Many efforts in the past decade have been made to improve the temporal resolution of in-situ TEM in order to reveal the dynamics of processes at the nanoscale. However, most processes occur at time scales in the micro- to femtosecond domain which is beyond the acquisition frequency of the TEM cameras (down to few milliseconds). Thus the salient details of sample dynamics such as defect formation, phase transformations, nucleation phenomena etc. are often inaccessible.
For time-resolved studies, a much higher temporal resolution is therefore required. This can be achieved by using short electron pulses in a pump-probe approach. Ultrafast TEM (UTEM) consists of a TEM combined with a pulsed laser (figure 1). A photo cathode in the electron gun is illuminated by a fs-laser to produce a photoelectron pulse with a duration of 2-10 ps. After laser excitation of the object (pump pulse), the photoelectron pulses serve as probes with a variable time delay after the excitation. Repeating this process at different pump-probe delays allows time-resolved studies.
In contrast to conventional TEM, the electron-electron interaction in one pulse is not negligible in UTEM and has to be studied in detail. The energy width (ΔE), temporal length (Δt) and different arrival times (t0) of the electron pulses on the specimen depend on many effects related to electron-electron interactions such as space charge limited current, Boersch effect, emission angles and trajectories, or filtering effects due to chromatic aberration of lenses. It is crucial to understand these as a function of the relevant experimental parameters (position and shape of the photocathode, laser power, Wehnelt bias) in order to optimize spatial and temporal resolution while preserving reasonable acquisition times.
Our experimental setup consists of a JEOL 2100 with a thermionic gun and Wehnelt electrode, combined with a femtosecond fiber laser. Measurements are based on the PINEM effect (photon-induced near-field electron microscopy), which occurs when pump and probe pulses are synchronized at the sample. It results in a change of the electron energy distribution due to inelastic electron scattering by the photonic near field around a sample that can be observed by EELS.
The ability of the pump-probe setup to precisely measure the arrival time of the electrons allows a deeper understanding of the emission pattern. A conical tantalum cathode with flattened tip positioned close to the opening of the Wehnelt shows two electron populations, i.e., an intense big halo and a central spot (figure 2). The arrival time of electrons from the outer halo is shifted with respect to the central spot; the time difference changes with the applied Wehnelt bias. These measurements enable us to decipher the emission areas and electron trajectories. The halo is attributed to shank emission from the side wall of the cone where electrons leave at larger angles. The central spot are electrons emitted from the flattened tip. Larger Wehnelt gaps cut shank emission so that only electrons from the tip reach the specimen. Here the emission resembles the one from a Ta disc where all electrons are emitted from the flat surface at any Wehnelt gap.
Furthermore, PINEM scans were measured at different pump-probe delays, giving the temporal evolution of the electron pulse. Repeating these scans at different experimental settings (UV intensity, Wehnelt bias) allows to extract Δt and ΔE (figure 3). For instance, an increasing UV power allows to shorten the acquisition time but increases space charge and Boersch effect. At high Wehnelt bias the energy width (ΔE) is narrow, allowing good spatial and energy but lower temporal resolution. A low bias gives the opposite: good temporal but lower energy resolution.
Such understanding of the electron dynamics allows us to define optimal settings for time-resolved experiments, which are always a compromise between temporal, spatial, and energy resolution as well as acquisition times. The detailed beam characteristics will be presented.
Figures:

Fig. 1: a) Schematic drawing of the ultrafast electron microscope. An infrared femtosecond laser is split into two beams. One is directly focused onto the specimen as pump pulse. The other is frequency-quadruplicated to UV and focused onto the photocathode to generate the electron pulses via photoemission. b) Photograph of the setup.

Fig. 2: Emission pattern of a flattened conical tip at different bias on the Wehnelt electrode. The arrival time of each population is shown in d) as a function of the applied bias. Electrons from the halo arrive later than electrons on-axis; the time difference shrinks with increasing bias until saturation.

Fig. 3: Electron pulse analysis of a disc-shaped filament. Temporal width as a function of the energy distribution. PINEM scans were taken at different Wehnelt bias, each with varying UV power.
To cite this abstract:
Kerstin Bücker, Matthieu Picher, Olivier Crégut, Thomas LaGrange, Bryan Reed, Sang Tae Park, Dan Masiel, Florian Banhart; Ultrafast transmission electron microscopy reveals electron dynamics and trajectories in a thermionic gun setup. The 16th European Microscopy Congress, Lyon, France. https://emc-proceedings.com/abstract/ultrafast-transmission-electron-microscopy-reveals-electron-dynamics-and-trajectories-in-a-thermionic-gun-setup/. Accessed: March 2, 2021« Back to The 16th European Microscopy Congress 2016
EMC Abstracts - https://emc-proceedings.com/abstract/ultrafast-transmission-electron-microscopy-reveals-electron-dynamics-and-trajectories-in-a-thermionic-gun-setup/