FEI is currently finishing the construction of a monochromized 200kV (S)TEM which aims at 0.1 eV resolution in Electron Energy Loss Spectroscopy (EELS). This requires improvement of the 200kV supply, improvement of the EELS spectrometer, and the incorporation of an electron beam monochromator. This microscope will be delivered to the National Centre for High Resolution Electron Microscope at the University of Delft.

The present 200kV supply contributes 0.2 eV to the energy resolution because of its instabilities, drift and high frequency ripple. in order to reduce this, mechanical as well as electrical damping elements are being added, and sources of cross-talk between the high frequency generator and the high tension are eliminated. in this paper, we present preliminary results obtained with a standard 200kV supply; we expect to present results of the improved 200kV supply at the conference.

The resolution of a present-day standard 200kV spectrometer is now typically 0.4 eV. in order to improve on this, GATAN has developed a HR-EELS spectrometer for the monochromized TEM.

The Gatan Enfina EELS spectrometer, featuring a new generation of detection system, was recently developed to upgrade or replace Gatan PEELS / DigiPEELS systems. Major advantages of the system include significant improvements to the detector point spread function, readout speed, sensitivity, readout noise, and increased number of detection channels. This design was until now not available for VG STEM systems because compatible detectors must not significantly degrade the UHV column vacuum and must survive column baking. in this abstract we report on a new detector system with comparable specifications to the standard Enfina system but designed for new or for upgrading existing Gatan UHV PEELS / DigiPEELS systems on the VG STEM.

The standard Enfina design has the scintillator, fibers, CCD, CCD socket board, flex-cables, thermo electric cooler and coupling grease in the shared vacuum of the EELS spectrometer and the electron microscope.

The ability to obtain elemental maps processed by using inelastically scattered electrons in a transmission electron microscope (TEM) or a scanning transmission electron microscope (STEM) is extremely useful in the analysis of materials, and semiconductor devices such as ULSI’s and GMR heads. Electron energy loss spectra (EELS) also give useful information not only to identify unknown materials but also to study chemical bonding states of the objective atoms. Hitachi developed an elemental mapping system, consisting of a STEM (Hitachi, HD- 2000) equipped with a two-window energy filter (Hitachi, ELV-2000), and performed realtime conventional jump-ratio images with nanometer resolution by in-situ calculation of energy-filtered signals [1]. Additional function of acquiring EELS along any lines on specimen has been developed in this system to investigate the energy loss near edge structure (ELNES).

Figure 1 shows a schematic figure of the two-window energy filter, consisting of two quadrupole lenses for focusing and zooming spectra, respectively, a magnetic prism spectrometer, a deflection coil and two kinds of electron beam detectors.

Energy filtering transmission electron microscopy (EFTEM) has matured into an important nanoanalytical technique in both materials and life sciences. One of the technique′s main advantages stems from the possibility to quickly form images that contain two-dimensional elemental information for most elements (Li to Pu) from relatively large specimen areas with nanometer resolution. Over the last years numerous and wide-ranging applications have demonstrated its almost unrivalled power for assessing typical materials science questions.

Both in-column or post-column filter microscopes have reached a high performance level, but future improvements are still desirable, in particular with respect to increased transmissivity, improved isochromaticity and higher detection efficiency. Current instruments allow to acquire elemental maps with 1-5 nm spatial resolution limited mainly by the aberrations of the microscope and by the quality of the specimen. As recently shown 0.4 nm resolution is feasible.

One major practical problem of EFTEM compositional mapping is the poor signal-to-noise ratio, which means that experimental parameters have to be chosen very carefully in order to optimize this ratio.

It is now becoming increasingly more commonplace to acquire three-dimensional EELS data sets, whether by parallel acquisition of a linear series of spectra to form an EELS spectrum-image, or by discretely sampling energy-loss space through a series of energy-filtered images to build an EFTEM image-spectrum. These data sets can contain large amounts of both spectral and spatial information, and, as a result, the distinction between the two acquisition modes is becoming increasingly blurred. Accordingly, approaches are required that facilitate both the techniques developed for EFTEM and EELS analyses, in order to obtain optimal information from the acquired data.

An example of such an approach is illustrated by means of characterisation of a Mn and Fe rich intermetallic particle in a cast aluminium alloy (Figs. a-f). A continuous EFTEM image-series was acquired from 0-800 eV loss, using a CM300 FEGTEM equipped with a 2k Gatan Imaging Filter.

The optimized performance of Co(CrTaPt) thin-film longitudinal magnetic recording media (for computer hard discs) depends critically on the grain size distribution and on intergranular Cr segregation that magnetically isolates the grains. These microstructural aspects of both model and commercial media have been extensively investigated at ∽1 nm resolution with quantitative elemental mapping by energy-filtered transmission electron microscopy (EFTEM) of Cr, Co, and O, but not Ta and Pt which unfortunately are not amenable to quantitative elemental mapping by EFTEM. Procedures for data acquisition and processing have been refined in order to provide robust methods for measuring intergranular compositions for statistically significant numbers of grains. Figure 1 shows results obtained with a Gatan imaging filter (GIF) interfaced to a LaB6 Philips CM30 for a carbon-capped 25-nm-thick film of Co80Cr16Ta4 grown on a Cr underlayer. Typical acquisition parameters have been described previously. to avoid the Cr underlayer, foil thicknesses t<0.2X are used, including the carbon overlayer (λ = inelastic mean free path ≌100 nm).

The term “quantitative“ means many things to many people. in imaging its value is often discounted to mean only that a more intense region can be said to have more of what is being measured than a less intense region. Others may be more stringent and require an approximately linear relationship between the measured property and image intensity. (This is most often the case in EFTEM and spectrum imaging) Some specimens contain internal references (such as elements with a known stoichiometric relationship) that are sufficient to get to a true concentration calibration for an image. However, in EELS imaging this still doesn’t allow us to get an understanding of our errors and rarely is it helpful in determining detection or quantification limits.

A competent high-school chemistry or physics student could tell us that a quantitative measurement includes both a measured value and an estimate of its error.

A software package has been developed that synthesizes entire energy-loss spectra for an arbitrary distribution of impurities embedded in an arbitrary matrix, for user specified sample parameters and acquisition conditions. The simulation of a complete spectrum facilitates the estimation of quantities such as the Minimum Detectable Mass (MDM) or the Minimum Detectable Number of atoms (MDN). This is addressed with particular attention to optimising the acquisition conditions to maximize the detectability and also to estimating the error in a typical EELS quantification. It is envisaged that this approach will have wide-ranging implications for the role of EELS in quantitative microanalysis.

Electron Energy Loss Spectroscopy (EELS) is widely recognized as a microanalytical tool capable of characterizing the elemental composition of materials with nm-level resolution and sensitivity approaching single atoms. Extracting quantitative information from EEL spectra is complicated by the fact that both the cross-sections and the pre-edge background used for inner-shell analysis vary widely.

Despite of the intensive studies, Colossal Magnetoresistance (CMR) phenomena occurring in manganese oxides is still not fully understood. Theoretical studies based on an ordered crystal phase such as the charge-ordering (CO) phase have shown some successes in reproducing experimental evidences. Recently it was, however, shown that such a CO phase included giant clusters which were as large as 100 nm. This indicates the importance of the nanometer-scale information about the electronic structure to understand the CO. La0.5Sr1.5MnO4 is one of the oxides under intense investigations. in the CO phase, the eg valence electrons were found to be ordered at Mn sites. The optical responses of the CO clusters, however, has not been studied. in this study we have attempted to obtain such information from CO clusters in La0.5Sr1.5MnO4 by angular-resolved electron-energy-loss-spectroscopy (EELS) using an in-column type Ω-spectrometer.

Last year we discussed the performance of a new parallel EELS spectrometer, the ENFINA™. The spectrometer optics have been considerably improved over past designs and the detector changed from a photodiode to a charge coupled device that was designed specifically for spectroscopy. in this work we present application examples that demonstrate some of the capabilities of the ENFINA™.

STEM spectrum imaging requires a spectrometer of high speed and sensitivity if useful data is to be collected in a short period of time. The ENFINA™ spectrometer achieves both of these through a radically improved point spread function, increased sensitivity, and fast detector readout. Useful spectrum images can be acquired in a few minutes. Figure 1 shows a dark-field STEM image of a typical semiconductor. A spectrum image containing 190 x 150 spectra was acquired from within the marked area.

The hexagonal boron nitride (h-BN) exhibits a phase transformation to the cubic one (c-BN), the latter of which is well known as the second hardest material so far, under high pressure. This transformation is accompanied not only by the geometrical change in the crystal structure, e.g. stacking sequence of the close-packed planes, but also change in the chemical bonding from sp2 to sp3 characters. Electron energy-loss spectroscopy (EELS) is a hopeful way to characterize the state of bonding around the phase boundaries etc., and will provide rich information on the transformation mechanism. However, the energy-loss near-edge structure (ELNES) is rather sensitive to the orientation of specimens, if the structure is anisotropic as in the case of h-BN. This is sometimes obstructive to evaluations of ELNES, whereas a detailed investigation on this peculiar effect is needed. in the present work, the angular dependence of ELNES in h-BN was extensively studied by both calculations with DV-Xα (Discrete Variational X α cluster) method and experiments, beside investigations of the c-BN.

Zirconium carbide belongs to the group of refractory compounds having NaCl structure with high chemical inertness and melting point, characteristics that have made these materials to play a prominent role as hard coatings and these properties are closely related to their electronic structure. The electronic configuration of ZrC is: (Xe) 5d2 6s2, one s electron being able to be promoted to a d orbital forming four equivalent sd3 hybrid orbitals, favoring covalent bonds maintaining, however, a percentage in ionic character of about 30%.

Some attention has been paid in the last two decades in studying the optical properties of transition metal carbides and nitrides by electron energy loss spectroscopy (EELS), being TiC and TiN the most extensively reported.

In this work the optical constants of commercial powder ZrC were obtained using a Gatan Parallel Detection Electron spectrometer (model 766) attached to the CM-200 transmission electron microscope (TEM).

EFTEM has become a very useful tool in characterizing material properties, because it allows studying local chemical and electronic properties of a specimen with nanometer level spatial resolution. There are two spectroscopic imaging methods have been established to obtain both the spatial and the spectra information, I (x, y, ΔE), which can be quantitatively evaluated. One technique, as known as spectrum-image method, is based on scanning the electron beam across the sample to acquire the EELS spectrum with two-dimensional information. Conversely, the other method retrieves the spectra in a specific position from a series of energy loss images, as known as imaging-spectrum method. The spectrum-image method can be achieved with better energy resolution, however, a long acquisition time. Compared to spectrum-image method, which acquires a series of spectrum with scanning probe, the advantage of imaging-spectrum method is the capability of simultaneously recording abundant information in a series of maps.

Energy filtered (EF) imaging and energy loss fine structure analysis have now become routine tools to characterize very complex microstructures at high spatial resolution. When combined with energy dispersive x-ray spectroscopy microanalysis in a field emission gun transmission electron microscope (TEM) and with samples prepared by a large array of techniques such as FIB or microtoming, these analytical techniques can tackle real-life and more fundamental problems in materials science. This paper presents some examples of this work that highlight the complementarity of information obtained either from EF imaging, near-edge structures and EDS microanalysis. These examples are taken from work in the development of battery materials for automotive applications and in the study of ultrathin magnetic layers. Other examples of high spatial resolution energy loss fine structure at interfaces of high-K dielectric-Si in interface are shown in elsewhere in these proceedings.

It is not uncommon for an electrical failure in a microelectronic device to be traced back to an individual resistive contact, such as a W contact lined by TiN to the underlying metal silicide [1]. Identifying the root cause of the defective cell then requires the capability of extracting interfacial chemistry and microstructure with near atomic resolution, which is achieved by recording EELS and EDX data either along 1-d lines or from 2-d arrays. EELS data are characterized by intrinsically low signal-to-background ratios and plural inelastic-scattering effects, which presents a challenge to the reliability of the methods used to extract quantitative information from a spectrum image. Commercially available software packages by Emispec and Gatan currently provide routines for power-law or polynomial background fits and allow the net counts under peaks or edges to be determined from each pixel. in both cases each spectrum in the series is processed in the same manner.

Energy filtered transmission electron microscopy (EFTEM) has been developed in recent years to be a highly flexible tool for spatially resolved microanalysis. The basic technique has been extended to allow fully quantitative mapping of single elements, or a large range of elements using an image-spectroscopy approach . All these techniques are still limited by the nature of the TEM as a structure projector; any elemental map is only a 2D projection of what is in reality a 3D structure. A fully 3D approach to EFTEM analysis would offer additional insights into many materials problems and make possible meaningful analysis of certain systems for the first time. Three-dimensional electron microscopy (electron tomography) offers one path for the reconstruction of 3D structures from 2D projections. Although this has been well developed for solving the structure of biological macromolecules and viruses its use in the physical sciences is quite new.

Recent developments in detector technology [1] for EELS and Energy Filtered TEM has made possible to obtain large number of spectra and energy filtered images during very short exposure times. This in turn opens the exciting possibility of studying time dependent processes in the electron microscope, during exposure to the electron beam as well as the study of different radiation sensitive samples which are being degraded during lengthily data recording. This kind of data recording generates a large amount of data and manual data analysis should be avoided in order to be able to fully benefit from the improved sensitivity and increased speed of these new detectors. Thus a fast, real-time data analysis system is highly desirable.

A system for real-time data analysis (spectra classification) of data generated from such a detector has been simulated in a program based on the object oriented C++ framework ROOT [2][3].

As techniques for electron energy-loss spectroscopy (EELS) reach a higher degree of optimization, detection limits for analyzing biological structures are approaching those predicted by theory. in favorable specimens, single atom detection is predicted for elemental maps acquired by means of the scanning transmission electron microscope (STEM) equipped with a field emission source, paralleldetection EELS and a spectrum-imaging system. to obtain such results, the electron detector should have a detective quantum efficiency close to unity and a well behaved point-spread function; such design features are now available with a cooled charge-couple device (CCD) array. The energy-filtering transmission electron microscope (EFTEM) provides a complementary approach to mapping elements occurring at higher concentrations but distributed over larger regions of the specimen. Use of an optimized CCD detector in the EFTEM now enables accurate quantitation in addition to high analytical sensitivity, albeit not at the single atom level.

There has been considerable interest in recent years in simulating the complete electron energy loss (EELS) spectrum to help the analyst make suitable judgments on experimental parameters such as collection angles and acquisition times. The spectrum is characterised by edges that arise from the excitation of inner shell electrons to the first available empty state. The edge threshold is approximately given by the inner shell binding energy. Each edge sits on the background of the tails of all the edges from inner shell excitations of lower energy (see Figl). The background from plasmons decays very quickly with energy and even the broadest plasmon makes a negligible contribution to the intensity for losses above about 150eV. The visibility of an edge (and also the detectability of the corresponding element) can be related to the intensity of the edge compared to the background although it is more common to use the jump ratio, which is the ratio of the intensity in the region after the edge E+B to the intensity of the background in the edge region, B.

Electron energy loss spectroscopy (EELS) probes electronic excitations of a solid on the atomic scale. The widespread availability of first-principles calculations has lead to an explosion of theoretical calculations of EELS spectra. Agreement between theory and experiment is generally reported to be good at the typical energy resolutions in commercial microscopes of 0.7-1.3 eV. However a brief survey of the X-ray absorption literature suggests that the anticipated introduction of monochromators, along with improvements in energy stability, and spectrometer resolution will unmask many more effects that cannot be predicted as precisely as they can be measured.

The shape and binding energy of a core excitation is determined by both the ground state electronic structure (initial state effects) and the reponse to the excited electron-hole (final state effects) (Fig. 1). Errors in the initial state, such as the systematic errors in band gaps (and hence band offsets) are inherent in the local density approximation eigenvalues used to simulate EELS spectra.