Beam Test

A TOF module built in accordance with the final design was assembled and tested using the $\pi$2 test beam line of KEK-PS. The test counter was set on a movable stage that enabled the position and incident angle of the beam on the counter to vary simultaneously so that the geometry of the Belle TOF detector was reproduced.
Figure shows the TOF time resolution as a function of beam position. A time-walk correction was applied at each position and the start-counter time-jitter of 35 ps was subtracted in quadrature. An intrinsic time resolution of approximately 80 ps was obtained over the whole counter. In the Belle experiment we expect some additional contributions to the time resolution:

• a 10 % degradation of the intrinsic resolution caused by the 1.5 T magnetic field as was observed in a beam test of a same-size TOF counter with a similar fine-mesh PMT,
• a 20 ps contribution due to the 4 mm beam bunch length and jitter in the RF signal used as the reference time, and
• a 20 ps contribution from time stretcher readout electronics.

Figure: Time resolution of a Belle TOF module as a function of hit position.

Including these errors in quadrature gives an expected overall time resolution that meets the Belle design goal of 100 ps.
Figure shows time-of-flight distributions for $\pi^+$ and protons at 2.5 GeV/c which indicate 5$\sigma$ separation of $\pi^+$/p at 2.5 GeV/c for the beam at a forward region. This implies approximately 3$\sigma$ $\pi$/K separation at 1.25 GeV/c. A tail at later times is attributed to a gain saturation of the PMT at high signal rates. A rate of minimum ionizing particles on a whole TOF counter was observed to be about 14 kHz during the test with a positive beam. At this rate and with a PMT high voltage of 1500 V, the average anode current was about 10 $\mu$A of a total base current of 60 $\mu$A. In order to minimize the saturation, the ratio of anode to base currents should be kept below 10 %. With the present IR and beam pipe design and assuming a PMT gain of $3 \times 10^6$, a Monte Carlo simulation predicts an anode current of 2.4 $\mu$A. This indicates that the beam test conditions described above were much severer than those expected in the Belle environment. In order to guarantee a further safety margin under high beam background conditions, the nominal base current was increased to 300 $\mu$A in the final design.

Figure: $\pi/p$ separation at 2.5 GeV/c.

Figure shows the effective light yield seen by a TSC PMT as a function of beam position. A minimum ionizing track produces more than 25 photoelectrons over the whole counter. This ensures a high efficiency of 98 % for TOF trigger even at a nominal discrimination level of 0.5 mips.

Figure: Light yield of a TSC counter as a function of hit position.