Energy spectra

In figures 2.22 and 2.23 we show the LS1 and LS3 calibration sources spectra at $\theta$=0 and averaged on the azimuthal interval $\phi$ $\sim$ $\pm$10. The spectra are background subtracted. The channels 241-256 have been excluded because they do not contain scientific data (see section 1.2.3). Typical 1-sigma error are plotted in three points.

The spectra show the expected features discussed in section 3.1.2. At 60 keV the spectrum is composed by the photo-peak and the characteristic X-ray escape peak. At 122 keV we observe no more escape peak, which theoretically should still be present but here is 'merged' with the photo-peak due to the detectors low energy resolution, and it begins to apparent a contribution from Compton scattered photons at the left of the photo-peak. As the energy increases the Compton continuum and a back-scattering component superimposed to it becomes more and more evident. The Compton edge is very smoothed as the energy increases, and the gap between it and the photo-peak is filled by multiple-scattered photons and the low-energy wing of the photo-peak itself. No signature of pair production absorption is observed, as expected, because the maximum source calibration energy, 662 keV, is lower than the 1022 keV required for this process to occur. Also, in the low-energy part of the 166, 392 and 662 keV spectra, peaks due to the secondary lines at 33-39, 24-28 and 32 keV (Tab. 2.1) are apparent. Again, the low detector resolution do not allow to see these blends as separate peaks.

In figures 2.24 and 2.25, spectra at 60 keV for LS1 and at 662 keV for LS3 taken on $\phi$ ranges between -50$\rm ^{\circ}$ and +50$\rm ^{\circ}$ at $\theta$=0. are shown. The indications that come out from these data is that, at least not too far from detectors axis, the overall shape of the spectral response does not depend significatively on source direction. It is also well seen the dependency of the overall efficiency on $\phi$, accordingly with what is seen in ratemeters data. In particular, in the 662 keV spectra there is indication that the backscattering peak area decreases proportionally to the photo-peak area. Thus, the backscattering contributed counts can assumed to be a constant fraction of the source photons over this angular range. We finally note that in the 60 keV spectra, a little feature is present at channel corresponding to the gain calibrator energy, indicating a slight change in PMT gain between the different measurements.

The preliminary analysis of these calibration spectra shows that, in order to reconstruct LS1 and LS3 spectral response functions, in addition to the modeling of the main spectral components (photo-peak, characteristic X-ray escape peak, Compton continuum, multiple scattering events), a satisfactory estimation and subtraction of the backscattering contribution due to interaction of source photons with the calibration room walls and floor is required. The better way to to this is to introduce calibration room walls and floor in the Monte Carlo model and then compare the simulations results with these calibrations. Because the MC was not still developed to a satisfactory stage, we addressed this problem by a straight comparison between LS line spectra and the ESTEC and ENEA sources calibration spectra described in previous section.

  

Figure 2.22: LS1 line spectra at the various calibration energies.

  

Figure 2.23: LS3 line spectra at the various calibration energies.

  

Figure 2.24: LS3 spectra of the 60 keV source at different azimuthal angles.

  

Figure 2.25: LS3 spectra of the 662 keV source at different azimuthal angles.