Calorimeter System

An excellent knowledge of the electron or photon energy is necessary for precision measurements like couplings within and beyond the standard model, or to resolve possible narrow resonances of new particles over a large background. A good energy resolution and a good linearity are needed for energies ranging from a few GeV up to a few TeV.

An important example is the possible discovery of the Higgs boson. If the Higgs mass is below 130 GeV the decay H → γγ is the most promising discovery channel. If the Higgs mass is larger and, in particular if it is at least twice the mass of the Z0-boson, the Higgs can be discovered in the H → Z0Z0 → e+e-e+e- decay channel. Even in this case the energy of one of the electrons can be as low as about 10 GeV. The possible observation of the Higgs boson therefore requires excellent measurements of electrons and photons from low to high energies.

Electromagnetic Calorimeter


The electromagnetic calorimeter is a sampling calorimeter made of lead absorbers bent to an accordion geometry and immersed in a liquid argon bath. Between two consecutive absorbers, a three layer kapton electrode is dividing the 2.1 mm wide gap into two equal parts and plays a twofold role: it distributes the 2000 V high voltage through the two external layers, and it collects the signal in the inner one.

Liquid argon calorimetry has been chosen for ATLAS because of its intrinsic linear behavior, stability of the response in time and radiation tolerance. The accordion geometry has been chosen because it allows very good hermeticity. Additionally, it minimizes inductances in the signal paths, allowing the use of the fast shaping which is needed to cope with the 25 ns interval between bunch collisions at the LHC. The total thickness is 24 radiation lengths X0 in the barrel and 26 X0 in the end-caps.

In the barrel region (|η < 1.4|), a readout electrode is segmented laterally in cells pointing to the interaction point; longitudinally, a cell is separated in three parts, front, middle and back, with granularities of η×φ = 0.003×0.1 for the front section to η×φ = 0.050×0.025 for the back section.

The barrel is constructed from two wheels, each one made of 16 modules, with a module covering a range of 0 < |η| < 1.4 and 2π/16 in φ direction. It shares its cryostat with the superconducting solenoid. The end-caps, covering the region of 1.4 < |η| < 3.2 share their cryostats with hadronic and forward liquid argon calorimeters. To evaluate the amount of energy lost in front of the calorimeters, both the end-caps and the barrel are complemented with presampler detectors, covering the 0 < |η| < 1.8 range. These presamplers basically are thin layers of argon equipped with readout electrodes but no absorber.

The electromagnetic end-cap calorimeter has a mechanical structure similar to the barrel calorimeter, but with absorbers arranged like the spokes of a bicycle wheel. However, whereas the barrel calorimeter uses only one type of absorbers and has a constant gap thickness, the end cap uses two types of absorbers, one for the outer wheel (1.4 < |η| < 2.5) and one for the inner wheel (2.5 < |η| < 3.2), with varying gap thicknesses requiring different HV values as a function of η to maintain a constant response with η.

Test-beam measurements show an energy resolution fulfilling the requirements of σE/E= 10%/√E ⊕ 0.4% ⊕ 0.3/E.

Hadronic Calorimeter


The hadronic calorimeter is also a sampling calorimeter. In the barrel part, where the environment is less harsh in terms of radiation dose, it consists of iron absorbers interleaved with plastic scintillator tiles. It is therefore referred to as TileCal. At larger rapidities, in the extended barrel sections, liquid argon technology is used, again because of its intrinsic radiation hardness. The hadronic end-cap calorimeter uses copper as absorber in a parallel-plate geometry while the forward calorimeter uses rod electrodes immersed in liquid argon in a tungsten matrix. In total, the hadronic calorimeter covers the range |η| < 4.9 with a thickness between 11 and 20 interaction lengths depending on η. The granularity is η × φ = 0.1 × 0.1, the total number of channels is on the order of 10000.

The principle of the TileCal is illustrated on the left. The iron tiles have a constant thickness of 3 mm and are placed perpendicular to the colliding beams and staggered in r direction. Both ends of the scintillating tiles are read out by wave length shifting fibers into two separate photo multipliers. The signals produced by the scintillating tiles have atypical rise time of a few ns and a width of about 17 ns. The shaper transforms the current pulses from the photo multipliers into unipolar pulses with a FWHM (full width at half maximum) of 50 ns. The electronic noise is of the order of 20MeV per cell.

Between the barrel cylinder and the extended barrels a 68 cm widegap provides a passage for the cables from the inner detector and the piping from the EM calorimeter. It also houses the readout electronicsfor the EM calorimeter.

The liquid argon calorimeters in the end-caps use 25 - 50 mm thick copper absorbers with gaps of 8.5 mm in between. These gaps are equipped with three electrodes. Thus the maximum drift space is ∼1.8 mm. The electronic noise ranges from 200 to 1100 MeV per channel. The forward calorimeter is integrated in the end-cap cryostat, with the front face about 5 m from the interaction point. It has to provide at least nine interaction lengths of active detector in a very short longitudinal space. Therefore it uses a high density technique of rods, regularly spaced in a metal matrix. In the front region this matrix is made from copper while at the back tungsten is employed. The 250 - 500 μm wide gaps between matrix and rod is filled with liquid argon. The electronic noise in a jet cone of ΔR=0.5 is ∼1 GeV in ET at |η| = 3.2 and drops quickly to 0.1 GeV at |η| = 4.6.