Comparison of detection limits for EELS and EDXS

Simultaneous measurements have been made in the scanning transmission electron microscope to determine relative microanalytical sensitivities for electron energy loss (EELS) and energy-dispersive x-ray spectroscopies (EDXS). A photodiode-array parallel detector was used for EELS, and an ultrathin window Si(Li) detector subtending a solid angle of 0.18 sterad was used for EDXS. Energy loss spectra were acquired in the first or second difference mode to reduce channel- to-channel gain variations and to detect weak signals on large backgrounds. Both EELS and EDXS data were analyzed using multiple least squares procedures to fit reference spectra; this gave not only the fitting coefficients but also their standard deviations. Results showed that EELS is the method of choice for analysis of elements with Z 10. For higher atomic numbers (10 Z 25) EDXS is preferable to EELS if the K edge is used. However, EELS has the higher sensitivity for most elements in this atomic number range if the L23 edge is analyzed. For sodium detection limits are approximately a factor of 10 better with EELS. As the atomic number increases this factor becomes smaller although EELS still has a significant advantage for phosphorus. The occurence of the "white-line" resonance at the L23 edge for Z > 20 boosts the advantage of EELS for calcium but at iron EELS and EDXS have about the same sensitivity. Calculated estimates of the relative sensitivities for EELS and EDXS are in reasonable agreement with the experimental data. of soft x-rays by the detector window.