Track Segments in Hadronic Showers: Calibration Possibilities for a Highly Granular HCAL
TTrack Segments in Hadronic Showers: CalibrationPossibilities for a Highly Granular HCAL
Frank Simon , for the CALICE Collaboration1- Max-Planck-Institut f¨ur Physik, F¨ohringer Ring 6, 80805 M¨unchen, Germany2- Excellence Cluster ’Universe’, Boltzmannstrasse 2, 85748 Garching, GermanyThe CALICE collaboration has constructed a highly granular hadronic calorimeterbased on small scintillator tiles read out with silicon photomultipliers. With this de-tector, data was taken at CERN and at Fermilab. The high granularity of the detectorallows the identification of minimum-ionizing track segments within hadronic showers,demonstrating the imaging capabilities of particle flow calorimeters. These tracks canbe used for the cell-by-cell calibration of such a calorimeter. The possibility to calibratea complete ILC calorimeter with such track segments in the absence of muons is alsoinvestigated. The goal of the CALICE experimental program is to establish novel technologies for calorime-try in future collider experiments and to record electromagnetic and hadronic shower datawith unprecedented three dimensional spatial resolution for the validation of simulationcodes and for the test and development of reconstruction algorithms. Such highly gran-ular calorimeters are necessary to achieve an unprecedented jet energy resolution at theInternational Linear Collider [2] using particle flow algorithms [3].The CALICE test beam setup [4] consists of three separate sampling calorimeters: asilicon-tungsten electromagnetic calorimeter (ECAL), an analog scintillator-steel hadroncalorimeter (AHCAL) and a tail catcher/muon tracker (TCMT). The scintillator cells inthe latter two detectors are individually read out by silicon photomultipliers (SiPMs) [5].This setup has been tested extensively in electron, muon and hadron beams at CERN andat Fermilab. Figure 1 shows the schematic setup of the CALICE detectors in the CERN H6test beam area, where data was taken in 2006 and 2007.The AHCAL consists of 38 layers, each with a 1.6 cm thick steel absorber plate and ascintillator layer build out of individual tiles housed in steel cassettes with a wall thickness
Sc1 Sc2Sc4 Sc3 Muon TriggerDrift ChambersCherenkovDetector Scintillators ECAL HCAL TCMTBeam
Figure 1: Schematic of the CALICE experimental setup at CERN, with electromagneticand hadronic calorimetry as well as a tail catcher and muon tracker downstream of thecalorimeters.
LCWS/ILC 2008 a r X i v : . [ phy s i c s . i n s - d e t ] F e b e a m identified track segments Figure 2: Event display of a hadronic shower in the AHCAL initiated by a 25 GeV π − .All cells in the calorimeter with an energy deposit above 0.4 MIP are shown, the identifiedtrack segments are highlighted in red. One of these tracks is the incoming π − before thefirst interaction. In addition two tracks of particles created in the shower are identified, oneof which crosses over into the TCMT.of 2 mm, resulting in a total absorber thickness of 2 cm per layer. The lateral dimensionsare roughly 1 × , the total thickness amounts to 4.5 nuclear interaction lengths. Thefirst 30 layers of the calorimeter have a high granular core of 10 ×
10 tiles with a tile sizeof 30 ×
30 mm , an outer core composed of 60x60 mm tiles, and border tiles with a size of120 ×
120 mm . The last 8 layers use only 60 ×
60 mm2 tiles in the core, and the large bordertiles. In total, this amounts to 7608 channels. The light in each scintillator cell is collectedby a wavelength shifting fiber, which is coupled to the SiPM. The SiPMs, produced by theMEPhI/PULSAR group [5], have a photo-sensitive area of 1.1 × containing 1156pixels with a size of 32 × µ m .In this paper, a preliminary analysis of hadronic data, which explores the possibilityto use the high granularity of the calorimeter to identify minimum-ionizing track segmentswithin hadronic showers, is presented. This capability of imaging calorimeters can also beused for the cell-by-cell calibration of the detector in a collider experiment. First simulationstudies within the ILD detector concept have been performed to further investigate this. The high granularity of the active layers in the hadronic calorimeter and the cell-by-cellreadout gives the CALICE detectors unprecedented imaging capabilities. This is exploitedto study the topology of hadronic events in detail. Track segments created by chargedparticles produced within the hadronic shower can be identified, provided the particles travelan appreciable distance before interacting again and are separated from other activity in thedetector.
LCWS/ILC 2008 -1 slope [K -0.1 -0.08 -0.06 -0.04 -0.02 0 0.02 0.04 - c e ll s / . K ± Mean: -0.036 0.001 ± RMS: 0.014
CALICE Preliminary +error error(hSlope - muSlope)/ -4 -3 -2 -1 0 1 2 3 40510152025 Gaussian fit:Mean = 0.161 = 1.014 σ CALICE Preliminary
Figure 3:
Left:
Distribution of temperature slopse of the cell response for track segmentsin hadronic showers.
Right:
Comparison of temperature slope determined from hadronicevents and from muon events, demonstrating consistency of the two calibration methods.Figure 2 shows an example event of a 25 GeV π − in the AHCAL, the other detectorsof the CALICE setup are not shown. In this particular event the track of the incoming π − before the first interaction is seen, together with two additional track segments. Shownare all cells with a recorded energy above 0.4 times the energy corresponding to the mostprobable value of the energy loss of minimum ionizing particles. The response of all cellsin the detector was calibrated as described in [6]. Track segments are found from isolatedhits, cells with an energy deposit above a threshold of 0.4 MIP that do not have an energydeposit above threshold in the direct next neighboring cells in the same layer.The track segments identified in hadronic showers have a high quality, and are suitablefor detector calibrations via the extraction of the most probable value of the energy loss ineach cell along the track. This feature was used to investigate the temperature dependenceof the detector response, exploiting the large range of ambient temperatures in the presentlyavailable dataset. Further detail on the first preliminary studies with a subset of the availabledata can be found in [6]. Figure 3 shows the distribution of the temperature dependenceof the cell response, determined from track segments in hadronic events. This distributioncontains 250 detector cells, a small subset of ∼
3% of all channels, given by the availablestatistics, which are only significant in the center of the detector near the shower axis. Alsoshown is the comparison between the temperature dependence extracted with the novelmethod of track segments in hadronic events compared to the extraction from calibrationdata taken with muons. The distribution of the difference of the two methods, normalizedby the measurement errors, shows good consistency of the two methods. This validates themethod of extracting calibration quality data from hadronic events.
The cell-by-cell calibration of a complete hadronic calorimeter for the ILC, with approx-imately 5 million scintillator cells in the barrel calorimeter, is a formidable task. Cosmicmuons will likely not be sufficient, since very high statistics are needed due to the high granu-larity, and the muon flux is limited in underground locations. In addition, the power-pulsing
LCWS/ILC 2008 cheme used for all ILC detector electronics only allows signal sampling for less than 1% ofthe time. Thus, alternative calibration mechanisms are necessary. To monitor short-termvariations of the SiPM response, a LED calibration system is foreseen, which is already usedin the CALICE AHCAL [6]. The study of long-term variations and the cell-by-cell intercal-ibration of the detector could potentially be performed using standard data events with themethod of track segment identification, as discussed in the present paper. The quality ofthe track segments that can be identified within hadronic showers is sufficiently good thatthese can be used for a cell-by-cell calibration of the detector, based on the reconstructedvalue of the most probable energy deposit of these minimum-ionizing particles. layer t il e s on t r a ck / e v en t CALICE Preliminary
Figure 4: Number of cells crossed by identifiedhadronic track segments versus depth in thehadronic calorimeter.A prerequisite for the use of hadronicevents as a calibration tool are sufficientstatistics also in the last layers of thecalorimeters, where the expected particlemultiplicity is significantly reduced due tothe large amount of material in front. Fig-ure 4 shows the number of cells crossedby identified minimum-ionizing tracks perevent as a function of the depth in theAHCAL. Taking a muon contamination ofthe beam of a few percent into account, thisshows that still one in ten events with a 25GeV hadron has a track in the last layer ofthe detector. The situation in the CALICEsetup corresponds to a realistic collider de-tector, since there is a 30 X silicon tung-sten ECAL upstream of the AHCAL. Therise of the number of identified tracks in thefirst few layers of the detector is in part dueto the tracking algorithm, which in this preliminary study can not track across detectorboundaries and thus does not find tracks that cross from the ECAL into the HCAL. Thisaffects the first six layers in particular, since a minimum track length of six layers is requiredin the algorithm. In addition, the higher particle density in the early parts of the showermakes tracking more difficult. Overall, on average 1.7 track segments with an average lengthof 10.5 layers are identified per event for 25 GeV π − [6].To evaluate the possibilities of tracking within the hadronic calorimeter of a future ILCdetector, a first study on simulated data for the ILD detector concept has been performed.This detector concept uses a SiW ECAL and a scintillator tile HCAL, very similar to theCALICE setup discussed here. The tracking algorithm developed for the CALICE HCALhas been applied to simulated data, such as e + e − → q ¯ q at √ s = 500 GeV and at the Z resonance, to see if tracking is in principle possible in ILC data events. First results areencouraging, and yield similar track length distributions as observed in the CALICE data.Further optimization is necessary to account for the presence of a strong magnetic field, andfor differences in the geometry. Currently, the total number of events needed to get sufficientstatistics for a calibration of the first 20 layers of the calorimeter seems to be in the orderof tens of millions when requiring at least ∼ √ s = 500 GeV due to the small e + e − cross section at full ILCenergy. At the Z resonance, the cross section is much more favorable. However, the lower LCWS/ILC 2008 nergy in combination with the high magnetic field reduces the statistics in the later layersof the detector in comparison to the situation in high energy events. The preliminary studiesindicate the need of the order of 1 fb − of Z resonance data for a precise intercalibrationof the first 15 layers of the HCAL barrel. The quality of the calibration in the later layerswill decrease with decreasing statistics in those detector cells. By combining the data ofseveral cells, the needed statistics can be reduced significantly, with the caveat of reducedgranularity of the calibration. Such a strategy might be an attractive option to monitorpossible long-term variations in the detector.These preliminary investigations suggest that track segments identified in the hadroniccalorimeter in regular data events can indeed be used for the cell-by-cell calibration of thedetector. However, the need for high statistics likely requires the availability of a large sampleof Z resonance data for this calibration approach. Possibilities for further improvementsof the track identification, which would lead to a reductions of the required integratedluminosity for calibration using the hadronic track segments technique, are currently understudy. The CALICE collaboration studies highly granular calorimeters for detectors at the futureInternational Linear Collider. We presented a first preliminary study of the tracking ca-pability of the scintillator tile analog hadron calorimeter with SiPM readout. Due to theextremely high granularity, the identification and reconstruction of minimum-ionizing tracksegments of charged particles created within hadronic showers is possible. These track seg-ments are used for detailed detector studies, such as the investigation of the temperaturedependence of the response, yielding results consistent with those obtained with muon data.The track segments can also be used for a cell-by-cell calibration of the calorimeter at afuture ILC detector. First proof-of-principle studies using standard data events simulatedfor the ILD detector concept suggest that track identification of sufficient quality is indeedpossible. Tracking within hadronic showers could be a viable calibration strategy, howeverlikely requiring a large Z resonance dataset to obtain sufficient statistics in each detectorcell. References [1] Presentation: http://ilcagenda.linearcollider.org/contributionDisplay.py?contribId=109&sessionId=22&confId=2628 .[2] J. Brau et al. , arXiv:0712.1950 [physics.acc-ph] (2007).[3] M. Thomson, J. Phys. Conf. Ser. , 092032 (2008).[4] G. Eigen, AIP Conf. Proc. , 565 (2006).[5] G. Bondarenko et al. , Nucl. Instrum. Meth.
A442 , 187 (2000).[6] F. Simon, arXiv:0811.2431 [physics.ins-det] (2008)., 187 (2000).[6] F. Simon, arXiv:0811.2431 [physics.ins-det] (2008).