Fig. 1. Schematic overview of the PEnELOPE system. An oscillator generates ultrashort pulses, that are picked, stretched, amplified and finally compressed down to 150 fs with 1 PW peak power and at 1 Hz repetition rate. Envisioned results are shown in light gray.

Fig. 2. Sketch of the HEPA I amplifier setup. The laser diode pump is shown in green and the extraction passes of the laser pulse are shown in red ray traces. Two 70 kW peak power laser diode assemblies are used to pump the laser head containing four $\text{Yb}^{3+}$: $\text{CaF}_{2}$ slabs in between two high pressure windows of the cooling assembly. The pump distribution is shown in the lower left. The laser pulse passes 12 times through the amplifier head before exiting HEPA I.

Fig. 3. (a) Simulated transmitted wavefront through the single-12-pass amplifier using Zemax and (b) measured transmitted wavefront using a SID4 wavefront sensor. The SR for the simulation is 95%, while 75% is estimated from the measured wavefront with fused silica slabs instead of the ytterbium doped gain medium in order to prove the optical setup limitation.

Fig. 4. (a) Small signal gain after double-12-passes for pump pulse durations between 1 and 4 ms and intensities up to $16.5~\text{kW}/\text{cm}^{2}$ is shown on the left. SSG values up to 900 are measured for the longest pump duration with a narrow-band 1030 nm cw source. (b) The right side shows the output energy as a function of the input for the three different spectral distribution cases A (black), B (blue) and C (green) in the case of $16.5~\text{kW}/\text{cm}^{2}$ and 4 ms. The single points for the cases B and C are measured without thermal roll-over (TRO). The dotted lines are extrapolations using the first three points of the individual curves to give a hint for the TRO impact. In order not to overload the graphs, one point gives the estimated error bars for case B. The corresponding gain (output divided by the input energy) is given explicitly to compare it to (a).

Fig. 5. Spectra (a) out of amplifier HGBA III and (b) after amplification with 24 passes. The cases A (black), B (blue) and C (green) correspond to 0.7 nm, 5 nm and 16 nm FWHM before injection into the 24 passes of the amplifier benchmark setup. The resulting FWHM bandwidth after amplification is 0.7 nm, 4.6 nm and 8.5 nm. The case C is still enough to support sub-150 fs pulses with a Fourier limit of 146 fs.

Fig. 6. (a) shows the output beam profile for the case C w/o TRO at 12.6 J. (b) shows the high spatial frequency content of an interferometric map in transmission through the four ytterbium doped slabs in a single pass. Clearly visible are the high frequency modulations and line like structures of inhomogeneities caused by small angle grain boundaries observed also in Ref. [21]. The black circle shows the total sampled area during the single-12-passes.

Fig. 7. First optimization of the beam quality after 12 passes using the 1064 nm cw source. (a) A zoom into the initial far-field with an SR of 12% after double-12-passes is shown. (b) shows the situation after only single-12-passes before polarization coupling takes place. An SR of 24% is found. (c) shows an optimized far-field spot with an SR of 76% after single-12-passes, as this is the initial design of the amplifier setup HEPA I. To get a better perception of the energy distribution in the vicinity of the central focal spot, the color scale (normalized intensity) fades into white for very low intensities.