Dark-field image contrast in transmission scanning electron microscopy: Effects of substrate thickness and detector collection angle

Abstract

An annular dark field (ADF) detector was placed beneath a specimen in a field emission scanning electron microscope operated at 30kV to calibrate detector response to incident beam current, and to create transmission images of gold nanoparticles on silicon nitride (SiN) substrates of various thicknesses. Based on the linear response of the ADF detector diodes to beam current, we developed a method that allowed for direct determination of the percentage of that beam current forward scattered to the ADF detector from the sample, i.e. the transmitted electron (TE) yield. Collection angles for the ADF detector region were defined using a masking aperture above the detector and were systematically varied by changing the sample to detector distance. We found the contrast of the nanoparticles, relative to the SiN substrate, decreased monotonically with decreasing inner exclusion angle and increasing substrate thickness. We also performed Monte Carlo electron scattering simulations, which showed quantitative agreement with experimental contrast associated with the nanoparticles. Together, the experiments and Monte Carlo simulations revealed that the decrease in contrast with decreasing inner exclusion angle was due to a rapid increase in the TE yield of the low atomic number substrate. Nanoparticles imaged at low inner exclusion angles (<150mrad) and on thick substrates (>50nm) showed low image contrast in their centers surrounded by a bright high-contrast halo on their edges. This complex image contrast was predicted by Monte Carlo simulations, which we interpreted in terms of mixing of the nominally bright field (BF) and ADF electron signals. Our systematic investigation of inner exclusion angle and substrate thickness effects on ADF t-SEM imaging provides fundamental understanding of the contrast mechanisms for image formation, which in turn suggest practical limitations and optimal imaging conditions for different substrate thicknesses.