An amplitude of the fluorescent peak was much harder to reproduce. The key to the correct model of the flourescent peak, especially at energies close to the Si absorption edge, is taking into account fluorescence from the gates. Immediately above Si edge absorption in the polysilicon gates becomes a significant factor. Of the absorbed photons 4.3% produce a fluorescent photon and roughly half of them goes into the bulk of silicon and gets detected, while another half is emitted into the upper hemisphere and is lost. Unlike flourescent photons that are emitted from the bulk silicon, the leftover charge corresponding to the difference in energy between primary and fluorescent photon cannot be detected when it is formed in the polysilicon gate. This means that approximately half of the photons fluoresced from the gates will be found in the fluorescent peak, while only a tiny fraction of the ones from the bulk will end up in the fluorescent peak. Because of this flourescence from the gates is by far a dominant factor in forming the flourescence peak. On Fig. 4.128 are shown the amplitudes of the escape and flourescence peaks both from experimental data and simulated eventlists. The experimental data had been collected at a synchrotron ring at BESSY using Crystal Monochromator Beamline with a device w102c3.
Solid lines on Fig. 4.128 show
the model predictions. An agreement between the data and the model is good.
A discrepancy in the amplitudes of the fluorescence peak reported previously
in the Preliminary Calibration Report is now resolved. It was caused mostly
by not taking into account fluorescence from the gate structure.
Figure 4.128: Intensities of experimentally
measured escape (*) and fluorescence (+) features as a function of energy.
Solid lines are the model predictions.
