In this paper we introduce a simple scheme to spectrally combine four single beams using three low-cost dielectric interference filters as combining elements. 25 ns pulses from four independent and actively Q-switched fiber seed-sources are amplified in a single stage fiber-amplifier. Temporally and spatially combined 208 W of average power and 6.3 mJ of pulse energy are obtained at two different repetition frequencies. A detailed analysis of beam quality as well as the thermal behavior of the combining elements is carried out and reveals mutual dependency.
©2009 Optical Society of America
The rapidly growing market of fiber lasers for material processing applications is based on a demand for reliable power-scalable laser sources. Although the history of Yb-doped high power fiber lasers is fairly young, huge achievements in scaling continuous-wave power of single-transverse mode fiber lasers have been reported [1–3]. Properties like high optical efficiencies and beneficial heat dissipation coming along with a compact and robust design made this tremendous increase in output power just possible. But even such impressive results do not mark the theoretical limits of fiber based continuous-wave lasers and amplifiers . While there seems to be room for technical development in the case of continuous-wave operated systems, pulsed fiber-based laser sources face severe physical limitations.
Inherent to the light-guiding nature of the microscopic fiber-core, spatially confined and therefore intense pulses may evoke nonlinearities during propagation. Effects like self-phase modulation (SPM), stimulated Raman scattering and four-wave mixing reduce the potential of such systems. With less doping-concentration and hence less active material than their solid-state counterparts, state-of-the-art amplifier fibers are restricted in terms of extractable pulse energy . Despite the low pulse energy, the energy density is very high in the core region, which can lead to a destruction of the fiber end-facet. To a certain extent, these limitations can be overcome by the use of large-core and short-length photonic crystal fibers (PCF), which are designed to avoid nonlinear effects and end-facet damage while increasing the energy storage capability [6–10].
Beside pondering on new fiber designs and thus improving the performance of single amplifier emissions, other techniques of scaling have to be taken into consideration. A promising approach is spectral beam combination, which is successfully demonstrated not only for continuous-wave fiber amplifiers [11,12] but for ns-pulsed systems as well [13,14]. The idea behind this method is to spatially combine a certain number of single beams by means of a wavelength-selective element. Up to now, dielectric reflection gratings have been the first choice since they are capable of handling high peak and average powers with combining efficiencies up to 99% in optimal configuration . However, such gratings are very expensive and the quality of the combined beam is connected with a narrow spectral bandwidth, which can be difficult to sustain for fiber-amplified ns-pulses due to SPM-induced spectral broadening . Recently, dielectric thin film steep-edge interference filters were used as a spatial-dispersion-free combining element to avoid deterioration of beam quality . A four-stage master oscillator power amplifier provides 2-ns pulses and is spectrally separated by filters into three beams, amplified and recombined. A combined average output power of 52 W and 1.9 mJ of pulse energy at different pulse repetition frequencies were obtained with a combining efficiency of ~90%. However, no beam quality factor was measured to highlight the advantage of interference filters as combining element versus competing dielectric gratings.
In this contribution, we introduce a simple scheme to spectrally combine four single beams using three off-the-shelf dielectric interference filters as combining elements. Nanosecond pulses from four independent and actively Q-switched fiber seed-sources, conveniently superposed in time by the use of an electronic delay, are amplified in a single-stage. The experiment is conducted at two different repetition frequencies. At 50 kHz, 208 W of combined average output power and more than 4 mJ of pulse energy are obtained while up to 6.3 mJ can be extracted at a lower repetition rate. To evaluate the usability of interference filters as combining element, beam quality and thermal behavior are monitored during the measurements. As a consequence of absorption within the filter, thermal induced wave-front distortions are causing a degradation of beam quality of the combined output beam. Arising limitations are discussed in detail.
2. Spectral beam combination: experimental results
2.1 Single laser system: Q-switched oscillator and fiber amplifier
The experiment is divided into two parts. The first part is about scaling average power at a higher repetition rate (50 kHz). In the second part the repetition rate of the whole system has been reduced down to 10 kHz with the intention to increase pulse energy. However, for both experiments the setup remains the same.
Each of the four single ns-pulsed fiber amplifiers is set up as illustrated in Fig. 1 . Basically, a single channel consists of an actively Q-switched oscillator and a high power fiber amplifier. The cavity of the oscillator is formed by a high reflective (HR) mirror and the 4% Fresnel reflection of the perpendicularly polished fiber end-facet. A short piece of Yb-doped 6 µm core single-clad (SC) fiber is core-pumped by a single-mode (SM) laser diode (LD) at 976 nm wavelength by means of a wavelength division multiplexer (WDM). Because of the high doping concentration and hence the high gain of the active fiber, the AOM-sided (AOM = acousto-optical modulator) end-facet has to be angle-polished in order to avoid parasitic lasing. A 1450 lines/mm transmission grating in combination with the HR-mirror is used to keep the spectral bandwidth narrow (< 0.5 nm) and to provide different central wavelengths for each oscillator, which is essential for the concept of spectral beam combining. Q-switched operation is achieved by implementing an AOM driven by a set of electronics. The digital function generator provides the right repetition frequency while the analog pulse generator can electronically delay the pulse build-up to compensate for differences in optical paths. Due to the fact that no polarization-maintaining fiber is used, a polarization beam splitter (PBS) has to be inserted inside the cavity to stabilize the polarization-state and avoid fluctuations of output power after the 40 dB optical Faraday isolator. A slit in front of the grating is used to block the zero diffraction order of the AOM to assure first order operation only.
The amplifier fiber is a 1.2 m long rod-type photonic crystal fiber with an ytterbium-doped 85 µm active core and a measured mode-field-diameter of the fundamental transverse mode of 71 µm at 1030 nm wavelength . The corresponding mode-field area of ~4000 µm2 minimizes nonlinear effects and eases the damage threshold by reducing the power density in the core. Pump light from the multi-mode (MM) laser diode is coupled efficiently to a 200 µm pump-core, which has a very high NA of 0.58 because of an embedded air-clad structure. The large active core can guide a few transverse modes, but stable fundamental mode operation (M2 ~1.2) can be achieved under proper signal mode matching conditions. At a pulse repetition rate of fREP = 50 kHz, the amplifier provides an output power of Pout > 50 W and a pulse energy of > 1 mJ.
2.2 Temporal and spatial combination
The concept of spectral beam combination depends on a wavelength-selective element. For this purpose, three small-bandwidth off-the-shelf interference filters (IF) are used to combine the four beams spatially as shown in Fig. 2a . The dielectric layer of this IF is ion-beam sputtered on a 2 mm N-BK7 substrate, has a high damage threshold of 100 mJ/cm2  and therefore is well suited for applications in the ns-regime. With a transmission window of ~3 nm and steep edges (Fig. 2b), a beam within this spectral range can pass the filter while another beam with a different wavelength outside this window is reflected. By tilting the IF the center of the window is shifted to shorter wavelengths and thereby adapted to the wavelength of the transmitted beam.
Once the right angle of incidence of the reflected beam is chosen, both beams are superposed in the near- and the far-field. The first two interference filters (IF1-2 and IF3-4) are readily adjustable since each beam contains only one wavelength. In the case of IF12-34, it is somewhat more complicated due to two wavelengths λ(B3), λ(B4) that have to pass the filter. For a high overall combining efficiency, a trade-off between loss of the transmitted and reflected beam has to be found. During the experiment, the overall efficiency of the interference filters is measured to remain constant at about 86% for all power levels.
To find the right temporal superposition of the pulses, differences in optical paths have to be pre-compensated for each channel by assigning an electronic delay to the AOM gating signal and hence shift the pulse build-up as illustrated in Fig. 3 . Once the right delay is chosen, the separated pulses of the beams B1 to B4 are overlapping temporally and pulse energy and pulse peak power are scaled up by a factor of four. Since the pulses have a relatively long duration of 25 ns (at fREP = 50 kHz), timing jitter induced pulse broadening of the combined pulse is no issue. The inset in Fig. 3 shows the different wavelengths of the beams covering a spectral range from 1027 to 1035.5 nm.
2.3 Scaling average power at 50 kHz repetition rate
In the first part of the experiment, the repetition frequency is set to 50 kHz. At this repetition rate and with the available SM pump power the Q-switched oscillator produces 25 ns pulses with an average output power of 100 mW. The photonic crystal fiber has a gain of > 500 and amplifies the average output power up to > 50 W. The laser slope of the spatially and temporally combined beam (Fig. 4 , left scale) has an efficiency of ~30% and shows only a slight roll-over effect caused by gain saturation in the applied main amplifier PCF . At maximum pump power the average output power is as high as 208 W, which corresponds to 4.2 mJ of pulse energy.
A small amount of the power is absorbed by the interference filters leading to a growth of temperature, mainly as a function of transmitted power. The resulting deteriorated beam quality of the combined output beam (Fig. 4, right scale) is linked to an increased filter temperature and with this to thermal induced wave-front distortions for the transmitted as well as for the reflected beams. This dependency will be discussed in the third chapter of this paper. At Pout = 208 W the peak temperature of IF12-34 is risen by 21 K, which leads to a deteriorated beam quality factor (4σ-method) of M2 = 2.3. A transversal beam profile in focal plane of the combined beam at maximum output power is set into Fig. 4.
2.4 Scaling pulse energy at 10 kHz repetition rate
In the second part of the experiment, the pulse repetition frequency is reduced down to 10 kHz to find out, what pulse energy is extractable out of a single-channel as well as the combined four-channel system. Because of the lower pulse repetition rate the Q-switched oscillator is now generating shorter pulses (18 ns) based on a higher degree of inversion in the single-clad fiber. Pulse energy vs. pump power of the amplified signal is plotted in Fig. 5 .
It shows a stronger roll-over at higher pump levels related to gain saturation through depletion of inversion, which is also the limiting factor in this part of the experiment. The maximum extractable pulse energy of the four-channel system is 6.3 mJ but it must be stressed that the uncombined fiber amplifiers yield even 7.3 mJ of pulse energy but 1 mJ is lost during the combining process itself. Under given setup-parameters, each fiber has reached the maximum energy storage capability at about 1.8 mJ. Thus, the extraction of 6.3 mJ pulse energy marks a strong improvement compared to an individual amplifier system and shows the scaling potential of the presented combining scheme.
3. The influence of the combining element on beam quality
The evaluation of the interference filter as combining element rests mainly on the conservation of beam quality during the combining process. Depending on the spectral bandwidth, a grating always adds divergence to the combined output beam and hence decreases beam quality. Filters lack this drawback but reveal also detrimental effects. Strictly speaking, the quality of the combined beam is not only defined by a perfect temporal superposition of the pulses but by the uniformity of spatially beam properties of the non-combined beams as well. Such properties can be beam size, pointing stability and wave-front characteristics like convergence and divergence, respectively. Variations of these properties among the single-channel beams will lead to a deterioration of beam quality of the combined output beam. This is valid for all beam combining techniques.
Figure 4, right-hand side shows an increase of the beam quality factor M2 versus rising pump power. Since each non-combined beam possesses a power-independent M2-value of ~1.2, the guiding property of the applied rod-type PCF is not the reason for the observed decrease in beam quality. The reason is rather related to a power-dependent dispersion of the single-channel beam properties in the combining process itself. A growth of temperature on IF12-34 can be observed at higher output powers. This heating causes two opposing effects. By absorbing a small amount of the transmitted power, the IF12-34 does induce not only a change in refractive index (dn/dT) but is causing bulges on both surfaces as well. This thermal lens leads to a convergence of the transmitted Beam B3-4. The reflected beam on the other hand is only influenced by the deformed surface resulting in a divergence of B1-2, contrary to the convergence of B3-4. This has repercussions on the combined beam once it is focused (e.g. M2-measurement). Firstly, diverging and converging beams result in different beam diameters leading to different focal beam waists, once they are focused. Secondly, these non-uniform wave-fronts are causing a shift of the focal positions due to the presence of an additional (thermal) lens. If the diverging and the converging beams are combined, both issues lead to the observed increase of the M2-value. Even more beam aberrations occur if the reflected beam is larger in size than the generated thermal lens. In this case only a small part of the reflected wave-front is influenced by the parabolic shape of the surface bulge. The consequence of this constellation is explained and underlined by experimental results in the next paragraph.
3.1 Further scaling considerations and experimental results
The heating of the interference filters, at this power level almost exclusively caused by the transmitted beam, is the limiting factor of the described setup. To increase the number of beams involved in the combining process and to improve the performance of the whole laser system a combining setup is proposed as shown in Fig. 6a . By following this approach, the generation of a thermal lens is minimized because each interference filter has to handle only one beam in transmission. Considering the combining stage only, maximum power and energy of each beam (Bx) is determined by the characteristics of the used interference filter. This comprises issues of absorption, surface quality, energy damage threshold and spectral steepness. A formation of thermal lenses is reduced with less absorption and an interference filter with a small transmission window and steep edges allows for high spectral channel density and leads to a higher maximum number N of individual emitters, which can be implemented in the combining process.
The reflected beam (Fig. 6a, orange line) is propagating a longer distance in space than the transmitted beam and if no beam size adaptation is implemented this beam will have a larger beam diameter at IF1 than the transmitted beam B1. This results in a deterioration of beam quality as discussed before. This deterioration is getting even worse if additionally a thermal lens is present. In this case, only a part of the mode area of the reflected beam will be exposed to the wave-front distortion of the spatially small thermal lens caused by the smaller transmitted beam B1 resulting in a very inhomogeneous wave-front aberration in total. To experimentally evaluate this influence, a ‘pump and probe’ experiment is set up as shown in Fig. 6b. The ‘pump’ beam has a diameter of ~3 mm - half the ‘probe’ beam diameter - and delivers up to 350 W of diffraction limited laser emission at 1060 nm. While increasing the ‘pump’ power, M2 and the focus shift of the reflected 5 W ‘probe’ beam is measured (Fig. 7 ).
Firstly, a shift of the focus position is noticeable, which can be explained by the thermal induced bulging of the surface leading to a general divergence of the beam. The reason for the increased M2-value is illustrated in Fig. 8 . The pictures of the transversal beam profiles originate from the M2-measurement at three different positions within the focal length while the ‘pump’ is at maximum power of 350 W. The first profile (a) is before the focus and clearly indicates the inhomogeneous wave-front and the interaction zone of the thermal lens induced bulging of the surface (marked red). This region of the beam is focused behind the ‘regular’ focus because of the diverging part of the wave-front. As a result, the now stretched Rayleigh range reduces the intensity in the focal region and with it the beam quality as well. To avoid this effect and to sustain beam quality during the combining process, the size of both beams have to be matched while thermal lensing has to be minimized by reducing the absorption of the filters. For the investigated interference filters, the power-limit is at about 100 W in transmission. Higher powers intensify thermal lens induced wave-front aberrations and hence trigger the whole series of described effects leading to a reduced beam quality.
Three interference filters are used to spectrally combine four fiber-amplified Q-switched laser systems in a straightforward setup. It provides 208 W of average output power and 4.2 mJ pulse energy at 50 kHz repetition frequency. By simply reducing the frequency down to 10 kHz, even 6.3 mJ pulse energy have been extracted. In comparison to recent results , four times more average power and three times more energy is achieved. In general, interference filters have the potential to maintain a high beam quality but during the combining process thermal-induced wave-front distortions of the non-combined beams lead to a deterioration of beam quality of the resulting combined beam. By reducing the absorption of the interference filters to a large extent, filters used as combining elements could excel as real competitors to dielectric gratings leading to an acceptable trade-off between lower combining efficiency and a spatial-dispersion-free beam quality. Within the scope of the presented concept, a much higher number of beams are conceivable for spectral combination.
This work has been partly supported by the German Federal Ministry of Education and Research (BMBF) with project 03ZIK455 “onCOOPtics.”
1. J. Limpert, A. Liem, H. Zellmer, and A. Tünnermann, “500 W continuous-wave fiber laser with excellent beam quality,” Electron. Lett. 39(8), 645 ( 2003). [CrossRef]
2. Y. Jeong, J. Sahu, D. Payne, and J. Nilsson, “Ytterbium-doped large-core fiber laser with 1.36 kW continuous-wave output power,” Opt. Express 12(25), 6088–6092 ( 2004), http://www.opticsinfobase.org/abstract.cfm?URI=oe-12-25-6088. [CrossRef] [PubMed]
3. D. Gapontsev, and I. P. G. Photonics, “6kW CW Single Mode Ytterbium Fiber Laser in All-Fiber Format,” in “Solid State and Diode Laser Technology Review” Albuquerque (2008).
4. J. W. Dawson, M. J. Messerly, R. J. Beach, M. Y. Shverdin, E. A. Stappaerts, A. K. Sridharan, P. H. Pax, J. E. Heebner, C. W. Siders, and C. P. J. Barty, “Analysis of the scalability of diffraction-limited fiber lasers and amplifiers to high average power,” Opt. Express 16(17), 13240–13266 ( 2008), http://www.opticsinfobase.org/oe/abstract.cfm?URI=oe-16-17-13240. [CrossRef] [PubMed]
5. F. Röser, D. Schimpf, J. Rothhardt, T. Eidam, J. Limpert, A. Tünnermann and F. Salin, ”Gain Limitations and Consequences for Short Length Fiber Amplifiers,” in Advanced Solid-State Photonics (Optical Society of America, Nara, 2007), paper WB22.
6. J. Limpert, O. Schmidt, J. Rothhardt, F. Röser, T. Schreiber, A. Tünnermann, S. Ermeneux, P. Yvernault, and F. Salin, “Extended single-mode photonic crystal fiber lasers,” Opt. Express 14(7), 2715–2720 ( 2006), http://www.opticsinfobase.org/oe/abstract.cfm?URI=oe-14-7-2715. [CrossRef] [PubMed]
7. C. D. Brooks and F. Di Teodoro, “Multimegawatt peak-power, single-transverse-mode operation of a 100µm core diameter, Yb-doped rodlike photonic crystal fiber amplifier,” Appl. Phys. Lett. 89(11), 111119–111122 ( 2006). [CrossRef]
8. O. Schmidt, J. Rothhardt, T. Eidam, F. Röser, J. Limpert, A. Tünnermann, K. P. Hansen, C. Jakobsen, and J. Broeng, “Single-polarization ultra-large-mode-area Yb-doped photonic crystal fiber,” Opt. Express 16(6), 3918–3923 ( 2008), http://www.opticsinfobase.org/oe/abstract.cfm?URI=oe-16-6-3918. [CrossRef] [PubMed]
9. F. Röser, T. Eidam, J. Rothhardt, O. Schmidt, D. N. Schimpf, J. Limpert, and A. Tünnermann, “Millijoule pulse energy high repetition rate femtosecond fiber chirped-pulse amplification system,” Opt. Lett. 32(24), 3495–3497 ( 2007), http://www.opticsinfobase.org/ol/abstract.cfm?URI=ol-32-24-3495. [CrossRef] [PubMed]
10. L. Dong, H. A. McKay, L. Fu, M. Ohta, A. Marcinkevicius, S. Suzuki, and M. E. Fermann, “Ytterbium-doped all glass leakage channel fibers with highly fluorine-doped silica pump cladding,” Opt. Express 17(11), 8962–8969 ( 2009), http://www.opticsinfobase.org/oe/abstract.cfm?URI=oe-17-11-8962. [CrossRef] [PubMed]
11. A. Sevian, O. Andrusyak, I. Ciapurin, V. Smirnov, G. Venus, and L. Glebov, “Efficient power scaling of laser radiation by spectral beam combining,” Opt. Lett. 33(4), 384–386 ( 2008), http://www.opticsinfobase.org/ol/abstract.cfm?URI=ol-33-4-384. [CrossRef] [PubMed]
12. 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), http://www.opticsinfobase.org/oe/abstract.cfm?URI=oe-17-3-1178. [CrossRef] [PubMed]
13. O. Schmidt, T. V. Andersen, J. Limpert, and A. Tünnermann, “187 W, 3.7 mJ from spectrally combined pulsed 2 ns fiber amplifiers,” Opt. Lett. 34(3), 226–228 ( 2009), http://www.opticsinfobase.org/ol/abstract.cfm?URI=ol-34-3-226. [CrossRef] [PubMed]
14. O. Schmidt, C. Wirth, I. Tsybin, T. Schreiber, R. Eberhardt, J. Limpert, and A. Tünnermann, “Average power of 1.1 kW from spectrally combined, fiber-amplified, nanosecond-pulsed sources,” Opt. Lett. 34(10), 1567–1569 ( 2009), http://www.opticsinfobase.org/ol/abstract.cfm?URI=ol-34-10-1567. [CrossRef] [PubMed]
15. K. Regelskis, K. Hou, G. Raciukaitis, and A. Galvanauskas, “Spatial-Dispersion-Free Spectral Beam Combining of High Power Pulsed Yb-Doped Fiber Lasers,” in Conference on Lasers and Electro-Optics/Quantum Electronics and Laser Science Conference and Photonic Applications Systems Technologies, OSA Technical Digest (CD) (Optical Society of America, 2008), paper CMA4.
16. Semrock: MaxLine Common Specifications, http://www.semrock.com/catalog/MaxLine_CommonSpecs.