SuperCluster | η
3.6 CMS Hadron Calorimeter
The CMS hadron calorimeter [86] (HCAL) surrounds the ECAL and the inner track- ing system and is partially installed inside of the superconducting solenoid and par- tially outside of it. Similarly to the ECAL, its primary purpose is to measure and fully contain certain particles. It complements the ECAL by further absorbing hadrons. It is designed to (a) acquire the location, time and magnitude of the energy deposited by the incident particles and (b) capture all hadrons so that the only particles origi- nating from the collisions and traveling past the HCAL are muons and neutrinos. It directly contributes to muon identification and precise measurements of jet kinemat- ics and missing transverse energy. It also contributes to the rejection of jets faking prompt isolated leptons and photons. It is vital for a number of important trigger paths including hadronic activity, large missing transverse energy, etc.
The CMS HCAL is a sampling calorimeter. Its active volume consists of 70,000 scintillating plastic tiles interleaved with absorber plates. The short nuclear inter- action length of λI = 16.42 cm, non-magneticity and reasonable cost were the key
properties making brass alloy the absorber material of choice for the compact de- sign of HCAL placed in a strong magnetic field. Brass is complemented by steel for greater structural strength. Wavelength shifting fibers transmit the signal to Hybrid Photodiodes for readout.
The CMS HCAL consists of 4 sub-systems, the HCAL barrel (HB), the HCAL endcap (HE), the hadron outer calorimeter (HO), and the hadron forward calorimeter (HF). The HE and HB lie inside the magnet and the HO and HF outside of it, see Figure 3.29.
2008 JINST 3 S08004
HF HE
HB HO
Figure 5.1: Longitudinal view of the CMS detector showing the locations of the hadron barrel (HB), endcap (HE), outer (HO) and forward (HF) calorimeters.
Table 5.1: Physical properties of the HB brass absorber, known as C26000/cartridge brass.
chemical composition 70% Cu, 30% Zn
density 8.53 g/cm3
radiation length 1.49 cm interaction length 16.42 cm
(Dh,Df) = (0.087,0.087). The wedges are themselves bolted together, in such a fashion as to minimize the crack between the wedges to less than 2 mm.
The absorber (table5.2) consists of a 40-mm-thick front steel plate, followed by eight 50.5- mm-thick brass plates, six 56.5-mm-thick brass plates, and a 75-mm-thick steel back plate. The total absorber thickness at 90 is 5.82 interaction lengths (lI). The HB effective thickness increases with polar angle (q) as 1/sinq, resulting in 10.6 lI at |h|=1.3. The electromagnetic crystal calorimeter [69] in front of HB adds about 1.1lIof material.
Scintillator
The active medium uses the well known tile and wavelength shifting fibre concept to bring out the light. The CMS hadron calorimeter consists of about 70 000 tiles. In order to limit the number of individual elements to be handled, the tiles of a givenflayer are grouped into a single mechanical scintillator tray unit. Figure5.5shows a typical tray. The tray geometry has allowed for construc- tion and testing of the scintillators remote from the experimental installation area. Furthermore,
– 123 –
Figure 3.29: Longitudinal-sectional diagram of the HCAL layout depicting the posi- tions of the four sub-detectors [67].
The HB covers the central region up to pseudorapidity of |η|<1.4. It is radially restricted atr = 1.77 m by the extent of the ECAL from the inside and atr = 2.95 m by the extent of the solenoid from the outside. It comprises 2,304 separately-readout towers with a granularity of ∆η×∆φ = 0.087×0.087 (∆θ×∆φ= 5◦×5◦). Each tower is composed of 17 scintillator tiles interleaved with 16 absorber plates. Traversing HB from the inside out, the absorber consists of a steel front plate 40 mm in thickness, eight brass plates 50.5 mm in thickness, six brass plates 56.5 mm in thickness, and finally a steel back plate 75 mm in thickness. The total absorber thickness is 5.82λIat
pseudorapidityη = 0 and grows to 10.6λIat pseudorapidity|η|= 1.3. The scintillator tiles are 3.7 mm thick, with the exception of the front and back tiles which are 9 mm thick.
The HB is complemented by an additional absorber-scintillator layer, called HO, outside the solenoid for improved hermeticity and containment. The HO matches the coverage and segmentation of the HB. The HO scintillators are 10 mm thick. The absorber is made of iron and is 18 cm thick. It serves as a “tail-catcher” containing hadron showers leaking through the HB and the solenoid. It increases the total effective thickness of the hadron calorimetry including ECAL to ≈10λI.
The HE covers the forward regions with pseudorapidities of 1.3 < |η| < 3.0.
It comprises 20,916 scintillator tiles in 18 layers with a granularity of ∆η×∆φ = 0.087×0.087 for |η| < 1.6 and of ≈ 0.17×0.17 for 1.6 ≤ |η| < 3.0. The absorber plates are made of brass and are 79 mm thick. The front scintillator tile is 10 mm, all the others are 3.7 mm thick. The total effective thickness of the hadron calorimetry is approximately 10λI — the same as for the HB + HO.
The HF is composed of 900 towers and 1,800 channels shared among two modules, one on each side. The modules are cylindrical structures with outer diameters of 1.3 m. They sit at |z|= 11.2 m and extend the pseudorapidity coverage to |η| <5.2.
They use a steel absorber with a depth of 1.65 m. Quartz fibers constitute the active volume. They emit ˇCerenkov light and transmit it to photomultipliers for readout.
They have a diameter of 0.6 mm and are arranged in a square grid with a pitch of 5 mm. The tower segmentation varies from ∆η×∆φ= 0.175×0.175 at |η|= 3.0 to
∆η×∆φ = 0.3×0.35 at|η|= 5.0.
The total resolution of the CMS hadronic calorimetry including both ECAL and HCAL can be expressed as a function of the incident particle energy as:
σ(E)
E = S
√E ⊕C, (3.9)
where S is the stochastic term and C the constant term. The values of these terms were estimated experimentally in test beams [87, 88]. We summarize the results in
Parameter HB, HE HF Unit Stochastic term S 84.7 198 %√
GeV
Constant term C 7.4 9 %
Table 3.3: CMS HCAL calorimetry performance from beam tests.
Table 3.6.
Figure 3.30 shows the jet pT resolution as a function of the jet pT for jets re- constructed using two different algorithms: the so-called calo-jets based solely on calorimeter deposits, and the so-called PF jets based on the particle flow (PF) algo- rithm that combines information from all the subsystems including the inner tracking.
The calo-jet reconstruction is generally more robust while the PF-jet reconstruction is generally more performant. The measured relative jetpTresolution ranges between approximately 18 % for low-pT calo-jets (pT ∼ 50 GeV), and approximately 8 % for high-pT PF jets (pT >200 GeV).
[GeV]
pT
50 100 200
resolution Tjet p
0 0.1 0.2 0.3
total systematic uncertainty MC truth (c-term added) MC truth
data
50 100 200
0 0.1 0.2 0.3
total systematic uncertainty MC truth (c-term added) MC truth
data
total systematic uncertainty MC truth (c-term added) MC truth
data
total systematic uncertainty MC truth (c-term added) MC truth
data
=7 TeV, L=35.9 pb-1
s CMS preliminary 2010
CaloJets R=0.5) (Anti-kT
≤ 0.5 η| 0 < |
=7 TeV, L=35.9 pb CMS
[GeV]
pT
50 100 200
resolution Tjet p
0 0.1 0.2 0.3
total systematic uncertainty MC truth (c-term added) MC truth
data
50 100 200
0 0.1 0.2 0.3
total systematic uncertainty MC truth (c-term added) MC truth
data
total systematic uncertainty MC truth (c-term added) MC truth
data
total systematic uncertainty MC truth (c-term added) MC truth
data
=7 TeV, L=35.9 pb-1
s CMS preliminary 2010
PFJets R=0.5) (Anti-kT
≤ 0.5 η| 0 < |
=7 TeV, L=35.9 pb CMS
Figure 3.30: JetpT resolution in bins of jetpT for jets with pseudorapidities|η|<0.5 as measured in dijet events selected in 35.9 pb−1 of pp collisions at √
s = 7 TeV collected in 2010 (black dots with error bars), compared to the so-called MC truth, the generator-level simulation before and after applying a correction based on the data measurements (dashed and solid red lines, respectively) [89]. The yellow bands represent the total systematic uncertainty on the corrected MC truth. The left panel corresponds to calo-jets, the right panel corresponds to PF jets.