Recently LACBED (Large Angle Convergent Beam Electron Diffraction) studies for identifying the nature of stacking faults has been reported in [1,2]. Here we report the LACBED study for a microtwin whose images are usually similar to those of an intrinsic or an extrinsic stacking faults (for this discussion, see [3]).

Observations: Thin foils of cobalt with the thickness of about 180 nm (f.c.c phase, a=0.354 nm) were examined by a Philips CM200. Fig. 1 shows strong beam dark field images of microtwins or stacking faults. Fig. 2 shows the bright field LACBED pattern taken near the area marked as a circle in fig. 1. The specimen height, from the convergent point of beams, was about 0.0586 mm and the convergent angle was 0.615 degrees.

Calculations and analysis: Analysis of fig. 1 alone indicates the encircled fault an extrinsic stacking fault.

Recent experiments with X-rays and high energy electrons have shown that image recovery from diffracted intensities is possible for non-periodic objects using iterative algorithms. Application of these methods to biological molecules raises the crucial problem of radiation damage, which may be quantified by Q = ΔE σi/σe, the amount of energy deposited by inelastic events per elastic event. Neutrons, helium atoms and low energy electrons below most ionization thresholds produce the smallest values of Q (see for TMV imaged at 60 eV). For neutrons (λ = 10-2Å, and deuterated, 15N-abelled molecules) Q is ∼3000 times smaller (∼50 times for λ = 1.8Å) than for electrons (80- 500keV) and about 4x 106 times smaller than for soft X-rays (1.5Å). Since σe for neutrons is about 105 times smaller than for electrons (and about 10 times smaller than for soft X-rays), a 105 times higher neutron dose is required to obtain the same S/N in a phase contrast image compared with electrons, if other noise sources are absent.

Three-dimensional reconstruction in TEM is an established technique in both electron diffraction crystallography and electron tomography [1,2]. By stereo lattice imaging, we refer to a different task: determination of unit cell parameters in 3D using only direct-space lattice images at two or more (often widely varying) orientations [3,4], This note concerns the uncertainties associated with lattice parameter determinations by this technique. in particular, for decreasing crystal thickness the broadening of reciprocal lattice spots increases one’s chances to visualize lattice fringes. At the same time, however, it introduces uncertainties in reciprocal lattice vector length and direction.

In the following discussion, sources from which uncertainties arise are assumed to be uncorrected. Three sources of uncertainty for reciprocal lattice vectors which lie in the plane of the specimen stage (and hence can be imaged directly as fringes) are, in order of decreasing relative size as follows:

The phrase “fringe visibility map” here refers to high-resolution images of uniform-thickness single crystal foils showing locally hemispheric deformation (i.e. bent into the shape of a watchglass), and to various mathematical analogs thereof. As the availability of “diffraction information” in direct-space form increases, for example with the availability of HREM and z-contrast lattice images, fringe visibility maps may serve as the direct space analog to Kikuchi maps, i.e. to diffraction pattern maps of Kikuchi line pairs as a function of specimen orientation, and their various cartoonifications.

Figures 1 to 3 are schematic Mathematica-generated fringe visibility maps for face-centered, body-centered, and diamond face-centered cubic lattices, respectively. Here we have turned the “locally hemispheric” concept around and drawn fringes visible when one is viewing a spherical particle of the appropriate thickness, along varying beam directions (rather than lattice orientations) parallel to the local radius vector on that sphere.

Introduction: Modern TEM/STEM instruments with field emission gun and various analytical equipment allow the measurement of interfacial properties by a range of modes and techniques without exchanging the specimen. It is therefore useful to think of a criterion to assess the sensitivity of each technique, e.g. for the location of light atoms, by a commonly applicable formula.

Sensitivity measures: A “quantitative” TEM experiment is defined here as consisting of a digital comparison (e.g. via data difference or cross correlation) between experimental (E) and simulated (S) microscopy signal. The matching quality R = f(E,S) is to be optimised (minimum for data difference, maximum for cross correlation). The target quantity of the Q-TEM experiment is assumed to be the location or occupancy of a light atomic column in projection surrounded by heavier atoms in the core of an interface making use of dynamical enhancement effects.

Grain boundaries of covalent bonding materials such as silicon, germanium and SiC require rather physical consideration than that of geometry to discuss the result of their electron microscopic observation. Dangling bond, for example, is expected least in number in the stable boundary (least dangling bond rule). [1][2] in practice, further precise consideration must be done both in geometry and in physics in the covalent bonding materials boundary. in the present study, geometrical discussion is combined together with that of the atomic bonding about diamond, SiC and ZnO grain boundaries. Chemical structure image of the modern high-voltage transmission electron microscope enabled to do it. An asymmetrical Σ3 boundary of diamond and a 6H-SiC boundary with 70.5 degree rotation are mainly observed.

The sliced specimen, 0.3mm in thickness, was mechanically thinned down to 0.02 mm and was finally polished using an argon ion milling. Final thickness of the specimen was estimated to be 3 nm in the region close to the edge of the small hole.

Grain boundaries in fee metals with low stacking fault energy are known to undergo extended relaxations that can at times lead to a thin layer of a different structure. in Cu, for example, it has been found that Σ3{112}} boundaries relax into a 9R phase or even a bec phase. in this work, we have used high resolution electron microscopy to investigate the atomic structure of Σ99 grain boundaries in mazed bicrystal films of Au. On {001} Ge single crystal substrates, Au bicrystals can be grown in two equivalent {110} orientation variants, related to each other by a 90° rotation about the surface normal. There is equal probability to nucleate either orientation variant in early stages of film formation and since there is no reason for preferential growth, the morphology of the bicrystal film consists of many grains in only two orientations. All grain boundaries are of the type Σ99 and display tilt character.

Thin films 10-200nm thick were grown by physical vapor deposition on {001} Ge substrate at 280°C and followed by annealing at 200°C to relieve stresses.

A variety of properties of polycrystalline TiO2, such as conductivity, creep and grain boundary diffusion, are strongly dependent on the atomic-level structure and chemistry of grain boundaries. TiO2 has been studied as a model dielectric system because of its relatively simple structure and well-understood point defect chemistry. Defect segregation in grain boundaries of polycrystalline and bi-crystal TiO2 have been studied by several groups and significant variations in solute segregation levels from boundary to boundary were observed. in this paper, we address this issue of anisotropic grain boundary segregation. We have measured solute segregation as a function of grain boundary misorientation to determine any correlation between segregation and misorientation.

Yttria-doped TiO2 polycrystalline samples were prepared by mixing 99.999% pure TiO2 powder with 0.1% mole percent 99.99% purity Y(NO3)3 (both powders are commercially available from Aldrich Co. or Alfa Co.), followed by uni-axial pressed to 200-400Pa and sintering at 1300 °C for about 5-7 hours.

We report a novel technique to accurately measure interfacial lattice displacement by forming an electron probe close to a specimen in a manner similar to that originally proposed by Gabor to record a hologram. This method is based on the quantitative analysis of the interference pattern of shadow images in coherent electron diffraction. The approach is unique in that there are no adjustable microscope parameters, the contrast is strong even when the fault is viewed edge-on, and a large number of shadow images of the fault corresponding to different Bragg reflections can be studied simultaneously. Since it is an interferometric technique, the spatial resolution of the measurement is not limited by the wavelength of the fast electrons. An accuracy down to 1pm has been demonstrated in measuring the displacement associated stacking faults and grain boundaries in superconductors. It is, to our knowledge, the highest that has been achieved in measurements of displacement vectors.

Quantitative analysis of a structure means to determine which atoms are where. The where is restricted by the lateral resolution, and the which is related to the sensitivity against differences of the atomic species. Electron holography is on the move to tackle both questions.

• where - Lateral Resolution Electron holography has surmounted conventional electron microscopy in that it has reached atomic resolution both in the amplitude and phase images: with our CM30FEG/UT Special Tuebingen electron microscope, the information limit of 0.09nm and the pixel number of the 2048_ - CCD camera allow to take holograms with image details smaller than 0.095nm; the quality of correction of the aberrations allows an interpretable resolution approaching 0. lnm both in amplitude and phase [1].

• which - Phase Detection Limit (PDL) The phase detection limit describes the minimum phase difference detectable well above a given noise level.

During the last five years the technique focal-series reconstruction has evolved to a powerful tool for investigating materials science problems in high-resolution transmission electron microscopy. Compared to the conventional interpretation of one single high-resolution image, the quantum mechanical electron wave function at the exit plane of the object (exit-plane wave function, EPW) is in many cases a better starting point for the materials analysis. The retrieval of the EPW is achieved on a routine basis by applying automated numerical procedures to a series of images taken from the same specimen area at different objective lens defocus values. The application of the reconstruction procedure allows one to remove numerically all the well known instrumental artifacts, such as nonlinear contrast formation or delocalisation effects due to spherical aberration and other parasitic aberrations. The reconstructed EPW gives thus direct insight into the atomic structure in the case of sufficiently thin objects and renders tedious image simulations of complicated defects unnecessary in many cases.

Recent progress in High Resolution Transmission Electron Microscopy makes it possible to investigate crystalline materials by phase contrast microscopy with a resolution close to the 80 pm information limit of a 300 kV field emission microscope'"". A reconstruction of the electron exit wave from a focal series of lattice images converts the recorded information into interpretable resolution. The present contribution illustrates some recent applications of this technique to interfaces.

Fig. 1 shows a reconstructed electron exit wave of a heterophase interface between GaN and sapphire. The experiment takes advantage of three factors: First, we resolved the GaN lattice in projection, which requires at least 0.15 nm resolution. The projection eliminates the stacking fault contrast that usually obscures lattice images in the commonly recorded projection. Thus, image interpretation is drastically simplified. Second, all atom columns at the interface and in the sapphire are resolvable with a smallest projected aluminum - oxygen spacing of 85 pm in the sapphire.

Polycrystalline ZnO was sintered with 5 mol.% Ga2O3 at 1350°C for lhour in air. Samples for transmission electron microscopy (TEM) were prepared by cold pressing, grinding, dimpling and Argon ion-milling.

ZnO crystallizes in the hexagonal wurtzite structure. Bright field and dark field images reveal lamellar defect structures which are not observed in undoped ZnO. Electron diffraction shows that the lamellar regions consist of heavily twinned ZnO. The twin boundaries (TB) of both twin variants are parallel to lattice planes of type { }. Lattice images along reveal narrow planar twin boundaries as well as a diffuse image contrast within the twin lamellae (Fig. la). to characterize the nature of the boundaries we recorded dark field images of the twinned regions using reflections of type ±{0002}. in combination with microdiffraction it can be shown that the diffuse boundaries are inversion domain boundaries (IDB) which alternate with the sharp TB. This microstructure is related to the polarity of ZnO.

In polycrystalline materials the trapping of charge at interfaces has a decisive influence on the electrical transport properties through the formation of electrostatic potential barriers. This can either be an intrinsic phenomenon or can be related to impurity segregation leading to complex defect centers. This plays a key role in technologically important systems, especially those having electrically active interfaces, e.g.,p-n junctions in semiconductors and grain boundaries in electroceramics.

Electroceramic oxides such as ZnO and SrTiO3 are common systems that exhibit the tendency of current control by internal potential barriers. While bulk measurements, either electrical (e.g. P-E, C-V, I-V), or optical (e.g. Raman) have contributed significantly to the understanding of charged interfaces, there are a very few direct observations of electrical activity at a nanometer level. For instance, it has been recognized that space charge and dopant segregation at the grain boundary are inter-related.

The potential variation, charge transfer and lattice distortion across grain boundaries in polycrystalline superconductors play a crucial role in the transport properties of the material. in particular, the grain interfaces often act as weak links to the superconducting current, which limits the critical current densities that may be obtained. in the case of the high Tc superconductor Bi2Sr2CaCu2O8 (Bi-2212), special interest is paid to (001) twist boundaries since, in part, they are dominant in c-axis aligned tapes and wires, but more interestingly because the superconducting critical current is found to be independent of twist angle for this system. Such grain boundary behavior is presently not understood, and is further puzzling since “impurity phases” of Bi-2223 or Bi-2201 (consisting of additional or missing Cu-Ca-0 planes, respectively) are commonly observed at the (001) twist boundaries. with the rich structure found at these types if interfaces, understanding the potential distribution may lead to new insight about the boundary properties.

EELS fine structures on characteristic core edges are closely related to the bonding type and structural environment around the excited atoms. Consequently, they constitute useful hints for investigating the electronic properties at interfaces. One major difficulty then lies in the extraction of the information relevant to the involved atoms, from that due to neighbouring atoms on both sides of the selected interface seen end on by the impinging electrons. We have investigated possible ways to reach this goal, using the spectrum-line method, which consists in acquiring sequences of spectra at well defined increments (0.3 nm) while scanning a subnanometer electron probe (0.7 nm) across the interface. This approach constitutes one step forward with respect to the spatial-difference (SD) method which processes spectra acquired on two adjacent areas, one encompassing the interface, the other one on a reference area nearby. A normalized version (NSD) has been introduced to extend its field of application from grain boundaries to heterophase interfaces in order to enhance their specific ELNES contribution

Grain boundary Cr segregation provides magnetic isolation between the crystals in sputtered Co thin films used for longitudinal magnetic recording. Magnetic isolation decouples individual grains in order to increase coercivity, lower media noise, and allow for higher recording densities. These sputtered Co alloys grow with a strong [1120] texture owing to a body centered cubic (BCC) Cr underlayer. The sputtered Cr seed-layer has a [001]BCC growth texture and the BCC {110}cr planes lattice match with the hexagonal (0002)cr planes. This produces the desired condition of the easy axis of magnetization, [0001]co, in the plane of the magnetic layer. However, since two orthogonal orientations of the {110}cr planes exist parallel to the [001]cr growth direction, it is possible to nucleate two hexagonal Co grains with orthogonal variants of the c-axis that will impinge to form a 90° grain boundary. The well defined crystallographic texture of the Co-alloy thin film with random angle and 90° grain boundaries presents an opportunity to study the effects of boundary character on Cr segregation in hexagonal crystals.

Temper-embrittlement in low-alloy steel has been attributed to the segregation of elements to the prior austenite grain boundary (PAGB). Auger electron spectroscopy (AES) has shown that the embrittlement of the PAGB is due to phosphorus segregation. The problem with this type of analyses is that only the PAGBs that fracture can be analyzed. More recent X-ray studies in a field emission gun scanning transmission microscope have shown that molybdenum, nickel, chromium manganese and phosphorus all segregate to PAGBs, and that these elements can be classified into two segregation groups; those PAGBs that contain nickel, manganese and phosphorus and those PAGBs that contain molybdenum, chromium and phosphorus. AES cannot determine conclusively, which segregation group is responsible for the embrittled PAGBs. The measurement of how embrittled the low-alloy steel has become, requires the knowledge, which PAGBs are embrittled and the percentage of these boundaries within a given volume.

In recent years, electron energy loss spectroscopy has provided high spatial resolution information on the elemental composition of interfacial reactions. with the use of field emission guns in conventional TEM or STEM, additional information on the chemical, structural and electronic state at grain boundaries and interfaces can also be obtained (e.g. Ref. 1 and 2). in the study of new types of materials for semiconductor applications such as high-K dielectrics, high spatial resolution information of this type is of critical importance as the dimensions of the structures are reduced and the ultimate electronic properties become size dependent. This paper discusses the analysis of near edge structure information obtained at the Si-Gd2O3 interface, a system with great potential for the next generation of gate material.

The preparation of the thin films and the sample preparation details are given elsewhere3 and only a summary of the characterization method and results are described here.

Ni, Pd and Cr thin films were grown on (100)SrTiO3 surfaces by molecular beam epitaxy at substrate temperatures of TNJ, pd=650°C and Tcr =150°C. Electron energy-loss spectroscopy (EELS) and high resolution transmission electron microscopy (HRTEM) were applied to investigate the local electronic structure and the atomic structure of the interfaces, respectively. Analytical microscopy was carried out with a parallel energy-loss spectrometer (PEELS766) attached to a dedicated scanning transmission electron microscope (STEM) operated at 100keV, which has a point resolution of 0.22 nm. HRTEM studies were performed on a Jeol JEM ARM 1250 operated at 1250keV (0.12 nm point resolution). Conventional TEM and HRTEM experiments showed epitaxial orientation relationships between the thin metal films and the substrate for each interface.

The electronic structure of the interfaces in terms of the site- and symmetry projected density of states (PDOS) above the Fermi-level can be extracted from the electron energy-loss near-edge structures (ELNES).