Comparison of Counter Configurations

The light yield obtained for each counter configuration can be found in Table 4. If not mentioned particularly, pions at 2.0 GeV/$c$ was injected at the center of the counter in the forward direction.
Fig. 38 shows $N_{pe}$ for each configuration as a function of x.

From the cofiguration tests, the following features of each type were found.
(1)Belle type provides the largest $N_{pe}$ ($>$23 pe can be detected for 2 GeV/$c$ positive pions), but has a strong dependence on the beam position.
(2)Integration Cube type gives reasonable $N_{pe}$ and the uniform position dependence. The uniformity is kept when changing the counter width.
(3)Mirror type provides $N_{pe}$ per PMT larger than for Integration Cube type, and a small position dependence. This type collects scattered light rather than directed light.

In the beam test, Mirror type provides reasonable photoelectrons and a small position dependence despite one PMT for readout. But, in the actual experiment where aerogel with n = 1.01 will be used, hence, the observd light yields degrade about two thirds. Therefore, Integration Cube type is our solution of the counter design now. The common characteristics for Mirror type and Integration Cube type is an air space called ``Integration Cube'' behind the aerogel. The integration cube is effective for the uniformity of $N_{pe}$ on the beam incident position.

表 4: Photoelectron yields measured for each counter configuration.

Beam	Belle	Integration	Mirror
direction	type	Cube type	type
Forward injection
$N_{pe}$ (right)	11.4	8.9	13.5
$N_{pe}$ (left)	11.6	9.8	-
$N_{pe}$ (sum)	23.0	18.7	13.5
Reverse injection
$N_{pe}$ (right)	-	7.3	9.3
$N_{pe}$ (left)	-	8.2	-
$N_{pe}$ (sum)	-	15.5	9.3

図 38: Comparison of the three counter configurations. For Mirror type, the PMT is located at x = -6 cm.