Interpretability of high-resolution transmission electron microscopy images.

High-resolution electron microscopy is a well-suited tool for characterizing the nanoscale structure of materials. However, the interaction of the sample and the high-energy electrons of the beam can often have a detrimental impact on the sample structure. This effect can only be alleviated by decreasing the number of electrons to which the sample is exposed but will come at the cost of a decreased signal-to-noise ratio in the resulting image. Images with low signal to noise ratios are often challenging to interpret as parts of the sample with a low interaction with the electron beam are reproduced with very low contrast. Here we suggest simple measures as alternatives to the conventional signal-to-noise ratio and investigate how these can be used to predict the interpretability of the electron microscopy images. We test the models on a sample consisting of gold nanoparticles supported on a cerium dioxide substrate. The models are evaluated based on series of images acquired at varying electron dose.

Beam induced heating in electron microscopy modeled with machine learning interatomic potentials.

We develop a combined theoretical and experimental method for estimating the amount of heating that occurs in metallic nanoparticles that are being imaged in an electron microscope. We model the thermal transport between the nanoparticle and the supporting material using molecular dynamics and equivariant neural network potentials. The potentials are trained to Density Functional Theory (DFT) calculations, and we show that an ensemble of potentials can be used as an estimate of the errors the neural network make in predicting energies and forces. This can be used both to improve the networks during the training phase, and to validate the performance when simulating systems too big to be described by DFT. The energy deposited into the nanoparticle by the electron beam is estimated by measuring the mean free path of the electrons and the average energy loss, both are done with Electron Energy Loss Spectroscopy (EELS) within the microscope. In combination, this allows us to predict the heating incurred by a nanoparticle as a function of its size, its shape, the support material, and the electron beam energy and intensity.

Quantifying noise limitations of neural network segmentations in high-resolution transmission electron microscopy.

Motivated by the need for low electron dose transmission electron microscopy imaging, we report the optimal frame dose (i.e.e-/Å2) range for object detection and segmentation tasks with neural networks. The MSD-net architecture shows promising abilities over the industry standard U-net architecture in generalising to frame doses below the range included in the training set, for both simulated and experimental images. It also presents a heightened ability to learn from lower dose images. The MSD-net displays mild visibility of a Au nanoparticle at 20-30 e-/Å2, and converges at 200 e-/Å2 where a full segmentation of the nanoparticle is achieved. Between 30 and 200 e-/Å2 object detection applications are still possible. This work also highlights the importance of modelling the modulation transfer function when training with simulated images for applications on images acquired with scintillator based detectors such as the Gatan Oneview camera. A parametric form of the modulation transfer function is applied with varying ranges of parameters, and the effects on low electron dose segmentation is presented.

Reconstructing the exit wave of 2D materials in high-resolution transmission electron microscopy using machine learning.

Reconstruction of the exit wave function is an important route to interpreting high-resolution transmission electron microscopy (HRTEM) images. Here we demonstrate that convolutional neural networks can be used to reconstruct the exit wave from a short focal series of HRTEM images, with a fidelity comparable to conventional exit wave reconstruction. We use a fully convolutional neural network based on the U-Net architecture, and demonstrate that we can train it on simulated exit waves and simulated HRTEM images of graphene-supported molybdenum disulphide (an industrial desulfurization catalyst). We then apply the trained network to analyse experimentally obtained images from similar samples, and obtain exit waves that clearly show the atomically resolved structure of both the MoS2 nanoparticles and the graphene support. We also show that it is possible to successfully train the neural networks to reconstruct exit waves for 3400 different two-dimensional materials taken from the Computational 2D Materials Database of known and proposed two-dimensional materials.