A study of the energy tail
  1. From the Kp2 energy distribution, one observes an obvious energy tail appearing in the higher side, which was not found in the E787 data.
  2. A rate dependence study (where the histogram is for the Ck+Cpi < 12 M/spill low rate events, while the black dots are for the high rate events) show this tail is very likely due to the increasing accidental hits when the rates go up.
  3. It is noted that the energy tail can also present when kaon decay in flight. A study of energy distribution as a function of the delay coincidence (tk-tpi) is conducted. The plots (where from top left to bottom right, the energy distribution for the events with tk-tpi<8ns is compared to those with tk-tpi within 8-16, 16-24, 24-32, 32-40 ns, respectively)  show no evidence on this possible correlation. Only observed is that the events with tk-tpi within 24-32 ns have an 1 MeV peak shift and higher energy tail.
  4. To increase the Range Stack efficiency, the FERA gate was shifted by about 10ns. There is a concern about the election energy from muon decay may contribute to the pion energy tail. A study of pion energy as a function of the muon life time is performed, no evidence for such a correlation is observed.
  5. There is a concern on the correlation with the spill timings. The correlation is truly there, but the contribution is on the opposite direction (see this plot , interestingly, we also found this correlation in 1998 data, which could be a part of the previously unknown contribution to the energy resolution). A further study shows that the accidential activities in the I-counter are not correlated with those in the Range Stack.
  6. Accidental hit in the I-counter can also contribute to the energy tail as shown in the scatter plot of Eic vs. Etot. A study by comparing the 1998 and 2002 data is done and the energy distributions for which the I-counter energy measurement (Eic) is not 1.5 MeV greater than the expected (Eicest) show a strong tail suppression.  Also note that the fraction of events with Eic-Eicest>1.5 MeV increases from 3.5% in 1998 to 7.8% in 2002. By checking the I-counter timings, we also find that part of the worse IC energy resolution correlates with the the IC timings (see the plot).
          I therefore conclude that the observed energy tail enhancement in the 2002 data is mostly due to the accidental hits in
          the I-counter. The cut on the I-counter energy measurement consistency will significantly reduce this effect at a cost of
          about 4.3% additional acceptance loss. Exploiting the TD or CCD information, we should be able to get the efficiency
          back.