APPLIED PHYSICS LETTERS 92, 181108 共2008兲

High peak power single frequency pulses using a short polarizationmaintaining phosphate glass fiber with a large core M. Leigh,1,2 W. Shi,1,a兲 J. Zong,1 Z. Yao,1 S. Jiang,1 and N. Peyghambarian1,2 1

NP Photonics Inc, 9030 S. Rita Road, Tucson, Arizona 85747, USA University of Arizona, 1630 East University Boulevard, Tucson, Arizona 85721, USA

2

共Received 8 February 2008; accepted 8 April 2008; published online 8 May 2008兲 We report a high power, single frequency 1.5 ␮m pulsed fiber master oscillator-power amplifier laser source using a large core, polarization-maintaining phosphate glass fiber. The phosphate glass fiber produces high powers with low nonlinearities using high Er–Yb codoping and short length. Using this fiber, we were able to generate single frequency pulses with peak powers up to 51.5 kW. This is the highest reported peak pulse power for eye safe single frequency fiber laser systems. © 2008 American Institute of Physics. 关DOI: 10.1063/1.2917470兴 High power fiber lasers have generated a great deal of interest due to their many advantages such as thermal management, efficiency, compactness, and fiber splices eliminating alignment. While fiber lasers are capable of general high power laser tasks such as machining, they are also capable of high precision tasks, such as gravitation wave detection,1 and tetrahertz generation.2 A common obstacle to the creation of fiber laser systems with both high quality and high power is the onset of nonlinearities which can greatly broaden the spectrum and modify the laser’s temporal properties, even in large-mode area fibers. Many methods have been developed to improve the quality of high power fiber lasers such as altering index profiles,3 using photonic crystal fibers,4 and suppressing stimulated Brillouin scattering5 and stimulated Raman scattering.6 Much of the research in single-mode high power fiber lasers has focused on the 1 ␮m region where Yb-doped fibers have reached kilowatt continuous wave 共cw兲 powers,7 and hundreds of watts with narrow linewidths.8 While much of the development of high powered fiber lasers has been at 1 ␮m, there is considerable interest in the 1.5 ␮m region, which is relatively eyesafe, and where there are many commercially available fiber components. The highest output power for a 1.5 ␮m fiber laser system was a multimode system9 that produced 1.2 MW peak power at 1567 nm with a 65 ␮m fiber, and an M 2 of 8.5. Single mode 1.5 ␮m systems have produced cw powers of over 100 W with single frequency operation.10,11 Pulsed single mode fiber laser systems at 1.5 ␮m have produced millijoule pulse energies12 and peak powers of 170 kW,13 but without any report of single frequency output. These systems used silica fiber, which can be problematic as Er–Yb doping often requires additives that increase numerical aperture.12 As an alternative to Er–Yb doped silica fiber, we have used an Er–Yb doped phosphate glass fiber for the final power stage of our amplifier chain. Phosphate glass fibers have higher rare-earth ion solubility than silica fibers, thus enabling shorter systems with lower nonlinearities. Phosphate fibers have previously been used to produce high cw power per unit length at 1 ␮m14 and 1.5 ␮m,15 as well as high peak pulse powers at 1.5 ␮m.16 By using a polarizationmaintaining 共PM兲, double-clad, large core, highly Er–Yb doped phosphate fiber 共LC-EYPF兲, we were able to create a兲

Electronic mail: [email protected].

0003-6951/2008/92共18兲/181108/3/$23.00

single frequency pulses at higher peak powers than previous single frequency systems. Our master oscillator-power amplifier 共MOPA兲 system consists of a single-frequency, high stability cw seed laser, a seed pulse generation system, and a three-stage amplifier system, as shown in Fig. 1. The seed laser is an NP Photonics Inc. fiber laser that utilizes highly Er–Yb doped phosphate fibers, producing a single frequency with an extremely narrow linewidth of 2 kHz,17 providing an excellent seed source. Our seed laser has a wavelength of 1550.67 nm at up to 100 mW. The seed laser is spliced to a PM isolator, which is then spliced to the first electro-optic modulator 共EOM兲 in the pulse generation system. The pulse generation system consists of an EOM to directly modulate the fiber laser, an Erdoped preamplifier to boost the power of these pulses, a narrow bandpass filter 共featuring a PM circulator and fiber Bragg grating兲 for removing out-of-band amplified stimulated emission 共ASE兲 generated by the amplifier, and a second EOM 共time synchronized to the first EOM兲 to remove the in-band ASE, increase the extinction ratio, and clean the pulse temporal profile. A low-frequency electrical function generator triggers home-made pulse generators, which are capable of generating electrical pulses from 0.7 ns to 12 ns at repetition rates up to 38 MHz. These nanosecond electrical pulses are then fed into rf driver amps, providing the electrical signal to drive the EOMs, which then modulate the seed laser. The output fiber from the second EOM was fusionspliced to an all-fiber chain of three amplifier stages sepa-

FIG. 1. Schematic of the pulsed MOPA system. The pulses from the first EOM are amplified by the preamplifier 共PA兲, filtered, then gated by EOM 2. The pulses then pass through two amplifier stages 共A1-A2兲 and the final stage with the LC-EYPF 共A3兲.

92, 181108-1

© 2008 American Institute of Physics

181108-2

Leigh et al.

Appl. Phys. Lett. 92, 181108 共2008兲

FIG. 2. Variation of average power and peak power with pulse repetition rate. The peak output occurs at 5 kHz.

FIG. 3. Measured variation of pulse energy and peak power with pump power 共points兲, and calculated fits 共lines兲.

rated by interstage commercial fiber-coupled PM band-pass filters 共for ASE rejection兲, isolators, and fiber taps 共for pulse shape, spectrum, and power diagnostics兲. The first amplifier stage consisted of 0.5 m of commercial, PM, Er-doped silica fiber with a 7 ␮m core, the second stage used 1 m of commercial, PM, double-clad Er–Yb codoped silica fiber, also with a 7 ␮m core, and the final stage being a PM LC-EYPF. Such amplifier gain staging facilitated performance optimization via improved ASE management and distribution of the thermal load,18 and the short lengths reduced nonlinear interactions. We used high power single mode pumps for the first stage, and multimode diodes for the pumping of the second stage, but kept the pump power low enough to avoid visible nonlinearities. At 20 kHz, the system produced pulses with ⬃3.5 ␮J of pulse energy, 70 mW of average power, and 356 W of peak power, with minimal pulse steeping and ⬍50 MHz linewidth, providing a high quality input into the final stage of the amplifier chain. The LC-EYPF fiber for the final power amplifier was designed and drawn in-house using phosphate fiber technology developed by NP Photonics. The fiber has a 15 ␮m core, the largest core for a single mode phosphate fiber, with an outer diameter of 122 ␮m. The core is doped with 3.01 ⫻ 1020 cm−3 of Er3+ and 14.6⫻ 1020 cm−3 of Yb3+. The bulk glass had a core index of 1.5606, a cladding index of 1.5597, and a lifetime of 7.2 ms. The fiber contains stress rods to produce a birefringence that maintains the polarization of the seed pulse. We used a length of 12 cm, which greatly reduces distortions compared to silica fiber amplifiers that can have lengths on the order of meters or tens of meters. One end of the 12 cm LC-EYPF was fusion spliced to commercial large core silica fiber, and then to its pump diodes and the second amplifier stage. The output end was angle-polished to reduce feedback. The fiber was co-pumped from the input end with multimode diodes, with a maximum total pump power for the final stage of 34 W. This pumping leads a large inversion of the final stage LC-EYPF, which also leads to pulse steepening due to gain compression. At lower powers of pump, 10 W and under, the pulse widths decreased relatively linearly with pump power. Meanwhile, above 10 W the pulse width gradually flattened out to 2 – 3 ns. Thus, for our MOPA system, in order to produce pulse widths at a few nanoseconds, it is necessary to start with pulse widths around 10 ns. To measure the power output of a laser, we used an Ophir PE9F-SH pyroelectric energy meter that handles pulse

repetition rates up to 20 kHz. The detector measures only the pulsed output, eliminating cw components such as ASE, stray pump, and in-band cw. We also used a fast optical detector, digitizing oscilloscope, and Newport power meter as a second measurement source at repetition rates below 20 kHz, and to measure the power at higher repetition rates. Both types of measurements were in good agreement. To characterize the power amplifier system, we used the maximum pump of 34 W and varied the repetition rate from 500 Hz to 1 MHz, as shown in Fig. 2. The highest peak power was 51.5 kW at a repetition rate of 5 kHz, which is an improvement by a factor of 12.6 over previous 1.5 ␮m single frequency pulsed fiber lasers.2,16,19 The highest average power was 1.66 W at 20 kHz. After setting the repetition rate to 5 kHz, we varied the pump power from 2 to 34 W. As shown in Fig. 3, both the pulse energy and peak power curves were well fit by a three parameter logistic function. The lack of a pronounced rollover indicates that higher output powers are possible. With the system operating at peak power, we measured the spectrum, linewidth, and M 2 of the output beam. The spectrum was measured with an optical spectrum analyzer, and we observed sideband suppression of ⬃40 dB. The beam maintained excellent single mode output, with an M 2 of 1.2 in each direction. The linewidth was estimated by using a scanning Fabry-Pérot interferometer with a 1.001 GHz free spectral range 共FSR兲 and a digitizing oscilloscope, as shown in Figure 4. We determined the linewidth

FIG. 4. Oscilloscope trace of Fabry-Pérot interference pattern of MOPA output 共relative intensity兲 combined with the scanning voltage 共trianglewave signal兲.

181108-3

from the FSR of the Fabry-Pérot and the envelope width of the interference fringes.2,16,20 For small pump powers and a final pulse width of 9.82 ns, the linewidth was 50 MHz, but broadened to 500 MHz when the pulse was compressed to 2.1 ns at higher pump powers. The resulting time-bandwidth products varied from 0.5 to 1, indicating near-transform limited linewidth. In conclusion, we used a PM, highly doped, large core phosphate glass fiber to produce 51.5 kW single frequency 1.5 ␮m pulses, which is a significant improvement. The performance of our demonstrated source is of special interest for nonlinear wavelength conversion, which generally requires a pump beam of high peak power and spectral brightness, linear polarization, and excellent beam quality. 1

Appl. Phys. Lett. 92, 181108 共2008兲

Leigh et al.

S. Höfer, A. Liem, J. Limpert, H. Zellmer, A. Tünnermann, S. Unger, S. Jetschke, H.-R. Müller, and I. Freitag, Opt. Lett. 26, 1326 共2001兲. 2 W. Shi, M. Leigh, J. Zong, and S. Jiang, Opt. Lett. 32, 949 共2007兲. 3 J. K. Sahu, J. Kim, S. Yoo, A. Webb, C. Codemard, P. Dupriez, Y. Jeong, J. Nilsson, D. Richardson, and D. N. Payne, Proc. SPIE 6389, 638909 共2006兲. 4 C. Brooks and F. Di Teodoro, Appl. Phys. Lett. 89, 111119 共2006兲. 5 P. Weßels, P. Adel, M. Auerbach, D. Wandt, and C. Fallnich, Opt. Express 12, 4443 共2004兲. 6 L. Zenteno, J. Wang, D. Walton, B. Ruffin, M. Li, S. Gray, A. Crowley, and X. Chen, Opt. Express 13, 8921 共2005兲.

Y. Jeong, J. Sahu, D. Payne, and J. Nilsson, Opt. Express 12, 6088 共2004兲. S. Gray, A. Liu, D. T. Walton, J. Wang, M.-J. Li, X. Chen, A. B. Ruffin, J. A. DeMeritt, and L. A. Zenteno, Opt. Express 15, 17044 共2007兲. 9 S. Desmoulins and F. Di Teodoro, Opt. Express 16, 2431 共2008兲. 10 C. Alegria, Y. Jeong, C. Codemard, J. K. Sahu, J. A. Alvarez-Chavez, L. Fu, M. Ibsen, and J. Nilsson, IEEE Photonics Technol. Lett. 16, 825 共2004兲. 11 Y. Jeong, J. K. Sahu, D. B. S. Soh, C. A. Codemard, and J. Nilsson, Opt. Lett. 30, 2997 共2005兲. 12 C. Codemard, C. Farrell, P. Dupriez, V. Philippov, J. Sahu, and J. Nilsson, C. R. Phys. 7, 170 共2006兲. 13 M. Savage-Leuchs, E. Eisenberg, A. Liu, J. Henrie, and M. Bowers, Proc. SPIE 6102, 610207 共2006兲. 14 Y. W. Lee, S. Sinha, M. J. F. Digonnet, R. L. Byer, and S. Jiang, Opt. Lett. 31, 3255 共2006兲. 15 T. Qiu, L. Li, A. Schülzgen, V. L. Temyanko, T. Luo, S. Jiang, A. Mafi, J. V. Maloney, and N. Peyghambarian, IEEE Photonics Technol. Lett. 16, 2592 共2004兲. 16 W. Shi, M. Leigh, J. Zong, Z. Yao, and S. Jiang, IEEE Photonics Technol. Lett. 20, 69 共2008兲. 17 C. Spiegelberg, J. Geng, Y. Hu, Y. Kaneda, S. Jiang, and N. Peyghambarian, J. Lightwave Technol. 22, 57 共2004兲. 18 Y. Wang, C.-Q. Xu, and H. Po, IEEE Photonics Technol. Lett. 16, 63 共2004兲. 19 C. E. Dilley, M. A. Stephen, and M. P. Savage-Leuchs, Opt. Express 15, 14389 共2007兲. 20 M. Leigh, W. Shi, J. Zong, J. Wang, S. Jiang, and N. Peyghambarian, Opt. Lett. 32, 897 共2007兲. 7 8