Abstract: Quantitative atomic resolution annular dark field scanning transmission electron microscopy (ADF STEM) has become a powerful technique for nanoparticle atom-counting. However, a lot of nanoparticles provide a severe characterisation challenge because of their limited size and beam sensitivity. Therefore, quantitative ADF STEM may greatly benefit from statistical detection theory in order to optimise the instrumental microscope settings such that the incoming electron dose can be kept as low as possible whilst still retaining single-atom precision. The principles of detection theory are used to quantify the probability of error for atom-counting. This enables us to decide between different image performance measures and to optimise the experimental detector settings for atom-counting in ADF STEM in an objective manner. To demonstrate this, ADF STEM imaging of an industrial catalyst has been conducted using the near-optimal detector settings. For this experiment, we discussed the limits for atomcounting diagnosed by combining a thorough statistical method and detailed image simulations.

Abstract: Annular dark-field (ADF) scanning transmission electron microscopy (STEM) has become widely used in quantitative studies based on the opportunity to directly compare experimental and simulated images. This comparison merely requires the experimental data to be normalised and expressed in units of fractional beam-current. However, inhomogeneities in the response of electron detectors can complicate this normalisation. The quantification procedure becomes both experiment and instrument specific, requiring new simulations for the particular response of each instrument's detector, and for every camera-length used. This not only impedes the comparison between different instruments and research groups, but can also be computationally very time consuming. Furthermore, not all image simulation methods allow for the inclusion of an inhomogeneous detector response. In this work, we propose an alternative method for normalising experimental data in order to compare these with simulations that consider a homogeneous detector response. To achieve this, we determine the electron flux distribution reaching the detector by means of a camera-length series or a so-called atomic column cross-section averaged convergent beam electron diffraction (XSACBED) pattern. The result is then used to determine the relative weighting of the detector response. Here we show that the results obtained by this new electron flux weighted (EFW) method are comparable to the currently used method, while considerably simplifying the needed simulation libraries. The proposed method also allows one to obtain a metric that describes the quality of the detector response in comparison with the ideal detector response.

Abstract: In the present paper, the principles of detection theory are used to quantify the probability of error for atom-counting from high resolution scanning transmission electron microscopy (HR STEM) images. Binary and multiple hypothesis testing have been investigated in order to determine the limits to the precision with which the number of atoms in a projected atomic column can be estimated. The probability of error has been calculated when using STEM images, scattering cross-sections or peak intensities as a criterion to count atoms. Based on this analysis, we conclude that scattering cross-sections perform almost equally well as images and perform better than peak intensities. Furthermore, the optimal STEM detector design can be derived for atom-counting using the expression for the probability of error. We show that for very thin objects LAADF is optimal and that for thicker objects the optimal inner detector angle increases.

Abstract: Zinc is routinely employed in the synthesis of InP quantum dots (QDs) to improve the photoluminescence efficiency and carrier mobility of the resulting In(Zn)P alloy nanostructures. The exact location of Zn in the final structures and the mechanism by which it enhances the optoelectronic properties of the QDs is debated. We use synchrotron X-ray absorbance spectroscopy to show that the majority of Zn in In(Zn)P QDs is located at their surface as Zn-carboxylates. However, a small amount of Zn is present inside the bulk of the QDs with the consequent contraction of their lattice, as confirmed by combining high resolution high-angle annular dark-field imaging scanning transmission electron microscopy (HAADF-STEM) with statistical parameter estimation theory. We further demonstrate that the Zn content and its incorporation into the QDs can be tuned by the ligation of commonly employed Zn carboxylate precursors: the use of highly reactive Zn-acetate leads to the formation of undesired Zn3P2 and the final nanostructures being characterized by broad optical features, whereas Zn-carboxylates with longer carbon chains lead to InP crystals with much lower zinc content and narrow optical features. These results can explain the differences between structural and optical properties of In(Zn)P samples reported across the literature, and provide a rational method to tune the amount of Zn in InP nanocrystals and to drive the incorporation of Zn either as surface Zn-carboxylate, as a substitutional dopant inside the InP crystal lattice, or even predominantly as Zn3P2.

Abstract: By combining statistical parameter estimation and model-order selection using a Bayesian framework, the maximum a posteriori (MAP) probability rule is proposed in this chapter as an objective and quantitative method to detect atom columns from high-resolution scanning transmission electron microscopy (HRSTEM) images. The validity and usefulness of this approach is demonstrated to both simulated and experimental annular dark-field (ADF) STEM images, but also to simultaneously acquired annular bright-field (ABF) and ADF STEM image data.

Abstract: An efficient model-based estimation algorithm is introduced to quantify the atomic column positions and intensities from atomic-resolution (scanning) transmission electron microscopy ((S)TEM) images. This algorithm uses the least squares estimator on image segments containing individual columns fully accounting for overlap between neighboring columns, enabling the analysis of a large field of view. To provide end-users with this well-established quantification method, a user friendly program, StatSTEM, is developed which is freely available under a GNU public license. In this chapter, this efficient algorithm is applied to three different nanostructures for which the analysis of a large field of view is required.

Abstract: The maximum a posteriori (MAP) probability rule for atom column detection can also be used as a tool to evaluate the relation between scanning transmission electron microscopy (STEM) image quality and atom detectability. In this chapter, a new image-quality measure is proposed that correlates well with atom detectability, namely the integrated contrast-to-noise ratio (ICNR). Furthermore, the working principle of the MAP probability rule is described in detail showing a close relation to the principles of model-selection methods.

Abstract: In this chapter, the principles of detection theory are used to quantify the probability of error for atom counting from high-resolution scanning transmission electron microscopy (HRSTEM) images. Binary and multiple hypothesis testing have been investigated in order to determine the limits to the precision with which the number of atoms in a projected atomic column can be estimated. The probability of error has been calculated when using STEM images, scattering cross-sections or peak intensities as a criterion to count atoms. Based on this analysis, we conclude that scattering cross-sections perform almost equally well as images and perform better than peak intensities. Furthermore, the optimal STEM detector design can be derived for atom counting using the expression of the probability of error. We show that for very thin objects the low-angle annular dark-field (LAADF) regime is optimal and that for thicker objects the optimal inner detector angle increases.

Abstract: Advanced materials often consist of multiple elements which are arranged in a complicated structure. Quantitative scanning transmission electron microscopy is useful to determine the composition and thickness of nanostructures at the atomic scale. However, significant difficulties remain to quantify mixed columns by comparing the resulting atomic resolution images and spectroscopy data with multislice simulations where dynamic scattering needs to be taken into account. The combination of the computationally intensive nature of these simulations and the enormous amount of possible mixed column configurations for a given composition indeed severely hamper the quantification process. To overcome these challenges, we here report the development of an incoherent non-linear method for the fast prediction of ADF-EDX scattering cross-sections of mixed columns under channelling conditions. We first explain the origin of the ADF and EDX incoherence from scattering physics suggesting a linear dependence between those two signals in the case of a high-angle ADF detector. Taking EDX as a perfect incoherent reference mode, we quantitatively examine the ADF longitudinal incoherence under different microscope conditions using multislice simulations. Based on incoherent imaging, the atomic lensing model previously developed for ADF is now expanded to EDX, which yields ADF-EDX scattering cross-section predictions in good agreement with multislice simulations for mixed columns in a core–shell nanoparticle and a high entropy alloy. The fast and accurate prediction of ADF-EDX scattering cross-sections opens up new opportunities to explore the wide range of ordering possibilities of heterogeneous materials with multiple elements.

Abstract: Nanoparticle self-assembly and epitaxy are utilized extensively to make 1D and 2D structures with complex shapes. High-resolution transmission electron microscopy (HRTEM) has shown that single-crystalline interfaces can form, but little is known about the strain and dislocations at these interfaces. Such information is critically important for applications: drastically reducing
dislocation density was the key breakthrough enabling widespread implementation of light-emitting diodes, while strain engineering has been fundamental to modern high-performance transistors, solar cells, and thermoelectrics. In this work, the interfacial defect and strain formation after selfassembly and room temperature epitaxy of 7 nm Pd nanocubes capped with polyvinylpyrrolidone (PVP) is examined. It is observed that, during ligand removal, the cubes move over large distances on the substrate, leading to both spontaneous self-assembly and epitaxy to form single crystals. Subsequently, atomically resolved images are used to quantify the strain and dislocation density at the epitaxial interfaces between cubes with different lateral and angular misorientations. It is shown that dislocation- and strain-free interfaces form when the nanocubes align parallel to each other. Angular misalignment between adjacent cubes does not necessarily lead to grain boundaries but does cause dislocations, with higher densities associated with larger rotations.

Abstract: We report an innovative method to explore the optimal experimental settings to detect light atoms from scanning transmission electron microscopy (STEM) images. Since light elements play a key role in many technologically important materials, such as lithium-battery devices or hydrogen storage applications, much effort has been made to optimize the STEM technique in order to detect light elements. Therefore, classical performance criteria, such as contrast or signal-to-noise ratio, are often discussed hereby aiming at improvements of the direct visual interpretability. However, when images are interpreted quantitatively, one needs an alternative criterion, which we derive based on statistical detection theory. Using realistic simulations of technologically important materials, we demonstrate the benefits of the proposed method and compare the results with existing approaches.

Abstract: A thorough understanding of the three-dimensional (3D) atomic structure and composition of coreshell nanostructures is indispensable to obtain a deeper insight on their physical behavior. Such 3D information can be reconstructed from two-dimensional (2D) projection images using electron tomography. Recently, different electron tomography techniques have enabled the 3D characterization of a variety of nanostructures down to the atomic level. However, these methods have all focused on the investigation of nanomaterials containing only one type of chemical element. Here, we combine statistical parameter estimation theory with compressive sensing based tomography to determine the positions and atom type of each atom in heteronanostructures. The approach is applied here to investigate the interface in coreshell Au@Ag nanorods but it is of great interest in the investigation of a broad range of nanostructures.

Abstract: The three-dimensional (3D) atomic structure of nanomaterials, including strain, is crucial to understand their properties. Here, we investigate lattice strain in Au nanodecahedra using electron tomography. Although different electron tomography techniques enabled 3D characterizations of nanostructures at the atomic level, a reliable determination of lattice strain is not straightforward. We therefore propose a novel model-based approach from which atomic coordinates are measured. Our findings demonstrate the importance of investigating lattice strain in 3D.

Abstract: In the present paper, a statistical model-based method to count the number of atoms of monotype crystalline nanostructures from high resolution high-angle annular dark-field (HAADF) scanning transmission electron microscopy (STEM) images is discussed in detail together with a thorough study on the possibilities and inherent limitations. In order to count the number of atoms, it is assumed that the total scattered intensity scales with the number of atoms per atom column. These intensities are quantitatively determined using model-based statistical parameter estimation theory. The distribution describing the probability that intensity values are generated by atomic columns containing a specific number of atoms is inferred on the basis of the experimental scattered intensities. Finally, the number of atoms per atom column is quantified using this estimated probability distribution. The number of atom columns available in the observed STEM image, the number of components in the estimated probability distribution, the width of the components of the probability distribution, and the typical shape of a criterion to assess the number of components in the probability distribution directly affect the accuracy and precision with which the number of atoms in a particular atom column can be estimated. It is shown that single atom sensitivity is feasible taking the latter aspects into consideration.

Abstract: High angle annular dark field scanning transmission electron microscopy (HAADF STEM) images provide sample information which is sensitive to the chemical composition. The image intensities indeed scale with the mean atomic number Z. To some extent, chemically different atomic column types can therefore be visually distinguished. However, in order to quantify the atomic column composition with high accuracy and precision, model-based methods are necessary. Therefore, an empirical incoherent parametric imaging model can be used of which the unknown parameters are determined using statistical parameter estimation theory (Van Aert et al., 2009, [1]). In this paper, it will be shown how this method can be combined with frozen lattice multislice simulations in order to evolve from a relative toward an absolute quantification of the composition of single atomic columns with mixed atom types. Furthermore, the validity of the model assumptions are explored and discussed.