Spectral-brightness optimization of an X-ray free-electron laser by machine-learning-based tuning

Abstract

A machine-learning-based beam optimizer has been implemented to maximize the spectral brightness of the X-ray free-electron laser (XFEL) pulses of SACLA. A new high-resolution single-shot inline spectrometer capable of resolving features of the order of a few electronvolts was employed to measure and evaluate XFEL pulse spectra. Compared with a simple pulse-energy-based optimization, the spectral width was narrowed by half and the spectral brightness was improved by a factor of 1.7. The optimizer significantly contributes to efficient machine tuning and improvement of XFEL performance at SACLA.

Keywords

• x-ray free-electron lasers
• machine learning
• beam tuning
• sacla
• spectral-brightness optimization
• single-shot inline spectrometers