Combined Calorimeter Combined Calorimeters Properties Part. V. L. Morgunov DESY and ITEP HCAL meeting, DESY, July. V.L. Morgunov HCAL meeting, DESY, July.
Combined Calorimeter Introduction Combined calorimeter means ECAL + HCAL + TCMT; and all their combination, such as: ECAL + HCAL only: length is of about 5 nuclear interaction units (conventional for ILD); ECAL + HCAL + first 9 layers of TCMT: length is of about 6 nuclear interaction units (proposed for ILD); ECAL + HCAL + TCMT: full CALICE possible calorimeter length is of about 9 nuclear interaction units. Position cuts ECAL electromagnetic calibration check HCAL electromagnetic calibration check Muon stability check Energy linearity Energy resolution Remark: We will discuss of three different types of combined calorimeters to get a sense about a calorimeter length needed for ILD. V.L. Morgunov HCAL meeting, DESY, July.
cge Mean x.37 Mean y.57 RMS x 5.59 RMS y 7.496 Integral.78e+6 Combined Calorimeter 3 Beam position cut Pictures show different positions at different runs during CERN run beam-test period of data taking. X-Y ECAL cge Mean x 3.743 X-Y ECAL cge Mean x -4.8 X-Y ECAL Mean y 9.4 Mean y -.888 5 5 RMS x 6.9 RMS y.58 Integral 4.5e+5 5 5 RMS x.34 RMS y 7.858 Integral 7.593e+5 3 5 3-5 - -5 - -5-5 -5 - -5 5 5-5 -5 - -5 5 5 - - -5 5 A small transverse size of the ECAL (8x8 cm ) in compare with hadron shower size leads to the shower energy losses at dead material surrounded of electromagnetic calorimeter. In the case of combined calorimeter such a losses are occurred near by the maximum of longitudinal distribution of hadron shower. Energy density [GeV/Lambda] 8 Longitudinal Curve [GeV/Lambda] 7 6 5 4 3 4 6 8 Calorimeter depth in nuclear length without position cut A beam impact point was chosen to be at the center of ECAL sensors structure in small window 4x4 cm to reduce an influence of energy losses. This cut is also slightly reduce of influence of gaps between ECAL wafers. A position cut is the main reason of small statistics for many runs; these runs were dropped out. V.L. Morgunov HCAL meeting, DESY, July.
χ / ndf 47. / 3 A -.789 ±.86 B.8 ±.764 C -5.994e-6 ±.353 Combined Calorimeter 4 ECAL calibration check, individual runs E_rec / E_beam.4.3.. ECAL calibration, linearity Energy resolution [GeV].5 ECAL calibration, resolution.99.98.5 Sigma(E) = A*sqrt(E) + B*E + C.97.96 3 4 5 A = 7.9 % B =.8 % C = -5 kev 3 4 5 ECAL linearity is inside ± % of accuracy. ECAL energy resolution is of about 8% with energy dependent term that is of about % and with zero noise term. I do not understand a reason of spread of the resolution points. They are out of statistical accuracy. Let say for now: that is an estimation of our systematical errors for the ECAL resolution for electrons. It might be the effect of the beam particle type selection. Need more investigation to be understood. V.L. Morgunov HCAL meeting, DESY, July.
Combined Calorimeter 5 ECAL calibration check, combined runs E_rec / E_beam.4. Sigma(E) / E.4.. χ / ndf 46.3 / 8 Stochastic.8 ±.97 Calibr..44 ±.787 Noise -8.7e-9 ±.653.8.6.98.4.96. 3 4 5 3 4 5 Combined runs means: one histogram was filled for all type of runs with the same beam energy. V.L. Morgunov HCAL meeting, DESY, July.
Mean x -.5 Mean y 3.3 RMS x 8.38 RMS y 6.455 Integral.47e+6 Mean x -.5 Mean y 3.3 RMS x 34.55 RMS y 3.3 Integral.47e+6 Mean x 4.95 Mean y.6 RMS x 9.7 RMS y 5.487 Integral.375e+6 Combined Calorimeter 6 HCAL calibration check, individual runs Electromagnetic shower energy for every particle was collected inside a cylinder around shower center with radius of 5 cm, and length of HCAL layers to reduce an effect of HCAL m 3 noise. Two beam position with the same beam energy at different runs. cgh Y position HCAL [mm] 4 3 X-Y HCAL cgh cgxyh 4 3 - - -3-4 -4-3 - - 3 X position HCAL [mm] Different sets of calorimeter cells are at work to get a positron shower energy. At some position the dead cells could be occurred near by the center of the electromagnetic shower in HCAL, at another one it is not the case. V.L. Morgunov HCAL meeting, DESY, July.
χ / ndf 47.5 / 3 A -.8 ±.8 B -.4 ±.597 C.839e-7 ±.6 Combined Calorimeter 7 HCAL calibration check, individual runs E_rec / E_beam..5.95 HCAL calibration, linearity Energ resolution (E_beam) [GeV].5.5 HCAL calibration, resolution Sigma(E) = A*sqrt(E) + B*E + C A =. % B =. % C =.8 kev.9 3 4 5 3 4 5 6 Two clear visible structures at the amplitude ratios are the consequence of the two different position of the beam spot. The effect is order of percent in energy amplitude. (it is not so big). In the case of the hadron beam such an effect will be reduces by the spreading of the position of each individual electromagnetic shower inside the hadron one up to be neglected. Clear visible effect of SiPM saturation with high energy density. HCAL linearity is inside ± % of accuracy without saturation effect; and with systematics that is of about %. HCAL energy resolution for positrons is of about % with energy dependent term that is of about % and with zero noise term. V.L. Morgunov HCAL meeting, DESY, July.
Combined Calorimeter 8 HCAL calibration check, combined runs E_rec / E_beam.4. Sigma(E) / E..8 χ / ndf 38.8 / 5 Stochastic.77 ±.878 Calibr..73 ±.37 Noise.55e-6 ±.669.6.98.4.96..94 3 4 5 6 3 4 5 Combined runs means: one histogram was filled for all type of runs with the same beam energy. V.L. Morgunov HCAL meeting, DESY, July.
Combined Calorimeter 9 ECAL and HCAL electron/positron resolution This picture should be never shown at public. Resolution (E_beam) [GeV].8.6.4..8 HCAL EM resolution ECAL EM resolution.6.4 ILD expected ECAL EM resolution. 3 4 5 As the result: the resolution of combined calorimeter for hadrons will be much worse then expected. V.L. Morgunov HCAL meeting, DESY, July.
Combined Calorimeter Muon energy deposition at calorimeters Muon track energy for every particle was collected inside a cylinder around track center with radius of 7 cm at HCAL to keep a small noise contribution. Muon energy deposited at calorimeters [GeV] 4 3.5 3.5.5.5 ECAL ECAL+HCAL ECAL+HCAL+TCMT HCAL Different markers are different types of runs. V.L. Morgunov HCAL meeting, DESY, July.
Combined Calorimeter Reconstructed energies with pure electromagnetic calibration The energies are the simplest sum of the responses measured in different parts of the detector in GeV. E sum = E ECAL [GeV ] + E HCAL [GeV ] + E TCMT [GeV ] Gaussian fit of this sum in range of ±σ around maximum probable energy value was used to define reconstructed energy and its resolution. Reconstructed energy [GeV] 8 6 4 8 Reconstructed energy [GeV] 6 4 4 6 8 4 6 8 All three detectors show almost the same behaviour in term of linearity with a slope of about.8. Later we will use this slopes to correct energy to the exact beam values. Slightly smaller amplitude is visible for the shortest calorimeter; which is mostly affected by longitudinal leakage. V.L. Morgunov HCAL meeting, DESY, July.
Combined Calorimeter Response linearity with pure electromagnetic calibration E_rec / E_beam..5..5.95 ECAL + HCAL+ + TCMT ECAL + HCAL + TCMT9 ECAL + HCAL (E_rec - noise) / E_beam..5..5.95.9.9.85.85.8.8.75.75.7.7 Left plot shows an initial state without taking into account a noise in all calorimeters. The right hand side plot shows the same amplitudes, but with subtraction of estimated noise (constant). Noise for ECAL+HCAL calorimeter was taken as. GeV.!!! Noise for ECAL+HCAL+TCMT9 calorimeter was taken as.3 GeV.!!! Noise for ECAL+HCAL+TCMT calorimeter was taken as.6 GeV.!!! Dramatic changes in the linearity of the calorimeters at low energies are clear visible. V.L. Morgunov HCAL meeting, DESY, July.
Combined Calorimeter 3 Linearity: One factor =. Different calorimeters are marked by different colors..*(e_rec-e_noise) / E_beam.4..98.96.94 The spread are mainly the systematic uncertainties due to beam quality at particular run, HCAL temperature correction map (that is not perfect yet) and accuracy of calibration on muons. A CALICE combined calorimeters (all of them) are linear in energy range 8 GeV at the level of 3%! Here we do not see yet any difference between rather different calorimeter devices. V.L. Morgunov HCAL meeting, DESY, July.
Combined Calorimeter 4 Energy resolution in natural scale [GeV] Different calorimeters are marked by different colors. One factor =.. Sigma(E_beam) [GeV] 4 8 ECAL + HCAL + TCMT9 ECAL + HCAL 6 ECAL + HCAL + TCMT 4 4 6 8 4 6 8 Resolution of shorter calorimeter is worse than resolution of the longer one, as it was expected. Difference is significant at high energies and it is reaching almost 5% at 8 GeV in natural scale. V.L. Morgunov HCAL meeting, DESY, July.
Combined Calorimeter 5 Energy resolution Different calorimeters are marked by different colors. One factor =.. Sigma(E_beam) [GeV] ECAL + HCAL ECAL + HCAL + TCMT9 ECAL + HCAL + TCMT The same picture as previous slide, to show: there is no difference at low energies, as it was expected. V.L. Morgunov HCAL meeting, DESY, July.
Combined Calorimeter 6 Energy resolution: comparison Sigma(E_beam) [GeV] 8 6 ECAL + HCAL + TCMT Atlas: A = 5%, B = 5.9% Atlas: A = 48%, B = 5.% 4 Katja, Marina data with FIP at the first 5 HCAL layers 4 6 8 4 6 8 Beam energy [Gev] Katja and Marina data were collected for the same combined calorimeter configuration, i.e. ECAL+HCAL+TCMT, but for events with shower starting points at the first five layers of HCAL. Longitudinal leakage at this case was minimized. The significant effect of a leakage on the resolution at high energies is clear visible here. V.L. Morgunov HCAL meeting, DESY, July.
Combined Calorimeter 7 Combined calorimeters: CMS banana and ATLAS line Energy measurement for 5 GeV pion in ECal vs. that in HB at CMS. Understanding the performance of CMS calorimeter by Seema Sharma at http://www.springerlink.com/content/8h5hk4r6636x/ Correlation between LAr and TileCal energy, GeV pion Study of the ATLAS central calorimeters response to pions in 4 combined test beam by V. Giangiobbe at http://arxiv.org/pdf/physics/68 V.L. Morgunov HCAL meeting, DESY, July.
Combined Calorimeter 8 CALICE combined calorimeter line HCAL+TCMT energy sum [GeV] 6 5 4 3 HCAL+TCMT energy sum [GeV] 6 5 4 3 by three factors 3 4 5 6 ECAL energy sum [GeV] 3 4 5 6 ECAL energy sum [GeV] Individual coefficients are: C ECAL =.; C HCAL =.3; C TCMT =.3 the same for any energy. Technique of calorimeters weightings was developed for ILC whole detector. See Calorimeter energy calibration using the energy conservation law at http://www.springerlink.com/content/r4556u46r8w67/ V.L. Morgunov HCAL meeting, DESY, July.
Mean x 7.437 Mean y.597 RMS x.39 RMS y.7654 Integral.94e+4 Combined Calorimeter 9 Energies correlations, GeV, leakage TCMT energy sum [GeV] enreh_tany 8 6 Energy ECAL+HCAL vs TCM gev TCMT energy sum [GeV] Energy TCM vs ECAL+HCAL gev 8 6 enreh_thd Mean x 9.86 Mean y.685 RMS x.68 RMS y.968 Integral 5848 4 4-4 6 8 ECAL+HCAL energy sum [GeV] EM calibration only 4 6 8 ECAL+HCAL energy sum [GeV] Weighting with 3 factors V.L. Morgunov HCAL meeting, DESY, July.
Mean x 33.5 Mean y 3.8 RMS x 34.6 RMS y.3 Integral.55e+4 Mean x 6.8 Mean y 8. RMS x 34.7 RMS y 6.98 Integral.895e+4 Combined Calorimeter Energies correlations, 8 GeV, leakage TCMT energy sum [GeV] 8 6 4 Energy ECAL+HCAL vs TCM gev enreh_tany TCMT energy sum [GeV] 8 6 4 Energy TCM vs ECAL+HCAL gev enreh_thd 8 8 6 4 6 4 4 6 8 4 6 8 ECAL+HCAL energy sum [GeV] EM calibration only 4 6 8 4 6 8 ECAL+HCAL energy sum [GeV] Weighting with 3 factors V.L. Morgunov HCAL meeting, DESY, July.
χ / ndf 43 / 79 p.5383 ±.56 p.65 ±.6 p -8.636e-9 ±.7 Combined Calorimeter About energy resolution fitting Sigma(E_beam) / E_beam.4.35.3.5..5..5 Standard representation of energy resolution is σ(e)/e = A/ E B C/E, where: A, B and C are dimensionless parameters; This equation (red curve) does not fit well to data at low energies. As a result: we cannot extract a correct parameters from this fit function. This is a property of a combined calorimeter with rather different sampling fractions of each part of it. A longitudinal motion of the maximum energy deposition to the ECAL with lowering energy of the beam leads to better stochastic term in common; and this fact we will try to reflect at the new equation for the energy resolution. Let us make stochastic term to be dependent on the energy logarithmically. σ(e)/e = (A + A E loge)/ E B C/E (green curve). This function is much better fit to the data. V.L. Morgunov HCAL meeting, DESY, July.
Combined Calorimeter Instead of conclusion This job is on going. V.L. Morgunov HCAL meeting, DESY, July.