STEM-in-SEM of biological specimens in correlation with TEM

Abstract text (incl. figure legends and references)

1. Introduction

Scanning transmission electron microscopy in scanning electron microscopes (STEM-in-SEM) at low primary electron energies (≤ 30 keV) yields a pronounced image contrast for weakly scattering biological materials [1] and provides a complementary technique to transmission electron microscopy (TEM) at high electron energies. SEMs are widely available and can be equipped with a STEM detector. Further advantages compared to TEM are reduced maintenance costs and less demanding alignment procedures. STEM-imaging can be combined with secondary electron imaging, rendering STEM-in-SEM a multimodal technique for imaging both surface and subsurface features [2].

2. Objectives

In this work, the competitiveness of STEM-in-SEM as an alternative to TEM is explored. The two techniques are compared with respect to their resolution and signal-to-noise ratio.

3. Materials and Methods

Thin microtome slices of A549 lung carcinoma cells exposed to silica nanoparticles (NPs) are used for 200 keV TEM imaging in a Philips CM200. Bright-field (BF) STEM-in-SEM images are taken from the same region of interest in a FEI Helios G4 FX at 30 keV. The two techniques are compared by evaluating the spatial resolution and the contrast-to-noise-ratio CNR = (S-B)/σ_B (S: signal, B: background intensity, σ_B: standard deviation of B) of the images. As the contrast in TEM depends also on defocus and aperture settings, the beam convergence and detection angles for the images in Figure 2 are chosen such that similar image signals can be expected due to the reciprocity theorem for STEM and TEM [3].

4. Results

Figure 1 shows images of the same region of a cell section with NPs. To improve the contrast, the TEM image (Figure 1a) was taken with a defocus 1.5 μm. A line scan over a small feature reveals a feature size of 2 nm, estimated by the full-width-at-half-maximum (FWHM). The same feature in the BF-STEM image (Fig. 1b) also shows a resolution of 2 nm.

The evaluation of the CNR is shown in Figure 2. A line scan through one of the dark particles reveals a more pronounced noise in the TEM image (Figure 2a) with CNR=3. The 30 keV BF-STEM image in Figure 1b shows a reduced noise with CNR=20.

4. Conclusions

STEM-in-SEM is a powerful technique for analyzing biological specimens, combining good CNR with a high spatial resolution. The resolution is sufficient for many applications making STEM-in-SEM well competitive to TEM.

Figure 1. Spatial resolution of 2 nm estimated by the FWHM of a linescan over a small feature. a) TEM image with a defocus of 1.5 µm. b) BF-STEM image of the same region.

Figure 2. Line scan over a NP. a) CNR=3 for the TEM image. b) CNR=20 for the BF-STEM image

[1] U Kaiser et al., Ultramicroscopy 111 (2011), p. 1239. doi: 10.1016/j.ultramic.2011.03.012

[2] C Sun et al., J Mater Sci 55 (2020), p. 13824. doi: 10.1007/s10853-020-04970-3

[3] CB Carter and DB Williams, "Transmission Electron Microscopy", Plenum Press, New York, 1996