Imaging built-in electric fields and light matter by Fourier-precession TEM

www.nature.com/scientificreports OPEN Imaging built‑in electric fields and light matter by Fourier‑precession TEM Tizian Lorenzen 1,4, Benjamin März 1,2,4, Tianhao Xue 1, Andreas Beyer 3, Kerstin Volz 3, Thomas Bein 1 & Knut Müller‑Caspary 1* We report the precise measurement of electric fields in nanostructures, and high-contrast imaging of soft matter at ultralow electron doses by transmission electron microscopy (TEM). In particular, a versatile method based on the theorem of reciprocity is introduced to enable differential phase contrast imaging and ptychography in conventional, plane-wave illumination TEM. This is realised by a series of TEM images acquired under different tilts, thereby introducing the sampling rate in reciprocal space as a tuneable parameter, in contrast to momentum-resolved scanning techniques. First, the electric field of a p–n junction in GaAs is imaged. Second, low-dose, in-focus ptychographic and DPC characterisation of Kagome pores in weakly scattering covalent organic frameworks is demonstrated by using a precessing electron beam in combination with a direct electron detector. The approach offers utmost flexibility to record relevant spatial frequencies selectively, while acquisition times and dose requirements are significantly reduced compared to the 4D-STEM counterpart. In recent years, nanostructured materials have come increasingly into the focus of research in the fields of information and energy technology. Porous organic materials with highly ordered structure and tunable functionalities, such as metal and covalent organic frameworks (MOFs/COFs) are investigated for their optoelectronic properties or applications in energy storage, catalysis and gas s torage1,2. Halide based perovskites are explored for their applications in solar cells and lasers3. The functional properties of such devices are fundamentally determined by the structure, i.e., the nanoscale particle shapes, pores and atomic configuration. Understanding the structure–property relationships is central when designing applications and searching for suitable candidate materials. In addition to deciphering the structure via transmission electron microscopy (TEM) with a spatial resolution down to a few tens of picometres, mapping the small built-in electric fields in semiconductor nanostructures such as p–n junctions4 remains a severe challenge for TEM. Conventional TEM imaging of light atoms in organic chemistry or structural biology always involves some form of compromise. Unfortunately, weakly scattering specimens show no image contrast in the absence of aberrations at zero defocus as they only shift the phase of the illuminating electron wave slightly. In these cases, deliberately introducing partly large aberrations through defocusing is effective in converting phase shifts into amplitude contrast. While this is widely accepted for improving contrast, it comes at the expense of image resolution and complicates direct interpretability. On top of that, organic and biological specimens are highly dosesensitive, making trustworthy structural imaging with a dose budget in the range of ten electrons per Å 2 very complicated. This is approximately three orders of magnitude less than in typical materials science applications. In the last decade efficient phase contrast generation has been developed in scanning TEM (STEM) by increasing the dimensionality of acquired data using segmented or pixelated detectors leading to 4D-STEM. In essence, 4D-STEM aims at collecting complete diffraction patterns at preferably high spatial frequency sampling, for each raster position (rx , ry ) of the electron probe in real space thereby leading to 4D-data sets. Differential phase contrast5–7 (DPC) or centre-of-mass (COM) imaging8, as well as a variety of ptychographic m ethods9–11 have proven to be effective phase contrast methods. Successful applications include the mapping of atomic and mesoscale electric12 and magnetic fields13, and high-contrast imaging of light matter at low electron dose14. Substantial efforts are currently put into the technological and conceptual development of multisegment D PC15 or pixelated detectors16–18, and into coping with the resulting tremendous data rates. Yet, today, fast STEM detectors 1 Department of Chemistry and Center for NanoScience, Ludwig-Maximilians-Universität München, Butenandtstr. 11, 81377 München, Germany. 2Louisiana State University Shared Instrumentation Facility (LSUSIF), 121 Chemistry and Materials Building, 4048 Highland Rd., Baton Rouge, LA 70803, USA. 3Department of Physics, Philipps University Marburg, Hans‑Meerwein‑Straße 6, 35032 Marburg, Germany. 4These authors contributed equally: Tizian Lorenzen and Benjamin März. *email: Scientific Reports | (2024) 14:1320 | https://doi.org/10.1038/s41598-024-51423-x 1 Vol.:(0123456789) www.nature.com/scientificreports/ still only have a limited number of segments while large pixelated detectors with 104–105 pixels are comparatively slow. This means that currently suitably large fields of view at sufficient real space samplings can only be achieved with DPC detectors with a few segments at most. In this work, motivated by the requirement of large-scale electric field mapping and high-contrast imaging of light matter, a technique to overcome these limitations is developed conceptually and in applications. Based on the theorem of reciprocity in optics, we demonstrate the imaging of built-in electric fields in a p–n junction, and the enhancement of low-dose image contrast in organic nanostructures, such as a COF. The kernel of the method involves the acquisition of sparse 4D data using conventional plane-wave illumination TEM to record real space images for different tilts of the incident electron beam. This is schematically shown in Fig. 1 and importantly maintains the large field of view of TEM. In the general field of microscopy including light-optical ptychography, this acquisition scheme is occasionally referred to as Fourier ptychography19. The obtained data is then subjected to advanced 4D-STEM evaluations such as DPC, COM and a variety of ptychographic algorithms. By combining the precession capability of a conventional TEM with an ultra-fast camera, acquisition of DPC data with 100 segments is demonstrated, for which otherwise 4–16 segments are currently common. In addition, the method overcomes the hardware-dictated sampling of diffraction patterns, restricts the electron dose to recording only those spatial frequencies that are expected to carry the most relevant information about the specimen, and does not suffer from hydrocarbon contamination arising from focused probes. Following the optical theorem of reciprocity, mapping the intensity of a diffraction coordinate k�⊥ in the STEM Ronchigram against the scan position is identical to recording a TEM image under plane wave illumination, tilted such that the lateral component of the wave vector equals k�⊥20–22. Considering an (...truncated)