Open Access
Add to BibSonomyAbstract
We have constructed a double clad neodymium doped fiber laser operating on the three-level 4F3/2→4I9/2 transition. The laser has produced 11.5 W at 925 nm with 55% slope efficiency when pumped at 808 nm, comparable to the best previous results for a double-clad fiber configuration on this transition. Higher power pumping with both 808 nm and 880 nm sources resulted in an output of 27 W, albeit at lower slope efficiency. In both cases, output power was limited by available pump, indicating the potential for further power scaling. To suppress the stronger four-level 4F3/2→4I11/2 transition we developed a waveguide that provides spectral filtering distributed along the length of the fiber, based on an all-solid micro-structured optical fiber design, with resonant inclusions creating a leakage path to the cladding. The waveguide supports large mode areas and provides strong suppression at selectable wavelength bands, thus easing the restrictions on core and cladding sizes that limited power scaling of previous approaches.
© 2016 Optical Society of America
1. Introduction
Fiber laser technology has seen remarkable maturation in the last decade, as illustrated by demonstrations of 3 kW diffraction limited output (diode pumped) [1], 10 kW single mode (tandem pumped) [2] and a 10 kW (multimode) system now commercially available [3]. These results approach the theoretical limits for round geometries [4], and have been based almost exclusively on ytterbium (Yb) as the active medium. The high efficiency of these devices derives from their small quantum defect and low saturation intensity, and in spite of their 3-level transition which causes a higher threshold.
Neodymium (Nd) doped materials have long been the active medium of choice for high energy bulk lasers, owing to a favorable 4-level transition around 1060 nm, and the first fiber laser reported [5] (by Snitzer, over fifty years ago) as well as the initial results from University of Southampton [6] used Nd.
Nd3+ remains attractive because of its richer set of energy levels, which enables operation at other wavelengths, particularly around 900-950 nm, providing the dominant transition around 1060 nm can be suppressed. The 3-level transition around 925 nm, useful for remote sensing (water vapor), Guidestar lasers [7] and underwater communications via harmonic conversion is the focus of the present work. This work is based on a waveguide design that we believe is particularly advantageous over prior designs in suppressing the 1060-1140 nm wavelength band.
The earliest 900-950 nm Nd3+ fiber lasers were limited by available diode power and a core pumped configuration. Alcock [8, 9] demonstrated the first fiber laser in this wavelength region, achieving broad tunability at a few mW and less than 10% slope efficiency (SE). Reekie [10] followed with similar power and improved SE of 36%. Cook [11] optimized the laser configuration and got nearly 50 mW with 58% SE. Dragic [12, 13] showed amplification on the 3-level transition and reported on the effects of co-dopants on wavelengths and relative strengths of the 3- and 4-level transitions.
Higher powers are possible with a double-clad pump configuration and wavelength discrimination against the 4-level transition. Bufetov [14], Soderlund [15], Soh [16,17] and Fu [18] all used a depressed-well waveguide to suppress the 4-level transition in a double clad Nd-doped fiber, achieving various output powers and slope efficiencies. The best results were those of Soh, who achieved a 2.5 W tunable laser with 41% SE.
A second approach pioneered by Dawson [7] uses a double clad fiber configuration where the ratio of pump cladding to core diameter is limited to achieve high pump intensity. This approach enables the fiber amplifier to operate with a high inversion. In this case, the gain in the 1060-1120 nm region can be very high, more than 10x higher than the gain in the 900-950 nm region. The preference for operation at 900-950 nm is achieved in this case by careful management of stray reflections and seeding in the 900-950 nm range. In an oscillator configuration a bulk filter can be added to generate high losses in the 1060-1120 nm range. Dawson achieved 15 W at 938 nm in a large core double clad MOPA configuration. Efficiency was low due to a core composition consisting only of silica, germanium and neodymium. This core composition greatly limits solubility of the Nd3+ ions and is exceptionally prone to concentration quenching. However, it does have the uncommon property of shifting the emission spectra of the Nd3+ ion approximately 24 nm longer in wavelength, which was essential to Dawson’s application.
Wang [19] spliced a passive Photonic Bandgap (PBG) fiber to a Nd-doped fiber to construct a laser with strong wavelength selection; core pumping yielded about 200 mW with 32% SE. As with the configuration described by Dawson, the gain at 1060-1120 nm was still very high.
Most recently, Laroche [20] and Leconte [21] used the same approach as Dawson, supplemented by wavelength selective feedback from Bragg gratings, to achieve outputs in the range of 20 W and SE ~58%. Bartolacci and Laroche also employed a more favorable core composition, based on a careful assessment of the doping concentration limits of the Nd3+ ions to minimize concentration quenching [22].
Further efficient power scaling requires a large mode area core fiber and strong suppression for the 4-level transition. Depressed well designs do not meet the former requirement. High inversion (tight pump cladding) will not meet the latter, and will eventually limit pump coupling or output beam quality if the core is enlarged and becomes multi-mode. Finally, it is doubtful that filtering elements operating only at the ends of the fiber will provide sufficient suppression of the 4-level transition at higher powers, especially if the high inversion condition is relaxed for the sake of efficiency (pump absorption).
We have developed a novel scalable waveguide that supports large mode areas and provides strong distributed suppression of the 4-level transition, thus removing limits on the size of the pump cladding and providing a path for power scaling on the 3-level transition.
2. Waveguide design
Our waveguide design grew out of previous work [23] on a short pass PBG fiber intended to truncate the Raman cascade in a Raman fiber laser. The PBG fiber is a versatile design, enabling strong spectral filtering [24,25] as well as higher order mode suppression [25–27]. However, concern with pump light guiding by, and possible Raman amplification in, the high index elements led to consideration of a hybrid structure.
The current waveguide is based on the Photonic Crystal Fiber (PCF), which consists of a lattice of index depressions, with one or more defects (missing depressions) comprising the guiding core. The fundamental mode guidance can be thought of qualitatively as being due to a reduced average index surrounding the core. By virtue of restricting the material selection and the stack-and-draw fabrication process, the index contrast between the core and lattice is determined by geometry and can be precisely controlled. The PCF structure supports large single mode operation because a low index contrast can be attained reliably.
In contrast with PBG fibers, in which the lattice elements are raised in index, in a PCF the fundamental mode guidance is not a resonant effect. (The PCF does possess bandgaps, with indices below that of the both the background glass and the lattice space filling mode; in a truncated lattice modes associated with the gaps are highly lossy, and will be ignored here.) To achieve wavelength filtering in our waveguide we replace several down-doped lattice sites with up-doped inclusions, away from the core and extending to the edge of the PCF lattice, as shown in Fig. 1. These are chosen to be resonant with the fundamental core mode at the wavelength we wish to filter: at resonance, the core mode couples to these inclusions. Crucially, because the core mode’s effective index is below that of the background glass, and there is no barrier between the up-doped inclusions and the background, light in the core is coupled out of the PCF lattice. That is, guidance is lost when the resonance condition is satisfied.
Fig. 1 Schematic of the waveguide. Open circles are the down-doped PCF lattice, filled circles are resonant out-couplers. The core is comprised of seven cells index matched to the background glass.
We simulated the waveguide via the Finite Element Method (FEM, using COMSOL) in the scalar approximation, which is suitable due the low index contrast of the structure. A comparison of scalar and vector models for a representative case shows that despite differences in fine detail the former is sufficiently accurate for design purposes, as verified by the match between our model and fabricated fibers (see below).
Considering first the PCF alone, without the filtering inclusions, the resulting fiber is in fact “few moded”, with the higher order modes (HOM) susceptible to stripping by coiling of the fiber. Modeling indicates that the fundamental core mode has an effective area of 323 μm2 (MFD = 20.3 μm) at 925 nm. Coiling the fiber to a 200 mm diameter induces in the fundamental mode a reduction of the effective area by less than 1% and a propagation loss of ~1 dB/km, while the first HOM suffers a loss of 150 dB/km (averaged over orientations).
For the up-doped inclusions we used elements having a graded-index (GRIN) profile. The properties of the GRIN are chosen so that its first higher order mode will be resonant with the fundamental core mode at wavelengths within the gain peak of the Nd3+ four level transition. This GRIN mode is chosen over the GRIN fundamental due to the fact that it has a cutoff and moderately higher dispersion, which lead to an interaction with the core that’s better localized in wavelength. In the scalar approximation this mode is doubly degenerate, however parity allows coupling of only one of the two to the core mode.
These GRINs are arranged six groups or ‘spokes’ of three inclusions. The different spokes are well isolated from each other, but within a spoke, coupling between the GRINs causes a splitting of their propagation constants. The net result of the six identical triplets of GRINs is a triplet of loss peaks. Coiling of the fiber will break the symmetry between the spokes, further scrambling the resonant wavelengths and reducing the contrast in loss band.
In Fig. 2 is shown the calculated effective indices for waveguide and some of its constituent elements, as well as the confinement loss of the structure assuming no air cladding. The dispersion of the GRIN mode is high compared to that of the core, due to its smaller size and higher index contrast, yielding a narrow resonance band with the core fundamental mode. The GRIN mode will couple also to any HOMs supported by the core, although the resonance will be shifted in wavelength according to the dispersions of the modes involved. We highlight that although the GRIN HOM has a cutoff with respect to the PCF lattice, its resonance with the core and the loss that is induced by it occurs at shorter wavelengths. This is due to the coupling of the GRINs to the higher index background glass. That is, the loss of the core mode is not due to delocalization of the GRIN mode in the PCF lattice, but to coupling to the region outside of it. Figure 3 shows the calculated effect on the loss band due to coiling of the fiber.
Fig. 2 Calculated relative effective indices of the core mode (red), the space-filling mode of an infinite PCF lattice (blue, dashed), and the resonant mode a single GRIN inclusion in the PCF lattice. The core confinement loss is plotted on the second axis.
Fig. 3 Calculated core mode loss for straight and coiled fibers. The block band contrast in coiled fiber is reduced due to shifts in the GRIN node effective indices, and it is wider.
We emphasize that core mode confinement is governed by the base PCF structure. While the raised index elements are similar to those of the PBG fiber, their role is not confinement but to provide a leakage path out of the core. They comprise a secondary waveguide structure neighboring the core, to which the core can couple when the resonance condition is satisfied, and which is directly coupled to the outer unstructured region of the fiber. The similarity to a PBG is accidental: the neighboring structure need not present a bandgap, nor need it be periodic, it can in fact be monolithic rather than composed of elements on the PCF pitch. This distinguishes the current waveguide from that of the hybrid PBG/PCF (Cerqueira [28], Coscelli & Poli & Alkeskjold [29–31], Goto [32,33], Xiao [34]); it is better compared to coupled core fibers (Wang [35], Ward [36], Ma [37], Fini [38,39]) and designs which make use of mode delocalization (see [40] and references therein), both of which are typically used for HOM suppression. The key elements of the present design are (i) a mechanism for core confinement, (ii) a high index outer region for stripping, and (iii) a means of resonantly coupling the two. In light of this, it’s also the case that core confinement need not be provided by a PCF structure. A depressed well large enough to accommodate a resonant side core would also be suitable, although large mode areas are more readily achievable with a PCF.
To substantiate the above points, we have simulated variations of the design, as shown in Fig. 4. Specifically, we examine the behavior when we break the periodicity of the GRIN inclusions by shifting the middle one by 1/5th of the pitch, and when we replace the GRINs with a monolithic slab having a mode count similar to the collection of GRINs. In all cases the side structure induces loss in the central core when the two become resonant, showing that the effect does not depend upon a PBG.
However, the spectrum reveals an advantage of constructing the leakage path as a string of small structures. The modes of the string are supermodes of the constituent GRINs, with sparse effective indices clustered around those of a single GRIN. In contrast to the monolithic slab, the loss feature due to the string of GRINs is spectrally isolated. This allows for targeting specific laser lines while leaving the rest of the gain spectrum undisturbed.
3. Experimental – fiber fabrication & characterization
We fabricated the fiber via the stack-and-draw method, the structure being defined by the arrangement of various canes in the preform. The base glass was pure fused silica, and the PCF and GRIN elements were canes with cores doped with fluorine and germanium, respectively, surrounded by silica. The PCF elements have an index contrast of −6.74 x 10−3 and an ID/OD ratio of 0.533, while the GRINS have a parabolic index profile (not truncated) with a peak contrast 3.07 x 10−2 and a 0.5 ID/OD ratio.
The fiber design calls for the core to have the same index as the base glass, but the Nd3+ doping required for gain raises the index. For this reason we separately fabricated the core canes by an iterative homogenizing process, starting with a heterogeneous collection of silica, Nd3+ and fluorine doped canes chosen to have an average index equal to that of pure silica. These are stacked to produce a preform that is then drawn to canes of a size suitable for repeating the process, and with each iteration the heterogeneous features are reduced in size. We estimate from modeling that in the final fiber feature sizes less than 250 nm result in negligible perturbation to the mode and its effective index. With two iterations, we achieved a feature size of approximately 125 nm. Figure 5 shows one of the homogenized canes, in which the features are clearly visible. The figure also shows the final drawn fiber, the core consisting of seven of these canes, on a pitch of 8 μm. While we do not have the capability to quantify the index variations in the drawn fiber, in the figure the core appears qualitatively homogeneous and matched to the pure fused silica of the rest of the structure. For reference, the PCF elements are down-doped by −6.7 x 10−3 and exhibit high contrast in the image.
Fig. 5 (a) One of the core canes, with a diameter of about 1.3 mm, and (b) the final drawn fiber, of 225 μm diameter.
Our starting Nd3+ material was an experimental highly doped test preform, codoped with aluminum to increase solubility, nominally at a ratio of 8:1 with respect to Nd (Optacore, Slovenia) The absorption at 808 nm was 200 dB/m, high enough to raise concerns over clustering [22]. The homogenization process necessarily dilutes the area averaged Nd3+ concentration of the core material, reducing it in this case to 21.3% of its original value, and yielding a nominal pump absorption of 42.6 dB/m. The final 125 nm feature size is sufficiently small to avoid affecting the core mode.
The finished fiber has a pitch of 8 μm, a core of 7 elements, a PCF lattice of four rings outside the core and six strings of GRINs flush with the outer edge of the PCF lattice. Outside of this structure there are three rings of pure fused silica, and one ring of hollow capillaries comprising an air cladding for the pump. Due to difficulty controlling the inflating pressure during the draw, the air cladding was limited to 0.4 NA. The air cladding surrounds a total of 217 elements yielding a ratio of 31 in clad:core area. Its width from vertex to vertex is 17 pitches, or approximately 118 μm from face to face.
It might be objected that the air cladding will confine whatever light is lost from the core due to our leaky structure, weakening the filtering effect. However, the fused silica region outside the lattice supports a multitude of modes that can be expected to be coupled by waveguide roughness or other non-uniformities, or by the distortions induced by coiling. This region therefore constitutes a bath, a large phase space (position & angle) from which light is unlikely to couple back to the isolated phase space point that is the core mode. As shown below, the fabricated fiber does indeed exhibit the desired filtering.
Single and double clad lengths of fiber were drawn by collapsing or inflating the cladding air holes, respectively. Core attenuation was measured using the single clad fiber to ensure that light lost from the core would not be trapped by the air cladding. This was measured by a standard cut-back method, modified by an imaging stage between fiber output and detector, incorporating an aperture to block any light trapped in the GRINs. The measured spectrum is shown in Fig. 6, along with the calculated block band for comparison. The calculated and measured block bands are in reasonable agreement with respect to location and average magnitude. Although effort was made to maintain a straight and unperturbed fiber for this measurement, the core and GRIN effective indices are sensitive to residual bending perturbations and fabrication variations, which would affect the resonances. Some of the differences may also be attributable to our use of a scalar model instead of a higher fidelity (but more time consuming) full vector treatment. In any case, the differences are similar to those expected from coiling, which tends to wash out fine details (see Fig. 3).
Fig. 6 Core attenuation measured on a single clad fiber. Two lengths of fiber were used to measure the block band (red) and background (blue). Also shown is the calculated ideal block band for a straight fiber.
Also evident from Fig. 6 is a relatively high background loss (up to ¼ dB/m) and a very high (1 dB/m) OH absorption feature at 1380 nm. These losses are much higher than those typically found in optical fibers. Simulations indicate that the core structure should be robustly low loss outside the block band, and the presence of the high OH peak suggests contamination. A simple single mode fiber drawn from the same starting material showed a much lower OH peak of around 50 dB/km, but also a broad background loss similar to that in the filtering fiber. We conclude that the background is inherent to the starting material, which as mentioned was a test piece from the vendor; fibers of other compositions supplied by the same vendor have low loss, and we expect that this can be mitigated with further development. As for the OH peak, we attribute that to cleanliness issues during the extensive handling during our homogenization process.
A further cut back measurement was made on fiber with inflated air cladding to assess the cladding absorption at the pump wavelength as well as background attenuation. As shown in Fig. 7, the background attenuation is acceptably low, around 25 dB/km. However, the peak cladding absorption around 808 nm is only 0.4 dB/m. This is significantly less than the value of 1.37 dB/m we expected from the absorption of the starting material, after accounting for the homogenization and the clad:core ratio.
We originally suspected this was due to the segregation of some fraction of the pump light to the region outside the PCF lattice. To investigate this, we injected 808 nm pump light into a 30 m long piece of the fiber, at a power low enough to avoid generating any ASE, and imaged the output face. As can be seen in Fig. 8, the transmitted light is concentrated in the region outside the lattice and the GRINs. (The light in the GRINs is much brighter than elsewhere because they are low loss and provide good confinement at the pump wavelength; but based on the GRIN sizes and NA we estimate that the entire group of GRINs will capture less than 1% of the incident pump light.) However, from the parameters of the lattice, we can attribute to it an effective NA of 0.051 at the pump wavelength of 808 nm, while the pump light was launched with a 0.22 NA. Given the ratios of NA and of the area outside the lattice to the total cross sectional area, we could expect that only about 1% of the pump would be prevented from reaching the core. Still, the lattice does constitute a barrier between the core and outer pump cladding, and may impede transfer between them of at least some low NA portion of the pump light. Another possibility is the existence of a population of pump cladding modes having low overlap with the core. This is well known to degrade the pump absorption efficiency of double clad fibers, but the effect depends on the cladding shape and is not expected to be strong for a hexagonal cladding under our experimental conditions [41]. We are currently conducting beam propagation simulations to determine the cause of the reduced cladding absorption. Preliminary results show a segregation of the pump similar to that seen in our experiment, but also indicate that a substantial portion of the pump will be absorbed before the segregation becomes pronounced. We have not experimentally quantified the pump absorption vs. distance from the fiber entrance; nevertheless, we note that the length of fiber required to achieve the targeted absorption of the pump was approximately three times what we expected, consistent with the absorption constant derived from our cutback measurement.
We are addressing the issues contributing to the background loss and OH peak in the core, as well as considering designs that will mitigate the segregation of pump light away from the core, and plan to fabricate an improved fiber in the near future.
4. Experimental – laser tests
We have built fiber lasers in several configurations using the fiber described in the previous section. Figure 9 below illustrates a block diagram of a representative experimental set-up for the results reported here. The fiber was coiled to 200 mm diameter loops both to minimize higher order mode content and to adjust the suppression to cover 1060-1140 nm light. When the fiber was more loosely coiled, the leakage band was slightly shorter than desired and laser oscillation at 1120 nm was observed. The cavity is formed by Fresnel reflection from the fiber cleave on one end and by a broadband (covering both the emission bands as well the pump) High Reflection (HR) mirror on the other. The HR mirror also reflected the pump laser enabling the pump to be double passed for greater absorption in a shorter length of fiber. The layout is shown schematically in Fig. 9.
For our first test we pumped at 808 nm, a 25 W fiber coupled diode laser module with a 100 µm/NA0.22 multimode output fiber (LIMO). The pump light was transported to the gain fiber via collimating and focusing lenses between which were placed two dichroic mirrors splitters (Semrock FF875-Di01, 35 degree angle of incidence) for isolating the fiber laser from the pump laser. Because the 808 nm pump delivery fiber is slightly smaller than the air cladding of the gain fiber and has much lower NA, a simple magnification = 1 arrangement (two 20 mm focal length lenses, Thorlabs AL2520) was used to launch the pump. Figure 10 shows the output laser power vs. coupled pump power. The Nd gain fiber length was 10m, resulting in an estimated pump absorption of 84%, including the effect of double passing the pump via the HR mirror at the far end. As shown in Fig. 10, we achieved a lasing slope efficiency of 54.8%, and a maximum output power of 11.5 W, limited by available pump power. Figure 11 shows the lasing spectrum, with high contrast of nearly 40 dB of contrast between the three- and four-level transitions, due to our filtering fiber design.
The output beam quality was somewhat worse than expected as assessed by the M2 method (measured via Gentec BEAMAGE-CCD12), which yielded values of 1.32 and 1.38 in orthogonal directions (see Fig. 12). This can’t be attributed to inherent features of the fundamental mode of the fiber: although the fundamental mode of the fiber has structure related to the surrounding PCF and GRIN elements, its overlap integral with a best fit Gaussian beam is approximately 99%. We conclude that either the fundamental mode is distorted due to thermal loading, or more likely, we have a small admixture of the first higher order mode. The latter might be addressed most easily by tighter coiling of the fiber, the HOM being much more sensitive to bending than the fundamental mode. Beyond that, it’s also possible to adjust the confinement due to the PCF lattice by changing its pitch and/or its down-doped inclusions, while adjusting the properties of the GRINs to maintain the desired block band. Perhaps more importantly, pump absorption and signal amplification will deposit heat in the core, causing thermal lensing that will tend to increase confinement and HOM content. For higher power operation it may be possible and necessary to take that into account in the waveguide design. However, these are well known challenges to Large Mode Area (LMA) fiber design in general [42], and are not specific our wavelength selective waveguide.
We were limited to 25 W pump at 808 nm, but we did have an additional pump available at another wavelength. This was a 100 W, 880 nm module (DILAS), 200 µm/NA0.22 multimode output fiber. In an effort to get a higher output power we decided to use both pump sources, pumping the fiber from both ends. Referring to Fig. 9, the 808 nm pump was moved to the end with the HR mirror, with two additional dichroics for isolation. The output coupler was again the Fresnel reflection from the cleaved fiber, and the 880 nm pump was launched from this output side. For the 880 nm pump, a de-magnification by 2x was used, via 40 mm (PCX) and 20 mm (Thorlabs AL2520) focal length lenses. Finally, since the Nd absorption is lower at 880 nm (see Fig. 7) we increased the fiber length from 10 m to 30 m. We found that the NA of the 880 nm pump delivery fiber was slightly higher than expected, there being a reduced intensity halo beyond the nominal 0.22 NA. That, along with the modest NA (0.4) of our pump cladding, resulted in a reduced coupling into the gain fiber of about 85%. Nevertheless, we achieved an output of 26.7 W at a slope efficiency of 35.3%, as shown in Fig. 13. For this test the pumps at 880 nm and 808 nm pumps were ramped up together, with a nearly constant 3:1 power ratio. The lower efficiency with 880 nm diode can be explained by higher background loss in the fiber accumulated in longer fiber length. Both 808 nm and 880 nm pumping laser results are limited by available pump diode power that could be coupled to the fiber.
5. Conclusions
We have designed and built a novel large mode double clad Nd doped silica fiber for operation on the three-level 4F3/2→4I9/2 transition of neodymium. The design features distributed spectral filtering to suppress the stronger four-level 4F3/2→4I11/2 transition, thus allowing efficient operation around 925 nm. Previously reported schemes relied on high pump intensities (via small clad core ratio) to make the gains similar on the two transitions; or a depressed well core design that loses guidance at the four-level transition. Neither of these approaches provides a power scaling path for single mode operation due to limits on either the core or cladding size, or both. In contrast, our approach achieves robust suppression of the four-level transition, independent of core and cladding sizes, thus lifting the restrictions on power scaling.
We have constructed a laser based on this fiber, and achieved performance in terms of slope efficiency (55% with 808 nm pumping) and output power (27 W with dual 808 nm & 880 nm pumping) comparable to or exceeding previous results for a double clad fiber laser operating on the three-level transition of neodymium. These results were pump power limited and showed no sign of deterioration at the highest pump powers available to us, indicating that further power scaling is possible.
We are currently working on a fiber with improved background losses and pump absorption, which we expect will enable further scaling of both power and slope efficiency.
Funding
This work was funded by Department of Energy’s Lawrence Livermore National Laboratory under a Laboratory Directed Research and Development grant (14-ERD-078).
Acknowledgments
This work was performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under Contract DE-AC52-07NA27344.
This document was prepared as an account of work sponsored by an agency of the United States government. Neither the United States government nor Lawrence Livermore National Security, LLC, nor any of their employees makes any warranty, expressed or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States government or Lawrence Livermore National Security, LLC. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States government or Lawrence Livermore National Security, LLC, and shall not be used for advertising or product endorsement purposes.
References and links
1. Y. Jeong, A. J. Boyland, J. K. Sahu, S. Chung, J. Nilsson, and D. N. Payne, “Multi-kilowatt single-mode ytterbium-doped large-core fiber laser,” J. Opt. Soc. Korea 13(4), 416–422 (2009). [CrossRef]
2. E. Stiles, “New developments in IPG fiber laser technology,” in 5th International Workshop on Fiber Lasers (2009).
3. I. P. G. Photonics, YLS series, http://www.ipgphotonics.com.
4. J. W. Dawson, M. J. Messerly, R. J. Beach, M. Y. Shverdin, E. A. Stappaerts, A. K. Sridharan, P. H. Pax, J. E. Heebner, C. W. Siders, and C. P. J. Barty, “Analysis of the scalability of diffraction-limited fiber lasers and amplifiers to high average power,” Opt. Express 16(17), 13240–13266 (2008). [CrossRef] [PubMed]
5. C. J. Koester and E. Snitzer, “Amplification in a fiber laser,” Appl. Opt. 3(10), 1182–1186 (1964). [CrossRef]
6. S. B. Poole, D. N. Payne, and M. E. Fermann, “Fabrication of low-loss optical fibres containing rare-earth ions,” Electron. Lett. 21(17), 737–738 (1985). [CrossRef]
7. J. W. Dawson, A. D. Drobshoff, R. J. Beach, M. J. Messerly, S. A. Payne, A. Brown, D. M. Pennington, D. J. Bamford, S. J. Sharpe, and D. J. Cook, “Multi-watt 589nm fiber laser source,” Proc. SPIE 6102, 61021F (2006). [CrossRef]
8. I. P. Alcock, A. I. Ferguson, D. C. Hanna, and A. C. Tropper, “Tunable, continuous-wave neodymium-doped monomode-fiber laser operating at 0.900-0.945 and 1.070-1.135 microm,” Opt. Lett. 11(11), 709–711 (1986). [CrossRef] [PubMed]
9. I. P. Alcock, A. I. Ferguson, D. C. Hanna, and A. C. Tropper, “Continuous-wave oscillation of a monomode neodymium-doped fibre laser at 0.9μm on the 4F3/2→4I9/2 transition,” Opt. Commun. 58(6), 405–408 (1986). [CrossRef]
10. L. Reekie, I. M. Jauncey, S. B. Poole, and D. N. Payne, “Diode-laser-pumped Nd3+-doped fibre laser operating at 938nm,” Electron. Lett. 23(17), 884–885 (1987). [CrossRef]
11. A. L. Cook and H. D. Hendricks, “Diode-laser-pumped tunable 896-939.5-nm neodymium-doped fiber laser with 43-mW output power,” Appl. Opt. 37(15), 3276–3281 (1998). [CrossRef] [PubMed]
12. P. D. Dragic, L. M. Little, and G. C. Papen, “Fiber amplification in the 940-nm water vapor absorption band using the 4F3/2 - 4I9/2 transition in Nd,” IEEE Photonics Technol. Lett. 9(11), 1478–1480 (1997). [CrossRef]
13. P. D. Dragic and G. C. Papen, “Efficient amplification using the 4F3/2 - 4I9/2 transition in Nd-doped silica fiber,” IEEE Photon. Technol. Lett. 11(12), 1593–1595 (1999). [CrossRef]
14. I. A. Bufetov, V. V. Dudin, A. V. Shubin, A. K. Senatorov, E. M. Dianov, A. B. Grudinin, S. E. Goncharov, I. D. Zalevskii, A. N. Guryanov, M. V. Yashkov, A. A. Umnikov, and N. N. Vechkanov, “Efficient 0.9μm neodymium-doped single-mode fibre laser,” Quantum Electron. 33(12), 1035–1037 (2003). [CrossRef]
15. M. J. Soderlund and S. K. T. Tammela, “A depressed-index cladding filter for suppressing 1.06μm emission in a Nd-doped silica fiber amplifier,” in Proceedings Optical Fiber Communication Conference (IEEE, 2004), pp. 3–5.
16. D. B. S. Soh, S. Yoo, J. Nilsson, J. K. Sahu, K. Oh, S. Baek, Y. Jeong, C. Codemard, P. Dupriez, J. Kim, and V. Philippov, “Neodymium-doped cladding-pumped aluminosilicate fiber laser tunable in the 0.9μm wavelength range,” J. Quant. Electron. 40(9), 1275–1282 (2004). [CrossRef]
17. S. Yoo, D. B. S. Soh, J. Kim, Y. Jung, J. Nilsson, J. K. Sahu, J. W. Lee, and K. Oh, “Analysis of W-type waveguide for Nd-doped fiber laser operating near 940nm,” Opt. Commun. 247(1), 153–162 (2005). [CrossRef]
18. L. B. Fu, M. Ibsen, D. J. Richardson, J. Nilsson, D. N. Payne, and A. B. Grudinin, “Compact high-power tunable three-level operation of double cladding Nd-doped fiber laser,” IEEE Photonics Technol. Lett. 17(2), 306–308 (2005). [CrossRef]
19. A. Wang, A. K. George, and J. C. Knight, “Three-level neodymium fiber laser incorporating photonic bandgap fiber,” Opt. Lett. 31(10), 1388–1390 (2006). [CrossRef] [PubMed]
20. M. Laroche, B. Cadier, H. Gilles, S. Girard, L. Lablonde, and T. Robin, “20 W continuous-wave cladding-pumped Nd-doped fiber laser at 910 nm,” Opt. Lett. 38(16), 3065–3067 (2013). [CrossRef] [PubMed]
21. B. Leconte, B. Cadier, H. Gilles, S. Girard, T. Robin, and M. Laroche, “Extended tunability of Nd-doped fiber lasers operating at 872-936 nm,” Opt. Lett. 40(17), 4098–4101 (2015). [CrossRef] [PubMed]
22. C. Bartolacci, M. Laroche, T. Robin, B. Cadier, S. Girard, and H. Gilles, “Effects of ions clustering in Nd3+/Al3+-codoped double-clad fiber laser operating near 930nm,” Appl. Phys. B 98(2), 317–322 (2010). [CrossRef]
23. B. Ward, “Solid-core photonic bandgap fibers for cladding-pumped Raman amplification,” Opt. Express 19(12), 11852–11866 (2011). [CrossRef] [PubMed]
24. G. Bouwmans, L. Bigot, Y. Quiquempois, F. Lopez, L. Provino, and M. Douay, “Fabrication and characterization of an all-solid 2D photonic bandgap fiber with a low-loss region (< 20 dB/km) around 1550 nm,” Opt. Express 13(21), 8452–8459 (2005). [CrossRef] [PubMed]
25. F. Kong, K. Saitoh, D. Mcclane, T. Hawkins, P. Foy, G. Gu, and L. Dong, “Mode area scaling with all-solid photonic bandgap fibers,” Opt. Express 20(24), 26363–26372 (2012). [CrossRef] [PubMed]
26. T. Murao, K. Saitoh, and M. Koshiba, “Multiple resonant coupling mechanism for suppression of higher-order modes in all-solid photonic bandgap fibers with heterostructured cladding,” Opt. Express 19(3), 1713–1727 (2011). [CrossRef] [PubMed]
27. A. Baz, L. Bigot, G. Bouwmans, and Y. Quiquempois, “Single-mode, large mode area, solid-core photonic bandgap fiber with hetero-structured cladding,” J. Lightwave Technol. 31(5), 830–835 (2013). [CrossRef]
28. A. Cerqueira S Jr, F. Luan, C. M. B. Cordeiro, A. K. George, and J. C. Knight, “Hybrid photonic crystal fiber,” Opt. Express 14(2), 926–931 (2006). [CrossRef] [PubMed]
29. T. T. Alkeskjold, “Large-mode-area ytterbium-doped fiber amplifier with distributed narrow spectral filtering and reduced bend sensitivity,” Opt. Express 17(19), 16394–16405 (2009). [CrossRef] [PubMed]
30. E. Coscelli, F. Poli, T. T. Alkeskjold, D. Passaro, A. Cucinotta, L. Leick, J. Broeng, and S. Selleri, “Single-mode analysis of Yb-doped double-cladding distributed spectral filtering photonic crystal fibers,” Opt. Express 18(26), 27197–27204 (2010). [CrossRef] [PubMed]
31. S. R. Petersen, T. T. Alkeskjold, F. Poli, E. Coscelli, M. M. Jørgensen, M. Laurila, J. Lægsgaard, and J. Broeng, “Hybrid Ytterbium-doped large-mode-area photonic crystal fiber amplifier for long wavelengths,” Opt. Express 20(6), 6010–6020 (2012). [CrossRef] [PubMed]
32. R. Goto, S. D. Jackson, S. Fleming, B. T. Kuhlmey, B. J. Eggleton, and K. Himeno, “Birefringent all-solid hybrid microstructured fiber,” Opt. Express 16(23), 18752–18763 (2008). [CrossRef] [PubMed]
33. R. Goto, E. C. Magi, and S. D. Jackson, “Narrow-linewidth, Yb3+-doped, hybrid microstructured fibre laser operating at 1178nm,” Electron. Lett. 45(17), 877–878 (2009). [CrossRef]
34. L. Xiao, W. Jin, and M. S. Demokan, “Photonic crystal fibers confining light by both index-guiding and bandgap-guiding: hybrid PCFs,” Opt. Express 15(24), 15637–15647 (2007). [CrossRef] [PubMed]
35. Z. Wang, T. Taru, T. A. Birks, J. C. Knight, Y. Liu, and J. Du, “Coupling in dual-core photonic bandgap fibers: theory and experiment,” Opt. Express 15(8), 4795–4803 (2007). [CrossRef] [PubMed]
36. B. G. Ward, D. M. Wright, C. J. Eicholt, and C. G. Carlson, “Photonic crystal fiber with resonant-coupling higher-order-mode suppression,” Proc. SPIE 7580, 758011 (2010). [CrossRef]
37. X. Ma, C. Zhu, I. N. Hu, A. Kaplan, and A. Galvanauskas, “Single-mode chirally-coupled-core fibers with larger than 50 µm diameter cores,” Opt. Express 22(8), 9206–9219 (2014). [CrossRef] [PubMed]
38. J. M. Fini, J. W. Nicholson, R. S. Windeler, E. M. Monberg, L. Meng, B. Mangan, A. Desantolo, and F. V. DiMarcello, “Low-loss hollow-core fibers with improved single-modedness,” Opt. Express 21(5), 6233–6242 (2013). [CrossRef] [PubMed]
39. J. M. Fini, J. W. Nicholson, B. Mangan, L. Meng, R. S. Windeler, E. M. Monberg, A. DeSantolo, F. V. DiMarcello, and K. Mukasa, “Polarization maintaining single-mode low-loss hollow-core fibres,” Nat. Commun. 5(5085), 5085 (2014). [CrossRef] [PubMed]
40. F. Stutzki, F. Jansen, H. Otto, C. Jauregui, J. Limpert, and A. Tünnermann, “Designing advanced very-large-mode-area fibers for power scaling of fiber-laser systems,” Optics 1(4), 233–242 (2014).
41. P. Koška, P. Peterka, and V. Doya, “Numerical modeling of pump absorption in coiled and twisted double-clad fibers,” IEEE J. Sel. Top. Quantum Electron. 22(2), 55–62 (2016). [CrossRef]
42. F. Jansen, F. Stutzki, H. J. Otto, T. Eidam, A. Liem, C. Jauregui, J. Limpert, and A. Tünnermann, “Thermally induced waveguide changes in active fibers,” Opt. Express 20(4), 3997–4008 (2012). [CrossRef] [PubMed]
