There is significant international effort focussed on developing ultra-high-power systems for next-generation laser facilities, such as the Extreme Light Infrastructure (ELI). Existing amplification methods are based on chirped-pulse amplification (CPA). However, the low damage threshold of conventional solid-state optics results in very large amplifiers and compressors. To overcome this challenge, we use stimulated Raman backscattering of a long pump laser in plasma to provide amplification for a low intensity seed pulse. Plasma has the advantage that it is already a broken down medium and therefore field intensities are not constrained as they are in conventional laser amplifiers. This offers the potential to reduce the size and cost of these devices significantly, while providing a possible route to reach exawatt powers, which will enable investigation of extremely high field physics.
Despite its advantages, efficient Raman amplification has not yet been demonstrated experimentally. Efficiencies are limited to only a few percent for seed energies of a few mJ, in contrast with theoretical predictions. Several phenomena lead to saturation or inhibit the amplification process – such as detuning, wavebreaking and particle trapping – depending on the amplification regime. Amplification is therefore highly sensitive to the conditions and parameters used. Raman amplification experiments are challenging, and careful planning is required to ensure that controlled and sustained amplification can take place. Numerical simulation is an essential ingredient to this preparation yet, like the experiments themselves, this is not a trivial task. The amplification process takes place over several millimetres, while structures on the short beat wavelength of the lasers need to be adequately resolved. Since particle kinetic effects are also important, a large number of particles are required. Simulation of the entire domain therefore requires significant computing resources, and therefore many investigations are only performed in 1-dimension. Moreover, the long propagation times involved allow numerical artefacts from processes such as grid heating or numerical dispersion to become significant. These can become pathological and artificially seed or disturb the amplification process.
Using state-of-the-art numerical techniques, we investigate the amplification of low- and high-intensity seed pulses in plasma, and compare their amplification growth rates and efficiencies with experimental results obtained by our group. The use of a chirped pump laser pulses is discussed and compared.