3LE 
and a new 
olutions are 
to be tilted 
dimensional 
;entric axis. 
Therefore, a 
yund, which 
38 spatially 
at following 
ting out the 
roscope are 
yentisch und 
1 bei extrem 
nicht um das 
standpunkte, 
r virtuellen 
"wegen. Nur 
zu verlassen. 
e ein Objekt 
kelt, die mit 
durch eine 
hematik der 
ralprojektion 
Mikroskops 
few options 
thers, optical 
e mentioned. 
is project we 
for further 
ing approach 
of focus and 
1agnification, 
so, the good 
r evaluation. 
]., 1987) it is 
ut how SEM 
ipter (Maune, 
There are some preconditions to spatial object reconstruction. 
The sensor has to be mathematically modelled, hence the 
imaging properties have to be known. Additionally, two or more 
images, of which the orientation parameters have to be supplied 
are required to derive three dimensional coordinates. These 
requirements raise two difficulties: First, ways to achieve 
different point of views have to be found, since it is impossible 
to move the microscope around the sample. Second, an object 
that serves as a calibration standard has to be used. Also for 
implementation, the calibration standard should approximately 
match the object size of later samples. Chapter 2.2 will 
introduce the nano positioning tilting table as a solution to the 
first obstacle. The recently developed microscopic calibration 
standard will be shown in chapter 2.3. 
When using magnifications of 1000x and higher in the SEM, 
parallel projection can be assumed (Hemmleb, 2001). 
Therefore, not only the interior and exterior orientation 
parameters will change, but also the mathematic approach to 
compute them. Details will be explained in chapter 3. 
2. HARDWARE COMPONENTS 
2.1 The Scanning Electron Microscope 
An electron beam generated on top of the microscope is 
accelerated with up to several thousand volts. Using a system of 
electronic lenses and coils, the beam is focused in the 
microscope column and scanned over a chosen area of the 
sample surface. The important signal component consists of so- 
called secondary electrons (SE), which are generated within the 
sample surface when hit by the electron beam. As the exit depth 
of SE is rather small, they carry high-resolution information. A 
positively charged detector collects the secondary electrons 
emitted at any scan position. According to the SE intensity at a 
given scan position, an analogous signal is generated and sent to 
a cathode ray tube, the screen. Mostly SE are responsible for the 
topographic contrast in SEM images as used here. This results 
mainly from the dependence of the SE yield on the tilt angle of 
the local surface normal relative to the incident beam. 
The shorter wavelength of electrons compared to visual light 
allows a resolution in the nanometer range. In combination with 
the introduced advantages, like depth of focus and good signal- 
to-noise ratio, SEM images are very interesting to 
photogrammetric applications. The law of projection to be 
applied is depending on the magnification. While the scanning 
electron beam can be considered as central projective, is should 
be interpreted as parallel projective above a magnification of 
1000x and more. The consequences to the mathematical models 
will be explained in chapters 2.3 and chapter 3. 
2.2 The Nano Positioning Tilting Table 
We already mentioned the necessity to simulate different points 
of view in order to derive 3D coordinates. As the specimen is 
positioned inside the vacuum chamber of the microscope, it is in 
a way a part of the microscope itself. In this special case, we are 
not able to move the camera around the object. Instead, we have 
to tilt the sample along a rotation axis in order to allow views 
from different positions. Due to the fact that the field of view in 
an SEM is extremely small, any translation of the sample is 
unwanted. Rotating the sample, respectively the sample stage, 
may result easily in a translation motion, if the area of interest is 
not positioned within the rotation axis, the so-called eucentric 
axis, as shown in fig.l. Therefore, a way to very accurately 
move the sample into the eucentric axis had to be found. 
  
  
   
rotating- 
| € | axis 
  
  
c) | d) 
Figure 1. The effects of rotating a sample variously positioned 
with regard to the rotation axis: Situation a) and c) show a 
shifting along the z-axis, resulting in problems to the focus. The 
sample is displaced along the x-axis from the rotation axis in 
situation b) and c), causing the area of interest to exit the field 
of view. In situation d) finally, only the tilting of the sample 
will be observed. 
A nano positioning tilting table has recently been developed to 
reliably move the sample into the desired position as depicted in 
Figure 1.d), even at magnifications higher than 50000x. The 
dimensions of the microscopes vacuum chamber limit the 
design of the tilting table. Figure 2 shows an actual photograph 
of the table. To give an idea of the table's dimension, a 10 Euro- 
Cent coin has been placed next to the table, in the lower right 
corner of the image. 
    
kleindiek 
gnanotechnik 
Figure 2. Photograph of the positioning tilting table, showing 
the rotation axis (red), the translation axes (blue) and enhancing 
the actual sample stage (green). 
With this nano-positioning tool, the operator of the microscope 
is able to move the area of interest interactively into the desired 
position. However, despite the accuracy and reliability of the 
positioning table it remains an iterative process, challenging the 
operators patience. 
-211- 
—————