1: The first step of the data reduction is an elimination of the noise and
2: obvious background events. First of all, the events with total photoelectrons
3: less than 1000, which corresponds to $\sim 100$~MeV, are selected.
4: Next, the following data reductions are applied:
5: (a) Events whose time difference since a previous event was less
6: than 50~$\mu$sec were removed in order to eliminate decay electrons
7: from stopping muons.
8: (b) Events with an outer detector trigger and over 20 outer PMT hits
9: were removed in order to eliminate entering events like cosmic ray muons.
10: %%Figure~\ref{fig:anti} shows typical outer PMT hit distribution
11: %%before and after OD triggered event cut.
12: %%Almost of the events with large outer PMT hit is removed by cut (b),
13: %%but those events are sometimes remained due to outer DAQ problem, therefore,
14: %%(c) events with over 20 outer PMT hit, called OD mismatch events,
15: %%are removed out.
16: (c) A function to categorize noise events is defined by the ratio
17: of the number of hit PMT's with $|Q| \leq 1.0$~p.e. and the
18: total number of hit
19: PMT's. Since typical noise events should have many hits with low charge,
20: events with this ratio larger than 0.4 were removed.
21: (d) A function which can recognize an event where most of its hits are
22: clustered in one electronics module [ATM] is defined.
23: It is the ratio of the maximum number of hits in any one module
24: to the total number of channels (usually 12) in one ATM module.
25: Events with a larger ratio than 0.95 were removed as they generally arose
26: due to local RF noise in one ATM module.
27: %% Figure~\ref{fig:noise} shows the typical (f) and (g)
28: %% distribution and cut criteria.
29: (e) Events produced by flasher PMT's must be removed.
30: These flasher events often have a relatively larger charge than normal events,
31: therefore they are recognized using the maximum charge value and the number of
32: hits around the maximum charge PMT. The slightly involved criteria is shown in
33: Figure~\ref{fig:fla}. (f) An additional cut to remove noise and flasher events
34: is applied using a combination of a tighter goodness cut($goodness \leq 0.6$)
35: and a requirement of azimuthal uniformity in the Cherenkov ring pattern. Good
36: events have a uniform azimuth distribution of hit-PMT's along the
37: reconstructed direction, while flasher events often have clusters of hit
38: PMT which lead to a non-uniform azimuth distribution.
39: %
40: \begin{figure}[hbpt]
41: \begin{center}
42: \includegraphics[scale=0.35]{data-analysis-1st/fla.eps}
43: \caption{The relation between max charge and the number of hit PMT's
44: in a $5 \times 5$ patch surrounding the PMT with the maximum charge.
45: At most 24 tubes in this patch (plus the one at the center) can be
46: activated. The upper plot shows a
47: typical distribution including active
48: flasher events, and the lower plot
49: shows a typical good data set. The cut region is the area above
50: the line in the upper figure.}
51: \label{fig:fla}
52: \end{center}
53: \end{figure}
54: %
55:
56: In this noise reduction, the number of candidate solar neutrino
57: events went from $3.43\times10^7$ to $1.81\times10^7$ after applying the
58: 2~m fiducial volume cut and constraining the energy region from 5.0 to
59: 20.0~MeV. Table~\ref{tab:reduc} shows the reduction step summary.
60:
61: The loss of solar neutrino signal by the noise reduction is
62: evaluated to be $0.8\%$ using a Ni(n,$\gamma$)Ni gamma source at several edge
63: positions of the inner detector. The difference between data and MC for this
64: source yields the
65: systematic error for this series of reductions; it's $\pm 1.0\%$ for our
66: solar neutrino flux
67: measurement. The loss is mainly due to the flasher cut.
68:
69:
70:
71:
72:
73:
74: