Single photo-electron distributionIn the previous section, only the relative gain of PMTs is discussed. In this section, we describe how to measure the absolute gain, particularly at single photo-electron level. We get only the charge information from the electronics output, in units of pico coulombs. We have to know how to get the number of photo-electron from this charge value. The single photo-electron distribution is used for this purpose. The method is the same as the Nickel calibration, because the number of photon-electrons observed in each PMT is almost one in Nickel calibration. Fig 1 shows the charge distribution of each hit-PMT in units of pico coulombs. The mean value in this distribution is 2.055 pico coulomb, which is equivalent to single photo-electron.
The ``occupancy'', which is hit rate of getting a single photo-electron signal, as defined by Eq(B.2) in Y.Koshio's PHD thesis on the page of publication list, is used in order to check the up-down symmetry of the detector. Fig 2 shows the distribution of average occupancy over PMTs belonging to each layer of the barrel part. From this figure, occupancy appears higher in the top and bottom layers than the middle layers because of reflections from the surface of PMTs, however, the distribution is symmetric. Further, we checked that real data and Monte Carlo data were consistent. This result is crucial for analysis of the day-night effect of solar neutrinos. Note that there are the layers with high occupancy in Fig 2. This occurs because PMTs belonging to these layers were older (i.e. made earlier) than others and the quality is different. These PMTs have different single photo-electron detection efficiency even if the relative gain is adjusted to be same. The difference is 20% on average, and this is corrected in the energy determination .
| ||||||||
|
revised on 1999/09/27 webmaster@suketto.icrr.u-tokyo.ac.jp |