Quantitative analysis of electron spectroscopic imaging series

Reactive multilayers examined by HRTEM and plasmon EELS chemical mapping

We examine chemical mapping of reaction phases in a Cu-Al multilayer system using low-loss electron energy loss spectroscopy spectrum imaging and image spectroscopy techniques. The sensitivity of the plasmon peak position and shape to various crystal structures and phases is exploited using postprocessing of spectra into second derivative plasmon maps and line scans. Analytical transmission electron microscopy is complemented by studies of the orientation relationship of the multilayer system using high-resolution electron microscopy of interfaces and selected area diffraction. The techniques have been applied to the Cu-Al multilayer sample and sharply bound epitaxial phases are found, before and after heat treatment.

Tomographic spectroscopic imaging; an experimental proof of concept

Recording the electron energy loss spectroscopy data cube with a series of energy filtered images is a dose inefficient process because the energy slit blocks most of the electrons. When recording the data cube by scanning an electron probe over the sample, perfect dose efficiency is attained; but due to the low current in nanoprobes, this often is slower, with a smaller field of view. In W. Van den Broek et al. [Ultramicroscopy, 106 (2006) 269], we proposed a new method to record the data cube, which is more dose efficient than an energy filtered series. It produces a set of projections of the data cube and then tomographically reconstructs it. In this article, we demonstrate these projections in practice, we present a simple geometrical model that allows for quantification of the projection angles and we present the first successful experimental reconstruction, all on a standard post-column instrument.

Optimized FIB silicon samples suitable for lattice parameters measurements by convergent beam electron diffraction

The aim of this paper is to check the effect of artefacts introduced by focused ion beam (FIB) milling on the strain measurement by convergent beam electron diffraction (CBED). We show that on optimized silicon FIB samples, the strain measurement can be performed with a sensitivity of about 2.5 x 10(-4) which is very close to the theoretical one and we conclude that FIB preparation can be suitable for such measurements in microelectronic devices. To achieve this, we first used CBED and electron energy loss spectroscopy (EELS) which provide a procedure permitting an exact knowledge of the sample geometry, i.e. the thickness of both amorphous and crystalline layers. This procedure was used in order to measure the FIB-amorphized sidewall layer. It was found that if the FIB preparation is optimized one can reduce this amorphous layer down to around 7 nm on each side. Secondly different preparation techniques (cleavage, Tripodtrade mark and FIB) permit to check if the surface damaged layer introduced by FIB influences the strain state of the sample. Finally, it was found that the damaged layer does not introduce measurable strain in pure silicon but reduces appreciably the quality of the CBED patterns.

EFTEM spectrum imaging at high-energy resolution

This paper deals with the application of high-energy resolution EFTEM image series and the corrections needed for reliable data interpretation. The detail of spectral information gained from an image series is largely determined by the intrinsic energy resolution. In this work we show that energy resolution values of as low as 0.8 eV in spectra extracted from EFTEM image series can be obtained with a small energy-selecting slit. At this resolution level aberrations of the energy filter, in particular the non-isochromaticity, can no longer be neglected. We show that the four most prominent factors for EFTEM image series data correction--spatial drift, non-isochromaticity, energy drift and image distortion--must not be treated independently but have to be corrected in unison. We present an efficient algorithm for this correction, and demonstrate the applied correction for the case of a GaN/AlN multilayer sample.

Automated spatial drift correction for EFTEM image series

Energy filtering transmission electron microscopy (EFTEM) is a widely used technique in many areas of scientific research. Image contrast in energy-filtered images arises from specific scattering events such as the ionization of atoms. By combining a set of two or more images, relative sample thickness maps or elemental distribution maps can be easily created. It is also possible to acquire a whole series of energy-filtered images to do more complex data analysis. However, whenever several images are combined to extract certain information, problems are introduced due to sample drift between the exposures. In order to obtain artifact-free information, this spatial drift has to be taken care of. Manual alignment by overlaying and shifting the images to find the best overlap is usually very accurate but extremely time consuming for larger data sets. When large amounts of images are recorded in an EFTEM series, manual correction is no longer a reasonable option. Hence, automatic routines have been developed that are mostly based on the cross-correlation algorithm. Existing routines, however, sometimes fail and again make time consuming manual adjustments necessary. In this paper we describe a new approach to the drift correction problem by incorporating a statistical treatment of the data and we present our statistically determined spatial drift (SDSD) correction program. We show its improved performance by applying it to a typical EFTEM series data block.

Some trends in microscope image processing

The present review tries to identify some trends among the multitude of ways followed by image processing developments in the field of microscopy. Nine topics were selected. They cover the fields of: signal processing, statistical analysis, artificial intelligence, three-dimensional microscopy, multidimensional microscopy, multimodality microscopy, theory, simulation and multidisciplinarity. A specific topic is dedicated to a trend towards semi-automation instead of full automation in image processing.

Plasmon energy mapping in energy-filtering transmission electron microscopy

In this paper, we demonstrate the two-dimensional mapping of plasmon energies by energy-filtering transmission electron microscopy. The maps are obtained from a series of energy-filtered images in the plasmon energy region. Examples are shown for a nano-crystalline Si-B-C-N ceramic. This material contains SiC and Si(3)N(4) grains as well as intergranular regions composed of hexagonal BN (h-BN) and turbostratic carbon (t-C). The different phases can be clearly identified by their specific plasmon energies. An energy resolution of < or =0.1eV is achieved. In addition, the plasmon map of an amorphous carbon film is used to visualize the non-isochromaticity of the Corrected Omega filter (90 degrees filter) of the SESAM2. A procedure is proposed for the correction of the non-isochromaticity.

A sub-50meV spectrometer and energy filter for use in combination with 200kV monochromated (S)TEMs

A high-energy resolution post-column spectrometer for the purpose of electron energy loss spectroscopy (EELS) and energy-filtered TEM in combination with a monochromated (S)TEM is presented. The prism aberrations were corrected up to fourth order using multipole elements improving the electron optical energy resolution and increasing the acceptance of the spectrometer for a combination of object area and collection angles. Electronics supplying the prism, drift tube, high-tension reference and critical lenses have been newly designed such that, in combination with the new electron optics, a sub-50 meV energy resolution has been realized, a 10-fold improvement over past post-column spectrometer designs. The first system has been installed on a 200 kV monochromated TEM at the Delft University of Technology. Total system energy resolution of sub-100 meV has been demonstrated. For a 1s exposure the resolution degraded to 110 meV as a result of noise. No further degradation in energy resolution was measured for exposures up to 1 min at 120 kV. Spectral resolution measurements, performed on the pi* peak of the BN K-edge, demonstrated a 350 meV (FWHM) peak width at 200 kV. This measure is predominantly determined by the natural line width of the BN K-edge.

Comparison of detectability limits for elemental mapping by EF-TEM and STEM-XEDS

The analytical sensitivity in terms of the signal-to-noise ratio (SNR) was investigated for elemental mapping by a transmission electron microscope equipped with an energy filter (EF-TEM) and a scanning transmission electron microscope with an X-ray energy dispersive spectrometer (STEM-XEDS). To compare the detectability limits of the elemental maps by the two techniques, homogeneous Cu-0.98+/-0.34 wt% Mn and Cu-4.93+/-0.49 wt% Mn thin specimens were used. Elemental maps can be considered as either an image or a spectrum. Therefore, the detectability limits of the elemental maps were characterized by the spectral SNR. To evaluate the detectability limits of the elemental maps with statistical confidence limits such as 1 sigma, 2 sigma and 3 sigma, the SNR values were reviewed from the statistical point of view. In STEM-XEDS mapping, the spectral SNR values improve as the specimen thickness increases since the signal intensity increases. Conversely, the spectral SNR in EF-TEM mapping is maximized at a certain thickness and then reduces as the thickness increases. To compare the two mapping techniques with regard to the analytical sensitivity, a method to estimate the minimum mass fraction (MMF) from measured signal and background intensities was developed. In this experimental approach, the MMF value can be evaluated by selecting the appropriate SNR value corresponding to the statistical confidence limits. In comparing the estimated MMF values from the two mapping approaches, EF-TEM mapping can be more sensitive than STEM-XEDS mapping up to specimen thicknesses <20-30 nm in the 1 sigma confidence limit and < approximately 50 nm in the 3 sigma limits. However, as the specimen thickness increases, the XEDS maps provide better detectability limits in the Cu-Mn dilute alloy specimens.

Mapping the chemistry in nanostructured materials by energy-filtering transmission electron microscopy (EFTEM)

Energy-filtered transmission electron microscopy (EFTEM) can be used to acquire elemental distribution maps at high lateral resolution within short acquisition times, which makes it quite efficient for a detailed characterization of nanostructures, as illustrated with examples concerning a nanostructured substituted La-based cermet compound and a nanoscale multilayer. In the first example, we show how phases in a rapidly cooled substituted LaNi5 can be visualized by recording jump ratio images. Secondly, EFTEM was capable of imaging individual nanoscale layers in a magnetic multilayer consisting of 2 nm terbium and 3 nm iron.

Image-spectroscopy--I. The advantages of increased spectral information for compositional EFTEM analysis

The acquisition of a series of energy-filtered TEM images over the energy-loss range of interest creates a three-dimensional data set comprising both spatial and spectral information. Such an image-series contains energy-loss information not available with conventional two- or three-window methods, allowing standard techniques for quantitative EELS analysis to be applied to extracted 'image-spectra'. The increase in spectral information enables improved ionisation edge background extrapolation and interactive image-spectrum analysis to be performed. In this paper, the many advantages of the image-spectroscopy approach are outlined by reference to an example of elemental segregation in an AlZnMgCu alloy.

Contamination and the quantitative exploitation of EELS low-loss experiments

Quantitative exploitation of the low-loss domain of electron energy loss spectra is based on an accurate determination of the corresponding signal intensity profile. This signal can be erroneous and contains artefacts as a result of sample contamination in the microscope, for example. The consequences of contamination on the signal intensity of the low-loss spectra are discussed. In the case of a carbonaceous contamination, a simple additional spurious signal can be considered, as has been demonstrated in the case of a Si single crystal, a highly oriented pyrolytic graphite (HOPG) and a strontium titanate single crystal (SrTiO3). The linear variation of the rate of contamination with time allows the implementation of a simple method based on the subtraction of the spurious signal in order to correct for the contamination effect. The relative errors induced by the carbonaceous contamination on the determination of the optical properties of SrTiO3 are estimated.

Elemental maps from EFTEM images using two different background subtraction models

Acquisition of a great number of energy-filtered images in a TEM (EFTEM) around the characteristic signal with a low energy-selecting slit allows display of the electron energy loss (EEL)-spectrum of regions of interest (ROIs) of a sample. These EEL-spectra can be submitted to the different treatments already in use for electron energy loss spectroscopy (EELS). In particular, it is possible to fit the experimental background with different mathematical models, using images acquired below and above a characteristic ionization edge. After this fitting, elemental maps can be computed by subtraction of the extrapolated/interpolated background from the characteristic images. In this work, we compared two mathematical models for background fitting-the Egerton power law and the log-polynomial law. We studied the low-energy region (40-150 eV) and a higher-energy region (350-600 eV) with the aid of software for interactive processing of EFTEM image series that we developed. The analyzed elements were the constitutive elements: iron, phosphorus, nitrogen, and oxygen in several biological materials. Two analytical TEMs, one equipped with a post-column and the other with an in-column spectrometer, were used. Our experimental results confirm that the power law is very sensitive to the value of the energy loss of the pre-edge images when the background is computed by extrapolation. The log-polynomial model is less sensitive than the power law model to the value of the energy loss of the pre-edge images in the low energy region. For the oxygen K edge at 535 eV, it gives the best fit when it is combined with the interpolation method. The use of programs that facilitate the handling of EFTEM image series, and the controlled calculation of the background under the characteristic images, represent a step forward in the generation of elemental maps.

An approach to an objective background subtraction for elemental mapping with core-edges down to 50 eV: description, evaluation and application

To image the distribution of a specific element in a specimen with an energy filtering TEM, the element-unspecific background under the core-edge has to be subtracted. The most commonly used procedure is the three-window power-law method leading to considerable systematic errors for low-energy core-edges. Here a new method is described which can be considered as a generalized difference method. Characteristic examples for element detection in biological specimens using this method are shown. The background under the core-edge can be described by one or two pre-edge windows as a polynome of third order. This function can be deduced from specimen areas that are not known to contain the element or from a second specimen used as a standard. Control experiments showed that background subtraction for on-overlapping core-edges in the low-loss region (50-200 eV) needs two pre-edge images, whereas at higher-energy losses (> 300 eV) only one pre-edge image is necessary. With the method described, objective elemental mapping becomes possible even for edges at 50-100 eV. This was proven for the M2,3-edge of iron at 60 eV. The detection of phosphorous was possible with a signal-to-noise ratio five times higher than when using the three-window method. Preliminary data showed that it should be possible to detect calcium with only one image before the edge.

Quantitative electron spectroscopic imaging studies of microelectronic metallization layers

The combination of focused ion beam (FIB) sample preparation and quantitative electron spectroscopic imaging is an ideal tool for the investigation of layered structures used in microelectronic metallization schemes. In the present work, Si3N4/Cu/Si3N4/SiO2/Si and Al/TiN/Ti/SiO2/Si metallization layers produced by physical vapour deposition are investigated. We apply series of energy filtered images in the low loss region for a mapping of the sample thickness which makes it possible to refine the parameters of the FIB process. We also show how series of energy filtered images in the core loss region can be used to obtain elemental distribution images and chemical bonding information on these samples on a nanometre scale. For materials with a small grain size and/or a strong variation in Bragg orientation, the intensity distribution of the elemental map is strongly influenced by the superimposed Bragg contrast. This detrimental effect can be reduced greatly by using hollow cone illumination, as is demonstrated for polycrystalline Cu. One striking feature observed in Cu layers prepared with FIB is strong, regularly arranged contrast variations caused by subsurface defects in the Cu grains. We suppose that these defects are a consequence of a strong interaction of Ga atoms from the FIB with Cu.