Open Access
Add to BibSonomy
Abstract
A diode laser module emitting 1.4 kW optical in-pulse power near 780 nm optimized for high (≥ 10%) duty-cycle operation in a micro-channel free design is presented. With full collimation, a beam quality with a nearly symmetric M2 of 205 × 295 (vertical × horizontal direction) for a wide range of pulse widths is found.
© 2021 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
High power mid-infra-red lasers are a scientific field of growing interest due to their importance for many applications in e.g. material processing, basic research (for example particle acceleration), spectroscopy or medicine [1,2]. Two important designs of such lasers, fiber laser (FL) and solid-state laser (SSL), require highly efficient, high power pump sources with high stability and reliability at low cost. These pump sources provide the optical energy to excite the gain material in a high-power laser. An emerging field is the use of SSLs based on Thulium (Tm) doped in e.g. yttrium-aluminum-garnet (YAG) as it can generate laser pulses with wavelength near λ = 2 µm and is, therefore, regarded to be eye safe making it especially suitable for medical and ranging applications. Furthermore, a laser working at a wavelength near λ = 2 µm can be used to pump chromium doped gain material providing laser radiation near λ = 3 µm [2]. Tm:YAG can be efficiently pumped by high power diode lasers (HPDLs) that emit around λ = 780 nm, with a quantum efficiency of about 200% (one pump photon generates two laser photons) [3]. Moreover, the long lifetime of the upper state of Tm:YAG (of about 10 ms) allows the optical power of long pump pulses to be accumulated in the crystal to enable high energy pulses. Therefore, efficient, high power, high brightness and highly reliable HPDL pump sources are needed, that operate reliably without performance loss at high duty cycles (dc).
Alternative modules consist of stacks or arrays of high-energy class diode lasers (HEC-DL) in bar format where each bar is equipped with its fast axis collimation (FAC) lens [4–9].
These bars are usually densely packed and cooled from their rear edges with only limited cooling power and are, therefore, restricted to low dc operation. Or they are packed between micro-channel coolers which allows these units to operate from pulsed up to continuous wave operation. However, micro-channel coolers have demanding requirements on the cooling water which can lead to increased cost and can be a reliability hazard. Diode modules and bar assemblies which operate at 780 nm were reported as well in [10–13].
This contribution reports an important progress in addressing this emerging need. Specifically, a HPDL pump module is presented emitting around λ = 780 nm with high pulse repetition rate f = 10 Hz (100 Hz) with a pulse width of τ = 10 ms (1 ms), respectively, e.g. a high duty cycle of 10%. The paper is a follow-on publication of our work [14] in which the optical in-pulse power could be significantly increased.
The reported stack module can also be prepared with diode lasers that do emit at other wavelengths. Therefore, the stack module can be thought of as a carrier design that provides thermal and electrical connections w/o monitoring capabilities.
2. Chip device construction and chip device life test
The single emitters (SE) for the current generation of stack modules are GaAs-based and make use of a vertical epitaxial design that is tailored for high conversion efficiency at high bias, as described in detail in [15]. They were processed using standard techniques into innovative wide-aperture (1200 µm) diode lasers with cavity length of 6 mm. Periodic structuring of the contact layer using implantation was used as a technique to prevent lasing in parasitic ring-modes [15]. High reliability of the chips at high optical output was achieved by low defect epitaxial growth, wafer processing and packaging, and by the use of robust facet passivation (German Patent No. DE10221952). Subsequently, the front and rear facets were AR and HR coated achieving 0.8% and 98% reflectivity, respectively. The SE were designed to operate with optical output power Popt = 60 W, equivalent to Popt = 360 W per 1-cm bar for the 6 × 1200 µm emitter configuration in [16], and to Popt = 435 W for the 8 × 1093 µm configuration in [17]. The operating power is significantly below the measured peak (failure) power of more than 160 W [15].
The SE were soldered with epitaxy-side down between thermal expansion-matched CuW (10:90) heatsinks in a sandwich-like configuration [18]. Two AlxOy spacers placed left and right next to the chip ensure a better parallelism and serve as protection against packaging-induced mechanical stresses. In that orientation the fast axis of the emitters is the vertical direction whereas the slow axis of the emitters is the horizontal direction.
In order to understand the reliability of the SE, exemplary wide-aperture single emitters were mounted and tested in single sandwich-style stack element packaging which is identical to that later integrated into the stack modules. Following normal practice in testing of quasi-continuous wave diode laser sources, we focused on pulsed testing, on the assumption that the cycling of temperature and current is the primary driver of accelerated failure [19]. At the time of writing, life testing was only possible using a simple clamp-contact “blind” aging (burn-in) rack with limited cooling, and no in-line power, wavelength or temperature monitoring. These initial proof-of-principle tests were therefore restricted to pulsed operation at around dc = 1%, and the available test equipment did not allow us to make a reliable estimate of operating temperature. In a first study, four chip devices were tested in a constant drive current mode, at I = 70 A driving current which leads to Popt = 60 W optical power.
Four other SE were tested in a step-stress mode where driving current was increased from I = 70 A to 105 A in 5 A steps. The step stress was performed to assess operating margin, being intended to show that failure-free operation occurs for powers at least 25% higher than the specified power. The aging was performed with a pulse width of τ = 1 ms and a pulse repetition rate of f = 15.15 Hz. For both current regimes, repeat characterization was performed approximately once a week with a pulse width τ = 1 ms and a pulse repetition rate of f = 10 Hz (dc = 1%). Figure 1 shows the development of the optical power in repeat testing as a function of pulse number. To date (Dec. 2020) 350 × 106 laser pulses ( = 6417 h operation with f = 15.15 Hz) were accumulated without failure, with step stress operating point successfully increased to around Popt = 90 W (1.5 × operating power). In constant current mode, gradual degradation in the optical power is observed. A linear fit gives a degradation of 4.85 mW per 106 shots and, therefore, tolerating a degradation of 20%, more than 2.4 GShots (2.4 × 109 shots) are extrapolated as wear-out life time which means 44000 h or 5 years of operation can be achieved. When the devices are assembled into the stack modules, heat extraction is via soldered joints to two water cooled heatsinks (DCBs), instead of the clamp contacting used in the initial aging test. Thermal resistance will be reduced, operating temperature will be reduced and thus even larger device life times are to be expected. It is also possible that other failure modes will arise in the integrated module, but these have not been observed in previous completed modules with comparable construction at λ = 940 nm during long term use in pumping solid state lasers [15]. Further systematic studies are required to determine acceleration factors and to confirm that comparable lifetime is maintained in complete stack modules.
Fig. 1. Temporal development of the optical in-pulse power as a function of accumulated pulse number for a constant current and a step-stress life test. Crosses: constant current, dots: step stress with increasing current as marked. Both groups were tested at 70 A as well.
3. Stack module design
Twenty-four SE are soldered one over each other (AuSn solder) along their vertical direction (the vertical direction of the SEs is the vertical direction of the stack module) and then soldered (AuSn solder) between two expansion matched Direct Copper Bonding (DCB) large-channel water coolers to extract heat from the sides of the stacked devices. Cooling along the length of the resonator and the large area for heat extraction as well as the large volume for heat spreading leads to thermal impedance that is at least 5 times lower than conventional bar-based passively cooled solutions, even when these are tailored for long pulse or high duty cycle operation [7]. The novel cooling approach therefore allows high-heat (high dc) operation without need for the cost and potential reliability hazard of micro-channel coolers.
The single waveguide emitter in the chip (current design has 120) are electrically connected in parallel and the SE in the stack are electrically connected in series. There might be the risk of a total failure if just one SE fails electrically. However, it has been seen in the lab (no systematic study done yet) that the failure of one SE (in part or total) does not break the serial connection. The efficiency and, therefore, the thermal load may change locally but the stack as a whole keeps functioning at reduced overall optical power. We refer the reader to our earlier study, where we present a detailed review of our novel stack module as compared to bar-based solutions [15].
Once the soldering of the stack and the DCBs is complete, a stack module is assembled meaning the stack is mounted on a base plate together with water and current connectors. Three temperature sensors at the rear edge of the stack, attached to top (emitter 2), center (emitter 12) and bottom (emitter 23) allow early warning monitoring, and report comparable temperatures throughout testing (typically less than 0.5 K temperature non-uniformity). For high brightness applications each SE was individually fast axis collimated (FAC, with a focal length of 1.2 mm) and all share a common slow axis collimation (SAC). The FACs were aligned and attached to a common side bar with UV epoxy. The residual vertical divergence per SE is about 2…3 mrad and the residual horizontal divergence of the stack module is about 3 mrad. Figure 2 shows photographs of the stack module, the single FACs and a false color image of a screen on which the fully collimated single beams of the stack module were guided on. It shows an equidistant spacing of the laser beams. The screen was placed ∼0.5 m away and, for the photograph, the common SAC has not been optimally aligned.
Fig. 2. Left: Fully assembled and ready-for-shipment 1.4 kW stack module with full vertical collimation (FAC). Common SAC not shown. The module is cooled with large channel DCB heatsink, so is a micro-channel free design. Middle: Magnified view of individual fast axis collimation (FAC) lenses. Right: false color photograph of the fully collimated beams of a stack module being guided on a screen (operational conditions: 1 ms, 100 Hz, 70 A).
For power scaling, the collimated modules can be interleaved or polarization beam combined and also be integrated into a fiber coupled module [18]. A power scaling by interleaving the single beams of two stack modules (see e.g. [8]) is discussed in the following. The active regions of the chip devices have a vertical distance between each other (pitch) of 3.16 mm. The vertical beam diameter right behind the FAC-lenses is about 800 µm…1000 µm which represents a vertical fill factor of the stack module’s aperture (23 × pitch times emitter width = 72.7 mm × 1200 µm) of ≤ 50% which is the maximum fill factor for interleaving. Hence, with a vertical displacement of the two stacks of pitch half the single laser beams of one stack module can be guided into the gaps between the laser beams of the other stack module. In former work of the group, this was done with a prism stack [8].
It should be noted that this module design can also be adapted to other wavelengths, for pumping other solid-state gain media [18].
4. Standard characterization procedure of the stack modules
A standard characterization procedure (SCP) was performed to obtain stack module performance and properties. It covers electrical, spectral and beam quality measurements. All measurements were performed with a duty cycle of 10% for two pulse conditions i.e. a pulse width of τ = 10 ms (1 ms) and a pulse repetition rate of f = 10 Hz (100 Hz) and the stack module was cooled with a flow rate of the cooling water of 8 l/min (4 l/min per DCB cooler) with a water supply temperature of ∼18°C.
4.1 Optical power P, voltage V, current I and conversion efficiency η
The in-pulse voltage U and the optical power Popt (calibrated against national standards, average optical power measured with a Gentec UP55N) were measured as a function of current I and the conversion efficiency $\eta = {P_{opt}}/UI$ was calculated as well. The current supply for the 1 ms, 100 Hz regime was an Amtron CS401 whereas the 10 ms, 10 Hz regime was driven by a Delta Elektronika SM45-140 with an external function generator. Figure 3 shows the voltage, free space optical power after collimation and the efficiency of a 1.4 kW stack module as a function of drive current at 10% duty cycle operation. Even though higher levels of heating occur for longer pulses (see spectral measurements below), operating temperatures are similar and broadly comparable light-current-voltage characteristics are observed. Specifically, at the maximum operational current of 70 A an average optical in-pulse power of Popt = 1.4 kW is emitted independently of the pulse width (τ = 1 …10 ms), for pulse energies of up to 14 J. The highest observed conversion efficiency is η = 50%, comparable to the highest conversion efficiency found for a single unstacked SE, which is also η = 50% [15]. Power is significantly increased over the Popt = 1.1 kW reported in [14] by improvements in the collimation procedure and optics, as well as by ensuring the smallest number of lost or damaged emitters.
Fig. 3. Voltage, optical power and conversion efficiency as a function of the drive current for a stack module tested under two pulse widths at a duty cycle of 10%.
4.2 Spectral characterization
Spectral characterization was performed by integrating the emitted light spatially with an integrating sphere which was connected, via a fiber, to a spectrometer (Ocean Optics HR4000). Saturation has been avoided adjusting the exposure time. Figure 4 shows the normalized spectra for the two pulse regimes for multiple driving currents.
Fig. 4. Spatially integrated spectra of the stack module for two pulse regimes and multiple driving currents. Top: τ = 10 ms, f = 10 Hz. Bottom: τ = 1 ms, f = 100 Hz.
The side-cooled design, large heatsink volume and large heat-extraction area of the stack module provides a slow thermal response which allows the user to operate across a range of pulse widths and duty cycle combinations with only small changes in performance. The parameter that is usually used to understand and compare the thermal performance numerically is the dynamic thermal impedance ${Z_{th}}$ which is calculated by tracking the wavelength shift as a function of peak generated heat ${P_{Diss}} = U \cdot I - {P_{opt}}$. The center emission wavelength is a good measure for the temperature and increases monotonically with the temperature of the active region (${T_{AR}}$). For the current design of the SE that is $\Delta \lambda /\Delta T = 0.256nm/K$. Table 1 gives the tuneability and ${Z_{th}}$ values for the two pulse regimes.
Table 1. Spectral tuning as a function of current and dissipated power and the thermal impedance for the two pulse regimes.
The calculated ${Z_{th}}$ match those found in former experiments with comparable stack modules utilizing λ = 940 nm SE [18,20] and suggest that a higher duty cycle and/or longer pulses would also be possible in the current stack module design. The difference in Zth between 1 ms and 10 ms is small (being 8 and 11 mK/W respectively), and leads to < 10 K difference in operation temperature, insufficient to cause large performance differences. The effective cooling has enabled CW operation to be demonstrated in comparable previous stacks operating at λ = 940 nm, at the cost of power and beam quality (thermal resistance is ∼ 4 × larger CW cf. pulsed) [20], but was not studied here.
Spectral fine tuning of the center wavelength of a stack module can be done by adjusting the current or the temperature of the supply water which also can be adjusted to ensure overlap of the center wavelength of many combined stack modules.
4.3 Beam propagation ratio M2
The third step in the SCP is a caustic measurement to obtain values for M2 in horizontal and vertical direction to judge beam quality. A vertical focusing lens (VFL) and a horizontal focusing lens (HFL) with a focal length fVFL = 300 mm and fHFL = 75 mm, respectively, were used to focus the fully collimated stack module onto a camera sensor (CS) which moved through the beam waist region to measure 2D-intensity images which, subsequently, were used to calculate the beam diameter as a function of propagation position. See Fig. 6 and [20] for more details of the technique.
Fig. 5. Sketch (not to scale) of the experimental set-up used to measure 2D-intensity profiles used to calculate the beam quality, e.g. caustic measurement. The fully collimated single laser beams from the stack (not shown in the sketch) are on the left side and radiation propagates to the right. Top: side-view to illustrate the stacked beams and their focusing and the positions of the 2D-images given as samples in Fig. 6. Dimensions of the camera sensor are given as well. Bottom: view from top to bottom. Sample images that are given in Fig. 6 were taken at positions -20 mm, -10 mm, 0 mm, 10 mm, 20 mm.
Fig. 6. 2D-intensity profile of the propagating beam from a single collimated stack module (camera images, linear grey-scale, intensity is normalized at each position) at a series of positions through the beam waist. Measured optical intensity is shown as a function of propagation position for the two pulse regimes of τ = 10 ms, f = 10 Hz (top) and τ = 1 ms, f = 100 Hz (bottom). The stack module is located at the left-hand side of the beam path, the beam propagates to the right. Position around the beam waist (see Fig. 5) from left to right: (z – z0) ∼ −20 mm, ∼ −10 mm, 0 mm, ∼ +10 mm and ∼ +20 mm.
The readout of the camera gives 2D-images which consist of 1164 px × 874 px where a pixel is 11 µm × 11 µm. Hence, the size of the images is 12.8 mm × 9.6 mm. The images have an 8-bit resolution. The power density on the camera sensor varies with propagation position. Therefore, saturation of the sensor’s AD-converter has been avoided by adjusting the exposure time and by the use of neutral density (ND) filters.
The sample images of Fig. 6 may give the reader an impression on how discontinuous the intensity distribution of the stack module’s radiation looks like. The intensity of each single laser is clearly distinguishable but do overlap which appears to contradict the concept of interleaving the beams of two stack modules. The reason is that the images were taken around the focal region and the single intensity spots of the single lasers appear to be shrank together due to the focusing lenses. The interleave would be done where the single beams are parallel with enough gap size in between.
The beam diameter, here assessed using a typical industrial 95% power content measure, was analyzed following the ISO11146 standard applied to the 2D-camera images, in a process that we term here as a beam caustic analysis. Nearly one image per mm was taken. Figure 7 shows the beam diameter around the beam waist and the beam diameter fitted to the Gaussian Beam Theory using $D(z) = {D_0} \cdot \sqrt {1 + {{((z - {z_0})/{Z_R})}^2}} $ with beam waist diameter, D0, position along the propagation axis, z, beam waist position z0 ( = 0 mm here) and Rayleigh length, zR.
Fig. 7. Measured beam diameter with 95% power content (symbols) and fit using Gaussian model (curve) for vertical (top) and horizontal (bottom) beam for a single stack module. A 2D- image was taken at each symbol position.
Table 2 shows the parameters for the two pulse regimes, τ = 1 ms, f = 100 Hz and τ = 10 ms, f = 10 Hz. In the vertical direction, the beam properties are similar, although the beam waist is slightly smaller for τ = 1 ms, f = 100 Hz. In the horizontal direction, the far field divergence strongly increases for τ = 10 ms, f = 10 Hz. Exemplary images from the caustic measurement illustrate the propagation of the intensity distribution around the beam waist region, and are shown in Fig. 6, for the two pulse regimes at I = 70 A, Popt = 1.4 kW (rated operation conditions). The close-to symmetric beam profile with M2 < 300 allows for a simple pulse shaping and potentially enables the realization of compact, low cost (in €/W) high energy class MIR laser systems [21]. M2 is significantly reduced below the M2 = 330 reported in [14] by improvements in the collimation procedure and optics.
Table 2. Beam propagation ratios M2-Parameter that were obtained from fitting Gaussian Beam Theory to the intensity distribution of the camera measurements with beam diameter of 95% power content.
5. Summary
High duty cycle and high brightness λ = 780 nm pump stack modules were presented that emit high intensities (in-pulse power Popt = 1.4 kW) at high duty cycles dc = 10% (τ = 10 ms, f = 10 Hz and τ = 1 ms, f = 100 Hz) based on novel wide-aperture single emitters in a customized side-cooled stack format. The side cooled format enables high duty cycle, flexible operation at a range of pulse conditions without drop in performance, and the novel wide-aperture emitter leads to a close to symmetric beam with M2 < 300, for simple beam forming using bulk optical elements. These stack modules are promising components as pump sources for future pulsed MIR systems, and can be used as free-space pumps or integrated into a fiber-coupled module [18]. Additional stack modules can easily be optically combined for power scaling, using interleaving and polarization beam combining. The diode laser design can also be applied at other wavelengths and, therefore, the pump modules can be used to pump other solid-state gain material (e.g. Yb:YAG – see [20]).
Funding
Bundesministerium für Bildung und Forschung (FKZ-03VNE2068E).
Acknowledgements
We thank Joachim Hein and Jörg Körner from the Institute for Optics and Quantum Electronics (IOQ), Friedrich-Schiller University in Jena, Germany for many helpful discussions. We also thank the German Federal Ministry for Education and Research (BMBF) for funding within the KMU-NetC program (Project HECMIR: FKZ –03VNE2068E).
Disclosures
The authors declare no conflict of interest.
Data availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
References
1. M. Yumoto, N. Saito, Y. Urata, and S. Wada, “128 mJ/Pulse, Laser-Diode-Pumped, Q-Switched Tm:YAG Laser,” IEEE J. Sel. Top. Quantum Electron. 21(1), 364–368 (2015). [CrossRef]
2. S. B. Mirov, I. S. Moskalev, S. Vasilyev, V. Smolski, V. V. Fedorov, D. Martyshkin, J. Peppers, M. Mirov, A. Dergachev, and V. Gapontsev, “Frontiers of Mid-IR Lasers Based on Transition Metal Doped Chalcogenides,” IEEE J. Sel. Top. Quantum Electron. 24(5), 1–29 (2018). [CrossRef]
3. J. Körner, T. Lühder, J. Reiter, I. Uschmann, H. Marschner, V. Jambunathan, A. Lucianetti, T. Mocek, J. Hein, and M. C. Kaluza, “Spectroscopic investigations of thulium doped YAG and YAP crystals between 77 K and 300 K for short-wavelength infrared lasers,” J. Lumin. 202, 427–437 (2018). [CrossRef]
4. W. Fassbender, H. Kissel, J. Lotz, T. Koenning, S. Patterson, and J. Biesenbach, “Reliable QCW diode laser arrays for operation with high duty cycles,” Proc. SPIE 10085, 1008509 (2017). [CrossRef]
5. M. Wölz, A. Pietrzak, A. Kindsvater, J. Meusel, K. Stolberg, R. Hülsewede, J. Sebastian, and V. Loyo-Maldonado, “Laser diode stacks: pulsed light power for nuclear fusion,” High Power Laser Sci. Eng. 4, e14 (2016). [CrossRef]
6. P. Zhang, X. Liu, Q. Zhu, and J. Wang, “Thermal characteristics of compact conduction-cooled high-power diode laser array packages,” Proc. SPIE 10085, 100850A (2017). [CrossRef]
7. A. Kindsvater, M. Schröder, E. Werner, S. Seidel, M. Wölz, and V. Loyo-Maldonado, “High duty-cycle, high-efficiency QCW stacks for medical applications,” Proc. SPIE 9733, 97330M (2016). [CrossRef]
8. R. Platz, B. Eppich, J. Rieprich, W. Pittroff, G. Erbert, and P. Crump, “High duty cycle, highly efficient fiber coupled 940-nm pump module for high-energy solid-state lasers,” High Power Laser Sci. Eng. 4, e3 (2016). [CrossRef]
9. R. Platz, C. Frevert, B. Eppich, J. Rieprich, A. Ginolas, S. Kreutzmann, S. Knigge, G. Erbert, and P. Crump, “Progress in high duty cycle, highly efficient fiber coupled 940-nm pump modules for high-energy class solid-state lasers,” Proc. SPIE 10513, 1051319 (2018). [CrossRef]
10. T. Koenning, K. Alegria, Z. Wang, A. Segref, D. Stapleton, W. Faßbender, M. Flament, K. Rotter, A. Noeske, and J. Biesenbach, “Macro-channel cooled high power fiber coupled diode lasers exceeding 1.2 kW of output power,” Proc. SPIE 7918, 79180E (2011). [CrossRef]
11. A. Bayer, A. Unger, B. Köhler, M. Küster, S. Dürsch, H. Kissel, D. A. Irwin, C. Bodem, N. Plappert, M. Kersten, and J. Biesenbach, “Multi-kW high-brightness fiber coupled diode laser based on two-dimensional stacked tailored diode bars,” Proc. SPIE 9733, 97330A (2016). [CrossRef]
12. T. Koenning, D. McCormick, D. Irvin, D. Stapleton, T. Guiney, and S. Patterson, “Narrow-line fiber-coupled modules for DPAL pumping,” Proc. SPIE 9348, 934805 (2015). [CrossRef]
13. R. Pandey, D. Merchen, D. Stapleton, D. Irwin, C. Humble, S. Patterson, H. Kissel, and J. Biesenbach, “Narrow-line, tunable, high-power diode laser pump for DPAL applications,” Proc. SPIE 8733, 873307 (2013). [CrossRef]
14. M. Hübner, M. Wilkens, B. Eppich, A. Maaßdorf, D. Martin, A. Ginolas, P. S. Basler, and P. Crump, “Kilowatt-class, High Duty Cycle, Passively Side-cooled 780 nm Diode Laser Stacks as Pumps for Pulsed MIR Lasers,” in Laser Congress 2020 (ASSL, LAC) (Optical Society of America, 2020), p. JTh2A.26. [CrossRef]
15. P. Crump, M. Wilkens, M. Hübner, S. Arslan, M. Niemeyer, P. S. Basler, D. Martin, A. Maaßdorf, A. Ginolas, and G. Tränkle, “Efficient, high power 780 nm pumps for high-energy class mid-infrared solid-state lasers,” Proc. SPIE 11262, 1126204 (2020). [CrossRef]
16. M. M. Karow, C. Frevert, R. Platz, S. Knigge, A. Maaßdorf, G. Erbert, and P. Crump, “Efficient 600-W-Laser Bars for Long-Pulse Pump Applications at 940 and 975 nm,” IEEE Photonics Technol. Lett. 29(19), 1683–1686 (2017). [CrossRef]
17. M. M. Karow, D. Martin, P. D. Casa, G. Erbert, and P. Crump, “Design progress for higher efficiency and brightness in 1 kW diode-laser bars,” Proc. SPIE 11262, 1126205 (2020). [CrossRef]
18. M. Hübner, B. Eppich, A. Maaßdorf, D. Martin, A. Ginolas, P. S. Basler, M. Niemeyer, and P. Crump, “High Duty Cycle, High Repetition Rate High Brightness Diode Laser Pulsed-Pump-Sources,” in Laser Congress 2019 (ASSL, LAC, LS&C), OSA Technical Digest (Optical Society of America, 2019), p. JW2A.29. [CrossRef]
19. E. Deichsel, D. Schröder, J. Meusel, R. Hülsewede, J. Sebastian, S. Ludwig, and P. Hennig, “Highly reliable QCW laser bars and stacks,” Proc. SPIE 6876, 68760K (2008). [CrossRef]
20. M. Hübner, I. Will, J. Körner, J. Reiter, M. Lenski, J. Tümmler, J. Hein, B. Eppich, A. Ginolas, and P. Crump, “Novel High-Power, High Repetition Rate Laser Diode Pump Modules Suitable for High-Energy Class Laser Facilities,” Instruments 3(3), 34 (2019). [CrossRef]
21. J. Hein and R. Bödefeld, “High Energy Class Mid-IR Laser Workshop (2020),” https://photonics-days-berlin-brandenburg-2020.b2match.io/agenda?session=59035 (last accessed 19.02.2021).
