FR9806099
CMS, the Compact Muon Solenoid
E. Locci SPP/DAPNIA , CE Saclay, F-91191 Gif-sur-Yvette
Abstract The CMS detector has been designed to detect cleanly the diverse signatures from new physics by identifying and precisely measuring muons, electrons and photons over a large energy range and at high luminosity. Therefore the design goals of CMS are: a good and redundant muon system, a high quality central tracking, an electromagnetic calorimeter with good energy resolution and high granularity.
189 1 Introduction sions of the detector are: a length of 21.6 m (very forward calorimeters ex- CMS is one of the two multipurpose cluded), a diameter of 5.9 m and a to- experiments which will be installed on tal weight of 14500 tonnes. The mag- the LHC machine. netic flux is returned via a 1.8 m thick The 2-photon and 4-lepton chan- iron yoke which is instrumented with nels are crucial for a discovery of a Higgs muOn stations. The barrel and the en- boson in the minimal supersymmetric capS cover a rapidity range of 6 rapid- model (MSSM); excellent detector res- ity units (\rj\ < 3.0). olution and acceptance are needed. Var- The LHC machine [1] parameters ious channels involve the tau lepton; impose some constraints on the perfor- precise impact parameter measurements manCes of the detector. The high nom- 34 2 1 play an important role here. Events inal luminosity (10 cm~ s~ ) of the with many high energy jets and large 14 TeV proton-proton collider yields a missing transverse energy are the most large interaction rate which must be obvious and model independent signa- reduced by an online event selection. tures in searches for the supersymmet- The detector and the information treat- ric partners of quarks and gluons; the ment must cope with a bunch crossing background due to jet events faking miss- frquency of about 40 MHz . Many su- ing transverse energy because of cracks perimposed minimum bias events (a w or tails in the energy resolution func- 70 mb) will require high granularity de- tion can be kept well below that from tectors with good time resolution. High physics background. The copious pro- granularity is the only way to reduce duction of B mesons at LHC opens the occupancy. In addition detectors will way for significant measurements of CP have to withstand high radiation lev- violation effects in the B system. Fur- els. The crystals and scintillators are thermore by observing the time devel- particularly sensitive to gammas whilst opment of b6 oscillations, the mixing silicone devices are highly sensitive to can parameter x be measured for val- neutrons. The estimated radiation lev- ues up to 20-25. els integrated over the first 10 years, For a high luminosity proton-proton where 5 • 105pb~1 are expected to vary machine it is natural to optimize first from 0.3 • 104 Gy and 2 • 1013 n/cm2 in the muon detection system. The re- the central region to 20 • 104Gy and 50 • 1013 n/cm2 quirement for a compact design led to in the forward region. the choice of a strong magnetic field. A complete description of the CMS The only practical magnet that can gen- detector (figure 1) can be found in [2]. erate a very strong magnetic field is a solenoid. A magnetic field of 4T guar- antees large bending power and good 2i J. lie -LrclCKing momentum resolution for high momen- tum without strong demands on the The design goal of the central tracking chamber space resolution. The large system (figure 2) is to reconstruct iso- radius of the solenoid allows the full lated high pt tracks with an efficiency calorimetry to be located inside the solend?lbetter than 95% and hiSh P1 tracks Hence the coil does not affect the calorime^ithin Jets with an efficiency of better ter performance. The overall dimen- than 90%- Important discoveries may
190 depend on the ability of the tracking (34° Lorentz angle) results in charge system to perform efficient b-tagging spreading over more than one pixel. Us- even at the highest luminosities. This ing charge sharing, resolutions of 15 /j.m has led us to add a few layers of silicon can be obtained. Since the electric and pixel detectors close to the interaction magnetic field are parallel in the end- vertex. Our design goal is to achieve an cap detectors, one cannot use charge impact parameter resolution at high pt sharing as in the barrel. of order 15 /jm. Microstrip detectors are the natural choice for the layers fol- 2.2 The Silicon Microstrip lowing the pixels. The outer tracker Detector layers employ microstrip gas chambers (MSGC). Because of the high magnetic The silicon tracker [4] is required to field a good momentum resolution is have a powerful vertex finding capabil- achievable with only a few, high pre- ity and must be able to distinguish dif- cision, detection planes. On the other ferent interaction vertices at full lumi- hand, good track finding capability sets nosity. The detector provides at least a lower limit on the number of detec- three measuring points on each track. tion planes. Good track finding per- Three layers of silicon microstrip de- formance can be obtained provided at tectors instrument the radial region. To least 10-12 points per track are recorded. avoid dead regions and allow easier align- This leads to a large number of chan- ment, the detector modules are assem- nels (about 95 M -channels) which guar- bled with a few millimeter overlap. The antee an occupancy of all tracker sub- high radiation levels at the LHC are units smaller than 5% at high lumi- serious obstacles to the long term op- nosity. A small number of detection eration of the silicon tracker system, planes results in a relatively small amountThe devices can be safely operated if of matter in front of the calorimeters the temperature is kept below 5° at all not degrading their resolution signifi- times even during maintenance. cantly. 2.3 The Microstrip Gaz Cham- 2.1 The Pixel Detector bers A silicon pixel detector [3] consisting The microstrip gas chambers (MSGC) of 2 barrel layers and 3 endcap layers [5] provide a minimum of 7 hits for high is placed close to the beam pipe with pt tracks over the entire acceptance. the task of: The barrel region is made of 12 layers. assisting in pattern recognition by The intermediate forward region com- providing 2 or 3 true space points per prises 5 detection layers, while the end- track over the full rapidity range of the cap region is made of 9 layers. In order main tracker, to guarantee good resolution for high improving the impact parameter res- momentum tracks, MSGCs in the bar- olution for b-tagging, rel are tilted by 16° to compensate for allowing 3-dimensional vertex recon- charge displacement due to Lorentz an- struction by providing a much improved gle (induced by the 4T magnetic field). z-resolution in the barrel part. The tilt causes a slight assymetry in In the barrel, the large angle of drift
191 reconstruction efficiency between posi- sagitta measurement in the iron yoke. tive and negative charged particles at The last 2 measurements are per- low transverse momenta only. The for- formed with the muon detector alone, ward region consists of an intermedi- independently of the inner tracker. The ate section, where MSGCs and silicon best resolution is obtained when all 3 detectors are arranged in a common methods are combined with the vertex structure, and an encap zone, where constraint of 15 ^xm. The redundancy only MSGCs are used. The length of of the measurements makes the system the strips vary from 5 to 25 cm. They robust against background. Thanks to have a pitch of 200 fim. the large bending power of the solenoid only a moderate resolution per plane is 2.4 The Tracker Performanceneec*ec' anc* tneref°re well-known cham- ber techniques can be used and the ac- The simulation of the tracker has been curacy of the alignment is not a strin- used to evaluate the track finding effi- gent requirement. ciency and the momentum resolution. Several different muon detection tech- The track finding efficiency for a track nologies are considered to provide the of pt > 2 and |T?| < 2.5 is 98% if it is required position determination. In the isolated and 96% if it is included in a barrel, where the expected occupancies 500 GeV jet. Figure 3 shows the mo- and rates are low (< 10Hz/cm2) and mentum resolution versus the pseudo- there is no appreciable magnetic field rapidity for 10 GeV to 1 TeV pt tracks. in the vicinity of most the muon sta- tions, a system of drift tubes (DTs) is the natural candidate. In the endcap 3 The Muon Detector region, cathode strip chambers (CSCs) The muon detector should fulfil 3 ba- have been chosen because of their abil- sic tasks : muon identification, trigger, ity of functionning in a high non-uniform j , , m. magnetic field. Furthermore LSCs can detectoand momentur is placem dmeasurement behind the .calorime 1 he muo- nwithstan . f d, hig, . h. rate, and, th,e signal, fro. m ters and the coil. It consists of 4 muon the anode wires provide good time res- stations interleaved with the iron re- olution for tagging the beam crossing. turn yoke plates. The magnetic flux in Fast dedicated trigger detectors with the iron provides the possibility of an excellent timing capability and reason- independent momentum measurement, able position resolution will also been and the number of multilayer muon sta- installed. Resistive plate chambers (RPCs) tions ensures the reliability of the sys- have been chosen for this purpose in tem. The thickness of absorber be- the barrel and in the endcap region. tween the stations prevents station-to- The barrel muon detector (figure 4) station correlations due to high energy is based on a system of 240 chambers muon radiation. The muon momen- of drift tubes arranged in 4 concen- tum may be measured in 3 ways: tric stations. Every station contains sagitta measurement in the inner a module of drift tubes and one layer tracker, of RPCs for MS3 and MS4 and 2 lay- bending angle measurement imme- ers for MSI and MS2. Each DT mod- diately after the coil, ule consists of twelve planar layers of drift cells: 8 layers parallel and 4 layers
192 perpendicular to the beam. To avoid the same layer and on the precision pointing cracks, chambers and differ- of the staggering inside the superlayer. ent stations are staggered in
193 4 The Calorimetry PbWO4 has a short radiation length and a small Moliere radius leading to 4.1 The Electromagnetic Calaranaie-ect and low volume ECAL ter It has fast scintillation emission. About 85% of the light is collected in 25 ns. The physics process that imposes the The emission spectrum is centered at strictest performance requirements on 510 nm. The temperature dependence the electromagnetic calorimeter is the of scintillation light yield has been mea- intermediate mass Higgs decaying into sured from -193 to +50°. The data are 2 photons. Thus the benchmark against well fitted by a quadratic function for which the performance of the electro- temperatures between -25 and +50°. magnetic calorimeter (ECAL) is mea- The temperature coefficient is -1.98% sured is the di-photon mass resolution. per degree at 20°. A temperature sta- The mass resolution has terms that de- bilisation to a few tenths of a degree pend on the resolution in energy and and monitoring system is thus required the 2 photon angular separation. The for a precision calorimeter. energy resolution Good radiation hardness has been demonstrated on full length PbWO4 crystals doped with niobium Substantial production capability ex- receive contributions from 3 terms: the ists already. It is a relatively easy crys- stochastic term a (w 3%), the constant tal to grow from readily available raw term b (« 0.5%) [7], an < d< the energy i"-"" ••> v~ v-v/u; [_*J' *" -' ""^- ^"^6,7 materiali s equivalent of noise c (approx 25 MeV/channelX'he previous drawback of low light In order to achieve a good energy res- yield is effectively overcome by the re- olution all the contributing terms have cent progress in the commercial pro- to be kept small and should be of the duction of large area Si avalanche pho- same order at the relevant photon en- todiodes. ergies. For the Higgs 2-photon decay the angular term in the mass resolu- 4.1.2 The Detector Design tion can become important. An angu- lar resolution of 50 mrad/VE~ can be The barrel (figure 5) granularity is 432- achieved by measuring the shower cen- fold in 194 equal numbers of crystals. The crystals 4.2 The Hadronic Calorime- are arranged such that gaps pointing ter at the interaction point either between crystals or between modular divisions The hadron calorimeter (figure 6) sur- be avoided. In both direction (r), 195 to be 6.4%. The assembly can be mon- ter will be established by exposing all itored by radioactive source and by in- the modules to test beams. Each tower jecting light from UV lamps. Absolute can be calibrated using QCD 2-jet events, calibration and linearity of the calorime- By requiring that 1 jet be detected in ter will be established by exposing sev- the VF and the other in another part eral modules to a hadron test beam, of the calorimeter system, the inter- The calibration can be transported to calibration between these calorimeter the CMS detector using radioactive source^stems can be verified on the basis and can be maintained at the level of of transverse energy balancing. The 1%. A light laser system will be used long-term response of each calorimeter to monitor the stability of the photode- channel is monitored using minimum- tectors and the associated electronics, bias events. Time variations of the PM Minimum bias events can be utilised to gain are monitored with a laser test maintain uniformity of response and to pulse system, monitor its time stability. Z-jet trig- gers can be used to provide calibration „, . . and the absolute energy scale. ^ Conclusion The design and the expected perfor- 4.3 The Very Forward Calori^e^ of the CMS detector have been presented. Some parts have evoluted The very forward calorimeter (VF) (fig- since the tech?ical Pr°p°sal J2J and sti11 ure 7) improves the measurement of may evolute before the final design. missing transverse energy and enables very forward jets to be identified [10]. These jets are a distinguishing charac- teristics of several important physics process. The VF is required to have moderate energy resolution, with a con- stant term less than 10% for single hadrons, and sufficiently fine granularity to tag forward jets. The VF is required to have a very fast response time and to be particularly radiation hard. The de- tector consists of copper absorber filled with 200 pirn plastic-clad quartz fibres. The fibres are parallel to the beam and are sandwiched between layers of ab- sorber to form wedge shaped modules. In the first 2 segments, the fibres are grouped to form a tower structure. The granularity matches the segmentation of the hadronic calorimeter. Air light pipes guide the cherenkov light to an array of photomultiplier tubes. Absolute calibration of the calorime- 196 Overall dSmeief t 14,59 m Q^rsilKsiftti t 21.86 m : 4 ~ Figure,.!: CMS design for LHC 197 Endcap • Barrel iiiiiiit ™t •\- • l_ T •3.0 V \ A 3.0 HSGCs $.m% Si-Strips o . j.. , Figure 4: Cross sections of the CMS Figure 2: Schematic of inner tracker ...... muon system showing placement of the barrel muon stations and endcap sta- tions. Dimensions are in meters _ WKh vertex point i • TP DESIGN :-• Oiol. DESIGN pf-lOObGeV &•• ;*•*—•" Py.300 GeV » 2 MO" 1 . Pr- 0Sevv..__i^^ 0 0.5 1.0 1.5 2.0 2.5 Figure 3: Momentum resolution of the CMS tracker Figure 5: Electromagnetic calorimeter 198 Figure 6: The tower structure of HCAL for the barrel Tail catcher Hadronic 97cm 2X Figure 7: One half of the quartz fibre VF showing the wedge structure 199