The CERES spectrometer covers the pseudorapidity region between 2.1 and 2.7 with full azimuthal symmetry. The schematic view of the whole upgraded CERES spectrometer is given above. Particle identification and tracking are based on two azimuthally symmetric RICH detectors with a Cherenkov threshold more than 32 - high enough to substantially suppress signals from the large number of hadrons produced in the collision.
There are three main features in the design of the CERES spectrometer :
The main coils between the two detectors provides an azimuthal kick for momentum and charge determination, leaving the polar angle unchanged. The correction coils shape the field in the RICH-2 radiator such that it is parallel to the particle trajectories (see the field lines in the lower part of the figure). Thus there is no influence of the magnetic field on the particles. The radiators are filled with CH4 at atmospheric pressure and a UV transparent window of calcium fluoride in RICH-1 and quartz in RICH-2 separates the radiator from the UV detectors.
The Cherenkov photons emitted by fast charged particles traversing the radiators are reflected back by the mirrors and registered in two UV-detectors located in their focal plane.These UV-detectors have a two-dimentional readout with about 50000 pads each.
The target consists of 8 Au disks of 25 micron thickness and 600 micron diameter. Such a structure keeps a large interaction rate while minimizing the amount of gamma conversions and secondary interactions.
For the Pb beam a stand-alone external tracking has been incorporated. It consists of a closely spaced doublet of two silicon-drift detectors (SiDC) located at approximately 10 cm downstream the target. They allow a precise vertex reconstruction, and help the pattern recognition of RICH-1 by providing an a priori knowledge of the ring center location.
A Pad chamber fullfils the similar task for RICH-2 as the two SiDC for RICH-1. It is placed behind the mirror of RICH-2, at a distance of 3.3 m from the target and consists of a MWPC with pad readout. The chamber covers the pseudorapidity interval between 2.0 and 2.7, i.e. slightly larger than the acceptance of the spectrometer.
The doublet of SiDC is located 10 cm after the target with a 15 mm spacing between them. Their acceptance is slightly larger than 1 unit of pseudorapidity. The SiDC provide the ( , ) coordinates of all charged particles produced in the collision. This allows a precise vertex reconstruction without any additional information from RICH-1. The SiDC are also used for the determination of the pseudorapidity density of charged particles, dNch/d.Each chamber is made of a silicon wafer and has a disc shape of 4" diameter. Electrons liberated in the silicon by fast particles traversing the detector drift radially, towards the outer edge of the chamber, which consists of an array of 360 anodes, where the electrons are collected. The drift time gives the radial coordinate r (or, equivalently, ), and the charge deposit in the anodes measures the azimuthal coordinate, . The detectors used in the 1996 run were the first of a new generation of 4" detectors. Concentric rings defining the drift field are actually approximated by polygons of 360 sides. The nominal value of the field is 500 V/cm resulting in a drift velocity of 10 m/s. The resolution is a function of r. The larger the drift distance of an electron cloud, the bigger the spread of the pulse between adjacent anodes, and, as a consequence, the better the position resolution in the direction. For a particle coming close to the anodes, the diffusion length is not large enough to spread the charges over several anodes. Consequently the resolution in is given by the binning size, namely 17/(12) ~5 mrad. For small r, the electrons diffuse both in the drift direction and in the direction and are shared by a number of adjacent anodes. The coordinate of a particle is obtained by calculating the center-of-gravity of the anode signals yielding a resolution in which is typically one order of magnitude better than the bin size, namely ~1.7 mrad.
The Pad Chamber is a multiwire proportional chamber (MWPC) with pad read-out. The layout is given in the Figure.The chamber is placed behind the mirror of RICH-2, at a distance of ~3.3 m from the target. It covers the pseudorapidity interval 2.0 < < 2.7, i.e. slightly larger than the acceptance of the spectrometer. The Pad Chamber vessel is directly attached to the RICH-2 end flange. The electrodes have an annular shape with inner radius Rin = 419 mm and outer radius Rout = 850 mm. The scheme is very similar to the one adopted for the UV detectors. The upstream cathode is a stainless steel mesh, made of wire of 50 m diameter with a pitch of 0.5 mm. It is made of one piece, the mesh being stretched between two concentring rings of G10. The following electrode is the multiwire anode. It is made out of 30 m diameter gold plated tungsten wires spaced by 3 mm. The anode to cathode distance is 5 mm. All wires are parallel to each other as shown in the Figure. This has the disadvantage of mechanical stresses due to the 100 g tension of each wire. Nevertheless the scheme was adopted in order to avoid dead area losses (a 10% loss would occur in a 16 sectors scheme separated by radial spokes as done in UV-2). To avoid staggering instabilities of the longest wires (~2 m long) the wires are glued to a small support (3 mm wide) at the middle. The downstream cathode is the pad electrode used for the readout and consists of ~28000 pads with a size of 7.6X7.6 mm2. Each pad is connected to a module at the outer side of the detector where preamplifiers are directly plugged. The preamplifier module contains 121 channels and covers the area of 11X11 pads. The full acceptance requires a total of 250 modules. The raw Pad Chamber data consist of the x and y pad coordinates and signal amplitude. The Pad Chamber is operated with a gas mixture consisting of argon + 10% carbon dioxide. A small voltage of +100 V is applied on the mesh cathode in order to collect the electrons liberated between the RICH-2 exit window, which is at ground potential, and the mesh cathode. This brings the ionization path to a total length of 27 mm. For the gas mixture used this corresponds to a total primary charge of ~250 electrons created by a minimum ionizing particle. However ~40% only of this charge is collected by the front-end electronics due to the trigger and read-out time sequence, resulting in an effective primary charge of ~100 electrons. The detector was operated at a voltage of VAC ~2000 V producing an average signal of 3.4 105 e, i.e. an electronic gain of 3.4 103.