By performing EELS in conjunction with the Z-contrast imaging techniqu
e in the scanning transmission electron microscope (STEM), detailed in
formation on the composition, chemistry and structure of materials can
be obtained with atomic resolution and sensitivity. This unprecedente
d resolution can be achieved by using an atomic resolution Z-contrast
image to first identify structural features of interest and then to po
sition the electron probe over the feature for spectral acquisition. T
his method greatly limits the number of spectra that need to be acquir
ed from a given specimen, thus reducing total acquisition time and spe
cimen stability problems. In addition, a key advantage of this methodo
logy is that the collection conditions for the spectrum can be tailore
d to produce an incoherent, atomic resolution spectrum that can be cor
related directly with the image. This means the image can be used as a
reference atomic structure for theoretical modeling of the spectral f
ine-structure. Using multiple scattering analysis, the reference struc
ture can be modified to reproduce the experimental spectrum, thus givi
ng a 3-dimensional structural determination that is sensitive to singl
e atom vacancies and impurities. In this paper, the application of thi
s combined EELS and Z-contrast technique to various materials problems
is described. In the case of the MBE growth of CdTe on Si, the combin
ed atomic resolution techniques allow the diffusion of Te into the sil
icon substrate to be identified and a novel graphoepitaxial grow-th me
chanism to be identified. For nanoscale iron particles used as a fuel
additive to reduce/enhance soot formation during combustion, measureme
nt of the distribution of iron oxidation states within the particles p
ermits a mechanism for soot formation to be proposed. The application
of the multiple scattering analysis techniques to the study of grain b
oundaries is described for tilt boundaries in TiO2 and SrTiO3. In both
cases, the multiple scattering analysis gives information not present
in the image and allows the 3-dimensional structure of each boundary
to be determined.