Open Access
Add to BibSonomyAbstract
Lower NA in large-mode-area fibers enables better single-mode operation and larger core diameters. Fiber NA has traditionally been limited to 0.06, mostly due to the control tolerance in the fabrication process. It has been recognized recently that transverse mode instability is a major limit to average power scaling in fiber lasers. One effective method to mitigate this limit is to operate nearer to the single-mode regime. Lower fiber NA is critical in this since it allows relatively larger core diameters which is the key to mitigate the limits imposed by nonlinear effects. We have developed a fabrication process of ytterbium-doped silica glass which is capable of highly accurate refractive index control and sufficient uniformity for LMA fibers. This process is also capable of large-volume production. It is based on a significant amount of post-processing once the fiber preforms are made. We have demonstrated 30/400 and 40/400 LMA fibers with a NA of ~0.028 operating very close to the single-mode regime. The second-order mode cuts off at ~1.2μm and ~1.55µm respectively. We have also studied issues related to bend losses due to the low NA and further optimization of LMA fibers.
© 2016 Optical Society of America
1. Introduction
Fermann first established that single-mode operation in a multimode fiber is possible by careful excitation of fundamental mode at the input [1]. A couple years later Koplow et al demonstrated that improved single-mode operation can be achieved by further taking advantage of mode-dependent losses in tightly coiled multimode fibers [2]. These two studies lay the foundation for large-mode-area (LMA) fibers, which have been critical for the commercial success of high-power fiber lasers in the past decade or so. Today, fiber lasers have dominated the marking and engraving market [3] and made significant progress in industrial cutting and welding by displacing CO2 lasers [4]. There is still a significant need for further power scaling of single-mode fiber lasers to average powers in the multiple kW region for industrial micro-machining, scientific research and defense applications.
In recent years, it has been recognized that transverse mode instability (TMI) imposes a significant limit on average powers of single-mode fiber lasers. At low average powers, single-mode fiber lasers can be well behaved by careful excitation and coiling of fibers which support few modes. At high average powers, TMI can develop, leading to significant output beam quality degradation [5]. The key mechanism is suspected to be stimulated thermal Rayleigh scattering [6–9]. The periodic intensity fluctuation due to mode interference would normally be harmless in passive fibers. But in active fibers, when combined with quantum defect heating and rapid transverse heat diffusion, the traveling modes’ inference pattern can lead to traveling temperature waves. The resulting refractive index change is ideal for mode coupling, which can be significant at high average powers. This can impose a severe limit on the average powers of single-mode optical fibers. The current maximum average power of directly-diode-pumped fiber lasers is ~3kW without observing TMI [10]. The current record of 10kW from a single-mode fiber laser had to use tandem pumping by fiber lasers at 1018nm to lower quantum defect in the last amplifier stage [11].
There are currently few ways to mitigate TMI. One very effective method is to operate as close to the single-mode regime as possible [9], because TMI does not exist in the single-mode regime. This has driven kW fiber lasers to use smaller core diameters of 25μm and 20μm in recent years [12]. Further reduction of NA can move fibers even closer to the single-mode regime while maintaining acceptable core diameters. Currently fabrication tolerance has limited commercial LMA fiber NA to 0.06 (ΔN = ~1.2 × 10−3). Recently, there have been efforts to refine fabrication methods to further lower fiber NAs. Fibers with a NA of ~0.04 (ΔN = 8 × 10−4) were used for the 3kw demonstration in [10,13]. A ytterbium-doped fiber with a NA of 0.038 (ΔN = 5 × 10−4) and core diameter of 35μm was also recently demonstrated [14]. Very recently, a 52µm-core fiber with an extremely low NA of 0.025 [15] and a 50μm-core fiber with a NA of 0.02 were also reported [16]. In the latter work, the ytterbium-doped glass was made using a sol-gel technique. The measured refractive index profile had both poor average index control and large local fluctuations, indicating the still inadequate refractive index control. The glass also had high background loss, resulting in a poor efficiency of barely ~40% with regard to the absorbed pump powers.
Using a technique involving first fabrication of a large number of ytterbium-doped preforms on a Modified Chemical Vapor Deposition (MCVD) system, then extraction of the doped core glass and a repeated stack-and-draw process to homogenize the glass, we have been able to controllably make a large volume of ytterbium-doped glass with refractive index very close to that of silica. Optimized ytterbium-doped phosphosilicate glass is used with additional boron doping to lower the refractive index, enabling high doping levels and low photo-darkening [17]. Glass made using a similar process was previously used to make ytterbium-doped leakage channel fibers [18] and photonic bandgap fibers [19].
In this work, we report ytterbium-doped double-clad LMA fibers with a low NA of 0.028, i.e. ΔN = 2.7 × 10−4, using ytterbium-doped phosphosilicate glass made at Clemson. Two double-clad fibers were made. The first fiber has a geometry close to 30/400. It has an estimated second-order-mode cut-off at ~1.2μm and is nearly single mode for ytterbium fiber lasers. The second fiber has a geometry close to 40/400. The second-order-mode cut-off is estimated to be at ~1.55μm and is double-moded for ytterbium fiber lasers. High laser efficiency was demonstrated in both fibers. M2 = 1.01 was measured in the 30/400 fiber and M2 = 1.06 for the 40/400 fiber. We have studied bend loss limits and issues related further optimization. The new fabrication technique effectively eliminates the current NA limits imposed by the fabrication tolerance, opening up many new possibilities for LMA fiber designs.
2. Experiments
The two fibers are coated with low-index acrylic coating to provide a pump NA of 0.46. The 30/400 fiber has a hexagonal core which is 31μm flat-to-flat and 34μm corner-to-corner. The pump guide also has a hexagonal shape, measuring 397μm flat-to-flat and 428μm corner-to-corner. The measured pump absorption is ~3dB/m at 976nm. The 40/400 fiber has a hexagonal core which is 42μm flat-to-flat and 46μm corner-to-corner. The pump guide has a hexagonal shape with 411μm flat-to-flat and 431μm corner-to-corner. The measured pump absorption is ~4.5dB/m at 976nm. The fiber cross sections of the fibers are shown in Fig. 1. The measured refractive index profile of the 30/400 fiber was given in Fig. 2. The pixilated structure of the doped core is clearly visible, but the individual pixels are small compared to the laser wavelength. They are not expected to be a problem.
Fig. 2 (a) 2D refractive index of the 30/400 fiber, (b) refractive index scan along X axis and (c) Y axis.
The bend loss of the 30/400 fiber was measured for bend diameters of 0.8m, 0.9m and 1m. White light covering from 350nm to 1800nm (Bentham WLS100) was free-space launched into the fiber. The transmitted spectrum was recorded by an optical spectrum analyzer and compared to the transmission at loose fiber coil in order to measure the bend loss, which is shown in Fig. 3(a). The missing data is due to the ytterbium absorption. The simulated bend loss using the bend loss formulas in [20] was obtained using NA = 0.0281, i.e. ΔN = 2.7 × 10−4, and a core diameter equaling the average of the flat-to-flat and corner-to-corner core dimensions. The simulated loss fits well with the measured loss. The estimated second-order-mode cut-off wavelength is ~1.2μm and third-order-mode cut-off wavelength is ~750nm. The measured third-order-mode cut-off wavelength can be found at the loss peak at ~700nm (see Fig. 3), which is associated with its estimated cut-off, to be ~750nm. This is consistent with the result obtained from loss simulation. This 30/400 fiber is therefore very close to the single-mode regime at the wavelengths of ytterbium fiber lasers. A coil diameter of >1m needs to be used to minimize bend loss.
Fig. 3 (a) The measured loss of the 30/400 fiber with simulated bend loss using NA = 0.285 and core radius of 32.5μm and (b) The measured loss of the 40/400 fiber with simulated bend loss using NA = 0.27 and core radius of 44μm.
The bend loss of the 40/400 fiber was measured for bend diameters of 0.9m, 1m and 1.1m which is shown in Fig. 3(b). The simulated bend loss was obtained using NA = 0.027, i.e. ΔN = 2.51 × 10−4, and a core diameter equaling the average of the flat-to-flat and corner-to-corner core dimensions. The simulated loss fits well with the measured loss. The estimated second-order-mode cut-off wavelength is ~1.55μm and third-order-mode cut-off wavelength is ~973nm.
A fiber laser at ~1030nm was constructed with 4m of the 30/400 fiber with two straight cleaved ends. The laser was emitted from both ends and collected for lasing efficiency measurement. The pump source is a diode laser emitting at 976nm. The fiber was coiled at 1.9m diameter. At bend diameters below 1.5m, some leaked laser light in the cladding can be seen. The measured lasing slope efficiencies are 78% and 86% versus launched and absorbed pump powers respectively (Fig. 4(a)). The M2 measurement was done by using a CCD camera to trace the laser beam propagation through a focusing lens and then taking a M2 curve fitting. The M2 value was measured to be 1.01 (Fig. 5), reflecting a perfect Gaussian mode at the output.
Fig. 4 Measured efficiencies of fiber lasers made with (a) 4m of the 30/400 fiber coiled at 1.9m in diameter and (b) 2.5m of the 40/400 fiber coiled at 1m.
Fig. 5 Measured beam quality of the fiber laser made from the 4m 30/400 fiber coiled at 1.9m in diameter.
A fiber laser at ~1035.5nm was also constructed with 2.5m of the 40/400 fiber with two straight cleaved ends. The fiber was coiled at 1m diameter. The measured lasing slope efficiencies are 77% and 85% versus launched and absorbed pump powers respectively (Fig. 4(b)). The M2 was measured to be 1.06 (Fig. 6). The output beam appears to be a really good Gaussian mode (see Fig. 6(c)). The ability to coil the 40/400 fiber to a smaller diameter of 1m is due to its larger core diameter.
Fig. 6 Measured beam quality of the fiber laser made from the 2.5m 40/400 fiber coiled at 1m in diameter.
3. Simulations
The lower NA may result in more robust single-mode operation; but it also causes significantly increased bend loss. It is therefore important to understand the tradeoffs. The critical bend diameter, defined as when the bend loss is 0.1dB/m, was simulated for various core diameters for NA = 0.0281 (see Fig. 7(a)). A slightly smaller critical bend diameter of ~0.92m can be obtained at ~55μm core diameter for this NA. This is consistent with what was measured for the 40/400 fiber. The critical bend diameter is actually larger for larger core diameters. This is due to the dependence of bend loss on both V value and core diameter. The critical bend diameter versus core diameter was also simulated for larger NAs of 0.03 and 0.035. A much smaller critical bend diameter of <0.5m can be obtained for NA = 0.035, which has a minimum of 0.49m at ~45μm core diameter. It needs to be noted that this smaller critical bend diameter comes at the cost of operating much further away from the single-mode regime. Thus a much lower TMI threshold is expected [9].
Fig. 7 (a) Simulated critical bend diameter versus core diameter for NA = 0.0281, 0.03 and 0.035, and (b) simulated NA versus critical bend diameter for core diameters of 20μm, 30μm, 40μm and 50μm.
We have also simulated the dependence of NA on the critical bend diameter for various core diameters (see Fig. 7(b)). For critical bend diameter of 0.5m, the minimum NAs are 0.0396, 0.0359, 0.0349, and 0.0348 for core diameters of 20μm, 30μm, 40μm, and 50μm respectively.
We have also calculated the mode field diameter (MFD) at 1030nm versus core diameter for the LMA fibers with NA = 0.0281 (Fig. 8). The shaded area is the single-mode regime and the dotted line is when MFD equals core diameter. The MFD reached a minimum when the core diameter was ~22μm. The MFD increases at smaller core diameter due to a loss of core guidance. It is worth noting that the MFD is much larger than the core diameter in the single-mode regime. This is a well-known feature of the single-mode regime, where there is significant amount of power outside the core. For the 30/400 fiber, the MFD is close to the core diameter, indicating there is still a significant amount of power outside the core. Far into the multimode regime, the MFD is much smaller than the core diameter and most of the optical power is in the core. In cases where only the core is doped, the doped area has much better overlap with the fundamental mode than with the next higher-order modes near the single-mode regime. This helps to improve single-mode operation and to raise the TMI threshold.
Fig. 8 Simulated mode field diameter (MFD) at 1030nm of a LMA fiber with NA = 0.0281 versus core diameter. The shaded area is single-mode regime. The dotted line is when MFD = core diameter.
4. Conclusions
Using a specially developed fabrication process, we have shown that ytterbium-doped fiber with NA as low as ~0.028 can be realized in a controlled fashion. This allows us to make an ytterbium-doped 30/400 fiber operating very close to the single-mode regime. A M2 of 1.01 was measured, indicating a perfect Gaussian mode as would be expected from operating near the single-mode regime. Good M2 of 1.06 was also measured in an ytterbium-doped 40/400 fiber which operated deeper in the two-moded regime, but demonstrated a smaller coiled diameter of ~1m, consistent with our simulation. Both fibers demonstrated excellent lasing efficiency. This fabrication technique will enable new LMA designs with significantly improved TMI threshold. The potential tradeoff between TMI threshold and the critical bend diameter was also studied.
Acknowledgment
This material is based upon work supported in part by the U. S. Army Research Laboratory and the U. S. Army Research Office under contract/grant number W911NF-10-1-0423 and W911NF-12-1-0332 through a Joint Technology Office MRI program.
References and links
1. M. E. Fermann, “Single-mode excitation of multimode fibers with ultrashort pulses,” Opt. Lett. 23(1), 52–54 (1998). [CrossRef] [PubMed]
2. J. P. Koplow, D. A. V. Kliner, and L. Goldberg, “Single-mode operation of a coiled multimode fiber amplifier,” Opt. Lett. 25(7), 442–444 (2000). [CrossRef] [PubMed]
3. “Laser Marketplace 2015,” Laser Focus World (2015).
4. D. Larcombe, “Fiber versus CO2 laser cutting,” Industrial Laser Solutions, November 2013.
5. T. Eidam, S. Hanf, E. Seise, T. V. Andersen, T. Gabler, C. Wirth, T. Schreiber, J. Limpert, and A. Tünnermann, “Femtosecond fiber CPA system emitting 830 W average output power,” Opt. Lett. 35(2), 94–96 (2010). [CrossRef] [PubMed]
6. A. V. Smith and J. J. Smith, “Mode instability in high power fiber amplifiers,” Opt. Express 19(11), 10180–10192 (2011). [CrossRef] [PubMed]
7. B. Ward, C. Robin, and I. Dajani, “Origin of thermal modal instabilities in large mode area fiber amplifiers,” Opt. Express 20(10), 11407–11422 (2012). [CrossRef] [PubMed]
8. K. R. Hansen, T. T. Alkeskjold, J. Broeng, and J. Lægsgaard, “Thermally induced mode coupling in rare-earth doped fiber amplifiers,” Opt. Lett. 37(12), 2382–2384 (2012). [CrossRef] [PubMed]
9. L. Dong, “Stimulated thermal Rayleigh scattering in optical fibers,” Opt. Express 21(3), 2642–2656 (2013). [CrossRef] [PubMed]
10. V. Khitrov, J. D. Minelly, R. Tumminelli, V. Petit, and E. S. Pooler, “3kw single-mode direct diode pumped fiber laser,” Proc. SPIE 8961, 89610V (2014). [CrossRef]
11. V. Fomin, M. Abramov, A. Ferin, A. Abramov, D. Mochalov, N. Platonov, and V. Gapontsev, “10 kW single-mode fiber laser,” in 5th International Symposium on High-Power Fiber Lasers and Their Applications, St. Petersburg, June 28-July 1, (2010).
12. K. Brar, M. Savage-Leuchs, J. Henrie, S. Courtney, C. Dilley, R. Afzal, and E. Honea, “Threshold power and fiber degradation induced modal instabilities in high power fiber amplifiers based on large mode area fibers,” Proc. SPIE 8961, 89611R (2014). [CrossRef]
13. C. Hupel, S. Kuhn, S. Hein, N. Haarlammert, J. Nold, F. Beier, B. Sattler, T. Schreiber, R. Eberhardt, and A. Tünnermann, “MCVD Based Fabrication of Low-NA Fibers for High Power Fiber Laser Application,” in Advanced Solid State Lasers, OSA Technical Digest (online) (Optical Society of America, 2015), paper AM4A.2.
14. D. Jain, Y. Jung, P. Barua, S. Alam, and J. K. Sahu, “Demonstration of ultra-low NA rare-earth doped step index fiber for applications in high power fiber lasers,” Opt. Express 23(6), 7407–7415 (2015). [CrossRef] [PubMed]
15. V. Petit, R. P. Tumminelli, J. D. Minelly, and V. Khitrov, “Extremely low NA Yb doped preforms (<0.03) fabricated by MCVD,” Proc. SPIE 9728, 97282R (2016). [CrossRef]
16. W. Xu, Z. Lin, M. Wang, S. Feng, L. Zhang, Q. Zhou, D. Chen, L. Zhang, S. Wang, C. Yu, and L. Hu, “50 μm core diameter Yb3+/Al3+/F− codoped silica fiber with M2<1.1 beam quality,” Opt. Lett. 41(3), 504–507 (2016). [CrossRef] [PubMed]
17. S. Suzuki, H. A. McKay, X. Peng, L. Fu, and L. Dong, “Highly ytterbium-doped silica fibers with low photo-darkening,” Opt. Express 17(12), 9924–9932 (2009). [CrossRef] [PubMed]
18. F. Kong, G. Gu, T. W. Hawkins, J. Parsons, M. Jones, C. Dunn, M. T. Kalichevsky-Dong, K. Wei, B. Samson, and L. Dong, “Flat-top mode from a 50 µm-core Yb-doped leakage channel fiber,” Opt. Express 21(26), 32371–32376 (2013). [CrossRef] [PubMed]
19. G. Gu, F. Kong, T. Hawkins, J. Parsons, M. Jones, C. Dunn, M. T. Kalichevsky-Dong, K. Saitoh, and L. Dong, “Ytterbium-doped large-mode-area all-solid photonic bandgap fiber lasers,” Opt. Express 22(11), 13962–13968 (2014). [CrossRef] [PubMed]
20. A. W. Snyder and J. Love, Optical Waveguide Theory (Chapman and Hall, 1983).
