This issue is dedicated to Regents' Professor John Maxwell Cowley, FRS, in recognition of his lifelong contributions to electron microscopy, diffraction, and crystallography. This collection of 22 peer-reviewed articles is based on presentations made at the international workshop entitled “Recent Developments and Applications of Atomic Resolution Electron Microscopy and Spectroscopy—A Silver Jubilee.” The workshop was held in the historic Old Main building on the campus of Arizona State University in early January 2003 and was attended by about 150 people. It celebrated the 25th anniversary of the Center for High Resolution Electron Microscopy, which was founded by John Cowley, and also served as an occasion to honor him on his 80th birthday. Keynote speakers included Sumio Iijima, discoverer of the carbon nanotube; Mikail Roco, senior advisor on nanotechnology for the National Science Foundation; and John Cowley himself, whose lecture addressed innovative ideas for achieving sub-Ångstrom resolution without requiring aberration correction.

A small probe centered on an atomic column excites the bound and unbound states of the two-dimensional projected potential of the column. It has been argued that, even when several states are excited, only the 1s state is sufficiently localized to contribute a signal to the high-angle detector. This article shows that non-1s states do make a significant contribution for certain incident probe profiles. The contribution of the 1s state to the thermal diffuse scattering is calculated directly. Sub-Ångstrom probes formed by Cs-corrected lenses excite predominantly the 1s state and contributions from other states are not very large. For probes of lower resolution when non-1s states are important, the integrated electron intensity at the column provides a better estimate of image intensity.

Ultrahigh-resolution imaging may be achieved using modifications of the off-axis holography scheme in a scanning transmission electron microscopy (STEM) instrument equipped with one or more electrostatic biprisms in the illuminating system. The resolution is governed by the diameter of a reference beam, reduced by channeling through a line of atoms in an atomic-focuser crystal. Alternatively, the off-axis holography may be combined with the Rodenburg method in which a four-dimensional data set is obtained by recording a nanodiffraction pattern from each point of the specimen as the incident beams are scanned. An ultrahigh-resolution image is derived by computer processing to give a particular two-dimensional section of this data set. The large amount of data recording and data processing involved with this method may be avoided if the two-dimensional section is derived by recording the hologram while the four beams produced by two perpendicular biprisms are scanned in opposing directions across the specimen by varying the voltages on the biprisms. An equivalent scheme for conventional TEM is also possible. In each case, the complex transmission function of the specimen may be derived and resolutions of about 0.05 nm may be expected.

Crystal structure of nano-scale precipitates in age-hardening aluminum alloys is a challenge to crystallography. The utility of selected area electron diffraction intensities from embedded precipitates is limited by double scattering via matrix reflections. This effect can be signally reduced by the precession technique, which we have used to collect extensive intensity data from the semicoherent, metastable η′-precipitate in the Al-Zn-Mg alloy system. A structure model in the space group P-62c is proposed from high-resolution microscopy and electron diffraction intensities. The advantages of using the precession technique for quantitative electron diffraction is discussed.

The resolution achieved in low-dose electron microscopy of biological macromolecules is significantly worse than what can be obtained on the same microscopes with more robust specimens. When two-dimensional crystals are used, it is also apparent that the high-resolution image contrast is much less than what it could be if the images were perfect. Because specimen charging is one factor that might limit the contrast and resolution achieved with biological specimens, we have investigated the use of holey support films that have been coated with a metallic film before depositing specimens onto a thin carbon film that is suspended over the holes. Monolayer crystals of paraffin (C44H90) are used as a test specimen for this work because of the relative ease in imaging Bragg spacings at ∼0.4 nm resolution, the relative ease of measuring the contrast in these images, and the similar degree of radiation sensitivity of these crystals when compared to biological macromolecules. A metallic coating on the surrounding support film does, indeed, produce a significant improvement in the high-resolution contrast for a small fraction of the images. The majority of images show little obvious improvement, however, and even the coated area of the support film continues to show a significant amount of beam-induced movement under low-dose conditions. The fact that the contrast in the best images can be as much as 25%–35% of what it would be in a perfect image is nevertheless encouraging, demonstrating that it should be possible, in principle, to achieve the same performance for every image. Routine data collection of this quality would make it possible to determine the structure of large, macromolecular complexes without the need to grow crystals of these difficult specimen materials.

The key contribution of electron microscopy methods to condensed matter spectroscopy is undoubtedly spatial resolution. So far this has mainly been manifest through electron energy loss spectroscopy in the 1-eV to 10-keV energy range and has not seriously challenged the dominance of optical, X-ray, and neutron spectroscopy methods over most of the vast field at lower energies. At frequencies up to a few megahertz, corresponding to energies of a few nanoelectron volts and below, direct excitation by pulsed electron beams or electric fields has proved effective. Prospects are discussed for extending spatially resolved spectroscopy to the intermediate energy region, mainly by combining the advantages of electrons with those of photons.

The first paper on the FFT multislice method was published in 1977, a quarter of a century ago. The formula was extended in 1982 to include a large tilt of an incident beam relative to the specimen surface. Since then, with advances of computing power, the FFT multislice method has been successfully applied to coherent CBED and HAADF-STEM simulations. However, because the multislice formula is built on some physical approximations and approximations in numerical procedure, there seem to be controversial conclusions in the literature on the multislice method. In this report, the physical implication of the multislice method is reviewed based on the formula for the tilted illumination. Then, some results on the coherent CBED and the HAADF-STEM simulations are presented.

Along with rapidly developing nanotechnology, new types of analytical instruments and techniques are needed. Here we report an alternative procedure for electrical measurements on semiconductor nanowhiskers, allowing precise selection and visual control at close to atomic resolution. We use a combination of two powerful microscope techniques, scanning tunneling microscopy (STM) and simultaneous viewing in a transmission electron microscope (TEM). The STM is mounted in the sample holder for the TEM. We describe here a method for creating an ohmic contact between the STM tip and the nanowhisker. We examine three different types of STM tips and present a technique for cleaning the STM tip in situ. Measurements on 1-μm-tall and 40-nm-thick epitaxially grown InAs nanowhiskers show an ohmic contact and a resistance of down to 7 kΩ.

A study of screw dislocations in hydride-vapor-phase-epitaxy (HVPE) template and molecular-beam-epitaxy (MBE) overlayers was performed using transmission electron microscopy (TEM) in plan view and in cross section. It was observed that screw dislocations in the HVPE layers were decorated by small voids arranged along the screw axis. However, no voids were observed along screw dislocations in MBE overlayers. This was true both for MBE samples grown under Ga-lean and Ga-rich conditions. Dislocation core structures have been studied in these samples in the plan-view configuration. These experiments were supported by image simulation using the most recent models. A direct reconstruction of the phase and amplitude of the scattered electron wave from a focal series of high-resolution images was applied. It was shown that the core structures of screw dislocations in the studied materials were filled. The filed dislocation cores in an MBE samples were stoichiometric. However, in HVPE materials, single atomic columns show substantial differences in intensities and might indicate the possibility of higher Ga concentration in the core than in the matrix. A much lower intensity of the atomic column at the tip of the void was observed. This might suggest presence of lighter elements, such as oxygen, responsible for their formation.

Heterogeneous catalysis is one of the oldest nanosciences. Although model catalysts can be designed, synthesized, and, to a certain degree, characterized, industrial heterogeneous catalysts are often chemically and physically complex systems that have been developed through many years of catalytic art, technology, and science. The preparation of commercial catalysts is generally not well controlled and is often based on accumulated experiences. Catalyst characterization is thus critical to developing new catalysts with better activity, selectivity, and/or stability. Advanced electron microscopy, among many characterization techniques, can provide useful information for the fundamental understanding of heterogeneous catalysis and for guiding the development of industrial catalysts. In this article, we discuss the recent developments in applying advanced electron microscopy techniques to characterizing model and industrial heterogeneous catalysts. The importance of understanding the catalyst nanostructure and the challenges and opportunities of advanced electron microscopy in developing nanostructured catalysts are also discussed.

Sintering of a palladium catalyst supported on alumina (Al2O3) in an oxidizing environment was studied by in situ transmission electron microscopy (TEM). In the case of a fresh catalyst, sintering of Pd particles on an alumina surface in a 500 mTorr steam environment happened via traditional ripening or migration and coalescence mechanisms and was not significant unless heating above 500°C. After the catalyst was used for the hydrogenation of alkynes, TEM coupled with convergent beam electron diffraction and electron energy loss spectroscopy analysis revealed that most of the Pd particles were lifted from the alumina surface by hydrocarbon buildup. This dramatically different morphology totally changed the sintering mechanism of Pd particles during the regeneration process. Catalytic gasification of hydrocarbon around these particles in an oxidizing environment allowed the Pd particles to move around and coalesce with each other at temperatures as low as 350°C. For catalysts heating under 500 mTorr steam at 350°C, steam stripped hydrocarbon catalytically at the beginning, but the reaction stopped after 4 h. Heating in air resulted in both catalytic and noncatalytic stripping of hydrocarbon.

John Cowley and his group at Arizona State University pioneered the use of transmission electron microscopy (TEM) for high-resolution imaging. Three decades ago they achieved images showing the crystal unit cell content at better than 4 Å resolution. Over the years, this achievement has inspired improvements in resolution that have enabled researchers to pinpoint the positions of heavy atom columns within the cell. More recently, this ability has been extended to light atoms as resolution has improved. Sub-Ångstrom resolution has enabled researchers to image the columns of light atoms (carbon, oxygen, and nitrogen) that are present in many complex structures. By using sub-Ångstrom focal-series reconstruction of the specimen exit surface wave to image columns of cobalt, oxygen, and lithium atoms in a transition metal oxide structure commonly used as positive electrodes in lithium rechargeable batteries, we show that the range of detectable light atoms extends to lithium. HRTEM at sub-Ångstrom resolution will provide the essential role of experimental verification for the emergent nanotech revolution. Our results foreshadow those to be expected from next-generation TEMs with CS-corrected lenses and monochromated electron beams.

Kinematical and two-beam calculations have been conducted and are compared to experimental precession data for the large unit cell crystal La4Cu3MoO12. Precession electron diffraction intensities are found to exhibit approximate two-beam behavior and demonstrate clear advantages over conventional SADP intensities for use in structure solution.

We have used in situ electron microscopy to observe the nucleation of Ge islands on lithographically patterned Si(001) mesas. Images were obtained at video rate during chemical vapor deposition of Ge, using a reflection electron microscopy geometry that allows nucleation to be observed over large areas. By comparing the kinetics of nucleation and coarsening on substrates modified by different annealing conditions, we find that the final island arrangement depends on the nature of the mesa sidewalls, and we suggest that this may be due to changes in diffusion of Ge across the nonplanar surface.

It is shown that room-temperature diffraction pattern spots and diffuse scatter can appear to change their size and appearance relative to reciprocal-space sublattice reflections when the scattering material corresponds in structure to a critical phase. Under such a condition, the material is considered to be continually on the verge of a phase transition and the diffraction spot will have no definite width, its apparent size in reciprocal space dependent on the strength of the scattering into the diffracted beam. It is thought that the materials described in the experiments—niobia-zirconia ceramic alloys—are capable of entering such a critical phase because of their recently suggested planar XY spin character. After first describing how the seemingly crystalline ceramic alloy can display XY-like behavior, we analyze the intensity dependence of the critical scattering from the alloy's oxygen superlattice using information-theoretic methods.

The microstructure and magnetic domain structure of a Co-CoO obliquely evaporated tape for magnetic recording are studied by analytical electron microscopy and electron holography, respectively. While the existence of Co and CoO crystallites is confirmed by energy-filtered electron diffraction, columnar structure of the Co crystallites surrounded by the densely packed CoO crystallites is visualized by an elemental mapping method with electron energy loss spectroscopy, and the crystal orientation relation among the Co crystallites is clarified by high-resolution electron microscopy. It is found that the neighboring Co crystallites have close crystal orientations. On the other hand, electron holography reveals the magnetic flux distribution in a thin section of the tape. Although there exists the background resulting from the effect of inner potential with thickness variation, the distribution of lines of magnetic flux is found to correspond well to the recorded pattern.

The structure of Xe precipitates with sizes in several nanometers embedded in Al is known to be stable and its structure is well confirmed. But knowledge about the structure of Xe precipitates with nanometer sizes is very limited. There are difficulties in observing such small structures embedded in a crystalline matrix. An off-Bragg condition is used to observe diffraction patterns, dark-field, and high-resolution transmission electron microscopy images. The structure of Xe precipitates with sizes of about 2 nm and smaller is observed and confirmed. They are in an fcc structure and their orientation relationship with the Al matrix is similar to that of larger crystalline Xe precipitates or in an undefined structure. The lattice spacing or atomic distance in such nanometer-sized Xe precipitates is smaller than those of larger Xe precipitates embedded in Al matrix. There is a trend that as the size becomes smaller, the precipitates are more likely to have an undefined structure.

We propose the formation of LEED patterns using a highly convergent beam forming a probe of nanometer dimensions. A reflection rocking curve may then be recorded in many diffraction orders simultaneously. Multiple scattering calculations show that the intensity variations within these rocking curves is as sensitive to the parameters describing the surface dipole layer as conventional I/V scans. However the data may be collected from areas sufficiently small to avoid defects and surface steps, radiation damage controlled by use of low voltages, and the information depth selected by choice of the (constant) voltage. We briefly discuss also the application of this method to oxides and the formation of atomic-resolution scanning images in an idealized instrument in which coherent diffracted LEED orders overlap.

About 1 monolayer of palladium was deposited onto a silicon (111) 7 × 7 surface at a temperature of about 550 K inside an ultrahigh vacuum transmission electron microscope, resulting in formation of Pd2Si nanoislands and a 1 × 1 surface layer. Pd clusters created from an excess of Pd atoms on the 1 × 1 surface layer were directly observed by in situ plan view high-resolution transmission electron microscopy. When an objective aperture was introduced so that electron diffractions less than 0.20 nm were filtered out, the lattice structure of the 1 × 1 surface with 0.33 nm spacing and the Pd clusters with a trimer shape were visualized. It was found that image contrast of the 1 × 1 lattice on the specific height terraces disappeared, and thereby an atomic structure of the Pd clusters was clearly observed. The appearance and disappearance of the 1 × 1 lattice was explained by the effect of the kinematical diffraction. It was identified that a Pd cluster was composed of three Pd atoms without a centered Si atom, which is consistent with the model proposed previously. The feature of the Pd clusters stuck at the surface step was also described.

InxGa1−xAs quantum dots in GaP(100) crystals prepared by the OMVPE technique are observed along the [011] direction with a newly developed 200-kV spherical aberration(Cs)-corrected HRTEM, a 200-kV annular dark-field (ADF)-STEM, and a 200-kV conventional HRTEM equipped with a thermal field-emission gun. The dots are 6–10 nm in size and strongly strained due to the misfit of about 9% with the GaP substrate and GaP cap layer. All of the cross-sectional high-resolution electron micrographs show dumbbell images of Ga and P atomic columns separated by 0.136 nm in well-oriented and perfect GaP areas, but the interpretable images are limited to those taken with the Cs-corrected HRTEM and ADF-STEM with Fourier filtering of the images. The Cs-corrected HRTEM and ADF-STEM are comparable from the viewpoint of interpretable resolution. A detailed comparison between the Cs-corrected HRTEM images and the simulated ones with electron incidence tilted by 1° to 5° from the [011] zone axis gives information on local lattice bending in the dots from the images around 0.1 nm resolution. This becomes one of the useful techniques newly available from electron microscopy with sub-Ångstrom resolution.