Advancing Time-Resolved Spectroscopies with Custom Scanning Units and Event-Based Electron Detection

Abstract

Direct electron detection is revolutionizing electron microscopy by offering lower noise, reduced point-spread function, and increased quantum efficiency. Among these advancements, the Timepix3 hybrid-pixel direct electron detector stands out for its unique ability to output temporal information about individual hits within its pixel array. Its event-based architecture enables data-driven detection, where individual events are immediately read out from the chip. While recent studies have demonstrated the potential of event-based detectors in various applications in the context of continuous-gun scanning transmission electron microscopes (STEM), the use of such detectors with standard scanning units remains underdeveloped. In this work, we present a custom-designed, Timepix3-compatible scanning unit specifically developed to leverage the advantages of event-based detection. We explore its performance in enabling spatially and temporally resolved experiments, as well as its seamless interfacing with other time-resolved instruments, such as pulsed lasers and electron beam blankers. Additionally, we examine the prospects for achieving enhanced temporal resolution in time-resolved experiments using continuous-gun electron microscopes, identifying key challenges and proposing solutions to improve performance. This combination of custom hardware with advanced detection technology promises to expand the capabilities of electron microscopy in both fundamental research and practical applications.

Figures (9)

• Figure 1: The time-dependent scanning unit in scanning transmission electron microscopes. (A) Scheme of the synchronous DACx driving the microscope coils using an amplifier circuit. The digital inputs come from a digital logic circuit, such as an FPGA. (B) The time-dependent waveforms of a normal rastering scan for both independent scanning directions. (C) Data is shown to the user as function of the electron position $(x', y')$. Although time-dependency is not explicit, both $y'(t)$ and $x'(t)$ are fundamental waveforms defined by the scanning pattern.
• Figure 2: Scheme of the Timepix3 and the scanning unit interface. (A) Schematic of the STEM is shown, including the scanning unit, an EELS spectrometer followed by a Timepix3, and a parabolic mirror for light detection (CL) and injection (EEGS). The scanning unit is responsible for driving the electron probe position of the STEM, as well as acquiring analog signals, typically from a photo-multiplier tube. By putting Timepix3 in the same clock domain, the time-of-arrival of the electron hits in the detector can be related to the current position of the focused electron probe. Several experiments can be performed using this Timepix3 - SU architecture, including (B) fast hyperspectroscopy or 4D-imaging using arbitrary but well-defined time-dependent waveforms in each one of the microscope scanning coils, (C) Electron-photon coincidence measurements, by either inputting an external sample stimulus, or by receiving it, e.g. CL photons and EELS electrons, and (D) time-resolved light-injection experiments, such as in nanothermometry or in electron energy-gain spectroscopy.
• Figure 3: Comparison between frame and event-based in Lissajous scanning for a fixed total frame time. Lissajous scanning can be used with frame-based detectors, but the minimum attainable detector dwell time limits the highest achievable Lissajous frequency. When a nanosecond-resolved electron detector is used, higher frequencies become possible. In the frame-based case, the electron position is fixed until the detector frame acquisition is finished, and the pixel dwell time is 30 times longer than in the event-based case, also indicated by the radius of the scatter sphere.
• Figure 4: Lissajous-based scanning patterns. (A) A series of Lissajous scans ($512 \times 512$ pixels) and their respective sinusoidal frequencies, from 5 kHz up to 120 kHz, displayed in real time to the user. Despite some degradation of the image, its features remain visible. However, at 120 kHz, the loss of reference prevented proper calibration. (B) Due to the circuit bandwidth, coil impedance rises enough to reduce the current through the coil, reducing the field of view, as can be seen by the curve. In the inset, the image of the region of interested obtained by normal rastering pattern.
• Figure 5: Comparison between Lissajous scanning and normal rastering. Lissajous scanning enables rapid, multi-resolution imaging. As the scan progresses, the spatial resolution improves in Lissajous scanning by collecting more sample points. The imaging process transitions from sparse to dense sampling, allowing traditional sparse-sampling techniques, such as inpainting and interpolation, to be applied during the initial phase of image acquisition. In both scanning patterns, the total deposited dose is the same, but because of the flyback time, normal rastering is $\sim$ 30% slower. Lissajous scanning operates at approximately 8 kHz. Scale bars are 100 nm.