Conventional electron microscopy has reached the point where atomic characterization of a wide range of samples is routine. A focused beam of accelerated electrons can provide a view of the crystal structure of materials through direct imaging or electron diffraction with sub-picometer spatial resolution. The beam is produced via photoemission, and the timescale associated with the "pulse" or packet of electrons dictates the temporal resolution. By combining such beams with attosecond (10–18 s) spectroscopy, where lasers compress the beam into a series of these pulses with shorter time scales, researchers Yuya Morimoto and Peter Baum at Ludwig-Maximilians-Universität München have developed a method for studying physical phenomena with unprecedented space–time resolution.

The technique relies on the use of lasers to modulate the propagation of the photo-emitted electron beam. The electron packets are periodically accelerated and decelerated in a dielectric membrane, depending on their arrival time relative to the laser cycle. The end result is the production of a "train" of attosecond length pulses. Characterization of this pulse-train beam reveals that the beam quality is not significantly different from the pre-compression beam, with the additional benefit that the best-compressed pulses within the train are up to 35 times shorter than has previously been demonstrated for atomic-scale diffraction.

To demonstrate the novel capabilities of this technique, Morimoto and Baum conducted two proof-of-concept experiments. The first investigates Bragg diffraction, a fundamental interaction between electrons and matter that involves the conversion of crystal momentum into electron momentum as an incoming electron is scattered by the crystal lattice. The researchers used attosecond-resolved diffraction of single-crystal silicon to see whether there is a delay associated with this scattering mechanism. By quantifying the time-dependent streaking of the diffraction spots they could deduce that no delay occurs that can be resolved at attosecond timescales, despite the various electronic and physical interactions between the beam and crystal.

The second experiment involves studying the interaction between light and matter using real-space imaging. They tuned a laser to generate an optical excitation wave across a large (130 × 150 μm) silicon membrane. The time-dependent intensity observed due to the attosecond pulses can then be linked to the relative phase associated with the optical wave, resulting in a phase map where each point on the sample has deflected the beam with a phase delay and amplitude based on the optical waveform at that specific location. In this case the phase was shown to advance continuously in a direction parallel to the excitation polarization and they were able to measure the travelling wave parameter. The ability to measure this interaction between light and matter is a potential route to probing nanophotonic phenomena that occur on ultrafast timescales.

These experiments mark the beginning of what this hybrid method has to offer regarding the observation of attosecond-scale processes. A host of experiments in photonics and condensed-matter physics are now possible due to the space–time resolution achievable with this technique. The authors mention potential modifications that could further increase the temporal resolution by shortening electron pulses or by isolating single pulses from a train for even more advanced photonics applications.

More information can be found in Nature Physics.