For microstructure analysis of magnetic materials, in particular, for a new class of Nd-Fe-B hard magnets, a number of magnetic imaging techniques can be applied to gain different microstructure information. For instance, SEM magnetic imaging can be successfully used for very thick samples from a bulk ones down to a sample thickness of 10-20 μ. Below this limit the resolution of SEM magnetic images generally is too low to visualize fine details of magnetic structure. For very thin samples (less then 0.5 μm) magnetic Lorentz microscopy methods in conventional TEM are very useful tools to reveal magnetic domains and their structure. Therefore, it seems that relatively thick films of 0.5-20 μm in thickness are left out of the experimental limits of the methods discussed.

We report on a novel, indirect TEM method that allows us to extend the limits of magnetic imaging for TEM non-transparent magnetic samples by viewing the surface stray fields and fringing fields generated by magnetic domains at specimen edge.

Semiconductor production using 8-inches wafers seems to be widely carried out in semiconductor factories. Many FE-SEMs are coming to be used for their particle/defect inspection. As semiconductor processing turns precise, the need of a high resolution FE-SEM has been rising all the more.

Now, JEOL announces a high resolution FE-SEM (8"FE-SEM) for 8-inches wafer inspection, which takes over the strong point of the semi-in-lens field emission scanning electron microscope (FE-SEM) JSM-6340F which has been getting expensive evaluation with its high resolution performance. The 8"FE-SEM possesses a resolution of 3.0nm at lkV of its acceleration voltage and 1.2nm at 15kV, as its semi-in-lens type object lens makes it possible to minimize the aberration.

High resolution observation and inclined wafer observation can be selected properly by changing the wafer holder, as such a holder exchange system has been usually used for transmission electron microscopes.

At the University of Melbourne we have been pursuing an ongoing program of investigation into the recovery of phase information from intensity measurements [1-9]. In this paper we consider the implications of this work in optical and electron microscopy.

Many objects of interest to biologists are phase objects which means that light is slowed and refracted in the object but not absorbed. Techniques such as phase-contrast microscopy and Nomarski differential interference contrast (DIC) microscopy are traditionally used to render the phase structure visible but do not directly map the phase distribution and are unable to produce quantitative data. These techniques also entangle the phase and amplitude information and so have limitations where both phase and amplitude information is required. In this paper, we demonstrate quantitative non-interferometric recovery of microscopic phase structure using incoherent illumination of the type commonly found in optical microscopes, and show that our results correlate with structure observed using Nomarski DIC techniques.

A confocal, all-reflecting, infinity corrected microscope was developed to improve infrared microspectrometer (IMS) systems. This system consists of a Fourier transform spectrometer, an infinity corrected all-reflecting microscope, a viewing system for visible light microscopy, a single remote confocal mask for aperturing in either a transmission or reflection mode and an infrared detector. The radiant energy passes through a confocal aperture mask as it travels from the source to the sample and through the same confocal aperture mask as it travels from the sample to the radiant energy detector. A single confocal mask insures that the sample defining mask's size, geometrical shape and orientation are always matched. Mask alignment is assured and simplified since only one adjustment is needed. An aperture beam-splitter is used for reflection measurements. This system simplifies the operation of the microspectrometer, permits the use of standard commercial infinity corrected microscope optical components and reduces the cost of IMS accessory microscopes.

Isotopic measurements via Time-of-Flight Secondary Ion Mass Spectrometry (ToF-SIMS) have generally not been considered as practical because of the low duty cycle at which ToF-SIMS instruments operate and the corresponding low data rate. The recent discovery of pre-solar material in meteorites has shown that large variations in isotopic ratios (several orders of magnitude for some elements) exist in small (∼1 μm), refractory meteoritic grains. These grains are ideal candidates for ToF-SIMS, which consumes little sample material, compared to dynamic, magneticsector SIMS. ToF-SIMS also allows for parallel detection of all species present in the sample; thus, multiple isotopic systems can be studied in one measurement. As a prerequisite to studying the isotopic composition of meteoritic materials, preliminary determinations of ratios for a number of elements have been made on materials of known isotopic composition. This allows us to investigate problems that may be unique to ToF-SIMS for the measurement of isotopic ratios.

In recent years there has been considerable interest in two and three-photon excited fluorescence in laser scanning optical microscopy. Because absorption is confined tot he focal plane of the objective, these techniques provide intrinsic optical sectioning without the use of a confocal aperture. In addition, photobleaching and phototoxicity are greatly reduced above and below the focal plane. We have adapted a two-photon microscope to utilize surface second harmonic generation (SHG) as a new contrast mechanism for nonlinear optical biological imaging.

Surface SHG was first described by Shen [1] and arises from the second order nonlinear susceptibility, χ(2). Signal will only arise from a non-centrosymmetric environment such as an interfacial region. Thus this technique has the potential to probe cellular membranes at high specificity. Further, since SHG results from an induced polarization and not absorption, photobleaching considerations are greatly reduced over fluorescence based methods.

The two photon microscope provides optical sectioning of fluorescent specimens with a resolution comparable to that obtained in confocal microscopy (see refs 2,3). However, the excited volume in 2-photon microscopy is limited to the focal volume (unlike conventional fluorescence microscopy where excitation occurs throughout the specimen). This means that photodamage is limited to the plane of section being examined. Thus, the light emitted from each point in the specimen depends on the amount of fluorochrome present without the problem of prior illumination (of other planes within the specimen) reducing the photon yield so a better signal to noise ratio can be obtained when examination of multiple image planes is needed. Since 2-photon excitation spectra are wide and chromatic aberration is eliminated (because the emitted light does not have to be focused on a pinhole), it is possible to excite several fluorochromes simultaneously and map their positions with high accuracy.

The laser confocal microscope (LCM) is now an established research tool in biology and materials science. In biological applications, it is usually employed to detect the location of fluorescent marker molecules and, under these conditions, detected signal levels from bright areas often represent <20 photons/pixel (assuming a standard 1.6 μs pixel time) while those from dark areas are likely to average <1 photon/pixel. Although this data rate limits the speed at which information can be derived from the specimen, saturation of the fluorophor, photobleaching of the dye, and phototoxicity often prevent it being increased by simply using more laser power. Over the past 10 years, the optical photon efficiency of commercial confocal instruments has improved significantly and it is now reaching the point where further improvement is becoming very expensive. The only component is which a significant improvement is still possible is the photodetector.

Two-Photon Excitation (TPE) microscopy is a powerful tool for the imaging of biological specimen. The most important advantages compared with conventional confocal microscopy are: intrinsic optical sectioning, reduced phototoxcicity and photobleaching, and an increased penetration into the specimen. The increased penetration arises from the use of the longer wavelength excitation light, in the red or near infra-red.

Here, a quantitative in-depth TPE imaging study is presented on dental biofilm. The biofilm contains a mixed culture of a large number of oral bacteria with a typical size of ≈ 1 μm diameter. They exhibit a complex microstructure and heterogeneous physiological and biochemical properties. The high density of the biofilms limits the depth range over which confocal images can be acquired to 25-40 μm.

Using a two-photon excitation microscope fluorescence intensity images could be recorded over the whole thickness (100 μm) of the biofilm sample.

Research in our lab is aimed at examining how signaling molecules affect the differentiation of neurons in the developing cortex. Most morphological studies rely on single time point images of cells within a population. It is obviously advantageous to examine changes in individual living cells over time. By combining 2-photon imaging with co-transfection of genes for green fluorescent protein (GFP) and neurotrophins in living neurons, we can asses the effects of these on individual living cells.

Individual neurons in brain slices from ferret visual cortex were transfected with GFP using particle mediated gene-transfer, or Biolistics. Using the Helios Gene Gun (Bio-Rad), DNA-coated gold particles are blasted into living brain slices. Neurons are transfected when a gold particle lands in the nucleus. Within 24 to 36 hours individual cells are well-filled with GFP allowing visualization of dendrites, spines, and axons.

A video-rate scanning two-photon excitation microscope (TPEM) has been successfully constructed and tested. The TPEM, based on a Nikon RCM-8000, incorporates a femtosecond pulsed laser, a pre-chirper, and a non-confocal detection box for ratio imaging. Fig. 1 shows the schematic layout of the main components of the instrument, each of which is briefly discussed below.

Laser System: A Tsunami Ti: Sapphire laser (from Spectra-Physics) is optically pumped by a 5 W green laser (Millennia from Spectra-Physics) and is capable of generating 100 fs pulses at a repetition rate of 82 MHz and an average power of 0.8 W. The output wavelength is tunable from 690 to 1050 nm with three optical sets, each covering part of the spectrum with some overlapping.

Pre-chirper: After leaving the Tsunami, the laser beam enters an optic unit known as a pre-chirper which pre-chirps laser pulses to compensate for the group velocity dispersion which will result when the laser beam goes through the microscope optics.

Hair cells, the mechanotransducers of the inner ears and lateral lines of vertebrates offer an excellent preparation for the study of presynaptic mechanisms at fast chemical synapses. They offer the compact, almost cylindrical shape of an epithelial cell without entangling dendrites and axons, and make afferent synapses directly from the cell body. The biology of hair cells demands an unusually high performance of the output synapse, requiring continuous, and very quickly modulated transmitter exocytosis. Release follows the small receptor potential generated by the transduction apparatus; more transmitter is released upon depolarization, less after hyperpolarization. In some hair cells, transmission can phase-lock to kilohertz stimulus frequencies, preserving temporal information across the synapse.

We have exploited the synaptic terminal-like features of these cells to explore both the physiology and ultrastructure of the machinery underlying neurotransmitter release.

Electron microscope tomography is a useful method for deriving three-dimensional (3-D) structures from tilt series. We used an intermediate, high-voltage electron microscope operated at 400 kV to study the 3-D structure of selectively stained or labeled mitochondria from diverse sources. We have examined the structure of neuronal mitochondria and mitochondria from brown adipose tissue (BAT) employing either single-axis or double-axis techniques over a tilt range of ±60° with 1° or 2° angular increments (Fig. 1). Although BAT mitochondria contains only lamellar cristae with no tubular cristae as we have observed in neuronal mitochondria, we found that the inner membrane of BAT mitochondria invaginates to form cristae only through narrow, tubular openings or crista junctions (Fig. 2).

Recent interest in the possibility of structural artifacts in chemically fixed mitochondria imaged in situ motivated us to perform electron tomography on cryofixed brain and brown fat tissue.

Electron microscopic tomography is proving to be a valuable tool for investigating the 3D structure and organization of cellular organelles. Important progress is being made in the application of the technique to frozen-hydrated material, but it is likely that success with thick specimens will be limited by the low contrast and beam sensitivity of naked biological material. Thus, optimizing procedures for fixing, embedding, staining, and selectively labelling cells for 3D electron microscopy remains a priority.

Tomography of chemically fixed and plastic-embedded rat-liver tissue and isolated mitochondria has shown that the cristae (the invaginations of the mitochondrial inner membrane) are pleiomorphic and connected to each other and to the surface of the inner membrane by tubular regions 30-40 nm in diameter. This basic design feature has important implications for the microcompartmentation of ions and molecules within this organelle.

The mitochondrial inner membrane contains the machinery of oxidative phosphorylation. This membrane has invaginations called cristae which vary widely in shape between organisms and between tissues in the same organism. Electron microscopic tomography indicates that, despite this pleiomorphism, there is a common design feature, namely, the cristal membranes connect to each other and to the periphery of the inner membrane by tubular regions 30-40 nm in diameter. This finding has important implications for the internal diffusion of ions, metabolites and macromolecules within mitochondria.

In some types of mitochondria, the cristae exhibit periodicity.1 In the case of the amoeba Chaos carolinensis, detailed analysis and modeling of thin-section images of mitochondria in starved cells indicate that the highly curved cristae correspond to periodic cubic surfaces. We are undertaking electron microscopic tomographic and crystallographic approaches to more thoroughly characterize these membrane phases and, in particular, establish the continuity of the internal compartments which they define.

Advances in both electron cryomicroscopy and tomographic reconstruction methods have made it possible to collect and analyze data from large macromolecular assemblies frozen without fixation, sectioning, or staining. The Spindle Pole Body (SPB) is the microtubule organizing center in the yeast Saccharomyces, and is involved in mitosis, nuclear movement, budding, and mating. We have used electron cryomicroscopy and tomographic reconstruction to investigate the 1500Å diameter, 600Å high cylindrical core structure produced by limited heparin treatment of isolated SPBs.

Tilt series' of SPB cores preserved in vitreous ice were taken under low electron dose conditions. These images have been analyzed using iterative correlation alignment and computed back projections for tilt data without fiducial markers. In these studies, electron micrographs were digitized, and translationally aligned using correlation methods.

The organization of the microtubule network of eukariotic cells is controlled by the centrosome. This organelle which duplicates once during the cell cycle comprises two centrioles and a dense filamentous material called pericentriolar material. The centrioles are cylinders of ≈ 0.25 μm in diameter and ≈ 0.5μm in length and are composed of 9 microtubule triplets at the proximal end and 9 doublets at the distal end. For the last 50 years the centrosome and the centrioles in particular have been the object of many optical and electron microscopy studies. But, in spite of their highly symmetric organization, their fine structure is still not well understood. The reconstruction presented here is a substantial step forward towards a more detailed understanding of their structure.

Centrosomes were isolated from a human lymphoblastoma cell line (1). 10 nm gold particles were applied to formvar-carbon coated grids to provide markers for the alignment procedure (2,3). Centrosomes were centrifuged onto the grids and negatively stained.

Smooth muscle is a machine consisting of working and supporting elements whose structure and 3D organization must be elucidated for the mechanics of shortening and tension generation to be understood. Based on longitudinal and serial transverse sections of rabbit portal vein it was suggested that the contractile elements of smooth muscle formed “mini-sarcomeres”, analogous to skeletal muscle, containing parallel arrays of 3-5 myosin filaments 1.6-2.2 um long. Observations at the light microscopic level were consistent with this idea. The past decade has seen little further investigation into the in situ ultrastructure of this or other smooth muscles, and the general applicability of these findings remains unknown. We have taken advantage of recent methodological advances, which can provide full 3D computer representations of cellular organization based on EM data, using guinea pig taenia coli muscle as a model system.

Serial transverse sections (Fig 1) were used to generate 3D reconstructions of the organization of the myosin filaments and their relation to dense bodies, actin bundles, mitochondria and other organelles.

Light microscopic imaging has experienced a renaissance in the past decade or so, as new techniques for high resolution 3D light microscopy have become readily available. Light microscopic (LM) analysis of cellular details is desirable in many cases because of the flexibility of staining protocols, the ease of specimen preparation and the relatively large sample size that can be obtained compared to electron microscopic (EM) analysis. Despite these advantages, many light microscopic investigations require additional analysis at the electron microscopic level to resolve fine structural features.

High voltage electron microscopy allows the use of relatively thick sections compared to conventional EM and provides the basis for excellent new methods to bridge the gap between microanatomical details revealed by LM and EM methods. When combined with electron tomography, investigators can derive accurate 3D data from these thicker specimens. Through the use of correlated light and electron microscopy, 3D reconstructions of large cellular or subcellular structures can be obtained with the confocal microscope,

Three-dimensional reconstructions from projections are usually ridden by noise from different sources. A common problem among many three-dimensional reconstruction techniques is the systematic absence of certain projections, but also the accidental absence of spurious projections. In these three-dimensional reconstructions such absences are visible as directional smearing due to convolution. Other convolution effects such as those due to the optics of the instrument used to record the data usually cause severe damping of high frequencies and even contrast reversal (common in images from electron microscopes).

Several approaches to overcome these ‘noise’ effects in three-dimensional reconstructions have been developed, but they usually suffer from the very small radius of convergence. Very easily and commonly, the refinement iterations end up stuck in a premature choice of a minimum. We have developed another algorithm, constrained maximum entropy tomography (COMET), that in practice has been shown to overcome these problems.