Correlative electron tomography and atom probe tomography have been carried out successfully on the same region of a commercial 8090 aluminum alloy (Al-Li-Mg-Cu). The combination of the two techniques allows accurate geometric reconstruction of the atom probe tomography data verified by crystallographic information retrieved from the reconstruction. Quantitative analysis of the precipitate phase compositions and volume fractions of each phase have been obtained from the atom probe tomography and electron tomography at various scales, showing strong agreement between both techniques.

Extended abstract of a paper presented at Microscopy and Microanalysis 2004 in Savannah, Georgia, USA, August 1–5, 2004.

Extended abstract of a paper presented at Microscopy and Microanalysis 2004 in Savannah, Georgia, USA, August 1–5, 2004.

Extended abstract of a paper presented at Microscopy and Microanalysis 2004 in Savannah, Georgia, USA, August 1–5, 2004.

The length scales on which materials microstructures are being formed, grown, and even designed are becoming increasingly small and increasingly three-dimensional. For such complex structures two-dimensional transmission electron microscopy (TEM) analysis is often inadequate and occasionally misleading. One approach to this problem is the modification of electron tomography techniques, developed for structural biology, for use in materials science. Energy-Filtered (EF) TEM elemental distribution images approximate to true projections of structure, and, as such, can be used to reconstruct the three-dimensional distribution of chemical species. A sample holder has been modified to allow the high tilt (±60°) required for tomography and a semiautomatic acquisition script designed to manage energy-loss acquisition. Tilt series data sets have been acquired from two widely different experimental systems, Cr carbides in 316 stainless steel and magnetite nanocrystals in magnetotactic bacteria, demonstrating single- and multiple-element tomography. It is shown that both elemental maps and jump-ratio images are suitable for reconstruction, despite the effects of diffraction contrast in the former and thickness changes in the latter. It is concluded that the image contrast, signal, and signal-to-noise ratio (SNR) are key to the achievable reconstruction quality and, as such, the technique may be of limited value for high energy loss/small inelastic cross section edges.