Continuing Towards Real-time Programmable Photoinjector Shaping for Electron and X-ray Sources

The next generation of augmented brightness X-ray free electron lasers (XFEL), such as LCLS-II, promises to address current challenges associated with systems with low X-ray cross-sections. A key component of XFELs is the photoinjector [1], which produces the electron beams (e-beams) whose phase-space determines the performance of the XFEL [2]. Fast and active beam manipulation is required to capitalize on this new generation of XFELs. We examine e-beam shaping using a machine learning (ML) implementation of real-time photoinjector laser manipulation. Our presentation will focus on the photoinjector laser software model, the associated ML models, and the optical setup. We anticipate this approach to not only enable active experimental control of X-ray pulse characteristics but could also increase the operational capacity of future e-beam sources and accelerator facilities.

[1] Gilevich, Sasha, et al. "The LCLS-II Photo-Injector Drive Laser System." 2020 Conference on Lasers and Electro-Optics (CLEO). IEEE, 2020.

[2] Lemons, Randy, et al. “Dispersion-controlled Temporal Shaping of Picosecond Pulses via Non-colinear Sum Frequency Generation.” arXiv, 2020.