Open Access
Add to BibSonomyAbstract
Robust fundamental mode propagation and amplification of picosecond pulses at 1.56µm wavelength is demonstrated in a core-pumped Er fiber with 1170µm2 effective area. Record peak power exceeding 120 kW, and 67 nJ pulse energy are achieved before the onset of pulse breakup. A small increase in input pulse energy results in a temporal collapse of the pulse center to 58 fs duration, with peak powers approaching 200 kW.
©2008 Optical Society of America
1. Introduction
High peak power fiber amplifiers are limited by nonlinearity due to the confined mode effective area (Aeff). Increasing Aeff in a single-mode fiber is limited by bend loss, whereas increasing Aeff in a multimode fiber makes the fiber more susceptible to microbend induced coupling of the fundamental mode to higher order modes which deteriorates the output mode quality. Therefore practical conventional step index multimode fibers have been limited to Aeff < 1000 µm2. Complex fiber designs are being pursued to increase the Aeff beyond 1000 µm2. These include the “rod-like” microstructured fibers [1], “leaky-channel” fibers [2, 3], and higher-order-mode fibers [4]. In this paper we show that core pumping a conventional large mode area (LMA) Er doped fiber with Aeff=1170µm2, such that both pump and signal propagate in the fundamental mode enables robust diffraction limited amplification. The large Aeff of the fiber enables direct amplification of picosecond pulses to > 120 kW before the onset of pulse breakup due to nonlinearities. This is approximately twice the peak power reported previously with a similar amplifier employing a Aeff=870 µm2 Er fiber [5, 6].
2. Fiber properties
We fabricated an Er doped fiber with 54 µm core, and 200 µm clad diameter. The index profile was measured interferometrically and showed good azimuthal symmetry which is necessary to ensure a large Aeff. The calculated fundamental mode (Fig. 1) had Aeff=1170 µm2. The core absorption was measured to be 27 dB/m at the pump wavelength of 1480 nm. This fiber has approximately 30% larger Aeff and greater than twice the doping concentration of the fiber used in Refs. [5, 6], which enables us to extract higher gain from a shorter length of fiber with lower nonlinearity.
3. Characterization of mode properties of Er gain module
The LMA amplifier gain module consisted of a standard single mode fiber pigtailed to the LMA Er fiber. It was core pumped with a 1480 nm Raman fiber laser. The splice between the single mode and LMA fiber was through a proprietary fiber based mode transformer that matched the fundamental modes between the two fibers at the signal and pump wavelengths. The launch efficiency into the fundamental mode of the LMA fiber was measured to be > 98%.
Coupling into higher order modes at the LMA fiber input due to an imperfect launch into the fundamental mode of the LMA Er fiber forms a source of discrete multi-path modal interference (MPI) at the fiber output. In addition, microbends and perturbations in the core can induce coupling of light from the fundamental mode to higher order modes along the length of the fiber giving rise to distributed MPI. The discrete and distributed MPI have been shown to severely impact the focusing and beam pointing stability of the output beam [7, 8]. It is therefore crucial to ensure single mode operation of the amplifier where MPI due to higher order modes is significantly suppressed.
The near field image of the Er amplifier is measured to have a uniform Gaussian-like mode profile (inset of Fig. 2). An M 2 < 1.1 was measured at all levels of amplification (cf. Fig. 2), demonstrating that the output of the fiber diffracts like a Gaussian beam.
We further analyzed the modal content of our amplifier output using a new Spatially and Spectrally resolved mode imaging technique, called S2 [8]. The spectral interference due to mode beating is recorded by scanning a single-mode-fiber probe coupled to an optical spectrum analyzer across the magnified image of the output beam. From the data, the various higher order modes contributing to the MPI can be imaged and their magnitude can be calculated. We used a broadband Er amplified spontaneous emission (ASE) source to seed the amplifier module and carried out the S2 measurement at moderate pump levels. Figure 3 shows the images of the residual higher order modes due to discrete MPI retrieved from this measurement. The LP02 and LP11 modes were the dominant higher order modes with strengths 22 dB and 25 dB below the fundamental mode. The larger strength of LP02 versus LP11 is probably due to the radial symmetry of input coupling which mainly excites even modes. The M2 and S2 measurements therefore confirm robust single mode excitation and propagation in the LMA Er fiber.
Fig. 3. Residual LP11 (a), and LP02 (b), mode intensity images retrieved from a S2 measurement at the output of the LMA Er amplifier.
We modeled the gain of the LP01, LP11, LP02, LP12, LP03, and LP13 modes using a two level rate equation. The gain fiber cross-section was divided into a rectangular mesh with each cell treated as a waveguide whose gain was determined by the local overlap of the signal with the inverted ions. The model accounted for the overlap of modes with the radial inverted ion population distribution created by the pump in the fundamental mode. It fully accounted for spatially dependent gain and saturation effects – however the effect of mode competition in the presence of multiple modes was not calculated. The pump power was 7 W, the input cw signal average power was 2.4 mW, and the gain length was 2 m–these conditions were chosen to be similar to those in our picosecond pulse amplification experiments presented in the next section. The gain per unit length as a function of fiber length is plotted for the various modes in Fig. 4. The gain of the LP01 mode was ~3.6 dB/m higher compared to the LP11 mode, and ~8.8 dB/m higher compared to the LP02 mode. The difference in gain increases with increase in mode order. The number of ASE modes supported in the amplifier is thus greatly reduced which enables large gains approaching 30 dB before limitation due to ASE growth. The differential gain also stabilizes the fundamental mode against transfer of power to higher order modes due to distributed MPI [9, 10]. This is an important reason why we observe stable amplification of the fundamental mode in our system.
4. Picosecond pulse amplification
We exploited the large Aeff of the Er fiber to study direct amplification of picosecond pulses to high pulse energies. The phase dispersion of step index large mode area fibers is dominated by the anomalous material dispersion in the 1.5µm wavelength region. At high intensities, the interplay of anomalous fiber dispersion with the Kerr nonlinearity results in higher order soliton-effect pulse compression which leads to the temporal collapse and breakup of the pulse [11]. This is a major limitation to directly amplifying short pulses to high energies with LMA fibers. In contrast, amplification in LMA fibers at ~1µm wavelength occurs in the “similariton” regime where normal material dispersion and Kerr nonlinearity act together to induce a linear chirp on the pulse that can be reversed in an external pulse compressor to retrieve high quality pulses. At 1.5µm wavelength, increasing the mode area and gain per unit length are the only available paths to further increasing pulse energy and power.
Our laser-amplifier setup is similar to that employed in Ref. [6] and its schematic is shown in Fig. 5. Picosecond pulses were generated in a modelocked erbium fiber laser that used carbon nanotubes as saturable absorbers [12]. It was followed by a single mode Er fiber amplifier which delivered pulses of 1.06-ps duration with 102-pJ energy at a repetition rate of 22 MHz. The spectrum, centered at 1562 nm, had a 2.4 nm full width at half maximum (FWHM). The pulses were coupled with a 7-W Raman fiber laser at 1480-nm using a 1480/1550 nm fused fiber coupler. The coupler output was launched into the LMA Er fiber using the proprietary mode transformer which, as mentioned previously, coupled the pump and signal into the fundamental mode. The LMA Er fiber was loosely coiled with a minimum diameter of ~40 cm with the last 0.5 m, which sees the highest intensity, held straight. This minimizes nonlinearity due to bend induced reduction in Aeff [13].
Amplification was studied in a 2 m long Er fiber section, where the length was optimized to maximize pulse energy and minimize soliton pulse compression. The input pulse energy, Ein, was fixed at 102 pJ for which the highest peak powers were obtained without pulse breakup. The output power, spectrum, and pulse autocorrelation were measured as a function of pump power. The change in average output power and pulse energy versus 1480 nm pump power is shown in Fig. 6(a). A maximum gain of 28 dB was achieved with 67 nJ pulse energy, and 1.5 W average power without any pulse break up. The slope efficiency was ~28%.
The pulse spectra (cf. Fig. 6(d)) showed strong broadening with increase in pulse energy, Ep, due to self phase modulation. The pulse duration shortened with increase in Ep due to pulse steepening, and the interplay between the anomalous fiber dispersion, and self phase modulation. The autocorrelation FWHM decreased from 1362 fs to 850 fs (corresponding to a compression by a factor of 1.6) with increase in pulse energy from 1.5 nJ to 67 nJ (squares in Fig. 6(c)). Figure 6(b) compares the autocorrelations recorded at the minimum (circles) and maximum (dashes) Ep. Although the autocorrelation narrowed, no pedestal was observed to indicate the onset of pulse break-up. A deconvolution factor of 1.62 was calculated from the pulse spectrum (assuming a uniform spectral phase), to relate the autocorrelation FWHM to the pulse FWHM. A minimum pulse duration of ~526 fs was obtained at the maximum pulse energy of 67 nJ, which implies a pulse peak power of 127 kW (defined as pulse energy divided by pulse FWHM, since the exact pulse shape is unknown). The autocorrelation of a 526 fs duration Sech pulse is broader at the peak and narrower at the base when compared to our measured trace. Therefore whereas our estimate of the pulse duration is not exact, it is a reasonable compromise. The >100 kW peak power is twice that reported earlier in Ref. [6], and is a record for the highest peak power picosecond pulses from a conventional fiber amplifier, prior to pulse break-up.
Increasing the input pulse energy by a few percent was sufficient to induce temporal collapse of the pulse due to the higher order soliton compression effect which was accompanied by pulse breakup, as evidenced by the formation of pedestals in the autocorrelation. The decrease in autocorrelation FWHM with output pulse energy is shown for the case when the input pulse energy was increased to 107 pJ (circles in Fig. 6(c)). The temporal collapse is much more dramatic (compression factor ~11), with a minimum pulse duration of 58 fs recorded for the central peak of the autocorrelation (solid line in Fig. 6(b)) at maximum output pulse energy. The spectrum at maximum pulse energy is shown in Fig. 6(d) (solid line). The 107 pJ input pulse energy was optimal for inducing the maximum collapse, with larger or smaller input energies resulting in longer duration central lobes. By comparing the second harmonic intensity we estimate that with Ein=107 pJ, the maximum amplified pulse intensity is ~60% higher at ~200 kW when compared to the case with Ein=102 pJ. Nonlinear frequency conversion could possibly be used to isolate the narrow pulse center from the pedestal to generate frequency shifted clean ultrashort pulses with a few nJ of pulse energy. A frequency doubled output would be an attractive replacement for Ti:sapphire oscillators.
Fig. 6. (a) Variation of average power and pulse energy versus pump power. (b) Intensity autocorrelation measured at low amplification (circles), and maximum amplification with E[1] in=102 pJ (dashes), and E[2] in=107 pJ (solid line). (c) Variation of autocorrelation FWHMwith output pulse energy, for E[1] in=102 pJ (squares), and E[2] in=107 pJ (circles). (d) Spectra recorded at low amplification (circles), and maximum amplification with E[1] in=102 pJ (dashes), and E[2] in=107 pJ (solid line).
5. Summary
Robust propagation and amplification of light in a diffraction limited fundamental mode with Aeff ~1200 µm2 in a conventional fiber has been demonstrated for the first time. This is enabled by the amplifier architecture where both the pump and signal beams are excited, and propagate, in the fundamental mode so that the perfect overlap between them results in differential gain for the fundamental mode while discouraging ASE and higher order modes.
The large Aeff has enabled direct amplification of ps pulses to a record 67-nJ pulse energy, and ~127-kW peak power, without pulse break-up. Slight increase in input pulse energy results in a temporal collapse of the pulse center with durations < 60 fs. Our system is a simple, all-fiber, alignment free source for ultrashort pulses with sufficient energy for direct use in nonlinear spectroscopy, microscopy, and imaging applications.
References and links
1. J. Limpert, A. Liem, M. Reich, T. Schreiber, S. Nolte, H. Zellmer, A. Tünnermann, J. Broeng, A. Petersson, and C. Jakobsen, “Low-Nonlinearity Single -Transverse-Mode Ytterbium-Doped Photonic Crystal Fiber Amplifier,” Opt. Express 12, 1313–1319 (2004). [CrossRef] [PubMed]
2. L. Dong, J. Li, and X. Peng, “Bend-Resistant Fundamental Mode Operation in Ytterbium-Doped Leakage Channel Fibers with Effective Areas Upto 3160 µm2,” Opt. Express 14, 11512–11519 (2006). [CrossRef] [PubMed]
3. L. Dong, J. Li, H. McKay, A. Marcinkevicius, B. Thomas, M. Moore, L. Fu, and M. Fermann, Proceedings of Conference on Lasers and Electro-Optics, San Jose (2008) (Postdeadline) (Optical Society of America, Washington DC, 2008).
4. S. Ramachandran, J. W. Nicholson, S. Ghalmi, M. F. Yan, P. Wisk, E. Monberg, and F. V. Dimarcello, “Light Propagation with Ultralarge Modal Areas in Optical Fibers,” Opt. Lett. 31, 1797–1799 (2006). [CrossRef] [PubMed]
5. J. Jasapara, M. J. Andrejco, A. D. Yablon, J. W. Nicholson, C. Headley, and D. DiGiovanni, “Picosecond Pulse Amplification in a Core Pumped Large-Mode-Area Erbium Fiber,” Opt. Lett. 32, 2429–2431 (2007). [CrossRef] [PubMed]
6. J. C. Jasapara, M. J. Andrejco, J. W. Nicholson, A. D. Yablon, and Z. Várallyay, “Simultaneous Direct Amplification and Compression of Picosecond Pulses to 65-KW Peak Power Without Pulse Break-Up in Erbium Fiber,” Opt. Express 15, 17494–17501 (2007). [CrossRef] [PubMed]
7. S. Wielandy, “Implications of Higher-Order Mode Content in Large Mode Area Fibers with Good Beam Quality,” Opt. Express 15, 15402–15409 (2007). [CrossRef] [PubMed]
8. J. W. Nicholson, A. D. Yablon, S. Ramachandran, and S. Ghalmi, “Spatially and Spectrally Resolved Imaging of Modal Content in Large-Mode-Area Fibers,” Opt. Express 16, 7233–7243 (2008). [CrossRef] [PubMed]
9. D. Gloge, “Optical Power Flow in Multimode Fibers,” Bell Syst. Tech. J. 51, 1767–1783 (1972).
10. M. E. Fermann, A. Galvanauskas, and G. Sucha, Ultrafast Lasers-Technology and Applications (Marcel Dekker, New York, 2003).
11. G. P. Agrawal, in Nonlinear Fiber Optics (Academic Press, 1995), Chap. (5) Optical Solitons, (11) Fiber Amplifiers.
12. J. W. Nicholson, R. S. Windeler, and D. J. DiGiovanni, “Optically Driven Deposition of Single-Walled Carbon-Nanotube Saturable Absorbers on Optical Fiber End-Faces,” Opt. Epxress 15, 9176–9183 (2007). [CrossRef]
13. J. M. Fini, “Bend-Resistant Design of Conventional and Microstructure Fibers with Very Large Mode Area,” Opt. Express 14, 69–81 (2006). [CrossRef] [PubMed]
