Abstract

Cryogenic detectors with excellent energy resolution and low energy threshold far beyond the level of semiconducting detectors open a variety of new. applications in physics including search for Dark Matter in the universe [2], neutrino physics [3], and IR-, UV- and X-ray astrophysics [4, 9]. Interdisciplinary fields where cryogenic detectors have already shown promising results are the detection of biomolecules [5] and X-ray spectroscopy at synchrotron beam lines [6] and in scanning electron microscopes (SEMs) [7]. For both, astrophysical and analytical use, the development of high resolution microcalorimeters based on iridium/gold phase transition thermometers and aluminum tunnel junctions for use in a compact and universal detector system was initiated.

Our cryogenic microcalorimeters consist of an absorber, a temperature sensor and a weak coupling to a heat sink. An X-ray photon interacts with the absorber and raises its temperature. The sensor measures the temperature increase and the system then, mediated by the coupling, relaxes back to its operating temperature.

Introduction

Radiation curing of polymers, by UV light or E-beam, is of growing interest in the state-of-the-art industry. Radiation curing is solvent free and leads to important energy and money savings. Improved materials characteristics, better control of doses and more accurate application allow radiation curing to take a big part of the surface coating industry and, for the last ten years, the polymeric and microelectronics (lithography) industry. But behaviors of polymers under radiant source is often empirically characterized. In this abstract, it is proposed a way to obtain accurate energy distribution from Monte Carlo simulations. For bulk or thin polymeric films, it is therefore essential to simulate energy distribution, in order to foresee what happens in the sample.

Results

CASINO is a single scattering Monte Carlo Simulation of electroN trajectory in sO lid -particularly designed for low energy beam interaction in a bulk and thin foil.

In the semiconductor industry the importance of Computer Aided Materials Analysis (CAMA) is growing rapidly due to ever decreasing feature sizes on semiconductor wafers. Due to the fact that the excitation volume is often larger than the typical feature size on a processed state-of-the-art semiconductor wafer it was necessary to extend SeSAME, a Monte Carlo simulator for arbitrary 2- dimensional geometries, to be able to handle arbitrarily shaped 3-dimensional structures. This is achieved by using a so called multi-quadtree approach for the discretization of 3-dimensional geometries.

A multi-quadtree consists of several parallel arranged modified quadtrees. This means, an arbitrarily shaped 3-dimensional structure is divided into parallel slices of constant geometry in direction perpendicular to the quadtree plane. The number of such 2-dimensional slices and the thickness of these slices is solely limited by the available memory. Therefore, really 3-dimensional structures may be resolved in principal with any resolution.

Here we demonstrate sub- 1.5 Å resolution in compositionally sensitive high-angle annular dark-field (HAADF) (“Z-contrast”) imaging. For the first time this has been achieved on a 200 kV field-emission transmission electron microscope (FEGTEM), the JEOL JEM-2010F. With a Gatan imaging filter, this type of instrument is then capable of both analytical imaging and electron energy-loss spectroscopy at similar spatial resolution as in the 300 kV dedicated STEM.

The Z-contrast imaging technique has a spatial resolution given by the size of the electron probe. When used to image periodic specimens and their defects, the effective incoherent nature of the Z-contrast method leads to higher resolution for given lens Cs, higher sensitivity to atomic number and easier qualitative image interpretation than in HRTEM.In practice, the ability to form a small (atomic resolution) probe depends on the brightness of the electron source, and achieving low enough levels of mechanical and electrical instabilities that otherwise incoherently broaden the probe.

The trend of replacing trench and stack capacitors in a dynamic random access memory (DRAM) with a planar configuration has stimulated the development of high dielectric constant materials with reliably low leakage current and high dielectric breakdown strength. In this regard, high dielectric constant materials, such as PbZrxTiyO3(PZT), BaTiO3, SrTiO3, PbTiO3, and BaxSr1-xTiO3, have been extensively investigated as dielectrics in the last few decades. Of these, the sol id-solution quaternary BaxSr1-xTiO3, (BST) combines the high dielectric constant of BaTiO3, with the structural stability of SrTiO3, is one of the most promising materials for DRAM cells in very large-scale integrated circuits. BST shows a paraelectric phase for x<0.7 at room temperature, which provides additional features such as no aging or fatigue effects from ferroelectric domain switching. However, so far, there have been few studies of the interfaces between BST and the substrates, particularly at the atomic-resolution level.

MBE techniques allow the fabrication of exceptionally sharp compositional changes such as delta doped layers in semiconductors. Producing these spatially confined doped layers is critical to many innovative device designs. The spatial confinement of these delta doped structures can be less than the measurement resolution of the currently standard SIMS and RBS techniques. This allows only the upper limits on the layer width to be measured. These SIMS and RBS methods are also inadequate for the two dimensional information desired for future device design and development. More recently developed techniques such as Scanning Capacitance Microscopy and spreading resistance measurement give two dimensional information but have similar spatial resolution issues. The Z-contrast nature of Annular Dark Field (ADF) imaging with the complimentary technique of Electron Energy Loss Spectroscopy (EELS) in the Scanning Transmission Electron Microscope (STEM) shows promise for two dimensional dopant profiling with spatial resolution on the atomic scale.

Electron energy-loss spectroscopy (EELS) in the transmission electron microscope (TEM) analyzes the energy distribution of the probe electrons after they have lost energy within the sample. The resultant energy-losses are characteristic of elemental, chemical, and dielectric properties and are typically measured in one of two ways. Parallel-detection EELS spectrometers (PEELS) acquire large energy ranges of the energy-loss spectrum simultaneously for rapid acquisition of spectral data at a single area. In contrast, the energy filtering TEM (EFTEM) acquires only a single energy band at once, but does so for thousands or even millions of image pixels simultaneously.

Spectrum imaging (SI) involves acquisition of detailed spectroscopic data sufficient for rigorous analysis at each pixel in a digital image. (Fig. la) A STEM EELS spectrum image “data cube” can be acquired by stepping a focused electron probe to each pixel and filling the spectrum image one spectrum at a time.

The ability of modern TEMs to acquire a series of energy filtered images opens up new possibilities in energy loss compositional analysis. In particular, an electron spectroscopic imaging (ESI) series may be treated as a 2-D array of spectra whose resolution is dictated by the step size of the image series, as illustrated in Fig (a). This allows standard spectroscopic analysis techniques to be used on the extracted ‘image-spectra’, such as the removal of plural scattering by deconvolution. Examples of this are given in Fig (b) and (c), which show how Fourier-log and Fourier-ratio deconvolution can be used to recover the single scattering distribution (SSD) from both the low-loss and core-loss regions from a Cr specimen. A pure elemental sample is ideal for testing the validity of such analysis techniques for quantitative compositional mapping, and more details of this method will be published elsewhere. Further, for many simple metal systems, such as steels and alloys, and for simple semiconductors it is possible to model the plasmon contribution using a simple Drude-Lorentz model.

The dependence of diffraction contrast on electron energy loss is an important relationship that needs to be understood because of its potential effect on energy-filtering transmission electron microscope (EFTEM) images. Often when either a two-window jump-ratio image or a three-window elemental map is produced diffraction contrast is not totally eliminated and contributes to the intensity of the final EFTEM image. Background removal procedures often are unable to completely account for intensity changes due to dynamical effects (i.e., elastic scattering) that occur between images acquired at different energy losses, leaving artifacts in the final EFTEM image.

In this study, the relationship between diffraction contrast and electron energy loss was investigated by obtaining EFTEM images of a bend contour in aluminum in 100 eV increments from 0 to 1000 eV (Fig. 1). EFTEM images were acquired a JOEL 2010F FEG TEM with a Gatan imaging filter (GIF) at a microscope magnification of 8 kX using a 1 eV/pixel dispersion, 2X binning (512 x 512) and exposure times ranging from 0.25 s for 0 eV energy loss up to 132 sec for 1000 eV energy loss.

The small electron probe available in a scanning transmission electron microscope (STEM), when combined with X-ray energy dispersive spectroscopy (XEDS), enables the acquisition of 2D compositional maps with high spatial resolution and sensitivity. In addition to the spectroscopic information available, images sensitive to atomic number (so called Z-contrast images) can be acquired simultaneously using high-angle annular dark field imaging (HAADF). These capabilities make the STEM an ideal tool to study elemental distributions in catalysts containing 1-2 nm sized metal clusters within a support of low atomic number. The material examined in this work comprises of mesoporous silica in a honeycomb structure with channels of ˜3 nm diameter (designated MCM-41) in which nano-clusters of the bimetallic Pd6Ru6 have been introduced into the channels using metal-cluster carbonyls as precursors. Details of the specimen preparation are similar to that outlined by Shepard et al. for a Cu-Ru bimetallic catalyst.

Fig. 1 shows (a) a bright-field (BF) image, (b) a HAADF image and (c-f) X-ray elemental maps acquired on a VG 603 STEM, operated at 300 keV with probe size of ˜1.4 nm (FWTM) and probe current of ˜0.6 nA.

A common impurity in polycrystalline commercial alumina is Si since silica is often added as a sintering aid. There have been several studies on impurities in alumina, and substantial degree of grain boundary segregation has been found to occur. Alumina-metal composites have found several applications especially in the electronic industry and consequently systems such as alumina-copper have been studied exhaustively. However, the role of impurities in metal-alumina composites has not been fully explored. Studies of Pt-alumina composites has shown that silicon is present at the interface of vacuum, diffusion bonded samples. Our earlier studies have shown that Si indeed segregates to the triple junctions and grain boundaries in alumina and in alumina-copper composites, often Si segregates to the metal-ceramic interfaces as well. This investigation concentrates on the distribution of silicon in copper-alumina in comparison to aluminum-alumina composites.

The metal-ceramic composites were prepared by diffusion bonding 99.99% pure metal foils sandwiched in between commercially available sintered alumina (∼99.5% pure) substrates for several hours in vacuum.

In conventional x-ray mapping, energy regions are defined to straddle the N x-ray spectral peaks of interest. At each pixel position in the digital image, counts are recorded from each energy region and thus give rise to N element maps of intensity distributions over the field of view. Fig. 1 shows an example set of x-ray maps from a polished test sample. It is clear from the images that there are circular regions of differing composition but it is not immediately obvious which regions have the same composition. Furthermore, in a real sample, phases are unlikely to appear in such regular patterns so the correlations between maps are in general much more difficult to notice.

It is possible to choose two x-ray maps and generate a 2D scatter plot showing the intensity values in both maps for all pixels in the image.

A topic of growing importance is the nano-structuring of inorganic materials by exploiting interactions with self assembling polymers. One example are multiphase polymers.

Driven by the mixing enthalpy, most polymers phase separate below their critical solution temperature. The components of block copolymers are covalently linked, restricting separation to microscopic length scales. Depending on composition, block copolymers form different equilibrium morphologies representing the energetically most favorable balance between- interfacial area and chain conformation.

Mass thickness contrast of polymers is often weak and conventional TEM can thus fail to characterize the morphology unless selective staining methods exist. This is the case for the poly(isoprene)-b-poly(ethyleneoxide) diblock copolymer system (PI-b-PEO) used as a structure directing agent and to provide solubility (see below). Fig. 1 shows the micro-domain morphology of PI-b-PEO where the rubbery phase is selectively stained with OsO4 .

Hybrid materials are obtained by addition of an inorganic precursor to a solution of PI-b-PEO in a "mutual" solvent.

A strength of the modern electron microprobe is its ability to provide 2D compositional information about materials. These images give the ability to observe features that might otherwise pass unseen. Elements at the trace element level are generally ignored due to the high detection limits imposed by mapping under “standard” EMP conditions.

Trace element mapping requires beam regulation at high (e.g. 300 nA) to very high (e.g. 3 μA) faraday cup currents, reliable beam and stage control, and suitable samples and mounting media. The ability to operate at high accelerating voltage to maximize Pk2/Bkg is desirable (Robinson and Graham, 1992), although we have encountered column difficulties above 25-30 kV.

We are mapping trace and minor elements including Y, Sc, P, Cr, Mn, Ca, in garnets. Fig. 1 shows Y, Sc and Cr maps (Spear and Kohn, 1996; Kohn, Spear and Valley, 1997), and Fig. 2 Y and Sc maps (Cameron, unpub. data), produced with a Cameca SX51.

Electron energy loss spectroscopy is the only direct technique that can semi-quantitatively determine the nature of the bonding in carbon thin films. To quantify the sp2/sp3 bonding fraction, the spectrum taken from the film must be compared to that of a suitable known standard. The bonding fraction can be analysed by studying the K ionisation edge in the electron energy loss spectrum. A method for quantifying the sp2 bonding fraction in an amorphous carbon film has been described by Berger et al (1988), where the area of peak of the film is compared with that of graphite. The principle of quantifying the edge structure is to obtain a ratio of the two peak areas using the following formula, (1), where fπis the ratio between the two π* peaks, Iπ. is the integral of the transition, and ΔE is the integrated counts for the normalising energy window. The superscripts s and u denote the standard and unknown spectra respectively.

Elemental mapping of Co-Cr-X based magnetic recording media at resolutions approaching 1 nm by energy-filtered transmission electron microscopy (EFTEM) can provide quantitative measurements of intergranular Cr segregation for correlation with magnetic properties and materials processing. The thin-film media present many challenges for EFTEM methods, such as diffraction contrast and closelyspaced edges. The goal of this work was to provide robust methods for mapping quantitative compositions in such materials. Results presented here are for a model material of 60 nm of Co84Cr12Ta4 on a 75 nm Cr underlayer; both films were d.c. magnetron sputtered onto a NiP-plated Al substrate pre-heated to 250°C. Other compositions and thinner layers (∼30 nm) have also been studied. EFTEM was performed on back-thinned, plan-view specimens with a Gatan Imaging Filter (GIF) interfaced to a 300 kV LaB6 Philips CM30. Optimized acquisition conditions have been detailed elsewhere. Besides core-loss image series, zero-loss I0 (slit width Δ=10eV), low-loss Ik (Δ=30eV), and unfiltered IT images were recorded, and maps of t/λ. = ln(IT / I0), where t is specimen thickness and λ. is the total inelastic mean free path, were produced.

The development of high-efficiency, low-price photovoltaic devices relies on epitaxially grown layers of ternary chalcopyrites. The question addressed in this study is to which extent the near stoichimetric composition of CuInS2 grown on hydrogen terminated Silicon, Si (111), varies cross the sample area. The method usually employed [1] for such analysis is Argon-Rutherford Backscattering (RBS), whose spatial resolution, however, is limited to the order of millimeters. This limit can readily be overcome by electron microscopic and electron spectroscopic means.

For observation and recording of the energy loss spectra in the transmission electron microscope, samples were carefully prepared by mechanic polishing and ion-milling. The thinning began with polishing the Si substrate on ALLED diamond lapping films, down to 10 μm and a final ion-milling in a Gatan ion-thinning equipment from the Si-side (Ar+ 3 kV, the incident angle less than 7°).Spectra were taken by means of a Gatan image filter (GIF - 100) in a Philips CM200 FEG electron microscope operated at 200 kV.

Combustion synthesis has led to many advances in materials science, in part via the synthesis of powders consisting of particles of nanometer dimensions. Particle morphology is a key concern regarding the powders produced, but also of comparable importance is particle composition. Electron energy loss spectroscopy (EELS) and energy-filtering imaging (EFI) can be used to interrogate the gas-phase combustion synthesis environment for elemental particle composition information. Once established, this diagnostic approach can be used to address control of particle composition and other issues associated with particle formation and growth in flames. The evolution of the particle morphology in a laboratory scale combustion synthesis facility can be examined by passing TEM grids directly through the combustion synthesis flame at various heights above the burner surface, as shown in Fig. la. For the current work, SiOx particle samples are obtained from a SilL/^/FL/Ar flame using a rapid probe insertion technique.

Anionic modification of silicate-based glasses with carbon has recently attracted the attention of many research groups due to the beneficial effect on mechanical properties and both thermal and chemical stability. These carbon containing glasses, termed silicon oxycarbide glasses, can be synthesized from various organically modified silica gels or from preceramic polymers after a pyrolysis process in an inert atmosphere. The oxycarbide network is often mixed with a free carbon phase, and so the local environment around C atoms exhibits a complex distribution . In the ideal silicon oxycarbide phase, carbon atoms would be present only in carbidic units (forming only Si-C bonds). The carbon phases and the chemical bonding in the silicon oxycarbide have been investigated by high resolution electron microscopy (HREM), Raman spectroscopy, X-ray photoelectron spectroscopy (XPS) and electron paramagnetic resonance (EPR). Electron Energy Loss Spectroscopy (EELS) is an important analytical technique that can be used to further characterize the local chemistry and electronic structure of these glasses.

GaN-based materials hold promise for electronic and optical applications. Lacking a suitable lattice matched substrate, microstructural investigation by electron microscopy has become an important player in the development of these materials. Electron microscopy investigations vary from standard defect analysis, to polarity determination, to evaluation of electronic structure. To date, no evidence of electron beam damage has been reported.

For this experiment, the beam current from the Cornell University VG501 lOOkV STEM was approximately 0.3nA. During acquisition of multiple PEEL spectra of the N K-edge, the intensity of the core loss decreased and eventually disappeared due to electron beam induced damage. Damage is a function of dose. Thus observation of damage is dependent upon magnification and thickness, i.e., electron dose and fractional change in signal.