Recently, the late R. Shiromizu carried out studies to establish the limits of metastability in the Cu-Be system. Con-currently, we engaged in parallel research, but with a broader emphasis on the microstructures concomitant with quenching, aging and reversion. A complete report covering 0.16 to 2.1 wt.% Be will be published elsewhere; certain unique structures are presented below.

Alloys of >0.8 wt.% Be quenched from the α (f.c.c.) region exhibit weak {110} tweed strain contrast. Direct lattice imaging did not detect discrete particles in this condition, whereas rather high densities of {100} G.P. Zone plates in regular stairstep array were revealed after aging at temperatures well below the G.P.Z. solvus.

An aluminium alloy (L65) was fatigue tested in a longitudinal tension compression mode at a frequency of 20 kHz with a mean strain of 1.65.10-3. The specimens used were simple cylinderical rods of fully hardened, solution heat treated or fully annealed aluminium alloy. The effects of the variation in heat treatment on the nature of fracture were investigated by examining the fracture surfaces of these specimens in a Cambridge S4-10 scanning electron microscope.

The fracture surfaces of the fully hardened alloy exhibited the characteristics of cleavage fracture (Stage I and Stage II cracking) in agreement with Forsyth's theories. Occasionally these specimens underwent an unusual mode of centre initiated failure for which the fatigue endurance was greater than that experienced following the surface initiation in similar specimens.

When deformed metals and alloys are annealed at sufficiently high temperatures, strain-free nuclei appear at preferential sites and grow predominantly by grain-boundary migration. A majority of these strain-free grains collectively exhibit specific types of crystallographic orientations which significantly differ from those of the deformed matrix. The texture thus formed is frequently related to the deformation texture by simple angular rotations about particular crystallographic axes. In contrast to most metals and alloys for which the recrystallization and deformation textures are distinctly different, Ti exhibits a similar texture in both the cold-rolled and recrystallized conditions; this texture is characterized by two symmetrical texture components such that the (1010) in each remains parallel to the rolling direction and the [0001] is tilted along the transverse direction between 30 and 40 deg from the rolling-plane normal.

Small amounts of nickel are added to titanium to improve the crevice corrosion resistance but this results in an alloy which has sheet fabrication difficulties and is subject to the formation of large Ti2Ni precipitates. These large precipitates can serve as local corrosion sites; but in a smaller more widely dispersed form, they can have a beneficial effect on crevice corrosion resistance. The purpose of the present work is to show that the addition of a small amount of Mo to the Ti-1.5Ni alloy reduces the Ti2Ni precipitate size and produces a more elongated grained microstructure. It has recently been reported that small additions of Mo to Ti-0.8 to lw/o Ni alloys produce good crevice corrosion resistance and improved fabrication properties.

Silicoflagellates are single-celled flagellated protists that secrete a relatively simple siliceous skeleton generally less than 100 microns in length. Silicoflagellate skeletons have been recovered from many types of marine deposits and have been utilized by geologists in biostratigraphic and paleoecologic analysis. The systematics of this group is established upon overall skeletal morphology. Recently, the importance of ornamentation found on the surface of these skeletons has been emphasized The surface of the silicoflagellate skeletons may be smooth, or ornamented with spines, warts, linear striations, ridges and complex reticulations (i.e. ridges of silica in which at least three converge to a point on the skeleton).

If aluminum is anodized in a bath, such as sulfuric acid, in which the anodic oxide coating has appreciable solubility, then the oxide layer may grow to a thickness of many microns. The resulting film can be regarded as consisting of two layers, a thin barrier layer adjoining the surface of. the metal, and an outer thick film containing pores. The latter convey the electrolyte to the barrier layer, permitting continued oxidation.

We have examined the structure of a number of sulfuric acid anodized films to assess the applicability of the two principal models of the pore structure. The older model represented the pores as rather regularly spaced, uniform channels running nearly normal to the oxide surface, separated by walls of amorphous alumina. Although conceived as starshaped rather than circular in cross-section, experimental evidence has pointed to the latter. In this ideal model, the porous film was made up of hexagonal cells of amorphous alumina, each containing a central pore.

An energy analyzing transmission electron microscope of the Möllenstedt type was used to measure the electron energy loss spectra given by various layer structures to a spatial resolution of 100Å. The technique is an important, method of microanalysis and has been used to identify secondary phases in alloys and impurity particles incorporated into epitaxial Si films.

Layers Formed by the Epitaxial Growth of Ge on Si Substrates Following studies of the epitaxial growth of Ge on (111) Si substrates by vacuum evaporation, it was important to investigate the possible mixing of these two elements in the grown layers. These layers consisted of separate growth centres which were often triangular and oriented in the same sense, as shown in Fig. 1.

Several techniques have been developed for the quantitative microchemical analysis of thin foils. These procedures require that in an alloy AB the measured intensity ratio, IA/IB frora the foil be independent of foil thickness. Cliff and Lorimer find that this condition is met in all cases they have examined whereas Jacobs and Baborovska show data where absorption of longer wavelength X-rays was apparent.

The present study was undertaken to evaluate these procedures for the combination of a scanning transmission electron microscope (JEOL JSEM 200) and an energy dispersive analyzer (Ortec).

A Philips EM-200 electron microscope has been modified to incorporate two electron guns as described by Ong and Gold, for operation in the conventional (TEM) and scanning (SEM) mode.

The second gun, the SEM gun, is mounted below the viewing chamber and uses the imaging system of the microscope as the probe forming electron optic. The electrons follow essentially the same path as the image forming electrons for TEM, only in the opposite direction.

The modifications to the microscope column are designed to provide the space for the scanning coil assembly, the various detectors, and the second gun. A spacer is located above the objective lens to house the transmission electron detector, another spacer, below the objective lens, incorporates the scanning coil and a secondary electron detector.

A scanning attachment has been designed for the Corinth electron microscope enabling either standard transmission images of thin specimens (TEM) or reflection scanning images of bulk specimens (SEM) to be obtained from the one instrument.

At the heart of the modified instrument is a lens which acts as both the final projector lens for the TEM and also as the objective lens for the SEM with a working distance of 5mm. Thus even when used as a SEM the whole length of the Corinth column is traversed by the beam and this presents the choice of two possible modes of operation, namely, a two lens mode in which only the first condenser and final projector are excited, or a three lens mode in which the TEM objective lens is also excited. The three lens mode provides a range of much higher gaussian demagnifications.

This paper discusses the latest developments and results obtained from a field emission scanning transmission microscope. Some features of the instrument have already been described. The microscope has been in routine operation for more than two years at accelerating voltages of 80 and 100kV. Over this period a number of (111) and (310) oriented tungsten tips have been fitted and operated for varying periods in order to access the minimum lifetime to be expected. No tip has yet failed in service, and the (111) tip in current use has given more than 500 hours of stable and trouble free operation, in a vacuum of 3 to 5 x 10-11 torr.

The illumination system of the conventional transmission electron microscope is designed to provide an adequate level of coherent illumination over the field of view being examined. The additional requirement of providing the maximum possible current into very small (sub-micron) focussed spots, as is required by X-ray microanalysis for example, necessitates a probe forming lens with a very short working distance and has previously been met by using either a mini-lens or the prefield of the objective lens when used in the Single-Field Condenser-Objective mode.

The SFCO gives the smallest spot size and aberration parameters but suffers from the disadvantage that the illumination and imaging fields are not independently controllable.

The transmission electron microscopic study of thin metal film interconnections on electronic integrated circuits requires a sophisticated in-situ capability. Electromigration, particularly, cannot be investigated solely by viewing static heat-treated or electrically stressed specimens ex-situ. The key to the early electromigration studies by Blech, and others, was the in-situ observation of material movement in anelectrical-powering-stage equipped TEM. Blech's approach was insensitive to structure effects because of an absence of heating capability; he therefore relied on joule heating and temperature gradient effects.

Many of the TEM studies of radiation damage in crystalline materials have been directed toward illucidating the nature, number density, and size distributions of the primary structural defects resulting from the displacement of atoms from their normal lattice sites, i.e., "black spots." The "black spots" have been identified as either nonresolvable dislocation loops or planar clusters of self interstitials or vacancies by TEM techniques and diffraction contrast theories that are described in the book by Hirsch, et al. and the reviews by Ruhle and Eyre. An indispensable part of the studies of point defect clusters is the identification of their character (vacancy or interstitial).

The creep behavior of oxide dispersed alloys can be described by the following equation:

where ε is steady state creep rate, σ applied stress, R gas constant, T temperature in °K, Qc activation energy for creep, n stress exponent, A and n are constants.

Two unusual observations in the case of O.D. alloys are that Qc and n take on values that have not been explained satisfactorily. This investigation was undertaken to explain the unusually high strength, high Qc and high stability (high value of n) in the oxide dispersed alloys.

A good strength, high temperature Ni-Al alloy had been produced by thermomechanically cycling sintered aluminium and nickel powders. The microstructure of the alloy consisted of a solid solution in which there was a fine dispersion of inert ceramic particles. A typical micrograph is shown in Fig 1 in which three distinct types of particle A, B and C are indicated. To explain the mechanical properties of the alloy it was necessary to determine the composition and structure of the particles.

As the particles were 300Å to 1200Å in size the only satisfactory method to use was combined electron microscopy and energy analysis. Thinned foils containing unobscured particles were used so that, by placing slit of the Mollenstedt analyser across it, the energy loss spectrum of each particle was measured directly.

Cathodoluminescence (CL)-mode scanning electron microscopy studies were made on n-type GaAs substrate crystals, and on liquid (LPE)- and vapor-phase-epitaxial (VPE) GaAs. The observations were carried out in a Stereoscan S4-10 equipped with a separate CL detector and signal processing channel. Broad-band detection was employed, with the CL radiation from the specimen focused directly onto a photomultiplier (RCA type 7102) with S-l spectral response. The CL was excited by 30 kV electrons, with beam currents typically about 10 nanoamperes.

Figure 1 shows a low dislocation-density (˜103/cm2) substrate crystal. The CL image contains only heavy horizontal striations, which may be attributed to varying concentrations of dopant formed during growth, and a few contrast effects related to surface features (cf., emissive-mode (SE) image).The high dislocation-density (˜105/cm2) GaAs substrate in Figure 2 shows in addition a distribution of discrete dark spots and short line segments, identified as emerging dislocations: Dislocation etching of the specimen revealed a distribution and density of dislocations comparable to the CL-image dark spots.

A coherent precipitate known as gamma prime develops in iron-nickel base alloys when alloying additions of aluminum and titanium are included. The gamma prime phase is an ordered fcc structure of Ll2 symmetry having an Ni3(A1, Ti) composition. Gamma prime is generally found as a fine spherical or cubic precipitate, uniformly distributed throughout the material. Imaging is usually difficult because lattice mismatches are small and only superlattice reflections of weak intensity are available for dark field imaging. This paper presents evidence that the gamma prime phase can be formed by discontinuous precipitation, developing a typical cellular structure.

An experimental alloy, Fe-14.9Cr-34.8Ni-l.5Al-l.98Ti-.004C (weight percent), has been given a solutionizing treatment of 1145°C for 1 hour followed by an air quench and an aging treatment of 815°C for 1 hour. Examination of the microstructure by TEM indicated that a cellular precipitate had developed extensively at grain boundaries.

As part of a high resolution electron microscopic study of the precipitation sequence at 130°C in aluminum - 3% copper alloy, we studied G. P. [1], G. P. [2] or θ”, and θ’ precipitates by the lattice imaging technique. This approach proved valuable for phase identification and distinction. Examples of cros sed-lattice and c-plane lattice images at θ’ platelets will be presented here.

Slices were spark cut nearly parallel to {001} from a melt-grown single crystal of aluminum -3.0 wt. % copper, previously solution treated at540°C and water quenched. Slices were aged at 130°C in argon and hand ground. Discs were chemically, electrolytically or in one case ion thinned, and examined at 100 kV in a slightly modified Philips EM 300 microscope.

In the aluminum-copper system, the intermediate phases (G. P. zones and θ’) separate parallel to cube matrix planes, so that the slice orientation chosen resulted in two sets of platelets being parallel and one set normal to the beam.

Investigation of the particulate matter contained in the water sample, revealed the presence of a number of different types and certain of these were selected for analysis.

An A.E.I. Corinth electron microscope was modified to accept a Kevex Si (Li) detector. To allow for existing instruments to be readily modified, this was kept to a minimum. An additional port is machined in the specimen region to accept the detector, with the liquid nitrogen cooling dewar conveniently housed in the left hand cupboard adjacent to the microscope column. Since background radiation leads to loss in the sensitivity of the instrument, great care has been taken to reduce this effect by screening and manufacturing components that are near the specimen from material of low atomic number. To change from normal transmission imaging to X-ray analysis, the special 4-position specimen rod is inserted through the normal specimen airlock.