Quantitative Chemical Mapping with Low-Voltage EDS Spectrum Imaging and Multivariate Statistical Analysis (MSA)

Abstract

One of the advantages of performing X-ray microanalysis at low (≤5 kV) operating voltages is that high spatial resolution (≤250 nm) chemical maps of the specimen can be acquired. Spectrum imaging, where a full spectrum is acquired for each pixel in a two-dimensional array, provides the most comprehensive method of characterization, as long as the sampling density (pixel size) is sufficiently smaller than the spatial resolution. Multivariate statistical analysis (MSA) methods are effective in reducing the large (typically ∼10 MByte) raw spectrum images to the modest (typically ∼100 kByte) data files that contain all of the statistically significant information of interest about the specimen. Preliminary analysis of a cross section of a computer chip from a major semiconductor company was previously performed using limited spectrum imaging capabilities available with the 4pi X-ray mapping module, which allowed for simultaneous acquisition of only 48 channels. MSA of images acquired with only a 960 eV portion of the spectrum containing the Al-K, Si-K and W-M lines showed that excellent contrast between the Si- and W-rich regions of the specimen could be achieved in spite of the strong overlap between Si-Kα and W-Mα (∼34 eV separation).