Towards jet reconstruction in a realistic dual readout total absorption calorimeter A Driutti, A Para, G. Pauletta, N. Rodriguez Briones & H. Wenzel Universities and INFN Trieste/Udine & Fermi National Acceleratori Laboratory CALOR2010 G.Pauletta

Outline Introduction: calorimetry in future lepton colliders – limitations and possible solutions Leakage and the Dual Readout (DR) correction Correcting hadron showers for DR and leakage in totally active calorimeters Correcting jets Correcting for magnetic field Related presentations: Hans Wenzel:Simulation of total absorption dual readout calorimetry: principles and performance for single hadrons and jets Hans Wenzel : Enhancing the performance of a total absorption crystal calorimeter with dual readout (DR) by employing Particle Flow techniques Adam Para: Detailed Studies of hadron shower modeling in GEANT4 CALOR2010 G.Pauletta

Introduction The chief purpose of high resolution hadron calorimeters at the future colliders is likely to be di- and multi-jet spectroscopy. The primary figure of merit will be mass resolution and the main limitation will be hadron energy resolution. Compensation by means of Dual Readout (DR) in totally active calorimeters addresses this limitation but the design of collider experiments requires careful optimization of various aspects: scientific, technical and financial and and, the requirements of a calorimeter inside the superconducting coil inevitably leads to a ‘too thin’ calorimeter with leakage fluctuations and magnetic corrections possibly limiting the resolution. A longitudinally segmented total absorption calorimeter offers several possible ways of addressing these limitations Progress in the simulation study of these ideas will be summarized in this presentation CALOR2010 G.Pauletta

Total cal. thickness: 120cm  ~5.5 lint  significant leakage For the simulation: The SiD calorimeter volume is filled with sensitive BGO-like material of uniform composition. Only the (finer) segmentation distinguishes the initial “EM” section from the remaining “Had” volume. Different compositions and segmentations were investigated. EM calorimeter: 6 - 8 layers 5-3 cm thick, with equal transverse segmentation. HAD calorimeter: 9 – 17 layers 10-6 cm thick, with equal transverse segmentation The muon system/tail-catcher is implemented as a 48 layer sampling calorimeter Total cal. thickness: 120cm  ~5.5 lint  significant leakage CALOR2010 G.Pauletta

Leakage and the DR correction * Data is simulated for incident pions and electrons. Both ionisation (Si) and Čerenkov (Ci) energy depositions are recorded for each calorimeter element i and summed to form total raw signals Sr and Cr signals, respectively. Assuming no energy loss for electrons, the corresponding raw signals are normalized to the incident electron energies and the corresponding normalization is applied to the pion signals: The calibrated S and C responses for showers from 20, 50, 100 GeV p- s which impact the calorimeter at 90o to the beam direction are shown. Note the the low - energy leakage tails on the S-distributions (in addition to a peak near zero from punch-through) * See Related presentations listed on slide 3 for details regardin DR correction CALOR2010 G.Pauletta

S/C correlations and the DR correction Plotting S/Ep vs. C/S leads to the correlation used for the DR correction. The need to account for leakage first arises when determining the correct correlation function. A rough method: Divide C/S distrib. into slices Plot S/Ep distrib. for each slice fit only leading edge to exclude leakage The resulting correlation is fitted with: S/Ep=a(C/S)2 + b(C/S) + c CALOR2010 G.Pauletta

Applying DR corrections: _______S______ (C/S)2 + b(C/S) + c Scorr= DR correction functions can be determined in this way as a function of pion energy. (It should be noted that these correction functions do not show pronounced energy dependence) Applying DR corrections: _______S______ (C/S)2 + b(C/S) + c Scorr= effectively corrects the peak energy However a large fraction of events (leakage tails) lose energy from leakage : the leakage energy fluctuates and the fractional fluctuation increases with energy until it exceeds the stochastic term and sets the limit on the achievable energy resolution CALOR2010 G.Pauletta

LK correction using the longitudinal segmentation (SLK algor.) Since leakage depends on shower evolution so that longitudinal segmentation is expected to be useful for leakage correction. At first approximation, use only the outermost layers: plot (S/Ep) against the fractional energy deposited in the two outermost layers. This correlation function is then used to correct for the leakage. The DR correction is then applied This correction results in a symmetric distribution with good resolution (~2%) CALOR2010 G.Pauletta

DR correction applied after muon correction This can be compared with leakage corrections obtained using the muon tail-catcher (MLK algorithm) For this correction, we use the correlation between the the fraction of energy deposited in the calorimeter and the energy detected in the tail-catcher. The correction is applied before the DR correction and also partially corrects for punch-through S corrected only for leakage detected in the tail-catcher DR correction applied after muon correction CALOR2010 G.Pauletta

However, algorithms we developed do not exploit all the information available: the depth segmentation of the energy deposited in our calorimeters can clearly be used to model the shower evolution. Leakage fluctuations depend on the starting point of the hadron shower (“Interaction Depth or ID”) the extension of the shower, so it is expected that full use of the segmentation will improve the leakage corrections The same information might also be expected to improve the DR correction As a first step in this direction, the data was subdivided according to the depth of the segment in which the hadron shower started (the “interaction depth” or ID) . Correlations for the DR and for the MLK corrections were then evaluated separately for each ID CALOR2010 G.Pauletta

The effect of corrections on the overall energy are shown here The effect of these corrections are shown as a function ID (note the improvement, particularly for the outer IDs ), The effect of corrections on the overall energy are shown here (note the reduction of the non-gaussian tails) CALOR2010 ALCPG09 G.Pauletta G.Pauletta 11

Applying corrections to jets - particularly in a realistic environment, is complicated by the mixed content and by jet reconstruction. The weak dependence of the DR correction should help and by restricting the study to jets from single W or W/Z to begin with, one can both try to implement corrections on reconstructed jets (as one must finally do in a realistic environment) and implement corrections independently of jet reconstrunction by summing over the whole event, safe in the knowledge that the event contains only jets. In this spirit: hadron DR and leakage correction functions are first applied to reconstructed jets new DR corrections are then evaluated for comparison, by summing for S and C over all elements of the calorimeter corrections for the effects of the solenoid’s magnetic field are derived from charged particle momenta measured using the tracking system CALOR2010 G.Pauletta

Leakage correction overestmated b c Reconstructed W jets corrected using single p- DR and LK corrections at the corresponding jet energies (slide 7) : fractional energy residuals after DR correction (b) and after DR + LK correction (c) Reconstructed W jets corrected using single p- DR and LK corrections at the corresponding jet energies : invariant mass distributions after DR correction (b) and after DR + LK correction (c) a b c Leakage correction overestmated G.Pauletta CALOR2010

Invariant mass after the DR correction (using new correction function) DR correction functions evaluated from W-jet data by summing over all calorimeter hits (no jet reconstruction). It is smaller for all jet energies and this could be a reflection of the smaller average energy but also of a larger em content. Invariant mass after the DR correction (using new correction function) No pronounced improvement in MW Leakage correction negligible CALOR2010 G.Pauletta

Correction for effect of Solenoid magnetic field These are corrected for by using the tracker – measured momenta. The influence of this correction on the W inveriant mass is plotted below as af function of solenoid field, both befor and after the DR correction (applied using pion correction functions) CALOR2010 G.Pauletta

B = 5 T Magnetic Field Correction applied CALOR2010 G.Pauletta

Conclusions Corrections for leakage and magnetic field effects will be essential in future collider calorimeters – DR calorimeters are no exception. From a study involving single hadrons, on concludes that: In DR calorimeters, leakage must be accounted for in different stages: first in determining the DR correction and then in correcting for energy loss. Longitudinal segmentation is very effective to this end but does not correct for punch-through. Muon detector/tailcatcher assemblies can account for punch-through but tend to over-correct for leakage Despite the differences in content and energy distributions, one finds no great difference between single hadron DR correction functions and those obtained from simulated jet data. However leakage corrections are decidely smaller. More work is needed to extend these studies to more realistic multijet environments. CALOR2010 G.Pauletta

Backup CALOR2010 G.Pauletta

For “clusters” (jet cone = 0,4) For “jets” (jet cone = 0,7) fractional energy residuals after DR correction using jet-based DR correction functions For “clusters” (jet cone = 0,4) For “jets” (jet cone = 0,7) CALOR2010 G.Pauletta

DR correlations from W- jet data CALOR2010 G.Pauletta

CALOR2010 G.Pauletta

CALOR2010 G.Pauletta