulses are not efficiently compressed with a pulse duration of 53 fs. Moreover the pulses exhibit obvious satellite structure with only 68% pulse energy contained in the main pulse peak. It indicates that the residual nonlinear phase is not fully compensated as a result of insufficient pre-chirping TGR ratio. In Fig. 4(b), we increase the ratio to 0 fs, and the measured temporal envelope of the compressed pulses shows pulse duration of 42 fs, where 83% of the total energy is contained in the main peak. The residual pedestal denotes that the nonlinear chirp is not totally compensated by the pre-chirper. As shown in Fig. 4(c), when the pre-chirping ratio is set to 0.2 fs, the shortest pulses with minimal pedestals are obtained. Almost all the TOD has been compensated with a pulse duration of 39 fs, leading to 94% of pulse energy in the main pulse peak. Moreover, the measured temporal envelope almost coincides with the calculated traces of transform-limited pulse, which indicates the cubic phase only exhibits in the negligible satellite structure. The temporal envelope and phase profiles verify a thorough compensation of TOD. However, if we further increase the pre-chirping TGR ratio to 0.5 fs, the satellite structure increases. As shown in Fig. 4(d), the measured temporal envelope indicates that additional TOD limits the pulse duration of 45 fs, where only 67% of the total energy is contained in the main peak.
Fig. 4 The measure temporal envelopes (blue solid), and calculated traces of transform-limited pulses (green dash) at 68-W output power varies with different pre-chirping ratios: (a) −0.3 fs, (b) 0 fs, (c) 0.2 fs, and (d) 0.5 fs. TGR: TOD to GVD ratio.
The evolution of temporal envelope with varied pre-chirping ratio presents the convincing evidence for the above assumption. Comparing the pulse durations at different pre-chirping TGR ratios indicates that an optimum ratio exists at 0.2 fs. In this case, both the TOD introduced by amplification and asymmetric spectral phase could be compensated. The optimization of the pre-chirping TGR ratio could significantly enhance the quality of the compressed pulses.
Because the self-similar evolution is dependent on the gain profile, the optimum pre-chirping TGR ratio varies with different output power. Particularly, both finite gain bandwidth and gain spectral shape force the asymmetric spectral evolution, which limits the generation of symmetric phase. Figure 5 shows the optimal temporal envelope and the spectra of the compressed pulses measured at three representative cases. For 40 W of output power, the spectrum broadens with a flat top and smooth edge, showing a feature of self-similar evolution. Almost all the nonlinear phase is compensated with pre-chirping TGR ratio of 0.05 fs. The pulse duration is reduced to 45 fs at output power of 60 W, where the pre-chirping TGR ratio is around 0.1 fs. The shortest pulse duration is obtained at 93.5 W of output power, where 92% of the total pulse energy is concentrated in the 33-fs main pulse peak. In this case the optimum ratio is around to 0.4 fs.
Fig. 5 Measured temporal envelope profile of the compressed pulses at different output powers. The corresponding spectra are shown in inset.
The finite gain bandwidth is observed on the spectral shape with increasing output power. The relatively steeper short wavelength side extends more slowly than the long wavelength due to the finite gain bandwidth. The increase of gain extended the spectrum towards the longer wavelength induced by SPM [24]. Simultaneously, the short wavelength side was slightly extended with a cut off wavelength at 980 nm due to the self-absorption [25]. With further increase of the output power, both finite gain bandwidth and self-phase modulation force the asymmetric spectral shape, which leads to an asymmetric phase. This nonlinear spectral phase is an important source for pulse broadening and shape distortion. Experimentally, the above-mentioned nonlinear spectral phase could be retrained by applying more pre-chirping TGR ratio.
4. Conclusion
In conclusion, we demonstrate a TOD managed nonlinear amplifier operating at a 55-MHz repetition rate, generating 93.5 W compressed pulse with an pulse duration of 33 fs, and 51-MW peak power. The laser system exploits self-similar amplification with a grism-based pre-chirper for optimizing the pre-chirping ratio, which significantly enhances the quality of the compressed pulses. A complete characterization of the high power laser in terms of temporal and spectral profiles prove the suitability of this system for high-resolution spectroscopy and optical frequency comb. The combination of this broadband femtosecond laser source with the optical parametric oscillator in singly-resonant configuration may facilitate the development of a high-power, mid-infrared femtosecond lasers at high repetition rates.
Acknowlegments
This work was partly supported by National Natural Science Foundation of China (NSFC) (11434005 & 11422434), National Instrumentation Program (2012YQ150092), and funds from Shanghai Science and Technology Commission (14QA1401600 & 14JC1401600).
References and links
1. T. Nakamura, I. Ito, and Y. Kobayashi, “Offset-free broadband Yb:fiber optical frequency comb for optical clocks,” Opt. Express 23(15), 19376–19381 (2015). [CrossRef] [PubMed]
2. C. Hu, T. Chen, P. Jiang, B. Wu, J. Su, and Y. Shen, “Broadband high-power mid-IR femtosecond pulse generation from an ytterbium-doped fiber laser pumped optical parametric amplifier,” Opt. Lett. 40(24), 5774–5777 (2015). [CrossRef] [PubMed]
3. A. Volpe, F. Di Niso, C. Gaudiuso, A. De Rosa, R. M. Vázquez, A. Ancona, P. M. Lugarà, and R. Osellame, “Welding of PMMA by a femtosecond fiber laser,” Opt. Express 23(4), 4114–4124 (2015). [CrossRef] [PubMed]
4. L. Shi, W. Li, H. Zhou, D. Wang, L. Ding, and H. Zeng, “Enhanced ultraviolet pulse generation via dual-color filament interaction induced phase-matching control,” Appl. Phys. Lett. 102(8), 081112 (2013). [CrossRef]
5. L. Shi, W. Li, H. Zhou, L. Ding, and H. Zeng, “Impact excitation of neon atoms by heated seed electrons in filamentary plasma gratings,” Opt. Lett. 38(4), 398–400 (2013). [CrossRef] [PubMed]
6. D. Oron, N. Dudovich, D. Yelin, and Y. Silberberg, “Narrow-band coherent anti-Stokes Raman signals from broad-band pulses,” Phys. Rev. Lett. 88(6), 063004 (2002). [CrossRef] [PubMed]
7. K. Kim, X. Peng, W. Lee, S. Gee, M. Mielke, T. Luo, L. Pan, Q. Wang, and S. Jiang, “Monolithic polarization maintaining fiber chirped pulse amplification (CPA) system for high energy femtosecond pulse generation at 1.03 µm,” Opt. Express 23(4), 4766–4770 (2015). [CrossRef] [PubMed]
8. L. Shah, Z. Liu, I. Hartl, G. Imeshev, G. Cho, and M. Fermann, “High energy femtosecond Yb cubicon fiber amplifier,” Opt. Express 13(12), 4717–4722 (2005). [CrossRef] [PubMed]
9. V. I. Kruglov, A. C. Peacock, J. M. Dudley, and J. D. Harvey, “Self-similar propagation of high-power parabolic pulses in optical fiber amplifiers,” Opt. Lett. 25(24), 1753–1755 (2000). [CrossRef] [PubMed]
10. A. Chong, L. Kuznetsova, and F. Wise, “Theoretical optimization of nonlinear chirped-pulse fiber amplifiers,” J. Opt. Soc. Am. B 24(8), 1815–1823 (2007). [CrossRef]
11. A. Ruehl, A. Marcinkevicius, M. E. Fermann, and I. Hartl, “80 W, 120 fs Yb-fiber frequency comb,” Opt. Lett. 35(18), 3015–3017 (2010). [CrossRef] [PubMed]
12. Y. Deng, C. Y. Chien, B. G. Fidric, and J. D. Kafka, “Generation of sub-50 fs pulses from a high-power Yb-doped fiber amplifier,” Opt. Lett. 34(22), 3469–3471 (2009). [CrossRef] [PubMed]
13. D. B. Soh, J. Nilsson, and A. B. Grudinin, “Efficient femtosecond pulse generation using a parabolic amplifier combined with a pulse compressor. II. Finite gain-bandwidth effect,” J. Opt. Soc. Am. B 23(1), 10–19 (2006). [CrossRef]
14. D. N. Papadopoulos, Y. Zaouter, M. Hanna, F. Druon, E. Mottay, E. Cormier, and P. Georges, “Generation of 63 fs 4.1 MW peak power pulses from a parabolic fiber amplifier operated beyond the gain bandwidth limit,” Opt. Lett. 32(17), 2520–2522 (2007). [CrossRef] [PubMed]
15. J. Zhao, W. Li, C. Wang, Y. Liu, and H. Zeng, “Pre-chirping management of a self-similar Yb-fiber amplifier towards 80 W average power with sub-40 fs pulse generation,” Opt. Express 22(26), 32214–32219 (2014). [CrossRef] [PubMed]
16. W. Liu, D. N. Schimpf, T. Eidam, J. Limpert, A. Tünnermann, F. X. Kärtner, and G. Chang, “Pre-chirp managed nonlinear amplification in fibers delivering 100 W, 60 fs pulses,” Opt. Lett. 40(2), 151–154 (2015). [CrossRef] [PubMed]
17. Y. Liu, W. Li, J. Zhao, D. Bai, D. Luo, and H. Zeng, “High-power pre-chirp managed amplification of femtosecond pulses at high repetition rates,” Laser Phys. Lett. 12(7), 075101 (2015). [CrossRef]
18. V. I. Kruglov, C. Aguergaray, and J. D. Harvey, “Propagation and breakup of pulses in fiber amplifiers and dispersion-decreasing fibers with third-order dispersion,” Phys. Rev. A 84(2), 023823 (2011). [CrossRef]
19. S. Kane and J. Squier, “Grating compensation of third-order material dispersion in the normal dispersion regime: sub-100-fs chirped-pulse amplification using a fiber stretcher and grating-pair compressor,” IEEE J. Quantum Electron. 31(11), 2052–2057 (1995). [CrossRef]
20. L. Kuznetsova, F. Wise, S. Kane, and J. Squier, “Chirped-pulse amplification of femtosecond pulses in a Yb-doped fiber amplifier near the gain narrowing limit using a reflection grism compressor,” in Advanced Solid-State Photonics, OSA Technical Digest Series (Optical Society of America, 2007), paper TuB3.
21. S. Zhou, L. Kuznetsova, A. Chong, and F. Wise, “Compensation of nonlinear phase shifts with third-order dispersion in short-pulse fiber amplifiers,” Opt. Express 13(13), 4869–4877 (2005). [CrossRef] [PubMed]
22. A. Galvanauskas, Ultrafast Lasers (CRC, 2002).
23. Y. Zaouter, D. N. Papadopoulos, M. Hanna, F. Druon, E. Cormier, and P. Georges, “Third-order spectral phase compensation in parabolic pulse compression,” Opt. Express 15(15), 9372–9377 (2007). [CrossRef] [PubMed]
24. P. Wan, L. M. Yang, and J. Liu, “High pulse energy 2 µm femtosecond fiber laser,” Opt. Express 21(2), 1798–1803 (2013). [CrossRef] [PubMed]
25. L. J. Kong, X. S. Xiao, and C. X. Yang, “Tunable all-normal-dispersion Yb-doped mode-locked fiber lasers,” Laser Phys. 20(4), 834–837 (2010). [CrossRef]
