Pulsed Electron Illumination and Beam Deflection Transfer Function Measurement using Multi-Trigger < 1 μs Exposures on the Merlin – Medipix Detector

Recent studies have shown that using a pulsed electron beam illumination, where the sample receives few-electron illumination bursts within ∼1 μs windows that are separated by much greater intervals (∼ 20 μs), leads to a decrease in sample damage for a given electron dosage in transmission electron microscope (TEM) observations [1]. Deceasing sample damage for a given dose, thus increasing effective real signal to noise, is important to characterize many important technological and biological materials, such as polymers, pharmaceuticals, proteins, and perovskites, which have particularly low critical electron dose thresholds. Conventionally such fast and repetitive exposures in an electron microscope are obtained using either pulsed laser stimulated electron emission, or perhaps a fast electrostatic deflector or shutter. However, here we explore an alternative scheme using a conventional electromagnetic beam deflection in a STEM / TEM combined with using the synchronous fast (< 1 μs) multi-trigger exposure mode of the Merlin detector [2]. Here the sample is placed in a section of an orbital path of the electron beam illumination, depicted in Figure 1(a), where the beam orbits at a frequency in the region of 50 kHz. The Merlin pixelated direct electron detector is synchronously triggered to only count electrons during each ∼1 μs or less interval when the beam passes over the region of interest in the sample. Each multi-trigger frame retrieved from the Medipix detector thus contains the sum of thousands of individual few electron number illuminations of the sample region. Thus, extraneous background signal is excluded from the collected diffraction data or image. For attainable beam current conditions in our S/TEM microscope (Hitachi HF-3300V), the simulated distribution of the time intervals between successive electron arrivals (Figure 1(b)) shows a non-exponential bound, giving the beam a distinctly different statistical characteristic to just reducing the beam current in a normal TEM illumination mode. For this example of a 50 kHz orbit, the illumination (Figure 1(c)) is dominated by single electron hits separated by at least 20 μs. This scheme was also used without a sample to directly determine the amplitude and phase response of the beam deflection to the voltage applied the STEM scan amplifier. This was achieved by projecting a magnified image of the sample plane on to the detector while supplying a sinusoidal a drive voltage to the scan amplifier in order to deflect a focused electron beam over the sample. This provided a direct measure of the drive-voltage to beam deflection transfer characteristic, including the phase delay, as presented in Figure 2. This direct measurement method avoids having to attempt to measure currents in the manufacturer supplied scan amplifier for example and avoids the need to use any models of the behavior of magnetic material in the vicinity of deflection coils. Furthermore, in comparison to knife edge frequency response methods, which we also investigated here, the Merlin-Medipix determined transfer function has much greater repeatability and hence reliability. A measured transfer characteristic for a fixed magnification mode in STEM is shown in Figure 2. While we see that the amplitude response follows that of a simple second order response system, the phase response does not match well with a simple second order filter model. Knowing the details of the experimentally measured transfer functions allow us to operate at the upper limits of the frequency scan system, and know phase offset that must be applied to our trigger signal to successfully acquire in the desired region of interest in the sample. While experiments have not yet been performed to determine whether sample damage reduction can be achieved with this illumination scheme, we have outlined, modelled and experimentally demonstrated aspects of a scheme that can achieve a sample illumination mode similar (though not identical) to pulsed laser assisted electron emission gun systems. It achieves this without requiring a specialized electron gun. We envisage that this scheme thus has the potential to help with imaging and characterization of highly beam sensitive materials in the TEM. (a) Schematic of experimental arrangement; (b) simulated distribution of times between successive electron hits in the sample; and (c) simulated illumination over the sample region for a 1 second low-dose (< 1 e- / Å) exposure. Measured beam-deflection to scan-voltage ratio vs frequency characteristic, determined using the arrangement shown in Figure 1.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.