Investigation of non-linear imaging in high-resolution transmission electron microscopy

Exit wave function reconstruction from two defocus images using neural network

Wave function reconstruction from one or two defocus images is promising for live atomic resolution imaging in transmission electron microscopy. However, a robust and accurate reconstruction method we still need more attention. Here, we present a neural-network-based wave function reconstruction method, EWR-NN, that enables accurate wave function reconstruction from only two defocus images. Results from both simulated and two different experimental defocus series show that the EWR-NN method has better performance than the widely-used iterative wave function reconstruction (IWFR) method. Influence of image number, defocus deviation, residual image shifts and noise level were considered to validate the performance of EWR-NN under practical conditions. It is seen that these factors will not influence the arrangement of atom columns in the reconstructed phase images, while they can alter the absolute values of all-atom columns and degrade the contrast of the phase images.

An improved iterative wave function reconstruction algorithm in high-resolution transmission electron microscopy

Exit wavefunction reconstruction is a powerful image processing technique to enhance the resolution and the signal-to-noise ratio for atomic-resolution imaging in both aberration uncorrected and corrected transmission electron microscopes. The present study aims to improve the performance of the iterative wavefunction reconstruction algorithm in comparison not only with its conventional form but also with the popular commercial Trueimage software for exit wavefunction reconstruction. It is shown that by implementing a wave propagation procedure for refining its image alignment, the iterative wavefunction reconstruction algorithm can be greatly improved in accurately retrieving the wavefunctions while keeping its original advantages, which allow the reconstruction be performed with less images and a larger defocus step in the data set of through-focus image series. In addition, calculations of this algorithm can be accelerated drastically by the graphic processing unit (GPU) hardware programming using the popular computer unified device architecture language, whose computing speed can be 25-38 times as fast as a central processing unit (CPU) program.

Applicability of non-linear imaging in high-resolution transmission electron microscopy

According to transmission cross-coefficient theory, the information limit of non-linear imaging in high-resolution transmission electron microscopy is, under certain conditions, far beyond that of linear imaging, which suggests the possibility of using high-frequency information for structural determination. In this article, we studied the information beyond the linear information limit by means of multislice method simulation, with AlN as an example, and more structural information was obtained by using part of the high-frequency information.