Minimization of dose as a criterion for the selection of imaging modes in electron microscopy of amorphous specimens.

A fundamental limitation in electron microscopy of organic specimens is radiation damage by the electron beam. To minimize damage it is necessary to have maximum information collection for a given dose. Various modes of operation of conventional and scanning transmission microscopes are compared with respect to their ability to detect small changes in specimen thickness or density with a given signal to noise ratio. Incoherent imaging is assumed, and this is expected to apply to amorphous specimens under a variety of microscope conditions. For either very thin or very thick specimens, the scanning transmission microscope is found to require nearly 10 times less dose than a conventional microscope for the same signal to noise ratio in the image. For specimens of intermediate thickness, scanning and conventional transmission electron microscopes are roughly equivalent.     

Charging effect in electron-irradiated ice

To maintain the original distribution pattern of diffusible elements in biological samples, electron probe microanalysis is carried out with frozen hydrated bulk specimens and cryosections, analysed at temperatures below 130 K. Ice has a very low intrinsic conductivity at this working temperature and surface- and space-charging appears, when uncoated specimens are irradiated with non-penetrating electrons. Although coating with a grounded conductor abolishes the surface potential, the build-up of an internal space-charge field is possible, depending on the sample thickness and beam voltage used. Consequently, the geometry of the X-ray source volume and the spectral distribution of the emitted continuous and characteristic X-rays are affected. To simulate the situation for microanalysis of frozen hydrated specimens the charging process in electron irradiated ice is studied by recording simultaneous specimen currents from the top and bottom of ice layer preparations. The external currents yield information on the build-up of internal space-charge fields which result from the balance of charge injection, storage, and transport. Irradiation of uncoated bulk specimens with a finely focused beam results in the build-up of a space-charge field close to the surface, which causes a reduction of the depth of microprobe analysis. In coated bulk specimens the induced conductivity renders possible a current flow to the front electrode, thereby limiting the space-charge field. Sections with an effective rear electrode will not charge appreciably if the electron range is larger than about half the section thickness.     

Beam heating of a moderately thick cold stage specimen in the SEM/STEM

A mathematical model for the electron beam heating problem of thick specimens (thickness less than the electron range) in the cold stage of SEM/STEM is proposed, and an analytic solution for the problem is provided. The use of the model is demonstrated by calculating the maximum temperature rise in a red blood cell mounted on a metallized polymer film, for a large range of operating conditions. With a probe size of 20 nm and probe current of less than 10 nA the temperature rise was found to be on the order of 5 K. This temperature rise increases proportionally to the probe current if all other variables are kept constant. Although the amount of energy dissipated in thick specimens is large relative to thin films, temperature rise can be minimized by mounting the thick sample on a metal stub or on a metallized film thus providing large heat transfer area through a high thermal conductivity substrate. Of course our calculations only have meaning if the specimen is in good thermal contact with the underlying support.     

Crystallographic analysis of thin specimens

Electron backscattering diffraction (EBSD) is commonly used on bulk samples for crystallographic material characterization. In this work, the technique was applied on transmission electron microscopy (TEM)-type thin specimens, prepared with a focused ion beam. Orientation maps were successfully collected on specimens made of a Cu₃Au copper–gold alloy. As compared to EBSD analysis on a bulk specimen, an improved pattern quality and a high spatial resolution (well below 10 nm) were obtained. Furthermore, a clear improvement of the signal-to-noise ratio with decreasing sample thickness was observed.     

Dynamic hydration effects in an electron microscope cold stage

Water can be a substantial proportion of the residual gas in modern electron microscopes even when frozen hydrated specimens are not used. During measurements of the mass thickness of thin collodion film specimens at low temperatures, it was found that a volatile surface layer (condensed water) modified the apparent rate of mass loss induced by radiation exposure. Mass loss can be enhanced by the presence of water (specimen “etching”), or mass loss can be masked by the dynamic adsorption of water to the specimen surface. The microscope or the grid can be a secondary source of the water; even with cold anticontaminator plates in the vicinity of the specimen, water can be desorbed by x-rays or backscattered electrons. In one typical situation, the mass loss rate appears reduced (due to water adsorption), but the ultimate damage is greater (due to etching). These results illustrate that care must be taken in interpreting mass thickness measurements made in the presence of water and that the lowest stage temperature does not necessarily produce the best observation conditions for all specimens.     

Thick specimens in the CEM and STEM. Resolution and image formation.

A theory of resolution and image formation is presented for thick amorphous specimens in transmission electron microscopes. Eight modes of operation are considered, four in the scanning transmission electron microscope (STEM) and four in the conventional electron microscope (CEM). A thick specimen is defined here as one in which the resolution of detail is limited by plural scattering of the electron beam. In practice this includes films on the order of a micron in thickness. An analytic theory of plural incoherent scattering is developed which is general with respect to material and beam voltage. The theory gives the distribution of elastically scattered electrons as a function of transverse coordinate and angles, and is directly applicable to optical systems. The theory applies to all thicknesses normally encountered, and includes thin specimens as well as thick specimens. Criteria are proposed for evaluation of the quality of microscope images, and the modulation transfer function is applied to determine some practical estimates of picture quality. The STEM is found to have distinct advantages over the CEM for thick specimens. For a carbon specimen one micron thick a STEM operating in bright field at 90 keV produces an image which is roughly equivalent to that of a CEM operating in bright field at 1 MeV. Improvement can be obtained in the CEM by filtering out eneryg-loss electrons which degrade resolution due to chromatic aberration. This results in a reduction in signal intensity and usable thickness, however.     

Dopant profiling of focused ion beam milled semiconductors using off-axis electron holography; reducing artifacts,extending detection limits and reducing the effects of gallium implantation

Focused ion beam (FIB) milling is one of the few specimen preparation techniques that can be used to prepare parallel-sided specimens with nm-scale site specificity for examination using off-axis electron holography in the transmission electron microscope (TEM). However, FIB milling results in the implantation of Ga, the formation of amorphous surface layers and the introduction of defects deep into the specimens. Here we show that these effects can be reduced by lowering the operating voltage of the FIB and by annealing the specimens at low temperature. We also show that the electrically inactive thickness is dependent on both the operating voltage and type of ion used during FIB milling.     

Preparation of (InGa)As/GaAs materials for TEM by one side non-rotation ion beam thinning

A technique is described for the preparation of thin specimens for transmission electron microscopy (TEM) of (InGa)As/GaAs multilayered materials. In this technique, a shielding method is used for selective-area perforation by ion beam thinning. Thin cross-sectional specimen slices are mechanically pre-thinned to about 30 μm and then thinned by ion sputtering from one side of the specimen at a time without rotation of the specimen stage. No direct ion sputtering occurs at the growth surface of the specimen so that a specimen with thin areas containing the desired near-surface structures can be obtained. The recipe for this technique is given in detail. A patterning method for increasing the size of the thin area for TEM investigation is also described. It is shown that a smooth surface can be obtained by sputtering without rotating the stage if obstacles that produce redeposits onto the sputtered surface are removed.     

Improving image quality by zero-loss energy filtering: quantitative assessment by means of image cross-correlation

Fourier ring correlation and root-mean-square contrast of pairs of images, taken under identical conditions, were used as criteria of image quality for comparing unfiltered with zero-loss energy-filtered imaging using a TEM equipped with a post-column energy filter. For three different specimens (amorphous carbon film, macromolecules in light negative stain, virus particles in deep negative stain) the dependence of these quantities on electron dose, specimen thickness and defocus was investigated. A model, based on simple assumptions, was used to describe quantitatively their dependence on electron dose and specimen thickness. It was found that energy filtering is most advantageous for low-dose imaging and small defocus values. The gain due to energy filtering strongly increases with specimen thickness, whereby the dependence is linear for light scattering elements. For thick specimens, the gain by energy filtering is more pronounced in the resolution range between 4 and 2 nm than for lower spatial frequencies.     

Preparation of high-quality ultrathin transmission electron microscopy specimens of a nanocrystalline metallic powder

This article explores the achievable transmission electron microscopy specimen thickness and quality by using three different preparation methods in the case of a high-strength nanocrystalline Cu-Nb powder alloy. Low specimen thickness is essential for spatially resolved analyses of the grains in nanocrystalline materials. We have found that single-sided as well as double-sided low-angle Ar ion milling of the Cu-Nb powders embedded into epoxy resin produced wedge-shaped particles of very low thickness (<10 nm) near the edge. By means of a modified focused ion beam lift-out technique generating holes in the lamella interior large micrometer-sized electron-transparent regions were obtained. However, this lamella displayed a higher thickness at the rim of ≥30 nm. Limiting factors for the observed thicknesses are discussed including ion damage depths, backscattering, and surface roughness, which depend on ion type, energy, current density, and specimen motion. Finally, sections cut by ultramicrotomy at low stroke rate and low set thickness offered vast, several tens of square micrometers uniformly thin regions of ～10-nm minimum thickness. As major drawbacks, we have detected a thin coating on the sections consisting of epoxy deployed as the embedding material and considerable nanoscale thickness variations.     

Specimen thickness dependence of hydrogen evolution during cryo‐transmission electron microscopy of hydrated soft materials

The evolution of hydrogen from many hydrated cryo‐preserved soft materials under electron irradiation in the transmission electron microscope can be observed at doses of the order of 1000 e nm^?2 and above. Such hydrogen causes artefacts in conventional transmission electron microscope or scanning transmission electron microscopy (STEM) imaging as well as in analyses by electron energy‐loss spectroscopy. Here we show that the evolution of hydrogen depends on specimen thickness. Using wedge‐shaped specimens of frozen‐hydrated Nafion, a perfluorinated ionomer, saturated with the organic solvent DMMP together with both thin and thick sections of frozen‐hydrated porcine skin, we show that there is a thickness below which hydrogen evolution is not detected either by bubble observation in transmission electron microscope image mode or by spectroscopic analysis in STEM electron energy‐loss spectroscopy mode. We suggest that this effect is due to the diffusion of hydrogen, whose diffusivity remains significant even at liquid nitrogen temperature over the length scales and time scales relevant to transmission electron microscopy analysis of thin specimens. In short, we speculate that sufficient hydrogen can diffuse to the specimen surface in thin sections so that concentrations are too low for bubbling or for spectroscopic detection. Significantly, this finding indicates that higher electron doses can be used during the imaging of radiation‐sensitive hydrated soft materials and, consequently, higher spatial resolution can be achieved, if sufficiently thin specimens are used in order to avoid the evolution of hydrogen‐based artefacts.     

Atom probe specimen preparation with a dual beam SEM/FIB miller

Dual beam scanning electron microscope/focused ion beam (SEM/FIB) methods complement electropolishing methods and enable specimens to be made from a wider range of materials. Several methods have been developed to fabricate specimens from different forms of materials, including thin ribbons, mechanically ground sheet and fine powders. In addition, FIB-based methods can be used in conjunction with electropolishing methods to improve the shape, surface finish and taper angle of specimens. Several lift-out (LO) methods have been developed for selecting specific microstructural features or other regions of interest such as phases, interfaces, grain boundaries, subsurface or implanted regions and interdendritic regions. These LO methods make use of an in situ nanomanipulator and platinum deposition to transfer and attach the lifted out volume to a post for final annular milling into a needle-shaped specimen. In order to improve the efficiency and to facilitate the LO procedure, some special specimen mounts that hold both the specimen and the support post at the appropriate working distance have been developed.     

Conventional and back-side focused ion beam milling for off-axis electron holography of electrostatic potentials in transistors

Off-axis electron holography is used to characterize a linear array of transistors, which was prepared for examination in cross-sectional geometry in the transmission electron microscope (TEM) using focused ion beam (FIB) milling from the substrate side of the semiconductor device. The measured electrostatic potential is compared with results obtained from TEM specimens prepared using the more conventional 'trench' FIB geometry. The use of carbon coating to remove specimen charging effects, which result in electrostatic fringing fields outside 'trench' specimens, is demonstrated. Such fringing fields are not observed after milling from the substrate side of the device. Analysis of the measured holographic phase images suggests that the electrically inactive layer on the surface of each FIB-milled specimen typically has a thickness of 100 nm.     

Cold-stage microscopy system for fast-frozen liquids

The least artifact-laden fixation technique for examining colloidal suspensions, microemulsions, and other microstructured liquids in the electron microscope appears to be thermal fixation, i.e., ultrafast freezing of the liquid specimen. For rapid-enough cooling and for observation in TEM/STEM a thin sample is needed. The need is met by trapping a thin layer ( approximately 100 nm) of liquid between two polyimide films ( approximately 40 nm thickness) mounted on copper grids and immersing the resulting sandwich in liquid nitrogen at its melting point. For liquids containing water, polyimides films are used since this polymer is far less susceptible to the electron beam damage observed for the commonly used polymer films such as Formvar and collodion in contact with ice. Transfer of the frozen sample into the microscope column without deleterious frost deposition and warming is accomplished with a new transfer module for the cooling stage of the JEOL JEM-100CX microscope, which makes a true cold stage out of a device originally intended for cooling specimens inside the column. Sample results obtained with the new fast-freeze, cold-stage microscopy system are given.     

Subject index to volume 10

The scanning transmission electron microscope with a field emission electron source operated at 100 kV allows X-ray microanalysis using electron probes as small as 1 to 2 nm. Measurements of the probe in a Vacuum Generators HB-501 STEM show that spherical aberration in the objective lens controls the probe size and shape at beam convergence half-angles of 10 mrad and greater typically used for X-ray microanalysis. A virtual objective aperture eliminates X-ray contributions from the probe-forming system, but must be aligned exactly to avoid asymmetrical broadening of the probe by spherical aberration. It is estimated that 5 nm X-ray spatial resolution can be achieved in low to medium atomic number materials. Even at this resolution however, probe broadening in the specimen controls the resolution; the main limitation is one of specimen preparation and a knowledge of the final specimen thickness. Determination of composition profiles near voids, dislocations and other individual defects in thin foils also requires a knowledge of the defect depth position and deconvolution of the probe and composition profiles.     

Deformation studies at sliding wear tracks in iron

Determinations have been made of strains on the surface and subsurface of specimens of high purity iron after different distances of sliding wear. The method involved the measurement of loss of intensity (contrast) of particular electron channeling lines obtained from small selected areas near the wear track. A calibration specimen, deformed plastically to a range of strain values, was used to relate the channeling line contrast loss to plastic strain. Strain maps lateral to the wear track and below the original surface were obtained after removing controlled thicknesses of iron by electropolishing. In all cases the maximum strain was found at the track center location at the surface and the strains decreased steadily with depth below the track. With a 50 g load the strains vanished at about 40 μm depth. Significant strains were found to exist outside the wear track boundaries. There was no indication of a soft or less-hardened surface layer in any of the specimens studied.     

Backscattered electron imaging and scanning transmission electron microscopy imaging of multi-layers

Experimental and theoretical results on image contrast of semiconductor multi-layers in scanning electron microscopy investigation are reported. Two imaging modes have been considered: backscattered electron imaging of bulk specimen and scanning transmission imaging of thinned specimens. The following main results have been reached. The image resolution of the multi-layers is, in both cases, defined by the probe size. The contrast, governed by density and atomic number differences, is affected by the size of the interaction volume in backscattered electron imaging and by the beam broadening in scanning transmission. Operating in the scanning transmission mode, the contrast of bright field images can be easily related to local variation in atomic number and density of the specimen while the dark field image contrast is strongly affected by electron beam energy, detector collection angles and specimen thickness. All these factors are able to produce contrast reversals that are difficult to explain without the support of a suitable simulation code.     

The quantitative characterization of SiGe layers by analysing rocking profiles in CBED patterns

Convergent beam electron diffraction is used for the quantitative determination of layer thickness, composition and strain in pseudomorphic Si/SiGe two- and three-layer systems grown by molecular beam epitaxy. By using plan-view specimens, we are able to avoid the influence of surface relaxation which generally complicates the determination of strains in cross-sectional specimens. For quantitative strain determination, rocking curves of Bragg lines in energy-filtered convergent beam electron diffraction patterns are analysed. The experimentally obtained rocking curves are compared with kinematical calculations by a computerized fit procedure. The resulting layer parameters are then further refined by a dynamical simulation. Results for the strains obtained with this technique are in good agreement with theoretical values. With this method layer thickness is measured down to monolayer precision. The accuracy of the strain analysis depends on the sequence and thickness of the layers.     

Quantitative X-ray microanalysis of biological specimens

Qualitative X-ray microanalysis of biological specimens requires an approach that is somewhat different from that used in the materials sciences. The first step is deconvolution and background subtraction on the obtained spectrum. The further treatment depends on the type of specimen: thin, thick, or semithick. For thin sections, the continuum method of quantitation is most often used, but it should be combined with an accurate correction for extraneous background. However, alternative methods to determine local mass should also be considered. In the analysis of biological bulk specimens, the ZAF-correction method appears to be less useful, primarily because of the uneven surface of biological specimens. The peak-to-local background model may be a more adequate method for thick specimens that are not mounted on a thick substrate. Quantitative X-ray microanalysis of biological specimens generally requires the use of standards that preferably should resemble the specimen in chemical and physical properties. Special problems in biological microanalysis include low count rates, specimen instability and mass loss, extraneous contributions to the spectrum, and preparative artifacts affecting quantitation. A relatively recent development in X-ray microanalysis of biological specimens is the quantitative determination of local water content.     

The use of a phi(rhoz) model for simultaneous determination of composition and thickness in analytical transmission electron microscopy

X-ray microanalysis using transmission electron microscopy (TEM) offers the possibility to perform quantitative analysis with high spatial resolution. Disadvantages are its low accuracy and the problem of preparing very thin specimens and the thin film standards, needed for the analysis and calibration. To calculate composition from the measured X-ray intensities, a peak-ratio method is usually applied, based on the thin film method by Cliff and Lorimer (Proceedings of the Fifth European Congress on Electron Microscopy, 1972, p. 140). We however, applied an entirely different approach, calculating the composition using a full matrix correction method based upon a phi(rhoz) matrix correction model as they are commonly used in EPMA measurements up to 40 kV. The validity of the model under TEM conditions was checked by performing bulk analyses on AlNi and AlTi samples and thin film analyses on an AlNi TEM specimen. In principle, both thin and thick specimens as well as light elements can be analysed this way. No major changes to the TEM set-up or simplifications to the model are needed, only an accurate beam current meter is required.