Experimental arrangement

This section briefly describes the experimental layout of the RICH-1 prototype detector. A full description of the apparatus, and the detector configurations used, can be found in [3].

The tests make use of two distinct geometrical configurations of detector prototype, referred to throughout as Configurations 1 and 2. Configuration 1 is a ${\frac{1}{4}}$ scale model of RICH-1. It is used with a total of 5, and later 7, HPDs to study the photon yield and Cherenkov angle resolution from aerogel and the photon yield from gas radiators. Configuration 2 is a full-scale prototype of the optical layout of RICH-1 and is used to measure the Cherenkov angle resolution with the C₄F₁₀ gas radiator. The geometry and characteristics of the two configurations are described below :

• Configuration 1 consists of a vessel of dimensions 800x600x600 mm. The vessel is aligned such that the beam passes through a block of aerogel, followed by 400 mm of air or C₄F₁₀ gas radiator. This is shown schematically in Figure 1.

  Figure: Schematic layout of Configuration 1, the ${\frac{1}{4}}$-scale prototype RICH-1 vessel.

  
  Cherenkov photons produced are reflected and focused by the mirror onto an array of HPDs. These are mounted on movable elements, placed 246 mm from the mirror centre, at an 18^o tilt relative to the beam.
• Configuration 2 uses two extension tubes to increase the active gaseous radiator length of Configuration 1 to 1 m. The mirror focal length is also increased to 1117 mm. The geometry is shown schematically in Figure 2.

  Figure: Schematic layout of Configuration 2, the full-scale prototype RICH-1 detector.

  

A three-dimensional coordinate system is defined for the purpose of reconstruction and alignment. The origin of the orthogonal axes is defined to lie at the point where the beam intersects the mirror surface. The z-axis then lies along the beam direction, pointing upstream. The x-axis is then defined to lie in the horizontal plane in the direction of the detector plane, with the y-axis pointing vertically upwards. The two detector configurations share common components which are :

• The radiators : Two types of silica aerogel samples are used, both with areas of 110 x 110 mm² perpendicular to the input beam direction. They are referred to here and in [3] as samples 1 and 2, with thicknesses of 18 and 11 mm respectively². Their optical transmissions were measured in the laboratory and are shown in Figure 3(d).

  Figure: (a) Transmission characteristics for glass and mylar, (b) mirror reflectivity, (c) HPD detector quantum efficiency and (d) measured and fitted aerogel transmission curves as funct ion of wavelength.

  
  The aerogel samples have nominal refractive indices of n $\sim$ 1.03, and are positioned 425 mm upstream of the mirror centre. Fits to the transmission curves shown in Figure 3(d), both return a clarity coefficient of 0.008 $\mu$m⁴/cm. The dimensions of the air and C₄F₁₀ gas radiators in each configuration are given in Table 1.
• The detector plane : The photodetector plane holds up to seven 61-pixel hybrid photo-diodes³, contracted from the DEP company. These are run at a common acceleration voltage of between 10 and 12 kV during normal operation [3]. The detectors are read out individually using VIKING VA2[5] amplifiers with a shaping time of $\sim$1.2$\mu$s. The detectors and their electronics are mounted onto a plate where, after initial alignment, their (x, y) positions and rotation angles are fixed. The orientation and numbering of the HPDs on the detector plane is indicated in Figure 4.

  Figure: A display of accumulated events, within a single run, shown on the front face of the detector plane of HPDs indicating the 7 detector positions. The ring is due to photons emitted in the C₄F₁₀.

  
  Filters are placed over some of the HPD input windows to restrict the detected wavelength. In these tests, borosilicate glass and mylar filters are used.
• The mirrors: These are mounted on the backplate of the vessel and held on three micrometer screws. They are adjusted to align the Cherenkov ring image onto the centre of the detector plane. The focal lengths of the mirrors used in each detector configuration are given in Table 1.
• The gas circulation system : The RICH vessel is connected to a gas recycling and purification system [3] that cycles C₄F₁₀ gas approximately once every hour. For these tests, a slight overpressure is maintained.

Table: Measured global geometry parameters for the three RICH configurations.

Quantity	Value
Mirror-to-beam Angle	18 ($\pm$2)o
Mirror Focal Lengths :
Configuration 1	236 ($\pm$2) mm
Configuration 2	1117 ($\pm$10) mm
Mirror-to-detector distance :
Configuration 1	245 or 250 mm
Configuration 2	1143 ($\pm$4) mm

The prototype RICH vessel was sited in the CERN T9 beam, which is a secondary beam derived from extracted PS protons, and provides charged particles of either polarity. Particle momenta from 2 $\rightarrow$ 15.5GeV/c are available. The particle type (e, $\pi$, K, p) can be separated using a CO₂ threshold Cherenkov counter, installed 30 m upstream of the prototype.

Incident beam particles are selected using two upstream and two downstream scintillation counters. The last two are placed 4 m apart and the smallest scintillator ( 4 x 4 mm²) defines the accepted beam size and an expected beam divergence of not more than 3 mrad. Scintillators 1 and 4 are 50 x 50 mm and cover the whole beam. The readout is triggered by the four scintillation counters in coincidence. During a typical extraction pulse of 200ms duration, 5 x 10³ beam particles enter the detector volume, resulting in $\sim$ 30 triggers.

In Configuration 2, the beam trajectory is measured using a silicon telescope. This consists of three planes of silicon-pad detectors. Each plane has 22x22 pads, with pad dimensions 1.3 x 1.3 mm². The planes are placed upstream of the prototype and separated by approximately 1 m. The purpose of the telescope is to allow event-by-event track direction measurement from which the Cherenkov angle can be more precisely measured, giving a beam trajectory uncertainty of approximately 0.38 mrad. This assumes perfect geometrical alignment, and is discussed later in Section 4. Analogue signals from this detector are read out using the same digitiser as the HPDs, as described in [3].