Open Access
Add to BibSonomyAbstract
We demonstrate soliton self-frequency-shifted, femtosecond-pulse amplification in a newly-developed, polarization-maintaining, Er-doped, very-large-mode-area fiber amplifier. The PM-VLMA Er fiber had a core diameter of 50 μm, an effective area of ~1050 μm2, and Er absorption of 50 dB/m. The measured birefringence beat length of the PM-VLMA Er fiber was 14.1 mm. The soliton wavelength could be shifted by more than 90 nm. The soliton generation process resulted in remarkably clean, 86 fs pulses with 21 nJ energy at 1650 nm and 244 kW peak power from an all-fiber, fusion spliced system without bulk-optics for pulse compression. The polarization extinction ratio of the soliton was greater than 40 dB, and the M2 was 1.1. The fully polarization-maintaining fiber laser system provides robust and stable soliton generation. Peak-to-peak variation in the soliton wavelength, measured over the course of an hour was only 0.03% and pulse energy variation was only 0.5%.
© 2016 Optical Society of America
1. Introduction
Emerging applications in ultrafast optics such as two photon polymerization, nonlinear microscopy, electro-optical sampling and terahertz imaging benefit from the excellent beam quality provided by fiber lasers, but will be further enabled with increases in pulse energy. Furthermore, eliminating the need for additional dispersive control of the pulse width will provide simple, direct amplification of femtosecond pulses in fiber, increasing reliability and decreasing cost.
Soliton self-compression and the soliton self-frequency shift enable a fiber-based approach to tunable femtosecond pulse generation with no need for bulk compression optics [1]. Solitons in fibers transfer energy from higher to lower frequencies due to the effect of intra-pulse stimulated Raman scattering (SRS). For solitons with pulse durations shorter than 100 fs, the optical spectrum broadens to such an extent that the longer wavelength tail experiences Raman amplification at the expense of power at shorter wavelengths [2]. This leads to an overall spectral shift of the soliton towards longer wavelengths, called the soliton self-frequency shift [3–5]. Using passively modelocked Er-doped fiber lasers starting at 1560 nm, together with single-mode, polarization maintaining fiber, the soliton wavelength could be shifted to as long as 1780 nm [6].
However, the energy within the Raman shifted soliton can be limited, as the peak power required for the fundamental soliton is proportional to the mode effective area and inversely proportional to the nonlinear refractive index. Due to the small effective area, typical soliton energies for single mode fibers are of the order of 1 nJ [4]. Therefore, to increase the soliton energy, various types of specialty fibers have been utilized, with an aim towards decreasing the fiber nonlinearity and increasing mode effective area. In general the strategy is to use large mode area (LMA) fibers to enable high energy, self-frequency shifted solitons [7–12]. For example, by utilizing hollow-core photonic bandgap fibers [7], which can have extremely low nonlinearity, or photonic crystal [8] or higher-order mode fibers [12], which can have ultra-large effective areas, solitons with peak powers in excess of 1 MW can be generated. Because soliton generation requires anomalous dispersion, most work on self-frequency shifted solitons has been at wavelengths of 1550 nm, although with specialty structures such as hollow-core fiber [7] or higher-order modes [11], high energy solitons can be generated at shorter wavelengths such as 800 nm, or 1 micron.
Most work on high-energy, self-frequency shifted solitons has utilized passive fibers for soliton generation. Thus, a high-energy femtosecond pulse is required at the launch into the soliton generating fiber. Alternatively, a rare-earth-doped, large-mode area fiber with anomalous dispersion can simultaneously provide amplification of a low power femtosecond seed pulse as well as self-frequency-shifted soliton generation, further simplifying the system architecture. In addition, a remarkable advantage of a soliton amplifier is sub 100 fs pulses can be easily achieved as the soliton is not limited by gain narrowing effects, in contrast to chirp-pulse, or divided pulse amplification.
Recently, soliton amplification and self-frequency shifting was demonstrated in an Er-doped very-large-mode-area (VLMA) fiber amplifier [12]. The VLMA fiber [13] had an effective area of approximately 1050 μm2. The large area enabled high energy, 17 nJ pulses with 110 fs pulse width, free of satellite pulses. However, this demonstration of self-frequency shifted soliton generation in the VLMA-Er fiber, as well as other prior experiments on self-frequency shifted solitons in passive LMA passive, utilized non-polarization maintaining fibers. Non-polarization maintaining fiber ultimately limits long-term environmental stability of the laser system. For long term stability, a fully-polarization maintaining system should be utilized. In this work, we demonstrate for the first time amplification and soliton self-frequency shifting of femtosecond pulses in a newly developed polarization-maintaining (PM) VLMA fiber amplifier. Using this fiber, we demonstrate stable, self-frequency shifted solitons at 1650 nm with 86 fs pulse width, 21 nJ pulse energy, and polarization extinction ratio greater than 40 dB.
2. Self-frequency shifted solitons in a polarization-maintaining, very-large-mode area, Er-doped fiber amplifier
A schematic of the experimental setup for self-frequency shifted soliton generation in the PM-VLMA Er-doped fiber amplifier is shown in Fig. 1. The core design of the polarization maintaining VLMA-Er fiber was similar to previous demonstrations of non-PM VLMA-Er fibers [13], with a core-diameter of 50 μm, and Er absorption of 50 dB/m at 1530 nm. The effective area of the fundamental mode at 1550 nm was ~1050 μm2 when the fiber was straight, however at the operating coil diameter of 25 cm, the effective area of the fiber was reduced to approximately 950 μm2. Polarization-maintaining operation was provided by stress rods, with a measured birefringence beat length of 14.1 mm. The PM-VLMA Er-doped fiber amplifier was recently used for single-frequency, microsecond pulse amplification at 1572.3 nm for CO2 LIDAR applications [14].
Fig. 1 Experimental setup for self-frequency shifted soliton generation in a PM-VLMA Er-doped fiber amplifier.
The PM-VLMA amplifier was core-pumped by a (non-PM) Raman fiber laser providing 20 W, single-transverse-mode output at 1480 nm [15]. Both pump and signal propagate in the fundamental mode of the PM-VLMA amplifier, providing high overlap between pump and signal and suppression of unwanted higher-order modes (HOMs) via differential gain [16]. As a result, unwanted HOMs typically only amount to 1 to 2% of output power [13]. Furthermore, core-pumping helps keep the amplifier short compared to cladding pumping, decreasing nonlinear impairments.
An all-PM, modelocked, femtosecond fiber laser was used as a seed source. The seed laser consisted of a SESAM modelocked oscillator seeding a parabolic, single-stage amplifier all pumped with one diode. The output of the seed laser produced 100mW average power at 100 MHz repetition rate and a linear chirped pulse output of about 600fs with a spectral width of 20 nm. The pulse was chirped such that it compressed within the first few centimeters of the PM-VLMA fiber to sub 100 fs pulse width.
The pump and signal were combined using a polarization-maintaining, fused-fiber wavelength-division multiplexer (WDM). The output end of the PM-VLMA Er fiber was angle cleaved. The fiber output was collimated, and a high-pass filter was used to transmit the soliton, while reflecting the residual Stokes orders from the Raman laser, the un-absorbed 1480 nm pump, and the fundamental signal pulse at 1560 nm.
As the pump power was increased, initially, the output power of the fundamental pulse increased. However, above a pump power of approximately 6 W, the soliton was observed to break away from the main pulse, and with increasing pump power, the soliton shifted to longer wavelengths, shown in Fig. 2(a). In Fig. 2(a), the un-shifted signal pulse at 1560 nm was removed using the high-pass filter. The PM-VLMA fiber length in this instance was 3.7 m.
Fig. 2 (a) Soliton shifting as a function of the pump power. The unshifted pulses at 1560 nm have been blocked by the high-pass filter. (b) Total signal output power (unshifted 1560 nm signal pulse plus soliton without high-pass filter) compared to power in the soliton, as a function of pump power.
Output power vs. pump power is shown in Fig. 2(b), comparing total signal output power vs. power only in the soliton. In this experiment, the high-pass filter was used to remove the unabsorbed 1480 nm pump, but not the fundamental 1560 nm un-shifted signal pulse, and the power in the soliton was estimated from the measured optical spectra.
Initially the soliton energy increased with increasing pump power, but above a pump power of 9 W, the soliton energy decreased. This energy dependence with input power is different than observed in soliton shifting experiments in passive fibers, where the soliton energy monotonically increases with increasing input power [8,9]. This difference in behavior can be understood by considering the effective length of fiber over which soliton shifting occurs. In a passive fiber, the effective length is constant with input power. In contrast, in the amplifier fiber, the effective interaction length increases with increasing pump power, leading to a decrease in soliton energy at high pump powers.
Properties of the soliton vs. pump power are shown in Fig. 3, for two different PM-VLMA Er-fiber amplifier lengths of 3.7 m and 3 m. Figure 3(a) shows energy in the soliton versus pump power. Figure 3(b) shows the center wavelength of the soliton vs. pump power, and Fig. 3(c) shows pulse full-width at half-maximum (FWHM). The pulse FWHM was measured using a home-built, two-photon autocorrelator. The pulse FWHM was estimated assuming a decorrelation factor of 1.54 for a sech2 pulse.
Fig. 3 Soliton properties for a 3.7 m long PM-VLMA Er fiber amplifier compared to a 3 m long amplifier. (a) Energy in the soliton, (b) center wavelength of the soliton, and (c) pulse FWHM of the soliton.
The longest center wavelength of the soliton obtained was 1672 nm, independent of the amplifier length. With the longer amplifier fiber, soliton shifting required slightly less pump power for a given target wavelength. However, the shorter amplifier fiber provided the highest soliton pulse energy of 21 nJ and shortest pulse width of 86 fs.
For both amplifier lengths the pulse width initially decreased, and then increased again with increasing pump power. We attribute the initial decrease in pulse width to the presence of the high-pass filter, which cut some of the soliton bandwidth for small wavelength shifts. In passive fibers, the frequency-shifted soliton pulse width increases with increasing fiber length [5]. However, as discussed above, the effective length of fiber through which the soliton propagates also increases with increasing pump power. Thus, the increase in pulse width with increasing pump power is consistent for solitons shifted in the amplifier fiber.
The spectrum, on a linear scale, and interferometric autocorrelation for the shortest pulse with 21 nJ energy and 86 fs pulse width from the 3 m long amplifier is shown in Figs. 4(a) and 4(b), respectively. Both the spectrum and autocorrelation were remarkably clean, features not typically associated with ultrashort-pulse fiber laser systems. The spectrum was free from ripple, and the autocorrelation did not show any indication of satellite pulses.
Fig. 4 (a) Spectrum, on a linear scale, and (b) interferometric autocorrelation, corresponding to the shortest pulse with 21 nJ energy from the 3 m long amplifier fiber. For comparison, the spectrum from the highest energy soliton obtained in a non-PM amplifier (see [12]) is shown in (c).
For comparison purposes, Fig. 4(c) shows the spectrum of a soliton with similar pulse energy and central wavelength, generated from a non-PM VLMA-Er fiber, such as reported in [12]. In those experiments, a polarization controller was used to optimize the polarization state before launch into the non-PM VLMA fiber. However, the spectrum of the soliton from the non-PM VLMA amplifier showed significant sub-structure, compared to the soliton from the PM-VLMA amplifier. The structure in the non-PM soliton is likely due to non-linear polarization mixing during the pulse compression and soliton generation process.
Figure 5 shows the result of the measurement of polarization extinction ratio (PER) of the soliton. To ensure that optics did not degrade the PER during measurement, the analyzing polarizer was placed immediately after the collimating lens, and before the high-pass pump filter, as it was found that the high-pass filter reduced the PER of the soliton substantially.
Fig. 5 Measurement of polarization extinction ratio. (a) PER at low power with the amplifier pumped to transparency. (b) PER of the soliton at 12 W of 1480 nm pump power.
Figure 5(a) shows the PER measurement of the amplifier at low power when pumped to transparency. Depending on wavelength, the signal PER varied between 20 and 30 dB. Figure 5(b) shows the PER at 12 W of pump power, with the soliton shifted to 1650 nm. At this pump power the PER of the residual, unshifted, 1560 nm pulse was approximately 15 dB. However, the PER of the soliton was measured to be greater than 40 dB. Only the polarization with high power undergoes wavelength shifting, and the soliton shifts to wavelength much longer than the launched seed laser. Consequently, the soliton shifting process in the PM amplifier provides substantial improvement in the PER.
The measurement of the beam profile of the soliton at 12 W pump power is shown in Fig. 6(a). M2 of the soliton was measured using a commercial device (ThorLabs M2 measurement system). The result, shown in Fig. 6(b), shows the soliton has near diffraction limited performance, with M2 = 1 0.1.
In addition to providing high polarization extinction, and a clean spectral, and temporal, profile, utilizing all polarization maintaining fibers makes the laser system robust against environmental fluctuations and provides long term stability of the soliton wavelength and energy, which is critical for using femtosecond lasers in industrial settings. The stability of the laser system is illustrated in Fig. 7. The soliton spectrum and pulse energy from a 3.7m long amplifier was monitored over the course of an hour. Figure 7(a) shows the soliton spectra measured over an hour overlaid on top of each other. Figures 7(b) and 7(c) plot the peak soliton wavelength and soliton pulse energy respectively. The peak-to-peak variation in the peak wavelength of the soliton over the course of the hour was only ~0.5 nm, or 0.03% of the center wavelength. The peak-to-peak pulse energy variation was < 0.1 nJ, or 0.5%.
Fig. 7 Stability of the soliton over time in a 3.7 m long PM VLMA amplifier. (a) Soliton spectra, measured over the course of one hour, and overlaid. (b) Peak soliton wavelength, and (c) soliton energy measured over the course of one hour.
3. Conclusions
In conclusion, we have demonstrated for the first time soliton generation and self-frequency shifting in a polarization-maintaining, very-large-mode area, Er-doped fiber amplifier. The newly developed PM-VLMA Er fiber had a core diameter of 50 μm, effective area of ~1050 μm2, and Er absorption of 50 dB/m. The measured birefringence beat length of the PM-VLMA Er fiber was 14.1 mm.
The solitons could be shifted to wavelengths as long as 1672 nm by adjusting the amplifier pump power. The maximum measured soliton energy was 21 nJ, and the pulse width was 86 fs, corresponding to a peak power of 244 kW. The beam quality of the soliton was near diffraction limited, with an M2 of 1.1. The peak-to-peak variation in the center wavelength of the soliton, measured over the course of an hour, was only 0.03%, and the pulse energy variation was 0.5%.
Importantly, the process of soliton generation provides substantial improvement in the polarization extinction ratio of the soliton, compared to the launched pulse, as the nonlinear process of the soliton generation shifts only the high-power pulse out of the wavelength range of the lower PER seed.
In these experiments, we did not use quantitative techniques to measure the higher-order mode content of the soliton. Interferometric measurements such as S2 [17] or C2 [18] imaging require a coherent beat between the fiber output and the interferometer reference. But the interferometer reference must overlap in wavelength with the fiber output, so such measurements will not work in the case of the wavelength shifting soliton. However, the higher-order mode content of the soliton should be substantially improved compared to the launched seed, for the same reasons as the PER of the soliton improves. Only the high power mode (the fundamental) shifts in wavelength, moving the soliton away from the wavelength of the launched HOMs. Indeed, the smooth nature of the soliton spectrum as well as the satellite-free interferometric autocorrelations are indications of exceedingly low HOM content in the soliton.
In summary, the polarization-maintaining, very-large-mode area, Er-doped fiber amplifier is an all-fiber, fusion-spliced laser system that provides spectrally and temporally clean, high-energy, wavelength tunable, environmentally stable, femtosecond pulses at the end of the fiber, without the need for bulk-optic compression.
References and links
1. M. Hofer, M. E. Fermann, A. Galvanauskas, D. Harter, and R. S. Windeler, “High-power 100-fs pulse generation by frequency doubling of an erbium ytterbium-fiber master oscillator power amplifier,” Opt. Lett. 23(23), 1840–1842 (1998). [CrossRef] [PubMed]
2. F. M. Mitschke and L. F. Mollenauer, “Discovery of the soliton self-frequency shift,” Opt. Lett. 11(10), 659–661 (1986). [CrossRef] [PubMed]
3. J. P. Gordon, “Theory of the soliton self-frequency shift,” Opt. Lett. 11(10), 662–664 (1986). [CrossRef] [PubMed]
4. J. H. Lee, J. van Howe, X. Liu, and C. Xu, “Soliton self-frequency shift: experimental demonstrations and applications,” IEEE J. Sel. Top. Quantum Electron. 14(3), 713–723 (2008). [CrossRef] [PubMed]
5. P. Govind, Agrawal, Nonlinear Fiber Optics, 4th ed. (Academic Press, 2007).
6. N. Nishizawa and T. Goto, “Compact system of wavelength-tunable femtosecond soliton pulse generation using optical fibers,” IEEE Photonics Technol. Lett. 11(3), 325–327 (1999). [CrossRef]
7. D. G. Ouzounov, F. R. Ahmad, D. Müller, N. Venkataraman, M. T. Gallagher, M. G. Thomas, J. Silcox, K. W. Koch, and A. L. Gaeta, “Generation of megawatt optical solitons in hollow-core photonic band-gap fibers,” Science 301(5640), 1702–1704 (2003). [CrossRef] [PubMed]
8. K. Wang, N. Horton, K. Charan, and C. Xu, “Advanced fiber soliton sources for nonlinear deep tissue imaging in biophotonics,” IEEE J. Sel. Topics. In Quantum Elec. 20, 6800311 (2014).
9. T. N. Nguyen, K. Kieu, D. Churin, T. Ota, M. Miyawaki, and N. Peyghambarian, “High power soliton self-frequency shift with improved flatness ranging from 1.6 to 1.78 μm,” IEEE Photonics Technol. Lett. 25(19), 1893–1896 (2013). [CrossRef]
10. S. Saint-Jalm, E. R. Andresen, P. Ferrand, A. Bendahmane, A. Mussot, O. Vanvincq, G. Bouwmans, A. Kudlinski, and H. Rigneault, “Fiber-based ultrashort pulse delivery for nonlinear imaging using high-energy solitons,” J. Biomed. Opt. 19(8), 086021 (2014). [CrossRef] [PubMed]
11. L. Rishøj, G. Prabhakar, J. Demas, and S. Ramachandran, “30 nJ, ~50 fs all-fiber source at 1300 nm using soliton shifting in LMA HOM fiber,” in Conference on Lasers and Electro-Optics, OSA Technical Digest (2016) (Optical Society of America, 2016), paper STh3O.3. [CrossRef]
12. A. Zach, W. Kaenders, J. W. Nicholson, J. Fini, and A. DeSantolo, “Demonstration of Soliton Self Shifting Employing Er3+ Doped VLMA- and HOM-Fiber Amplifiers” in Conference on Lasers and Electro-Optics (CLEO) (OSA, 2015), paper ATu2M.6.
13. J. C. Jasapara, M. J. Andrejco, A. DeSantolo, A. D. Yablon, Z. Varallyay, J. W. Nicholson, J. M. Fini, D. J. DiGiovanni, C. Headley, E. Monberg, and F. V. Dimarcello, “Diffraction-limited fundamental mode operation of core-pumped very-large-mode-area Er fiber amplifiers,” IEEE J. Sel. Top. Quantum Electron. 15(1), 3–11 (2009). [CrossRef]
14. J. W. Nicholson, A. DeSantolo, M. F. Yan, P. Wisk, B. Mangan, G. Puc, A. W. Yu, and M. A. Stephen, “High energy, 1572.3 nm pulses for CO2 LIDAR from a polarization-maintaining, very-large-mode-area, Er-doped fiber amplifier,” Opt. Express 24(17), 19961–19968 (2016). [CrossRef] [PubMed]
15. J. W. Nicholson, M. F. Yan, P. Wisk, J. Fleming, F. DiMarcello, E. Monberg, T. Taunay, C. Headley, and D. J. DiGiovanni, “Raman fiber laser with 81 W output power at 1480 nm,” Opt. Lett. 35(18), 3069–3071 (2010). [CrossRef] [PubMed]
16. Z. Várallyay and J. C. Jasapara, “Comparison of amplification in large area fibers using cladding-pump and fundamental-mode core-pump schemes,” Opt. Express 17(20), 17242–17252 (2009). [CrossRef] [PubMed]
17. J. W. Nicholson, A. D. Yablon, S. Ramachandran, and S. Ghalmi, “Spatially and spectrally resolved imaging of modal content in large-mode-area fibers,” Opt. Express 16(10), 7233–7243 (2008). [CrossRef] [PubMed]
18. D. N. Schimpf, R. A. Barankov, and S. Ramachandran, “Cross-correlated (C2) imaging of fiber and waveguide modes,” Opt. Express 19(14), 13008–13019 (2011). [CrossRef] [PubMed]
