Performance of ACC

The performance of prototype ACC modules has been tested using the $\pi$2 beam line at KEK PS. Typical pulse-height distributions for 3.5 GeV/c pions and protons measured by an aerogel counter, with $n$ = 1.015 and read by two 2.5 in. FM-PMTs, are shown in Figs. (a) and (b), with and without a magnetic field of 1.5 T, respectively. The number of photoelectrons ($N_{pe}$) for 3.5 GeV/c negative pions is measured to be 20.3. In the case of Fig. (b) we used a preamplifier and applied a higher HV to get almost the same gain as that without a magnetic field. Pions and protons are clearly separated by more than 3$\sigma$. It was found that cracks in the aerogel do not make a difference in the light yield.

Figure: Pulse-height spectra for 3.5 GeV/c pions (above threshold) and protons (below threshold) obtained by a single module of ACC in (a) non-magnetic field and (b) a magnetic field of 1.5 T. Silica aerogels with $n$ = 1.015 were stacked to form the module.

Figure: Pulse-height spectra in units of photoelectrons observed by barrel ACC for electrons and kaons. Kaon candidates were obtained by $dE/dx$ and TOF measurements. The Monte Carlo expectations are superimposed.

After ACC was installed into the Belle detector in December 1998, the initial calibration of the detector was carried out using cosmic rays. The Belle detector has been rolled into the interaction point and commissioned with $e^+e^-$ beams since May 1999.
Figure shows the measured pulse height distribution for the barrel ACC for $e^{\pm}$ tracks in Bhabha events and also $K^{\pm}$ candidates in hadronic events, which are selected by TOF and $dE/dx$ measurements [53]. The figure demonstrates a clear separation between high energy electrons and below-threshold particles. It also indicates good agreement between the data and Monte Carlo simulations [52].
A careful calibration for the pulse height of each FM-PMT signal has been performed with $\mu$-pair events. Figs. (a) and (b) show the average number of photoelectrons $<N_{pe}>$ for each counter row in the barrel ACC and each layer of the end-cap ACC, respectively. In the barrel ACC each row has on average 60 boxes and the row number is given from left to right in Fig. . The layer number of the end-cap ACC is given from the inner to the outer side. The light yield for the $\mu$ tracks depends on the refractive index of aerogel radiators, size and number of FM-PMTs attached on the counter module, and geometry of the counter module box. The light yield ranges from 10 to 20 for the barrel ACC and from 25 to 30 for the end-cap ACC, high enough to provide useful $\pi/K$ separation.

Figure: Average number of photoelectrons $<N_{pe}>$ for (a) each counter row in barrel ACC and (b) each layer in end-cap ACC.