Quantitative X-Ray Microanalysis of Heterogeneous Materials Using Monte Carlo Simulations

Monte Carlo simulations are used to obtain new results of x-ray microanalysis of sample types frequently encountered in practical analytical situations such as a vertical layer embedded in a homogeneous matrix and a spherical particulate deposited on a substrate. The simulations show that a 10-nm layer of boron in a steel matrix can be imaged using backscattered electrons and detected using x-ray microanalysis with a field emission scanning electron microscope even with an electron beam energy equals to 20 keV and also that these simulations can be useful to estimate the optimum acceleration voltage to perform such analyses. For a carbon spherical particulate located on the top of a gold substrate, it is shown that x-ray emission and electron backscattering are a strong function of the diameter of the particulate and also of the electron beam energy. Finally, a new method to determine the thickness of a thin film deposited on a substrate is proposed that does not require the measurement of the beam current. That technique can also be used for a spherical particulate deposited on a substrate.     

High-Spatial-Resolution Low-Energy Electron Beam X-Ray Microanalysis

 Performing X-ray microanalysis at beam energies lower than those conventionally used (< 10 keV) is known to significantly improve the spatial resolution for compositional analysis. However, the reduction in the beam energy which reduces the X-ray interaction diameter also introduces analytical difficulties and constraints which can diminish the overall analytical performance. This paper critically assesses the capabilities and limitations of performing low beam energy, high spatial resolution X-ray microanalysis. The actual improvement in the spatial resolution and the reduction in the X-ray yield are explored as the beam energy is reduced. The consequences for spectral interpretation, quantitative analysis and imaging due to the lower X-ray yield and the increased occurrence of X-ray line overlaps are discussed in the context of currently available instrumentation.     

X-Ray Microanalysis of Thin Surface Films and Coatings

 The paper gives an overview of the present knowledge in the field of X-ray analysis of surface films and more generally stratified specimens. The aim of the paper is not to report the details and formulas of the available quantitative procedures, but to concentrate on the general ideas and orders of magnitude illustrating the capability and limits of the method, and on the optimal adaptation of the operating conditions to every particular problem. The various specific pitfalls which can be encountered are pointed out, in particular the fluorescence effects when using high-energy X-ray lines, or the anomalies due to chemical bonding, absorption uncertainties, and contamination effects when soft radiations are employed.     

X-Ray Microanalysis of Real Materials Using Monte Carlo Simulations

Monte Carlo simulations have been used to obtain new results that aid in the microanalysis of sample types frequently encountered in practical analytical situations such as rough surfaces and embedded inclusions. It is shown that the hypothesis that the peak to background is constant on a rough surface is not always true, especially for x-ray lines of low energy and that the peak to background ratio is very sensitive to the electron probe diameter. It is also shown that for embedded inclusions that the shape of the (z) curves differs significantly from the (z) curves of bulk homogeneous materials.     

Indirect Measurement of Biologically Important Compounds by Means of X-Ray Microanalysis

 The main objective of our work was to investigate the possibility and usefulness of indirect methods in X-ray microanalysis for the quantification of biologically important compounds. Metallothionein-like proteins (MT-like proteins) from kidney and liver, rich in sulfur were chosen as an indicator of heavy metal presence in cells and their environment. Tissues from goldfish (Carassius auratus gibelio) were sampled after short and prolonged periods of exposure to Co⁺², CrO₄ ⁻², Pb⁺², Cu⁺² and control treatment and prepared for histochemical staining for peroxidated thiolate groups. Commonly used –S–S– bonds dye (Nitro Red) was replaced with iodine atoms and they were quantified at L line by means of X-ray microanalysis combined with SEM. After fish treatments with heavy metal solutions changes in MT-like proteins and in I atom contents were expected. There was statistically significant decrease in MT-like proteins level in kidney after lead treatment (Pb/C = 0.62). In liver a statistically significant increase in MT-like proteins concentration was observed after chromium, cobalt and lead ions treatment in comparison to control animals. The following ratios were noted: 3.04 for Cr/C, 2.18 for Co/C and 2.10 for Pb/C. Our finding indicates that the method of indirect measurement of MT-like proteins in fish and other animal tissues is possible. The concentration of iodine atoms is above their detection level by EDS and their changes are possible to identify. During histochemical procedures it is worth taking into account sample preparation methods which might disturb the quality and quantity of the analysed material.     

Energy Dispersive X-Ray Fluorescence Analysis and X-Ray Microanalysis of Medieval Silver Coins

 Austrian medieval silver/copper coins were investigated at their surfaces by energy dispersive X-ray fluorescence analysis (EDXRF) and at the cross-sections by X-ray microanalysis in the scanning electron microscope (SEM/EDX) in order to estimate the error occurring when corroded objects of art and archaeology are analyzed on the surface by non-destructive methods. Additionally, Ag/Cu-standards were treated in diluted sulphuric acid and the depletion of copper on the surface was measured by EDXRF. By calculating the ratio of the Ag-K/Ag-L intensity the process of blanching could be studied.     

Monte Carlo Simulation of Electron Transport and X-Ray Generation. II. Radiative Processes and Examples in Electron Probe Microanalysis

Theoretical models for Monte Carlo simulation of radiative processes, i.e. bremsstrahlung and characteristic x-ray emission, are presented. Possible strategies for simulating electron transport are briefly described. For mechanisms involving energy loss and angular deflections, difficulties for strict implementation of accurate numerical differential cross sections still remain due to the strong correlations between these variables. Practical solutions for the case of inelastic collisions and bremsstrahlung emission are described. Comparisons of simulation results with experimental data for several problems of interest in electron probe microanalysis are presented.     

Microcalorimeter Detectors and Low Voltage SEM Microanalysis

The development of the microcalorimeter energy-dispersive X-ray spectrometer (µ-cal EDS) offers a significant advancement in X-ray microanalysis, especially for electron beam instruments. The benefits are especially pronounced for low voltage (5kV) X-ray microanalysis in the field emission scanning electron microscope (FE-SEM) where the high energy resolution of the µ-cal EDS minimizes the peak overlaps among the myriad of K, L, M and N lines in the 0–5keV energy range. The availability of L- and M-shell X-ray lines for microanalysis somewhat offsets the absence of X-ray lines traditionally used above 5keV energy. The benefits and challenges of the µ-cal EDS will be discussed, including P/B ratio for characteristic X-rays, collection angle, count rate capability and the impact of polycapillary X-ray optics on microanalysis.     

Electron Probe Microanalysis of HfO2 Thin Films on Conductive and Insulating Substrates

Quantitative electron probe microanalysis of highly insulating materials is a complicated problem, partially solved by coating samples with grounded thin conductive layers or using novel scanning electron microscopy (SEM) techniques, such as low-voltage and/or variable pressure SEM. In this work, some problems of quantitative X-ray microanalysis of thin HfO₂ films, in particular the possibility to determine mass thickness correlated to the density of the layer material, are discussed. For comparison, Al₂O₃, Ta₂O₅ and TiO₂ films grown onto both semiconductive Si and insulating quartz substrates were also analysed. All the films studied were synthesized by atomic layer deposition method.     

Pile-Up Correction for Improved Accuracy and Speed of X-Ray Analysis

Despite advances in electronic design, pile-up artefacts are still common in EDX spectra and can lead to false element identifications, inaccurate correction for peak overlap and losses of counts that give poor quantitative results. With the capability to do spectrum imaging there is increasing temptation to work at count rates far beyond the correction capability of pile-up inspection electronics. Fundamental limitations due to noise are explained and a new correction procedure is described that implements a comprehensive channel-by-channel correction for pile-up. Practical examples are given that demonstrate the range of application of the new algorithm and show that, with correction, count rates at least 4 times higher can be used with no sacrifice in performance.     

Microanalysis of Glass Containing Alkali Ions

 The common problems connected with alkali ion migration during EPMA were studied on glasses containing nearly all possible alkali ions (Na, K, Rb, Cs). Binary silica glasses were prepared by melting from a very pure batch in Pt crucible. The glasses were carefully polished using alcohol to prevent surface corrosion by water and they were stored in vacuum. The specimens were coated with carbon layers approximately 30-nm thick and exposed to a 50- keV electron beam of 100 μm diameter. It was found that all alkali ions migrate under the electron beam, but the rate of the migration depends on the current density. The decay curves (characteristic X-ray intensity versus time) are similar in shape in all cases. The decay curve shows two transport regimes, the first being linear-like, the second being the exponential-like. The first transport regime busts into the rapid alkali migration after a time known as the incubation period. The period is in general longer for the larger-alkali ions size. It was found that even large rubidium and caesium ions migrate inside the glass with the same mechanism as sodium and potassium ions. While for K, Rb, and Cs ions the incubation periods were observed under the suitable experimental conditions, binary glass containing Na exhibits no observable incubation period. Except for the binary Na₂O + SiO₂ glass, the suitable experimental conditions for reliable quantitative EPMA can be found.     

Design and Application of a Set of Vitreous Standards for EDS-SEM Microanalysis of Melting Shop Slags

Non-metallic inclusions in steel are formed due to interactions between liquid steel and its environment in the melting shop (refractories and slags). Particles not trapped in the slag remain in the cast steel, impairing steel properties in service. Inclusion composition varies and certified standards covering the whole compositional range are unavailable.The present study describes the preparation of a set of homogeneous, glassy and crystalline standards with a wide compositional range for use in analysing non-metallic steel inclusions. For this, materials were designed based on knowledge of the compositional range of interest in the Al₂O₃–SiO₂–CaO–MgO–CaF₂ system, which is the matrix of the most frequent steel inclusions. The composition and microstructural homogeneity of the reference materials was verified.     

X-Ray Emission from Thin Films on a Substrate – Calculation and Experiments

 Monte Carlo simulation of electron transport in solids is widely used in electron microscopy, spectroscopy and microanalysis. The reliability of physical models incorporated in a Monte Carlo code is usually checked by comparing with experimental results. Elastic or inelastic collisions are usually considered as the basic interactions of electrons with atoms. In our Monte Carlo code the single scattering model is employed for simulation of X-ray emission from thin films of Au on the Si substrate. The electron beam energy was in the range 10–30 keV, the take-off-angle was 40°. The simulated values of X-ray production were calculated in our Monte Carlo code using several models of ionisation cross-sections. For the emitted intensities the depths of inelastic collision and X-ray absorption were taken into account, then the k-ratios were calculated. These data were compared with experimental values of k-ratios calculated from X-ray intensities of Au M and Au L characteristic lines. We followed mainly the dependence of the k-ratios of the film element on film thickness. The film thickness was in the range 0.05–1 μm. Reasonably good agreement was found for dependences of X-ray intensity on film thickness in the whole energy range and for both lines, especially for Powell’s model of the ionisation cross-section.     

Electrically Cooled SiLi Detectors for Application in X-Ray Equipment

The results of the development of electrically cooled SiLi detectors are presented for applications in X-ray equipment. The characteristics of such detectors compare well with those of the detectors cooled by liquid nitrogen, and to obtain these characteristics only 60 W of electric energy is needed instead of liquid nitrogen.     

Energy Dispersive X-Ray Spectrometry by Microcalorimetry for the SEM

 Analytical X-ray spectrometry for electron beam instruments has advanced significantly with the development of the microcalorimeter energy dispersive X-ray spectrometer (μcal EDS). The μcal EDS operates by measuring the temperature rise when a single photon is absorbed in a metal target. A cryoelectronic circuit with electrothermal feedback and a superconducting transition edge sensor serves as the thermometer. Spectral resolution approaching 4.5 eV for high energy photons (6000 eV) and 2 eV for low energy photons below 2000 eV has been demonstrated in energy dispersive operation across a photon energy range from 250 eV to 8 keV. Spectra of a variety of materials demonstrate the power of the μcal EDS to solve practical problems while operating on a scanning electron microscope platform.     

Quantitative Bulk and Trace Element X-Ray Mapping Using Multiple Detectors

X-ray mapping using energy dispersive spectroscopy or wavelength dispersive spectroscopy is a very popular characterisation tool for determining the elemental distribution in materials. Furthermore, quantitative X-ray mapping has become a very powerful technique enabling reliable quantitative results that can be an order of magnitude better than traditional analysis. Quantitative X-ray mapping is also far superior to regions of interest X-ray maps where low levels of an element or elemental overlaps are present. The one major drawback with X-ray mapping is the time required to obtain a high resolution X-ray map with good statistics at low levels of concentration. The use of multi-detectors, and just developed dual turret detectors for X-ray mapping, allows improvement in performance at low levels without compromising quantification quality and precision of traces, even in the presence of overlaps. However, for quantitative X-ray mapping to work properly, the characteristics of each detector must be accurately determined so that the final quantification of the individual detectors can be summed. To accomplish this effectively, the full spectrum at each pixel for each energy dispersive detector should be saved. As a final check for consistency between detectors, a technique was developed that involves assigning a different red-green-blue colour for each detector for the same element. By doing this, when we combine the three maps of the same element, we should obtain a grey scale map that indicates total correlation between the three detectors at the most critical final stage of quantification. To reduce contrast noise and further improve the quality of quantitative X-ray mapping images, a filter referred to as a “speckle filter” has been developed that allows the eye to see a more correct elemental concentration relationship.     

WDXRF, EPMA and SEM/EDX Quantitative Chemical Analyses of Small Glass Samples

Quantitative chemical analysis is a powerful tool for the classification and discrimination of small glass fragments in forensic and archaeometric analytical issues. In the choice of the analytical method the small dimensions and the irregular shape of the fragments must be accounted for. Several accurate and highly sensitive instrumental techniques successfully employed for the chemical analysis of small glass samples (AAS, ICP-AES and FAAS) are destructive, and therefore unacceptable for these applications. X-ray analytical techniques such as WDXRF, EPMA and SEM/EDX provide means for rapid, non destructive and sensitive multi-elemental analysis of glass fragments. In this study the analytical sensitivity and accuracy of WDXRF, EPMA and SEM/EDX for the analysis of small commercial glass fragments are investigated. The results are compared and discussed in order to establish a relation between the minimum size of glass samples and the accuracy and sensitivity of the analytical methods.     

Detector Calibration and Measurement of Fundamental Parameters for X-Ray Spectrometry

Energy-dispersive X-ray spectrometry offers the opportunity for a fast and reliable determination of the composition of a specimen. For fundamental parameter based quantification, detailed knowledge of the X-ray generation cross sections and of the detection system’s efficiency is required. The detection efficiency is determined comparing the measured and calculated spectrum of undispersed synchrotron radiation (SR) from the electron storage ring BESSY II in the spectral range from 0.1 keV to 100 keV. Alternatively, monochromatized SR in the spectral range from 0.1 keV to 60 keV is used to determine the detection efficiency with a typical relative uncertainty of 1% to 2% by direct comparison with a reference detector. Employing well-calibrated detectors and monochromatized SR of well-known radiant power and high spectral purity, fluorescence yields have been investigated and resonant Raman scattering was studied as an example of a basic effect creating spectral background.     

Electron Probe Microanalysis of Nitrogen in Glasses and Glass Ceramics

 A homogeneous (phosphate glass) and a heterogeneous (glass ceramic) material were chosen to develop a method for electron microprobe analysis of nitrogen in glasses and glass ceramics. The metaphosphate (50 mole-% P₂O₅, 25 mole-% BaO, 25 mole-% Rb₂O) base glass and the silicate glass (composition of the base glass: 6.9 mole-% Al₂O₃, 21.3 mole-% MgO, 47.3 mole-% SiO₂, 5.1 mole-% CaO, 19.4 mole-% AlN) were melted and cooled down to room temperature. In a second step, the two materials were nitrided by applying dry NH₃ in a special tube furnace. Up to 5 wt.-% of nitrogen could be introduced. The ability of energy dispersive (EDX) as well as wavelength dispersive (WDX) electron probe microanalysis to analyze low contents of light elements in combination with highly concentrated heavy elements was tested both for glasses and for glass ceramics. Measuring conditions had to be optimized to get reliable analytical results as well as to avoid radiation damage of the glasses which may occur especially in the case of wavelength dispersive X-ray analysis. The results were compared with two different analytical methods: inductive coupled plasma analysis and a commercial nitrogen-oxygen-analyzer (the specimen is decomposed by heating and the released gases are analyzed spectroscopically (O) and electrically (N)).     

Modification of a Field Emission Source SEM For Quantitative X-Ray Microanalysis

Abstract

We have modified a field emission source SEM in an attempt to obtain a quantitative x-ray microanalysis capability. This report describes how emission current instabilities, characteristic of cold field emission cathodes, were effectively circumvented. The technique involves integrating the electron beam current and terminating x-ray data acquisition at a selected value of integrated current.

Working well in the current ranges available on most electron beam instruments, the scheme may be universally applicable to any charged beam instrument in which some event is, or can be made, dependent on the integrated beam current.