Abstract

The characterization of attosecond pulses is crucial for attosecond metrology. In this work, we investigate the isolated attosecond pulse reconstruction with the all-optical method. The results show that this method can characterize isolated attosecond pulses with a duration shorter than 50 attoseconds. Moreover, we develop a deep learning scheme to characterize isolated attosecond pulses. Through supervised learning, the deep neural network learns the mapping from the photon spectrograms to attosecond pulses. It allows complete characterization of the amplitude and phase of isolated attosecond pulses. Compared to the conventional principal component generalized projections algorithm, the reconstruction with our neural network shows superior quality and robustness to noise. Also, the reconstruction computation time is significantly reduced to a few seconds.

© 2023 Optica Publishing Group

Deep learning reconstruction of ultrashort pulses

Tom Zahavy, Alex Dikopoltsev, Daniel Moss, Gil Ilan Haham, Oren Cohen, Shie Mannor, and Mordechai Segev
Optica 5(5) 666-673 (2018)

Deep learning in attosecond metrology

Christian Brunner, Andreas Duensing, Christian Schröder, Michael Mittermair, Vladimir Golkov, Maximilian Pollanka, Daniel Cremers, and Reinhard Kienberger
Opt. Express 30(9) 15669-15684 (2022)

Fast phase retrieval for broadband attosecond pulse characterization

Jiacan Wang, Fan Xiao, Li Wang, Wenkai Tao, Xiaowei Wang, and Zengxiu Zhao
Opt. Express 31(26) 43224-43233 (2023)

Complete reconstruction of ultra-broadband isolated attosecond pulses including partial averaging over the angular distribution

Thomas Gaumnitz, Arohi Jain, and Hans Jakob Wörner
Opt. Express 26(11) 14719-14740 (2018)

Robustness of the ePIE algorithm for the complete characterization of femtosecond, extreme ultra-violet pulses

Mario Murari, Giacinto D. Lucarelli, Matteo Lucchini, and Mauro Nisoli
Opt. Express 28(7) 10210-10224 (2020)