Intelligent single-shot full-field characterization over femtosecond pulses

Area of Science:

• Optics and Photonics
• Ultrafast Science
• Machine Learning Applications

Background:

• Conventional femtosecond (fs) pulse characterization relies on nonlinear optical effects, limiting the measurement of weak pulses.
• Existing methods often require iterative phase retrieval or spectrometer-based readout, hindering single-shot full-field characterization of high-repetition-rate fs pulse trains.

Purpose of the Study:

• To develop a novel method for single-shot full-field characterization (ISFC) of high-repetition-rate femtosecond (fs) pulse trains.
• To enable the measurement of weak fs pulses by avoiding nonlinearities in the characterization process.

Main Methods:

• Linear spectral shearing interferometry (LSSI) is employed to generate spectral shear through purely linear effects.
• A well-trained fully-connected neural network reconstructs both the intensity and phase of fs pulses from temporal interferograms in a single shot.
• The method is validated using a megahertz fs pulse train with picojoule-level energy.

Main Results:

• Successful demonstration of LSSI-based ISFC for characterizing weak fs pulses.
• Achieved single-shot full-field characterization of high-repetition-rate fs pulse trains.
• Resolved the switching dynamics of a programmable spectral filter at a megahertz frame rate.

Conclusions:

• LSSI-based ISFC offers a powerful new tool for characterizing weak fs pulses, overcoming limitations of nonlinear methods.
• The technique is suitable for high-repetition-rate fs pulse trains, including potential applications for ultraviolet and attosecond pulses.