393 23
field) of the SEM, as shown in. Fig. 23.15. The aluminum
stub was mapped with the specimen plane located at the
manufacturer’s specified ideal working distance for this
SEM/EDS. The Al intensity map encoded with a pseudo-
color scale shows that the transmission varies from maxi-
mum (100 %) along the top of the image to values in the
range 60–70 % at the bottom. The top-to-bottom asymme-
try in this map reveals that the collimator on this EDS sys-
tem is actually misaligned, since with proper orientation the
maximum of the collimator transmission should be at the
image center (coincident with the optic axis). The transmis-
sion function of the collimator as a function of vertical dis-
tance along the SEM optical axis can be determined by
repeating the mapping at different working distances. The
graph in. Fig. 23.15 shows the intensity measured at the
center of each map, revealing a decrease of approximately
40 % as the working distance was increased from 10 to
20 mm. This collimator thus allows high transmission from
a large volume of space, with dimensions of at least
3 × 2.5 × 10 mm, so that any X-rays produced in this volume
with a line-of-sight to the EDS detector will contribute to
the measured spectrum.
Optimizing the EDS spectrum measured from a rough,
irregularly shaped surface requires careful consideration of
the selection of the location on the specimen to be measured.
The analyst must be aware of the location of the EDS relative
to the measured location to avoid the situation illustrated in. Fig. 23.16a, where the beam location leads to an X-ray path
that must pass through the bulk of the specimen to reach the
EDS, leading to extremely high absorption. Ideally, using a
specimen stage with several rotation axes, a rotation about a
vertical axis will bring the feature of interest to directly face
the EDS, thus minimizing the absorption, as shown in
. Fig. 23.16b. A further rotation about a horizontal axis
places the feature perpendicular to beam to minimize back-
scattering and remote X-ray excitation (. Fig. 23.16c). Note
that although backscattering is minimized by establishing
normal beam incidence (effectively a zero tilt angle), the
backscattered electrons are broadly emitted with a cosine dis-
tribution so that while the majority are emitted at high angles
there still remains a small but significant fraction emitted at
low angles to the surface that may strike nearby features and
excite the surrounding materials, contributing to the spec-
trum measured at the beam impact position.
020040060080010001200810121416182022
Working distance (mm)Al counts/pixel01020304050607080901001 mm. Fig. 23.15 X-ray mapping experiment to
determine extent of collimator acceptance. Large
scale low magnification map (3 × 2.5 mm) of an
aluminum stub. The graph shows the intensity
measured at the center of a series of such maps
recorded at different working distances
23.5 · Best Practices for Analysis of Rough Bulk Samples