Artificial intelligence for online characterization of ultrashort X-ray free-electron laser pulses

• DOI: 10.48550/arxiv.2108.13979
• Publication Date: 2021-01-01
• Authors: Kristina Dingel; Thorsten Otto; Lutz Marder; Lars Funke; Arne Held; Sara Savio; Andreas Hans; Gregor Hartmann; David Meier; Jens Viefhaus; Bernhard Sick; Arno Ehresmann; Markus Ilchen; Wolfram Helml
• Keywords: Attosecond; Streaking; Free electron model; Characterization; Coulomb explosion

X-ray free-electron lasers (XFELs) as the world's brightest light sources provide ultrashort pulses with a duration typically in order of femtoseconds. Recently, they have approached and entered attosecond regime, which holds new promises for single-molecule imaging studying nonlinear ultrafast phenomena such localized electron dynamics. The technological evolution XFELs toward well-controllable precise metrology processes has been, however, hampered by diagnostic capabilities characterizing at frontier. In this regard, spectroscopic technique photoelectron angular streaking successfully proven how to non-destructively retrieve exact time-energy structure XFEL on single-shot basis. By using artificial intelligence techniques, particular convolutional neural networks, we here show can be leveraged from its proof-of-principle stage routine diagnostics even high-repetition-rate XFELs, thus enhancing refining their scientific accessibility all related disciplines.