Faults and foibles of quantitative scanning electron microscopy/energy dispersive x-ray spectrometry (SEM/EDS)

Abstract

Scanning electron microscopy with energy dispersive x-ray spectrometry (SEM/EDS) is a powerful and flexible elemental analysis method that can identify and quantify elements with atomic numbers > 4 (Be) present as major constituents (where the concentration C > 0.1 mass fraction, or 10 weight percent), minor (0.01≤ C ≤ 0.1) and trace (C < 0.01, with a minimum detectable limit of ≈± 0.0005 - 0.001 under routine measurement conditions, a level which is analyte and matrix dependent ). SEM/EDS can select specimen volumes with linear dimensions from ≈ 500 nm to 5 μm depending on composition (masses ranging from ≈ 10 pg to 100 pg) and can provide compositional maps that depict lateral elemental distributions. Despite the maturity of SEM/EDS, which has a history of more than 40 years, and the sophistication of modern analytical software, the method is vulnerable to serious shortcomings that can lead to incorrect elemental identifications and quantification errors that significantly exceed reasonable expectations. This paper will describe shortcomings in peak identification procedures, limitations on the accuracy of quantitative analysis due to specimen topography or failures in physical models for matrix corrections, and quantitative artifacts encountered in xray elemental mapping. Effective solutions to these problems are based on understanding the causes and then establishing appropriate measurement science protocols. NIST DTSA II and Lispix are open source analytical software available free at www.nist.gov that can aid the analyst in overcoming significant limitations to SEM/EDS.