John T. Armstrong

A new EPMA method for fast trace element analysis in simple matrices

Abstract It is well known that trace element sensitivity in electron probe microanalysis (EPMA) is limited by intrinsic random variation in the X-ray continuum background and weak signals at low concentrations. The continuum portion of the background is produced by deceleration of the electron beam by the Coulombic field of the specimen atoms. In addition to the continuum, the background also includes interferences from secondary emission lines, “holes” in the continuum from secondary Bragg diffraction, non-linear curvature of the wavelength-dispersive spectrometer (WDS) continuum and other background artifacts. Typically, the background must be characterized with sufficient precision (along with the peak intensity of the emission line of interest, to obtain the net intensity for subsequent quantification), to attain reasonable accuracy for quantification of the elements of interest. Traditionally we characterize these background intensities by measuring on either side of the emission line and interpolate the intensity underneath the peak to obtain the net intensity. Instead, by applying the mean atomic number (MAN) background calibration curve method proposed in this paper for the background intensity correction, such background measurement artifacts are avoided through identification of outliers within a set of standards. We divide the analytical uncertainty of the MAN background calibration between precision errors and accuracy errors. The precision errors of the MAN background calibration are smaller than direct background measurement, if the mean atomic number of the sample matrix is precisely known. For a simple matrix and a suitable blank standard, a high-precision blank correction can offset the accuracy component of the MAN uncertainty. Use of the blank-corrected-MAN background calibration can further improve our measurement precision for trace elements compared to traditional off-peak measurements because the background determination is not limited by continuum X-ray counting statistics. For trace element mapping of a simple matrix, the background variance due to major element heterogeneity is exceedingly small and high-precision two-dimensional background correction is possible.

CALCZAF, TRYZAF and CITZAF: The Use of Multi-Correction-Algorithm Programs for Estimating Uncertainties and Improving Quantitative X-ray Analysis of Difficult Specimens

Much work has been done in the last 50 years in developing comprehensive correction algorithms for quantitative electron microbeam analysis. A number of correction methods – both theoretically and empirically based – have been used, incorporating or parameterizing various compilations of physical constants like mass absorption coefficients and mean ionization potentials. There is no single, universally accepted correction method used for microbeam analysis and the correction programs employed by the commercial instrument manufacturers give different results for the same input data.

Strategies for Low Accelerating Voltage X-ray Microanalysis of Sub-Micrometer Features with the FE-EPMA

X-ray microanalyses of sub-micrometer features require that a lower accelerating voltage be used. This reduction in the accelerating voltage reduces the penetrating distance of the beam electrons and thereby reduces the analytical volume. Two general strategies have been proposed for achieving the smallest analytical volume. The first approach involves significantly reducing the accelerating voltage to minimize the electron interaction volume. For many element systems, the optimum accelerating voltage is in the 5-8kV range, assuming a field emission electron gun is used for the analysis. A lower kV will reduce the penetrating depth of the electrons further. However with lower kV, the diameter of the electron beam becomes a critical factor in determining the analytical area, and with a lower kV, the beam diameter becomes larger. Mostly this comes from the higher required beam current used due to the lower X-ray production rate at the lower kV. This low-kV strategy also typically requires a different set of X-ray lines be used, since some of the more commonly used X-ray lines are not generated at these lower accelerating voltages. However, it can produce a very small analytical volume (Fig. 1a).

Comparative Performance of SDD-EDS and WDS Detectors for Quantitative Analysis of Mineral Specimens: The Next Generation Electron Microprobe

With the development of silicon drift detectors (SDD), energy dispersive x-ray emission spectrometry is now capable of making compositional measurements of comparable precision as those that have been routinely obtained by wavelength dispersive spectrometers on electron microprobes [e.g., 1]. Large area silicon drift detectors, when placed near the specimen, subtend significantly larger solid angles than WDS detectors and so can make these measurements at considerably lower beam currents, making them particularly well-suited for use at the conditions ordinarily employed for scanning electron microscopes. The significantly lower resolution of SDDs compared to WDS spectrometers (although significantly improved over Si(Li) EDS detectors, particularly at low energies) limits their ability to make high precision measurements of minor element concentrations and precludes their determination of trace element concentrations. Additionally there are a number of severe x-ray peak overlaps that require a higher degree of energy/wavelength resolution than is possible with SDDs, despite the high quality of available EDS spectral deconvolution software. For these cases, WDS spectrometers are demonstrably superior in performance.

A New EPMA Method For Fast Trace Element Analysis In Simple Matrices

Traditionally Electron Probe Micro Analysis (EPMA) has relied upon precise characterization of the continuum intensities adjacent to the emission line of interest using Wavelength Dispersive Spectrometry (WDS) for determination of the background under the peak. Recent improvements including new hardware designs with large area Bragg crystals, new software methods implementing exponential and polynomial interpolations to more accurately characterize the curvature of the background, and aggregated spectrometer signals to improve sensitivity, have enabled the EPMA to attain detection limits as low as 2 to 3 PPM in some materials.[1]

Improving Trace Element Analysis Precision By Not Using Off-Peak Measurements

It is well known that trace element sensitivity in Electron Probe Micro Analysis (EPMA) is limited by intrinsic random variations in the x-ray continuum background produced by the deceleration of the electron beam by the Coulombic field of the specimen atoms. Typically, this continuum must be characterized with sufficient precision (along with the peak intensity of the emission line of interest), in order to attain reasonable sensitivity for the elements of interest. Generally we characterize these intensities by measuring on either side of the emission line and interpolate the intensity underneath the peak to obtain the net intensity.

What's Still Missing with the Fluorescence Corrections and Should We Care?

Of the corrections employed for quantitative electron microprobe analysis, the correction for effects of fluorescence by characteristic x-ray lines has changed the least since the early re-parameterization of Castaing’s original correction by S.J.B. Reed in 1965 [1], and formal correction for the effects of continuum fluorescence is still seldom applied with no consensus parameterization employed by the standard software provided for EDS and WDS analysis systems. With regard to the characteristic fluorescence correction, part of the reason for the lack of change is the brilliance of Reed’s formulation of a useful set of equations. However, although he noted in his original paper that a number of simplifications had been made due to lack of fundamental measurement data and to aid in calculation and he and others later proposed substitution of better data sets and equations [e.g., 2,3], most correction programs still retain the original simplified expressions. Alternate methods have been proposed based on use of parameterized expressions for the primary φ(ρz) distribution coupled with numerical integration of the fluorescing and fluoresced x-ray path lengths [e.g., 4-6] and from Monte Carlo calculations [e.g., 7-10]. These alternate expressions can be used or modified to calculate correction for fluorescence in thin films, layered specimens, particles, inclusions, and across phase boundaries. But none of these are yet in common use and most still involve assumptions or simplifications to avoid the very extensive set of calculations a truly rigorous fluorescence correction would require. Moreover there are virtually no sets of measurements of k-ratios in well characterized systems where fluorescence effects are important to determine if the various correction algorithms actually work [e.g. 11].

Use of Multilayer Correction Procedures to Improve the Quantitative Electron Microprobe Analyses of Coated Insulating Specimens

Evaporating or sputtering thin layers of conducting materials onto specimens for electron microscopy and microbeam analysis have long been employed to improve electrical and thermal conductivity, signal to noise ratio, and spatial resolution [1]. However, such surface coatings affect the relative emitted x-ray intensities and complicate quantitative analyses. Traditionally, when performing quantitative microprobe analysis, surface coatings are limited to thin layers of light elements (usually carbon), and samples and standards are coated at the same time or with thickness monitors to insure that both have the same overlayer thickness [2]. When such procedures are not followed, the errors in quantitative analysis can easily exceed 5%, and even when carefully undertaken, multi-percent level variations in analytical results can still be caused by spatial variations in coating thickness [3,4].