12 1 An Introduction to EELS
was no strong focusing lens, the instrument lacked spatial resolution (the beam
diameter at the specimen was about 10 μm) and was not used extensively because
similar spatial resolution and better energy resolution could be obtained by infrared
spectroscopy.
1.3.1 Energy-Selecting (Energy-Filtering) Electron Microscopes
Instead of recording the energy-loss spectrum from a particular region of spec-
imen, it is sometimes preferable to display a magnified image of the specimen
(or its diffraction pattern) at a selected energy loss. This can be done by utilizing
the imaging properties of a magnetic field produced between prism-shaped pole-
pieces, as first demonstrated by Castaing and Henry (1962) at the University of
Paris. Like the normal unfiltered TEM image, a plasmon-loss image was found
to contain diffraction contrast due to differences in elastic scattering, but in suit-
able specimens it also conveyed “chemical contrast” that was useful for identifying
different crystallographic phases (Castaing, 1975). Installed in various laborato-
ries, the Castaing–Henry filter was also used to record spectra and images from
inner-shell energy losses (Colliex and Jouffrey, 1972; Henkelman and Ottensmeyer,
1974a; Egerton et al., 1974). At the University of Toronto, Ottensmeyer reduced the
aberrations of his filter by curving the prism edges, a design that was eventually
incorporated into a TEM by the Zeiss company.
In order to maintain a straight electron-optical column, the Castaing–Henry filter
uses an electrostatic mirror electrode at the electron gun potential, an arrange-
ment not well suited to high-voltage microscopes. For their 1-MeV microscope at
Toulouse, Jouffrey and colleagues adopted the purely magnetic “omega filter,” based
on a design by Rose and Plies (1974). In Berlin, Zeitler’s group improved this sys-
tem by correcting various aberrations, resulting in a commercial product: the Zeiss
EM-912 energy-filtering microscope (Bihr et al., 1991). The omega filter has since
been incorporated into TEM designs by other manufacturers.
An alternative method of energy filtering is based on the scanning transmis-
sion electron microscope (STEM). In 1968, Crewe and co-workers in Chicago built
the first high-resolution STEM and later used an energy analyzer to improve their
images of single heavy atoms on a thin substrate. At the same laboratory, Isaacson,
Johnson, and Lin recorded fine structure present in the energy-loss spectra of amino
acids and nucleic acid bases and used fading of this structure as a means of assessing
electron irradiation damage to these biologically important compounds.
In 1974, the Vacuum Generators Company marketed a field-emission STEM
(the HB5), on the prototype of which Burge and colleagues at London University
installed an electrostatic energy analyzer (Browne, 1979). Ferrier’s group at
Glasgow University investigated the practical advantages and disadvantages of
adding post-specimen lenses to the STEM, including their effect on spectrometer
performance. Isaacson designed an improved magnetic spectrometer for the HB5,
while Batson showed that superior energy resolution and stability could be obtained
by using a retarding-field Wien filter. Various groups (for example, Colliex and