The Varian Clinac 2100C accelerator (fig.2) has an exit window, scattering foils, a beam monitor ion chamber, a mirror, a ``shielding plate'', X and Y jaws and an electron applicator. This particular accelerator at the Wisconsin Comprehensive Cancer Center in Madison has thicker scattering foils (same as in an older Clinac 1800) in order to match the depth-dose curves of an earlier machine. It also has the relatively thick-walled gold-plated mica ion chamber and a type II applicator cone with 4 relatively thin scrapers[1]. All measured data are from Carol Wells and Rock Mackie of the University of Wisconsin.
Figures 11, 12, 13, 14 and 15 present the spectra and the angular distributions at the phantom surface and comparison between calculated and measured dose distributions along with dose components contributed from each beam defining component from 6, 9, 12, 15 and 18 MeV electron beams respectively.
There is excellent agreement between calculations and measurements. All dose measurements have been done with a small diode detector.
Note that there are two peaks in the electron spectra for all beams. The low-energy peak is caused by electrons passing through a 6 mm aluminum applicator scraper and remaining in the beam[1] based on a detailed study which separated the spectra from each scraper component.
Note that the total bremsstrahlung tails in the depth-dose curves are contributed mainly by the contaminant photons in the beams.
In order to explain experimental observations[11] on the influence of an applicator design we also simulate the 9 MeV beam with a newer applicator design (Type III). The energy and angular distributions with this new applicator cone (Type III) from a 9 MeV beam are shown in figure 16 which should be compared to figure 12 for the Type II applicator. The scattered electrons from the applicator are greatly reduced and the second low-energy peak disappears. The surface dose contributed by scattered electrons from the applicator is reduced from 17% with Type II applicator to 8% with the Type III applicator.
By calculation it was found that
the peak could be made to disappear using an additional piece of lead on the
top of this part of the applicator.
Figure
17 presents the result of a 9 MeV beam with our
modified applicator design.
The calculated depth-dose curve with the
lead in place was virtually identical to that of the new design applicator
shown in figures
16 and 17 respectively.
The calculated dose profiles at 
for these three different applicators along with dose components
are shown in figure 18.
After this ``discovery'',
we learned that Varian sells an upgrade kit which has a
similar design and performs the same function.