next up previous contents
Next: Pileup Measurements and Modelling Up: Energy Scale and Spectral Previous: Results

Off-Peak Features of the Spectral Redistribution Function

  The most important off-peak features of the CCD response function are the silicon escape and fluorescence peaks that emerge when the energy of incoming irradiation exceeds the silicon absorption edge energy. The intensity of those peaks is a function of energy, and hence the ideal way to experimentally characterize these features is to illuminate the CCD with a monochromatic beam and collect data at different energies. This has been done at the synchrotron storage ring at BESSY using the Crystal Monochromator Beamline (KMC). The CCD which was tested there was w102c3. Data appropriate for fluorescence analysis were taken at energies ranging from 2 to 6 keV at an operating temperature of -120 C. After obtaining a histogram of the standard combination of grades at each energy a single gaussian fit was made for each of three peaks: main peak, escape and fluorescence. The number of counts for both escape and fluorescence peaks normalized to the main peak are shown in Fig 4.8 as a function of energy. A simulation of the CCD response for the same energies was run using a model based on the Monte-Carlo method. For each energy point 1,000,000 photons were thrown into the CCD structure, the simulated histograms were then treated in exactly the same way as experimental data.

The model predictions are shown in the same Figure as solid lines.


  
Figure 4.8: Intensities of experimentally measured escape (*) and fluorescence (+) features as a function of energy. Solid lines are the model predictions.
\begin{figure}
\vspace{3.5in}
\special{psfile=calReport/gyp/figures/esc.ps
angle=270
hscale=50
vscale=50
voffset=290
hoffset=45
}\end{figure}

For the escape peak the agreement between model and experiment is excellent. For the fluorescence peak it is reasonable, although there is an obvious discrepancy indicating that the model introduces some systematic error. The reason for that will be investigated later. This error does not have any practical significance, taking into account the very low number of counts in the fluorescence peak


next up previous contents
Next: Pileup Measurements and Modelling Up: Energy Scale and Spectral Previous: Results

Mark Bautz
11/20/1997