This book (Practical Electron Microscopy and Database) is a reference for TEM and SEM students, operators, engineers, technicians, managers, and researchers.

=================================================================================

In general, there are many drawbacks induced by the specimen contamination in EM analysis, for instance:
         i) Preferentially absorb low-energy X-rays emitted from the specimen. This effect is especially harmful when analyzing light elements.
         ii) Increase the X-ray and EELS background and thus reduce the signal-to-background ratio.

EDS ZAF mode and Phi-Rho-Z mode are reasonably used in SEM-EDS analysis, while Cliff-Lorimer mode is generally applied in TEM-EDS measurement.

For SEM-EDS analysis, in most light-element cases, it is better that a low accelerating voltage (e.g. ≤10 kV) is used since the depth of penetration of the electrons is smaller and the amount of absorption is reduced if a lower accelerating voltage is used. In the SEM-EDS measurements on bulk materials, beam energy of 20 keV is a good compromise between both the requirements of obtaining adequate overvoltages and of minimizing absorption in the specimen. However, such beam energy may still mask the presence of light elements that only produce X-ray energies below 2 keV since their high overvoltage (e.g. U >10) will lead to deep penetration of the incident electrons into the specimen, and thus the absorption can be as high as 50-99%. To avoid missing possible light elements present at low concentrations, spectrum acquisitions should be repeated at beam energies in the range of 5-10 keV.

For TEM-EDS measurements, the k_AB factors are experimentally determined from well-investigate, homogeneous standards. The quality of EDS analyses depends significantly on the accuracy of the k_AB values so that it is very important these values are carefully quantified using several standards for each element. For the elements with atomic numbers greater than 12, the k_AB factors can now be determined with an error in the range of 1 to 4%, while the determination of k_AB factors for light (smaller Z) elements is normally less accurate.

The k_AB factors for light elements (low energy X-rays) are normally affected strongly by some detector parameters such as the absorption coefficients and the thicknesses of the Be, Au and Si layers in the detector mentioned on page 3777. On the other hand, the x-ray yields for light elements are much lower than those for heavy elements. Therefore, EDS measurement becomes less efficient for lighter atoms (say Z < 30). However, EDS measurements of light elements and many 3d transition elements can still be accomplished at accelerating voltage of 1.5 kV [1].

[1] E. D. Boyes, Adv. Mat. 1998, 10, 1277.