c

. Fig. 23.22  Particles deposited on a thin (~20-nm) carbon film supported on a copper grid shown at various magnifications. a Nominal 200 X, field width 351 by 263 micrometers; b nominal 1 kX, field width 70 by 53 micrometers; c nominal 5 kX, field width 14 by 11 micrometers

The Importance of Beam Placement

The placement of the beam on the particle relative to the EDS detector position can have a strong effect on the measured spectrum. For particles of intermediate size where the interaction volume is contained within the particle, placing the

. Fig. 23.23  Schematic illustration of the effect of beam placement on a particle on the length of the absorption path to the EDS detector

beam on the side of the particle away from the EDS results in an extended absorption path. The generated X-rays must pass through a large mass of the particle to reach the detector. This effect is illustrated schematically in .  Fig. 23.23, where absorption paths for X-rays generated near the maximum penetration of beam electrons in the particle are compared for three beam positions: top center, and at positions directly facing and away from the EDS. Because absorption follows an exponential dependence on path length and becomes increasingly significant for low energy photons below approximately 4 keV, the low energy portion of the spectrum will show the strongest absorption effects. Spectra recorded at the top of a 5-μm-diameter particle and on the side away from the EDS are compared in .  Fig. 23.24 demonstrating a factor of two difference in the intensity measured at the

energy of Mg K-L2,3 (1.254 keV). For a beam placed on the curved side of the particle facing the EDS, the absorption

path length is actually reduced relative to the top center, leading to extra emission from the low energy photons, creating

about 30 % excess for Mg K-L2,3, as shown in the comparison of spectra in .  Fig. 23.25. .  Figure 23.26 shows the impact of

beam placement on the accuracy of quantitative analysis, as discussed below. It is thus critical that the analyst is always aware of the relative position of the EDS in the SEM image when choosing locations to analyze. A reliable way to locate the EDS is to record an X-ray spectrum image and examine selected X-ray intensity maps, as shown in .  Fig. 23.27. The general rule that the source of the apparent illumination

Counts

10000

1000

100

Photon energy (keV)

Counts

appears to come from the position of the detector immediately reveals the true EDS position. For this particular particle, the eclipsing effect of the extended absorption path on the backside of the particle is readily apparent, even in the relatively energetic Co K-L2,3 (6.930 keV) intensity map.

Overscanning

“Overscanning” is a strategy by which the beam is continuously scanned over the particle (or rough surface, heterogeneous material, etc.) while the EDS spectrum is being collected. If the magnification control of the SEM is continuously variable, then the size of the scan raster can be adjusted to bracket the particle, minimizing the fraction of the time that the beam spends on the surrounding substrate, as shown in .  Fig. 23.28 (yellow box). Alternatively, the scan raster can be adjusted to fill as much of the particle image as possible while remaining within the bounds of the particle, thus avoiding direct beam placement on the substrate, as shown in

.  Fig. 23.28 (green box).

23 While useful for gaining qualitative information on the constituents of a particle, overscanning only has utility if the particle (or other target object) is homogeneous. Overscanning discards valuable information on any possible inhomogeneous

structure within the particle. As described below, overscanning is NOT a means by which an “average” composition can be found by quantitative microanalysis.

23.6.3\ X-ray Spectrum Imaging:

Understanding Heterogeneous

Materials

X-ray spectrum imaging (XSI) involves collecting a complete EDS spectrum at every pixel location visited by the scanned beam. When applied to particle analysis, the XSI provides the analyst with an abundance of information which can be recovered by post-collection processing of the XSI datacube. Composition heterogeneity down to the single pixel level can be detected and interpreted. The particle X-ray intensity images shown in .  Fig. 23.27a, b were extracted from an XSI, and the localization of Ti and Mo in inclusions is immediately obvious. Solidification dendrites

are outlined in the Ti K-L2,3 map, a feature that is not obvious in any of the other elemental intensity maps. Software

tools enable the analyst to extract spectra that are representative of individual components of the microstructure, such

Counts

. Fig. 23.26  Quantitative analysis­ performed at the static beam locations indicated; relative errors observed for the normalized concentrations (oxygen calculated by stoichiometry)

EDS

. Fig. 23.27  a SEM BSE image and X-ray intensity maps for Al K-L2,3, Cr K-L2,3, and Co K-L2,3. Note the shadowing effect of the particle thickness causing attenuation of X-ray intensity. b X-ray intensity maps for

Mo L3-M4,5, and Ni K-L2,3, and a color overlay with Mo L3-M4,5 = red, Ti K-L2,3 = green, and Ni K-L2,3 = blue. Note the localization of Ti and Mo in inclusions, and the solidification dendrites outlined by Ti