Phosphorus co-doping is known to reduce clustering levels of rare earth ions in silica hosts. In this paper, ytterbium-doped silica fibers with ~8.9wt% Yb2O3, up to ~4700dB/m peak core absorption at 976nm, and low photo-darkening are demonstrated using high phosphorus co-doping. Measured gain as high as ~7dB/cm is demonstrated in the fiber.
©2009 Optical Society of America
Fiber lasers have been a great commercial success in the last few years in a wide range of applications with average powers from a few watts to multi-kilowatts. At high average powers, the fiber geometry with its large heat-dissipating surface very close to a small active core provides diffraction-limited output and almost total elimination of the thermal issues typically related to solid state lasers. It is, however, still difficult for fiber lasers to provide the high peak powers and high pulse energies required in many material processing applications due to their low nonlinear thresholds comparing to those in solid state lasers.
In recent years, significant effort has been directed towards finding fiber approaches which support single mode operation and large effective mode area at the same time [1–7]. Most nonlinear thresholds, with exception of nonlinear self-focusing, scale with the effective mode area. Solutions are sought to overcome the fundamental conflicts between large effective mode area and single mode operation. Significant progress has been made in this area. On the other hand, most nonlinear thresholds also scale inversely with effective nonlinear length. A shorter fiber amplifier with the same gain is better for producing high peak power pulses. One way to achieve this is by reducing the ratio of pump to doped-core area in double-clad fibers to increase pump absorption [2,3]. This can indeed help as long as the maximum gain available from the rare earth doped glass at the desired wavelength is not reached. This approach, however, requires higher brightness pumps. The higher NA pump guides made from air-hole-cladding ease this brightness requirement somewhat [2,3]. In addition to pump guide design, further increase of rare earth doping levels will be able to provide shorter fiber amplifiers regardless of fiber designs.
The limits to higher rare earth doping level in glass are ion clustering and eventual phase separation. Clustering can lead to cooperative up-conversion, which not only creates channels for energy loss but also produces more energetic photons which can cause secondary effects such as photo-darkening through color center formation in the glass. Cooperative up-conversion can be a significant problem for most rare earth ions with multiple f-shell transitions. Doping levels of rare earth ions such as Er3+ ions are typically limited to below 1000 mol ppm in silica fibers. Yb3+ ions, on the other hand, has a simple two-level f-shell transition and can be doped up to few thousands mol ppm. Up-conversion in Yb3+ ions, however, can happen through a virtual state . Even though energy loss through this process is not typically significant in ytterbium-doped fiber lasers, however, if color centers are produced by the few energetic photons, they can lead to an increase of background loss at the lasing wavelength, which, in turn, leads to a degradation of laser performance over time. Photo-darkening effect in ytterbium-doped fibers has been studied extensively in the last few years [9,10].
It is well known that rare earth ions have much higher solubility in phosphate glass [11,12]. Yb2O3 has been doped to 12wt% recently in phosphate glass, and the glass was with 9032dB/m absorption at 976nm without significant signs of photo-darkening. Unfortunately, phosphate glass can only be made with crucibles, which typically introduce very high levels of impurities. This not only leads to higher background losses, but also lowers damage thresholds; both are significant problems for high power fiber lasers. The high background losses also limit phosphate fibers to be used in short lengths. This further limits them to low power fiber lasers with low overall heat load that can be dissipated in a short fiber. In addition, the low transition temperature of ~500°C for phosphate glass, comparing to ~1200°C for silica, also makes them less reliable for high power lasers, where a much more elevated temperature is frequently expected in the active core. A glass host, which is capable of having high doping concentrations, low impurity levels, and compatibility to silica fibers to enable the use of mature technologies developed for telecommunications, is, therefore, highly desirable.
Recently Yb2O3 of up to 11wt% was doped in phosphosilicate fibers in . Photo-darkening loss of ~30dB/m at 500nm was measured in a phosphosilicate fiber with 1.3wt% Yb3+, i.e. ~1.5wt% Yb2O3 in . Aluminum and phosphorus doped silica fibers with Yb2O3 of 0.45 mol%, i.e. ~3wt% Yb2O3, was demonstrated in . Photo-darkening loss of ~15dB/m at 633nm was measured in this fiber . We will demonstrate, in this work, highly ytterbium-doped silica fibers with phosphorus and aluminum co-doping, which were fabricated with an optimized vapor deposition process for high phosphorus doping. Efficient and reliable ytterbium-doped fibers with ~8.9wt% Yb2O3, peak core absorption of ~4700dB/m at 976 nm are demonstrated with saturated photo-darkening loss of ~36dB/m at 675nm, corresponding to 1.8–3.6dB/m at 1.05µm, at ~50% inversion. In fibers with ~3.7wt% Yb2O3, peak ytterbium absorption of ~3200dB/m at 976 nm, and negligible saturated photo-darkening loss of ~0.8dB/m at 675nm, corresponding to 0.04–0.08dB/m at 1.05µm is demonstrated. Low background loss of less than 0.05dB/m at 1.3µm is regularly obtained in all fibers. These fibers can also be doped with a high level of boron. This allows refractive indexes close to that of silica to be achieved even at very high rare earth doping levels, a critical requirement for most large core fiber designs.
2. General characteristics of fabricated fibers
A large number of fibers have been fabricated for this study. The fabrication process has been continuously improved for maximum phosphorus incorporation in silica glasses over the last few years in order to improve solubility of rare earth ions. The method is IMRA proprietary and cannot be disclosed in detail. Aluminum was also added to further improve rare earth solubility. We have also developed a process to incorporate boron in these glasses to enable additional control of refractive index of the doped glass to be very close to or below that of the silica. Boron is incorporated in an additional process separately from that of phosphorus incorporation due to a reaction of the precursors at room temperature. This ability to make highly rare earth doped silica glasses with refractive index close to that of the silica is very important for implementing large core designs. Many of the fabricated fibers have been used in amplifiers and lasers, including the two highly doped fibers studied in this paper, providing demonstrated efficient and reliable operation over long time.
Typical normalized absorption and emission cross sections of ytterbium-doped aluminosilicate and ytterbium-doped phosphosilicate fibers were first characterized with fibers fabricated at IMRA. They are shown in Fig. 1(a). They were obtained by curve fitting to the measured data as described in . Both absorption and emission of phosphosilicate fibers are similar to that of ytterbium-doped phosphate glass. This implies that most ytterbium ions are near phosphorus sites. Similar absorption and emission cross sections of ytterbium-doped phosphosilicate fibers have also been shown in . The absorption of ytterbium-doped phosphosilicate fibers typically has three peaks instead of two in ytterbium-doped aluminosilicate fibers. Details of characteristics of Yb3+ ions in silica fibers were first studied in [15,16]. The emission peaks of ytterbium-doped phosphosilicate fibers are at shorter wavelengths comparing to that of ytterbium-doped aluminosilicate fibers, due to a narrower Stark split, which will be discussed later. Lifetime measurements in a commercial ytterbium-doped aluminosilicate fiber and one of our own ytterbium-doped phosphosilicate fibers are shown in Fig. 1(b), giving excited state lifetimes of 0.6ms and 1.35ms respectively. These results are consistent with what are in literatures, considering 20% or so variations are typically expected depending on compositions and fabrication conditions.
Energy diagram of Yb3+ ions in glasses is given in Fig. 2(a). Net cross sections at various inversion levels, normalized against that at the gain peak around 976nm at 100% inversion, are plotted in Fig. 2(b) for ytterbium-doped aluminosilicate fiber and Fig. 2(c) for ytterbium-doped phosphosilicate fiber. These figures were obtained through similar curve fitting routine described in . The gain peak in the phosphosilicate fibers around 1010–1030nm has relatively higher gain and is broader than that in the aluminosilicate fibers, ideal for amplification of ultra short pulses with broad spectral width. The two emission peaks in the wavelength from 1000nm to 1150nm resulting from transitions e→b and e→c move significantly closer to the shorter wavelength in the phosphosilicate fiber. The energy levels of Yb3+ ions in the fibers can also be obtained from curve fittings. These are given in table 1 for the aluminosilicate and the phosphosilicate fibers, respectively. It can be clearly seen that the Stark level split is noticeably smaller in the phosphosilicate fibers, resulting in emission wavelength moving towards shorter wavelengths.
Two fibers, fiber 1 and fiber 2, with peak ytterbium absorption of ~3200dB/m and ~4700dB/m respectively are studied in this paper. The overlap between the dopant and the guided fundamental mode has been factored out, so the absorption indicates core glass absorption that corresponds to that for 100% dopant and mode overlap. This convention is used throughout this paper unless stated otherwise. Composition at the core center of the preforms of the two fibers was measured by an Energy Dispersion Spectroscopy (EDS) setup on a scanning electron microscope. Preforms were used to get a large sample area for better precision. The composition measurements have an estimated error of ±1.5%. The core center of fiber 1 with ~3200dB/m absorption has 3.7wt% Yb2O3, 14.8wt% P2O5, 5.9wt% Al2O3, and 75.6wt% SiO2. The core center of fiber 2 with ~4700dB/m absorption has 8.9wt% Yb2O3, 22.4wt% P2O5, 11.5wt% Al2O3, and 57.2wt% SiO2. The refractive index profiles of the two fibers are shown in Fig. 3(a). Measured absorption spectra are shown in Fig. 3(b) for fibers 1 and 2. Fiber 1 has a second mode cutoff wavelength of ~0.7µm from simulation using a mode solver and the measured refractive index profile, and a measured background loss of ~40dB/km at 1.3µm (not factoring out spatial overlap between guided mode and core). Fiber 2 has a second mode cutoff wavelength of ~1µm and a measured background loss of ~35dB/km at 1.3µm (not factoring out spatial overlap between guided mode and core). Ytterbium excited state lifetimes were also measured for the two fibers. They are shown in Fig. 4 and well fitted to exponential decays. The measured lifetimes are 1.1ms and 0.75ms for the fibers 1 and 2 respectively. Fiber 1 has relatively higher P2O5 content with wt% ratio of P2O5 to Al2O3 of ~2.5. Both its absorption spectrum and lifetime are closer to that of phosphosilicate fibers. Fiber 2 has a relatively lower wt% ratio of P2O5 to Al2O3 of ~1.9. Both its absorption spectrum and lifetime are closer to that of aluminosilicate fibers.
Two sets of additional ytterbium-doped fibers with a constant phosphorus content of ~15wt% P2O5 and ~4wt% Yb2O3, were fabricated. The first set has no boron, but has various levels of aluminum up to ~10wt% Al2O3. The second set has a fixed level of boron (amount was not characterized because EDS cannot detect boron easily), but has a varying level of aluminum up to ~10wt% Al2O3. Photo-darkening in all the fibers was characterized. It was found that the level of aluminum does not influence photo-darkening within the experimental accuracy. From the same set of data, it was also concluded that boron doping can help to reduce photo-darkening by a factor of ~2 in dB/m. The two fibers studied in this paper do not contain boron.
3. Photo-darkening measurements
A large number of fibers with various compositions, fabrication conditions and peak ytterbium absorption ranging from 1500–5000dB/m, were tested for photo-darkening with a 675nm LED with the continuous core-pumping setup shown in Fig. 5. A short fiber length of less than 30mm was typically spliced at both ends in the test setup. A commercially available arc fusion splicer was used to splice the ytterbium-doped fibers and standard silica fibers used in the test setup. Splice losses of the fibers were typically less than 0.3dB for each joint if the fiber core sizes matched with each other and appropriate splicing conditions were used. Two detectors were used to compensate fluctuations of the output of the 675nm LED. Appropriate filters were used before the signal detectors to isolate any unwanted pump powers for continuous monitoring. An ytterbium amplifier simulator  was used first to calculate 975nm pump power for a saturated inversion of ~50%. Since photo-darkening loss is very sensitive to inversion levels [9,10], sufficient pump power was always used to ensure inversion saturation for good repeatability. Results from two fibers are reported here.
The measured photo-darkening losses are shown in Fig. 6(a). For comparison, a commercial ytterbium-doped single mode fiber with a peak core absorption of ~380dB/m (Nufern) was tested and the result is also shown in Fig. 6. The overlap between the doped core and the guided fundamental mode for 675nm has been also factored out for the losses shown in Fig. 6(a), representing a dopant and mode spatial overlap of 100%. Photo-darkening loss usually takes few hours to reach saturation. The saturated photo-darkening losses at 675nm are 0.8dB/m, 0.8dB/m and 36dB/m for the commercial fiber, fiber 1 and fiber 2 respectively. Despite the phosphosilicate fiber with ~3200dB/m peak absorption having near an order of magnitude higher ytterbium concentration, it has a photo-darkening loss level similar to that of the commercial ~380dB/m fiber. Photo-darkening spectra were also measured in the two phosphosilicate fibers and are shown in Fig. 6(b), normalized against the loss at 675nm. The long wavelength loss was not well resolved for the fiber 1. The loss ratio at wavelengths of 675nm and 1.05µm is estimated to be 10–20 using the measured data. Saturated photo-darkening losses at 1.05µm of 0.04–0.08dB/m and 1.8–3.6dB/m respectively were estimated for the two phosphosilicate fibers 1 and 2, representing the lowest photo-darkening loss at these doping levels ever reported in silica fibers. This saturated photo-darkening loss represents the highest possible loss in these fibers by a 975nm pump, which can produce a maximum inversion of ~50%. Since photo-darkening loss is a strong function of inversion and inversion in a high power double clad fiber is much lower, a much lower photo-darkening loss is expected in high power fiber lasers and amplifiers for the two phosphosilicate fibers. Photo-darkening loss dynamics for pump on and off was also measured for fiber 1 and is shown in Fig. 7. The photo-darkening loss recovered by ~50% after pump was switched off, and increased again at slightly higher rate for the second pumping cycle. Photo-darkening loss dynamics was also studied in phosphorus and aluminum co-doped silica host in .
4. Gain and amplifier measurements
Gain was also measured for the two phosphosilicate fibers with a setup similar to that in Fig. 5. An ytterbium fiber ASE source and an OSA were used instead of the LED and detector. Transmission loss was first measured by cutting back. Net gain could then obtained by subtracting out the measured loss from the measured gain. Gains of amplifiers made from 3cm long fiber 1 and 2cm long fiber 2, seeded by the ytterbium ASE source peaked at ~1030nm, were monitored for 44 and 136 hours respectively from the first pump-on at ~50% inversion (see Fig. 8), showing high unit gain levels approaching that of ytterbium crystals of ~5.3dB/cm and ~7dB/cm respectively with negligible degradation over the test periods (the overlap was factored out). This type of performance was also confirmed by numerous oscillator and amplifier tests using phosphosilicate core glass in fibers with core diameters ranging from 5µm to 80µm [4,18,19].
To further characterize amplifier performance, fiber amplifiers were built with the two fibers with counter-pumping configuration. Lengths used were 16cm and 9cm for fibers 1 and 2 respectively. Pump wavelength was at 975nm. Input signal was from an ytterbium ASE source with 12mW of average power. The peak wavelengths at the output of the amplifier were measured to be 1025nm and 1026nm for fibers 1 and 2 respectively. Measured slope efficiencies were 84% and 66% for fibers 1 and 2 respectively (see Fig. 9). Splice loss is factored out. Residue pump power is, however, not accounted for.
5. Discussions and conclusions
One of our recent works has used a short piece of an ytterbium-doped phosphosilicate fiber with ~3600dB/m absorption at 976nm in an efficient mode-locked oscillators and another 0.5m long fiber amplifier with a pump cladding diameter of 125µm and a 0.07NA and 15µm core with ~2400dB/m absorption at 976nm to achieve a repetition rate of ~1GHz . Frequency-stabilized combs were demonstrated without further amplification. These wide-spaced frequency-stabilized combs are required for frequency references and high resolution spectroscopic measurements. Ytterbium-doped cores with refractive index closely matched to that of silica glass are also used in our active leakage channel fiber demonstrations in [4,19].
To summarize, using phosphorus co-doping we have demonstrated reliable and efficient highly ytterbium-doped silica fibers with low photo-darkening. These fibers are expected to further increase peak power performance of fiber lasers and amplifiers for all fiber designs, and provide more robust and reliable alternative to ytterbium-doped phosphate glass fibers for single frequency lasers.
References and links
1. M. E. Fermann, “Single-mode excitation of multimode fibers with ultrashort pulses,” Opt. Lett. 23, 52–54 (1998). [CrossRef]
2. 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]
3. C. D. Brooks and F. Di Teodoro, “Multi-megawatt peak power, single-transverse-mode operation of a 100µm core diameter, Yb-doped rod-like photonic crystal fiber amplifier,” Appl. Phys. Lett. 89, 111119–111121 (2006). [CrossRef]
4. L. Dong, T. W. Wu, H. A. McKay, L. B. Fu, J. Li, and H. G. Winful, “All-Glass Large-Core Leakage Channel Fibers,” IEEE J. Sel. Top. Quantum Electron. 15, 47–53 (2009). [CrossRef]
5. S. Ramachandran, J. W. Nicholson, S. Ghalmi, M. F. Yan, P. Wisk, E. Monberg, and F. V. Dimarcello, “Light propagation with ultra large modal areas in optical fibers,” Opt. Lett. 31, 1797–1799 (2006). [CrossRef] [PubMed]
6. F. Di. Teodoro, J. P. Koplow, S. W. Moore, and D. A. V. Kliner, “Diffraction-limited, 300kW peak-power pulses from a coiled multimode amplifier,” Opt. Lett. 27, 518–520 (2002). [CrossRef]
7. R. Sims, V. Sudesh, T. McComb, Y. Chen, M. Bass, M. Richardson, A.G. James, J. Ballato, and A.E. Siegman, “Diode-pumped very large core, gain guided, index anti-guided single mode fiber laser,” Advanced Solid State Photonics, paper WB3, Denver, February 2009.
8. A. V. Kir’yanov, Y. O. Barmenkov, I. L. Martinex, A. S. Kurkov, and E. M. Dianov, “Cooperative luminescence and absorption in ytterbium-doped silica fiber and the fiber nonlinear transmission coefficient at 980nm with a regard to the ytterbium ion-pairs’ effect,” Opt. Express 14, 3981–3992 (2006). [CrossRef] [PubMed]
10. S. Jetschke, S. Unger, U. Röpke, and J. Kirchhof, “Photodarkening in Yb doped fibers: experimental evidence of equilibrium states depending on the pump power,” Opt. Express 15, 14838–14843 (2007). [CrossRef] [PubMed]
11. Y. W. Lee, S. Sinha, M. J. F. Digonnet, R. L. Byer, and S. Jiang, “Measurements of high photodarkening resistance in heavily Yb3+doped phosphate fibers,” Electron. Lett. 44, 14–16 (2008). [CrossRef]
13. A. V. Shubin, M. V. Yashkov, M. A. Melkumov, S. A. Smirnov, I. A. Bufetov, and E. M. Dianov, “Photodarkening of aluminosilicate and phoshosilicate Yb-doped fibers,” CLEO Europe, paper CJ5, 2007.
14. S. Jetschke, S. Unger, A. Schwuchow, M. Leich, and J. Kirchhof, “Efficient Yb laser fibers with low photodarkening by optimization of the core composition,” Opt. Express. 16, 15540–15545 (2008). [CrossRef] [PubMed]
15. H. M. Pask, R. J. Carman, D. C. Hanna, A. C. Tropper, C. J. Mackechnie, P. R. Barber, and J. M. Dawes, “Ytterbium-doped silica fiber lasers: versatile sources for the 1–,1.2µm region,” IEEE J. Sel. Top. Quantum. Electron. 1, 2–13 (1995). [CrossRef]
16. R. Paschotta, J. Nilsson, A. C. Tropper, and D. C. Hanna, “Ytterbium-doped fiber amplifiers,” IEEE J. Quantum. Electron. 33, 1049–1056 (1997). [CrossRef]
17. X. Peng and L. Dong, “Temperature dependence of ytterbium-doped fiber amplifier,” J. Opt. Soc. Am. B , 25, 126–130 (2008). [CrossRef]
18. I. Hartl, H. A. McKay, R. Thapa, B. K. Thoimas, A. Ruehl, L. Dong, and M. E. Fermann, “Fully stabilized GHz Yb-fiber laser frequency comb,” Advanced Solid State Photonics, post-deadline paper MF9, Denver, February 2009.
19. L. B. Fu, H. A. McKay, S. Suzuki, M. Ohta, and L. Dong, “All-Glass PM Leakage Channel Fibers with up to 80µm Core Diameters for High Gain and High Peak Power Fiber Amplifiers,” Advanced Solid State Photonics, post-deadline paper MF3, Denver, February 2009.