Quantitative Electron Probe Microanalysis of Fe at Low Accelerating Voltage Using the Lα and Lβ X-ray Lines

Abstract

For more than six decades, electron probe microanalysis has been employed to determine the composition of unknown samples at ~micron-size domains. The introduction of the field emission gun (FEG) has allowed the reduction of the electron beam size. Electron imaging by microprobe can now reach a sub-micron resolution allowing the clearer imaging of sub-micron features such as inclusions in larger phases. To improve the X-ray lateral spatial analytical resolution, however, the interaction volume of the electrons inside the sample also needs to be reduced. This can only be done by decreasing the accelerating voltage from typically 15 kV to 5-7 kV, reducing the typical size of the electron interaction volume from ~1-2 µm to ~0.2-0.7 µm (depending upon material composition). However, by reducing the accelerating voltage, the commonly utilized X-ray lines (the Kα lines) of many of the elements of the periodic table up to Z=34, cannot be excited and thus cannot be used for analysis. Instead of using the K lines, the L lines of these elements must be used. However, for some elements, these L X-ray lines are not well known and complex chemical and physical effects can occur, making their use difficult in quantitative analysis. This is especially the case for the transition metals with the main L lines (and for many of the rare-earth elements with the main M lines). Several effects can influence the emission of X-rays produced by an electronic transition involving outer electronic shells. In the case of the transition metals, the Lα and Lβ X-ray lines involve a partially filled 3d electron shell which is subject to strong bonding and hybridization with electrons from the neighboring atoms. This results in a change of the state of the outer electronic orbitals and then changes the energy of the electrons and the electronic transition probabilities associated with these shells. Consequently, the value of the different atomic parameters (mass absorption coefficient, fluorescence yield, Coster-Kronig factor, …) not only depend of the studied element but also of the composition and nature of the specific material, both of the unknown and of the standard used to generate the k-ratio.Our work was focused on the transition metal iron as it is a commonly encountered element in many materials of interest. Gopon et al [1] showed that the Fe Ll line had some utility, but its weak intensity was a drawback. Here we investigate the possible utility of Fe Lα and Lβ. The atomic parameters associated with the emission and absorption of the Fe Lα and Lβ X-ray lines were reevaluated on 15 iron-bearing standards of well-known composition (silicates and oxides). Trends are deduced from these measurements in order to predict the value of the atomic parameters in relation to the composition of the material and thus allowing their use for correct quantitative analysis. The mass absorption coefficients (MAC) of the Fe Lα and Lβ X-ray lines were first evaluated. For each standard, the L X-ray spectra of these lines were recorded with three waveleng...

Metal impact and vaporization on the Moon's surface: Nano‐geochemical insights into the source of lunar metals

“…To provide the combination of chemical and spatial resolution required to analyze lunar npFe 0 , we have employed low-voltage/high-resolution electron probe microanalysis (HR-EPMA) coupled with atom probe tomography (APT) to characterize npFe 0 obtained from Apollo 16 lunar regolith (sample 61501,125). While EPMA allows the identification and fully quantified elemental analyses of the larger (~0.5 to 1 lm) metallic components (Gopon et al, 2013;Hughes et al, 2018;Moy & Fournelle, 2017), APT offers the unique ability to simultaneously characterize the trace, minor, and major elements present at near atomic scale resolution, and has the potential to determine the isotopic compositions of sub-micrometer-scale phases in three dimensions (Daly et al, 2018;Gopon et al, 2017Gopon et al, , 2020Gopon, Douglas, et al, 2022;Gopon, Douglas, Auger, et al, 2019;Valley et al, 2014).…”