As part of the project, an efficient manufacturing processes for the envisioned multicore fiber with 10x10 cores have to be developed. Two different strategies have been realized. A first preform for a 4x4 core fiber has been produced based on a sintering process. First experiments showed, that the necessary geometrical accuracy could not be achieved with this process due to the shrinking of the preform during the drying and sintering process. Therefore, a second preform has been produced by deep hole drilling of fused silica. This process allows for easy and fast manufacturing of the preform and allowed for a first 4x4 core fiber. This fiber has been used for subsequent experiments on coherent combining.
Based on the fiber described above, we have realized a first coherently combined amplifier. We have realized a 1:16 beam splitter based on a segmented mirror with zones of different reflectivities. This splitter can split the incoming beam into a 4x4 beam array with equidistant and parallel beams. This array can then be directly imaged to the fiber end-facet. Due to the high quality of the optical setup, this leads to a coupling of all the beams into the respective cores of the multicore fiber. After amplification, the beams are recombined using a setup that is basically similar to the splitting stage. To ensure correct phasing of the beams during the combination steps, an integrated piezo array with 16 actuators was designed and acquired. Mirrors were attached to each actuator to provide individual path length adjustment, and, therefore, phasing of each beam. In a first experiment, we achieved a combined output power of 70W at a combination efficiency of 80% for 40ps pulses.

We have also investigated algorithms for detecting phase differences between the beams at the output and provide feedback signals to the piezo array to ensure constructive interference during combination. We have developed and implemented a novel phase modulation scheme that is scalable to a higher number of beams while taking into account the bandwidth limitations of the piezo actuators of about 10kHz. This scheme has also been mathematically proven.

Recently, we also tested the average power capabilities of the mentioned fiber and could achieve an output power of 1.4kW.
For temporal pulse combination, we implemented a first example of electro-optically controlled divided-pulse amplification (EDPA), which does not require a pulse splitting stage. Due to the enhanced temporal distribution in comparison to a single pulse, a reduction of accumulated nonlinear effects and an improved energy extraction from the active medium is expected. Additionally, in contrast to previous concepts, EDPA has the potential to counteract certain performance-reducing phenomena, such as gain saturation. Thus, EDPA is a promising technique for the generation of high-energy ultrashort pulses.
In a low-power proof-of-principle experiment, 4 subsequent picosecond pulses emitted by a fiber-oscillator with a repetition rate of 108 MHz have been combined successfully. While a high combining efficiency of more than 95% has been achieved, the built system has also shown a low relative intensity noise of 0.54%. Moreover, a good temporal contrast, i.e. the energy of the combined pulse divided by the energy of the strongest pre-pulse, of about 29dB has been measured. Supported by these achievements, the next step has been to stack a burst of 8 subsequent pulses, which has shown similar results. For the upcoming application at a high-power laser, a size-optimized EDPA setup additionally enabling a further improved stability has been designed and assembled.
The development of multi-kW average power ultrafast lasers and their coherent combination requires low-absorption substrates and coatings for the beam combining, transport and polarization control. Investigations on available and off-the-shelf components, such like 50/50 beam splitters, thin-film polarizers and wave plates, were carried out using our high-power lasers. Most of the components show significant heating starting from 1 kW average power, which leads to thermal lensing deteriorating the beam quality, where the beam splitters turned out to be the most critical component. Regarding the substrate, two types of optical glass and calcium fluoride have been identified to support this power level without thermal lensing or substrate heating. Optical components based on these substrates have been acquired and were testes again for power handling capabilities. Here, the choice of a thin coating turned out to be the best solution, as the absorption of the coating material remains the limiting factor. Beam splitters based on near-normal incidence feature the thinnest coatings compared to thin-film-polarizers or beam splitters at 45° angle of incidence. Based on this result, beam splitters for use at near-normal incidence have been ordered for testing.