A method for spectral combination of lasers with extremely high spectral density is introduced, enabling greater than 80% and theoretically approaching 100% spectral density utilization with no degradation in beam quality. Experiments demonstrating the utility of our method are described, cumulating in a demonstration of a compact, packaged laser with photonic-crystal-fiber-rod amplifiers at 0.5-MW peak power and 0.15-nm wavelength spacing. Our method is potentially scalable to many 100’s of channels within the gain bandwidth of high average power or peak power rare earth doped fiber lasers at any wavelength in a compact footprint and uses only reflective optics and gratings.
©2011 Optical Society of America
In this paper we introduce the concept of chirp-precompensated spectral beam combination with extremely high spectral densities. Our method (which is the subject of several pending US patents) allows spectral combination of laser outputs with reasonable (but not necessarily perfect) beam quality with a spectral packing density of greater than 80% and theoretically approaching 100% spectral utilization. This enables use of relatively broadband lasers with reasonable beam quality to achieve very high channel counts in a spectral beam combination system, while keeping within the gain bandwidth of typical rare earth doped fiber lasers. This can ultimately allow for average power scaling with 100’s of channels of kW class CW lasers or extremely high multi-MW peak powers within relatively narrow spectral windows (< a few nm) with pulsed MW peak power class fiber lasers. This combination can be achieved with all-reflective optics, mitigating thermal effects that may plague transmissive optics based systems. In addition, the system can be implemented in relatively small footprints and in polarization independent configurations.
Past spectral beam combination (SBC) demonstrations, rely on either diffraction gratings (in reflection or transmission) or volume Bragg gratings (VBG) and are able to produce diffraction limited beams only for lasers with narrow bandwidth and precisely fixed wavelengths [1–6]. Narrowband lasers are required because in the case of SBC of broadband lasers, different spectral components in each beam are spatially dispersed by the dispersive optical element (grating or VBG), and as a consequence, the beam quality of the resulting output beam degrades. The resultant angular dispersion of the various spectral components is known as spatial chirp. In addition to spatial chirp related issues with combination of broadband lasers, when performing SBC of narrowband lasers with wavelengths that are not well controlled or drift over time, the combined beam quality also degrades due to the same spatial effects causing the individual beams to angularly point in correspondence to the drift in the laser wavelength. The requirement on keeping the wavelength highly stable and the complexity associated with mitigating Stimulated Brillouin Scattering resulting from use of narrowband fiber amplifiers makes these spectral combination options less attractive. Other SBC designs mask the spatial chirp problem by either restricting the dispersion angle to be much less than the beam divergence angle, or by widely separating wavelengths between adjacent elements relative to the overall wavelength content (linewidth and wavelength drift) [1–5]. Neither of these techniques is amenable to extremely high spectral density, high beam quality spectral beam combination for power scaling. The recently demonstrated dual-grating SBC concept also suffers from this limitation with the problem manifesting itself in the near field rather than the far field . As a result past diffraction grating based SBC techniques cannot pack wavelengths tightly together, and though VBGs have been shown to allow 0.25 nm channel spacing , they still require the individual wavelengths to be very spectrally narrow and accurately maintained resulting in lower fill-factor utilization of available spectrum. By pre-spatially-chirping beams in the opposite direction as would be applied by the combining grating to compensate exactly for the chirp it would induce, our beam combiner is insensitive to the accuracy of individual wavelengths and their linewidths, and greatly increases the spectral utilization efficiency (fill factor between wavelengths) to a large fraction of the wavelength spacing.
2. Principle of operation
The basic principle behind the chirp-precompensated SBC (CP-SBC) as Fig. 1 illustrates, consists of two identical gratings with a staircase mirror. The first grating, which we call the pre-chirp grating, simultaneously adds the correct amount of pre-chirp to all the laser beams angularly. This prechirp compensates for the chirp induced by the second (combining) grating, which combines the beams and their spatially spread wavelength components into one single beam, with little or no residual spatial chirp. Central ray propagation of the various wavelength components of three individual beams through the CP-SBC is shown in Fig. 1, with propagation from left to right. Collimated input beams are incident on the pre-chirp grating at appropriate incident angles similar to the traditional SBC input beam array (inset of Fig. 1), but with the wavelengths arranged in reverse order compared to the arrangement for single grating combination. The pre-chirp grating disperses the different wavelength beams apart and fans out the individual beams spatially based on their bandwidth. After the pre-chirp grating, a “staircase” mirror rearranges the beams between the gratings to their proper order and sets the correct spacing for beam combining at the second grating. The lenses (curved mirrors are used in practical high power systems to avoid thermal issues associated with transmissive optics) on either side of the staircase mirror form an image relay pair with the object and image planes positioned at the plane of the two gratings. The first lens thus points all of the spectrally spread wavelength content to propagate parallel and also focuses each originally collimated beam onto a different step of the staircase mirror. The bandwidth per laser that the CP-SBC passes, (i.e. the amount of spatial wavelength spread that is tolerable, which we call “passband”), is determined by the size of an individual stair step (nominally ~mm in dimension in practice) and the focused spot size (or line-focused spot size in the case of experiments using cylindrical lenses to reduce intensity on the mirror) on the staircase mirror. With arbitrarily initially large collimated beams and a tight focus on the staircase mirror, the passband can theoretically approach the wavelength spacing between elements. Practical design concerns limit the passband to between 70% and 90% of the channel spacing (i.e. the edge of two consecutive laser spectra are spaced by ~10%-30% of the spacing between channels desired). The second lens re-collimates the beams and points the beams to be incident on the combining grating at the appropriate angle for efficient combination. The appropriate pre-chirp as induced by the first grating is maintained, which is then canceled out by the second combining grating, resulting in an combined output beam with the same beam quality as individual input beams.
3. Experimental demonstrations
3.1 Proof of concept experiment
We have conducted a series of experiments to demonstrate the utility of our method. First, using a single tunable laser source we demonstrate the wide passband of the CP-SBC configuration and compare it to the traditional single grating SBC (Fig. 2 ). The proof of concept CP-SBC system in this experiment is designed to have a 50-GHz (0.188-nm at ~1064 nm) channel spacing and uses a pair of 1200-line/mm ruled gratings and a gold coated staircase mirror. Figure 2(a) is a schematic top view of the experimental setup. A single frequency distributed Bragg fiber laser which is continuously tunable over 1 nm and loosely collimated is used as the input. The laser beam has a focused spot size of 30-µm diameter located approximately 2.5 cm in front of the collimator. The slowly diverging beam passes over a “scraper mirror” and to the first collimating lens which directs the beam at the correct angle to the pre-chirp grating (upper grating in Fig. 2(a)). Both pre-chirp and combining gratings are operated at the Littrow condition. The beam diffracts from the pre-chirp grating and is pointed by moving the non-diffracting axis of the grating such that the beam travels at a slightly downward angle back through the first lens at slightly lower spot, ultimately focusing the beam onto the staircase mirror. The scraper mirror reflects the pre-chirped beam, separating it from the original laser beam and pointing it towards the staircase mirror. The beam then hits the staircase mirror at the appropriate step, is directed towards the second lens where it is re-collimated before the combining grating. The beam diffracted off of the combining grating is collimated, and in a typical CP-SBC would be the output beam (combined with several other output beams of varying wavelengths). However, in this experiment, the combined beam is allowed to pass through the second lens again, which focuses it onto a slit of width of 1.35 times the 1/e2-diameter of the far field beam. Power transmission through the slit is a measure of combined beam quality in the dispersion plane.
To simulate different channels, the fiber-laser collimator is mechanically translated across the input plane (upper left rays of Fig. 2(a)) in steps of the designed array spacing of 696 µm. At each step the laser wavelength is swept to determine the passband of the combiner. The resulting transmission through the slit discussed earlier is an indicator of the passband of the CP-SBC system. Figure 2(b) shows tuning over ± 20 GHz at a variety of input “steps” (literally steps on the staircase mirror) does not degrade the transmission through a diffraction limited slit, indicating that CP-SBC allows extremely high density spectral packing. Ripples on the flat part of the passband and unequal passbands among the channels are caused by fabrication imperfections and the loose tolerances on the first generation staircase mirror used in this experiment. To demonstrate the difference in performance between CP-SBC and other techniques, a traditional single grating combiner was set up in a configuration shown in Fig. 2(c) with the same slit. Transmission through the slit is less than 5% when the laser is tuned by 20 GHz, as shown in Fig. 2(d).
Beam quality and combined beam brightness as a function of the input beam linewidth can be calculated for the two different SBC setups used in this experiment. Figure 3 clearly shows the CP-SBC is superior, demonstrating a 4x improvement in brightness as linewidth is increased.
3.2 Demonstration with very large bandwidth lasers
Our CP-SBC technique is also applicable to spectral combination of lasers with arbitrarily large bandwidths and non-uniform spectral structures within those bandwidths, e.g. spectral beam combining of multi-kW multi-nm-linewidth fiber lasers. We demonstrate the utility of the technique in the case of very large bandwidth lasers by using the gratings and the lenses of the CP-SBC in the same configuration as the first experimental demonstration (Fig. 2(a)), but we replace the staircase mirror with physically longer steps as shown in Fig. 4(a) to allow for the greater spatial spread of the combined beam. We couple a fraction of the power from three multimode-fiber coupled 976 nm diode pumps into three single mode fibers and perform CP-SBC on the three single mode fibers with very broad, independent spectra. One of the three diode pumps is wavelength locked and has a linewidth of about 0.2 nm, while the other two pumps are highly multimode with lines spanning over 5 nm each as seen in Fig. 4(c). Traditional SBC (as in Fig. 2(c)) of the three fibers yields an output beam, seen in Fig. 4(f) that is 30 times diffraction limited in the dispersion plane, while the single output beam (Fig. 4(e)) from our CP-SBC is 1.2 times diffraction limited.
3.3 Packaged system implementation
To prove the utility of the technique in a real world, fieldable laser system, we have implemented the CP-SBC in a compact packaged laser with all-reflective optics. The CP-SBC system combines the pulsed output of photonic-crystal-fiber amplifier rods, which we reported in . Pulse width, repetition rate, and the average power per rod are 2.5-nsec, 50 kHz, and 31 W, respectively. Output pulse energy per rod of greater than 0.6 mJ is produced in this experiment. This packaged laser is currently being assembled, and will ultimately contain two polarization combined sets of 4 amplifier channels spectrally combined, but maintains a CP-SBC combiner design which is expandable to allow combination of 8 spectral channels (or 16 channels if polarization combination is also incorporated in the same CP-SBC architecture). We show in Fig. 5 the preliminary results of combining two amplifier chains of this laser. The wavelengths of the rod amplifier outputs are spaced by 0.15 nm (40 GHz), and are set by tuning the individual master oscillator seeds to the appropriate wavelengths for the individual amplifiers. The wavelength spacing is chosen in anticipation of the eventual linewidth of the self-phase modulated pulses when the rod amplifiers are operated at their intended pulse energy of 2 mJ (~2MW peak power for nanosecond pulses) . Because of the large passband of the CP-SBC, no fine wavelength control is necessary over the lifetime of the laser.
For this packaged laser, the CP-SBC is designed for compactness, high peak power, and high average power (hundreds of Watts) as well as extremely high combining efficiency. To prevent optical damage and thermal aberrations which plague transmissive optical components, only reflective optics are used in the beam combiner. A pair of multilayer dielectric gratings  with 1740 lines/mm is used for combination and easily handles the combined peak power of 0.5 MW and average power of ~60W in this demonstration with no indication of degradation due to electric field or thermal effects. Cylindrical reflective mirror pairs are used on either side of the staircase mirror so that the beams are line-focused to reduce the laser intensity at the stair-case mirror. This is an improvement compared to the higher intensity produced by the small, round spot produced by spherical optics. When aligned appropriately the use of these cylindrical optics have no impact on beam quality or CP-SBC performance while allowing higher power operation without risk of optical damage on the staircase mirror. The staircase mirror is a monolithic glass structure (shown in the picture of Fig. 5) and each step is coated with a high damage threshold, highly reflective dielectric coating. The size of the steps and height of the mirror is only determined by the arrangement of the beams required and by the spectral width and power of the various channels to avoid clipping and damage to the mirror “steps.” Beam paths are folded numerous times within the limited footprint of the CP-SBC region of the laser by way of dielectric coated mirrors. Combination efficiency is extremely high, with the overall transmission for each beam through the CP-SBC measured at 91%. The diffraction efficiency (97% per grating) accounts for the majority of the losses and the remainder are made up by the mirror reflectivity losses of 0.2% per mirror, for more than 20 total bounces required for the compact package. Figures 5(a) and 5(b) show the output spectrum and the pulse profile of the combined beam, respectively. It should be noted that the spectrum is measured with an optical spectrum analyzer and that the y-axis units are in dB. The near and far fields are well overlapped as shown by the beam profiles in Fig. 5(c) and 5(d). Accounting for the correctable common astigmatism in the near field caused by the slight misalignment of the cylindrical mirrors used within the CP-SBC, the beam quality of the combined beam is the same as the beam quality of the individual beams (M2<1.2).
In summary, we introduced a method for spectral beam combination of a large number of lasers with tight spectral spacing. The lasers are able to optimize the use of available spectrum by not requiring narrow linewidths and this acceptance of broader bandwidths allows looser tolerances on laser spectral performance compared to other SBC schemes. We validated the method in a number of experiments and implemented it in a packaged, fieldable nanosecond-pulsed-amplifier system. The CP-SBC addresses the spectral-usage shortcoming of past SBC methods, and is scalable to many 100’s of channels within the gain bandwidth of kW class Yb:fiber lasers or other rare earth doped fiber lasers, in a relatively compact footprint. For these muli-100-kW average power systems, thermal management will become the major issue, rather than concerns about linewidths of individual beams and issues associated with amplifying such narrow spectra to high powers. We demonstrated a system with only high damage threshold mirrors and reflective gratings, and we believe with proper engineering and area scaling, these optics are suitable for use in muli-100-kW systems as well as multi tens of MW peak power pulsed systems. Lastly, we point out that recent progress in the fabrication of multilayer dielectric gratings has produced gratings with greater than 99% diffraction efficiency, making thermal issues even more manageable. Also any thermal distortion of the grating does not degrade the combining efficiency to first order, and any resulting aberration is common to all beams and is correctable with a common adaptive optical element for the combined laser output.
The authors wish to acknowledge Aaron Potter, Michael Hemmat, and Mark Weber for their assistance in building the pulsed fiber amplifiers for the packaged laser system.
References and links
1. V. Smirnov, L. Glebov, D. Drachenberg, A. Jain, I. Divliansky, G. Venus, and C. Spiegelberg, “Phase Locking and Spectral Combining of Fiber Lasers by Volume Bragg Gratings,” in Fiber Laser Applications, OSA Technical Digest, paper FWB2 (2011).
2. S. J. Augst, J. K. Ranka, T. Y. Fan, and A. Sanchez, “Beam combining of ytterbium fiber amplifiers,” J. Opt. Soc. Am. B 24(8), 1707–1715 (2007). [CrossRef]
3. O. Andrusyak, V. Smirnov, G. Venus, V. Rotar, and L. Glebov, “Spectral combining and coherent coupling of lasers by volume Bragg gratings,” IEEE J. Sel. Top. Quantum Electron. 15(2), 344–353 (2009).
4. C. Wirth, O. Schmidt, I. Tsybin, T. Schreiber, T. Peschel, F. Brückner, T. Clausnitzer, J. Limpert, R. Eberhardt, A. Tünnermann, M. Gowin, E. ten Have, K. Ludewigt, and M. Jung, “2 kW incoherent beam combining of four narrow-linewidth photonic crystal fiber amplifiers,” Opt. Express 17(3), 1178–1183 (2009). [CrossRef] [PubMed]
5. T. H. Loftus, A. M. Thomas, P. R. Hoffman, M. Norsen, R. Royse, A. Liu, and E. C. Honea, “Spectrally beam-combined fiber lasers for high-average-power applications,” IEEE J. Sel. Top. Quantum Electron. 13(3), 487–497 (2007).
6. P. Madasamy, D. R. Jander, C. D. Brooks, T. H. Loftus, A. M. Thomas, P. Jones, and E. C. Honea, “Dual-grating spectral beam combination of high-power fiber lasers,” IEEE J. Sel. Top. Quantum Electron. 15(2), 337–343 (2009).
7. F. Di Teodoro, M. K. Hemmat, J. Morais, and E. C. Cheung, “High peak power operation of a 100μm-core, Yb-doped rod-type photonic crystal fiber amplifier,” Fiber Lasers VII: Technology, Systems, and Applications, Proc. of SPIE Vol. 7580, 758006 (2010).
8. B. W. Shore, M. D. Perry, J. A. Britten, R. D. Boyd, M. D. Feit, H. T. Nguyen, R. Chow, G. E. Loomis, and L. Li, “Design of high-efficiency dielectric reflection gratings,” J. Opt. Soc. Am. A 14(5), 1124–1136 (1997). [CrossRef]