Practical Electron Microscopy and Database

The current direct electron detectors are mainly produced by three companies: Gatan, FEI, and Direct Electron. These detectors, often referred to as Direct Detection Devices (DDD), yield significantly better images compared to film. All three types utilize similar sensor technology, where electrons strike a lightly doped silicon epilayer supported by a more heavily doped silicon substrate. Each frame of the exposure is continuously read out in a rolling-shutter mode, creating a “movie” of the capture process.

K2 and its application:
         Raw electron diffraction patterns can be collected using a K2 direct electron detector at a rate of 400 frames per second. 4DSTEM can be acquired by scanning a 6 nm electron probe across the sample in a two-dimensional grid with a 2.5 ms dwell time per step and a 20 nm step size. The electrostatic lens above the probe-forming aperture can be adjusted to achieve a fluence of approximately 1 electron per square angstrom (~e⁻ Å⁻²) per scan step. Diffraction patterns can be recorded by using a Gatan K2-IS direct electron detector, with an effective camera length of 575 mm. Super-resolution movies can be recorded using a Gatan K2 Summit detector, with a nominal pixel size of 1.04 Å per pixel (0.52 Å in super-resolution mode), and each movie, comprising 20 frames, deliveres a total electron dose of 27.8 e⁻ Å⁻² per image.

Figure 0011 shows the DQE values for integrating and counting modes, particularly the comparisons and the range of 40–60% at half Nyquist and around 25% at Nyquist frequency. This figure compares DQE at 300 keV across different detectors, showing variations in DQE as a function of spatial frequency for models such as the DE-20, Falcon II, and K2 Summit. On the other hand, the frame rate of the K2 detector is approximately ten times faster than those of the other two detector brands.

Table 0011 lists the typical microscope parameters used for acquisition of high-resolution images and movies with the three direct electron detector types. Each detector has an optimal dose rate where DQE reaches its peak. At higher dose rates, all detectors tend to saturate, rendering each frame effectively devoid of useful information. Conversely, at very low dose rates, multiple frames must be combined to achieve sufficient exposure, but these combined frames often become dominated by electronic readout noise. In practical terms, the DQE as a function of dose rate is relatively broad, allowing for a range of acceptable exposures. The saturating dose rate is significantly higher—by several hundred times—when the detector operates in integrating mode compared to counting mode. This allows both the Falcon and DE detectors to function in integrating mode at dose rates up to 2 or 3 electrons per pixel per frame, while in counting mode, the dose rate should remain below 0.01–0.025 electrons per pixel per frame. [1] In practice, the total exposure time for the Falcon and DE detectors typically falls within 1–3 seconds. In contrast, the K2 detector generally requires exposures between 6 and 16 seconds, and potentially longer with larger pixel sizes. As a result, the Falcon and DE detectors are well-suited for microscopes with less stable cold stages or notable stage drift.

[1] R.A. Crowther, The Resolution Revolution: Recent Advances In cryoEM, MRC Laboratory of Molecular Biology, Cambridge, United Kingdom, 579, pp.2-445 (2016).