All glass leakage channel fibers have been demonstrated to be a potential practical solution for power scaling in fiber lasers beyond the nonlinear limits in conventional large mode area fibers. The all glass nature with absence of any air holes is especially useful for allowing the fibers to be used and fabricated much like conventional fibers. Previously, double clad active all glass leakage channel fibers used low index polymer as a pump guide with the drawbacks of being less reliable at high pump powers and not being able to change fiber outer diameter independent of pump guide dimension. In this work, we demonstrate, for the first time, ytterbium-doped double clad all glass leakage channel fibers with highly fluorine-doped silica as pump cladding. The new all glass leakage channel fibers have no polymer in the pump path and have independent control of fiber outer diameters and pump cladding dimension, and, therefore, enable designs with smaller pump guide for high pump absorption and, at the same time, with large fiber diameters to minimize micro and macro bending effects, a much desired features for large core fibers where intermodal coupling can be an issue due to a much increased mode density. An ytterbium-doped double clad PM fiber with core diameter of 80μm is also reported, which can be coiled in 76cm diameter coils.
©2009 Optical Society of America
Fiber lasers have demonstrated their great potential in commercial single mode laser systems of high average powers, largely due to their ease of achieving a single spatial mode  and excellent heat dissipation from their large surface just few hundred micrometers away from the active area. The extremely long, typically a few meters, and slender geometry provides a surface in close proximity to the active media of at least two orders of magnitude more than that in a typical disc laser. Powerful ultra short pulses, on the another hand, have shown great benefits in material processing due to their high peak powers and avoidance of the much slower but detrimental thermal effect. It is, however, very challenging to increase peak powers available from fiber lasers to the levels of solid state lasers while providing robust single mode output. The tight confinement of the guided modes has also proved to be highly limiting due to their low nonlinear thresholds. In recent years, photonic crystal fibers (PCF) have been developed to provide a much larger fiber core diameter over that in conventional fiber designs and much increased nonlinear thresholds while maintaining good single mode output [2–5]. Active PCFs with core diameters of 100µm have been demonstrated to provide reasonable single mode operation . PCFs with core diameters beyond 40µm cannot be bent without severe loss penalty and are typically made in straight rods of ~1m long. The short length can limit the available gain from a fiber amplifier, leading to complex configurations with multi-stage amplifications . A polarized output is also required in many applications, especially in wavelength conversions and CPA systems. Recently, a PM PCF rod with core diameter of 70µm was demonstrated to address this issue .
On the other hand, most nonlinear thresholds also scale with the inverse of amplifier length. A shorter amplifier with the same gain is better at producing high peak power pulses. One way to achieve this is by reducing the ratio of pump to doped-core areas in double-clad fibers to increase pump absorption when gain saturation is not an issue. Since most large core fibers tend to be more sensitive to micro and macro bending, which can lead to mode degradation and loss due to intermodal coupling enhanced by the higher mode density, large fiber diameters are desirable to ease this . Air-clad pump guides are used in [3–5] to allow fiber diameter adjustment to be independent of pump guide. The removal of polymer from pump path in these fibers with air-clad pump guides, compared to conventional double clad fibers, also improves fiber laser reliability at high pump powers. Inclusion of air holes in fibers does lead to some compromise in geometry controls in fiber fabrication. Furthermore, these air holes need to be sealed or spliced at both ends in fiber lasers, which also increases the chance of mode distortion, especially at large core diameters.
All glass leakage channel fibers (LCFs) have been demonstrated to be a promising solution to achieve large effective mode areas with an all glass structure very similar to that of conventional fibers [7,8]. Previously, we have demonstrated ytterbium-doped double clad LCFs with low index polymer pump cladding. In this work, we will further demonstrate, for the first time, ytterbium-doped double clad all glass LCFs with highly fluorine-doped silica pump cladding to provide a pump NA of ~0.28, which not only removes all polymer from the pump path but also enables designs with large fiber diameters to ease micro and macro bending effects in large core LCFs while allowing a much smaller pump guide for high pump absorptions. Direct amplification of 10ps pulses to over 1MW peak powers is demonstrated for the first time in those fibers to the best of our knpowledge. Both non-PM and PM LCFs are demonstrated in this configuration. Their performance in fiber lasers is also demonstrated in this work.
2. Non-PM Leakage Channel fibers
All LCFs in this work have a refractive index difference between the background and the low index feature of Δn = 1.2 × 10−3. The low index features were made of slightly fluorine-doped silica. Two non-PM fibers, LCF1 and LCF2, shown in Fig. 1(a) and 1(b) respectively, were fabricated. LCF1 has a 50µm core diameter, feature size to pitch ratio d/Λ = 0.7, and a rounded hexagonal pump guide. The pump cladding was made of highly fluorine-doped silica (see the dark ring in Fig. 1) with a dimension of 222µm by 250µm, providing a pump NA of 0.28. LCF1 has a pump absorption of ~12dB/m and an outer diameter of 450µm coated with regular high index coating. Coating was used for the protection of glass surface for reliability. With the highly fluorine-doped pump cladding, coating will not see the high optical power from the pump. Both the non-circular inner boundary of the pump cladding and the two rings of low index features around the doped core help pump mode mixing in order to provide an enhanced pump absorption. LCF2 was designed for an improved bending performance with an inner ring d/Λ = 0.8 and outer ring d/Λ = 0.7. It has a 47µm core diameter, a rounded hexagon pump guide with a dimension of 238µm by 256µm, a pump NA of 0.28, a pump absorption of ~12dB/m, and an outer diameter of 538µm coated with regular high index coating, shown as the outermost layer in Fig. 1(b).
In the first experiment, 5.2m LCF1 was used in a coil of 60cm diameter. Both LCF1 and LCF2 can operate in fundamental mode in short straight fibers without any bending. This bending diameter was chosen to be large enough not to cause excessive transmission loss for the fundamental mode. A mode-locked fiber oscillator delivering 15mW 13.6ps near transform-limited pulses at 1037nm and 10MHz was used to seed the amplifier, similar to that described in . The amplifier was counter-pumped by a 100W multimode 980nm pump laser diode delivered in a 200µm core fiber with 0.22NA. A slope efficiency of ~68% was achieved, shown in Fig. 2(a) . Expected SPM spectral broadening is clearly seen in Fig. 2(b). A maximum 30W, 3µJ, 220kW peak power and 4.5nm FWHM were obtained. The fiber dispersion is ~–30ps/nm/km at 1.05µm, dominated by material dispersion and is not expected to have much effect on temporal pulse shape in this short length of fiber and narrow spectral width. Peak power was calculated by dividing pulse energy by pulse FWHM width. No sign of Raman scattering was observed at longer wavelengths even at the maximum peak powers. The measured spectra are plotted in the Fig. 2(b) at various output peak powers. A maximum peak power of 90kW can be obtained with negligible spectral broadening of <0.5nm. The amplifier provided a gain of 34dB at the maximum output power. The mode of the amplifier fiber is given in Fig. 3 , showing the expected fundamental mode of the fiber.
To achieve higher average powers in the second experiment, we used another mode-locked fiber oscillator which operates at 1037nm delivering 60mW near transform-limited 15ps pulse at 48MHz in an amplifier based on a 4.8m long LCF2, counter-pumped by a 200W multimode pump with a delivery fiber of 200µm core and 0.22NA. A smaller coil diameter of 52cm diameter was used in this case due to the slightly larger d/Λ of the inner ring features which lowers critical bend radius. The amplifier performance is shown in Fig. 4 . A maximum average power of 98W (2µJ pulse energy, 137kW peak power, 33dB gain) with a slope efficiency of 71% was achieved with negligible ASE. Coating at the pump end started to burn at high pump powers, this lead to the apparent roll-off in Fig. 4. This can be mitigated by an improved pump coupling arrangement. The experiment was repeated with a different sample of 3.5m long LCF2, also shown in Fig. 4. A slope efficiency of 75% was achieved in this case. The output mode was measured at 14W output power and is shown in the inset in Fig. 4. The new LCFs allow us to reach ~100W without significant SPM broadening using a much simpler scheme of direct single-stage amplification of a fiber oscillator without the need of pre-chirping used in [10,11].
In a third experiment, the output from a mode-locked oscillator providing 120mW, 78MHz and 15ps Gaussian pulses was down-counted to 1MHz. The output from the down-counter was then amplified to >30mW at 1MHz. The much lower repetition rate allowed us to reach much higher pulse energy with the pump powers available to us. A 4m long LCF2 coiled to a 50cm diameter was used in this experiment. Pulse energy as high as 11.2µJ, 11.2W average power, was obtained. The output spectra are shown in Fig. 5(a) , along with simulated spectra from the generalized NLS equation without considering Raman and self-steepening effects. The simulation used the following parameters considered appropriate for the experimental conditions, γ = 0.11 W/km, β2 = 20 ps2/km, β3 = 4.4 × 10−2 ps3/km, gain α = −1.74 1/m, and an input Gaussian pulse with peak power of 1kW and FWHM of 10ps. The effect of SPM is well predicted by the simulation. Autocorrelation of the input pulse was measured. Gaussian shape with a FWHM of 14.2ps was measured, implying a Gaussian pulse with a FWHM of ~10ps. Autocorrelation traces of the output at various pulse energies were measured in a separate experiment using a similar LCF. It was confirmed that the measured autocorrelation traces remained unchanged at pulse energies all the way up to the eventual pulse breakup, which happened at 10dB spectral bandwidth >~20nm and proceeded by observable SRS. Simulated pulse shapes at various peak powers are given in Fig. 5(b), showing that Gaussian pulse shape remains unchanged up to 1MW peak power and confirming the experimental results. This allows us to estimate a peak power of 1.1MW at the maximum output pulse energy of 11.2µJ, the highest peak power at this short pulse duration directly generated in fibers according to the best of the authors' knowledge. Simulated chirp is also shown in Fig. 5(c). Spectral bandwidth measured at the –10 dB points is given in Fig. 6(a) , along with the simulated result, showing good agreement. A wide scan of the output spectra at the higher pulse energies are shown in Fig. 6(b), showing some sign of SRS at 11.2μJ. Recently, similar SPM-broadened pulse formation with longer pulse was observed in PCFs .
3. PM Leakage Channel fibers
In the first experiment with all glass PM LCFs, a passive all-glass PM LCF was fabricated with a core diameter of 50µm to demonstrate the basic characteristics of PM LCFs. A high d/Λ = 0.9 was used for a smaller critical bend radius. Two stress elements with a refractive index of ~13 × 10−3 below that of the background silica glass were introduced instead of the usual features on either sides of the core to provide birefringence (see Fig. 7(a) and 7(b)). The fiber has an outer diameter around 885µm and is coated with standard acrylic coating. The measured near field imaged with a single lens is shown in Fig. 7(c) at the output of a 1.8m long sample. Due to the much higher d/Λ = 0.9 for this fiber, some bending was necessary for the fundamental mode operation in short length of this fiber, unlike LCF1 and LCF2 where no such bending is necessary even in a short straight length of less than 1m. The output was robustly single mode, however, in a 30m long sample of this passive LCF (see Fig. 7(d)) coiled in 40cm diameter coils (This small coiling diameter of 40cm was required due to limited space requirement in a subsequent oscillator experiment). Bend loss was measured using a 5m sample coiled in various diameters. The result is show in Fig. 8(a) . The critical bend radius for 1dB/m loss is expected to be ~11cm by our FEM simulation , matched very well to the measured 10.5cm. PER was characterized at the output of a 1.8m long sample with an ytterbium ASE source as the input, ensuring that the input was aligned with the fast axis and then measuring the output with an output polarizer aligned with the fast and slow axis respectively. The measured PER was >15dB over the entire range of the ASE source (see Fig. 8(b)). The polarization mode beating was also measured with both input and output polarizers aligned at 45 degree to the fast axis. The result is also shown Fig. 8(b) and the corresponding birefringence was calculated to be ~8 × 10−5. PER was also measured at the output of the 30m long. The result is shown in Fig. 8(c), again showing >15dB PER over the entire range of the ASE source.
An ytterbium-doped all-glass PM LCF with a core diameter of 80µm was then fabricated (see Fig. 9(a) ). The active PM LCF also has a fluorine-doped pump cladding, providing a pump NA of ~0.28. Low index features with an inner layer d/Λ of 0.8 and an outer layer d/Λ of 0.7 were used. This active PM LCF has a pump guide of ~400µm (flat-to-flat) and a fiber outer diameter of ~835µm and is coated with standard acrylic coating. Pump absorption was estimated to be ~12dB/m. The mode field diameter was measured to be ~62 ± 2µm, by imaging the output mode onto a CCD camera and calibrating against the pump guide dimension which was measured independently. The output from a fiber oscillator was again used for seeding in a counter-propagation pumping configuration with a 4m long PM LCF. The fiber oscillator provides transform-limited pulses of ~40mW average power and 14.2ps pulse width at 10MHz repetition rate. The seed was directly injected into the 4m long fiber with both ends angle-polished. The fiber was in a single coil 76cm in diameter with a length of straight section at each end. The average power at the amplifier output is plotted in Fig. 9(b) versus launched pump powers from a pump diode at ~976nm in a 200µm diameter delivery fiber. A slope efficiency of ~74% was measured along with a maximum single path gain in excess of 30dB. A maximum average power of ~27.4W was achieved before ASE suppression increased beyond 30dB. This gave 14.2ps and ~3 times transform-limited pulses of 2.74µJ pulse energy and ~190kW peak power. This is a record peak power to our knowledge for direct amplification of near transform-limited pulses with minimal SPM broadening. Previously, pulses at 47MHz and 200kW peak power were achieved only through pre-chirping to compensate SPM . M2 values were also characterized at various output power levels and found to be below 1.35 for the entire output power range. A near field mode patterns is shown in the inset of Fig. 9(b). The achieved SPM suppression as a result of large effective mode area is further illustrated in Fig. 10(a) , showing the small difference in pulse spectra measured at various amplifier outputs up to 27.4W, or 230kW peak power. PER was also characterized at various output power and found to be above 15dB for the entire range of output powers. A wide spectral scan at the maximum output power of 27.4W is also shown in Fig. 9(b), showing the substantial ASE suppression and absence of any Raman scattering. In a separate experiment with a slightly larger coil diameter of 88cm, a slightly higher slope efficiency of 78% was obtained. It is worth noting that some reduction of effective mode area due to bending is expected in the coiled section of this fiber .
4. Summary and conclusions
To summarize, we have demonstrated both non-PM and PM all glass ytterbium-doped LCFs with highly fluorine-doped silica pump cladding, a further step towards a practical, robust and reliable solution offering single mode operation with a large effective mode area for high power fiber lasers in an all glass configuration. The new all glass LCFs are easy to fabricate and use, while providing a large effective mode area, polarization preservation, high gain, and an improved ability to be coiled.
References and links
1. E. Snitzer, “Proposed fiber cavities for optical masers,” J. Appl. Phys. 32(1), 36–39 (1961). [CrossRef]
2. J. Limpert, A. Liem, M. Reich, T. Schreiber, S. Nolte, H. Zellmer, A. Tünnermann, J. Broeng, A. Petersson, and C. Jakonsen, “Low-nonlinearity single-transverse-mode ytterbium-doped photonic crystal fiber amplifier,” Opt. Express 12, 1313–1319 (2004). [CrossRef] [PubMed]
3. J. Limpert, N. Deguil-Robin, I. Manek-Hönninger, F. Salin, F. Röser, A. Liem, T. Schreiber, S. Nolte, H. Zellmer, A. Tünnermann, J. Broeng, A. Petersson, and C. Jakobsen, “High-power rod-type photonic crystal fiber laser,” Opt. Express 13, 1055–1058 (2005). [CrossRef] [PubMed]
4. O. Schmidt, J. Rothhardt, T. Eidam, F. Röser, J. Limpert, A. Tünnermann, K. P. Hansen, C. Jakobsen, and J. Broeng, “Single-polarization ultra-large-mode-area Yb-doped photonic crystal fiber,” Opt. Express 16(6), 3918–3923 (2008). [CrossRef] [PubMed]
5. C. D. Brooks and F. Di Teodoro, “Multi-megawatt peak-power, single-transverse-mode operation of a 100 µm core diameter, Yb-doped rod-like photonic crystal fiber amplifier,” Appl. Phys. Lett. 89(11), 111119–111121 (2006). [CrossRef]
6. M. E. Fermann, “Single-mode excitation of multimode fibers with ultrashort pulses,” Opt. Lett. 23(1), 52–54 (1998). [CrossRef]
7. L. Dong, J. Li, H. A. McKay, A. Marcinkevicius, B. K. Thomas, M. Moore, L. Fu, and M. E. Fermann, Robust and practical optical fibers for single mode operation with core diameters up to 170µm, ” CLEO, post-deadline paper CPDB6, san Jose, 2008.
8. L. Dong, T. W. Wu, H. A. McKay, L. Fu, J. Li, and H. G. Winful, “All-Glass Large-Core Leakage Channel Fibers,” IEEE J. Sel. Top. Quantum Electron. 15(1), 47–53 (2009). [CrossRef]
9. I. Hartl, G. Imeshev, L. Dong, G. C. Cho and M. E. Fermann, “Ultra-compact dispersion compensated femtosecond fiber oscillators and amplifiers,” CLEO, paper CThG1, 2005.
10. J. Limpert, N. Deguil-Robin, I. Manek-Hönninger, F. Salin, T. Schreiber, A. Liem, E. Röser, H. Zellmer, A. Tünnermann, A. Courjaud, C. Hönninger, and E. Mottay, “High-power picosecond fiber amplifier based on nonlinear spectral compression,” Opt. Lett. 30(7), 714–716 (2005). [CrossRef] [PubMed]
11. Y. Zaouter, E. Cormier, P. Rigail, C. Hönninger, and E. Mottay, “30W, 10µJ, 10ps, SPM-induced spectrally compressed pulse generation in a low non-linearity ytterbium-doped rod-type fiber amplifier,” Proc. SPIE 6453, 64530O (2007). [CrossRef]
12. A. M. Thomas, D. Alterman, and M. S. Bowers, “High peak power short-pulse fiber lasers for material processing, ” SPIE Photonics West, 7195–43, San Jose, 2009.
13. J. Limpert, N. Deguil-Robin, I. Manek-Hönninger, F. Salin, T. Schreiber, A. Liem, E. Röser, H. Zellmer, A. Tünnermann, A. Courjaud, C. Hönninger, and E. Mottay, “High-power picosecond fiber amplifier based on nonlinear spectral compression,” Opt. Lett. 30(7), 714–716 (2005). [CrossRef] [PubMed]