We report a pulsed, fiber-amplified microchip laser providing widely tunable repetition rate (7.1–27 kHz) with constant pulse duration (1.0 ns), pulse energy up to 0.41 mJ, linear output polarization, diffraction-limited beam quality (M2<1.2), and <1% pulse-energy fluctuations. The pulse duration was shown to minimize nonlinear effects that cause temporal and spectral distortion of the amplified pulses. This source employs passive Q-switching, single-stage single-pass amplification, and cw pumping, thus offering high efficiency, simplicity, and compact, rugged packaging for use in practical applications. The high peak power and high beam quality make this system an ideal pump source for nonlinear frequency conversion, and we demonstrated efficient harmonic generation and optical parametric generation of wavelengths from 213 nm to 4.4 µm with Watt-level output powers.
© 2006 Optical Society of America
Numerous applications of lasers require nanosecond-duration pulses with high repetition rate (1–100 kHz), moderate pulse energy (0.1–10 mJ), linear polarization, and diffraction-limited beam quality. Examples include materials processing, remote physical sensing, chemical detection, and nonlinear frequency conversion. In many cases, the ability to vary the repetition rate while maintaining constant values for the other optical parameters is highly desirable. For instance, in laser marking, the pulse energy and peak power are optimized for the given material properties; varying the repetition rate with constant pulse duration and beam quality would allow the marking characteristics to be maintained while varying the feed rate or the track. Q-switched solid-state lasers (e.g., Nd:YAG) have an undesirable coupling of the repetition rate to the pulse duration and beam quality. Typically, increasing the repetition rate causes a concomitant increase in pulse duration (decrease in peak power) because the gain does not fully recover between pulses and results in a degradation in beam quality (increase in M2) because of thermal effects in the gain medium .
The goal of the present work was to develop a laser system that addresses the problems of conventional, Q-switched, solid-state lasers by providing pulses with a widely tunable repetition rate but fixed duration, pulse energy, peak power, beam quality, and polarization state. Furthermore, this source should have a simple architecture and be compact, rugged, and efficient for use in practical applications. To accomplish this goal, we seeded a mode-filtered, Yb-doped, polarization-maintaining (PM) fiber amplifier with the output of a passively Q-switched, microchip laser with a widely tunable repetition rate (4–27 kHz) and fixed pulse duration (1.0 ns).
Passively Q-switched microchip lasers provide nanosecond-duration pulses in a simple and rugged architecture. Obtaining high output pulse energy requires use of a large pump-beam -waist (to increase the excited volume) and a correspondingly high pump power (to reach the threshold pump intensity). Using this approach, pulse energies of 0.25–1 mJ have been obtained directly from microchip lasers [2, 3]. This approach has several limitations: (1) The repetition rate is restricted to relatively low values (≤1 kHz) because the average power is limited by the thermal handling capability of the microchip; (2) The pump laser has to be modulated to achieve the required low repetition rate, which lowers efficiency and increases system complexity; (3) The beam quality is often degraded by the soft focusing of the pump beam and by thermal effects in the microchip; and (4) Long microchips with undoped end-caps are often used to improve thermal management, which favors oscillation on multiple longitudinal modes and contributes to decreased stability. Consequently, a master oscillator - power amplifier configuration is often preferred , in which the pulse parameters (duration, repetition rate, beam quality) of the seed laser can be optimized, with subsequent amplification in a power amplifier. We used amplifiers based on mode-filtered, Yb-doped fibers, which allow simultaneous generation of high average and peak powers [5–7] and offer numerous practical advantages.
Because our laser system employs passive Q-switching, single-stage single-pass amplification, and cw pumping of the oscillator and amplifier, it offers high efficiency, simplicity, and compact, rugged packaging for use in practical applications. We show that our seed laser provides an optimum pulse duration for wide tuning of the repetition rate. In addition, the high output peak power and diffraction-limited beam quality of the mode-filtered fiber amplifier make this system an ideal pump source for nonlinear optical interactions: we have frequency converted this source to efficiently generate radiation from the mid-IR (via optical parametric generation) through the deep-UV (via harmonic generation) with Watt-level output powers.
2. Laser system description and results
2.1. Experimental apparatus
The 1064 nm seed laser was based on a diffusion-bonded Nd:YAG/Cr:YAG microchip with a cavity length of ~2 mm. The microchip was pumped at 808 nm by a cw, fiber-coupled diode laser whose output power was varied from 0.4 to 1.3 W. By optimizing the pump-beam focusing and microchip mount, we achieved stable operation (timing jitter <0.5%) over a wide range of repetition rate (4–27 kHz). The pulse duration and pulse energy depend primarily on the size of the pump-beam waist, not the pump power; the pump-focusing optics and mounting of the microchip were thus designed so that variation of the waist was minimized over our range of pump power. Specifically, we used a relatively large pump waist, which minimized the change in the thermal lens with pump power. In addition, we mounted the microchip in temperature-controlled holder with low thermal resistance, which minimized temperature variations. As a result, the output pulse energy and duration were constant for repetition rates between 4 and 27 kHz, with values of 1.0±0.1 ns full width at half-maximum (FWHM) and 3.2±0.3 µJ, respectively. Stable operation at lower repetition rates could be obtained using the same microchip with a modulated pump , but we employed cw pumping for simplicity in the present work. We note that our approach for achieving constant output pulse duration and energy from the microchip laser is limited to relatively low pulse energies, but this limitation is acceptable for the master oscillator - power amplifier configuration employed in the present work.
In some experiments, we employed an alternative Nd:YAG seed microchip laser (discontinued JDSU DualChip) that provided a shorter pulse duration (0.68 ns FWHM). This laser had a fixed repetition rate of 33.7 kHz and a pulse energy of 20 µJ.
The output of both seed lasers usually contained a secondary after-pulse with ~10% of the energy of the primary pulse (a common feature of microchip lasers). The secondary pulse occurred ~5–20 ns after the first pulse, had a longer duration (~2.0–4.0 ns FWHM), and was at a slightly different wavelength (0.18 nm red-shift, corresponding to a second longitudinal mode). Except where noted, all of the measurements of the amplified-pulse parameters reported below have had the contribution of the secondary pulse subtracted.
The output of the microchip laser was optically isolated and launched into the core of a double-clad, Yb-doped, PM (Panda design) fiber amplifier. The fiber (Nufern) had a core diameter of 30 µm, core numerical aperture (NA) of 0.06, round inner-cladding with a diameter of 250 µm, and length of 9.9 m. The fiber was end-pumped with the ~976 nm output of a fiber-coupled diode bar (Apollo). The pump-coupling efficiency was measured using a short piece of the gain fiber to be 85%. The Yb-doped fiber was coiled on orthogonal spools with a bend radius of 3.5 cm to suppress high-order modes via bend-loss-induced mode filtering .
The amplifier output was characterized spectrally, temporally, and spatially . Spectral measurements employed an optical spectrum analyzer (Ando) with 0.02 nm resolution. The pulse temporal profile was measured using a 10 GHz InGaAs photodiode (EOT) and a 4 GHz oscilloscope (Agilent) with a measured instrument-response function of 150 ps FWHM. Beam quality (M2) was determined by focusing the output beam, measuring the irradiance distribution using a CMOS camera (DataRay) at various positions along the beam path, and fitting the standard deviations in the two axes (x and y) with the propagation equation. Pulse-energy and average-power measurements were made using a pyroelectric detector and a thermopile, respectively (Coherent). The polarization extinction ratio (PER) was determined by measuring the energy transmitted through a calcite polarizer as a function of polarizer angle.
2.2. Laser-system performance
Figure 1 shows the output pulse energy and average power as a function of launched pump power at five repetition rates between 7.1 and 27.4 kHz. The highest pulse energy (0.41 mJ) was obtained at the lowest repetition rate, and the maximum pulse energy was lower at higher repetition rates. The average output power was largely independent of repetition rate. The maximum average power was 7.8 W. As seen in Fig. 1, the laser system could be operated at a constant pulse energy by varying the pump power as the repetition rate was varied. Pulse-to-pulse energy fluctuations were <1% at all repetition rates (dominated by the microchip laser).
Figure 2 shows the pulse duration as a function of repetition rate at six pulse energies. The pulse duration was 1.05±0.07 ns FWHM (identical to the seed laser), essentially independent of repetition rate. The peak power was ~210 kW at 0.28 mJ and ~270 kW at the maximum pulse energy of 0.41 mJ.
The decoupling of the pulse duration, pulse energy, and beam quality from the repetition rate shown in Fig. 2 is a unique feature of our architecture, which combines the tunable repetition rate and short pulses of a passively Q-switched microchip laser with the high gain, high-power capability, and diffraction-limited beam quality of a mode-filtered fiber amplifier. Although fiber amplification of microchip lasers has been reported previously at both 1064 nm [5, 7, 10, 11] and 1535 nm , these systems operated at a fixed repetition rate. More recently, frequency tuning was demonstrated over a more restricted range of repetition rates, but few details of the optical pulse parameters were provided . Decoupling of the pulse duration from the repetition rate has also been reported in actively Q-switched lasers , but this approach relies on intra-cavity nonlinear frequency conversion and thus cannot be used for the fundamental wavelength (1064 nm); furthermore, this approach requires complex electronic drive signals, and the pulse energy still varies with the repetition rate. The system described in the present work offers significantly more versatility and generality.
In all experiments, the polarization of the seed laser was aligned along the slow axis of the PM fiber. The amplifier output was highly linearly polarized, with a PER of >22 dB over the full range of repetition rate and pulse energy, as shown in Fig. 3.
M2 was measured to be <1.2 in both the x and y axes at 0.16 mJ pulse energy and 27.4 kHz repetition rate. Similar beam quality was obtained previously with this fiber amplifier and a different microchip-laser seed source . As discussed in Section 3, this diffraction-limited beam quality is critical for optical frequency conversion, as well as for applications that require tight focusing, such as micromachining or printing, or minimum beam divergence, such as remote sensing.
Spectral measurements showed the onset of nonlinear processes at a peak power of >20 kW (dependent on repetition rate), the most important being stimulated Raman scattering and four-wave mixing (FWM) . These processes resulted in only modest distortion of the pulse temporal profile at the highest energies (Fig. 2 insets). The spectral width of the amplified pulses increased almost linearly with the output energy (and thus with the peak power), independent of repetition rate. At the maximum output power, the spectral width was <0.15 nm (<40 GHz) FWHM. The spectra also exhibited broad, low-amplitude spectral features due to FWM; even though they remained more than 30 dB below the 1064 nm peak, the integrated pulse energy of these features was up to half of the total output energy. A complete experimental and theoretical study of the FWM process, including detailed spectral and temporal characterization of the amplified pulses, will be reported separately.
We observed extremely low stimulated Brillouin scattering (SBS) power: the backward-propagating power was <0.1% of the forward propagating power. Amplified spontaneous emission (ASE) accounted for <7% of the average output power at the lowest repetition rate and less at higher repetition rates. The data shown in Fig. 1 had the ASE contribution subtracted.
2.3. Impact of the seed pulse duration and spectrum
Previously published studies on high-power, pulsed fiber amplifiers have pointed out the importance of the duration of the seed pulses in controlling the nonlinear effects [5–7,10,11]. A recent study using two different microchip seed lasers showed that SBS dominated the spectral broadening in the case of long pulses (2.3 ns FWHM), whereas self-phase modulation (SPM) dominated in the case of shorter pulses (0.38 ns FWHM) . The shorter pulses also experienced larger temporal distortions during amplification.
In order to determine an “optimum” seed pulse format, we compared the temporal profile (Fig. 4) and spectral width (Fig. 5) of the amplified pulses for two different seed-pulse durations (0.68 ns and 1.05 ns FWHM), both at high repetition rate (33.7 kHz and 28.2 kHz respectively). In combination with the previous experiments employing 0.38 ns and 2.3 ns FWHM seed lasers, these results confirm that minimum distortion of the output pulse temporal and spectral profiles are obtained for seed pulses of 1 ns duration.
As seen in Fig. 4, the 1.0 ns microchip laser did not exhibit an after-pulse (because of the design of the microchip and the laser settings), whereas the 0.68 ns microchip laser had an after-pulse with a duration of 1.95 ns and a peak power ~8% that of the primary pulse. As a consequence of its much lower peak power, this after-pulse was amplified without noticeable spectral broadening (Fig. 5, right panels). With increasing pump, the relative energy of the amplified after-pulse decreased because of gain depletion by the primary pulse. In previous experiments with a non-PM fiber, we found that the very weak backward-propagating SBS pulses had spectral shifts corresponding to the wavelength of the secondary pulse, not to that of the primary pulse. This result confirms that the pulse duration of the first pulse (<1 ns) was too short to generate SBS, whereas the peak power of the amplified after-pulse was too low to efficiently generate SBS.
3. Nonlinear frequency conversion
3.1. Harmonic generation
We generated the second, third, fourth, and fifth harmonics of the amplified 1064 nm beam using an experimental setup similar to that described in Ref. 16. The amplifier output beam was passed through a narrow bandpass filter (4 nm FWHM, Semrock) centered at 1064 nm and was then focused to a 250 µm spot size (1/e 2 diameter) using a plano-convex lens with a 200 mm focal length. LBO and BBO crystals (Table 1) were located as close as possible to the fundamental beam waist. For fifth-harmonic generation, a wave plate (half-wave at 532 nm, full-wave at 355 nm) was placed prior to the final BBO crystal to rotate the polarizations to be parallel. The various wavelengths were separated by a fused-silica Pellin-Broca prism, and the powers were corrected for the measured reflection loss at each wavelength. The output powers at each harmonic are shown in Fig. 6, and the maximum conversion efficiencies are listed in Table 1.
We used a half-wave plate to adjust the polarization of the 1064 nm beam, allowing us to optimize the conversion efficiency to the second and fourth harmonics. For the third and fifth harmonics, however, no effort was made to re-optimize the polarizations of the different beams between the crystals, so that the corresponding conversion efficiencies were not optimized. Nonetheless, the results shown in Fig. 6 confirm that the high peak power and diffraction-limited beam quality of the mode-filtered fiber amplifier make this system an ideal pump source for nonlinear optical interactions. In particular, the conversion efficiency of second-harmonic generation (532 nm) was up to 74%. We note that the system employed high-reliability, angle-tuned crystals, a simple optical setup (a single lens), and a single pass through the crystal train.
3.2. Tunable optical parametric generator
We used the spectrally filtered 1064 nm pulses to pump an optical parametric generator (OPG) in order to generate tunable near-IR (λ s=1.4–1.7 µm, signal) and mid-IR (λ i=2.8–4.4 µm, idler) radiation. The OPG was based on an uncoated, 50 mm long, 0.5 mm thick, multi-grating, periodically poled LiNbO3 (ppLN) crystal [17–19]. The crystal was held at 150°C to avoid photorefractive damage. The individual gratings were 0.5 mm wide, with periods ranging from 28 µm to 32 µm in 0.5 µm steps. Translating the crystal perpendicular to the pump beam allowed the signal and idler wavelengths to be tuned in discrete steps (Fig. 7). The pump beam was focused to a 1/e 2 waist diameter of 280 µm. We separated the signal beam from the non-converted pump beam using a 45° long-pass dichroic mirror and a filter (long-pass, cut-off at 1400 nm). The idler beam was measured behind a germanium filter placed directly at the exit of the ppLN crystal.
We measured the conversion efficiency as a function of pump power at λ s=1473 nm. The threshold irradiance was 100 MW/cm2, and the pump-to-signal conversion efficiency saturated at 34% (corrected for transmission of the exit face of the crystal) for a pump irradiance of 300 MW/cm2. These values are slightly better than previously published values based on a similar crystal and pulse duration [18–20], which is likely attributable to the diffraction-limited beam quality of the present laser system. Proper focusing of the pump beam is critical for achieving maximum OPG conversion efficiency, especially at high repetition rates and long idler wavelengths, where thermal lensing due to residual absorption of the idler beam can be very detrimental . We generated signal and idler powers up to 1.8 W and 0.8 W, respectively, as shown in Fig. 7; the corresponding total efficiency was up to 55% inside the crystal.
Figure 8 shows the ratio of idler to signal output powers (P i and P s, respectively) as a function of λ s. The expected variation of this ratio, deduced from the Manley-Rowe relation and taking into account the ppLN absorption at λ s and λ i (α s and α i, respectively), is P i/P s=λ s/λ i exp[-(α s-α i)L]. The very good agreement between this equation and the measurements shows that the high peak power and high beam quality of the fiber amplifier allow optimal pumping of the ppLN crystal. We conclude that a mode-filtered fiber amplifier is an ideal pump source for high-average-power OPGs.
The signal bandwidth of the OPG depends on several parameters: the bandwidth increases when the signal and idler come closer to degeneracy, with increasing pump intensity (due to higher gain ), and with increasing divergence of the pump beam (due to non-collinear interactions ). Consequently, for a given signal wavelength, a diffraction-limited pump leads to minimum bandwidth for the generated beams. The measured signal bandwidth as a function of λ s is shown in Fig. 8. The measured bandwidths are slightly narrower than those reported in previous experiments based on microlaser pump sources with poorer beam quality [19, 20] and closely match the model taking into account the pump intensity and divergence .
We have demonstrated a versatile laser system with a rigorous decoupling of repetition rate and pulse duration over a wide range of operating conditions. We varied the repetition rate in the range of 7.1–27 kHz with a nearly constant pulse duration of 1.0 ns, pulse energy up to 0.41 mJ (~270 kW peak power), stable linear polarization state (>22 dB PER), diffraction-limited beam quality (M2<1.2), and pulse-to-pulse energy fluctuations of <1%. These laser specifications are useful for materials processing and many other applications. Further power scaling is possible by increasing the repetition rate and/or pulse energy. The maximum pulse energy at high repetition rates was limited by the available pump power, and we obtained a pulse energy as high as 0.46 mJ at 3.6 kHz. The microchip laser has recently been improved and operated at repetition rates up to 56 kHz with no change in pulse duration or stability. Finally, larger-core and/or shorter fibers will provide higher pulse energies .
The laser system employs a simple architecture (passive Q-switching, single-stage single-pass amplification, and cw pumping of the oscillator and amplifier) and is therefore suitable for use in practical applications. In particular, due to the very high peak power and diffraction-limited beam quality, this laser is an ideal pump source for nonlinear optical frequency conversion: we efficiently generated the second, third, fourth, and fifth harmonics (532–213 nm), and we demonstrated Watt-level tunable radiation in the near-IR (1.4–1.7 µm) and mid-IR (2.8–4.4 µm) in a single-pass OPG. Such systems are highly suitable for applications such as materials processing, remote sensing, chemical detection, and optical counter-measures.
References and links
1. W. Koechner, Solid-State Laser Engineering, 5th ed., (Springer-Verlag, Berlin, Germany, 1999), Chaps. 7 and 8
2. J. J. Zayhowski, C. Dill III, C. Cook, and J. L. Daneu, “Mid- and high-power passively Q-switched microchip lasers,” in Advanced Solid-State Lasers, M. M. Fejer, H. Injeyan, and U. Keller eds., Vol. 26 of OSA Proceedings Series (Optical Society of America, Washington, D.C., 1999), pp. 178–186.
3. T. Taira, Y. Matsuoka, H. Sakai, A. Sone, and H. Kan, “Passively Q-switched Nd:YAG microchip laser over 1-MW peak output power for micro drilling,” in Conference on Lasers and Electro-Optics/Quantum Electronics and Laser Science Conference and Photonic Applications Systems Technologies, Technical Digest (CD) (Optical Society of America, 2006), paper CWF6.
5. F. Di Teodoro, J. P. Koplow, S. W. Moore, and D. A. V. Kliner, “Diffraction-limited, 300-kW peak-power pulses from a coiled multimode fiber amplifier,” Opt. Lett. 27, 518–520 (2002). [CrossRef]
6. M. Y. Cheng, Y. Chang, A. Galvanauskas, P. Mamidipudi, R. Chankatoti, and P. Gatchell, “High-energy and high-peak-power nanosecond pulse generation with beam quality control in 200-µm core highly multimode Yb-doped fiber amplifiers,” Opt. Lett. 30, 358–360 (2005). [CrossRef] [PubMed]
7. R. Farrow, D. Kliner, P. Schrader, A. Hoops, S. Moore, G. Hadley, and R. Schmitt, “High-peak-power (>1.2MW) pulsed fiber amplifier,” in Fiber Lasers III: Technology, Systems and Applications, A. Brown, J. Nilsson, D. Harter, and A. Tünnermann, eds., Proc. SPIE6102, 138–148 (2006).
8. J. P. Fève, N. Landru, and O. Pacaud, “Triggering passively Q-switched microlasers,” in Advanced Solid-State Photonics, C. Denman, ed., Vol. 98 of OSA Proceedings Series (Optical Society of America, Washington, D.C., 2005), pp. 373–378.
9. J. P. Koplow, D. A. V. Kliner, and L. Goldberg, “Single-mode operation of a coiled multimode fiber amplifier,” Opt. Lett. 25, 442–444 (2000). [CrossRef]
10. C. Brooks and F. Di Teodoro, “1-mJ energy, 1-MW peak-power, 10-W average-power, spectrally narrow, diffraction-limited pulses from a photonic-crystal fiber amplifier,” Opt. Exp. 13, 8999–9002 (2005). [CrossRef]
11. F. Di Teodoro and C. Brooks, “Multistage Yb-doped fiber amplifier generating megawatt peak-power, subnanosecond pulses,” Opt. Lett. 30, 3299–3301 (2005). [CrossRef]
12. F. Imkenberg, J. Barenz, H. D. Tholl, A. Malinowski, K. Furusawa, and D. J. Richardson, “Microchip laser master-oscillator Er/Yb-doped fiber-power-amplifier emitting 158 µJ pulses with a duration of 4.5 ns,” in Conference on Lasers and Electro-Optics/Europe, Technical Digest (Optical Society of America, 2003), p. 628.
14. L. A. Eyres, J. J. Morehead, J. Gregg, D. J. Richard, and W. Grossman, “Advances in high-power harmonic generation: Q-switched lasers with electronically adjustable pulse width,” in Solid Sate Lasers XV: Technology and Devices, H. J. Hoffman and R.K. Shori, eds., Proc. SPIE6100, 349–358 (2006).
15. G. P. Agrawal, Nonlinear Fiber Optics (Academic, San Diego, CA, 1995).
16. D. A. V. Kliner, F. DiTeodoro, J. P. Koplow, S. W. Moore, and A. V. Smith, “Efficient second, third, fourth, and fifth harmonic generation of a Yb-doped fiber amplifier,” Opt. Commun. 210, 393–398 (2002). [CrossRef]
17. L. E. Myers, R. C. Eckardt, M. M. Fejer, R. L. Byer, and W. R. Bosenberg, “Multigrating quasi-phase-matched optical parametric oscillator in periodically poled LiNbO3,” Opt. Lett. 21, 591–593 (1996). [CrossRef] [PubMed]
18. M. J. Missey, V. Dominic, P. E. Powers, and K. L. Schepler, “Periodically poled lithium niobate monolithic nanosecond optical parametric oscillators and generators,” Opt. Lett. 24, 1227–1229 (1999). [CrossRef]
19. J. P. Fève, B. Boulanger, B. Ménaert, and O. Pacaud, “Continuous tuning of a microlaser-pumped optical parametric generator by use of a cylindrical periodically poled lithium niobate crystal,” Opt. Lett. 28, 1028–1030 (2003). [CrossRef] [PubMed]
20. O. Pacaud, J. P. Fève, and L. Lefort, “Mid-infrared laser source with high average power and high repetition rate,” in Advanced Solid-State Photonics, C. Denman, ed., Vol. 98 of OSA Proceedings Series (Optical Society of America, Washington, D.C., 2005), pp. 438–443.