Detection and quantification of low energy, low level electron energy loss edges

The detection and quantification of low energy, low level edges presents several problems for Electron Energy Loss Spectrometry. These edges can have very low "jump" ratios making recognition of the presence of the edge difficult. For quantitative analysis the onset energy of the edge must also be identified. Both of these problems can be overcome by the routine use of a second difference (top-hat) filter [1]. The next major problem is that of background subtraction. The standard a E- r curve fit results in over subtraction of the background for spectra of a Li K edge [1]. The same curve constrained to fit the background at high energy as well as low energy can result in under subtraction of the background. Modifying the standard fit to a(E+ m)-r- cE and constraining the curve to fit at higher energy shows promise and is easy to implement. The "c" term allows for variation of the power law with energy however the most critical term appears to be the "m" term. For very low level edges another problem is encountered in the subtraction of the background associated with the electron detection system. The detector background level for a yttrium- aluminium- garnet/ Photodiode Array detector has been observed to vary in a non-linear way with incident electron intensity. This affects both the pre-edge background curve fitting and the results of edge count integration