Introduction

This paper reports results from a prototype Ring Imaging Cherenkov (RICH) counter which uses an array of hybrid photo-diode detectors (HPDs) to detect the Cherenkov photons. This is a prototype of the upstream RICH detector of the LHCb experiment [1]. The LHCb experiment is a single-arm spectrometer, recently approved to run at the CERN Large Hadron Collider. The data were collected during the spring and summer of 1997 at the CERN PS facility.

The LHCb experiment is designed to make precision measurements of CP violation in B decays. The experiment utilises two RICH detectors. The upstream detector (RICH-1) covers polar angles between approximately 25 and 330 mrad, and the downstream detector (RICH-2) covers polar angles between approximately 10 and 120 mrad. The RICH detectors are crucial in providing charged $\pi$/K separation for particles with momenta between 1 and >  100 GeV/c. The detectors reduce backgrounds in selected final states, eg. B_d⁰ $\rightarrow$ $\pi^{+}_{}$$\pi^{-}_{}$, B_s⁰ $\rightarrow$ D_s^$\pm$K^$\mp$, hence minimising systematic errors on CP violation measurements. They are also used to provide an efficient flavour tag of B mesons using kaons.

The RICH-1 detector [1,2,3] combines gas and aerogel radiators to provide $\pi$/K separation over the momentum range . The aerogel is placed against the entrance window, with a second C₄F₁₀ gas radiator behind it. With a refractive index of n=1.03, a 5 cm thick slab of aerogel provides low momentum $\pi$/K separation up to $\sim$ 10 GeV/c. It gives an expected mean number of $\sim$ 15 detected photoelectrons for a fully saturated ($\beta$ = 1) track. The C₄F₁₀ gas radiator (with n=1.0014) is approximately 95 cm long, covers the momentum range up to 65  GeV/c, and gives an expected mean number of detected photoelectrons of $\sim$55. A spherical mirror of focal length 1 m is common to both the gas and aerogel radiators. The mirror is tilted by $\sim$ 250 mrad to reflect Cherenkov photons out of the spectrometer acceptance and onto the photodetector plane. The RICH-2 radiator comprises approximately 180 cm of CF₄ gas.

The LHCb RICH counters use arrays of photon detectors to observe the Cherenkov light. HPDs are the chosen baseline option for the LHCb photon detectors as they provide large-area coverage, whilst the silicon segmentation gives the necessary detector granularity [4]. HPDs have low noise, a high quantum efficiency and good single-photon sensitivity. However, when operated at 20 kV with a gain of $\sim$5000, the signal charge resulting from one photoelectron is relatively small. The devices used in the tests reported here were manufactured commercially by DEP¹. They have an hexagonal array of 61 pixels. The pixels have an area of approximately 2 x 2 mm² which is the same as that proposed for LHCb, but with a much smaller active photocathode area of $\sim$ 190 mm².

The aim of the prototype RICH tests described here is to measure, in a realistic beam environment, the following :

• The signal-to-noise ratio of the HPDs and whether they are able to detect the relatively small numbers of photons with an acceptable efficiency.
• Photon yields from aerogel, air and C₄F₁₀ radiators and to compare that observed with expectations from theory and simulation. The aim is to ensure enough redundancy to allow robust pattern recognition in the full LHCb RICH-1 detector.
• To reconstruct the Cherenkov angle, and understand the contributions to its measurement resolution. The aim is to determine the overall precision of the RICH so as to verify the $\pi$/K separation performance of the LHCb detector as a whole.

This paper presents detailed analyses of data from the RICH-1 prototype and comparisons with simulation. It is structured as follows. Section 2 briefly describes the features of the RICH-1 prototype, its electronic readout and accompanying silicon telescope. Section 3 describes the data selection criteria and photon detector performance whilst Section 4 describes the technique of geometrical alignment of the individual elements. Details of the detector simulation are given in Section 5. An estimation of the photon yields from aerogel and gas radiators are described in Section 6. Section 7 presents the reconstruction of the mean Cherenkov angle for the aerogel and C₄F₁₀ radiators, and contributions to the angular resolution of these measurements. Finally, a summary is given in Section 8, including plans for future work.