Search

A study is presented of low load indentations made into GaAs/AlAs superlattices and Si/SixGe1−x strained layer superlattices. We have studied the indentation sites using high resolution scanning electron microscopy (SEM) which allows us to distinguish between the individual layers due to the differences in the secondary electron (SE) yields of the materials. Transverse cleavage through the indentation sites reveals the pattern of the deformed multilayers, showing in particular a region of compressive strains of up to 30%.

Many of the topics in valence spectroscopy originally pioneered with TEM equipment have now been taken over by STEM. Notable exceptions are the studies by J. Silcox and his colleagues of plasmon dispersion and of the guided modes in thin slabs where retardation and other relativistic effects arise. The growing use of STEM valence EELS on increasingly complex nanostructures and the development of dielectric excitation theory for quantitative interpretation of the results were recently reviewed. The spatial resolution approaches the atomic scale, exceeding by orders of magnitude the capability of near field spectroscopy. By using the cathodoluminescence (CDL) signal, optical precision in STEM energy loss measurement can also be achieved, particularly in the difficult region below 5 eV or even 1 eV. High spatial resolution can be preserved provided that an intermediate stage of carrier diffusion is not involved in the radiative decay of excitations.Figure 1 shows STEM CDL images of an Ag particle obtained from the dipolel=1 and quadrupolel=2 surface plasmon excitations

The bulk loss function Im(-l/ε (ω)), a well established tool for the interpretation of valence loss spectra, is being progressively adapted to the wide variety of inhomogeneous samples of interest to the electron microscopist. Proportionality between n, the local valence electron density, and ε-1 (Sellmeyer's equation) has sometimes been assumed but may not be valid even in homogeneous samples. Figs. 1 and 2 show the experimentally measured bulk loss functions for three pure silicates of different specific gravity ρ - quartz (ρ = 2.66), coesite (ρ = 2.93) and a zeolite (ρ = 1.79). Clearly, despite the substantial differences in density, the shift of the prominent loss peak is very small and far less than that predicted by scaling e for quartz with Sellmeyer's equation or even the somewhat smaller shift given by the Clausius-Mossotti (CM) relation which assumes proportionality between n (or ρ in this case) and (ε - 1)/(ε + 2). Both theories overestimate the rise in the peak height for coesite and underestimate the increase at high energies.

Secondary electron signals with 1 nm spatial resolution and 1 eV energy resolution obtained in a STEM are presented and discussed. Reflection SE images of oxidised Cu show oxide islands and details of their interaction with surface steps. Stopping power theory identifies inner shell excitations and well localised valence excitations rather than plasmon rather than plasmon excitations as the dominant SE generation mechanisms and accounts for the spatial resolution obtained.

The growth and structure of defects in electron-irradiated Ge has been followed at room temperature and at higher temperatures between 200° and 300°C. At room temperature, small interstitial loops of approximately edge character on {311} planes are observed and eventually grow into dendritic structures. At the higher temperatures, rods elongated along the 〈110〉 directions form first but then tend to widen into plates of interstitial type on {311} planes with a displacement vector R = 0.04〈611〉. It is possible that all of these defects are of the same type and correspond to aggregates of Ge interstitials, but the possibility of impurity effects cannot be ruled out. The defects tend to transform into perfect dislocation loops on annealing.

The application is described of various bright field and dark field imaging methods to the study of catalyst and other small particles in the presence of scattering effects from the support. A major advance has been possible with the use of annular dark field, Z contrast and microdiffraction techniques in the STEM. Microanalysis of individual particles can now be carried out almost routinely. Methods of imaging surface structure are under development and it may also be possible to extract useful information about valence electron properties.

Careful analysis of high resolution electron micrographs of amorphous materials, making use of a simple optical simulation technique, provides a means for discriminating between ‘meaningful’ and ‘meaningless’ images, and for detecting ordered regions as small as 15 A in the ‘meaningful’ ones. Such regions, comprising only about 100 atoms, have been directly observed in ‘amorphous’ carbon, and it is possible that they exist in a range of other ‘amorphous’ materials.

The importance of the diffuse background in weak-beam electron micrographs is discussed. In typical cases its magnitude, as measured by a displaced aperture technique, is about 0-005 relative to the incident beam intensity and roughly equal to the ’coherent’ (Bragg) background. The role of various diffuse scattering mechanisms is estimated and it is concluded that the magnitude of the diffuse background cannot be adequately explained by thermal diffuse scattering or by disorder scattering from, for example, surface contaminant films. A major contribution arises from scattering due to electronic excitation. The quality of weak-beam images could perhaps therefore be improved by energy filtering.

The ability of REM imaging to detect surface and sub-surface features is illustrated with results obtained in a STEM on fractured GaAs surfaces and oxidised Cu surfaces. A qualitative interpretation of some features of these images can be given on simple ideas of Bragg reflection and phase contrast at titled surfaces. However, analysis of energy loss spectra, supported by Bloch wave calculations, suggests an effective penetration depth of about 3 nm even under surface resonance conditions. It appears that a more quantitative interpretation of the image data in terms of local surface structure and composition should be based on a theory for the elastic and inelastic scattering of evanescent waves.

The energetics of multiply twinned particles (MTPs) are investigated using elasticity theory. This allows the homogeneous strain models to be critically compared with the disclination model for the strains in decahedral particles and with a new model for the strains in icosahedral particles based on inhomogeneous elasticity. The overall energy balance between MTPs and single crystals is then evaluated, including the significant cost of elastially distorting the surface and using two extreme models of the faceting. The results of this analysis indicate that icosahedral MTPs will be more stable than single crystals for small sizes only for strong faceting conditions, decahedral MTPs being true intermediaries between the two. Experimentally observed stress-relief mechanisms provide indirect evidence for the inhomogeneous strain models.

Kinematical electron microscope images to be expected from the Polk continuous random-network model and from a polycrystalline cluster have been computed. Dark-field images of both models show qualitatively similar speckle patterns. Fringes are observed in bright-field images of thin polycrystals (t≲ 40A), but they disappear in thicker samples and are not observed in any of the Polk-model images. The experimental observation of fringes therefore cannot be explained by the current theory, possibly because it does not take account of the angular spread of illumination. Except perhaps in very thin specimens, it seems impossible at present to draw any reliable conclusions about the local atomic arrangement in amorphous materials by simple examination of electron micrographs.

Intensity profiles corresponding to transmission electron microscope images of planar defects in crystals have been calculated for a range of values of g. R, where g and R have their usual meaning. The results show that when two-beam theory is used, the defects will be effectively invisible if R.g differs from an integer by less than 0·02. When systematic many-beam theory is used, this criterion may be modified. In addition, many-beam profiles have been calculated for a number of specific cases relating to a stacking fault in silicon, and the results compared with experiment.

The profiles of screw dislocations and stacking faults in thin crystals, when imaged in the back-scattered mode of the scanning electron microscope, have been calculated using dynamical diffraction theory. These results may be extended to thick crystals by the inclusion of a suitable background term to allow for multiple scattering effects. The theoretical contrast of dislocation images in thick crystals is similar to that in calculated channelling patterns. The influence of probe size and beam divergence on the image is discussed, and it is apparent that the use of higher accelerating voltages offers advantages not only because of increased gun brightness but also because of more favourable diffraction and aberration conditions. It is concluded that to detect dislocations in a thick crystal using scanning microscopes with conventional electron guns and detectors the scanning times required are unreasonably long. However, in view of the improved contrast available from more efficient detectors, i...

The image contrast from amorphous materials studied by dark field or by interference electron microscopy using tilted illumination is quantitatively discussed in terms of the microcrystallite model and of the random network model using a network model for SiO2. The experimental observations of bright spots of diameter > 10 A. in dark field and of lattice fringes in interference micrographs are well explained by the microcrystallite model but cannot be accounted for by the random network model. The random network gives rise to a number of interesting diffraction effects, however, in particular to very small rather intense bright spots of diameter

It has been known for several years that anomalous structures can occur in the early stages of particulate growth. Ino1 and Ino and Ogawa2 for gold on rocksalt and Allpress and Sanders3 for nickel on mica found small polycrystalline nuclei which they interpreted as multiply-twinned particles (MTPs). These particles can be considered as 5 or 20 f.c.c. tetrahedra joined by twin boundaries to give decahedra or icosahedra (Fig. 1). As the tetrahedra do not assemble to a completely space filling structure, either elastic strains or grain boundaries are required to complete the particles. More recently, MTPs have been observed in the early stages of growth for all f.c.c. metals in certain conditions. Gillet4 has recently reviewed the field whilst Hayashi et al.5 give extensive results for f.c.c. metals in argon smokes. The possible significance of MTPs in heterogeneous catalysts was first suggested by Allpress and Sanders6 and investigated by Avery and Sanders7. They found, however, that at most 2% of particles 100–350 A in size for Ni, Au, Pt and Pd on γ-alumina were multiply twinned. We report here the presence of considerable numbers of multiply-twinned particles in various heterogeneous metal catalysts, in particular silver on α-alumina. The particles can have quite large dimensions up to ∼2,000 A. Attention is drawn to the possible significance of some special sites on these particles which have no counterpart on single crystal surfaces, and to the unusual distribution of surface faces.

Although the very first work on the dynamical theory of electron diffraction (Bethe[1]), dealt with RHEED geometry (the so-called Bragg case), progress in developing a completely satisfactory theory has been rather slow. Recent spectacular results e.g. REM images of surface structures, Osakabe et al[2,3], and RHEED oscillations in MBE, Neave et al[4], together with the increasing number of studies with clean surfaces, may however spur the theoreticians. In somewhat similar circumstances several decades ago, rapid advances took place in transmission electron diffraction (the Laue case) largely as a result of the enormous growth in transmission microscopy, Hirsch et al[5]. Real space, diffraction contrast images and later, high resolution structure images, of both perfect and imperfect crystals can now be interpreted in enormous and quantitative detail almost as a matter or routine.

The Beginnings of Electron Microscopy. Edited by Peter W. Hawkes. Academic: 1985. Pp.633. 88, 67.