next up previous
Next: event selection Up: Critical test: total neutral Previous: Critical test: total neutral

total energy spectra

Figure 5: Summed energy from all reconstructed forward clusters in an event. The sample includes all events with between 2 and 7 clusters, a single recoil and no observed neutral energy in the barrel.
\begin{figure}\begin{center}\mbox{\epsfxsize =13cm\epsffile{totEinclusive.eps}}\end{center}\end{figure}

Figure 6: Summed energy from all reconstructed forward clusters in an event, weighted by the the tagger-cpv coincidence condition (solid histogram) and accidental condition (dashed histogram). The sample includes all events with between 2 clusters, a single recoil and no observed neutral energy in the barrel.
\begin{figure}\begin{center}\mbox{\epsfxsize =9cm\epsffile{totEcoinc.eps}}\end{center}\end{figure}

Figure 7: Difference spectrum from Fig. 6 which estimates the spectrum that would be obtained from neutral tagged events free from accidental effects. The fit is to a sum of two Gaussians with parameters given in Table 1.
\begin{figure}\begin{center}\mbox{\epsfxsize =9cm\epsffile{totEdiff.eps}}\end{center}\end{figure}


Table 1: Parameters of two-Gaussian fit shown in Fig. 7. Statistical errors are at the level of the least significant digit.
parameter Gaussian 1 value Gaussian 2 value
height 6378 4191
mean (GeV) 4.732 4.355
r.m.s. (GeV) 0.266 0.495

A test of above procedure would be to look at the total energy from all reconstructed showers in the forward calorimeter. This is a critical test, as can be seen from the total forward energy spectrum for all events shown in Fig. 5. The shape shows a threshold around 3 GeV coming from the online level-3 trigger, followed by a monotonically decreasing intensity with energy. Although the maximum energy is roughly consistent with the end-point energy of 5.65 GeV, there is very little evidence of an enhancement in the region of subtended by the tagger 4.39 - 5.38 GeV. Fig. 6 shows the same spectrum weighted by the tagger-cpv coincidences weight for tagging counter 10. The dashed histogram in the same figure indicates the corresponding spectrum for the accidentals for counter 10. The difference spectrum is shown in Fig. 7. The sum of two Gaussians gives an excellent empirical description of the spectral line-shape, with approximately equal numbers of events in the narrow spike at the maximum and the low-energy tail.


Table 2: Parameters of a Gaussian fit to the narrow peak in the total forward photon energy spectrum for subtracted spectra corresponding to each individual tagging channel. Statistical errors are at the level of the least significant digit.
tagging channel nominal mean (GeV) mean (GeV) r.m.s (GeV)
0 5.360 5.200 0.280
1 5.310 5.160 0.288
2 5.210 5.120 0.278
3 5.190 5.060 0.276
4 5.140 5.030 0.270
5 5.090 4.960 0.273
6 5.040 4.910 0.263
7 5.020 4.880 0.254
8 4.940 4.820 0.261
9 4.880 4.770 0.264
10 4.830 4.730 0.267
11 4.770 4.670 0.259
12 4.710 4.600 0.250
13 4.640 4.550 0.251
14 4.590 4.510 0.253
15 4.550 4.480 0.253
16 4.510 4.430 0.248
17 4.460 4.380 0.255
18 4.410 4.350 0.246

Figure 8: Total energy spectrum from the forward calorimeter for five-cluster events with one recoil and no neutral energy in the barrel. The spectra shown correspond to the lowest (solid) and highest (dashed) tagging channels in the experiment. The calculated mid-energies for these two tagging channels are indicated with the vertical lines.
\begin{figure}\begin{center}\mbox{\epsfxsize =13cm\epsffile{totE5clust.eps}}\end{center}\end{figure}

Fits like the one shown in Fig. 7 have been performed for all of the tagging channels. All channels exhibit a spectral shape similar to the one shown in the figure. Table 2 shows the fit parameters for the narrow Gaussian peak, and compares it with the energy calculated based upon the tagging channel. While not perfect, the agreement between the peak energy and the mid-energy for the given tagger channel is impressive when it is remembered that the energy calibration was based entirely on the invariant mass scale of two-cluster combinations. This is the first time that the total energy scale in the LGD has ever been verified.

The above spectra are all taken for the case of two-cluster events. Similar plots are obtained for events with higher multiplicities, albeit with slightly larger r.m.s. values as expected based on the known calorimeter resolution. Fig. 8 compares the subtracted total-energy spectra for the highest and lowest counters in the tagging focal plane. The correspondence with the energy computed based upon the tagger channel is similar to the two-cluster case, systematically low by about 100 MeV. An offset of this order is expected in order to account for the energy carried away by the recoil proton.

A comparison of the shapes of the two spectra in Fig. 8 shows that there is a significant tail on the low side of the resolution function, perhaps indicating some loss of energy out the sides of the detector combined with the effects of dead channels that appeared throughout the run, but mainly explained by the distribution of the kinetic energy transfered to the recoil proton. The tail is less prominent in the spectrum from the low energy end of the tagger only because it is truncated by the online level-3 threshold around 3 GeV. The widths of the narrow peak listed in Table 2 are in good agreement with the r.m.s. expected on the basis of the empirical resolution function [2] for single showers

\begin{displaymath}
\frac{\sigma_E}{E} = 0.036 + \frac{0.073}{\sqrt{E}}
\end{displaymath} (1)

whose value varies from 265 MeV at 4.41 GeV to 310 MeV at 5.36 GeV. An interesting study that has not yet been performed would be to look for a correlation between the energy transfered to the recoil proton and the missing energy in the forward that gives rise to the low-energy tail in Figs. 7-8.


next up previous
Next: event selection Up: Critical test: total neutral Previous: Critical test: total neutral
Richard T. Jones 2004-09-14