Two-tone 1064 nm fiber amplifiers having both cold (16°C) and pump induced temperature zones co-seeded with narrow linewidth 1064 nm and broad linewidth 1040 nm photons have been shown to have a power enhancement factor between 1.6 and 1.8 relative to the optimum single-tone 1064 nm amplifier while maintaining an efficiency of 65% or greater. The output power and efficiency of 1064 nm narrow linewidth two-tone amplifiers is dependent on the length of the gain fiber, the narrow to broad linewidth seed ratio, the wavelength of the broad linewidth seed and the temperature of the gain fiber.
© 2010 OSA
A high power laser system having a diffraction limited beam is desirable for long range strategic applications. To simplify such a system composed of multiple fiber amplifier legs, it is necessary to increase the output power of the individual fiber legs. For narrow linewidth amplifiers, the major impediment to accomplishing this is Stimulated Brillouin Scattering (SBS). Numerous methods have been suggested or implemented to mitigate SBS to include large-mode area fibers , thermal gradients [2,3], stress , along with various manipulations of the acoustic properties of the core or cladding regions of the gain fiber [5,6]. A promising method to suppress SBS in the long wavelength gain region involves the usage of a tandem pumping two-stage brightness enhancing technique . Specifically, a multiplicity of 976 nm diodes are first converted to several 1010 nm single mode lasers which are then combined to pump a low area ratio double clad fiber for amplification of the 1070 nm signal. The result is a reduced nonlinear length relative to that obtained when the 1070 nm amplifier is pumped directly with 976 nm diodes. This technique has enabled SBS free amplifiers in the 1 kW regime and SRS free amplifiers at 10 kW. Another method to suppress SBS is a two-tone technique where a narrow linewidth fiber amplifier is co-seeded with a broad linewidth seed . In a co-pumped configuration under proper selection of wavelengths and seed power ratio, it has been shown theoretically that the SBS process can be mitigated and an increased output power can be obtained as a result of power transfer from the broad to the narrow linewidth tones due to laser action [9,10]. Ideally, it is desirable for the power transfer to be completed by the end of the gain fiber. For the two-tone amplifier system (976 nm pump, 1040 nm broad linewidth seed, and 1064 narrow linewidth seed) studied in this paper, ideal intensity profiles enacting complete power transfer would appear as shown in Fig. 1 . Utilization of a broad linewidth co-seed avoids the possibility of reaching the SBS threshold prematurely.
As seen in Fig. 1 below, the intensity of 1064 nm in the gain fiber tends to mimic that found for counter-pumped amplifiers. An amplifier having a 1064 nm intensity distribution as shown in Fig. 1 will have a shorter effective length for SBS and will therefore have higher output power. By two-tone seeding, the benefits of counter-pumping may be achieved using the readily available monolithic and proven components for the co-pumping system. Recently, two-tone seeding was investigated at lower powers by co-seeding an amplifier having 10 m of 10/125 PM Yb-doped double clad gain fiber with narrow linewidth 1064 nm and broad linewidth 1045 nm radiation . An increase in SBS threshold was found for the two 1045 nm/1064 nm seed ratios of 10.8 and 5.4. The increase in SBS threshold could not be quantified because the experiment was pump limited at approximately 8.5 W of 976 nm radiation. In this work, the results of experiments performed at higher powers are presented.
2. Investigation of two-tone amplifier trade space
The performance enhancement of two-tone amplifiers relative to single-tone amplifiers depends on the length of the gain fiber, the 1040 nm/1064 nm seed ratio, and the temperature distribution in the gain fiber. The parameter space associated with two-tone amplifiers was explored as fully as possible for lengths of gain fiber between 1 and 10 m. The gain fiber in both the two-tone and the comparison single-tone amplifiers was spooled in two temperature zones with the temperature zone furthest from the end of the fiber at 16°C and the temperature zone closest to the end of the fiber at a pump induced temperature. Two-tone amplifiers co-seeded with narrow linewidth 1064 nm and broad linewidth 1040 nm radiation were compared with the optimum single-tone 1064 nm amplifier to determine the power enhancement factor and the optimum operating point.
2.1 Experimental methodology
A high power fiber amplifier pumped with wavelength stabilized 976 nm radiation was co-seeded with narrow linewidth 1064 nm and broad linewidth 1040 nm signals. A schematic of the two-tone system is shown in Fig. 2 . The two seeds were first amplified to a maximum of 3-4 W prior to entering the wavelength division multiplexer (WDM). The combined power of the two seeds was limited by the WDM which was rated at 3 W. From high resolution spectra, both seed sources were found to be spectrally pure with amplified spontaneous emission power less than 40 dB below the signal and not measurable. From the WDM, the two seeds were injected into the high power amplifier and eventually into Nufern generation 7 PM 25/400 Yb-doped double-clad gain fiber. The amplifier was measured to have a beam quality given in terms of M2 of 1.1.
The same amplifier was utilized for both single- and two-tone experiments with the difference being the number of seeds injected into the WDM. For both single- and two-tone amplifiers, the unabsorbed pump was removed from the amplifier output via a dichroic, M1. For two-tone amplifiers, 1040 nm in the output was removed via a spike filter, M2, which transmitted 1064 nm and reflected virtually all other wavelengths emitted by the high power
Yb amplifier under test. For single-tone amplifiers, the following quantities were recorded for each pump level: unabsorbed pump power, 1064 nm output power, and backward tap coupler power. For two-tone amplifiers, 1040 nm output power was recorded in addition to the quantities listed above, see Fig. 3 .
In order to further enhance the suppression of SBS, the gain fiber for both single- and two-tone amplifiers was placed in two temperature zones. The temperature zone furthest from the output end of the gain fiber was cooled to a temperature of approximately 16°C. The temperature zone closest to the output end of the gain fiber was allowed develop a pump induced thermal gradient. For both single- and two-tone amplifiers, the gain fiber was placed in the two temperature zones in such a way so as to equalize the SBS gain in each temperature zone, thereby maximizing the output power that would be achieved before SBS occurred. For single-tone amplifiers, it was found that placing 57.5% of the gain fiber in the cold zone and 42.5% in the zone of pump induced temperature gradient resulted in equal SBS in the two temperature zones. For two-tone amplifiers, the intensity of 1064 nm along the length of gain fiber is different than that for single-tone amplifiers with much greater intensities occurring toward the end of the gain fiber, see Fig. 1. In order to equalize the SBS gain in each temperature zone for this case, approximately 82% of the gain fiber must be placed in the cold zone with the remainder, roughly 18%, in the pump induced temperature zone.
To determine the set of optimum operating parameters for two-tone amplifiers, various lengths of gain fiber as well as 1064 and 1040 nm seed levels were explored. Two-tone amplifiers having lengths of gain fiber between 1 and 10 m were seeded with 1064 nm seed powers that varied between 39.5 and 236 mW and with 1040 nm seed powers that varied between 295 to 2456 mW. Only seed combinations resulting in amplifiers having less than 20% 1040 nm in the output or of roughly 60% or greater efficiency were investigated. Each amplifier was run up in power until the SBS threshold was reached. This was done by monitoring the power out of the backward tap coupler. At the SBS threshold, which was defined to be the point where the SBS peak was approximately 20 dB above the Rayleigh peak, the maximum power from the amplifier was recorded. Comparison single-tone experiments, seeded with approximately 2.3 W of 1064 nm, were performed at each length of gain fiber as well.
2.2 Experimental results
2.2.1 Profile of 1040 and 1064 nm along the gain fiber
The physical processes responsible the power enhancement found in two-tone amplifiers can be illustrated by examining the power profile of the 1040 and 1064 nm along the length of the gain fiber for two distinct cases. The two-tone amplifier represented in Figs. 4a -4b is seeded with 42.5 mW of 1064 nm and 1436 mW of 1040 nm. It is representative of two-tone amplifiers that are lightly seeded in 1064 nm and heavily seeded in 1040 nm (high 1040 nm / 1064 nm seed ratio). As one can see, the power profile of 1064 nm is initially suppressed by growth of 1040 nm in the initial segment of the gain fiber. Approximately 5 m down the gain fiber, power transfer from the 1040 nm to the 1064 nm starts to accelerate with power transfer being nearly complete 10 m down the gain fiber. At lengths of gain fiber less than 10 m, power transfer is not complete resulting in an increasing amount of 1040 nm in the output of the amplifier for shorter lengths of gain fiber. The two-tone amplifier represented in Figs. 4c-4d is seeded with 785 mW of 1064 nm and 1436 mW of 1040 nm (lower 1040 nm / 1064 nm seed ratio). It is representative of two-tone amplifiers that are more heavily seeded in 1064 nm. The power profiles shown in Figs. 4c and 4d are different than those in Figs. 4a and 4b primarily because the 1064 nm is competing on a more favorable basis with the 1040 nm for the gain because of the increased seed level. The net result is that the 1064 nm tends to be less suppressed in the initial segment of the gain fiber with the power transfer between the 1040 and 1064 nm accelerating 3 m down the gain fiber. With respect to amplifier performance, the output power of the amplifier with a lower 1040 nm/1064 nm seed ratio will be less since the effective length for SBS is greater resulting in a lowered SBS threshold. In addition, at all lengths of gain fiber, the efficiency of such an amplifier will be greater since the output will contain less 1040 nm.
2.2.2 Effect of co-seeding with 1040 nm
The effect of the level of 1040 nm seeding (or the size of the 1040 nm/1064 nm seed ratio) on the SBS threshold or the power enhancement of two-tone amplifiers was investigated. Trends associated with the level of 1064 and 1040 nm in the amplifier output along with the 1064 nm amplifier efficiency are shown as a function of the 1040 nm seed level (from 0 to 2.45 W) for a fixed level of 1064 nm seed in Figs. 5a and 5b. Figure 5a shows measurements associated with a 7 m amplifier seeded with 0.485 W of 1064 nm and Fig. 5b shows measurements associated with a 9 m amplifier seeded with 0.065 W of 1064 nm. The following trends are associated with both amplifiers as the seed level of 1040 nm increases: the level of 1040 nm in the output of the amplifier increases, the output power of 1064 nm increases, and the efficiency of the 1064 nm amplifier decreases. The increase in the 1064 nm output power occurs because the 1040 nm in the gain fiber is able to more effectively suppress the intensity profile of 1064 nm in the gain fiber as the 1040 nm seed level increases resulting in a decrease in the effective length for SBS (Fig. 4a versus Fig. 4c). But, as the 1040 nm is seeded more heavily, power transfer to the 1064 nm becomes less and less complete resulting in a greater amount of 1040 nm in the output of the amplifier with a subsequent decrease in 1064 nm efficiency, Fig. 4b versus Fig. 4d. Because the amplifier shown in Fig. 5a is seeded a factor of 7.5 more 1064 nm than is the amplifier shown in Fig. 5b, the increase in 1040 nm doesn't have as great an effect since the 1064 nm continues to compete favorably for the gain. That is, the efficiency decreases only slightly from 76% to 72% and the amount of 1040 nm in the output of the amplifier increases to approximately 10% at the highest 1040 nm seed level. For the more lightly seeded amplifier shown in Fig. 5b, the amount of 1040 nm in the output of the amplifier increases to approximately 40 W and the efficiency of the amplifier drops off from roughly 82% to 55% at the highest seed level due to an inability of the 1064 nm to compete favorably for gain. What is also clear from both cases is that as the seed level of 1064 nm increases relative to the 1040 nm seed (or the 1040 nm / 1064 nm seed ratio decreases), the 1064 nm output power falls and the 1064 nm efficiency increases since the power transfer from 1040 nm to 1064 nm occurs in a shorter length of gain fiber, Fig. 4d versus Fig. 4b. As a result, the 1040 nm is unable to suppress the 1064 nm intensity profile in the gain fiber as effectively (Fig. 4c versus Fig. 4a) resulting in an increase in effective length for SBS and a decrease in the SBS threshold.
2.2.3 Optimum operating point for two-tone 1064 nm amplifier
The efficiency and output power of a 1064 nm two-tone amplifier is dependent on the seed levels for 1040 and 1064 nm as well as the length of the gain fiber for the experimental configuration described in this paper. Shown in Table 1 are the best two-tone amplifiers as a function of the gain fiber length within each efficiency range. Each amplifier was driven to the SBS threshold.
Upon examination of Table 1, a surprising trend for the experimental configuration considered is that for the higher 1064 nm amplifier efficiencies, the optimum length of gain fiber (for maximum 1064 nm power) appears to 7 m versus a length of 5 m which is optimum for single-tone amplifiers. This can be explained by the fact that the optimum 1064 nm/1040 nm seed ratio needs to be higher at the shorter lengths of gain fiber (5 and 6 m) in order to keep the amount of 1040 nm in the output low and the efficiency high, see Table 2 . This is the case since the length of gain fiber required for an efficient transfer of power from the 1040 nm to the 1064 nm decreases as the seed ratio 1064 nm/1040 nm increases. This is shown clearly in Figs. 4b and 4d. The net result is a decrease in suppression of the 1064 nm (see Figs. 4a and 4c), an increase in the effective length for SBS, and finally, a decrease in the 1064 nm SBS limited output power. Another trend in Table 1 that was shown before is that the output power of the 1064 nm tends to increase with decreasing efficiency. This is caused by a decrease in the effective length for SBS due to improved suppression of 1064 nm in the gain fiber when higher seed levels of 1040 nm are used. Finally, the best 1064 nm output power achieved from the two-tone amplifiers investigated was 88.8 W at an efficiency of 64.6% out a 5 m fiber amplifier seeded with 0.3 W of 1040 nm and 0.485 W of 1064 nm.
Table 2 shows the power enhancement achievable for two-tone amplifiers with efficiencies ≥ 70% at 1064 nm for lengths of gain fiber between 5 and 10 m. Upon examination of Table 2, it is apparent, that the best power enhancement factor, 1.6, for a two-tone amplifier relative to the optimum 70% efficient 5 m single-tone amplifier occurs when the gain fiber has a length of 7 m. (A power enhancement factor of 1.8 was found for the 5 m amplifier having an efficiency of 64.4% described above.) Relative to single-tone amplifiers at the same length of gain fiber, a maximum power enhancement factor of 2.3 was found to occur at lengths of 7 and 8 m.
2.3 Investigation of the effect of temperature of the second heating zone on efficiency and output power
An investigation was carried out on three different 7 m two-tone amplifiers having 1064 nm efficiencies in excess of 70% to see the effect of keeping the entire gain fiber cold versus having two temperature zones as described above. The seed levels of the amplifiers involved are shown in Table 3 . When the entire gain fiber was kept cold, Fig. 6a , it was found that the output power achievable prior to reaching the SBS threshold was significantly less than when the last 18% or so of the gain fiber was allowed to achieve a pump induced temperature gradient, Fig. 6b.
The output power was found to decrease, for this case, between 18 and 37.7% as the level of the 1064 nm seed increased. This is as expected since the contribution to SBS when the gain fiber is kept at one temperature occurs at one frequency versus two resulting in the SBS threshold being reached at lower 1064 nm output powers. More surprising was the fact that the efficiency of the 1064 nm amplifier was also found decrease between 7.6 and 19.8 percent when the entire gain fiber was kept cold. The decrease in 1064 nm efficiency appeared to be greatest for amplifiers seeded heaviest with 1064 nm. This may be due to temperature dependence of Yb absorption and emission cross-sections. Further investigation of this will be the subject of another paper.
For all cases, in two-tone amplifiers, when the 1040 nm seed level is increased for a fixed seed level of 1064 nm a significant increase in the SBS threshold is observed due to a decrease in the effective length for SBS. A potential downside of such amplifiers is lower 1064 nm efficiency since power transfer from the 1040 nm to the 1064 nm is typically not complete due to the greater length of gain fiber required. In this case, the SBS limited power increases but the amplifier efficiency at 1064-nm decreases. Correspondingly, in two-tone amplifiers, when the 1064 nm seed level is increased for a fixed 1040 nm seed level, typically a more complete power transfer from 1040 nm to 1064 nm occurs. The net result in the later case is an increase in the effective length for SBS which results in a decreased SBS threshold. Thus, we have a performance trade off the SBS limited power decreases but the 1064 nm amplifier efficiency increases. Finally, the properties of a 1064 nm two-tone amplifier are ultimately determined by the profiles of the 1040 and 1064 nm signals within the gain fiber.
Two-tone amplifiers comprised of shorter lengths of gain fiber (5 and 6 m) are particularly plagued by an incomplete power transfer from 1040 nm to 1064 nm leading to a significant amount of 1040 nm in the output along with decreased 1064 nm amplifier efficiency. The percentage of 1040 nm in the output of the amplifier can be decreased at all lengths by increasing the seed ratio of 1064 nm/1040 nm which has the net effect of shifting the point where the power transfer is accelerated to a shorter length of gain fiber. It is also possible to adjust the longitudinal power profiles of the 1040 and 1064 nm in the gain fiber by changing the wavelength of the broadband seed because of the spectral dependence of the absorption and emission cross-sections. An increase in the absorption cross-section or emission cross-section of the broadband seed decreases and shifts the point of power transfer from 1040 to 1064 nm to an earlier position in the gain fiber. This has the effect of increasing the effective length of SBS which in turn decreases the SBS limited 1064 nm output power and increases the 1064 nm amplifier efficiency. The converse is true for a decrease in the absorption or emission cross-sections. Any change to the wavelength of the broadband seed needs to be done carefully since the absorption and emission cross-sections of Yb have different dependencies on wavelength. For example, a broadband wavelength of 1050 nm would shift the point of power transfer to 1064 nm to a later position in the fiber whereas a broadband wavelength of 1028 nm would shift the point of power transfer to a an earlier position in the fiber. Finally, it may also be possible to control the percentage of 1040 nm in the amplifier output by adjusting the temperature of the end of the gain fiber. This latter idea will be investigated in future work.
In summary, for the two-tone amplifier configuration described above with the gain fiber in two temperature zones, the first at 16°C and the second at a pump-induced temperature, the optimum length of the gain fiber appears to be 7 m for a minimum 70% efficient 1064 nm amplifier. At a gain fiber length of 7 m, a maximum output power on the order of 78-80 W can be achieved with a resulting power enhancement factor of 1.6 relative to the optimum single-tone amplifier which occurs at a gain fiber length of 5 m. Higher output powers are achievable from a shorter gain fiber, i.e., on the order of 90 W from a 5 m gain fiber, resulting in a power enhancement factor of 1.8 but with decreased efficiency of 64.4% due to incomplete power transfer from 1040 nm to 1064-nm at the amplifier output.
The authors would like to acknowledge Dr. Iyad Dajani and Mr. Clint Zeringue for their ideas on the usage cold and pump induced temperature zones as a way to mitigate SBS.
References and links
1. J. P. Koplow, D. A. Kliner, and L. Goldberg, “Single-mode operation of a coiled multimode fiber amplifier,” Opt. Lett. 25(7), 442–444 (2000). [CrossRef]
2. D. P. Machewirth, Q. Wang, B. Samson, K. Tankala, M. O'Connor, and M. Alam, “Current developments in high-power, monolithic, polarization maintaining fiber amplifiers for coherent beam combining applications,” Proc. SPIE 6453, 64531F (2007). [CrossRef]
3. Y. Jeong, J. Nilsson, J. K. Sahu, D. N. Payne, R. Horley, L. M. B. Hickey, and P. W. Turner, “Power scaling of single frequency ytterbium-doped fiber master-oscillator power-amplifier sources up to 500 W,” IEEE J. Sel. Top. Quantum Electron. 13(3), 546–551 (2007). [CrossRef]
4. A. Wada, T. Nozawa, D. Tanaka, and R. Yamauchi, “Suppression of SBS by intentionally induced periodic residual-strain in single-mode optical fibers,” in Proceedings of the 17th ECOC, 1991, 25–28.
5. M. J. Li, X. Chen, J. Wang, S. Gray, A. Liu, J. A. Demeritt, A. B. Ruffin, A. M. Crowley, D. T. Walton, and L. A. Zenteno, “Al/Ge co-doped large mode area fiber with high SBS threshold,” Opt. Express 15(13), 8290–8299 (2007). [CrossRef] [PubMed]
6. M. D. Mermelstein, M. J. Andrejco, J. Fini, C. Headley, and D. J. DiGiovanni, “11.2 dB SBS gain suppression in a large mode area Yb-doped optical fiber,” Proc. SPIE 6873, 68730N (2008). [CrossRef]
7. B. Shiner, “Recent technical and marketing developments in high power fiber lasers,” in Tech Focus: Fiber Lasers and Amplifiers: Concepts to Applications, CLEO Europe, Munich, Germany, 2009.
8. T. Bronder, I. Dajani, C. Zeringue, and T. Shay, “Multi-tone driven high-power narrow linewidth rare earth doped fiber amplifier,” US Patent 7764720.
9. I. Dajani, C. Zeringue, T. J. Bronder, T. Shay, A. Gavrielides, and C. Robin, “A theoretical treatment of two approaches to SBS mitigation with two-tone amplification,” Opt. Express 16(18), 14233–14247 (2008). [CrossRef] [PubMed]
10. I. Dajani, C. Zeringue, and T. M. Shay, “Investigation of nonlinear effects in multitone-driven narrow-linewidth high-power amplifiers,” IEEE J. Sel. Top. Quantum Electron. 15(2), 406–414 (2009). [CrossRef]
11. C. Lu, I. Dajani, C. Zeringue, C. Vergien, L. Henry, A. Lobad, and T. Shay, “SBS suppression through seeding with narrow linewidth and broadband signals: experimental results,” Proc. SPIE 7580, 75802L (2010). [CrossRef]