Atom probe field ion microscopy is the most suitable technique for determining local chemical composition changes during nanocrystallization processes of amorphous alloys. In this talk, we report atom probe field ion microscopy (APFIM) and high resolution electron microscopy (HREM) studies on nanocrystallization processes in Fe-Si-B-Nb-Cu soft magnet and Fe-Nd-B-Co-Ga exchange spring magnet.

Fe or Co based alloys with nanocrystalline microstructure show excellent permeability because the net magnetocrystalline anisotropy is significantly reduced when the grain size becomes smaller than the ferromagnetic exchange length. Fe-Si-B-Nb-Cu alloy is the pioneering nanocrystalline soft magnetic material invented by Yoshizawa et al. in 1988 [1]. Our previous works [2,3] reported evidence for clustering of Cu prior to the onset of the crystallization reaction. However, in the previous studies, it was not confirmed that these Cu clusters work as heterogeneous nucleation sites for a-Fe primary crystals.

The ability to locate an individual atom on a surface, remove it in a controlled fashion, and determine its chemical identity makes the atom-probe field-ion microscope an extremely powerful tool for the analysis of solid surfaces. By itself, the field ion microscope has contributed significantly to our understanding of surface atomic structure, single-atom surface diffusion, and the detailed interactions that occur between atoms and defects on surfaces.1 When used in combination with the atom-probe mass spectrometer there have been several additional areas within the traditional definition of "surface science" where the chemical identification capability of the atom probe has led to new insights. In this paper these applications are reviewed focusing on two specific areas: surface segregation in intermetallic alloys and chemical reactions on metal surfaces.

The equilibrium distribution of component species in the near surface region of solid solution alloy may be different from the distribution in the bulk.

Control and optimization of materials’ properties will increasingly depend on the ability to tailor both the structure and chemical composition of the materials at the nanoscale level (100 nm and less). One specific class of “nanostructured” materials is the multilayer film (MLF) structure, which is formed by alternate deposition of two or more different elements or compounds and has a wide range of applications in, for example, the field of data recording and storage. In order to understand MLF structures well enough to improve and optimize their properties, characterization at the atomic level with regard to both structure and chemistry is essential. Interface quality is one of the most important parameters due to the possibility of intermixing between layers. It is thus desirable to obtain information about the interface morphology in these materials at the highest possible resolution.

The distribution of small amounts of alloying elements or impurities to interfaces can have a major influence on the properties of a material and the atom probe field ion microscope technique is well suited for the study of such segregation effects, where structural and compositional changes often take place over regions only a few nanometers wide.

Many features of the atom probe technique make it very useful in studies of interfaces:

1. i) high spatial resolution, about lnm laterally and 0.2 nm in depth

2. ii) equal sensitivity for all elements, including the lightest

3. iii) no specimen contamination, surface artifacts from the specimen preparation are removed by field evaporation,

4. iv) low temperature operation, negligible diffusion during analysis

A ncessary prerequisite for any systematic analysis is the ability to prepare suitable specimens. The volume accessible for analysis by APFIM is orders of magnitude smaller than a typical grain size, so a specimen preparation technique was developed that could bring features of interest sufficiently close to the apex (see Fig. la).

Trace additions of Ag and Mg in Al-Cu and Al-Li-Cu alloys change the precipitation processes drastically. In Al-Cu-Mg-Ag alloy, plate-like Ω phase precipitates on the {111} matrix planes. Similarly, additions of Ag and Mg to Al-Li-Cu alloy cause uniform dispersion of the T1 precipitate on the {111} matrix planes. Ω and T1 phases are structurally and morphologically similar, and a similar mechanism is expected to work in both alloy systems to enhance uniform dispersion of the plate-like precipitates on the {111} planes. A previous atom probe work [1] reported that Ag and Mg are segregated at the αΩ. interfaces in an Al-Cu-Mg-Ag alloy. In an Al-Li-Cu-Mg-Ag alloy, Ag and Mg atoms were found to be incorporated with the T, phase [2]; however, because of the thinness of the T1, plate, the exact location of these atoms with respect to the T1, phase could not be determined sucessfully by a conventional atom probe.

Segregation of boron, carbon and nitrogen to grain boundaries in a nickel based model alloy has been investigated using atom probe field ion microscopy (APFIM). The material corresponds to a commercial alloy (Inconel 600), but contains lower levels of alloy additives and impurities. The major composition was Ni-16Cr-10Fe (wt.%). The alloy was solution annealed at 950°C for 10 min, which resulted in a grain size of 20 μ. Subsequently heat treatments for 1 h at temperatures of 550°C, 600°C and 700°C were applied. TEM investigation showed that the heat treatment at 700°C resulted in precipitation of intergranular chromium-rich carbides. The other temperatures were obviously too low and the aging times too short to cause precipitation, since carbides were not observed in the materials heat treated at 550°C and 600°C.

As the grain size was about 100 times larger than the accessible depth of APFIM analysis (≃200 nm), much care had to be taken in preparing samples containing a grain boundary close to the tip apex.

The last several decades have seen an explosion in the number of experimental techniques being developed to characterize materials. In particular, there has been exponential growth in the coupling of microscopic methods with conventional spectroscopic characterization tools. This has been due mainly to the renaissance in microscopy over the last quarter century and the advent of significant computational power “on a bench”.

Compositional imaging is just a new term for what was over a half century ago called “Chemical Microscopy”. Although then the term referred to optical microscopy, today the notion of compositional imaging refers to a method by which we can correlate chemical or elemental composition with a microscopic image feature.

Because many of these microcharacterization methods have developed from many different disciplines, the “local” language used often appears to be untranslatable from one technique to the next without considerable pain and difficulty.

Historically, electron energy-loss spectroscopy (EELS) in the electron microscope has evolved from work performed in biological as well as materials science and physics laboratories. The task of developing EELS compositional imaging for biological applications has been particularly challenging for two reasons. First, biological structures are radiation-sensitive and second, elements and compounds of interest are distributed sparsely in a multi-component matrix. For example, the physiologically important element, calcium, is typically present in cells at only 10-100 atomic parts per million.3,5 Even, when individual molecular components of a cell are isolated, elemental concentrations can still be low. Thus, double-stranded DNA contains only six phosphorus atoms per nanometer of its length; and protein molecules typically contain less than 0.5 per cent sulfur. This demand for high sensitivity in biological applications has provided an impetus to improve electron optics, to develop flexible image acquisition methods and to optimize spectral processing.

Secondary ion mass spectrometry (SIMS) is based upon the energetic ion bombardment of surfaces resulting in in the emission of sputtered particles, including both atomic and molecular ions. The use of mass spectrometric detection provides a highly versatile and sensitive tool for surface and thin film microanalysis. The scope of the technique includes a diversity of analysis modes including:

1. 1. Elemental Depth Profiling (dynamic SIMS),

2. 2. Laterally Resolved Imaging (ion microprobe or ion microscope analysis),

3. 3. Image Depth Profiling (combination of modes 1 and 2 providing 3-D images),

4. 4. Molecular Monolayer Analysis and Imaging (static SIMS),

5. 5. Sputtered Neutral Mass Spectrometry (post-ionization).

Much of the early work in dynamic SIMS centered on depth profiling and imaging techniques, with an emphasis on applications to electronic materials. SIMS has made extensive contributions to semiconductor materials science since the 1960's, including the development of new devices and processes, and in failure analysis.

Many recent applications in light microscopy involve the use of multiple fluorophores or the delineation of signals arising from spectrally distinct sources. In microspectroscopy, it is always desirable to illuminate fluorescently-labeled microscopic specimens with monochromatic light as the narrowest possible excitation wavelength range usually results in the highest emission signal-to-noise ratio. Generation of polychromatic light from an arc lamp and selection of the excitation wavelength by interference filters or monochrometers are the most common techniques for excitation microspectrofluorometry. Emission spectroscopy is usually done with filter wheels, monochrometers, or interferometers inserted between the microscope detection port and the detector. This presentation will be directed toward other, less frequently-used, approaches for spectral scanning of the specimen in the light microscope. Three topics will be considered: 1) the use of acousto-optical tunable filters and lasers for rapid, narrow-band, excitation wavelength selection; 2) the use of holographic notch filters for rejection of unwanted excitation laser light; 3) using liquid-crystal tunable filters for emission scanning.

Equipping a transmission electron microscope with an energy-filter offers extraordinary advantages for the characterization of both materials science and biological samples. Besides improvements for TEM imaging and electron diffraction like better contrast and resolution, elemental mapping using inner-shell ionizations has become the main application of EFTEM. Elemental maps with resolution down to 1 nm and elemental sensitivities down to a single monolayer have been reported. Using a Philips CM20 equipped with a Gatan Imaging Filter (GIF) for our experimental work, we acquired elemental maps with the three window method (A.E-1r background extrapolation from two pre-edge windows) and jump ratio images (division of post-edge image by a pre-edge image). One important application of EFTEM is the detection of secondary phases in materials e.g. precipitates and grain boundary phases. For example, fig.la shows the TEM-image of a 10%Cr steel with the secondary phases mostly invisible.

Differential interference contrast (DIC) light microscopy, particularly when coupled with digital image processing, is a powerful tool for the high-resolution microscopy of unstained, transparent biological specimens and can equally well be applied to semiconductor measurements. We show analytically, and with images of diatoms, plant cells and protoplasts, that switching the polarization of the incident light by 90 degrees, changes the image highlights found in conventional DIC images into shadows and vice versa (1). Using a ferroelectric liquid-crystal modulator, this switching can be done at frame rates, synchronized to the camera. By subtracting alternate frames, a stream of difference DIC images is created. We call this technique Pol Mod DIC. Subtraction of alternate images is carried out efficiently by frame buffer operations and amounts to massively parallel synchronous detection. A similar method has been applied to confocal microscopy (2).

GaN and the related AlN semiconducting materials have recently attracted considerable attention because of their versatile applications for optoelectronics. Large stresses are present in the GaN/AlN thin-film heterostructure and can exceed GPa's. They originate from a large lattice mismatch between the substrate(sapphire)/GaN (14%) and the AlN/GaN interface (-2.7%) in quantum well structures. The strain is expected to induce local piezoelectric fields in these polar materials. It is essential for a further development of GaN based thin films to fully understand the formation and the local strength of these fields. Previous studies show that there are unusual large fluctuations of the patterns in lattice images present across GaN/AlxGa1-xN and GaN/InyGa1-yN quantum wells. They are attributed to compositional fluctuations and have been studied by quantitative high resolution electron microscopy. Differences between the strain profiles and the electrostatic (scattering) potential profiles were observed.

Data from digital electron probe x-ray microanalysis (EPXMA) images can be used to generate high resolution profiles of Ca binding within sarcomeres of vertebrate striated muscle. While ratios of elemental peak to bremstrahlung at each point (Hall method) have been used in many studies to allow comparison of data taken from different samples, this procedure has limitations for our application. A different normalization procedure is described here which provides a means of assessing variation among elemental and bremstrahlung profiles obtained from different images and samples.

Freeze dried cryosections of chemically skinned frog or rabbit skeletal muscle were prepared as previously described. Digital EPXMA images were collected for 24 to 36 hours using a Zeiss EM910 STEM and an Oxford ExL2 microanalysis system. Calcium and bremstrahlung counts were summed within one pixel wide masks placed at successive positions along each half sarcomere, as previously described, using an automated routine in IPLAB on a Power Mac 7100.

Since the first demonstration by Cosslett and Duncumb, X-ray mapping by an electron probe microanalyzer (EPMA) has become a most popular approach in microanalysis because elemental distributions of constituents in a bulk sample can be displayed visually. The major disadvantage of EPMA mapping is poor spatial resolution (∼ 1 μm). The spatial resolution of X-ray microanalysis can be improved to a few nanometers using electron transparent thin-specimens in the analytical electron microscope (AEM). However, X-ray count rates from thin specimens are strictly limited because of the improved spatial resolution (i.e. smaller interaction volume) and the poor collection efficiency of X-ray. To obtain reasonable counts for accurate quantification in the AEM, extraordinarily long mapping times are required. Therefore, quantitative X-ray mapping is rarely attempted in the AEM. However, these limitations can be overcome by use of intermediate-volt age instruments combined with field-emission guns to increase the beam current-density, careful stage design to maximize the X-ray collection efficiency and the peak-to-background ratio, and ultrahigh vacuum system to reduce contamination.

Here we demonstrate sub- 2Å resolution, in compositionally sensitive Z-contrast imaging, for a 200 kV field-emission transmission electron microscope (FEGTEM). This method has shown great promise for determining the atomic structure of interfaces in such materials as ceramics, semiconductors and high-Tc superconductors.

The Z-contrast technique involves focusing an electron probe at a specimen surface (which is oriented with a zone-axis along the beam direction). The transmitted signal, at an annular detector, is displayed on a VDU, rastered synchronously with the probe. Coherent, contrast reversal effects are suppressed in this geometry; images approximate to a convolution of the probe intensity profile with the square of the projected specimen potential. The Z2 dependence means that images can, with care, be interpreted as chemically sensitive maps of the projected structure. Furthermore, electron energy loss spectroscopy (EELS) can be performed simultaneously with Z-contrast imaging to yield additional compositional information.

Experiments and theoretical work on spinodal decomposition of thin films of binary polymer blends indicate large changes in phase behavior and dynamics relative to the bulk behavior due to the complex interactions of the polymer-polymer, polymer-air, and polymer-substrate interfaces. Currently used techniques are able to sense growing domain sizes but are not always able to sense local two and three dimensional variations within domains. We are using a new characterization approach to these thin films utilizing the Scanning Transmission X-ray Microscope (STXM) at beamline XIA at the National Synchrotron Light Source (NSLS)2. We extract component thickness maps of each polymer in a sample region. This allows for quantitative investigation of the dynamics of polymer phase separation and should allow the determination of important parameters affecting the domain formation in these thin films.

Rubber toughening was first developed about seventy years ago as a method of increasing the impact resistance of brittle glassy polymers. Since that time, a wide variety of techniques have been developed to produce discrete dispersions of rubber within a glassy polymer matrix. We are exploring a new route to rubber-toughen polymer through the non-equilibrium process of mechanical alloying (high-energy ball milling). Here we have blended poly(methyl methacrylate) (PMMA) with polyisoprene (PI) (see the insets in Fig. 1 for chemical structures) and have characterized the resultant blends with Scanning Transmission X-ray Microscopy (STXM) in order to obtain chemical, as well as morphological, information about the blends.

The STXM technique utilizes tuneable soft x-rays that are focused with a diffraction zoneplate to a microprobe of 45 nm full width at half maximum. Thin sections of the sample are placed at the microprobe and the transmitted photon intensity is measured.

The intimate blending of two incompatible polymeric materials constitutes one of the most important fields of study in polymer science. A vast arsenal of techniques, ranging from copolymerization to melt-blending, has been assembled to mix intrinsically immiscible polymers and optimize the physical properties of the resultant system. For some systems, however, even established techniques cannot produce useable blends. To overcome this limitation, non-equilibrium mechanical alloying (high-energy ball milling of two or more dissimilar powders) has been employed to produce blends of poly(ethylene terephthalate) (PET) and Vectra (73/27 oxybenzoate/oxynaphthoate). Characterization of these blends by TEM is confounded by the absence of a preferential staining agent, in which case Scanning Transmission X-ray Microscopy (STXM) provides a valuable alternative in examining the morphology and chemistry of these blends.

The x-ray microscope utilizes a diffraction element called a zone plate to focus soft x-rays into a microprobe. A thin section of sample is placed in this microprobe, and the transmitted photon intensity measured.

An O-N-O layer within a DRAM is designed to be an insulator between a single-crystal silicon substrate and a poly-crystalline silicon gate. The nominal structure of the layer is SiO2-Si3N4-SiO2. A small amount of phosphorus is doped in the poly-crystalline silicon gate to improve the electrical characteristics of the device. Transmission Electron Microscopy (TEM) was first used to view the micro structure in the layer. A bright field image gave clear contrast in the O-N-O layer, however, TEM images do not provide direct information about the elemental distribution. In this study X-ray elemental mapping was thus used to show this distribution.

The specimen was cut from a 16M-DRAM device and thinned using the cross-section ion milling method. The O-N-O layer was examined using a JEM-201 OF FE-(S)TEM equipped with a JEOL UTW-EDS detector. Fig. 1 shows a high resolution TEM image of the O-N-O layer.