Ytterbium-doped solid-core photonic bandgap fiber amplifiers operating at the long-wavelength edge of the ytterbium gain band are reported. The low-loss bandgap transmission window is formed in the very low gain region, whilst outside the bandgap, large attenuation inhibits the exponential growth of amplified spontaneous emission in the huge-gain 1030-1100 nm region. Hence parasitic-lasing-free, high-power amplification with a marked efficiency is enabled. A 32 W output at 1156 nm with a 66% slope efficiency and 30 W output at 1178 nm with a 58% slope efficiency were successfully obtained. To our knowledge, these are the highest output powers generating from active photonic bandgap fibers, as well as from ytterbium-doped fiber lasers at these wavelengths.
©2009 Optical Society of America
Yellow-orange light sources are required in various fields such as medical applications, high-resolution spectroscopy, and laser guide star (LGS). Especially the LGS application requires a high-power (>10 W), diffraction-limited source at 589 nm . For such applications, fiber laser-based sources are desirable because of fiber’s excellent properties such as high brightness, efficient heat dissipation, robustness, and so on. Frequency doubling of 1150-1200 nm fiber Raman lasers [2,3] and recent bismuth-doped fiber lasers [4,5] are being explored. However, both lasers need core-pumping with a 1 μm high-brightness source such as an Yb fiber laser. In addition, a fiber Raman laser particularly suffers bandwidth broadening and resultant low conversion efficiency in the frequency-doubling stage.
Ytterbium (Yb)-doped silica fiber itself has a very broadband fluorescence amounting up to 1200 nm and thus also has been extensively investigated for an 1150-1200 nm source directly pumped by a laser-diode (LD) [6–11]. The net small-signal gain in a double-clad Yb fiber can be >10 dB around 1180 nm . However, the huge gain ranging over a 1030-1100 nm region creates strong amplified spontaneous emission (ASE) and can lead to parasitic lasing, limiting the utilizable gain and power scaling [6,10]. Thus long-wavelength Yb fiber laser/amplifier has been a challenging work. We proposed employing a high-Q resonator with a small in-operation gain to restrain the ASE gain, and demonstrated 6.5 W non-polarized  and 3W linearly polarized  1178 nm Yb fiber lasers with ~50% slope efficiencies. However, this method is not effective for amplifier. Kurkov et al. reported an 1180 nm Yb fiber laser with an output power of 800 mW, where the Yb fiber was heated to ~200 °C to reduce the ASE gain by increasing reabsorption . Yb fiber laser pumping has also been pursued to spectrally limit the ASE [9,10], whilst it loses the benefit of LD-direct pumping. But, for the higher power operation, a breakthrough to overcome the ASE/parasitic lasing limit has been keenly demanded.
Here, we report, for the first time to our knowledge, high-power, high-efficiency amplification in the 1150-1200 nm region by air-clad, Yb-doped solid core photonic bandgap fibers (Yb-PBGFs). As high as 32 W output power at 1156 nm with a 66% slope efficiency and 42 dB ASE suppression were obtained. Furthermore, by a refined low-loss Yb-PBGF, 30 W non-polarized and 25 W linearly-polarized 1178 nm outputs, with 58% and 56% slope efficiencies respectively, were also achieved without any sign of ASE.
2. Yb-doped PBGFs for 1150-1200 nm operation
Rare-earth-doped PBGF laser has been firstly demonstrated in 2006 . Recently >6 W Bragg-fiber structured PBGF laser was also reported . Those Yb-PBGFs are targeting anomalous dispersion and large mode area, and both lasers operate in the normal spectral region of Yb fibers (1030-1100 nm). On the other hand, the most important advantage of PBGF will be sharp-cut in-fiber, distributed filtering [14,15]. Hence the gain spectrum can be designed artificially, which is desirable for special wavelength operation. In 2008, Pureur et al. reported a zero-line 977 nm Yb-PBGF laser with an output power of 130 mW with the ASE above 1000 nm inhibited successfully . Yb-PBGFs for long-wavelength operation have also been investigated and ASE suppression in the huge-gain region was reported [17,18]. Here, we report a long-wavelength Yb-PBGF operation far beyond the power level of previous long-wavelength Yb fiber lasers as well as PBGF lasers.
In the 1030-1100 nm region or at the zero-line transition, the gain of Yb fibers can be high (typically >10 dB/m) and the impact of fiber loss on the efficiency is moderate. On the other hand, the gain at 1150-1200 nm is much smaller. The emission cross section of an Yb-doped aluminosilicate fiber at 1178 nm is σe~0.01×10-20 cm2, only ~1/60 of that at 1030 nm. With an Yb concentration of 5000 ppmwt (N=0.39×1020 cm-3), considered as the maximum concentration with unserious photodarkening , the small-signal gain is at most Nσe/2~0.8 dB/m (when the pump absorption is fully saturated with 976 nm pumping). As a general theory on energy extraction in gain media, the ratio of small-signal gain to loss should be more than 10 for 40-50% extraction efficiency . Hence, for high-efficiency operation, the loss of the bandgap guidance is required to be much lower than 0.1 dB/m. In non-doped solid-core PBGFs, the attenuation in the 1 μm range can be as low as 0.01 dB/m . However, such low attenuation is very difficult to reach in a PBGF with a doped core, because fluctuations, due to the impact of the doping on the bandgap guidance properties, can significantly increase the confinement loss. Thus, it is a great challenge to achieve high-efficiency PBGF lasers/amplifiers in the 1150-1200 nm region.
The Yb-PBGFs fabricated for this research are similar to that recently reported by Olausson et al. and the details were given in . The cross section of an Yb-PBGF is shown in Fig. 1(a). In the core at the center Yb and fluorine are codoped so that the refractive index is closely matched to silica. The core is surrounded by 9 rings of Ge-doped silica rods arranged in a triangular lattice (pitch Λ) which creates a bandgap allowing for confinement in the core. B-doped rods are introduced on either side of the core to create birefringence (on the order of 10-4), while maintaining confinement via total internal reflection . Here we used two fibers, Yb-PBGF-A and B, with a small difference of dimension and thus bandgap transmission window [Fig. 1(b)]. For Yb-PBGF-A, Λ=9.5 μm and the Ge-rod and B-rod diameters are ~8.3 μm and ~10 μm with the index contrasts of 0.035 and -0.006, respectively. Yb-PBGF-B has just scaled dimensions for Λ=10.1 μm. Yb-PBGF-A was fabricated in 2007 and has a bandgap centered at 1140 nm with a propagation loss of Λ0.1 dB/m across the main window. Lately fabricated Yb-PBGF-B has a bandgap centered at 1210 nm with a lower loss of <0.02 dB/m by refining the process. Here the losses were measured by cut-backing 30-100 m long fibers coiled on a 32-cm diameter dram. The confinement loss outside the bandgaps (stop bands) is too high to be measurable (at least >400 dB/m), realizing perfect inhibition of ASE in the huge-gain region. The mode-field diameter (MFD) is ~10 μm at the signal wavelengths (1156 nm and 1178 nm) for both fibers. Each fiber is designed for cladding pumping with an air-clad diameter of 220 μm and a numerical aperture of 0.57 at 976 nm. The pump absorption is 1.1 dB/m at 976 nm.
The emission cross section of Yb-doped fiber increases steeply as the wavelength shorter (dashed-dot curve in Fig. 2). In order to avoid growth of ASE/parasitic lasing at any wavelength shorter than the 1150-1200 nm region, fine position adjustment of the short-wave cut-off was done by coiling the fiber. The white light transmission of each fiber with different coiling diameters (one turn for each) is shown in Fig. 2. Here we used butt coupling of MFD-matched fibers at both ends to excite/probe only the core mode . Thus, small difference of the transmitted power for different diameters can be seen. By bending the transmission bandwidth narrows and the shorter-wavelength edge red-shifts faster than the blue shift of the longer-wavelength edge, which is a characteristic property in PBGFs . The presence of the B-doped rods can cause some anisotropy in the bending loss, but it was not clear in the measurement due to twist of the fibers. The loss increase in the bandgap peak is not seen even with a 10 cm diameter (see also Fig. 4 (b)). This is partly because the transmission window is the third-order bandgap with low bend susceptibility . To avoid accidental fiber fracture by bend stress, in amplification we used coiling diameters of 10 cm for Yb-PBGF-A. Yb-PBGF-B has a bandgap at a longer wavelength and thus mild coiling by a 26 cm diameter was applied. The calculated effective small-signal gain spectra (dashed curves in Fig. 2) show excellent profiles for long-wavelength operation. The gain of the Yb fiber in the shorter wavelength region, which amounts to >35 dB/m at ~1030 nm with no bandgap structure, is perfectly eliminated by the stop band.
3. High-power Yb-PBGF amplification
The experimental setup of the high-power long-wavelength Yb fiber amplifier is shown in Fig. 3. Non-polarized fiber Raman lasers (FRLs) at 1156 nm and 1178 nm were used for the seed sources. To avoid undesired coupling between the amplifier and seed source in the huge-gain 1030-1100 nm region, the seed was routed by two sharp-cut dichroic mirrors (R<1% for 1000-1100 nm and R>99.8% for 1150-1300 nm). The Yb-PBGF was coiled on a dram with the aforementioned diameter. The length of the fiber section, put straight on each side, was minimized (<30 cm). Both fiber ends were sealed and angle-polished by ~6 deg. The fiber was backward cladding-pumped by a fiber-coupled LD at 976 nm with a launched pump power of ~50 W. The fraction of pump light trapped to the Ge-rods in the clad, which cannot be used for exciting Yb ions in the core and passes through the fiber, was estimated to be ~12%. The amplified output was separated by a dichroic mirror (R>99.8% for 1020-1230 nm) and the core-mode output was measured after a pinhole put in the near-field position.
Yb-PBGF-A was used for amplification at 1156 nm and 1178 nm. The fiber length was 24 m with a 10 cm coiling diameter (~75 turns). The seed powers were ~5 W at 1156 nm and ~3 W at 1178 nm. The output power evolutions as functions of the pump power and the spectra at the maximum output powers are shown in Fig. 4. In 1156 nm amplification, the maximum output power was 32 W at a 53.8 W pump power. This is the record high output power from PBGFs, as well as from long-wavelength Yb fiber lasers. More marked is the high slope efficiency (66%). Considering the pump trapping to the Ge-rods, the net slope efficiency is estimated to be as high as 75%. ASE cannot be seen in the huge-gain 1030-1100 nm region but weakly exists around 1120 nm, as can be expected from the effective gain spectrum [Fig. 2 (a)]. Because of the high ASE suppression (42 dB), higher output power can be expected only by using higher-power LD.
On the other hand, in 1178 nm amplification, the ASE around 1120 nm grew rapidly and parasitic lasing started at ~30 W pump power. Hence the maximum output power was limited to 9.1 W with the slope efficiency of ~33% and 25 dB ASE suppression. Those poorer properties than at 1156 nm clearly indicate the difficulty of the long-wavelength operation. In addition to the low efficiency predicted in Section 2, due to the high saturation power (1.6 W) at 1178 nm, the upper-state population of Yb ions is not depleted enough and ASE experiences high gain. At 1156 nm, the twice larger emission cross section (σe~0.02×10-20 cm2) dramatically changes the behavior. The 0.1 dB/m propagation loss of Yb-PBGF-A is low enough for high efficiency operation. Owing to the low saturation power (0.8 W), depletion of the upper-state population occurs well and limits the gain for the ASE.
We calculated the fiber-length and loss dependences of the available output power by using a rate-equation model . Here the Yb concentration of 0.38×1020 cm-3 and signal-core overlap factor of 0.84 (calculated by a mode solver based on the finite element method) are used. It is assumed that the pump light, except for the trapped fraction (12%), is absorbed homogenously. For simplicity, the seed power is put as the same (4 W). As shown in Fig. 5 (a), while the 24 m fiber length is almost optimum for 1156 nm operation, a longer fiber is required for 1178 nm operation, because the pump absorption is saturated due to weak depletion of the upper-state population by the signal. This interplay is more serious with higher pump power. The extraction efficiency, defined as the signal power increase divided by the absorbed pump power, is more susceptible to the propagation loss for 1178 nm operation than 1156 nm [Fig. 5 (b)], as predicted. Thus, a lower loss, longer fiber is required as the signal wavelength is longer.
The performance of the 1178 nm amplification was dramatically improved by Yb-PBGF-B. According to the calculation, a longer fiber (37 m length) was used. The coiling diameter was 26 cm (~44 turns) and the seed power was ~4 W at 1178 nm. Figure 6 shows the result. The maximum output power was 29.7 W at a 49.5 W pump power, and the slope efficiency amounted up to 58% (66% if pump trapping is considered). The extraction efficiency is estimated to be 59%. No sign of ASE can be seen in the output spectrum, indicating the shape of the gain spectrum is appropriate for 1178 nm operation. The near-field beam profile of the amplifier output is shown in the inset of Fig. 6 (b). The beam quality was measured to be M2<1.1.
Polarization-maintaining amplification was also successfully demonstrated. By incorporating a polarizer before seeding the Yb-PBGF (see Fig. 3), a linearly-polarized seed was prepared with a power of 2 W. The polarization direction was adjusted to match with the fast axis of the fiber (across the B-doped rods). The maximum output power of 25.3 W was obtained. The 56% slope efficiency is similar to that for non-polarization amplification (58%) in spite that the seed power was halved, indicating fully saturated amplification. The polarization extinction ratio was measured to be >13 dB at an 8 W output power, and can be kept even with stress applied to the fiber by fingers. This is suitable enough for frequency conversion to the yellow-orange spectral range.
We demonstrated LD directly pumped, high-power, high efficiency Yb fiber amplifiers operating in the long-wavelength edge (1150-1200 nm) by the advanced PBGF technology. The bandgap guidance in the Yb-doped core is formed in the long-wavelength edge and the ASE growth in the shorter wavelength region is inhibited. A 32 W output at 1156 nm with a 66% slope efficiency and 30 W output at 1178 nm with a 58% slope efficiency were successfully obtained. To our knowledge, these are the highest output powers generating from active PBGFs, as well as from Yb-doped fiber lasers at these wavelengths. The ASE was highly suppressed so that the output was limited only by the pump power and thus further power scaling can be expected. The ultralow attenuation property (<0.02 dB/m) of the Yb-PBGF is the reason for the marked high-efficiency operation even at the low-gain wavelengths.
The presented power-scalable long-wavelength Yb-fiber source will be useful for various applications, especially for laser guide star (LGS). At present the main architecture for 589 nm LGS sources is sum frequency generation of two Nd:YAG lasers at 1064 nm and 1319 nm . However, the near-future large-aperture telescope will require multiple diffraction-limited 589 nm sources with >50 W output powers in CW or quasi-CW operation . A fiber-based source will be much preferable for such high-average power operation as well as for number scaling with equivalent beam quality. The Yb-PBGF will be a promising solution for the next-generation LGS.
The authors thank Y. Hayano for the fruitful discussion on LGS.
References and links
1. N. Saito, K. Akagawa, M. Ito, A. Takazawa, Y. Hayano, Y. Saito, M. Ito, H. Takami, M. lye, and S. Wada, “Sodium D2 resonance radiation in single-pass sum-frequency generation with actively mode-locked NdYAG lasers,” Opt. Lett . 32, 1965–1967 (2007). [CrossRef] [PubMed]
2. Y. Feng, S. Huang, A. Shirakawa, and K. Ueda, “589 nm light source based on Raman fiber laser,” Jpn. J. Appl. Phys . 43, L722–724 (2004). [CrossRef]
3. D. Georgiev, V. P. Gapontsev, A. G Dronov, M. Y Vyatkin, A. B. Rulkov, S. V. Popov, and J. R. Taylor, “Watts-level frequency doubling of a narrow line linearly polarized Raman fiber laser to 589 nm,” Opt. Express 13, 6772–6776 (2005). [CrossRef] [PubMed]
4. E. M. Dianov, A. V. Shubin, M. A. Melkumov, O. I. Medvedkov, and I. A. Bufetov, “High-power cw bismuth-fiber lasers,” J. Opt. Soc. Am . B 24, 1749–1755 (2007). [CrossRef]
5. A. B. Rulkov, A. A. Ferin, S. V. Popov, J. R. Taylor, I. Razdobreev, L. Bigot, and G. Bouwmans, “Narrow-line, 1178nm CW bismuth-doped fiber laser with 6.4W output for direct frequency doubling,” Opt. Express 15, 5473–5476 (2007). [CrossRef] [PubMed]
6. J. Ota, A. Shirakawa, and K. Ueda, “High-power Yb-doped double-clad fiber laser directly operating at 1178nm,” Jpn. J. Appl. Phys . 45, L117–L119 (2006). [CrossRef]
7. M. Adachi, K. Kojima, T. Yoda, and K. Hayashi, “CW orange beam generation by 1160-nm fiber laser system,” presented at 53rd Spring Meeting of the Japan Society of Applied Physics and Related Societies, Tokyo, Japan , 22–26 Mar. 2006. [PubMed]
8. S. Sinha, C. Langrock, M. J. F. Digonnet, M. M. Fejer, and R. L. Byer, “Efficient yellow-light generation by frequency doubling a narrow-linewidth 1150 nm ytterbium fiber oscillator,” Opt. Lett . 31, 347–349 (2006). [CrossRef] [PubMed]
9. A. S. Kurkov, V. M. Paramonov, and O. I. Medvedkov, “Ytterbium fiber emitting at 1160 nm,” Laser Phys. Lett . 3, 503–506 (2006). [CrossRef]
10. A. S. Kurkov, “Oscillation spectral range of Yb-doped fiber lasers,” Laser Phys. Lett . 4, 93–102 (2007). [CrossRef]
11. H. Maruyama, A. Shirakawa, and K. Ueda, “1178nm linearly-polarized all fiber laser,” in Conference on Lasers and Electro-Optics/Pacific Rim 2007, Technical Digest (CD) (Optical Society of America, 2007), paper TuA4. [CrossRef]
13. S. Février, D. D. Gaponov, P. Roy, M. E. Likhachev, S. L. Semjonov, M. M. Bubnov, E. M. Dianov, M. Y. Yashkov, V. F. Khopin, M. Y. Salganskii, and A. N. Guryanov, “High-power photonic-bandgap fiber laser,” Opt. Lett . 33, 989–991 (2008). [CrossRef] [PubMed]
15. T. Taru, J. Hou, and J. C. Knight, “Raman gain suppression in all-solid photonic bandgap fiber,” in European Conference and Exhibition on Optical Communication 2007, Berlin, Sep. 2007, paper 7.1.1. [CrossRef]
16. V. Pureur, L. Bigot, G. Bouwmans, Y. Quiquempois, M. Douay, and Y. Jaouen, “Ytterbium-doped solid core photonic bandgap fiber for laser operation around 980 nm,” Appl. Phys. Lett . 92, 061113 (2008). [CrossRef]
17. C. B. Olausson, C. I. Falk, J. K. Lyngsø, B. B. Jensen, K. T. Therkildsen, J. W. Thomsen, K. P. Hansen, A. Bjarklev, and J. Broeng, “Amplification and ASE suppression in a polarization-maintaining ytterbium-doped all solid photonic bandgap fibre,” Opt. Express 16, 13657–13662 (2008). [CrossRef] [PubMed]
18. R. Goto, K. Takenaga, K. Okada, M. Kashiwagi, T. Kitabayashi, S. Tanigawa, K. Shima, S. Matsuo, and K. Himeno, “Cladding-pumped Yb-doped solid photonic bandgap fiber for ASE suppression in shorter wavelength region,” in Optical Fiber Communication Conference 2008, Technical Digest (CD) (Optical Society of America, 2008), paper OTuJ5. [CrossRef]
19. J. Koponen, M. Söderlund, H. J. Hoffman, D. Kliner, and J. Koplow, “Photodarkening measurements in large-mode-area fibers,” Proc. SPIE 6453, 64531E (2007). [CrossRef]
20. K. Ueda, “The prospects of high power fiber lasers,” Rev. Laser Eng . 29, 79–83 (2001).
21. J. K. Lyngsø, B. J. Mangan, and P. J. Roberts, “Polarization maintaining hybrid TIR / bandgap all-solid photonic crystal fiber,” in Conference on Lasers and Electro-Optics 2008, Technical Digest (CD) (Optical Society of America, 2008), paper CThV1.
23. A. Hardy and R. Oron, “Signal amplification in strongly pumped fiber amplifiers,” IEEE J. Quantum Electron . 33, 307–313 (1997). [CrossRef]
24. C. Boyer, B. Ellerbroek, M. Gedig, E. Hileman, R. Joyce, and M. Liang, “Update on the TMT laser guide star facility design,” Proc. SPIE 7015, 70152N (2008). [CrossRef]