Open Access
Add to BibSonomy
Abstract
Rare earth co-doping of erbium with optically inactive ytterbium or lanthanum ions was investigated in silica fibers for resonant pumping of amplifiers and lasers. The co-dopants are optically inactive when pumped at 1475 nm and therefore their main role is to reduce erbium clustering in solution-doped fibers. We have demonstrated increased efficiencies compared to fibers where no co-dopants are used. With this technique, a single mode Er-La fiber with an Er ion concentration of 0.98x1025 ions/m3, exhibited a slope efficiency of 79.6% in a core-pumped master oscillator-power amplifier (MOPA) with a signal wavelength of 1560 nm.
© 2019 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
Erbium doped fiber lasers are attractive candidates for high energy lasers (HELs) because they operate at wavelengths that are both safer to the eye and in a high atmospheric transmission window [1]. However, a significant issue for erbium doped fibers occurs when detrimental clustering effects cause quenching. This includes both cooperative upconversion and pair-induced quenching. These processes reduce efficiency for fibers with the high concentrations of Er that are needed for clad-pumped HELs. We have investigated the use of codoping with Er and rare earth ions such as La and Yb (RE:Er) where the rare earth ion co-dopants are used to help disperse the Er ions in order to avoid these clustering effects. This technique has been used previously with success [2–6], but specific fiber design parameters and the optimal concentration of RE co-dopant, and other dopants in order to increase Er efficiency have not been fully discussed.
Aluminum is added to silica fibers to increase Er solubility and to reduce clustering. Upon introduction into the silica lattice, the Al ion generally resides in a tetrahedral [AlO4/2]- configuration at concentrations typical for RE doped fibers [7]. Rare earth ions tend to cluster near these sites due to charge compensation. Therefore, Er co-doping with another RE species can help reduce the probability of Er pairs, as each species competes for proximity to the Al sites. This technique may also enable us to reduce aluminum concentration in the fibers, which is beneficial since Al2O3 has a high refractive index and thus increases the core NA. Furthermore, the core glass will devitrify at high Al concentrations.
Both La and Yb ions are optically inactive and take no part in lasing for resonant pumping directly into the Er 4I13/2 manifold and lasing of erbium at ~1500-1600 nm, in contrast to pumping at 980 nm into the Er 4I11/2 level where Yb competes with the Er ions. Therefore, these species are good candidates for use as passive “spacers” to separate Er ions. In this work the concentrations of precursors used for the synthesis and the ratios of RE:Er and Al/Er in solution-doped fibers were optimized in order to increase slope efficiency.
2. Experiment
All fibers in this work were prepared by solution doping, where RE and Al chlorides were dissolved in methanol, filtered and doped into the porous silica soot preform cores in situ on the MCVD lathe. Precursor concentration ratios of RE:Er were investigated from 2:1 to 8:1, and two ratios of Al/Er were used, 50/1 and 75/1, where the highest ratio of aluminum corresponds to 0.55 M in the precursor solution. As a control two fibers were also produced with Al/Er = 75/1 but no RE co-dopant, and 2 commercial Coherent-Nufern fibers were also investigated.
The refractive index profiles of the preforms were measured with a Photon Kinetics P104 Preform Analyzer, and that of the fibers with an Interfiber Analysis IFA100 Optical Fiber Analyzer. Background optical absorption measurements were made by the cutback method with a Photon Kinetics 2300 Fiber Analysis System. Erbium core absorption at 1532 nm was measured by the cutback method using an Agilent 83437A EELED source and an Ando 6315 optical spectrum analyzer (OSA). Core Er absorption serves as a good measure of Er concentration (and indeed, commercial fiber is sold by core absorption values), and we will refer to Er core absorption in dB/m throughout the paper. The actual concentrations of Er, Al, La and Yb were also obtained using electron probe microanalysis (EPMA).
Amplifier performance was evaluated with a master oscillator-power amplifier (MOPA) using a 1475 nm pump laser diode with powers up to 300 mW, and a 1560 nm signal laser diode amplified with a Thorlabs EDFA100S erbium doped fiber amplifier. The slope efficiency with respect to absorbed pump power was measured for increasing pump power up to 300 mW and increasing signal input powers from 12mW to 76mW until near saturation. A Semrock FF01-1535-LP/25 sharp cutoff long pass filter on the output separated the amplified signal from the unabsorbed pump. Slope efficiency was measured at each fiber length as the fiber was cut back, and the initial injected pump and signal powers were measured by cleaving the Er-doped fiber as close as possible to the splice with the single mode fiber (SMF). Amplifier gain spectrum measurements over the C band were performed using an Agilent 81689A tunable laser (for 0.6 mW signal) and the Thorlabs EDFA for 12 and 76 mW to obtain increased input signal power. The pump power was 300 mW at 1475 nm for these experiments. A LabVIEW program was created to control the MOPA and gain experiments.
3. Results and discussion
Table 1 shows the properties of the fibers studied. Note that there are two groups with Al/Er ratios of 50/1 and 75/1, and ratios of co-dopant to erbium, RE:Er of 4:1 and 8:1 with one additional fiber with RE:Er = 2:1. There are two groups of Er concentrations ~1 – 1.6 x 1025/m3, corresponding to core absorptions of ~20-28 dB/m, and ~3 x 1025/m3, corresponding to core absorptions of ~52-65 dB/m.
Table 1. Er core absorption, Er and Al concentration, co-dopant and Al/Er ratio, NA and slope efficiency with 12 mW input signal power for the fibers of this study.
In Fig. 1 we have plotted five absorption spectra for three different fibers. This includes two La:Er co-doped fibers, one with a La:Er ratio of 4:1 and an Er core absorption of 60.4 dB/m and one with an La:Er ratio of 8:1, and an Er core absorption of 21.3 dB/m. There is one Yb:Er fiber with a ratio of Yb:Er of 4:1, and a core absorption of 54.6 dB/m. The absorption spectra for these fibers are plotted along with two Er fibers having no rare earth co-dopant: one with a core absorption of 20.3 dB/m and one with a core absorption of 58.5 dB/m. As discussed, all fibers in this figure have aluminum in them to solubilize the RE ions, and all fibers had an Al/Er ratio of 75/1, which represents the maximum ratio for the fibers in this work. The corresponding aluminum concentrations in mol % are listed in Table 1. The absorption has been normalized to the peak value for comparison. The black curve is the Er fiber with no RE co-dopant and a core absorption 20.3 dB/m and an Al2O3 concentration of 1.806 mol %. The width of the absorption spectrum is the smallest of the fibers presented. The La:Er 8:1 fiber with a similar Er core absorption (21.3 dB/m) and Al concentration (1.691 mol %) has resulted in a slight increase in the width of the absorption spectrum (red curve). At the highest Er core absorptions we compare the La:Er 4:1 (red curve), Yb:Er 4:1 (blue curve), and an Er fiber with no RE co-dopant (black curve). Here a concomitant increase in Al concentration was given to the fibers in order to maintain an Al/Er ratio of 75/1. The Al2O3 molar concentrations in these fibers are greater than 4% for these fibers (Table 1). It is evident that the width of the absorption spectra for all fibers has increased but there is essentially no difference in the width among them. It is well known that aluminum doped silica fiber is effective in broadening both the absorption and gain spectra [7] for silica fibers, and that effect is evident here, where it impacts the spectra more than the addition of a RE co-dopant. In Fig. 1(b), background absorption spectra for representative La:Er, Yb:Er and Er fiber with no co-dopant are presented. We can see that the co-dopants do not affect the background absorption in the fibers and they all have a very low background attenuation of less than 50 dB/km. The rising background at shorter wavelengths for the Yb:Er fiber is due to the absorption for Yb absorption band at 975 nm.
Fig. 1 (a) Normalized absorption spectra for RE co-doped fibers compared with a fiber with no co-dopant; (b) background loss for the fibers.
In Fig. 2 we plot efficiencies of the fibers as a function of erbium core absorption for input signal powers of 76 mW (Fig. 2(a)) and 12 mW (Fig. 2(b)). Here, two ratios of aluminum/erbium were investigated, 50/1 (black symbols) and 75/1 (red symbols). For comparison we have included two of our erbium solution doped fibers with an Al/Er ratio of 75/1 and no La or Yb co-dopant and two commercially available erbium doped fibers from Coherent-Nufern: EDFL-1480-HP (30 dB/m) and SM-ESF-7/125 (62.9 dB/m). All fibers are single mode at the signal wavelength of 1560 nm.
Fig. 2 Slope efficiency with respect to absorbed pump power with (a) 76mW signal input power and (b) 12 mW input signal power for RE:Er co-doped fibers with varying ratios of RE:Er from 2:1 to 8:1 and ratios of Al/Er of 50/1 and 75/1.
The two erbium-doped fibers that were made with no RE co-dopant included both low and high Er core absorptions of 20.3 dB/m and 55.8 dB/m, respectively (red circles and Table 1). In Fig. 2(b) for input signal power of 12 mW, we can see that the lower Er concentration fiber with no RE co-dopant and Al/Er = 75/1 had a relatively high slope efficiency of 64% (red circle). By maintaining similar Er and Al concentrations and adding a RE co-dopant to the fibers we investigated its effects on efficiency. It can be seen that adding La at nearly identical Er concentrations and the same Al/Er = 75/1 has served to increase the efficiency for all La:Er ratios (red squares) from 64% to as high as 73.5%. Although this is clear evidence that the RE co-dopant increases efficiency, we have found that there is only a modest increase in going from La:Er of 2:1 to 8:1.
Also included in Fig. 2(b) are two Yb:Er co-doped fibers that were prepared with slightly increased erbium concentrations of 26.3 and 28.2 dB/m (black triangles). These two fibers had a lower Al/Er ratio of 50/1 (and therefore less aluminum) than for the erbium-doped fiber without RE co-dopant, and two ratios of Yb:Er; 4:1 and 8:1. In the case of the fiber with 4:1 Yb:Er ratio, the efficiency has increased from 64% for no co-dopant to 68.3%, indicating that co-doping has been effective as a substitute for aluminum in the fiber to reduce erbium clustering. A further increase in Yb:Er of 8:1 has resulted in a reduced efficiency of 62.7%. The reduced efficiency value from the case of the Yb:Er 4:1 fiber may indicate that higher ratios of RE:Er would benefit more from greater concentrations of Al in the fiber to further solubilize the larger total RE loadings, and that the concentration of aluminum in the Al/Er = 50/1 fibers should be increased.
The highest concentration Er doped fiber with a core absorption of 55.8 dB/m and no RE co-dopant had an efficiency of 53% (Fig. 2(b) red circle and Table 1). This fiber had the highest Al/Er ratio of 75/1, and therefore the highest aluminum concentration. A corresponding fiber with approximately the same erbium and aluminum concentrations, and a Yb co-dopant added in the ratio of 4:1 (Fig. 2(b) red triangle and Table 1) had an increased efficiency to a value of 64.5%. And so by adding the co-dopant, the efficiency has again been greatly increased. When the aluminum concentration is reduced to a Al/Er ratio of 50/1 while maintaining the 4:1 Yb:Er ratio the efficiency is reduced to 58.4% (Fig. 2(b) black triangle and Table 1). But this is still a higher efficiency than the fiber with no co-dopant.
Also presented in this higher Er core absorption region of Fig. 2(b) are two La:Er 4:1 co-doped fibers with a 50/1 Al/Er ratio (black square) and a 75/1 ratio (red square). The fibers had increased slope efficiencies of 54.8% (50/1 Al/Er) and 59.1% (75/1 Al/Er) respectively. The co-doped fiber prepared with 75/1 Al/Er ratio indicates a similar increase in efficiency as the Yb:Er co-doped fiber. In both cases these increased efficiencies represent a substantial improvement when it is considered that the Er core absorptions (and thus concentrations) are higher for these two fibers than for the fiber with no RE co-dopant by as much as 10 dB/m (Table 1). And so, as in the case of lower erbium concentrations, the reduced aluminum concentration (Al/Er = 50) and use of a RE co-dopant of RE:Er 4:1, has given increases in efficiency.
The rare earth co-dopants have high refractive indices of n = 1.877 for La2O3 and n = 1.881 for Yb2O3, while for Al2O3, n = 1.653. Therefore, their addition in the fiber will increase the NA depending on concentration. However, these RE co-dopants are added in low concentrations of less than 0.336 mol% (Table 1) relative to the concentration of Al2O3 (Table 1), thus having a much less pronounced effect on NA. This is important because we can reduce aluminum concentration, and add RE co-dopants with resultant relatively small increases in fiber NA. This is illustrated in Table 1 by a comparison of the high concentration Er fiber with no co-dopant and 4.220 mol% Al2O3 and the Yb:Er 4:1 fiber that has 3.044 mol% Al2O3 (Table 1). Both fibers have a similar NA, 0.148 and 0.152 respectively, but the Yb:Er fiber has increased efficiency. Without the RE co-dopant, an increase in efficiency could only be achieved by adding large concentrations of Al2O3 alone, with the penalty of a much higher NA.
We note that the ytterbium Yb3+ and erbium Er3+ ionic radii are close in value at 100.8 pm and 103 pm respectively, while the ionic radius of lanthanum La3+ at 117.2 pm is significantly larger. However, in view of the efficiency values for the fibers with nearly equal Al, Er, and Yb or La concentrations, the difference in ionic radii does not appear to have a bearing on the co-dopants ability to reduce clustering.
By increasing the input signal power to a high value of 76mW, we were able to increase the slope efficiency for all fibers in this study Fig. 2(a). This represents a situation where the amplifier is near saturation. In Fig. 2(a) we plot efficiency vs. erbium core absorption results for the same fibers as previously in the case of lower input signal power. The higher efficiencies for all fibers here are understandable; at high signal powers, we are able to extract more power from the fibers. The greater efficiencies for rare earth co-doping at lower erbium and aluminum concentrations was still maintained at this higher signal power (Fig. 2(a) red squares). However, the differences in efficiency for rare earth co-doping vs. fibers with no co-dopants at higher erbium core absorptions (>50 dB/m region) are evident, and there does not seem to be a benefit to co-doping. The fiber with the highest core absorption (and thus largest Er concentration) in this study, was a La:Er 4:1 fiber with a core absorption of 65.2 dB/m (black square). Here the efficiency has dropped substantially from the other fibers in this region, and it is clear that the 50/1 Al/Er ratio is not high enough for this RE doping level.
As discussed the benefit to RE co-doping is maintained in the lower Er core absorption region. In this case our 8:1 La:Er co-doped fiber prepared with a 75/1 Al/Er ratio had the highest slope efficiency in this study (Fig. 2(a) red square), with a value of 79.6% (Fig. 3). This fiber had an NA = 0.11 and a mode field diameter calculated to be 11.1 µm at 1560 nm. Here, the Er core absorption is 21.3 dB/m with a corresponding ion concentration of 0.98x1025 ion/m3 as determined by EPMA (Table 1).
Fig. 3 Slope efficiency and fiber refractive index profile (inset) for La-Er doped fiber with La:Er concentration ratio of 8:1.
The RE co-doping approach has been investigated by past researchers but with limited information on fiber design. Aiso et al. reported a high slope efficiency of 75% for a La:Er co-doped fiber with an erbium core absorption of 24 dB/m [6]. The fiber was reported to be doped with germanium and aluminum, but concentrations and ratios for those species were not given. And so we have obtained a modest improvement from their high efficiency value with a comparable Er core absorption. We do note that the highest efficiency reported for a heavily Er doped fiber (58 dB/m) using the co-doping technique was reported by Lin et al., who achieved a slope efficiency of ~80% in a Yb-Er co-doped fiber using an in-band core pumped MOPA [2]. They did not specify the concentration of Yb in the fiber or any other dopants. This high value for efficiency was achieved for a very large signal input value of 1.5 W, confirming that very high signal input powers are needed to achieve high slope efficiencies. Only two fibers were studied in that work. Here we have presented more detail on the concentration values of the constituents to aid in fiber design to optimize this approach.
We have measured gain spectra for fibers in order to ascertain any differences that may occur with RE co-doping. Figure 4 contains representative data at two Er core absorptions and Al concentrations, and we have included our two Er doped fibers that have no RE co-dopant. These latter fibers had an Al/Er ratio of 75/1, and thus the highest Al2O3 concentration, and two core absorption values, ~20 dB/m (open black circles) and ~55dB/m (solid black circles). Because we have maintained the 75/1 Al/Er ratio, the high Er core absorption fiber has a large Al2O3 concentration of 4.220 mol%. Also included in the figure is our La:Er 8:1 co-doped fiber (hollow red squares), and our Yb:Er 4:1 co-doped fiber (solid red triangles). These fibers also have core absorptions of ~20 dB/m and ~55 dB/m, respectively, and were prepared with an Al/Er ratio of 75/1, and so the spectra can be compared with the fibers without co-dopants. The gain spectra were obtained at a low input signal power of 0.6 mW in order to observe any difference that occur due to the co-dopants.
Fig. 4 Gain spectra for Er, La:Er, and Yb:Er co-doped fibers. All fibers have an Al/Er ratio of 75/1. The open points correspond to fiber with Er absorptions of ~20 dB/m and the closed points to ~55 dB/m
At higher signal powers of 12 mW and 76 mW, the gain curves are flattened for all fibers (not presented) and no distinguishable shape differences exist. We can see in Fig. 4 that in the case of the lower Er core absorption (~20 dB/m), both the Er fiber (open black circles) and the La:Er fiber (open red squares) have the same shape throughout the spectrum, where a maximum gain occurs in the 1530 nm region and then a gain flattening occurs in the 1560 nm region. The only difference in the two spectra being that the La:Er fiber has a higher gain across the spectrum. And so it is evident that the width and shape of the gain spectrum has not been altered by the RE co-dopant. This same general shape for Er fiber and a co-doped fiber is also evident when looking at the two curves for the higher Er core absorptions of ~55 dB/m. Here the Er fiber (solid black circles) has the same shape as the Yb:Er fiber (solid red triangles). This situation is similar to our comparison of absorption spectra presented earlier. The Al2O3 in the fibers dictates the shape and width of the gain spectra far more than the RE co-dopant [8]. For the higher Er core absorption spectra it is evident that an increase in the gain at longer wavelengths occurs for both of the fibers. Here the peak at 1530 nm is reduced and the maximum gain occurs in the region near 1560 nm. This behavior is typical for the situation where less Er ions are inverted in the fiber [9]. We expect this, as these fibers are more heavily Er doped than in the previous case.
4. Conclusions
In conclusion, we have used RE co-doping of erbium fibers in a solution doping technique and have found that slope efficiencies are increased by the addition of these RE “spacers” which help to reduce erbium clustering. The benefits of RE co-dopants requires a suitable Al concentration in the fiber. This was best illustrated at lower erbium concentrations and high Al/Er ratio of 75/1, where increases of La:Er ratios improved efficiency. For the case of near saturated input signals, the benefits of rare earth co-doping at lower erbium concentrations are maintained, but this benefit is not well maintained as the Er concentrations are increased. We note that this method of co-doping may have its greatest benefit for clad pumped fiber lasers, where the greatest possible power needs to be extracted for a situation of minimal signal powers. Finally, we note that we have achieved a core pumped single mode slope efficiency of 79.6% using these optimized conditions for a fiber with an Er core absorption of 21.3 dB/m, 0.94x1025 ions/m3.
Funding
High Energy Laser Joint Technology Office (Project 15-S&A-0553)
References
1. C. C. Baker, E. J. Friebele, A. A. Burdett, D. L. Rhonehouse, J. Fontana, W. Kim, S. R. Bowman, L. B. Shaw, J. Sanghera, J. Zhang, R. Pattnaik, M. Dubinskii, J. Ballato, C. Kucera, A. Vargas, A. Hemming, N. Simakov, and J. Haub, “Nanoparticle doping for high power fiber lasers at eye-safer wavelengths,” Opt. Express 25(12), 13903–13915 (2017). [CrossRef] [PubMed]
2. E.-L. Lim, S. U. Alam, and D. J. Richardson, “Optimizing the pumping configuration for the power scaling of in-band pumped erbium doped fiber amplifiers,” Opt. Express 20(13), 13886–13895 (2012). [CrossRef] [PubMed]
3. M. A. Jebali, J.-N. Maran, and S. LaRochelle, “264 W output power at 1585 nm in Er-Yb codoped fiber laser using in-band pumping,” Opt. Lett. 39(13), 3974–3977 (2014). [CrossRef] [PubMed]
4. N. V. Kiritchenko, L. V. Kotov, M. A. Melkumov, M. E. Likhachev, M. M. Bubnov, M. V. Yashkov, A. Yu Laptev, and A. N. Guryanov, “Effect of ytterbium co-doping on erbium clustering in silica-doped glass Laser,” Phys. 25, 025102 (2015).
5. M. Nakazawa and Y. Kimura, “Lanthanum codoped erbium fibre amplifier,” Electron. Lett. 27(12), 1065–1067 (1991). [CrossRef]
6. K. Aiso, Y. Tashiro, T. Suzuki, and T. Yagi, “Erbium Lanthanum co-doped fiber for L-band amplifier with high efficiency, low non-linearity and low NF”, in Optical Fiber Communication Conference and International Conference on Quantum Information, OSA, paper TuA6. (2001).
7. K. Arai, H. Namikawa, K. Kumata, T. Honda, Y. Ishii, and T. Handa, “Aluminum or phosphorus co-doping effects on the fluorescence and structural properties of neodymium-doped silica glass,” J. Appl. Phys. 59(10), 3430–3436 (1986). [CrossRef]
8. W. J. Miniscalco, “Optical and Electronic Properties of Rare Earth Ions in Glasses,” in Rare-Earth-Doped Fiber Lasers and Amplifiers, 2nd ed, M. J. F. Digonnet, ed. (Marcel Dekker Inc. 2001).
9. P. F. Wysocki, “Erbium-Doped Fiber Amplifiers: Advanced Topics,” in Rare-Earth-Doped Fiber Lasers and Amplifiers, 2nd ed, M. J. F. Digonnet, ed. (Marcel Dekker Inc. 2001).
