Electron Spectroscopic Tomography for Materials Science

Abstract

Tomographic 3D-reconstructions from tilt series of projections differ in two major aspects when moving from biochemistry (e.g.) to materials science applications: (I) Crystalline inorganic materials fail the weak phase image formation approximation and suffer from dynamic and nonlinear contrast contributions (which depend for both CTEM and HREM on atomic lattice plane orientation). (II) exposure tolerance increases by two orders of magnitude. We therefore propose energy filtered tomography with inelastic electrons (electron spectroscopic imaging, ESI) as a new method for 3D-reconstruction of free or buried subsurface nanostructures or nanoparticles. Energy filters are not used here to improve SNR by removing inelastic background , but to provide the necessary projection relationship between a 3D concentration distribution and the (processed) image intensity. Ideally, a 4-dimensional data space is recorded for a single tilt axis tomography experiment, D=D(δE, Θ,x,y), (with 5E as energy loss, Θ as tilt, and (x,y) as image coordinates). The material has to withstand exposures of T = t × Ntilt × Nloss (t = 5...20sec per image, Ntilt >12 directions, Nloss = 5....20 energy loss values).