Log-polynomial background subtraction in energy-filtered TEM

Abstract

An inverse power relation is widely used to model the background under inner shell ionization edges in both electron energy-loss spectrometry (EELS) and energy-filtered transmission electron microscopy (EFTEM). Proper background subtraction is necessary to obtain correct core-loss integrated intensities in EELS or elemental maps in EFTEM. However, the empirical inverse power relation often does not accurately model the background when losses at slightly lower energies interfere with the edge of interest, or in general, when energy losses are less than ∼200 eV. In such cases, the background of EELS spectra has been successfully fitted as a linear least-squares fit to a polynomial, usually a quadratic of the form: log(I) = A + BX + CX2, where I = intensity and X = log(energy loss). In the present study, this alternative background model, the “log-polynomial,” has been applied to pre-edge images for background subtraction from post-edge energy-filtered images.