Effective L-Series Mass Absorption Coefficients for EDS

Abstract

Energy dispersive X-ray spectroscopy has insuf®cient resolution to separate the individual lines of low energy L-series peaks. However, the mass absorption coef®cients for L and L radiation differ signi®cantly for elements with atomic numbers between 21 and 32. Effective mass absorption coef®-cients for the entire L-shell emission were determined by measuring the variation of the X-ray intensities emitted from pure element standards, as a function of the accelerating voltage, and ®tting the experimental data with a theoretical curve using the XMAC software. These experimentally determined mass absorption coef®cients were compared with average values calculated on the basis of the theoretical line intensities, taking into account the primary vacancy generation and the radiationless Coster-Kronig transitions.The component lines of the L-series are not completely resolvable by energy dispersive spectrometry for X-ray energies less than approximately 3 keV. This results in the L and L peaks appearing as a single peak with an energy shift $ 10 eV with respect to the exact L line position. Use of the undeconvoluted L peak for quantitative analysis may therefore lead to large errors in cases where mass absorption differs signi®cantly for the individual lines [1±3]. The elements principally affected are those lying between atomic numbers 21 (scandium) and 32 (germanium), for which the energy of the L line is slightly greater than the ionization energy of the L 3 subshell. Due to the proximity of the absorption edge the mass absorption coef®cients for L radiation in these elements are up to a factor of six higher than for L.An additional source of error in analysis constitutes the uncertainty in the reported values of the mass absorption coef®cients for the L lines of the elements from Ti to Zn [4 ± 6]. The self absorption coef®cient for atoms bonded with atoms of other elements can furthermore vary noticeably from that in the pure element because the electron transition probabilities and X-ray absorption properties are in¯uenced by modi®cations in the structure of the valence band caused by alloying. The 3d transition elements, in which the valence band is incompletely ®lled, are most strongly affected and the self absorption of Ni L in alloys such as Ni-Al or Ni-Zn, for example, is consequently weaker than in the pure metal [7,8].Because of the uncertainties in analysis using soft X-rays, a method has been developed to determine precise values for mass absorption coef®cients by measuring the variation of the emitted line intensity with accelerating voltage. The experimentally measured values of the X-ray emission rate per unit beam current are compared to a theoretical curve computed by the XMAC software [8,9], which is based on the XPP model of X-ray generation. Using an iterative procedure to optimize the ®t a value for the mass absorption coef®cient can be obtained.Two alternative approaches may be used to perform quantitative energy dispersive analysis using compound L-series peaks. In the ®rst, the L intensity is e...

Determination of Mass Attenuation Coefficients of Th, U, Np, and Pu for Oxygen KαX-Rays Using an Electron Microprobe

“…An earlier version of XMAC was used by Pouchou and Pichoir to obtain the MACs in specific cases where tabulated values at that time were affected by large uncertainties, namely for the X-ray lines of C, N, and O in highly absorbing elements, for some X-ray lines close to an absorption edge and for the L-lines of transition metals (Sc to Cu) which are affected by resonant absorption (see Pouchou & Pichoir, 1988, for further details). XMAC was also used to determine the MACs for L-lines (L α + L β ) of transition metals measured with an energy-dispersive spectrometer (Rickerby & Wachter, 2000), and more recently, it has been used to determine the MACs for metal L-lines of Fe silicides (Gopon et al, 2013) and of Ni silicides (Llovet et al, 2016), where absorption appears to abnormally depend on silicide composition. A similar program was developed and used by Mackenzie (1991) to improve the analysis of the oxygen content of bismuth strontium copper oxide samples.…”

An inter-laboratory comparison of EPMA analysis of alloy steel at low voltage

“…For a given X-ray line and element, the mentioned programme determines the MAC by i) considering it as a free parameter in the XPP model, and ii) fitting the predictions of the model to measurements of the X-ray intensity as a function of beam energy. In the case of Cr, the Lα and Lβ peaks appear as a single (Lα+Lβ) line, therefore, the use of the L doublet for analysis will lead to an additional source of uncertainty, as the MAC for the Lα line differs significantly from that of the Lβ line (see e.g., Ref [11]). To assess the effect of using different MACs on the resulting concentrations, the k-ratios obtained by laboratory 1 for sample P13 were re-evaluated using the PAP model with the MACs shown in table 13.…”