W.J. de Ruijter

Accurate measurements of mean inner potential of crystal wedges using digital electron holograms

Abstract The mean inner potential of a solid is a fundamental property of the material and depends on both composition and structure. By using cleaved crystal wedges of known angle, combined with dogital recording of off-axis electron holograms and with theoretical calculations of dynamical effects, the mean inner potential of Si (9.26±0.08 V), MgO (13.01±0.08 V), GaAs (14.53±0.17 V) and PbS (17.19±0.12 V) is measured with high accuracy of about 1%. Dynamical contributions to the phase of the transmitted beam are found by Bloch wave calculations to be less than 5% when the crystal wedges are titled away from zone-axis orientations and from major Kikuchi bands. The accuracy of the present method is a factor of 3 better than previously achieved by reflection high-energy electron diffraction and electron interferometry. The major causes of uncertainty were specimen imperfections and errors in phase measurement and magnification calibration.

Detection limits in quantitative off-axis electron holography

Abstract The phase of an electron wave is altered by electric and magnetic fields as it passes through a specimen. This phase change can be accurately quantified from off-axis electron holograms acquired using a slow-scan CCD camera, and small changes can be observed over small dimensions. Expressions for the precision of the phase estimate, which is limited by shot noise, have been developed. These include most of the real experimental parameters. It is found that the typical precision of practical phase measurements is better than π/100 for spatial resolutions of 1–3 nm, in good agreement with the theoretical optimal phase precision. In order to attain such small errors the effects of geometric distortion, which can introduce phase differences of up to π, must be carefully corrected.

Methods to measure properties of slow‐scan CCD cameras for electron detection

New ways to measure detection properties of slow‐scan charge‐coupled device (CCD) cameras suitable for electron microscopy, mainly based on the statistics of single‐electron events, are discussed. The experiments concentrate on the newly introduced Gatan 679 slow‐scan CCD camera. It has been established that for this instrument (if equipped with a thin YAG scintillator) 20% of the true intensity is recorded at the highest detectable spatial frequency (Nyquist frequency), limited mostly by the modulation transfer function of the YAG scintillator. The detection quantum efficiency is 0.45 for single 100‐kV electrons and 0.15 for single 400‐kV electrons, and is approaching unity for intensities higher than ten electrons. Furthermore, nonlinearity of the response is smaller than deviations in the image intensity due to shot noise. Examples are presented illustrating the detection properties of slow‐scan CCD cameras for electron imaging, which also include high dynamic range and negligible geometric distortion.

Practical autoalignment of transmission electron microscopes

Abstract We present a working method for automatic correction of beam misalignment, defocus and objective lens astigmatism based upon computer-induced beam tilts which result in image shifts that are measured and used to adjust the microscope. The performance of the method has been tested on 100 and 400 keV microscopes, and it proves to work reliably over a wide range of magnifications for both weak phase and amplitude contrast specimens. The performance of the alignment procedure is instrumentation, specimen and magnification dependent. The reproducibility for alignment on the coma-free axis was measured to be less then 0.1 mrad at magnification M = 500 000 and less then 2 mrad at M = 10 000. The reproducibility for focusing and stigmation was measured to be less then 3 nm at M = 500 000 and less then 50 nm at M = 10 000.

Autotuning of a TEM using minimum electron dose

Abstract A new method is proposed to measure and correct defocus, astigmatism and beam tilt misalignment of a transmission electron microscope (TEM) automatically. The method is applicable to low-dose, high-resolution electron microscopy if the specimen may be described as a weak-phase object. Defocus and astigmatism are estimated from two different pairs of images. These are obtained, respectively, by tilting the illumination in two perpendicular directions over two equal but opposite angles. Beam tilt misalignment is estimated from three images, one formed without beam tilt and two with equal but opposite tilt angles. In all cases use is made of the cross-spectra of the images involved. An optimal measuring strategy has been designed and tested with simulations for the estimation of defocus and astigmatism. The design shows that the TEM should be tuned in two steps. The first step adjusts the defocus and astigmatism to 1.5 (in Sch), with a precision of 1. The second step adjusts the TEM with high precision. The simulations show that defocus and astigmatism can be estimated with a precision of 0.02 from images with an average intensity of 6500 electrons per nm 2 . First measurements show a promising agreement of measured and simulated cross spectra.

Measurement of lattice-fringe vectors from digital HREM images: experimental precision

Abstract The practical measurement of lattice-fringe spacings and angles recorded in digital high-resolution electron micrographs is evaluated experimentally. The method is based on a statistical estimation procedure and involves computer analysis of reciprocal-space parameters. This work concentrates on the analysis of images recorded with slow-scan CCD cameras, but alternative methods of image pick-up are also briefly considered. The method has been successfully applied to images recorded with electron doses smaller than 1eA 2 and for sample dimensions as small as 8Aacross. The practical precision depends on specimen characteristics, electron dose and the size of the measurement area and is in the range of 0.001–0.05Afor lattice spacings, and 0.1°–0.5° for lattice-plane angles. Finally, the technique is demonstrated in studies of a catalyst system, of a reduced surface oxide phase and of TiO 2 and TiN particles.

Applications of electron holography to the study of interfaces

Abstract Electron holography has been applied to a variety of layered structures to assess its usefulness for supplying information about composition profiles across heterogeneous interfaces. The phase of the exit-surface electron wave, which to a first approximation is dependent upon the mean inner potential and the specimen thickness, was extracted from electron holograms acquired from suitable cross-sectional multilayer specimens. Line profiles from the reconstructed phase images were analyzed to obtain information about interface diffuseness and layer width with a spatial resolution of about 5 A. Using spatial averaging parallel to the interface, increased measurement precision was obtainable in some special cases. Differences in interdiffusion widths between Mo-Si and Si-Mo interfaces in an Mo/Si multilayer structure were confirmed, and the width of the amorphous layer at Si 3 N 4 grain boundaries was measured to be about 12 A. It was concluded that off-axis electron holography represented a useful complementary technique for characterizing interfaces.

Progress towards quantitative high-resolution electron microscopy

Abstract In recent years there has been a great upsurge of applications involving quantitative high-resolution electron microscopy, in particular comparing experimental micrographs with image simulations for determination of defect structures. Emphasis has been given to the determination of experimental parameters, the utilization of slow-scan CCD cameras for digital recording and extraction of quantitative structural and chemical information. More attention to surface cleanliness is needed to improve signal quality and the possibility of electron irradiation damage should not be overlooked. Issues related to adoption of a reliability of R -factor are briefly discussed.

Quantitative Electron Holography

Electron holography is necessarily a two-step process. First, the object and reference waves must be coherently superimposed in recording the electron hologram. Second, reconstruction of the hologram is required in order to extract the desired phase and/or amplitude of the electron wave which has passed through the object. Historically, hologram recording has been carried out photographically, and the subsequent reconstruction has been done optically.261,451 In addition to inevitable time delays, the use of these techniques in electron holography can result in unsuspected artefactual information, thus motivating the utilization of alternative recording and processing methods that offer greater reliability and the possibility of more accurate quantification.