Nanocrystal segmentation in scanning precession electron diffraction data.

Scanning precession electron diffraction (SPED) enables the local crystallography of materials to be probed on the nanoscale by recording a two-dimensional precession electron diffraction (PED) pattern at every probe position as a dynamically rocking electron beam is scanned across the specimen. SPED data from nanocrystalline materials commonly contain some PED patterns in which diffraction is measured from multiple crystals. To analyse such data, it is important to perform nanocrystal segmentation to isolate both the location of each crystal and a corresponding representative diffraction signal. This also reduces data dimensionality significantly. Here, two approaches to nanocrystal segmentation are presented, the first based on virtual dark-field imaging and the second on non-negative matrix factorization. Relative merits and limitations are compared in application to SPED data obtained from partly overlapping nanoparticles, and particular challenges are highlighted associated with crystals exciting the same diffraction conditions. It is demonstrated that both strategies can be used for nanocrystal segmentation without prior knowledge of the crystal structures present, but also that segmentation artefacts can arise and must be considered carefully. The analysis workflows associated with this work are provided open-source. LAY DESCRIPTION: Scanning precession electron diffraction is an electron microscopy technique that enables studies of the local crystallography of a broad selection of materials on the nanoscale. The technique involves the acquisition of a two-dimensional diffraction pattern for every probe position in an area of the sample. The four-dimensional dataset collected by this technique can typically comprise up to 500 000 diffraction patterns. For nanocrystalline materials, it is common that single diffraction patterns contain signals from overlapping crystals. To process such data, we use nanocrystal segmentation, where a representative diffraction pattern is constructed for each individual crystal, together with a real space image showing its morphology and location in the data. This reduces the dimensionality of the data and allows unmixing of signals from overlapping crystals. In this work, we demonstrate two methods for nanocrystal segmentation, one based on creating virtual dark-field images, and one based on unsupervised machine learning. A model system of partly overlapping nanoparticles is used to demonstrate the segmentation, and a demanding case for segmentation is highlighted, where some crystals are not discernible based on their diffraction patterns. To obtain a more complete nanocrystal segmentation, we add an image segmentation routine to both methods, and we discuss benefits and limitations of the two methods. The demonstration data and the used code are provided open-source, so that it can be used by everyone for analysis of nanocrystalline materials or as a starting point for further development of nanocrystal segmentation in scanning precession electron diffraction data.

Entropic Comparison of Atomic-Resolution Electron Tomography of Crystals and Amorphous Materials.

Electron tomography bears promise for widespread determination of the three-dimensional arrangement of atoms in solids. However, it remains unclear whether methods successful for crystals are optimal for amorphous solids. Here, we explore the relative difficulty encountered in atomic-resolution tomography of crystalline and amorphous nanoparticles. We define an informational entropy to reveal the inherent importance of low-entropy zone-axis projections in the reconstruction of crystals. In turn, we propose considerations for optimal sampling for tomography of ordered and disordered materials.

Electron tomography of III-V quantum dots using dark field 002 imaging conditions.

We present an evaluation of electron tomography of buried InAs quantum dots using dark field 002 imaging conditions. The compositional sensitivity of this imaging condition gives strong contrast among III-V materials of differing compositions and, in principle, should allow an accurate 3D model of the buried structures to be produced. The large extinction distance allows specimens several hundred nanometres in thickness to be examined and reduces the effect of strain contrast in the images, with the advantage that it can be performed using conventional transmission electron microscopy techniques. A two-beam condition must be maintained for all images, and the presence of other strong diffraction effects at certain specimen orientation results reduces the number of orientations available for tomography by approximately 10%. The data presented here are limited due to a lack of angular range in the data set but we find that an acceptable 3D model of a buried quantum dot may be produced by imposing cylindrical symmetry on the data set.

Quantitative off-axis electron holography of GaAs p-n junctions prepared by focused ion beam milling.

Focused Ion beam (FIB) prepared GaAs p-n junctions have been examined using off-axis electron holography. Initial analysis of the holograms reveals an experimentally determined built-in potential in the junctions that is significantly smaller than predicted from theory. In this paper we show that through combinations of in situ annealing and in situ biasing of the specimens, by varying the intensity of the incident electron beam, and by modifying the FIB operating parameters, we can develop an improved understanding of phenomena such as the electrically 'inactive' thickness and subsequently recover the predicted value of the built-in potential of the junctions. PACS numbers: 85.30.De.

Imaging flux vortices in type II superconductors with a commercial transmission electron microscope.

Flux vortices in superconductors can be imaged using transmission electron microscopy because the electron beam is deflected by the magnetic flux associated with the vortices. This technique has a better spatial and temporal resolution than many other imaging techniques and is sensitive to the magnetic flux density within each vortex, not simply the fields at the sample surface. Despite these advantages, only two groups have successfully employed the technique using specially adapted instruments. Here we demonstrate that vortices can be imaged with a modern, commercial transmission electron microscope operating at 300kV equipped with a field emission gun, Lorentz lens and a liquid helium cooled sample holder. We introduce superconductivity for non-specialists and discuss techniques for simulating and optimising images of flux vortices. Sample preparation is discussed in detail as the main difficulty with the technique is the requirement for samples with very large (>10microm), flat areas so that the image is not dominated by diffraction contrast. We have imaged vortices in superconducting Bi(2)Sr(2)CaCu(2)O(8-delta) and use correlation functions to investigate the ordered arrangements they adopt as a function of applied magnetic field.

3D imaging of nanomaterials by discrete tomography.

The field of discrete tomography focuses on the reconstruction of samples that consist of only a few different materials. Ideally, a three-dimensional (3D) reconstruction of such a sample should contain only one grey level for each of the compositions in the sample. By exploiting this property in the reconstruction algorithm, either the quality of the reconstruction can be improved significantly, or the number of required projection images can be reduced. The discrete reconstruction typically contains fewer artifacts and does not have to be segmented, as it already contains one grey level for each composition. Recently, a new algorithm, called discrete algebraic reconstruction technique (DART), has been proposed that can be used effectively on experimental electron tomography datasets. In this paper, we propose discrete tomography as a general reconstruction method for electron tomography in materials science. We describe the basic principles of DART and show that it can be applied successfully to three different types of samples, consisting of embedded ErSi(2) nanocrystals, a carbon nanotube grown from a catalyst particle and a single gold nanoparticle, respectively.

Fabrication and characterization of TiN-Ag nano-dice.

TiN-Ag nanocomposite was synthesized by dc arc-plasma method. Microstructures of TiN-Ag nanocomposite were carefully characterized by powder X-ray diffraction method and transmission electron microscopy, and nano-morphologies by three-dimensional electron tomography. It was found that the surface of nanocrystalline TiN matrix was densely covered by finely dispersed Ag nanoparticles, and it was found that they were physically attached but not chemically bonded from their orientation relationships.

3 D characterization of gold nanoparticles supported on heavy metal oxide catalysts by HAADF-STEM electron tomography.

Living on the edge: Three-dimensional reconstructions from electron tomography data recorded from Au/Ce(0.50)Tb(0.12)Zr(0.38)O(2-x) catalysts show that gold nanoparticles (see picture; yellow) are preferentially located on stepped facets and nanocrystal boundaries. An epitaxial relationship between the metal and support plays a key role in the structural stabilization of the gold nanoparticles.

TEM characterization of Ge precipitates in an Al-1.6at% Ge alloy.

The growth mechanism and morphology of Ge precipitates in an Al-Ge alloy was characterized by a combination of in-situ transmission electron microscopy, high-resolution transmission electron microscopy and three-dimensional electron tomography. Anisotropic growth of rod-shaped Ge precipitates was observed by in-situ transmission electron microscopy over different time periods, and faceting of the precipitates was clearly seen using high-resolution transmission electron microscopy and three-dimensional electron tomography. This anisotropic growth of rod-shaped Ge precipitates was enhanced by vacancy concentration as proposed previously, but also by surface diffusion as observed during the in-situ experiment. Furthermore, a variety of precipitate morphologies was identified by three-dimensional electron tomography.

Gold and iodine diffusion in large area perovskite solar cells under illumination.

Operational stability is the main issue hindering the commercialisation of perovskite solar cells. Here, a long term light soaking test was performed on large area hybrid halide perovskite solar cells to investigate the morphological and chemical changes associated with the degradation of photovoltaic performance occurring within the devices. Using Scanning Transmission Electron Microscopy (STEM) in conjunction with EDX analysis on device cross sections, we observe the formation of gold clusters in the perovskite active layer as well as in the TiO2 mesoporous layer, and a severe degradation of the perovskite due to iodine migration into the hole transporter. All these phenomena are associated with a drastic drop of all the photovoltaic parameters. The use of advanced electron microscopy techniques and data processing provides new insights on the degradation pathways, directly correlating the nanoscale structure and chemistry to the macroscopic properties of hybrid perovskite devices.

A novel 3D absorption correction method for quantitative EDX-STEM tomography.

This paper presents a novel 3D method to correct for absorption in energy dispersive X-ray (EDX) microanalysis of heterogeneous samples of unknown structure and composition. By using STEM-based tomography coupled with EDX, an initial 3D reconstruction is used to extract the location of generated X-rays as well as the X-ray path through the sample to the surface. The absorption correction needed to retrieve the generated X-ray intensity is then calculated voxel-by-voxel estimating the different compositions encountered by the X-ray. The method is applied to a core/shell nanowire containing carbon and oxygen, two elements generating highly absorbed low energy X-rays. Absorption is shown to cause major reconstruction artefacts, in the form of an incomplete recovery of the oxide and an erroneous presence of carbon in the shell. By applying the correction method, these artefacts are greatly reduced. The accuracy of the method is assessed using reference X-ray lines with low absorption.

Enhanced quantification for 3D SEM-EDS: using the full set of available X-ray lines.

An enhanced method to quantify energy dispersive spectra recorded in 3D with a scanning electron microscope (3D SEM-EDS) has been previously demonstrated. This paper presents an extension of this method using all the available X-ray lines generated by the beam. The extended method benefits from using high energy lines, that are more accurately quantified, and from using soft X-rays that are highly absorbed and thus more surface sensitive. The data used to assess the method are acquired with a dual beam FIB/SEM investigating a multi-element Ni-based superalloy. A high accelerating voltage, needed to excite the highest energy X-ray line, results in two available X-ray lines for several elements. The method shows an improved compositional quantification as well as an improved spatial resolution.

3D electron microscopy in the physical sciences: the development of Z-contrast and EFTEM tomography.

The rapid advances in nanotechnology and the ever decreasing size of features in the microelectronics industry brings with it the need for advanced characterisation with high spatial resolution in two and three dimensions. Stereo microscopy allows some insight into the three-dimensional nature of an object but for true quantitative analysis, one has to turn to tomography as a way to reconstruct a three-dimensional object from a series of two-dimensional projections (images). X-ray tomography allow structures to be imaged at relatively large length scales, atom probe tomography at the atomic level. Electron tomography offers an intermediate resolution (of about 1nm) with a field of view of hundreds of nm making it ideal for the characterisation of many nanoscale devices. Whilst electron tomography has been used in the biological sciences for more than 30 years, it is only now being applied to the physical sciences. In this paper, we review the status of electron tomography, describe the basis behind the technique and some of the practicalities of recording and analysing data for tomographic reconstruction, particularly in regard to solving three-dimensional problems that are encountered in materials science at the nanometre level. We present examples of how STEM dark-field imaging and energy-filtered TEM can be used successfully to examine nearly all types of specimens likely to be encountered by the physical scientist.

Image-spectroscopy--I. The advantages of increased spectral information for compositional EFTEM analysis.

The acquisition of a series of energy-filtered TEM images over the energy-loss range of interest creates a three-dimensional data set comprising both spatial and spectral information. Such an image-series contains energy-loss information not available with conventional two- or three-window methods, allowing standard techniques for quantitative EELS analysis to be applied to extracted 'image-spectra'. The increase in spectral information enables improved ionisation edge background extrapolation and interactive image-spectrum analysis to be performed. In this paper, the many advantages of the image-spectroscopy approach are outlined by reference to an example of elemental segregation in an AlZnMgCu alloy.

Image-spectroscopy--II. The removal of plural scattering from extended energy-filtered series by Fourier deconvolution.

The increased spectral information obtained by acquiring an EFTEM image-series over several hundred eV allows plural scattering to be removed from loss images using standard deconvolution techniques developed for the quantification of EEL spectra. In this work, both Fourier-log and Fourier-ratio deconvolution techniques have been applied successfully to such image-series. Application of the Fourier-log technique over an energy-loss range of several hundred eV has been achieved by implementation of a novel method that extends the effective dynamic range of EFTEM image-series acquisition by over four orders of magnitude. Experimental results show that the removal of plural scattering from EFTEM image-series gives a significant improvement in quantification for thicker specimen regions. Further, the recovery of the single-scattering distribution using the Fourier-log technique over an extended energy-loss range is shown to result in an increase in both the ionisation-edge jump-ratio and the signal-to-noise ratio.

Exploring the benefits of electron tomography to characterize the precise morphology of core-shell Au@Ag nanoparticles and its implications on their plasmonic properties.

In the design and engineering of functional core-shell nanostructures, material characterization at small length scales remains one of the major challenges. Here we show how electron tomography in high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) mode can be applied successfully to perform nano-metrological characterization of Au@Ag core-shell nanostructures. This work stresses the benefits of HAADF-STEM tomography and its use as a novel and rigorous tool for understanding the physical-chemical properties of complex 3D core-shell nanostructures. The reconstructed Au@Ag core-shell architecture was used as an input for discrete dipole approximation (DDA)-based electrodynamics simulations of the optical properties of the nanostructures. The implications of localized surface plasmon spectroscopy as well as Raman-enhanced spectroscopy are analysed.

Ultrafast electron diffraction pattern simulations using GPU technology. Applications to lattice vibrations.

Graphical processing units (GPUs) offer a cost-effective and powerful means to enhance the processing power of computers. Here we show how GPUs can greatly increase the speed of electron diffraction pattern simulations by the implementation of a novel method to generate the phase grating used in multislice calculations. The increase in speed is especially apparent when using large supercell arrays and we illustrate the benefits of fast encoding the transmission function representing the atomic potentials through the simulation of thermal diffuse scattering in silicon brought about by specific vibrational modes.

Finite element simulations of electrostatic dopant potentials in thin semiconductor specimens for electron holography.

Two-dimensional finite element simulations of electrostatic dopant potentials in parallel-sided semiconductor specimens that contain p-n junctions are used to assess the effect of the electrical state of the surface of a thin specimen on projected potentials measured using off-axis electron holography in the transmission electron microscope. For a specimen that is constrained to have an equipotential surface, the simulations show that the step in the projected potential across a p-n junction is always lower than would be predicted from the properties of the bulk device, but is relatively insensitive to the value of the surface state energy, especially for thicker specimens and higher dopant concentrations. The depletion width measured from the projected potential, however, has a complicated dependence on specimen thickness. The results of the simulations are of broader interest for understanding the influence of surfaces and interfaces on electrostatic potentials in nanoscale semiconductor devices.

Low voltage STEM imaging of multi-walled carbon nanotubes.

We present high magnification STEM images of multi-walled carbon nanotubes recorded with a 5 keV electron beam using a Helios Dual Beam microscope and a dedicated multi-segment transmission (STEM) detector. Images of carbon nanotubes recorded with bright-field (BF), annular dark-field (ADF) and high angle annular dark-field (HAADF) signals all show high contrast features, with internal structures 1-2 nm in width clearly revealed in the STEM images. Thicker regions of the nanotubes appear to show an unusual contrast reversal when comparing ADF and HAADF images. An understanding of the image contrast, and its dependence on thickness, is obtained by computing simulations of the ADF and HAADF images using Monte-Carlo software taking into account electron scattering in the nanotube.

Extended ptychography in the transmission electron microscope: possibilities and limitations.

The extended-ptychographical iterative engine (e-PIE) is a recently developed powerful phase retrieval algorithm which can be used to measure the phase transfer function of a specimen and overcome conventional lens resolution limits. The major improvement over PIE is the ability to reconstruct simultaneously both the object and illumination functions, robustness to noise and speed of convergence. The technique has proven to be successful at optical and X-ray wavelengths and we describe here experimental results in transmission electron microscopy supported by corresponding simulations. These simulations show the possibilities - even with strong phase objects - and limitations of ptychography; in particular issues arising from poorly-defined probe positions.