Aberration-corrected scanning transmission electron microscopy (STEM) has become an indispensable tool for characterizing atomic-scale structure in materials and devices. In STEM, a finely focused electron probe is scanned across the specimen and transmitted and/or scattered electrons from a localized material volume are detected by the post specimen detector(s) as a function of raster position. By controlling the detector geometry, we have a lot of flexibility in determining the contrast characteristics of STEM images and the formation mechanisms involved. We have developed an area detector which we refer to as the "Segmented Annular All Field (SAAF)" detector and which is capable of atomic-resolution STEM imaging. This area detector can obtain many simultaneous atomic-resolution STEM images which are sensitive to the spatial distribution of scattered electrons on the detector plane.
By taking the difference between the images from diametrically-opposed detector segments, we can form what are known as differential phase contrast (DPC) images. DPC STEM images approximate the gradient of the object potential (= fields) taken in the direction of the diametrically-opposed detector segments. In this presentation, we will discuss our recent and on-going researches of local electromagnetic field characterization in materials and devices using DPC STEM.

Nanoscale ferroelectrics have emerged as leading contenders for the next generation of non-volatile memory devices, nanoscale actuators, energy conversion and storage devices, and many other applications. Vital to the practical success of these devices is a clear understanding of the fundamental nature of ferroelectric order at reduced dimensions. The physical picture of ferroelectric order at nanoscale length scales has continued to evolve from early work indicating the complete quenching of polar order to more recent studies that have explored the disordering of local polar distortions or the emergence of vortex polarization states. In addition, the physical mechanisms governing this size-dependent behavior remain a subject of some contention, with previous reports implicating depolarization fields, internal strains, and other driving forces. Progress on these questions has been hindered both by the difficulty of synthesizing monocrystalline nanomaterials with well-controlled sizes, morphologies, and surface structures and by a reliance on ensemble measurements, which obscure material behavior through statistical averaging. Here, we examine ferroelectric ordering in individual nanocrystals synthesized via colloidal chemistry using atomic-resolution transmission electron microscopy, off-axis electron holography, and piezoresponse force microscopy. We examine two types of ferroelectric nanomaterials in this study: 5-8 nm nanocrystals of the ferroelectric semiconductor GeTe and 5-15 nm nanocubes of the traditional oxide ferroelectric BaTiO₃. Through-focus exit wave reconstructions of these particles calculated using the Gerchberg-Saxton algorithm allow sub-Angstrom ferroelectric distortions to be imaged on a local scale in individual particles, leading to spatial maps of ferroelectric order at nanoscale dimensions. These maps are complemented with direct imaging of ferroelectric polarization in BaTiO₃ nanocubes using off-axis electron holography. These results indicate a robust, linearly ordered ferroelectric polarization in both GeTe and BaTiO₃ down to dimensions of at least 5 nm. Comparison of highly conducting GeTe with insulating BaTiO₃ using Raman spectroscopy and X-ray atomic pair distribution function analysis highlights the roles of both depolarization effects and size and shape-dependent internal strains in driving ferroelectric size effects. In addition to our current results, prospects for the discovery of monodomain vortex polarization states, the analysis of ferroelectric phase transitions at the single-particle level, and other emerging directions will be discussed.

Atom probe tomography (APT) involves the atom-by-atom field evaporation of a needle-shaped specimen, in order to provide a reconstruction of the identity and position of each atom [1]. The influence of different phases in an APT needle on the electrostatic field surrounding it [2] has been described theoretically but never previously measured experimentally with nm spatial resolution.
Here, we apply off-axis electron holography in the transmission electron microscope (TEM) to study an electrically biased Fe needle that contains Y₂O₃ precipitates. We examine differences between phase images recorded with different voltages applied between the needle and a counter-electrode, in order to remove contributions to the recorded signal from mean inner potential and magnetic contributions to the phase.
Our results are used to provide maps of the electrostatic potential around the needle with a spatial resolution of approximately 1 nm, both in projection and in three dimensions by acquiring tomographic tilt series of electron holograms. Our measurements are compared with predictions for the effect of precipitates in metallic specimens on the equipotential contours surrounding them [3].
[1] Miller, M. K. Atom probe tomography: analysis at the atomic level., Kluwer Academic/Plenum Publishers, New York, 2000.
[2] Vurpillot, F.; Bostel, A.; Blavette, D. Appl. Phys. Lett. 76 (2000) 3127-3129.
[3] Oberdorfer, C.; Schmitz, G. Microsc. Microanal. 17 (2010) 15.

Potentiometry by means of off-axis electron holography is an unique technique, which allows the measurement of the spatially resolved electric potential down to the nanometer scale. Under kinematical diffraction conditions and in the absence of magnetic fields, the electron beam passing through the specimen experiences a phase shift with respect to a vacuum beam. This phase signal is proportional to thickness and potential.
When this technique is applied to model systems like p-n junctions, typically measured phase differences and hence built-in voltages are lower than values expected for the structures. In some semiconductors like Si or GaAs this mismatch is small and can mostly be explained by damages to the surfaces during preparation (structural changes, Fermi level pinning). However, in GaN the observed voltages only are 10-20% of the expected ones.
We conducted electron holographic experiments on GaN p-n junctions in needle-shaped specimen. This geometry restricts possible electric current paths to the needle itself. We observed a logarithmic dependency of the measured potential difference on the intensity of the probing electron beam. A reduction of the illumination dose rate by 3 orders of magnitude gave an increase of roughly 2.5x of the observed potential difference. The measured behavior can be quantitatively explained, when the specimen is modeled like a solar cell under open-circuit conditions.
From Si and GaAs it is known, that expected voltages are most often obtained, when the surface of the specimen is coated with conductive carbon [1,2]. Within the solar-cell model this corresponds to short-circuit conditions, where the voltage drop at the p-n junction equals the drop at the non illuminated junction. The investigated GaN-needles were already covered by conductive carbon contamination during the preparation. After changing the conductivity of the surfaces of the GaN-needles by plasma-cleaning and wet etching, we observed no change in illumination dependency. We conclude from this, that in the contrary to Si the damaged surface layers in GaN bear a large contact resistance and thus a short circuiting is not possible by such simple measures as carbon coating in this material.
Also effects of beam damage were noted over time. A drop of the measured potential by a factor close to two occurred within one hour during illumination with a dose rate of 332 e/nm²s at 300kV. We attribute this effect to the generation of point defects, as no changes in the specimens structure were directly observed.
The experiments clearly show for the first time, that especially the generation of electron-hole pairs and their flux must be considered and controlled in electron holographic experiments in order to obtain quantitative results.
We acknowledge support from the German Research Foundation (DFG) within CRC 787.
[1] D. Cooper et al., J. Appl. Phys. 101, 094508 (2007).
[2] M. R. McCartney et al., Appl. Phys. Lett. 80, 3213 (2002).

Electron wavefields can be brought to a focal point smaller than an atom, enabling small volumes of matter to be probed and characterized. The resultant scattered wavefield contains a wealth of information about the specimen that can be detected selectively to isolate and ‘image&’ specific structural information. This talk will describe different methods to achieve this to solve the atomic structure of nanostructured materials. It will illustrate these with a range of materials applications, such as the determination of the atomic structure and stability of nanoparticle facets [1]; the determination of the local atomic structure of “chessboard&’ nanostructures in lithium-based titanate perovskites [2]; and the measurement of local polarity, dopant concentration and atomic-scale morphology in semiconducting nanowire quantum wells [3,4]. Several methods using focused electron wavefields will be described, including: (i) A new approach for the determination of centrosymmetric structures from the direct observation of structure factor phases from features in convergent beam electron diffraction patterns [5]. (ii) Methods for the quantitative interpretation of the intensity in atomic resolution imaging and diffraction data for the measurement of local atomic and electronic structure. (iii) Pseudo-confocal scanning transmission electron microscopy methods which record the scattered intensity in a plane conjugate to the specimen (as opposed to the diffraction plane) for obtaining depth and chemical information [6,7].
[1] H. Katz-Boon, C. Rossouw, M. Weyland, A. M. Funston, P. Mulvaney, J. Etheridge, Nano Letts (2011) 11, 273-278 and manuscript submitted.
[2] R. Withers, L.N. Bourgeois, A. Snashall, Y. Liu, L. Noren, C. Dwyer, J. Etheridge, Chem. Mater. (2013) 25 190minus;201 and Y. Zhu et al manuscript in preparation.
[3] C.L. Zheng, J. Wong-Leung, Q. Gao, H.Tan, C. Jagadish, J.Etheridge, Nano Letts (2013) 13, 3742-3748
[4] H. Kauko, C.L. Zheng, Y. Zhu, S. Glanvill, C. Dwyer, A.M. Munshi, B.O. Fimland, T.J. van Helvoort, J. Etheridge, App Phys Letts (2013) 103 232111.
[5] P.N.H. Nakashima, A.F. Moodie, J. Etheridge, Proc Nat Acad Sci (2013) 110 14144-14149 (2013)
[6] J. Etheridge, S. Lazar, C. Dwyer, G.A. Botton, Phys Rev Letts (2011) 106, 160802.
[7] C.L. Zheng, Y. Zhu, S. Lazar, J. Etheridge, Phys Rev Letts (2014) 112, 166101.

Modern aberration-corrected scanning transmission electron microscopes have enabled the simultaneous collection of atomically resolved signals relating to coherent scattering (bright field and annular bright field imaging), structural information based on thermal scattering to large angles and known as high-angle annular dark-field (HAADF) imaging, bonding information using electron energy-loss spectroscopy (EELS), and element identification using both EELS and energy dispersive x-ray spectroscopy. Atomic resolution imaging based on secondary electron (SE) signals was also demonstrated in 2009 [1], but the technique is only slowly growing in use. While these SE signals are highly surface sensitive due to the narrow escape depth of electrons with energies <50eV [2,3], atomic resolution SE imaging of surface structures that differ from a simple bulk termination has not been previously demonstrated. The ability to simultaneously record surface sensitive SE and bulk dominated HAADF signals at atomic resolution makes the problem of surface structure registration to the bulk lattice highly tractable, which is a distinct advantage over other scanning probe methods. However, interpretability of the atomic resolution SE signal relies on the development of first-principles simulations of SE images.
We have recently studied the c(6×2) reconstruction on the (100) surface of single crystal SrTiO₃ through simultaneous atomic resolution SE and HAADF imaging and complementary HREM imaging. Preliminary analysis indicates that the registration between the surface structure and bulk of the previously reported structure, primarily refined from surface x-ray diffraction and scanning tunneling microscopy (STM) experiments [4], is incorrect. Interpretation of the experimental SE measurements from first principles is now possible using a recently developed quantum mechanical model to simulate the SE images. This approach takes into account the probability and angular distribution of electrons that are ejected from atoms in the specimen when ionization of both core and semi-core electrons occurs [5]. Our preliminary simulations of a newly proposed structure of the SrTiO3-<100>-c(6×2) reconstruction are in good agreement with the bulk-subtracted experimental SE data, and consistent with previously reported data from STM, Auger spectroscopy, and x-ray diffraction measurements. The structure solved by SE imaging is also stable in density functional theory simulations, and is on the thermodynamic convex hull of known reconstructions on SrTiO₃ <100>.
References
[1] Y Zhu et al., Nat. Mater. 8 (2009) p. 808.
[2] H Seiler, J. Appl. Phys. 54 (1983) p. R1.
[3] A Howie, J. Microsc. 180 (1995) p.192.
[4] CH Lanier, et al., Phys. Rev. B 76 (2007) 045421
[5] HG Brown et al., Phys. Rev. B 87 (2013) 054102.

Accurate crystallographic determination is essential for gaining key insights into material behavior. X-ray and neutron based diffraction methods, for example, provide excellent precision and accuracy, but have comparatively poor spatial sensitivity. In contrast, scanning transmission electron microscopy (STEM) can directly probe real space crystallographic information with atomic resolution. Although STEM has advanced significantly in recent years, accurate distance information has remained elusive due to specimen drift and scan system distortion.
In this talk, we will demonstrate that revolving STEM (RevSTEM) enables direct, accurate, and precise length measurements at the unit-cell scale. By measuring distortion from a set of images acquired with rotated scan coordinates, we reverse the effects of drift from every frame that are then averaged together. The final image has enhanced signal to noise, which enables precise determination of atom column positions, Using silicon as a calibration, we remove residual distortion introduced by the scan system to obtain an accurate and precise distance scale. This calibration is then applied to subsequent imaging of other samples. As a test of accuracy, we will show that lattice parameters of pure Bi₂Te₃ and Bi₂Se₃ obtained from STEM match those from X-ray powder diffraction to within 0.1 %.
Further, we will demonstrate that the approach can even be used to determine alloy composition by applying Vegard&’s Law to the STEM measured lattice parameters using a mixed Bi₂Te₃-Bi₂Se₃ crystal. We will also discuss a comprehensive analysis of the atomic structure employing atomic resolution chemical spectroscopy to determine site preference of the atomic species. We will show that Se resides almost exclusively within the middle layer,Te(2), of the Bi₂Te₃ quintuple - Te(1)-Bi-Te(2)-Bi-Te(1) - which are each held together by van der Waals forces. We will show that within the alloy unit-cell, bond lengths vary between those of the pure samples, as expected. The van der Waals gap, however, is found to vary anomalously by increasing relative to the pure compounds. Finally, we will discuss the general applicability of the approach to other materials systems.

There are mainly two complementary imaging modes in transmission electron microscopy (TEM): Conventional TEM (CTEM) and scanning TEM (STEM). In the CTEM mode the specimen is illuminated with a plane electron wave, and the direct image formed by the objective lens is recorded in the image plane. STEM is based on scanning the specimen surface with a focused electron beam and collecting scattered electrons with an extended disk or ring-shaped detector. In our contribution we introduce ISTEM (imaging STEM), a new TEM imaging mode [Phys. Rev. Lett. 113 (2014) 096101] which combines STEM illumination with CTEM imaging. We use a CCD camera to acquire images formed with the focused electron beam scanning over the specimen. As the acquisition time of the CCD-camera is equal to the area scan time, the images corresponding to all the probe positions are summed up. The wave functions for different electron beam positions occur at different times, so that they cannot interfere and corresponding images are summed up incoherently. Thus, ISTEM exploits an improvement in resolution obtained by switching the spatially coherent illumination to highly incoherent illumination. The gain in resolution can easily be understood for the case of a completely incoherent illumination where the transfer function is given by the autocorrelation of the coherent transfer function, whereby the maximum spatial frequency transferred by the system is increased by a factor of two. In our contribution we will present a simulation study showing that ISTEM generally allows extending the point resolution of CTEM imaging beyond the diffraction limit. We will also reveal by image simulation that this new TEM mode is more robust against chromatic aberration, which allows overcoming the conventional information limit of a microscope. These calculations are confirmed by experimental data for GaN along the [1-100] and [11-20] directions taken on our TITAN 80/300 microscope with a conventional information limit of 80 pm, where we resolved Ga and N columns at a distance of 63 pm. Thus, ISTEM combines advantages of STEM imaging such as improved point resolution with advantages of the CTEM imaging mode while avoiding disadvantages of STEM. In STEM, the precision for determining atom column positions is limited by scan noise which is caused by errors in positioning the electron probe, and the resolution is influenced by the finite source size effect. In contrast, ISTEM images do neither suffer from scan noise nor is the image resolution influenced by the finite source size. Furthermore, we will show by theoretical considerations that ISTEM is independent of lens aberrations of the probe forming system, but only depends on the radius of the probe forming aperture. Due to the principle of reciprocity, ISTEM can be made equivalent to annular bright field STEM using a ring-shaped condenser aperture, promising ultra-high resolution imaging of light elements by avoiding scan noise and source size effect.

Analysis of nanoparticles by TEM is a very useful characterization method. However, by nature of TEM instrument design placing sample grids under high vacuum, the number of samples that can be looked at in a given time period is very limited. Indeed, for each sample placed on a TEM grid, the TEM has to undergo a complete cycle of losing and re-establishing vacuum conditions. Having the capability to insert multiple samples at once inside the TEM and then acquiring all images without breaking vacuum for each one, would result in a tremendous increase in the number of samples that can be analyzed per hour.
For over 10 years, SCIENION&’s picoliter liquid dispensing technology has been used to deposit arrays of different sample solutions onto a wide variety of substrates. Applying this well-proven technology to dispense multiple liquid samples onto a single TEM grid is an obvious extension.
In this presentation, we illustrate the process by which one can deposit up to 100 different samples onto a single 3 mm TEM grid&’s 1 x 1 mm window. In summary, analytical suspensions of nanoparticles are loaded in the instrument using a 96-well plate. For each sample, the instrument aspirates a small aliquot of a few microliters, and dispenses 50-100 picoliters onto a precise, indexed location inside the TEM window. The process is repeated until all samples have been deposited inside the grid in different locations. The most suitable TEM windows have to be selected according to the nature of the nanoparticles to be investigated. The hydrophilicity of the window material is important to allow high quality spot formation as well as homogeneous distribution of the nanoparticles inside each spot.

We use angle- and polarization-resolved cathodoluminescence (CL) imaging spectroscopy to study coherent and incoherent radiation processes in single-crystalline silicon wafers and silicon photonic crystals. Samples are excited using a 30 keV electron beam, and the emitted radiation is collected using an aluminium half-parabolic mirror that is placed between the electron column and the sample. A micromanipulation stage allows accurate alignment of the mirror focus with the electron beam spot. Angle-resolved data are collected using a two-dimensional CCD imaging camera placed in the Fourier plane of the imaging optics, in combination with band-pass filters. Polarization is analysed by rotating a quarter-wave plate in combination with a polarizer in the emitted light beam.
We first analyse the CL emission from a single-crystal Si(100) wafer. The angular CL distribution is composed of a dipolar distribution due to transition radiation and a Lambertian emission profile resulting from the radiative recombination of electron-hole pairs inside the Si wafer. We find that incoherent radiative emission accounts for 25% of the collected light at a wavelength of 400 nm, increasing to 85% at 900 nm.
Next, we determine the full polarization state of the emitted CL in the 600-900 nm spectral range by performing measurements of the angular distribution at six different combined orientations of the quarter-wave plate and polarizer. We separate the polarized and unpolarized components of the emission from bulk Si and reconstruct the polarization vector at each emission angle. The polarizing effect of the Si/air interface on the CL emission generated inside the wafer is clearly observed.
Finally, we introduce the use of a scanning pinhole inside the optical beam path which enables measurements that combine both high angular and spectral resolution, allowing for analysis of strongly dispersive geometries. We analyze the spatially-resolved angular and spectral CL emission profiles from photonic crystal waveguides and cavities that we make using electron-beam lithography using a silicon-on-insulator wafer. From the data, we construct the photonic crystal bandstructure and determine the modal field distributions in the photonic cavities in the 1300-1700 nm spectral range.

Porous silicon has presented a lot of interest since its discovery in 1956, thanks to a wide range of applications in electronics, optoelectronics, photonics, chemical sensors, biosensors, etc. Many different geometries and morphologies of porous silicon can be obtained, depending on the fabrication parameters. Erbium doped porous silicon is particularly interesting for its luminescent properties at room temperature and could be a relevant material for efficient and cost-effective silicon-based optoelectronic devices. To characterize the geometries and size of pores of a few nanometers as well as the erbium localization, Electron Tomography (ET) is a key technique.
ET is a non-invasive technique to perform 3D imaging of objects of a few hundreds nanometers with a spatial resolution of about one nanometer. A TEM (Transmission Electron Microscope) is used in scanning HAADF (High Angle Annular Dark Field) mode to acquire 2D images of the object under different tilt angles. After careful alignment of the 2D image stack, one can reconstruct the 3D object using analytical or algebraic algorithms such as SIRT (Simultaneously Iterative Reconstruction Technique). Simulations show that SIRT gives almost perfect reconstructions with 180 projections in a 180° angular range. In practice, the resolution of the 3D reconstruction is far from identical to that of the 2D images. Differences between simulations and practice can be explained mostly by the misalignment of the image stack and the presence of noise. Here we propose an automatic alignment and a parameter-less denoising of the projections that help to enhance the resolution and the quality of the 3D reconstruction. This technique has been applied to erbium-doped porous silicon samples. It shows high reconstruction quality and high reliability to perform an efficient segmentation and quantitative analysis of the 3D structures and geometry of the pores. Compared to state of the art complex reconstruction algorithms such as Total Variation Minimization (TVM) based algorithms and Discrete Algebraic Reconstruction Technique (DART), the complete procedure used has the great advantage to be totally parameter-less, i.e. user-independent, and much faster than TVM or DART.
The principle of the technique and its comparison to classical alignment and reconstruction methods will be discussed as well as the quantitative analysis of the 3D reconstruction of erbium-doped porous silicon sample. The 3D reconstruction reveals clusters of a few nanometers, with localization depending on the curvature of the nearest pores surface. The surface curvature dependence is quantitatively demonstrated with a statistical determination of the surface curvature next to the clusters.

LaNiO₃ (LNO) is a perovskite of great importance in complex oxide electronics. Its low resistivity at room temperature and high chemical stability make it an ideal electrode candidate for many applications in complex oxide-based devices. Strain, oxygen vacancies and their mutual interplay are key aspects to understand the transport properties of these oxides, since they might affect the Ni-O hybridization. In this study, we perform a thorough analysis of these aspects with high spatial resolution TEM.
We have studied LNO thin films of different thicknesses (14 nm and 35 nm) grown on several substrates that allow studying a wide range of compressive (LAO and YAO) and tensile (LSAT and STO) strain states. Aberration corrected HRTEM, HAADF-STEM, atomic resolution EELS mapping, electron diffraction and image simulation studies have been carried out. Strain states in the films have been studied by Geometric Phase Analysis (GPA) of the high-resolution images.
The presence of an oxygen deficient monoclinic phase has been detected in the LNO strained films. This perovskite-related superstructure occurs when oxygen vacancies order along a given crystallographic direction, consisting in the loss of the two apical oxygen atoms in the Ni octahedra for alternating columns of octahedra and square planes along the c axis. Different crystallographic orientations were prepared for TEM observation. Electron diffraction patterns, contrast modulation in HRTEM images and Z contrast in HAADF images are consistent with this vacancy ordering. Image simulations (both HRTEM and HAADF) support these findings. We report on the effect of the strain state of the film and the film thickness on the occurrence and orientation of the monoclinic superstructure.

Nanodomains in ferroelectrics are attractive due to their applications on the ultra-small electric, optical, actuator devices and nonvolatile memory. Recently, the in-situ nanodomain investigations by the Scanning Probe Microscope (SPM) and the sub-atomic resolution Transmission Electron Microscopy (TEM) achieved great successes, revealing numerous novel ferroelectric nanodomain structures such as the self-similar nested bundles, the nanovortex array and the stabilized charged domains. In this work we improved an in-situ system in TEM to study the nanodomain structures in the free-standing nanopillars. With Cs-corrected TEM, different types of nanodomain structures were studied in three dimensions. And the unique properties of the nanodomain structures are discussed as well.

General transmission electron microscopy (TEM) is not suitable for observing liquid specimen due to high vacuum environment in a TEM. To overcome this problem, previous studies using several type of liquid cell with Si₃N₄ viewing window. However, thick thickness Si₃N₄ layers and liquid surrounding the sample are main challenges for atomic-resolution imaging. A novel graphene liquid cell that encapsulating a liquid film between two layers of graphene is expected that could enable atomic-resolution observation of liquid specimens including chemical reaction or nucleation and growth due to outstanding characteristics of graphene. Graphene, a one-atom-thick planar sheet of carbon atoms, can provide a high contrast images in a TEM and sealing any type of material including liquid and gas phase materials because of high flexibility. In order to investigate nucleation and growth behavior of core-shell nanoparticle, we employ the graphene liquid cell. In bimetallic nanoparticles, composed of two different metal elements, the particle size and the structure of bimetallic nanoparticles can affect their catalytic properties including activity and selectivity of catalysts. So it is important to understand growth mechanism of bimetallic nanoparticles. We observed the growth of Cu shells on Au nanoparticles in the TEM using the graphene liquid cell. Graphene liquid cell is prepared by encapsulating Au nanoparticles as templates and Cu growth solution between two graphene layers. Cu nanoparticle growth in the TEM employed the reduction of Cu²⁺ ion via electron beam illumination. The Au seed nanoparticles and the stock Cu growth solution were observed successfully trapped between the two graphene layers. The growth of Cu shells on Au nanoparticle was initiated by electron beam irradiation and size and shape of the synthesized Au-Cu core-shell nanoparticles are determined by Au seed nanoparticles. We have observed the formation process of Cu shells on Au nanoparticle with atomic-resolution using graphene liquid cell. We speculate that this method can be applied to other bimetallic nanoparticles in order to design the desired bimetallic nanoparticle.

Hydroxyapatite [Ca₁₀(PO₄)₆(OH)₂, HAp] is one of the primary constituents in hard biological tissues and in its synthetic form has significant potential uses in the medical industry, mainly for its applications in bone regeneration. The apatite lattice is very tolerant of substitutions, and many kinds of cations can be introduced into the lattice. In particular, doping with rare-earth ions is of great interest because the material can be used as a fluorescent probe in the medical industry. Cathodoluminescense (CL) is a very efficient technique to study the luminescent properties of several materials locally. When adapted in an SEM, it permits one to obtain luminescent properties at great magnifications, sometimes overlooked in bulk measurements. In this work, we present an in-depth analysis of the CL response of europium-doped hydroxyapatite powders in order to demonstrate the dopant distribution in the host. Samples were synthesized using combustion synthesis at different pH values and the same amount of Eu precursor. CL spectra show UV-blue emissions between 350 - 440 nm attributed to transition 4f⁶5d¹-4f⁷(⁸S_7/2) of the Eu²⁺, and orange-red bands in the range 575-700 nm generated by transitions ⁵D₀ → ⁷F_J of the Eu³⁺. Monochromatic CL images acquired at 425 and 616 nm show inhomogeneity in the spatial distribution of Eu²⁺ and Eu³⁺ in the samples, revealing bright regions at the edge of the HAp:Eu microstructures for the Eu³⁺ emission. XPS measurements confirm the presence of Eu⁺³ and Eu⁺² ions in the samples, with a concentration ratio Eu⁺²/Eu⁺³ of about 0.22 that corresponds to a preferential incorporation of Eu⁺³ on the surface.

The correlation between a material&’s luminescence properties and its nanoscale morphology, microstructure and local chemistry offers great benefit in the understanding of many technologically important materials and devices. This has encouraged a growing interest in performing cathodoluminescence (CL) microscopy at high spatial resolution in a STEM microscope [1]. Here, we use combined electron energy loss spectroscopy (EELS) and CL analysis of the GaN / InGaN quantum well (MQW) from a light emitting diode (LED) to investigate the role of In clustering in luminescence efficiency; sub-nanometer compositional information is correlated with the luminescence from individual quantum wells with the MQW.
III-nitride semiconductors are technologically important materials with GaN / InGaN MQWs being the source of light emission in current generation blue and white LEDs. However, efficient white LEDs based entirely of III-nitride semiconductors remain elusive due to poor efficiency at green emission wavelengths, the so-called ‘green-droop&’. The reason for low efficiencies at high In content is not well understood; auger electrons fluctuations in quantum well and In composition are two of the many proposed mechanisms. Thus, understanding how the structure, composition and luminescence (intensity and emission wavelength) of MQW structures are correlated is ultimately very important for improving device performance.
For the results of this paper, compositional information was obtained from sub-nanometer scale EELS analysis using means of MLLS fitting to extract the composition and the local luminescence was measured simulatnously using CL. The CL light was acquired using miniature elliptical mirrors (solid angle of about 7.3 sr) integrated into the tip of a conventional cryogenic TEM holder. Light is coupled out of the holder through two optical fibres to an optical spectrometer fitted with a PMT and CCD detectors. This combined EELS / CL system offers the advantage of the best in spectral resolution (up to 4 meV), spatial resolution analysis and sensitivity to microstructural changes. Simultaneous EELS / CL data was collected with the sample at -171C minimizing the influence of the electron beam on the sample and increasing the spatial resolution of the Cl data as result of the enhanced rate of radiative recombination within the QWs. The analysis was carried out across the GaN / InGaN MQWs and superlattice layers where each layer is just a couple of nanometers wide. Variations of the luminescence quantum efficiency by more than an order of magnitude was observed between regions separated by only a few tens of nanometers and free extended defects; we analyse how the composition measured by EELS affects the luminescence.
[1] Zagonel L. F., Mazzucco S., Tence&’ M., March K., Bernard R., Laslier B., Jacopin G., Tchernycheva M., Rigutti L., Julien F. H., Songmuang R. and Kociak M., Nanoletters 11, 2011, 568

Long persistence SAEDB phosphors consist of luminescent rare earth atoms, i.e. Dy²⁺ and Eu³⁺, embedded in a B-incorporated Sr₄Al₁₄O₂₅ matrix. We have previously demonstrated that the addition of boron into the strontium aluminate system dramatically extends the persistence duration from several minutes to more than 8 hours [1]. However, the exact mechanism by which B extends afterglow is still unclear, and the potential for such materials to be exploited in energy-efficiency applications motivated our study.
As a key step in a broader effort to understand the mechanisms of persistence in phosphors, we combined electron-based spectroscopy techniques by means of cathodoluminescence (CL) and electron energy-loss spectroscopy (EELS) in a dedicated scanning transmission electron microscope (STEM) [2]. This unique approach enables us to probe locally their optical and chemical heterogeneities down to the nanoscale, and therefore to correlate spatially the cation homogeneity and luminescence uniformity in SAEDB phases. Because the 4f⁶5d¹ agrave; 4f⁷ transition of Eu²⁺ engenders luminescence in the visible light spectrum, we probed locally the spatial distribution of this luminescence, after excitation by energy absorbed from the probe electrons. We could then map the B chemical distribution in the oxide matrix, by probing the B-K core-loss edge by EEL spectroscopy.
The particles studied were synthesized by a modified Pechini process, yielding powders of Sr₄Al₁₄O₂₅ doped with 1 mol% Eu²⁺, 1 mol% Dy³⁺, and 7 mol% B (Sr₄A₇EDB). The crystal structure, as well as the cationic stoichiometry, was first characterized by means of XRD, high-resolution TEM imaging, and ICP analysis. While a 2-D CL map in the spectral range of 413-436 nm appeared relatively homogeneous, a CL-filtered-map in the 277-300 nm spectral range was not uniform. A CL peak centered at 288 nm with a FWHM of ~65 nm was precisely localized on one side of the probed particle. This strong cathodoluminescence feature is consistent with the strong absorption of Sr₄A₇EDB in the region below 350 nm, with absorption peaks at 250nm, 288 nm, and 350 nm. In this spectral region, the control specimen, which did not contain boron, absorbed with much lower intensity, and no sharp features above the weaker background absorption. We thus report a correlation between luminescence and the spatial distribution of boron in Eu²⁺ and Dy³⁺ co-doped boron-incorporated strontium aluminate (SAEDB) phosphors.
[1] M.G. Eskin, Sabanci University, M.Sc. Thesis (2011).
[2] L. Zagonel et al., Nanoletters, 11 (2011) 568
Acknowledgement: The research leading to these results has received funding from the European Union Seventh Framework Programme under Grant Agreement 312483 - ESTEEM2 (Integrated Infrastructure Initiative - I3)

Thermoelectric materials have the capability to directly generate power using the Seebeck effect or refrigerate using the Peltier effect, without the need of additional moving parts. The figure of merit, ZT, defines the efficiency of thermoelectric materials: ZT=(S²σ/k)T, where S is the Seebeck coefficient, σ is the electrical conductivity, k is the thermal conductivity and T is the absolute temperature at which the properties are measured [1][2]. For the thermoelectric device to be competitive, an average value of the ZT larger than 1 is needed within the application range - usually from room temperature to 900°C. Achieving this goal requires a detailed understanding of the atomic and electronic structure of the materials involved.
The Half-Heusler alloys are a promising group of materials for thermoelectric applications. They are ternary semiconductor or metallic materials, having the stoichiometry XYZ (1:1:1), with X and Y transition metals (e.g., X=Ti, Zr, Hf and Y=Ni, Co) and Z a metal or metalloid (Sn, Sb). The electronic structure, and charge carrier concentration and scattering may be manipulated by substitution on three crystallographic sites in order to enhance the thermoelectric properties.
In this contribution, we study a series of Half-Heusler compounds with the starting composition TiNiSn, but where we allow the following substitutions: Ti with Zr, Hf; Ni with Co; and Sn with Sb. The crystal structure and the chemical composition are investigated with electron diffraction and Energy Dispersive X-ray Spectroscopy analyses, while the transition metal oxidation state and d-state occupancy are tracked by Electron Energy Loss Spectroscopy [3][4].
[1] G. S. Nolas, J. Sharp and H. Goldsmid, Thermoelectrics: Basic Principles and New Materials Developments, Springer, New York, 2001;
[2] D. Rowe, Thermoelectrics Handbook: Macro to Nano, CRC Press, Boca Raton, 2006;
[3] D. H. Pearson, C. C. Ahn and B. Fultz, PRB Vol 47 (1993);
[4] Y. Shao, C. Maunders, D. Rossouw, T. Kolodiazhnyi, G. A. Botton, Ultramicroscopy Vol 110 (2010).

In the search for new solar cells with higher efficiencies and lower production costs, thin films have shown great potential and could eventually replace silicon. As the films are on the nanometer scale, and the interface is strongly dependent on the processing methods, the electronic properties across the interface become important.
In this project, we study interfaces in different types of ZnO/Cu₂O heterostructures. Films of Cu₂O are deposited onto ZnO substrates by sputtering, and TEM samples are prepared by cutting, grinding and polishing. The structure and composition of the heterostructure interfaces are studied using a probe-corrected and monochromated FEI Titan 60-300 STEM. Furthermore, using high energy resolution monochromated EELS we investigate the dielectric properties and band gap variations across the interfaces.
The critical voltage of Cerenkov radiation for ZnO is 79 kV, and to avoid this Cerenkov limit the microscope is operated at a high tension of 60 kV. The goal is to measure the band gap close to the interface, an also consider the inelastic delocalization of the electrons [1] [2].
References:
[1] Egerton, R. F., Ultramicroscopy 107 (2007) 575-586
[2] Stöger-Pollach, M., Ultramicroscopy 145 (2014) 98-104

CeO₂ (ceria) doped with aliovalent cations such as Gd³⁺ and Sm³⁺ is a common solid state O^2- conductor in solid oxide fuel cell (SOFC) electrolyte research due to its high ionic conductivity at low and intermediate temperatures (300 °C - 550 °C) [1]. However, at such temperatures the resistance contribution from grain boundaries in polycrystalline electrolytes also becomes considerable due to an intrinsic electrostatic space charge potential barrier emanating from grain boundary cores. This argument generally assumes grain boundaries have similar properties and ignores the potential differences in grain boundary structure [2]. And because atomic structure is thought to be a principal component of the origin of the grain boundary space charge potential [3], the authors hypothesize that because grain boundary structures vary, their electrostatic potentials vary as well, manifesting as variation in other grain boundary properties including composition and electrical conductivity. In practice, there may be many different types of grain boundaries in a polycrystalline ceramic and it is important to provide statistical context to the atomic resolution information derived from TEM. In this contribution we employ orientation imaging performed with a nanometer-sized probe in the scanning transmission electron microscope (STEM) to determine the character of hundreds of grain boundaries. This data provides a rapid statistical distribution of the grain boundary character which can then be used to guide the atomic resolution structural and compositional analysis by imaging and electron energy-loss spectroscopy. All measurements were performed on a JEOL ARM 200F aberration-corrected TEM/STEM equipped with an ASTAR precession nanodiffraction system. Data are presented on the grain boundary distribution and statistical prevalence, correlated with point defect segregation at grain boundaries in a polycrystalline Gd/Pr doubly-doped ceria specimen. Preliminary data shows that the dopant segregation varies with grain boundary angle.
1. Fergus, J., et al., eds. Solid oxide fuel cells: materials properties and performance. CRC Press, 2008.
2. N. Shibata , F. Oba , T. Yamamoto, Y. Ikuhara. Structure, energy and solute segregation behaviour of [110] symmetric tilt grain boundaries in yttria-stabilized cubic zirconia, Philosophical Magazine84:23 2381 (2004).
3. Tschöpe, A. Interface Defect Chemistry and Effective Conductivity in Polycrystalline Cerium Oxide. Journal of Electroceramics14 5-23 (2005).
4. This material is based upon work supported by the National Science Foundation Graduate Research Fellowship Program under Grant No. DGE-1311230, and NSF DMR-1308085. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Science Foundation.

Matter is a three-dimensional (3D) agglomeration of atoms. The properties of materials are determined by the positions of the atoms, their chemical nature and the bonding between them. Therefore, reaching atomic resolution in 3D has been the ultimate goal in the field of electron tomography for many years.
In addition to the use of discrete tomography [1-3], one of the possibilities to perform electron tomography with atomic resolution is by applying reconstruction algorithms based on compressive sensing. The methodology was applied to high angle annular dark field scanning transmission electron microscopy (HAADF-STEM) images acquired from Au nanorods. In the final 3D reconstruction, the Au crystal lattice was reproduced without using prior knowledge on the atomic structure [4]. Going further than determining the positions of atoms is the aim to determine the type of individual atoms in hetero-nanoparticles. Again using a combination of HAADF-STEM and compressive sensing, we were able to distinguish individual Ag from Au atoms in core-shell Au@Ag nanorods, even at the metal-metal interface [5].
Also energy dispersive X-ray (EDX) mapping can be combined with electron tomography and the use of the Super-X system, consisting of 4 EDX detectors symmetrically arranged with respect to the sample is hereby very beneficial. EDX tomography was applied to obtain qualitative information concerning a galvanic replacement reaction in AuAg nanoparticles and to investigate the 3D composition of Fe and Co in nanodumbbells [6,7]. Quantification 3D EDX results can be considered as the next crucial step.
Finally, by combining electron tomography with electron energy-loss spectroscopy at high energy resolution, we were able to determine the valency of the Ce ions in CeO_2-x in 3D. These unique experiments revealed a clear facet-dependent reduction shell at the surface of ceria nanoparticles, invisible to modern high resolution TEM structural imaging techniques [5].
We kindly acknowledge Prof. L. M. Liz-Marzán and Prof. K. Soulantica for the provision of the samples.
[1] S. Van Aert, K. J. Batenburg, M. D. Rossell, R. Erni, G. Van Tendeloo, Nature 470 (2011) 374
[2] S. Bals, M. Casavola, M. A. van Huis, S. Van Aert, K. J. Batenburg, G. Van Tendeloo, D. Vanmaekelbergh, Nano Lett. 11 (2012) 3420
[3] B. Goris, S. Bals, W. Van den Broek, E. Carbo-Argibay, S. Gomez-Grana, L. M. Liz-Marzan, G. Van Tendeloo, Nature Mater. 11 (2012) 930
[4] B. Goris, A. De Backer, S. Van Aert, S. Go#769;mez-Gran#771;a, L. M. Liz-Marzán, G. Van Tendeloo, S. Bals, Nano Lett. 13 (2013) 4236
[5] N.Liakakos, Ch. Gatel, Th. Blon, Th. Altantzis, S. Lentijo-Mozo, C. Garcia-Marcelot, L.-M. Lacroix, M. Respaud, S. Bals, G. Van Tendeloo, K. Soulantica Nano Lett.14 (2014) 2747
[6] B. Goris, L. Polavarapu, S. Bals, G. Van Tendeloo, L. M. Liz-Marzán Nano Lett. 14 (2014) 3220

The ever increasing complexity of the 3-D architectures of modern nanodevices drives the demands to image nanoscale features in 3-D. ADF-STEM tilt-series tomography can reconstruct materials with nanometer and even atomic-scale resolution; however it often requires days to align and reconstruct the tomograms. With the development of aberration correctors where the depth of focus has been reduced to the sub-10nm region, 3-D imaging by depth sectioning becomes a possibility. It potentially allows for a reconstruction of nanomaterials by simply recording a through-focal series. Unfortunately, it has been demonstrated that both ADF-STEM and bright-field scanning confocal electron microscopy (BF-SCEM) have cones of missing information that produce excessive elongation artifacts in the resulting reconstructions. Here, we explore the inelastic counterpart of BF-SCEM—scanning confocal electron energy loss microscopy (SCEELM). This technique is free of the missing-information cone and resulting elongation artifacts. In a microscope without post-specimen chromatic aberration (C_c) correction, this method has a dose efficiency comparable to that of ADF-STEM depth sectioning if valence-loss signals are used. However, the efficiency can be increased by a factor of 10-100 with post-specimen C_c correction by parallel acquisition of SCEELM signals in spectroscopy mode. It can potentially enable a rapid and reliable 3-D reconstruction of materials with sub-10 nanometer depth resolution in C_c-corrected confocal TEMs^1,2,3,4.
¹ HLX is supported by the Center for Functional Nanomaterials, Brookhaven National Laboratory, which is supported by the U.S. Department of Energy, Office of Basic Energy Sciences, under Contract No. DE-AC02-98CH10886.
² Scanning Confocal Electron Energy-Loss Microscopy Using Valence-Loss Signals, H. L. Xin*et al,Microscopy and Microanalysis, 19, 1036-49 (2013)
³Three-dimensional imaging in aberration-corrected electron microscopes, H. L. Xin* and D. A. Muller, Microsc. and Microanal. 16, 445 (2010)
⁴Aberration-corrected ADF-STEM depth sectioning and prospects for reliable 3D imaging in S/TEM, H. L. Xin* and D. A. Muller, J. Electron Microsc. 58, 157 (2009)

Interfacial atoms located between metal nanoparticles and supports are proposed active sites, because of their distinct physical and chemical properties [1,2]. However, the atomistic details are difficult to resolve in the interface; the lack of knowledge has been a major obstacle toward unraveling their roles in chemical transformations. Here we report the detection of interfacial Au atoms on the rutile (TiO₂) (110) surfaces thanks to the improved spatial resolution and depth of focus brought by aberration corrected scanning transmission electron microscopy (STEM).
Au on TiO₂ is selected because it shows remarkable catalytic activity as the sizes of Au particles reduce to ~3 nm or below, for the oxidation of CO. The Au catalysts are typically prepared by Au precipitation on titania support, followed by calcination in air or reduction under H₂ at elevated temperatures. Extensive study has been done concerning the mechanism of CO oxidation catalyzed by gold and the role of interfacial gold atoms. However, direct observation of interfacial Au atoms has not been reported before. A major obstacle is the TiO₂ support surface, which is often highly complex, undetermined and varies at the nanometer scale. With atomic resolution images recorded at different focuses along TiO₂ [001], we have reconstructed the 3D intensity profiles of interfacial atoms. The results lend to direct support to the presence of interfacial Au atoms, embedded in a single interfacial layer.
The experiment started with forming epitaxial Au nanocrystals (NCs) on rutile (110) by e-beam evaporation deposition followed by annealing in air. The sizes of the Au NCs ranged from 3.5 to 12 nm in width depending upon the annealing conditions. The samples are observed by aberration corrected (AC) scanning transmission electron microscopy (STEM) using the JEOL 2200FS installed at the Center for Microscopy and Microanalysis, Frederick Seitz Materials Research Laboratory, at 200kV. The microscope is capable of resolving atoms separated by 1 Å.
With the interface imaged along TiO₂ [110] and [001], results show that a single interfacial layer forms between Au NCs and the TiO₂ (110) surface, with Au atoms embedded inside the layer. The interfacial Au atoms are located in 3D by depth sectioning, and they are separated from Ti-O according to their intensities. A detailed description of interactions between Au NCs and the TiO₂ support is made possible from the results.
References
1. Akita, T., M. Kohyama, and M. Haruta, Electron Microscopy Study of Gold Nanoparticles Deposited on Transition Metal Oxides. Accounts of Chemical Research, 2013. 46(8): p. 1773-1782.
2. Widmann, D. and R.J. Behm, Active Oxygen on a Au/TiO2 Catalyst: Formation, Stability, and CO Oxidation Activity. Angewandte Chemie International Edition, 2011. 50(43): p. 10241-10245.

Solving the crystal structure of a material using X-ray based techniques is not often an easy task, especially when dealing with nanocrystals. Recent advances in electron microscopy have eased the way to explore the three-dimensional structure of a material using novel TEM-based methods. Three-dimensional electron diffraction tomography (3D EDT) has been recently developed to simplify the electron diffraction data collection process. The program-controlled TEM-based technique is based on fast 3D reciprocal space fine scanning allowing subsequent reconstruction of high resolution reciprocal space volume¹. The 3D data set can be further used for the unit cell determination, quantitative intensities extraction and ab initio structure solution. The main advantage is that the 3D electron diffraction data set can be acquired on any randomly oriented crystal unlike the time-consuming conventional electron diffraction acquisition method.
We demonstrate how the 3D EDT technique can be used for fast automated acquisition and processing of a 3D electron diffraction data set in order to solve the crystal structure of any individual sub-micrometer crystals² or nanocrystals³ on the atomic scale. We also show how the use of 3D EDT can be extended, using small-angle mode, to probe mesoscopic structures such as self-assembled nanoparticle arrays⁴. This novel technique opens new possibilities to investigate not only challenging inorganic materials but perhaps also a wide range of organic structures.
1. M. Gemmi and P. Oleynikov, Z. Krist., 2013, 228, 51-58.
2. D. Xu, Y. Ma, Z. Jing, L. Han, B. Singh, J. Feng, X. Shen, F. Cao, P. Oleynikov, H. Sun, O. Terasaki, and S. Che, Nat. Commun., 2014, 5, 4262.
3. A. Mayence, J. R. G. Navarro, Y. Ma, O. Terasaki, L. Bergström, and P. Oleynikov, Inorg. Chem., 2014, 53, 5067-72.
4. A. Mayence, D. Wang, G. Salazar-Alvarez, P. Oleynikov, and L. Bergström, Nanoscale, 2014.

Scanning transmission electron microscopy (STEM) high angle annular dark field (HAADF) imaging has become popular for materials electron tomography for a number of reasons: Firstly image contrast is easily interpreted and the intensity varies monotonically with thickness (at least for reasonable mass-thickness levels). Secondly, HAADF signal is sensitive to atomic number and provides a measure of the local composition. Thirdly, and of importance to this paper, it can be combined easily with analytical techniques such as energy-dispersive xray spectroscopy (EDS), electron energy loss spectroscopy (EELS), and electron diffraction.
When EDS or EELS is combined with tomography, after suitable reconstruction, spectral information is available at every 3D real space location (voxel). Many modern microscopes are now fitted with large solid angle x-ray detectors and when symmetrically disposed around the optic axis offer an efficient means to record EDS spectrum-images as a function of tilt to build up a 4D data set. We have investigated the best use of such detectors, taking into account shadowing and absorption and we will present examples of 3D compositional maps from a number of materials systems. In a similar way, EELS and tomography can be combined to enable 3D compositional maps to be reconstructed (using core loss data) and 3D optical / dielectric maps (using low loss spectra). We will illustrate both but concentrate on the latter where, for example, a 3D imaging technique has been developed to visualise the 3D excitation of localised surface plasmons.
It is also possible to combine tomography with diffraction to enable 3D orientation data to be acquired at each 3D real space position. Scanning precession electron tomography (SPET) uses a scanned beam to record a precession electron diffraction (PED) pattern at every real space pixel, forming a 4D ‘diffraction-image&’. By acquiring a tilt series, both real and reciprocal spaces can be reconstructed for every phase in the volume of interest.
In each of these techniques, the data acquired is likely to be a mixed signal (with overlapping spectral modes in one, overlapping diffraction patterns in the other). In each case that mixed signal needs to be ‘unmixed&’ in an objective fashion. We now use on a routine basis multivariate statistical analyses and in particular non-negative matrix factorisation (NMF) which decomposes each raw signal into a linear sum of positive signal components (e.g. spectral modes or distinct diffraction patterns) weighted by a positive loading (spatial distribution map). Without data reduction techniques of this kind, processing multi-dimensional data sets become intractable. In addition, we will discuss compressed sensing techniques as a framework for incorporating prior knowledge about the object under investigation. This restricts the number of possible solutions in reconstruction space and minimise artefacts in the final tomogram.

Off-axis electron holography is a powerful technique for recording the phase shift of the high-energy electron wave that passes through an electron-transparent specimen in the transmission electron microscope. The phase shift is, in turn, sensitive to the electrostatic potential and magnetic induction in the specimen. Recent developments in the technique have included the use of advanced specimen holders with multiple electrical contacts to study nanoscale working devices, the application of electron holographic tomography to record three-dimensional potentials with nm spatial resolution and the use of ultra-stable transmission electron microscopes to achieve sub-2π/1000-radian phase sensitivity. We are currently working on the application of off-axis electron holography to the measurement of electrostatic potentials and electric fields around electrically-biased atom probe tomography needles. Each experiment typically involves applying a voltage between a needle and a counter-electrode. The recorded phase shift can be analyzed either by fitting the phase distribution to a simulation based on two lines of opposite charge density or by using a model-independent approach that involves contour integration of the phase gradient to determine the charge enclosed within the integration contour. Both approaches often require evaluation of the difference between phase images acquired for two applied voltages, in order to subtract the mean inner potential (and sometimes also the magnetic) contribution to the phase. On the assumption of cylindrical symmetry, the three-dimensional potential and field around such a needle can be determined from the results. We are also working on a model-based approach that can be used to reconstruct the three-dimensional magnetization distribution inside a specimen from a series of phase images recorded using electron holography. In order to develop the technique, we are generating simulated magnetic induction maps by projecting three-dimensional magnetization distributions onto two-dimensional Cartesian grids. We use known analytical solutions for the phase shifts of simple geometrical objects to pre-compute contributions to the phase from individual parts of the grids, in order to simulate phase images of arbitrary three-dimensional objects from any projection direction, with numerical discretization performed in real space in order to avoid artifacts generated by discretization in Fourier space without a significant increase in computating time. This forward simulation approach is used in an iterative model-based algorithm to solve the inverse problem of reconstructing the three-dimensional magnetization distribution in the specimen from a tomographic tilt series of phase images. The use of such a model-based approach avoids many of the artifacts that result from using classical tomographic techniques based on backprojection, as well as allowing additional physical constraints to be incorporated.

Contemporary aberration-corrected scanning transmission electron microscopy (STEM) enables to probe material properties at a resolution of 80pm and below. However, mapping of atomic electric fields in nanoelectronics remains a challenge. Progress in this field relies on the differential phase contrast (DPC) technique which uses segmented detectors to detect the field-induced angular deflection of the STEM beam via a shift of the central part of the diffraction pattern (the ronchigram), causing a characteristic asymmetry in opposite segments (Nature Phys. 8, 611 (2012); Ultramicroscopy 2, 251 (1977)). The established interpretation of DPC relies on the assumptions that the ronchigram is homogeneously filled and shifted as a whole.
First we focus on the validity of these assumptions. In simulation and experiment we show that ronchigrams neither exhibit homogeneous intensity even for thinnest specimens (1-5nm), nor are they shifted as a whole. Contrary, the effect of electric fields is a complex redistribution of intensity inside the ronchigram which needs to be captured by pixelated detectors such as CCD cameras.
Second, we present a simple but stringent quantum mechanical interpretation: In recording the 2D intensity distribution in the diffraction pattern, we detect the squared modulus of the specimen exit wave function in momentum space. Hence the intensity I(p_x,p_y) in a certain pixel of the CCD is proportional to the probability that the corresponding momentum (p_x,p_y) is observed. Thus the centre-of gravity-type summation <p>=int; I(p_x,p_y) dp_xdp_y yields the expectation value <p> for the momentum, independently of the complexity of the diffraction pattern. We demonstrate how the momentum transfer can be related to the electric field convolved with the intensity of the STEM probe in thin specimens. In exploiting Maxwell's equations, we show how our method can pave the way to mapping charge and electron densities.
Third, we present case studies for GaN and SrTiO₃ to demonstrate the applicability of this concept. In a simulation study for GaN a set of 80x80 ronchigrams (corresponding to the raster of the STEM probe) has been simulated using the multislice method. These data allow for determining <p>, the electric field, charge- and electron density, and for a comparison with the density functional theory based counterparts. In an experimental study of SrTiO₃, a unit cell was rastered with an aberration-corrected STEM probe at 20x20 pixels, and ronchigrams have been recorded for each raster position. Using this 4D data set, we demonstrate the experimental applicability of our method by quantitatively determining both momentum transfer and electric field within the SrTiO₃ unit cell. By calculating the divergence of the electric field, we furthermore show the ability to image light atoms such as oxygen, while mapping charge densities in experiment is currently hindered by technical constraints such as scan noise.

We show that a straightforward interpretation of differential phase contrast and holography from materials containing electric fields is problematic. It is confirmed that electric fields in free space deflect the beam in the expected manner and that this is a sufficient explanation for the holography signals in this case. In contrast to this, beam deflections are minimal in polarised perovskites and thus free electric fields are not the main source of any differential phase contrast (DPC) or holographic signal. In contrast to this, strong contrast asymmetries are noted within the primary diffraction disc in areas where strong fields are present. It is found that this diffraction contrast results from polarisation, and this will be the dominant signal contributing to either differential phase contrast or electron holography signals from many materials subjected to electric fields. Thus, any interpretation of differential phase contrast or electron holography of electric fields in materials that does not take explicit account of diffraction contrast in the form of asymmetries within the primary disk is flawed. It is shown that the separation of the effects of E and P fields on the electron beam is best performed using scanned diffraction, and it is suggested that this would be an essential counterpart to future experiments using holography or differential phase contrast in order to provide an unambiguous interpretation of the signals

The strong focus on more efficient energy use and conversion, on more efficient transportation, and on environmental protecting technologies relies heavily on the advancement of (new) functional nanomaterials and nanosystems. At any stage in research and development, studies of these nanomaterials&’ structure, properties, and function are critical, including detailed atomic-scale insights.
Progress in technology and methodology has made scanning / transmission electron microscopes (S/TEM) powerful and indispensable tools for characterizing nanostructures. However, studies e.g. at room temperature and/or under standard high vacuum conditions might be inadequate to investigate the actual functional state of a material or system, whose properties depend on varying operating or environmental conditions.
Fortunately, in recent years the technology has also been significantly advanced to enable in situ studies while maintaining high-resolution imaging and analytical capabilities when applying in situ stimuli to functional nanomaterials, such as temperature, current, gas etc. Implementation of differential pumping apertures in an aberration corrected electron microscope (ETEM [1]) enables environmental studies, e.g. oxidation, reduction, or corrosion experiments [2]. Further development in in situ heating stages with faster settling time and more accurate knowledge of experimental conditions [3] enables quantitative atomic-scale studies at elevated temperatures in any (gaseous) environment.
Recent application examples will be presented to highlight in situ S/TEM capabilities and possibilities. As an example, utilizing FEI&’s NanoExtrade; [3] and ChemiSTEMtrade; [4] solutions, dynamic EDS studies at elevated temperatures on Boron/Nickel composite nanowires are essential to understand a potential growth model based on coalescence and mobility of Nickel nanoparticles inside the Boron nanowires [5]. Also, reversible pattern evolution of ferroelectric domains in BaTiO₃ (BTO) have been observed in heating cycles between room temperature and temperatures above the critical transition temperature (Tc) [6]. Recent relevant ETEM applications will be reviewed as well [1,2].
[1] http://www.FEI.com/ETEM
[2] J. R. Jinschek, Chemical Communications 50, 2696 (2014)
[3] http://www.FEI.com/NanoEx
[4] P. Schlossmacher et al., Microscopy Today 18 (4), 14 (2010)
[5] B. Barton, et al., Advances in Imaging and Electron Physics, submitted (2014)
[6] A. Schilling et al. Microsc. Microanal. 20 (3), 1560 (2014)

The application of electrical fields can enable the accelerated consolidation of materials during field assisted sintering, such as spark plasma sintering or flash sintering. Although such techniques are already employed for the synthesis of a wide variety of microstructures with unique macroscopic properties, a fundamental understanding of the atomic-scale mechanisms that lead to enhanced densification in the presence of electrical fields and/or currents is mostly absent from the literature. In situ transmission electron microscopy experiments will be reported that were designed to investigate densification mechanisms in the absence and presence of electro-static fields. Specific focus is on effects such as dielectric breakdown of passivating oxide layers, neck formation between individual nanoparticles as a function of applied electrical field strength, 3D sintering of Y-stabilized ZrO₂ while exposed to electrostatic fields, and strength of individual ceramic powder agglomerates that can limit densification. Such experiments allow the direct structural and chemical characterization of densification mechanisms on the atomic length scale. For Y-stabilzed ZrO₂, it was found that densification temperatures and time frames are significantly reduced in the presence of a critical electrical field strength but absence of any currents. The presentation will highlight the current experimental capabilities and report recent results.
Research is supported by the University of California Laboratory Fee Program (12-LR-238313) and the US Army Research Office (program manager: Dr. S. Mathaudu) under grant W911nf-12-1-0491-0.

Dendritic growth, recognized as a paradigm of non-equilibrium process, is commonly found in nature. However, it still lacks the fundamental understanding of branching mechanisms due to limited capability of imaging nanoscale dendritic growth in solution in situ. With the advances in liquid cell transmission electron microscopy (TEM), we studied the solid-liquid reactions as well as nanocrystal formation. We loaded a precursor solution of iron nitrate, olylamine, oleic acid and benzyl ether into the liquid cell and electron beam induced dendrites growth is observed. The crystal structure and chemical composition have been investigated by high-resolution TEM, electron energy loss spectroscopy (EELS) and energy dispersive x-ray spectroscopy (EDS). The local iron and oxygen ion concentration gradient between the dendritic tip and surrounding solution have been captured. Critical insights have been provided on the solution-based dendritic growth of oxide nanomaterials.
+C. H. C. and W. I. L. contribute equally to the work
We used TEM facilities of the National Center for Electron Microscopy (NCEM) at the Molecular Foundry of Lawrence Berkeley National Laboratory, which is supported by the U.S. Department of Energy (DOE) under Contract No. DE-AC02-05CH11231. HZ thanks the funding support from U.S. DOE Office of Science Early Career Research Program. CHC and WIL acknowledge the funding support from Ministry of Science and Technology (MOST) in Taiwan (No. 103-2917-I-009-185 and NSC 102-2119-I-009-502).

Nanoparticle growth, chemical reactions or biochemical activity often occur in the presence of a liquid. To study liquid samples in an electron microscope, several liquid cell designs have become commercially available in recent years that enable materials to be imaged in a carefully controlled liquid environment within the vacuum of a TEM. However, all suffer from a few key limitations that do not allow for atomic-resolution imaging or spectroscopy: 1) two Si₃N₄ layers (50-500 nm thick) used as electron transparent windows and 2) the thickness of the liquid surrounding the sample. Electron energy-loss spectroscopy (EELS) is degraded by multiple scattering events in the thick window layers, and the strong core-loss signals associated with the presence of Si and N. In addition to the increased sample thickness, radiation damage is a fundamentally limiting factor when examining beam sensitive materials and /or hydrous samples in TEM. It has been shown that coating the specimen with carbon or lowering the temperature have positive effects on radiation damage by reducing electrostatic charging, mass loss, loss of crystallinity, or defect formation rate. However, further reduction of radiation damage is needed for characterization of biological samples.
In this contribution, we will present a novel approach of encapsulating liquid containing samples in monolayers of graphene. This not only allows biological samples to be directly imaged at atomic resolution in a native liquid state without limitations from the window thickness, but also enables nm-scale analysis using EELS to quantify reactions in an aqueous environment. It will be shown that the energy deposited by the incoming electrons is dissipated by graphene from the area irradiated at a rate equivalent to the beam current of several electrons per Ų per second, reducing the effects of radiation damage and allowing for high-resolution characterization of beam sensitive materials. As a model system, ferritin molecules are examined, and it will be shown that individual Fe atoms or polypeptide of unstained protein can now be resolved in a liquid environment. EELS elemental identification of ferritin molecules with 1 nm resolution is achieved. By carefully controlling the induced electron dose rate, reactions, such as liquid/ gas phase transition can be quantified at selected locations in the graphene liquid cell at nm resolution

In bulk metals, mechanical strain is known not to influence electrical and thermal transport. However, their high volume fraction of grain boundaries and different deformation mechanisms may make nanocrystalline metals violate classical physics. To investigate this hypothesis, we developed an experimental approach, where we performed thermal and electrical conductivity measurements on 100 nm thick freestanding nanocrystalline aluminum films in situ inside a transmission electron microscope (TEM). We present experimental evidence of decrease in thermal conductivity and increase in electrical resistivity as a function of uniaxial tensile strain. In-situ TEM observations suggest that that grain rotation induced by grain boundary diffusion is the dominant deformation mechanism in these thin films. We propose that diffusion causes rise in oxygen concentration resulting in increased defects at grain boundaries. Presence of oxygen only at the grain boundaries is confirmed by energy dispersive spectroscopy. Increased defect concentration by mechanical strain at grain boundary causes the change in thermal and charge transport.

Over the years, Transmission Electron Microscopy (TEM) has been a primary characterization tool to understand the structure-property relationship of most of the materials. In most of the studies, the specimens are investigated post-mortem. There have been several successful attempts to carry out in situ TEM experiments wherein dynamic changes in a specimen are investigated while applying a stimulus like heating, electrical bias, mechanical deformation or exposing to a reactive environment. With the advancements in TEMs and micro-electro-mechanical systems (MEMS), the area of in situ TEM has progressed extensively over the last decade. Here, we show the application of MEMS based devices developed in-house to carry out in situ heat-treatment and environmental TEM studies. For environmental TEM studies, a controlled environment is established inside the TEM by one of the following approaches: the open type, using a differentially pumped vacuum system where the reactive gases are spread around the specimen area of the TEM; and the closed type, using a windowed environmental cell. Our studies are based on the closed environmental cell called the nanoreactor. The nanoreactor consists of two silicon chips facing each other with thin electron-transparent silicon nitride membranes. One half of the nanoreactor (bottom half) is embedded with a Pt coil for resistive heating. Using one half of the nanoreactor, we have demonstrated the growth of S-type precipitates in Al alloy 2024 while heating in the range of 180 - 250 °C. When closed with the other half, we have succeeded in studying the localized corrosion of Al alloy 2024-T3 at room temperature. The initial experiments were carried out by gluing both the halves of the nanoreactor together.
One of our recent developments is a holder that allows assembling a leak-tight nanoreactor without any gluing. The main advantage of using such a holder is the possibility to dissemble the nanoreactor to carry out any further analysis like chemical mapping and tomography after a chemical reaction with the surrounding environment. Furthermore, the contamination due to gluing is also avoided. In the present study, using this holder, we demonstrate the corrosive attack of an Al alloy that has been heat-treated at 250 °C in an environment of oxygen bubbled through aqueous HCl of pH 3. After corrosion, STEM-tomography has been carried out to understand the propagation of the corrosive attack. These sort of experiments are critical to understand the performance of engineering materials in service conditions. It is now possible to correlate the microstructural changes happening during the service of an engineering alloy, due to any processes that can heat the material like welding or higher operating temperatures, with the changes in the environment.

It is very common to synthesize and use some nanomaterials nowadays. Because most nanomaterials are prepared in the laboratory, especially for those synthesized using chemical methods in aqueous environments, it is difficult to observe and understand the reaction process accurately. Herein, employing transmission electron microscopy (TEM), aberration-corrected scanning transmission electron microscopy (STEM) and liquid holder techniques, we investigate the reaction of calcium sulfate with water. Using TEM, we obtain unambiguous evidence that the morphology and microstructure of calcium sulfate hemihydrates (Bassanite, CaSO₄bull;0.5H₂O) and calcium sulfate dihydrate (Gypsum, CaSO₄bull;2H₂O) are both changed when reacting with deionized (DI) water. Bassanite reacts with water to form gypsum, with the microstructure changing from powder to nanowires with needle or rod shape; when gypsum reacted with DI water, the microstructure changed to nanowires without any compositional or crystallographic variation, as proven by electron diffraction. However, the reaction of gypsum with water was a complex process that did not reveal new information on the microstructure evolution.
Using liquid holder under TEM mode, the structure transformation of the gypsum was tracked and recorded in real time. The gypsum powder was observed to dissolve in DI water. The dissolution of nanowires that formed during reaction of gypsum powder and water started when the nanowires were exposed to electron beam. Initially, fractures formed inside the nanowire. And then the size of the nanowires decreased as they dissolved in the water until they were invisible in TEM.

In this study, we design, simulate and fabricate a micro-electro-mechanical system (MEMS) based experimental setup, which is capable of performing mechanical tests inside a transmission electron microscope (TEM) at elevated temperature up to 1000 °K. The MEMS device is fabricated on silicon-on-insulator (SOI) wafer and has integrated heaters, force sensors and thermal actuators. The device can be co-fabricated with thin films deposited on the wafer, as long as they can be patterned and subsequently released from the substrate to create freestanding uniaxial tension specimens. Young&’s modulus, fracture stress as well as stress-strain relationship of thin films at high temperatures can be demonstrated to visualize deformation and fracture mechanisms inside the TEM. The device is demonstrated on metal films as well as graphene, molybdenum di-sulfide and boron nitride. Addition of microelectrodes in the device also allows in-situ measurement of thermal and electrical conductivity in-situ inside the TEM while the specimen microstructure is modulated by external stimuli such as temperature or stress.

A detailed analysis of thermal decoherence in electron beams is presented. We find that thermal noise during the interaction of the electrons with distant materials leads to observed stochastic deflection angles and losses of visibility in self-interference experiments that are relatively insensitive to the conductivity of the material. Interestingly, the electron undergoes angular broadening even in symmetric configurations, such as when it moves along the axis of a hollow tube. The presented formalism can be readily applied to arbitrary materials and interaction geometries. Besides their fundamental interest, these results are of interest for the improvement of spatial resolution in future electron microscopes.

Localized electric fields that are induced optically exhibit unique phenomena of fundamental importance to nanoplasmonics. In recent years, they have been considered for efficient photovoltaic and light harvesting devices, single molecule detection, biomolecular labeling and manipulation, and surface enhanced Raman scattering. Success has been made in developing experimental methodologies to probe the effect of their presence, but it remains difficult to directly and robustly image the optically induced near-fields, both in space and time. Herein, we introduce a novel imaging methodology that can directly map the near-fields of nanoplasmonics with spatiotemporal resolutions that were not possible before. Ultrafast transmission electron microscopy (UTEM) enables the direct visualization of the electric fields as they rise and fall within the duration of the excitation laser pulse (few hundreds of femtoseconds) with several nanometers of spatial resolution. This imaging approach is based on an inelastic photon-electron interaction process, where the probing electrons gain energy equal to the integer multiple of the photon quanta (2.4 eV in these experiments). This new phenomenon in electron energy loss spectroscopy and its fundamentals are discussed. Furthermore, we present images, and movies, of the near-fields of particle dimers, nanoparticles with different sizes and shapes, particle ensembles and standing-wave plasmons at the step edges of layered-graphene strips. These results establish UTEM as a tool with unique capabilities to approach nanoplasmonics.

In the first experimental imaging of Au atoms and nanometer sized Au particles using aberration-corrected STEM, copious movement of particles, including both attractive and repulsive forces, was immediately apparent [1]. Numerical calculations, by us and by others, confirmed this hypothesis and also identified a repulsive force for sub-nanometer impact parameters, [2] but did not yield a detailed understanding of the mechanisms that drive these forces. In this work, we calculated the time dependent response for a 2nm-size gold sphere within the bulk and on its surface, during and after the passage of a relativistic electron. We then evaluated the Lorentz Force on the sphere due to both electric and magnetic fields acting on charges and currents induced in the sphere. We found many fascinating results: First, as we expected, surface plasmons at optical frequencies produce forces at femto-second times which are likely mediated by the emission of photons into the EM fields. Second, at frequencies around 25 eV, we noticed a second plasmonic instability that apparently originates with the excitation of 5d electrons in Au. This instability also produces a strong peak in the inelastic electron scattering, and has been noticed in calculations of the dynamical response.[3] We found that these features produce a weak repulsive force at small distances. At higher energies, we notice a strongly confined wake pattern on the sphere having a wavelength that is significantly shorter than the particle diameter. Finally, at atto-second times the external time-like electric and magnetic fields of the fast electron interact with induced charges and currents within the sphere to produce strong attractive and repulsive forces. As the swift electron approaches the sphere, the attractive interaction of the applied electric field, with surface charges, is reduced by the increasing divergence of the field, while the repulsive interaction of the applied magnetic field, with induced surface currents, remains largely unchanged, producing net repulsive forces.
[1] P. E. Batson, A. Reyes-Coronado, R. G. Barrera, A. Rivacoba, P. M.
Echenique, J. Aizpurua, Nano Letters 11 (2011) 3388.
[2] A. Reyes-Coronado, R. Barrera, P.E. Batson, P. Echenique, A. Rivacoba and J. Aizpurua, PRB 82 235429 (2010). F. J. Garcia de Abajo, PRB 70
115422 (2004).
[3] I. G. Guturbay, J. M. Pitarke, I. Campillo, A. Rubio, Comp. Mat. Sci.
22 123-128 (2001).
We acknowledge financial support from the Basic Energy Sciences Division of the Department of Energy, Award #DE- 334 SC0005132, the Department of Industry of the Basque Government through the ETORTEK project inano, the Spanish Ministerio de Ciencia e Innovacion through Project No. FIS2010- 19609-C02-01, and the Consejo Nacional de Ciencia y tecnologí (Mexico) through Project No. 82073.

The progress in (scanning) transmission electron microscopy development had led to an unprecedented knowledge of the microscopic structure of functional materials at the atomic level. Additionally, although not widely used yet, electron holography is capable to map the electric and magnetic potential distributions at the sub-nanometer scale. Nevertheless, in-situ studies inside a (scanning) transmission electron microscope ((S)TEM) are extremely challenging because of the much restricted size and accessibility of the sample space and the large magnetic field at the sample position.
Here, we introduce a concept for a dedicated in-situ (S)TEM with a large sample chamber and several large uniform access ports for flexible multi-stimuli experimental setups. We report about the electron optical performance of a JEOL JEM-2010F retro-fitted with two Cs correctors converted from a dedicated low-voltage high-resolution (S)TEM into a large-chamber in-situ microscope. Both correctors are aligned to act as a corrected Lorentz lens in conventional as well as in scanning mode. We demonstrate a maximum resolving power of about 1 nm in conventional imaging mode and substantially better than 5 nm in scanning mode while providing an effectively usable “pole piece gap” of 70 mm. Additionally, we show a new portable sample stage solution, contributing to the flexibility of the "lab in the gap".
Finally, we discuss the possibilities to further improve the electron optical performance of large sample chamber (S)TEMs with nowadays state-of-the-art technology.

The interaction of nanosecond pulsed laser with materials has many important technological and commercial applications ranging from Pulsed Laser Deposition to dental applications and even tattoo removal. Theoretical models predict that two main mechanisms by which removal of materials from underlying substrates can occur. The first is direct heating and evaporation, and the second is thermo-mechanical stress driven fracture. Both mechanisms drive materials removal and droplet ejection during laser ablation. However, experimental methods with rapid time and spatial resolution are required to validate the models and directly visualize the ablation process. We have conducted nanosecond pulsed laser experiments on plan-view TEM samples of nickel thin films on (100) silicon substrates using the Dynamic Transmission Electron Microscope (DTEM). Time-temperature profiles for the investigated interface structure were calculated using the COMSOL multiphysics package by solving the Fourier heat diffusion equation. It was experimentally observed that the nickel film begins dewetting within the first 20 ns and form psuedolinear wires in areas with temperatures that are above the melting temperature of the nickel thin-film. As time proceeds, the substrate fractures and nanometric particles were ejected from the sample. We will show the experimental results capturing the evolution of the nanosecond laser ablation with high temporal and spatial resolution in combination with COMSOL simulation data modeling the time-temperature profiles. This work is supported by University of California Laboratory Fee award under Contract No. 12-LR2383.

Direct observation at Atomic revel was realized by using aberration corrected transmission electron microscope (Cs-corrected TEM/STEM). Especially, since an aberration correction system has been practically established [1-3], electron microscopes with the system drastically enhanced analysis capability in a scanning transmission electron microscopy (STEM) and structural study in transmission electron microscopy (TEM).
Recently, we developed a new atomic resolution electron microscope of GRAND ARM (JEM-ARM300F) with several new technologies (which were called G-Features) as a new platform for a super high-resolution microscopy. The developed microscope is equipped with an ultra-stable cold field emission gun and new aberration correctors for probe and image forming systems. These correctors were using dodeca-pole lens and special electron trajectory which was developed in R005 project [3]. The spherical aberration is compensated by two three-fold astigmatism fields [1] generated in two dodeca-poles. As a further optical innovation, an electron trajectory is expanding toward a specimen in the corrector and largely expanding between the condenser min lens and the transfer condenser mini-lens (CM-CMT) or between the objective min lens and the transfer objective mini-lens (OM-OMT) [3]. The expanding trajectory enables us to reduce disturbance in above elements resulting that extra chromatic aberration and noise from the corrector can be reduced. We call the system as ETA (Expanding Trajectory Aberration) optical system.
We newly developed two types of pole pieces of an objective lens for this microscope. They are FHP (full high resolution pole piece) for ultra-high resolution configuration and WGP (wide gap pole piece) for analytical high resolution configuration. The stability of the microscope in TEM with FHP was tested by a lattice fringe of a crystal specimen and Young&’s fringe for a thick specimen showing beyond (50 pm)-1 spatial information in these images. Ga-Ga dumbbells separated by 63 pm for a GaN [211] specimen was resolved in STEM high angle annular dark field imaging. As a further challenging, a sub-50 pm resolution was imaged in HAADF. 47-pm separation was conformed in the intensity line profile, and a spatial information of (47 pm)-1 was detected in the power spectrum. Moreover, we also developed dual SDD detecting system for highest sensitivity of X-ray analysis which was optimized WGP. This new detector system has about 3 times higher sensitivity than HR pole pies with 100mm² SDD of ARM200F.
The microscope of “GRAND ARM” have capability to image sub-50 pm resolution. The stability must be effective not only ultra-high-resolution imaging but also robust data acquisition for analysis and structure study. The microscope will be new platform for an atomic resolution study.
[1] Haider, M., Uhlemann, S., Schwan, E., Rose, H., Kabius, B., and Urban, K. ; Electron microscopy image enhanced. Nature, 392, 768-769. (1998)

The dynamic transmission electron microscope (DTEM) is a photo-emission TEM capable of nanosecond-scale time-resolved electron imaging and diffraction. DTEM is distinguished from other photo-emission TEM techniques in that each electron pulse contains enough electrons (~10¹⁰) to form an image or diffraction pattern [1], making it well-suited for studying the irreversible transformations that are central to materials processing. The technique has been used to study a variety of phase transformations, ranging from solidification of metals [2] to the crystallization of phase change materials [3]. The DTEM at Lawrence Livermore National Laboratory, originally built to capture single images, is now capable of generating multiple nanosecond-scale electron pulses spaced over several microseconds. An electrostatic deflector placed below the TEM imaging system deflects each image to a different part of the CCD camera, overcoming the ~ms limit on the refresh rate of the camera.
Laser crystallization of amorphous semiconductors is an important processing path for electronic devices. Laser heating enables extremely high heating rates to highly localized area of a device. It is possible to achieve a variety of grain sizes and textures, including large grain sizes, which are important for the processing of thin film transistors. Large grains are believed to develop through “explosive” crystallization, where heat released from crystallization promotes further propagation of the crystal growth front.
Single-shot DTEM has been used to study the crystallization kinetics of amorphous Ge during laser crystallization [4-5]. In regions of explosive crystallization, the crystal growth rate was estimated to be roughly 8 ± 2 m/s [5]. In our current work, multi-frame DTEM is used to re-examine the crystallization kinetics of amorphous semi-conductors during laser heating. The nine-frame movies captured during a single crystallization event in Ge thin films allows the crystalline growth front to be accurately tracked even as the growth front velocity exceeds 10 m/s. This has revealed that growth rates remain steady during explosive crystallization in spite of significant temperature gradients established during laser heating. Simulated temperature profiles are used to connect measured crystallization rates with kinetic models growth. The role of “amorphous melting” on the growth front velocity and on the extreme nucleation rate gradients that arise during laser annealing will be discussed.
This work performed under the auspices of the U.S. Department of Energy, Office of Basic Energy Sciences, Division of Materials Sciences and Engineering by Lawrence Livermore National Laboratory under Contract DE-AC52-07NA27344.
References
[1] LaGrange et al., Micron43 (2012) 1108
[2] McKeown, et al., Acta Mater65 (2014) 56
[3] Santala et al., Appl Phys Lett102 (2013) 174105
[4] Nikolova et al., Appl Phys Lett97 (2010) 203102
[5] Nikolova et al., Phys Rev B87 (2013) 064105

Brittle to ductile transition (BDT) of silicon at macro-scale has been extensively investigated. However, the effect of size on BDT remains unclear. In order to study BDT, a new bending stage is proposed to test silicon beams at micro/nano scale at different temperatures. In case of uniaxial tension, any flaw within the entire volume of the specimen may cause fracture before yielding. The new stage minimizes uniaxial tension while maximizing pure bending stress to avoid premature fracture. Long supporting beams are implemented during straining to allow the ends of the bending sample to approach each other with minimal resistance, thus minimizing the tension on the sample. Moreover, to overcome specimen handling and misalignment problems reported in literature, the stage and the specimen are designed to be co-fabricated together. The force and displacement sensors are built on the same structure with the specimen to minimize the experimental setup size, allowing both quantitative and in-situ tests to be performed in SEM and TEM chambers.
This study introduces an analytical model of the stage which predicts the deflection and stresses generated in the specimen beam upon loading with the aid of a piezo-actuator. A numerical model is built using finite element (FE) package to analyze the mechanical behavior of the stage to eliminate the assumptions used in the analytical model. Good agreement is found between analytical and numerical results. Therefore, the analytical model can be used to further optimize the stage performance.

Bulk silicon is known to be brittle at low temperatures, but is ductile above 540C. It has been postulated that at small scale, the brittle to ductile transition (BDT) transition may occur at lower temperatures. To test this possibility, tension and bending experiments were carried out at room temperatures on small scale samples in the literature. The results were inconclusive, i.e., tension samples were ductile but bending samples were brittle. In order to resolve this paradox, we carried out bending tests on single crystal silicon beams with widths varying from 8 micro meters to 720 nm at various temperatures. We found, BDT temperature indeed decreases with size, reaching to 293C for 720 nm sample. We then developed a dislocation based model of silicon subject to pure bending. The model accounts for the stress gradient and high Peierls stress of silicon. The model offers a closed form relation between the yield stress of silicon and its size scale. Approximately, the yield stress increases inversely as the square root of the length scale under bending, but it is size independent under uniaxial tension. At small scale, due to reduced flaw size, the fracture stress is also high. Thus, at small scale, the yield stress may be less than fracture stress under uniaxial tension, but higher than fracture stress under bending at low temperatures, and the sample is brittle under bending and ductile under tension. With increased temperature, flow stress decreases for both the cases due to reduction of Peierls stress, and both tensile and bending samples behave as ductile. The theoretical model and the BDT experiments together resolve the current paradox in the literature.

Current passage through thin nanowire can lead to atoms displacements at high current densities. This phenomena is called electromigration [¹]. The electromigration process in Pt nanobridges at elevated temperatures was investigated by in situ transmission electron microscopy (TEM), using a FEI Titan microscope operating at 300 keV. The combination of micro-electro-mechanical-system (MEMS) based heater as a substrate for nanobridge and a special electrical sample holder, built in-house, allows to visualise changes in a specimen under both dynamic conditions, i.e. heating and current passage.
Polycrystalline Pt nanobridges (500 nm width, 1000 nm length) with a thickness of 15 nm were produced by e-beam evaporation onto the MEMS heater with a flat centre (600-nm-thick freestanding silicon nitride membrane) [²]. The MEMS chip has 6 contacts - 4 for the heater spiral and 2 for the electrical measurements. The resistance change as a function of the bridge temperature was investigated by doing I/V measurements in a range ±50 mV. These measurements were done with temperature increments of 100C (up to 5000C). Based on deviations of a linear behaviour of the resistance-temperature curve several substrate temperatures were selected for electromigration experiments. To allow the system to equilibrate, the sample was maintained at the target value for 3 minutes. The I/V measurements were done in bias-ramping mode, i.e. gradual increase from 0 V to a predefined value 400-800 mV, followed by a decrease back to 0 V, and optional continuation into the negative range, after it a new cycle with higher maximum voltage was done [³].
Grain growth starts at 2700C, which was observed to lead to a gradual reduction of the electrical resistance by 40%. After the grain growth temperature coefficient of resistance (TCR) was found to be 1.47×10^-3 0C^-1 which is 2.5 times smaller than TCR for bulk Pt. The smaller conductivity (and TCR) for thin film in comparison with bulk material can be explained by additional surface scattering and grain boundary scattering of the electrons [⁴].
The electromigration experiments at various temperatures show the same tendency: material transport occurs from the cathode to the anode side, which can be explained with the electron-wind force. But at higher substrate temperatures the process of holes formation is enhanced in the areas of contact pads. These holes lead to irreversible bridge destruction. As expected, at higher temperatures lower power is needed to break the nanobridge.
Acknowledgement: The authors gratefully acknowledge NIMIC and ERC project 267922 for support.
1. Ho, P. S.; Kwok, T. Rep Prog Phys 1989, 52, (3), 301-348.
2. Van Huis, M.A.; Young, N.P.; Pandraud, G.; Creemer, J.F.; Vanmaekelbergh, D.; Kirkland, A.I.; Zandbergen, H. W. Advanced materials 2009, 21, (48), 4992-4995.
3. Kozlova, T.; Rudneva, M.; Zandbergen, H. Nanotechnology 2013, 24, 505708.
4. Mayadas, A. F.; Shatzkes, M.; Janak, J. F. Appl Phys Lett 1969, 14, 345.

Cu₂Se CdSe (NCs) nanocrystals are well-established colloidal nanomaterials due to their semiconductor properties, p- and n- type respectively, and other such as optoelectronic and thermoelectric. In particular, antifluorite Cu₂Se phase when heated above a threshold temperature becomes superionic (SI) so that Cu cations are able to diffuse randomly with liquid-like mobility around the rigid Se sublattice. By further temperature increase, SI Cu₂Se NCs expel free Cu species from the lattice, forming then Cu-vacancies in cation sublattice, with consequent change of stoichiometry into Cu_2-xSe, and modifying its electronic structure. In this direction, the formation of these Cu-vacancies (free holes) can active the oscillation of positive charge carriers, resulting in a near-infrared absorption band of NCs in the range 900 nm to 1600 nm (1.38 eV to 0.77 eV). Besides, changes in the electronic structure of Cu-depleted Cu_2-xSe materials can be correlated to the appearance of a plasmon absorption band in the electron energy loss (EEL) spectrum, showing a peak at about 1.1 eV detectable by high resolution EEL Spectrometry.
The free Cu species expulsion from Cu_2-xSe NCs at high temperature can be exploited to perform solid state cation exchange reactions of nano-compounds at local scale by simple thermal activation. Here, we present the results of thermally driven cation exchange at solid state between SI Cu₂Se and wurtzite CdSe nanorods (NRs) and nanowires (NWs) when deposited randomly on the same substrate (carbon film of TEM grid). After in-situ thermal treatment (400°C), under TEM high vacuum (1.5x10^-5 Pa), SI Cu_2-xSe shows the expected Cu-depletion (x = 0.15); on the other hand, wurtzite CdSe undergoes a pervasive chemical transformation, revealing a total loss of Cd species with concomitant Cu substitution. Total exchanged-CdSe NRs and NWs experience also a complete structural transformation from wurtzite to antifluorite crystal structure preserving the close-packing direction of Se atoms in the structures, namely [0001]_hcp and [111]_fcc.
These studies of as-obtained cation exchange evidence how: 1) the shape features and textural relationships of both nanoparticles species are preserved; 2) in absence of SI Cu₂Se NCs, CdSe NRs or NWs do not undergo any transformation in the same thermal range; c) the cation exchange reaction occurs also in regions not irradiated by the electron beam, confirming that the as-obtained cation exchange is a simple thermally driven process.
Acknowledgments.
This work is in part supported by ERC TRANS-NANO project with contract number 614897.

The ground-potential monochromator recently introduced by Nion [1] is giving 10 meV resolution in electron energy loss spectra (EELS) recorded at 60 keV, with a 1-2 Å wide electron probe. By approaching the energy resolution of commonplace optical techniques like Raman spectroscopy, but combined with orders of magnitude better spatial resolution in the electron microscope, many new types of experiments are becoming possible.
We are at the beginning of the exploration. Some of the first results will be reported by our collaborators elsewhere in this session [2] including the detection of optical phonons at energy losses ranging from 70 to 360 meV in a wide variety of materials [3]. Studying the spatial extent of the phonon interaction shows the signal has two components: a “delocalized” component that is relatively well understood in terms of classical dielectric theory, and a “localized” signal associated with high-angle scattering. The delocalized signal has applications for radiation sensitive samples—in particular for hydrogen detection.
This talk will focus on future directions such as isolating the localized signal for exploring the vibrational properties of materials at near-atomic spatial resolution, detecting hydrogen also at high spatial resolution, isotopic separation at the atomic scale, and using the added optical complexity that comes with the monochromator to generate vortex beams of a single orbital angular momentum for studying magnetic properties of materials—possibly also at excellent energy and spatial resolutions.
The localized vibrational signal should be enhanced by selecting scattering angles of several mrad and higher, and studying vibrational signatures from boundaries and individual defects may soon be within reach. Also, at scattering angles greater than about 50 mrad, the energy transfer that accompanies incoherent Rutherford scattering provides another highly informative signal. The atomic nucleus essentially behaves as free, and can be “weighed” by the magnitude of the energy transfer to it [4]. The motion of the nucleus broadens the loss peak by Doppler shift, and this could potentially be used to study the nuclear motion. The instrumental requirements for this type of spectroscopy are especially demanding - excellent energy resolution must be combined with large EELS acceptance angles - but we are making good progress in this direction too.
[1] O.L. Krivanek et al., Microscopy 62 (2013) 3-21.
[2] P.A. Crozier et al., MRS April 2015 session ZZ (same session)
[3] O.L. Krivanek et al., Nature 514 (2014) 209-212.
[4] T.C. Lovejoy et al., Microsc. Microanal. 20 (suppl. 3) (2014) 558-559.

We have recently shown that it is now possible to perform vibrational spectroscopy in the electron microscope using monochromated electron energy-loss spectroscopy [1]. While techniques such as Raman and IR spectroscopy have been routinely used to characterize vibrational modes in solids and surfaces for many years, the vibrational spectrum probed with focused fast electron beams is completely unexplored. This new form of spectroscopy is enabled by a unique monochromator allowing energy resolutions in the range 10-20 meV to be routinely achieved. Moreover, the monochromator is interfaced to an aberration corrected scanning transmission electron microscope (Nion UltraSTEM) with an electron probe size of about 1 Å. The combination of both high energy and spatial resolution provides a completely new tool for research in materials, by allowing vibrational spectroscopy to be combined with traditional atomic resolution STEM imaging techniques. Results will be presented on the application of vibrational spectroscopy to different classes of materials, including ceramics, polymers and catalysts. The spatial resolution of the vibrational signal is a topic of ongoing research and appears to show both localized and delocalized components. Variations in the spatial resolution of optical phonon signals at surfaces and interfaces will be presented. Changes in the phonon spectrum with particle size will be investigated. The delocalized component of the vibrational signal opens up the possibility of detecting vibrational excitations in the so-called aloof beam mode. Such an approach may have applications in radiation sensitive material by allowing the surface character of chemical functional groups such as hydroxides and carbonates to be determined, while avoiding knock-on or radiolytic damage.
[1] Krivanek, O. L., T. C. Lovejoy, N. Dellby, A. Toshihiro, R. W. Carpenter, P. Rez, S. E., J. Zhu, P. E. Batson, M. J. Lagos, R. F. Egerton and P. A. Crozier (2014). "Vibrational spectroscopy in the electron microscope." Nature 514: 209-2012.
[2] The authors acknowledge support support from NSF to purchase the instruments (NSF DMR 1308085, NSF MRI-R2 959905) and research support from and DOE DE-SC0004954 and DE-SC0005132. The authors acknowledge John M. Cowley Center for High Resolution Microscopy at Arizona State University.

Recent developments in monochromated electron energy-loss spectroscopy have demonstrated that energy resolutions in the range 10 - 20 meV can be routinely achieved [1]. When combined with the small probe-forming capability of a scanning transmission electron microscope, the combination of high energy resolution with spatial resolution approaching 1 A opens up new avenues for materials nanocharacterization. Here we explore some of the new forms of information that can be obtained on oxide systems using a series of doped CeO₂ materials as model systems. In this system, optical, electrical and catalytic properties can be readily altered by changing the type and concentration of lower valence cation dopants such as Gd, Pr, Ca and Co [2]. Moreover extended defects such as grain boundaries can play an important role in controlling materials properties. Recently, we demonstrated a significant change in the electrical conductivity of grain boundaries in Gd/Pr doubly-doped ceria which we speculated—based on spatially-resolved STEM EELS analysis of grain boundary Pr content could be caused by small polaron electronic conduction [2] along networked grain boundaries. This change in macroscopic electrical conductivity is expected to accompany a change in the electronic structure (e.g. inter-bandgap electronic states associated with the Pr). In this contribution we employ ultra-high energy resolution EELS in a monochromated Nion UltraSTEM100 to explore interband states in ceria doped with high levels of Pr. If these states are observed in bulk grains then we will look for localized interband states at grain boundaries in materials with lower doping levels. The dielectric of the material at the grain boundary will influence electrical properties. Low-loss EELS may be employed to explore variations in the dielectric response in the grain boundary region. Cerenkov radiation may complicate the interpretation of the spatially resolved spectral data, but may be constant to a first approximation for rare earth doping. Careful analysis of valence loss spectra may also reveal changes in bonding across the boundary. interband state may only be present at the intergranular phase so high spatial resolution will be critical. Along with inter-bandgap states we will also explore detection of optical phonons at grain boundaries, particularly in ceria doped with relatively light element such Ca.
1. Krivanek, O. L., et al. (2014). "Vibrational spectroscopy in the electron microscope." Nature 514: 209-2012.
2. Tuller, H. L., Bishop, S. R. Point Defects in Oxides: Tailoring Materials Through Defect Engineering. Annual Reviews of Materials Research41 369 (2011).
4. Support from NSF DMR 1308085, NSF MRI-R2 959905, DOE DE-SC0004954, DOE DE-SC0005132, ASU LECSSS & CHREM, NSF GRFP DGE-1311230, and NSF DMR-1308085. Opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the NSF.

Nanometer-scale semiconductors that contain a few intentionally added impurity atoms can provide new opportunities for controlling electronic properties. However, since the physics of these materials depends strongly on the exact arrangement of the impurities, or dopants, inside the
structure, and many impurities of interest cannot be observed with currently available imaging techniques, new methods are needed to determine their location. We combine electron energy loss spectroscopy with annular dark-field scanning transmission electron microscopy (ADF-STEM) to image individual Mn impurities inside ZnSe nanocrystals. While Mn is invisible to conventional ADF-STEM in this host, our experiments and detailed simulations show consistent detection of Mn. Thus, a general path is demonstrated for atomic-scale imaging and identification of individual dopants in a variety of semiconductor nanostructures.

With the development of aberration correctors and monochromators in transmission electron microscopy, the improved stability of electron optics and brighter electron sources, atomic-scale chemical analysis of a wide range of materials has become routinely available and widely accessible. Electron energy loss spectrometry (EELS) in the scanning transmission electron microscope (STEM) is the technique of choice to combine both spatial and chemical information, via elemental mapping in crystals at the atomic level [1]. The unoccupied density of states probed in core-loss EELS contains large amounts of information on the electronic structure and related properties of the material investigated. While for elemental mapping, the full absorption edge can be selected, additional information on the electronic structure and properties can be obtained by selecting a specific feature in the fine structure (FS) of the energy loss near edge structure (ELNES). These FSs arise from transitions to unoccupied states of a particular energy, thereby allowing to identify localized states at the atomic scale.
The relevance of such FS mapping will be shown for the case of cuprate superconductors. The localization of holes in these compounds is of primary interest to shed light on their physical properties from the nano- to the macro-scale. Contrary to most probes that integrate over all CuO₂ planes in the structure, such as X-ray absorption spectroscopy, STEM-EELS allows to be spatially selective [2,3]. In the present study, the chain-ladder superconductor Ca₁₁Sr₃Cu₂₄O₄₁, which consists of CuO₂ chains and Cu₂O₃ ladders, is investigated by STEM-EELS at the atomic level. Spectroscopic information at the O-K and Cu-L_2,3 edges is in good agreement with XAS results available in the literature. The hole distribution is studied from the O-K pre-edge ELNES, composed of the O-2p hole band, and the upper Hubbard band [4]. Using the chemical mapping technique available from STEM-EELS spectrum imaging, we show that the holes are mostly concentrated within the CuO₂ chains of the structure. The possibility to quantify the relative hole concentration within the chains and the ladders is evaluated, and discussed in comparison with available results from other characterization techniques. In addition, simulations were performed to assess the influence of channeling effects. The electronic structure fingerprint in the ELNES, combined with atomic resolution in the aberration corrected STEM, opens the way to better understand the electronic properties of high T_c superconductors [5].
[1] S. J. Pennycook and C. Colliex, MRS Bull. 37, 13 (2012)
[2] N. Gauquelin et al., Nat. Comm. 5 (2014)
[3] J. Fink et al., Phys. Rev. B 42, 4832(R) (1990)
[4] A. Rusydi et al., Phys. Rev. B 75, 104510 (2007)
[5] The experimental work has been performed at the Canadian Center for Electron Microscopy, a national facility supported by NSERC and McMaster University.

The capabilities of energy dispersive X-ray spectrometry (EDS) using silicon drift detectors (SDDs) in TEM and SEM are explored in relation to other complementary EM-based analysis techniques. Various materials science problems were studied, e.g. silicon-based and III-V semiconductors, magnetic nanostructures [1], core-shell particles, coated multiwall carbon nanotubes [2] and life science samples. Although SDD-EDS is well established, the attainable high spatial resolution and accuracy of quantitative analysis continues to impress.
Optimal photon detection geometry in the microscope, including a large solid angle, a high take-off angle, good collimation and a suitable sample holder, is crucial for efficient EDS and will be explained. Multi-detector arrangements allow the realization of large solid angles at useful take-off angles [3, 4, 5]. At high take-off angle absorption and shadowing effects can be minimized. Good collimation is necessary to ensure a high signal to background ratio. This geometric optimization in combination with high beam current and aberration correction allows the identification of single atoms in graphene at 0.1sr [6]. Higher solid angle in STEM offers quantitative analysis down to 0.02at% (Rb in an Orthoclase mineral standard) in reasonable measurement time [7]. An annular detector design in SEM achieves a solid angle of more than 1sr at superb take-off angles of 50° to 70°. This enables element analysis of electron transparent samples in SEM on the nm scale and fast EDS of samples with high topography or/and high radiation sensitivity and large areas of interest [8]. For quantitative EDS of electron transparent samples the relative Cliff-Lorimer-method and the absolute Zeta-factor-method can be used. To correctly interpret EDS element maps on the atomic scale though, simulations of relevant scattering and radiation effects are necessary [9], similarly to interpreting EELS data.
EDS of electron transparent samples in SEM can be combined with other emerging complementary SEM-based techniques: micro-XRF allows trace analysis for higher Z elements at low spatial resolution and Transmission Kikuchi Diffraction offers crystallographic analysis on the nm-scale [10].
[1] C. Brombacher et al., Appl. Phys. Lett. 97 (2010) 102508.
[2] S. Hermann et al., Microelectronic Engineering Vol. 87 (3) (2010) 438-442
[3] S. von Harrach et al., Microsc. Microanal. 15 (Suppl. 2) (2009) 208.
[4] P. Schlossmacher et al., Microscopy Today 18(4) (2010) 14-20.
[5] R. Terborg et al., Microsc. Microanal.16 (Suppl. 2) (2010) 1302-1303.
[6] T.C. Lovejoy et al., Appl. Phys. Lett. 100 (2012) 154101.
[7] Z. Gainsforth et al., Microsc. Microanal. 20 (Suppl. 3) (2014) 1682-1683.
[8] S. Rades et al. RSC Adv., 2014, 4(91), (2014) 49577-49587.
[9] B. D. Forbes et al., Phys. Rev. B 86 (2013) 024108.
[10] R. R. Keller, R. H. Geiss, J. Microscopy 245 (2012) 245.

Recently, the observation of a #64257;rst-order magnetic/electronic transition in certain Pr-based perovskite cobaltites, such as Pr_0.5Ca_0.5CoO₃, has attracted significant attention. More specifically, a simultaneous metal to insulator transition (MIT), a sharp drop in magnetic moment and a change in the electronic structure has been reported to occur at around 90 K. It was suggested that the low-temperature phase in Pr_0.5Ca_0.5CoO₃ is stabilized by a shift of the mixed valence Co³⁺/Co⁴⁺ toward pure Co³⁺, enabled by a valence change of Pr³⁺ to Pr⁴⁺ by transferring electrons from the Pr 4f levels into hybridized Co 3d-O 2p orbitals. The small difference in energy of the two Pr valence states makes this transition possible. This hypothesis has been confirmed by the occurrence of a Schottky peak in heat-capacity measurements and x-ray absorption spectroscopy at the Pr L₃ edge.
In this contribution, we will present an atomic-scale study of (Pr_1-yY_y)_0.70Ca_0.30CoO₃ using simultaneous high-angle annular dark field and annular bright field imaging, electron energy-loss spectroscopy and in-situ cooling experiments. The valence state transition in (Pr_1-yY_y)_0.70Ca_0.30CoO₃ occurs at a transition temperature T_MIT~135 K for y=0.15, and the in-situ cooling experiments are conducted at 90 K. At room temperature, we find oxygen vacancy ordering in polycrystalline (Pr_1-yY_y)_0.70Ca_0.30CoO₃. We will demonstrate that the electron transfer occurs from Pr to Co below the transition temperature using in-situ cooling experiments and show that oxygen vacancy ordering vanishes as a result of the Co valence state transition. The effects of oxygen mobility, sample homogeneity and the impact on the observed transition will also be discussed.

We use angle-and polarization-resolved cathodoluminescence imaging spectroscopy to determine the local modes and angular radiation profile of nanophotonic structures at deep-subwavelength resolution. We determine the resonant electric and magnetic plasmonic dipole and quadrupole modes of single Au plasmonic nanoscatterers and patch antennas. The dipolar and quadrupolar modes can be selectively excited, depending on the electron beam position. Strong angular beaming is observed from these antenna&’s, with the angular profile determined by the superposition of radiation by multiple plasmonic modes that are coherently excited by the electron beam.
Using a rotating quarter-wave plate and a linear polarizer in the optical beam path we retrieve the full polarization state of the CL emission in the far field as a function of emission angle. We demonstrate this novel type of CL polarimetry on plasmonic bull&’s-eye gratings milled into single crystal gold. For central excitation, these gratings are driven in phase, leading to an azimuthally symmetric emission pattern and a radial polarization distribution. When excited off-center, the emission pattern shows multiple lobes and alternating regions in angular space in which the emission varies between circular and linear polarization. We then analyse bull&’s eyes with chiral structures of different handedness, and find their chirality reflected in the emitted CL distribution

Design of complex metal nanoparticles offers a great playground for plasmonic nanoengineering. Here we show that it is possible to resolve plasmon resonances from ultraviolet to near infrared by modifying the metal nanoparticle morphologies from solid Ag nanocubes to hollow AuAg nanoframes and multiwalled hollow AuAg nanoboxes by using electron energy loss spectroscopy (EELS) in a monochromated scanning transmission electron microscope (STEM). We experimentally demonstrate that structural modifications, i.e. void size and final morphology, are the dominant determinants of plasmonic properties in hollow AuAg nanostructures, while compositional variations allow us fine tuning of plasmon resonances. Electron energy loss spectroscopy (EELS) mappings of localized surface plasmon resonances (LSPRs) reveal an increased plasmon energy and intensity field inside the voids of hollow AuAg nanostructures along with the enhanced extends of plasmon intensities around the nanostructures. Simulations by discrete dipole approximation (DDA) are consistent with obtained results and reveal the effects of structural nanoengineering on plasmonic properties of hollow metal nanostructures. Possibility of tuning the LSPR properties of hollow metal nanostructures in a wide energy range by modifying the void size/shell thickness is shown by Mie scattering theory calculations, which reveal that void size is the dominant factor for tuning the LSPRs. As a proof of concept for enhanced plasmonic properties, we show the effective giant label free sensing of bovine serum albumin (BSA) of single-walled AuAg nanoboxes in comparison with Au NPs demonstrating their excellent performance for future biomedical applications.

Monochromated Scanning Transmission Electron Microscopy (STEM) and Electron Energy Loss Spectroscopy (EELS) has proven to be an effective technique in studying Local Surface Plasmon Resonances (LSPR) at the nanoscale. Qualitative changes in the local density of optical states, due to changes in the dimensions of nanomaterials, can be imaged with EELS due to a high spatial resolution. Effects of resistive and capacitive coupling between neighboring metal nanoparticles can be measured by quantification of shifts in LSPR spectral peaks. In this work we use Monochromated EELS to study how the chemistry, microstructure, and local dielectric environment changes the behavior of LSPR&’s. First, the effect of grain boundaries on the dampening of LSPR&’s was studied. Colloidal, single-crystal Au nanorods and lithographically-defined polycrystalline Au nanoantennas are used as a model system. It was observed that a majority of the resonant modes in single-crystal Au nanorods have a quality factor that is 1.5 to 2.5 times greater than in the polycrystalline nanoantennas. We also apply the same technique to noble metal alloys, doped metal oxides, and other plasmonic materials to better understand the effect of dopant and alloying elements on the dampening of LSPR&’s.
[1] Work at the Molecular Foundry was supported by the Office of Science, Office of Basic Energy Sciences, of the U.S. Department of Energy under Contract No. DE-AC02-05CH11231.

Metamaterials based on nanoscale split ring resonators (SRRs) have been proposed as potential amplifiers of surface plasmons by stimulated emission of radiation (spasers). Recent theoretical work demonstrates that an array formed by four upright U-shaped SRRs, deployed in a toroidally symmetrical cell, with the lower bar of the U parallel to the radial direction, can function as optically active spasers based on surface plasmon toroidal dipole resonance modes [1].
Single SRRs exhibit magnetic resonance modes, with frequency and mode order dependent on the dimensions of the lower beam on the U shape and on the height of the prongs [2]. When combined in a toroidal cell the SRRs show clear plasmon resonances for both magnetic and toroidal modes as was shown by numerical simulations [1], [3]. Analysis with visible and near infra-red light has been performed on vertical arrays of U-shaped SRRs, with a lower beam length of approximately 100 nm, and plasmonic resonance modes have been detected in these arrays through dips in the transmission spectra of light passing through the metamaterial. The SRRs show an increase in resonance wavelength with an increase in the length of the lower bar of the U structure and an increase in mode order with an increase in prong length [2].
In this work using electron beam lithography we fabricate vertical arrays of four U-shape SRRs arranged in a toroidal configuration on a 50 nm thick silicon nitride membrane. Using a state of the art TEM-STEM system equipped with an electron monochromator and a high-resolution electron energy loss spectrometer, in combination with deconvolution to improve the energy resolution (approaching the 10 meV) [4], we probe the SRR arrays. We thus excite multiple resonant plasmon modes and detect the resonances as low energy losses in the electron beam. We acquire high-resolution spectral maps of multiple bright and dark magnetic resonance modes and toroidal resonance modes. Our results demonstrate distinct toroidal resonances in the visible spectral range that can be easily tuned to longer wavelengths by changing the physical dimensions of the SRRs, opening the possibility for plasmonic devices and applications across multiple ranges of the electromagnetic spectrum.
References
[1] Y.-W. Huang, W. T. Chen, P. C. Wu, V. A. Fedotov, N. I. Zheludev, and D. P. Tsai, Sci. Rep., vol. 3, Feb. 2013.
[2] W. T. Chen, C. J. Chen, P. C. Wu, S. Sun, L. Zhou, G.-Y. Guo, C. T. Hsiao, K.-Y. Yang, N. I. Zheludev, and D. P. Tsai, Opt. Express, vol. 19, no. 13, pp. 12837-12842, Jun. 2011.
[3] Y.-W. Huang, W. T. Chen, P. C. Wu, V. Fedotov, V. Savinov, Y. Z. Ho, Y.-F. Chau, N. I. Zheludev, and D. P. Tsai, Opt. Express, vol. 20, no. 2, pp. 1760-1768, Jan. 2012.
[4] E. P. Bellido, D. Rossouw, and G. A. Botton, Microscopy and Microanalysis, vol. 20, no. 03, pp. 767-778, Jun. 2014.

Picosecond and femtosecond spectroscopy allow for a detailed study of carrier dynamics in nanosctructured materials [1]. In such experiments, a laser pulse usually excites several nanostructures at once. However, spectroscopic information may also be acquired using pulses from an electron beam in a modern scanning electron microscope (SEM), exploiting a phenomenon called cathodoluminescence (CL) where electrons are promoted from the conduction band to the valence band upon impingement of the high energy electron beam onto a semiconductor. This approach offers several advantages over the usual optical spectroscopy. The multimode imaging capabilities of the SEM enable the correlation of optical properties (via CL) with surface morphology (secondary electron mode) at the nanometer scale [2] and the large energy of the electrons allows the excitation of wide-bandgap materials.
Here, we present results obtained on a field emission time-resolved cathodoluminescence scanning electron microscope. The microscope can either be operated in CW mode by heating up the emitter (Schottky emission), or in time-resolved mode by illuminating the field emission gun with a femtosecond UV laser, so that ultrafast electron pulses are emitted through the photoelectric effect. In both modes, a space resolution of 10 nm is demonstrated. The collected cathodoluminescence signal is dispersed in a spectrometer and analyzed with a CCD camera (CW mode) or an ultrafast STREAK camera to obtain 10 ps time resolution (TR mode).
Quantitative CW cathodoluminescence was first used to quickly map ZnO and GaN-based nanowires and identify their local energy structure. Then, time-resolved cathodoluminescence measurements were carried out on specific regions in order to measure local lifetimes and carrier diffusion within the nanowires.
[1] Shah, J. Ultrafast Spectroscopy of Semiconductors and Semiconductor Nanostructures, Ch. 8 (Springer, Berlin, 1999).
[2] Reimer, L. Scanning Electron Microscopy, Ch. 1 (Springer, Berlin, 1998).
[3] M. Merano et al., Nature 438, 479 (2005).

Stimulated by both instrumental (monochromators, detectors) and methodological (signal deconvolution and processing) advances, fast electron based spectroscopies have demonstrated their unique potential in probing surface plasmons (SP) of metallic nanostructures. Their nanometer spatial resolution and ability to probe so-called dark modes have given Electron Energy Loss Spectroscopy (EELS) and Cathodoluminescence spectroscopy (CL) a central role in experimental nano-optics. Today, these techniques are used to investigate nanostructures of increasing complexity in which the particle morphology, the substrate, or the interparticle interactions strongly influence their optical response[1]. Several examples of recent breakthroughs in combined electron/optical spectroscopy techniques such as Electron Energy Gain Spectroscopy demonstrated in ultrafast Transmission Electron Microscopes or surface plasmon three-dimensional imaging push forward the need and development of novel simulation techniques.
In this context, we have developed a novel simulation technique allowing to describe thoroughly the interaction of fast electrons with metallic nanostructures. Building on the 3D Green Dyadic Method, our technique yields accurate predictions of the energy losses and CL photon emission consecutive to the interaction of a moving charge with a metallic nanostructure. It can be applied to nanostructures of arbitrary morphology, both penetrating and non-penetrating trajectories and rigorously takes into account the dielectric response of the substrate. Several examples will be presented which show an excellent agreement with recent experimental results. The influence of the substrate on the EELS spectra will be addressed. EELS spectra and maps (Fig. 1-a), CL spectra, maps and radiation patterns (Fig. 1-b) of several gold nanostructures from well-known textbook examples (nanoprisms, rods...) to more complex architectures (nanoporous films, particle aggregates, Fig. 1-c-d) will be presented. Finally, the potential of our technique will be illustrated on complex scenarii involving electron/photon interactions[2].
Acknowledgement: The authors acknowledge financial support from the European Union under
the Seventh Framework Programme under a contract for an Integrated Infrastructure Initiative
Reference 312483-ESTEEM2 and the National Research Agency under the program ANR HYNNA (ANR-10-BLAN-1016)
[1] Surface Plasmon Damping Quantified with an Electron Nanoprobe
M. Bosman, E. Ye, S. F. Tan, C. A. Nijhuis, J. K.W. Yang, R. Marty, A. Mlayah, A. Arbouet, C. Girard, and M. Y. Han
Scientific Reports, 3, 1312, (2013)
[2] Electron energy losses and cathodoluminescence from complex plasmonic nanostructures : spectra, maps and radiation
patterns from a generalized field propagator
Arnaud Arbouet; Adnen Mlayah; Christian Girard; Gérard Colas des Francs
N. J. Phys, in press, (2014)

Tiny confined optical modes of high-quality Q factor are fundamental building blocks in the field of photonics, for instance for the development of ultrahigh sensitive sensors and quantum optical devices. Imaging their electromagnetic fields is intrinsically difficult as the main part of the information about the field is stored in evanescent contributions. Although techniques such as near-field scanning optical microscopy are able to investigate these evanescent contributions they are restricted to surface imaging. Here, we demonstrate the imaging of high- Q modes confined in dielectric planar photonic crystal nanocavities using high-energy resolution low-loss EELS. The technique is based on the Vavilov-Cherenkov (VC) effect, in which the coupling of a moving charge with the local optical density of states in a dielectric medium induces energy loss. Unlike earlier studies applied this effect to imaging local fields using “aloof mode” EELS in the vacuum [1], here we probe the mode inside the solid matrix, bringing particular challenges of poor signal-to-noise. Taking the data in monochromated scanning TEM mode we retrieve information about the intensity distribution and the local state of polarization of the optical field at the dense core of the cavity with a spatial resolution a factor of 30 better than that given by the light optical diffraction limit for the structures.
Results will be presented for a so-called L3 photonic crystal cavity that is defined by a line defect of three missing holes inside a triangular lattice of holes (lattice constant a = 460 nm) in a 220 nm thick lithographically prepared silicon membrane. With the membrane perpendicular to the electron beam, only transverse magnetic (TM) modes contribute to the electron energy loss via the VC effect; as theoretically predicted the transverse electric (TE) cavity mode is not observed, thus proving that the EELS signal is not subject to spurious residual luminescence unlike that which might occur for cathodoluminescence. By then tilting the membrane towards the electron beam there is coupling to TE modes, giving a small but unambiguous peak at the 0.82 eV energy-loss that corresponds to the 1500 nm wavelength cavity mode. The spatial distribution of this peak&’s intensity further matches to the theoretical intensity distribution of the electric field modulus in the cavity. While initial results for one sample tilt have been taken using the SESAM instrument at Stuttgart [2], data with more sample orientations will be presented from the monochromated Titan Themis just installed at EPFL. By doing so, theoretical analysis shows it should ultimately be possible to determine the electric field at particular local positions along the electron&’s trajectory. This creates the possibility of making 3D tomographic analyses of the electric field in optical cavities, in a highly-spatially resolved and noninvasive way.
[1] Cha et al. PRB 81, 113102
[2] Le Thomas et al. PRB 87, 155314

Ternary InGaN compounds show great promise for light-emitting diode (LED) applications because of bandgap energies (0.7-3.4 eV) that can be tailored to have emission wavelengths spanning the entire visible spectrum. Complex III-N device heterostructures have been incorporated into GaN nanowires (NWs) recently, but exhibit emission linewidths that are broader than expected for their corresponding planar counterparts using photoluminescence (PL) spectroscopy [1]. It is thus critical to understand how the structural and optical properties interplay, using combined spectroscopic methods that can resolve different localized signals at the nanoscale within analytical scanning transmission electron microscopy (STEM).
In this work, multiple InGaN/GaN quantum dots (QDs) embedded NW LEDs grown on Si(111) substrates by molecular beam epitaxy, with an additional p-AlGaN electron-blocking layer (EBL) incorporated to mitigate electron overflow, were characterized by STEM. Nanometer-resolution STEM-cathodoluminescence (CL) spectral imaging on single NWs was performed using a custom-made system on a VG HB-501 STEM [2]. Individual NWs showed diverse optical responses, but most NWs exhibit one main emission peak centered at 500-600 nm in the yellow-green. Spectral features consisting of multiple sharp peaks (25-50 nm at FWHM) spanning a wavelength range of ~100 nm arise from within the QD active region, identified using the high-angle annular dark-field (HAADF) signal collected concurrently, and show an apparent spatial dependence of the spectral shifts. This is consistent with the PL, indicating that the broad emission originates from within single NWs and is not an inhomogeneous broadening. Subsequent atomic-resolution HAADF imaging on an aberration-corrected STEM on the same NWs were acquired to evaluate their structural properties, such as the size and morphology of the 10 QDs within the NW active region. High-resolution STEM-electron energy-loss spectroscopy (EELS) spectrum imaging was used together with MLLS fitting to extract the In-distribution for quantifying the In-composition. Direct spatial-spectral correlation can be made between shifts in the emission wavelength from the STEM-CL to the relative In-composition between successive QDs from the STEM-EELS. The luminescence intensity within NWs shows a direct relationship to the thickness of the GaN shell surrounding the InGaN/GaN heterostructure and the presence of an AlGaN shell formed during the growth of the AlGaN EBL due to lateral diffusion. Both enhance the in-plane confinement of carriers, hence reducing non-radiative surface recombination. Additionally, absence or even poor coverage of the additional AlGaN shell drastically changes the luminescence spatial symmetry. The effects of the presence of built-in and piezoelectric fields on the emission intensity distribution will also be discussed.
[1] Nguyen et al. (2012) Nano Lett. 12, 1317.
[2] Zagonel et al. (2011) Nano Lett. 11, 568.

The incredible improvements in instrumentation for electron microscopy over the last decade have opened new windows into the nanoscale world. Beyond simply making smaller, brighter electron probes, recent efforts have pushed the boundaries in other directions. Driving down the electron energy spread, creating beams in particular momentum states and pushing the limits of temporal scales have all but redefined the meaning of electron microscopy itself.
Complementary to these advancements in electron sources are the detector improvements required to support these unique applications. Focusing just on forward scatted electrons, the ultimate detector for the electron microscope would allow the collection of every transmitted electron while simultaneously recording its energy and momentum with both spatial and temporal resolution limited only by the uncertainty principle. While we are still a long way off from this goal, strides have been made to drive up the collection and detection efficiency of EELS systems while at the same time pushing the energy resolution, stability and time resolution to match the source advances.
In this presentation, we will review the current state of detectors for electron microscopy with an emphasis on systems for EELS measurements. By comparing with the ideal, we will explore how close we are getting and what strides are being made to close the performance gap.

In a transmission electron microscope (TEM), materials are probed by electron plane waves, but alternative beams are possible and especially cylindrical harmonics are an interesting option. Here, the plane waves are replaced by waves with a typical azimuthal phase factor exp(i m phi;), phi; being the angle in the plane perpendicular to the optical axis and m the so-called topological charge. Such waves also referred to as vortex waves, carry an orbital angular momentum (OAM) m#295; as well as a quantized magnetic moment mmu;_B due to the electron charge [1]. Electron vortex beams have been successfully applied to rotate nanoparticles [2] and characterize chiral crystals [3].
Here, we use the magnetic field created at the tip of a long ferromagnetic rod, approximating a magnetic monopole field [4], to create high purity and high intensity electron vortex beams with sub-aring;ngström resolution. After carefully tuning the cross section of the rod by focused ion beam (FIB) milling, OAM values very close to one can be obtained. This is revealed by the measurement of the phase shift caused by the magnetic field close to the tip of the rod using holography in field-free conditions. The presence of an electron vortex was verified through a focal series experiment and by cutting the defocused probe with a sharp edge.
The interaction of such an electron beam with an antiferromagnetic sample will be discussed in detail. Special attention is paid to the atomic resolution magnetic dependence of the EELS, linked to electron magnetic chiral dichroism (EMCD).
[1] K. Bliokh et al., Phys. Rev. Lett. 99 (2007) 190404.
[2] J. Verbeeck, et al., Advanced Materials (2013), 25, p1114-1117.
[3] R. Juchtmans et al., Submitted to PRL (2014).
[4] A. Béché et al., Nature Physics (2014), 10, p. 26-29.

X-ray magnetic circular dichroism is a well-established method to study element specific magnetic properties of a material, while electron magnetic circular dichroism (EMCD), which is the electron wave analogue to XMCD, is scarcely used today. Recently discovered electron vortex beams, which carry quantized orbital angular momentum (OAM) L, are promised to also reveal magnetic signals [1]. Since electron beams can be easily focused down to sub-nanometer diameters, this novel technique provides the possibility to quantitatively determine local magnetic properties with unrivalled lateral resolution. In order to generate the spiralling wave front of an electron vortex beam with an azimuthally growing phase shift of up to 2π and a phase singularity in its axial centre, specially designed apertures are needed [2]. Dichroic signals on the L₂ and L₃ edge are expected to be of the order of 5% [3,4].
For this purpose, we have successfully implemented a spiral aperture into the condenser lens system of a FEI Titan³ 80-300 transmission electron microscope (TEM) equipped with an image spherical aberration corrector. This setup allows for the generation of focused electron vortex beams with user-selectable OAM that can be used as probes in scanning TEM (STEM). Since for such spiral apertures, the different OAM are dispersed along the beam direction, tuning the focal length of the condenser lens allows for a selection of the OAM.
First investigations were focused on probing the presence of an EMCD signal with such vortex beams on different thin ferromagnetic Ni and Fe films and on hard magnetic L1₀ ordered FePt nanocubes. Here, the diameter of the L = 0 probe (= full with at half maximum, FWHM) is 0.14 nm, whereas the |L| = 1 probes have diameters of roughly 0.3 nm.
However, first experiments did not provide any evidence for differences in the absorption edges in the electron-energy-loss spectra (EELS) generated by vortex beams with different OAM. In contrast, comparative experiments on FePt nanocubes utilizing a classical EMCD setup have revealed magnetic dichroism in these samples.
In order to understand the lack of any dichroic signal when using spiral apertures, the generation and propagation of the vortex wave function and the spatial distribution of the OAM were simulated for the given experimental setup. The simulations reveal that although the OAM is largely localized (in all three dimensions) symmetrically around the geometrical focal points, the superposition of the selected vortex state (e.g., L = +1) with contributions from adjacent vortex states (e.g., L = 0 and L = -1) results in a suppression of the total OAM. Possible escape routes to this problem will be discussed.
[1] J. Verbeeck et al., Nature 467 (2010), p. 301-304.
[2] J. Verbeeck et al., Ultramicroscopy 113 (2012), p. 83-87.
[3] P. Schattschneider et al., Ultramicroscopy 136 (2014), p. 81-85.
[4] J. Rusz and S. Bhowmick, Phys. Rev. Lett. 111 (2013), 105504.

Electron magnetic chiral dichroism (EMCD) is the electron wave analogue of X-ray magnetic circular dichroism (XMCD). It offers the possibility to study magnetic properties at the nanoscale in a transmission electron microscope (TEM) [1]. In a “classical” EMCD setup, the sample is illuminated with a plane electron wave and acts as a beam splitter. Although this method is well established, so far only very few EMCD spectra were obtained from individual nanoparticles [2,3].
We report on “classical” EMCD measurements on individual FePt nanocubes with a side length of roughly 20 nm and compare our experimental findings with simulations. L1₀ ordered FePt is a particularly important material, since it offers the highest magneto-crystalline anisotropy among the oxidation-resistant hard magnets. It is therefore a promising materials candidate for future high density magnetic data storage media [4]. The L1₀ ordered FePt nanocubes were prepared by inert gas condensation. Prior to the spectroscopic measurements the samples were exposed to an external magnetic field of H_ext = 7.5 T applied perpendicular to the substrate. Thereby, all cubes with [001] zone axis orientation should give a maximum dichroic response (with the same sign for all particles), since the magnetization is then (anti)parallel to the electron beam.
The dichroic signals at the L₂ and L₃ edges are expected to be as small as just 10 % of the total scattering intensity. Due to these very small intensity differences in the electron energy loss spectra (EELS), a proper microscope alignment, a good reproducibility, a sufficiently stable sample and a decent signal-to-noise (S/N) ratio pose extreme experimental demands. The experiments were performed on a FEI Titan³ 80-300 microscope equipped with an image C_S corrector, and binned-gain acquisition of the EEL spectra was used to optimize the S/N ratio [4]. Our experiments were supported by simulations utilizing the WIEN2k program package [5], based on which FePt cubes with an optimal thickness were chosen for highest EMCD signals [6]. The likewise conducted experiments indeed reveal a small but reproducible dichroic signal that agrees well with the results of the theoretical calculations.
[1] P. Schattschneider et al., Nature 441 (2006), p. 486.
[2] J. Salafranca et al., Nano Lett. 12 (2012), p. 2499.
[3] Z.Q. Wang et al., Nature Comm. 4 (2013), p. 1395.
[4] M. Bosman and V. J. Keast, Ultramicroscopy 108 (2008), p. 837.
[5] K. Schwarz and P. Blaha, Computational Materials Science 28 (2003), p. 259.
[6] S. Löffler and P. Schattschneider, Ultramicroscopy 110 (2010), p. 831.

Almost all material properties are defined by electronic states. Not only do they define, e.g., optical, electrical, and magnetic properties, they also give rise to the bonds between atoms, thereby affecting properties such as stiffness, cohesion, or adhesion. Despite their great importance, the direct mapping of individual atomic orbitals in bulk specimens has not been possible so far. In this work, we demonstrate a technique based on the combination of scanning transmission electron microscopy (STEM) and electron energy loss spectrometry (EELS) to select transitions to specific orbitals and map them in real space with sub-Ångström resolution. The probe electrons can
exchange energy and momentum with the sample. The change in momentum results in a spatial intensity modulation which can be seen directly. The energy spectrum, which is determined by EELS, exhibits characteristic peaks at the different transition energies of the material. The energy loss near edge structure is strongly influenced by the density of states. Thus, a narrow energy window can be used to select transitions to orbitals corresponding to a particular energy. As a model system, a 20 nm thick Rutile (TiO2) single crystal was used. Rutile has a tetragonal unit cell with a crystal field which gives rise to an energy splitting of the Ti conduction states into a t2glike and an eg-like band [1]. This was subsequently exploited to selectively map the transitions to states with eg symmetry [2]. A clear asymmetry was found which corresponds to the direction of the Ti-O bonds and is in strong contrast to what one would expect for isolated atoms without any crystal field. In addition to the experiments, simulations were performed to be able to better understand and cross-check the results. The inelastic scattering was described based on density functional theory data calculated using WIEN2k [3]. The elastic scattering was taken into account by a multislice approach [4]. The simulated maps are found to be in excellent agreement with the experimental
data. Based on both simulations and experiments, we also discuss the limits of the approach. In particular, we study the dependence of the maps on the point group symmetry of the scattering atom and the sample thickness. In addition, the effect of intensity, drift, and damage, as well as possible approaches to overcome the issues they cause, are discussed. The possibility to map orbital information in real space with atomic resolution opens new possibilities for material science. In particular, it paves the way for directly studying the electronic structure at, e.g., defects, interfaces, or quantum dots, all of which are of paramount technological importance. [1] Jiang et al., Acta Cryst A 59 (2003) 341; [2] Löffler et al., in preparation; [3] Blaha et al., TU Wien (2001); [4] Kirkland, Plenum Press (1998). Financial support by the Austrian Science Fund (FWF) under grant number I543-N20 and SFB F45. FOXSI is gratefully acknowledged.

In this talk, we report a direct experimental real-space mapping of magnetic circular dichroism with atomic resolution in aberration-corrected scanning transmission electron microscopy (STEM). Using an aberrated electron probe with customized phase distribution, we reveal with electron energy-loss (EEL) spectroscopy the checkerboard antiferromagnetic ordering of Mn moments in LaMnAsO by observing a dichroic signal in the Mn L-edge. The aberrated probes allow the collection of EEL spectra using the transmitted beam, which results in a magnetic circular dichroic signal with intrinsically larger signal-to-noise ratios than those obtained via nanodiffraction techniques (where most of the transmitted electrons are discarded). The novel experimental setup presented here, which can easily be implemented in aberration-corrected STEM, opens new paths for probing dichroic signals in materials with unprecedented spatial resolution. This research was supported by the Center for Nanophase Materials Sciences (CNMS), which is sponsored at Oak Ridge National Laboratory by the Scientific User Facilities Division, Office of Basic Energy Sciences, U.S. Department of Energy (JCI), by the Swedish Research Council and Swedish National Infrastructure for Computing (NSC center) (JR), and by the Materials Sciences and Engineering Division Office of Basic Energy Sciences, U.S. Department of Energy (MAM, ARL), and by UT-Battelle, LLC, under Contract No. DE-AC05-00OR22725 with the U.S. Department of Energy (CTS, RRV).

In this work, we have developed the site-specific electron energy-loss magnetic chiral dichroism (EMCD) method for magnetic structure determination in magnetic materials with non-equivalent crystallographic sites at a nanometer scale. It&’s the first work to experimentally demonstrate that the fast electron as a new source can be used to determine magnetic structure for a wide range of materials, which is generally considered to be accomplished by neutron diffraction. Compared with previous EMCD works in which EMCD was just used for detecting the ferromagnetic signals of materials, we fundamentally raise the EMCD technique to the new level of magnetic structure determination.
In the example of NiFe2O4, we achieve comprehensive magnetic structure information using the site-specific EMCD method under the assumption of no magnetic information known previously. The magnetic structure information we obtain includes site-specific total magnetic moment, site-specific orbital to spin magnetic moment (mL/mS) ratio and total magnetic moment of a unit cell. Our method is testified to be valid by comparing our results with those obtained by theoretical calculations and other experimental techniques such as XMCD and neutron diffraction.
The site-specific EMCD method we developed has its unique values compared with previous magnetism characterization techniques such as XMCD and neutron diffraction. In comparison, using transmitted electron in site-specific EMCD method, we can reach a high spatial resolution, and get site-specific and element-specific magnetic information, as well as distinguish the orbital and spin magnetic moments. For example, using site-specific EMCD method our work first reports the experimentally determined mL/mS ratios of Fe atoms in octahedral and tetrahedral sites.
In the technical aspects, the extremely strong EMCD signals acquired using site-specific EMCD method are at least three times high than the previous reported EMCD spectra, which allow us to do quantitative analysis. We first did the quantitative works on EMCD spectra to obtain total magnetic moments (the sum of spin and orbital magnetic moments) for atoms at different sites.
In sum, our work opens the door of using fast electrons to determine magnetic structures for a wide range of magnetic materials in a nanometer scale. Site-specific EMCD may benefit much not only to the fundamental research of magnetic states and behavior in complex magnetic materials, but also to revealing the magnetic structure in nanostructures or interface of the composite magnetic films.