Smarter algorithm boosts X-ray pulses from laser plasma by 10-fold
Researchers used artificial intelligence to optimize extreme ultraviolet light sources created by firing lasers at plasma, achieving a tenfold increase in useful radiation. The advance could accelerate adoption of compact, high-speed imaging systems for materials research and industrial inspection that currently rely on expensive, building-sized facilities.
Originaltitel: Batch Bayesian optimization of attosecond betatron pulses from laser wakefield acceleration
Abstract Laser wakefield acceleration can generate a femtosecond-scale broadband X-ray betatron radiation pulse from electrons accelerated by an intense laser pulse in a plasma. The micrometer-scale of the source makes wakefield betatron radiation well-suited for advanced imaging techniques, including diffraction and phase-contrast imaging. Recent progress in laser technology can expand these capabilities into the attosecond regime, where the practical applications would significantly benefit from the increased energy contained within the pulse. Here we use numerical simulations combined with batch Bayesian optimization to enhance the radiation produced by an attosecond betatron source. The method enables an efficient exploration of a multi-parameter space and identifies a regime in which a plasma density spike triggers the generation of a high-charge electron beam. This results in an improvement of more than one order of magnitude in the on-axis time-averaged power within the central time containing half of the radiated energy, compared to the reference case without the density spike.