CMS Conference Report

Transcription

1 Available on CMS information server CMS CR 2004/067 CMS Conference Report 20 Sptember 2004 The CMS electromagnetic calorimeter M. Paganoni University of Milano Bicocca and INFN, Milan, Italy Abstract The CMS electromagnetic calorimeter is a very challenging detector which aims at providing high precision calorimetry in the LHC environment. It consists of about PbWO crystals that have to operate reliably for at least 10 years, in a high radiation environment. The readout electronics must sustain high data rates. In the last year new radiation hard readout electronics and a new cooling system were adopted and successfully tested. A large effort was also devoted to test the first modules performance in an electron beam, to validate the monitoring system and the calibration strategy. With more than one third of the barrel modules produced, the calorimeter is well into its construction phase. An overview of the calorimeter design and of its construction status will be given, as well as the predicted performance at the LHC. Presented at Xth Vienna Conference on Instrumentation, Vienna, February 16-21, 2004 on behalf of the CMS-ECAL collaboration

2 1 Introduction The Large Hadron Collider (LHC) will deliver proton-proton collisions at a centre-of-mass energy of 14 TeV, with a maximum design luminosity of 10 cm. s The basic requirements for a detector to be operated at LHC are a fast response to match a crossing rate of 40 MHz and a high granularity to cope with the 20 events and 1000 tracks produced on average per bunch crossing. In addition it must be radiation hard because it will be exposed, in 10 years of running at the highest luminosities, to a neutron fluence of about 10 n/cm and to a dose of about 10 Gy in the barrel and to doses up to a factor 100 larger in the endcaps. One of the key issues at LHC is the search for the Higgs boson. The golden channel to discover a Higgs with mass between 100 and 150 GeV/c is the decay. This channel has been considered the benchmark for the electromagnetic calorimeter by the CMS collaboration [1]. 2 The ECAL Structure The CMS collaboration has chosen lead tungstate (PbWO ) crystals for its electromagnetic calorimeter ECAL [2], because of their excellent energy resolution. Thanks to the high density (8.28 g/cm ) and the small radiation length (0.89 cm) of PbWO, the calorimeter is very compact and can be placed inside the magnetic coil needed for the tracker. The small value of the Molière radius (2.2 cm) well matches the very fine granularity needed by the high particle density of the events at LHC. The fast scintillation mechanism (80% of the light is emitted within 25 ns) allows the crystals to be used at the LHC crossing rate of 40 MHz. The drawbacks of PbWO are the low light yield (100 photons/mev, 0.2% with respect to NaI:Tl), which imposes a multiplication mechanism in the photodetector, and the strong temperature dependence of the crystal response (1/L.Y. dl.y./dt - 1.9%/ C). A phase of intensive R&D into the mass production of crystals took place between 1998 and 2000, to reach the optimal crystal quality and uniformity. Crystal growth methods and doping techniques were carefully optimized in order to improve the radiation hardness of the PbWO crystals. The radiation dose at the calorimeter front face foreseen for the LHC running at high luminosity varies from 0.15 Gy/h in the centre of the barrel up to 15 Gy/h in the endcaps. The only effect of the ionizing radiation is the creation of colour centers, due to oxygen vacancies or other defects in the crystals. This reduces the crystal transparency without affecting the scintillation mechanism. The amount of radiation damage depends on the radiation dose rate and saturates after a small integrated dose. The crystal transparency partially recovers in few hours. The loss of transparency is typically 3% for a dose rate of 0.15 Gy/h and can be precisely monitored and corrected for, by injecting laser light inside each crystal. In Fig. 1 the loss of transparency is shown as a function of the wavelength T(%) i n i t i a l a f t e r i r r a d i a t i o n w a v e l e n g t h ( n m ) Figure 1: The loss of light transmission after irradiation, as a function of the wavelength. The ECAL is made out of PbWO crystals. They are arranged into a barrel, covering the central rapidity region ( 1.48) and two endcaps, which extend the coverage up to 3. In the barrel, crystals with a tapered shape, cm front face and 23 cm length are positioned at a radius of 1.29 m. Hence the total depth is 26 2

3 X and the transverse granularity in and is given by In the endcaps, crystals with cm front face and 22 cm length are positioned at a distance from the interaction point of 3.17 m along the beam line. Tilts of 3 both in and in give the structure a geometry slightly off-pointing from the interaction region, in order to improve the hermeticity of the detector. The crystal production for the ECAL barrel has started at the Borogoditsk Techno-Chemical Plant (BCTP), in Russia. Up to now about crystals have been received in the two regional centers of Rome and CERN (see Ref. [3] for the details of the assembly procedure in the regional centers). On reception the crystals are characterized in an automated crystal quality control system. Then they are inserted in fiber glass alveola, which group 5 2 crystals in a single unit, named submodule. A module is the result of the assembly of 40 or 50 submodules, as can be seen in Fig. 2. Figure 2: The mounting of submodules inside an ECAL module. A supermodule is made out of 4 modules, with a total number of 1700 crystals. The whole ECAL barrel is made out of 36 supermodules. At the beginning of 2004, already 48 modules had been produced and 8 supermodules assembled. 3 The ECAL Physics Goals The Higgs mass resolution in the channel depends linearly on the photon energy resolution achieved.. The energy resolution is parametrized as: The target values for the ECAL are: 2.7% GeV for the stochastic term, limited by the photoelectron statistics; 200 MeV for the noise term corresponding to a reconstructed cluster, which depends on the photodetector dark current, the electronics noise and, at high luminosity, the event pile-up; 0.5% for the constant term, which is related to the longitudinal shower containment, the uniformity of the light collection in the crystals and the precision of the intercalibration. At high energies the most relevant contribution to the energy resolution comes from the constant term. The challenging goal of keeping it very small can be reached only provided that the intercalibration between crystals is very precise and there is a very tight control, at the level of few per mille, on instabilities and non-uniformities in the detector response. This puts severe requirements on the control of the temperature stability (cooling system) and on the following of the radiation damage (monitoring system). 4 The ECAL Photodetectors ECAL requires fast, radiation hard photodetectors with an internal amplification, and capable of operating inside a strong magnetic field (4 T). 3

4 For the barrel the Avalanche Photo Diodes (APD), solid state devices, have been developed together with Hamamatsu Photonics. The quantum efficiency around the PbWO emission peak ( = 420 nm) is about 80%. The principle of operation of the APD is visible in Fig. 3. γ E Si 3 N 4,, SiO 2, contact p ++ photon conversion p e - acceleration n e - multiplication n - e - drift n ++ e - collection contact Figure 3: The principle of operation of the APD. Each crystal is coupled to two APDs, for a total area of 50 mm and consequently 4000 photoelectrons are produced per GeV of deposited energy. A multiplication of about 50 is achieved inside a very small region of high electric field (about 380 volts over 5 m). The closeness of this region to the front surface and its thinness make the nuclear counter effect negligible and allow the device to operate even in an intense transverse magnetic field. The gain depends strongly on both the bias voltage (1/M dm/dv 3%/V) and the temperature (1/M dm/dt -2.4%/ C), which imposes severe stability requirements on both parameters: 0.05 C and 20 mv. The noise term is expected to be about 150 MeV at the beginning of LHC and to double after 10 years of running, because of the rise of the dark current under neutron irradiation. More than 90% of the final production has been already completed by Hamamatsu Photonics. In the endcaps the neutron fluence is too high for the APDs, but the direction of the magnetic field, parallel to the photodetectors, allows the use of Vacuum PhotoTriodes (VPT). They are single stage photomultipliers with a fine metal grid anode, an active area of about 280 mm, 20% quantum efficiency and a gain of about 10. More than 40% of the VPT final production has been completed by the RIE corporation (St. Petersburg, Russia). 5 The ECAL Readout Electronics Full custom integrated circuits were developed for the ECAL readout, because of the very challenging requirements on radiation hardness, high speed (40 MHz), wide dynamic range (30 MeV to 1 TeV) and very low noise. In 2002 the ECAL electronics design was reviewed and modified to reduce the very high costs of the optical data links and to profit of the new radiation hard 0.25 m CMOS technology, which allows the use of ASICs inside the detector. In the new architecture digitized data are stored in the front-end electronics (memories of 5 FENIX chips). The trigger process includes the transfer of the trigger primitives to the off-detector electronics, their processing by summing up the energies in trigger towers made of 5 5 crystals and the transfer of the trigger decision back to the front-end. The whole trigger process takes place within 3 s. On reception of a L1 trigger, the data from the 25 crystals of the trigger tower are transferred serially to the off-detector electronics. All the transfer processes take place by means of a Gigabit Optical Link (GOL) and three optical fibers (used respectively for the transfer of the trigger primitives, for the transfer of the data and for the clock and control signals). This allows a strong decrease of the electronics cost, because of the simplification of the off-detector system and the reduction, by a factor 3/25, of the number of optical links, with respect to the previous design. Each trigger tower is made of five Very Front End (VFE) cards, one Low Voltage Regulator (LVR) card, one Mother Board and one Front End (FE) card. In the VFE card, the signals coming from the APDs are amplified by a Multi Gain Pre Amplifier (MGPA) and digitized by a 40 MHz, 12 bits ADC. Both MGPA and ADC chips have been produced in the new 0.25 m CMOS technology. A more detailed discussion of the electronics can be found in Ref. [4]. 4

5 6 The Cooling and Monitoring Systems Due to the dependence on the temperature of both the crystal light yield and the APD gain, the temperature of the crystals must be kept stable at the level of 0.05 C. The cooling system [5] couples thermally the on-detector electronics cards, through gap pad and gap filler, to aluminum bars in which water pipes are embedded. The aim of this system is to remove, as much as possible, the power dissipated by the electronics ( 2.5 W/channel). Extensive laboratory tests, performed during the summer of 2003, validated the cooling system, by showing that the required stability of 0.05 C could be achieved. In situ calibration of ECAL will be obtained by physical events with W and Z [6]. It is estimated that, at low LHC luminosity, about 35 days will be needed to intercalibrate crystals with a 0.3% precision. The evolution of the calibration coefficients, due to the radiation damage, will be followed, between two in situ calibrations, by means of the monitoring system [7]. It consists of two lasers, which distribute four wavelengths (440 nm, 495 nm, 709 nm and 800 nm) to the single crystals. The stability of the lasers is precisely checked by means of radiation hard reference PN diodes, which have a stability better than 0.1%. The precision of the follow-up of the calibration coefficients with the monitoring system is at the level of 0.4%. 7 The Testbeam Results In the summers of 2002 and 2003, respectively 100 channels with the old electronics and 50 channels with new electronics were exposed to electron beams of various energies, for many LHC irradiation cycles. Both testbeams proved to be successful system tests of the ECAL and validated the cooling and the monitoring system, the integration between mechanics and electronics and the testbeam calibration procedure. Testbeam results showed that, even for crystals not pre-calibrated on a beam, the calibration can be extrapolated from the laboratory light yield measurements, with 4.5% accuracy. Furthermore, a detailed analysis of the crystal irradiation showed the validity of the relation, where and are the signals obtained with the electron beam and with the laser light during irradiation, while and are the starting values (see Fig. 4). Figure 4: The correlation between the signal from 120 GeV electrons and the signal from laser light, during irradiation at the testbeam, for a dose rate of 0.25 Gy/h. The fitted exponents of this power law for the different crystals were found to have a spread of 6.3%, around an average value of = 1.55, as shown in Fig. 5. For a dose rate of 0.25 Gy/h the drops in the signals from 120 GeV electrons and from laser light were measured 5

6 Figure 5: Dispersion of the fitted values for 19 crystals. to be respectively about 5%, and about 3%. 8 Summary The electromagnetic calorimeter of CMS is well into the construction phase, which is expected to last for three years. The electronics architecture has undergone a major revision and new cooling and monitoring systems have been designed and validated. From testbeam results it has become clear that the full potentiality of PbWO crystals can be exploited for a precision calorimetry at the LHC. References [1] CMS Collaboration, The CMS Technical Proposal, CERN/LHCC 94-38, [2] CMS Collaboration, The CMS Electromagnetic Calorimeter Technical Design Report, CERN/LHCC 97-33, [3] M. Diemoz, The electromagnetic calorimeter of the CMS experiment, Nucl. Phys. B Proc. Suppl., 125 (2003) 100. [4] G. Dissertori, Electromagnetic Calorimetry and e/ performance in CMS, Proceedings of the 2003 LHC Symposium, Fermilab, May 2003, CMS CR-2003/024. [5] P. Govoni, Performance of the cooling system of ECAL CMS, Proceedings of the 8 Conference on Astroparticle, Particle, Space Physics, Detectors and Medical Physics Applications, Como, Oct [6] R. Paramatti, Calibration of the CMS electromagnetic calorimeter, Nucl. Phys. B Proc. Suppl., 125 (2003) 107. [7] M. Dejardin et al., The CMS ECAL monitoring system, CMS IN-2002/12. 6