ECAL performance evolution plots Contact: 1

Plots from CMS-DP calDPGResults#CMS_DP_2012_027 2

3 Top plot: Relative response to laser light (440 nm) measured by the ECAL laser monitoring system, averaged over all crystals in bins of pseudorapidity, for the 2011 and 2012 data taking periods. The response change observed in the ECAL channels is of the order of a few percent in the barrel, while it reaches up to 25% in the most forward end cap regions used for electron and photon reconstruction. The response change is up to 60% in channels closest to the beam pipe. This is an update of the plot appearing in CMS-DP-2012/015, and includes measurements taken during These measurements are used to correct the physics data. Bottom plot: Instantaneous luminosity versus time

Correlation between relative change of response from ECAL monitoring data, observed between 6-SEP-11 (end of technical stop) and 4-OCT-11, and the absorption coefficient μ std induced in crystals by a standard γ irradiation [1] performed during Quality Control: one entry per crystal, for the fraction of crystals in the ring 2.4 < η < 2.6, where this information is available. The presence of a correlation proves that a fraction of the signal loss is due to a change in crystal transmission from ionizing radiation damage. The positive intercept of the correlation hints at the presence of a fraction of signal loss, uncorrelated with μ std, which could be caused by other sources, like losses in VPT response or cumulative losses from hadron-specific changes in crystal transmission. [1] The CMS ECAL group, P. Adzic et al., 2010 JINST 5 P

Crystal properties, induced absorption measurements and simulations results 5

6 Correlation between proton fluence and induced absorption from 24 GeV/c proton irradiations of CMS ECAL end cap crystals performed in 2010 and 2011 (colored symbols), superposed to a correlation [1, 2] obtained from a 20 GeV/c proton irradiation of CMS ECAL barrel crystals. Possible effects due to the irradiation rates, which are higher with respect to the exposure during CMS running, and the uncertainties affecting the results are addressed in [1]. [1] M. Huhtinen et al., Nucl. Instr. Meth. A 545 (2005) [2] G. Dissertori et al., Nucl. Instr. Meth. A 622 (2010) 41-48

7 Left plot: layout of the EE crystal matrix studied in the test beam to ascertain the effect of hadron damage. Indicated are the induced absorption values for each crystal in [m -1 ] from irradiations with 24 GeV/c protons with fluences up to 6 x p/cm 2. The lower values are for crystals whose damage was partially annealed [1]. Right plot: Preliminary electron energy resolution as a function of beam energy for an array of 3x3 non- irradiated crystals (green symbols), and for two arrays of 3x3 crystals in the matrix (same color for data points and for the corresponding array frame), for a central impact within 10 mm x 10 mm. The data are uncorrected for irradiation fluence uncertainties, while the pedestal noise has been subtracted and the channel intercalibration for 3x3 clusters has been performed at 50 GeV. The total energy has been reconstructed using a 3x3 cluster, by fitting the corresponding spectrum with the Crystal Ball function. The ordinate in the plot corresponds to the sigma of the crystal ball function over the peak position. The analysis method and the results are still preliminary. The energy resolution is clearly degraded with respect to a non-irradiated matrix. [1] A. Singovski, paper N29-4, IEEE NSS 2011 and E. Auffray et al., CMS CR-2011/257

8 For the simulation, the light collection efficiency (EFF) versus distance Z from the front face of EE crystals is calculated with LITRANI [3]. Scintillation photons are simulated as emitted isotropically, with timing and wavelength properties as in the ECAL TDR. Undamaged crystals are attributed a typical absorption length versus wavelength as measured for EE crystals, which is then applied uniformly inside a crystal. The procedure of FNUF determination is identical to the one used for crystal Quality Control during ECAL construction, and it is applied to EFF vs Z for a given induced absorption μ IND. The difference FNUF(μ IND =0) - FNUF (μ IND ) versus μ IND is plotted as a red line. Generally speaking, this value slightly depends on the initial transparency of the unirradiated crystal. [1] The CMS ECAL group, P. Adzic et al., 2010 JINST 5 P03010 [2] E. Auffray et al., Nucl. Instr. Meth. A 486 (2002) [3] X.F. Gentit, Nucl. Instr. Meth. A 486 (2002) The data points represent the correlation between the absorption coefficient μ std induced in crystals by a standard γ irradiation [1] performed during Quality Control, and the change in FNUF caused by the same irradiation, where the FNUF is the slope of a linear fit applied to the Light Yield measurements as a function of the longitudinal position between 13 X 0 and 4 X 0. It is expressed in relative change per radiation length [2].

9 The data points represent the correlation between induced absorption from 20 GeV/c proton irradiations of CMS ECAL barrel crystals and the induced loss of Light Output for cosmic muons traversing the crystal laterally, as defined by the trigger counter geometry in [1]. A LITRANI [2] simulation is performed to obtain the light collection efficiency (EFF) versus distance Z from the crystal front face for a given value of μ IND, as in slide 17. The simulation of the average distribution of cosmic muons along the crystal, N(Z), is performed using the dimensions of the experimental setup [1]. The distribution is not uniform, it has a maximum in the center of the crystal, and sharply decreases towards both ends of the crystal. The overall Light Output versus μ IND is then obtained as ∫ EFF(Z, μ IND )xN(Z)dZ [1] P. Lecomte et al., Nucl. Instr. Meth. A 564 (2006) [2] X.F. Gentit, Nucl. Instr. Meth. A 486 (2002) 35-39

10 Each LO loss value has been obtained by comparing a non-irradiated crystal to an irradiated one, using the same photomultiplier, and applying to it the same high-voltage in both cases. Since each crystal had a different Light Output before irradiation, the LO used in the comparison has been corrected to take this difference into account. Concerning the simulation, the same applies as for the plot of slide 19, expect that, in this case, the average distribution of energy depositions along the crystal is based on a Geant4 simulation of 50 GeV electrons hitting the front face of the crystal. The data points represent the correlation between induced absorption from 24 GeV/c proton irradiations of CMS ECAL end cap crystals and the induced loss of Light Output (LO) for 50 GeV/c electrons entering the crystal longitudinally.

11 The amplitude for each electromagnetic shower as a function of μ IND is calculated as A(μ IND ) = ∫ EFF(Z, μ IND )xE(Z)dZ The distribution of relative degradation A(μ IND )/A(μ IND =0) is obtained for each given μ IND value. The effective sigma is defined as half the shortest interval that covers 68% of the distribution, and the mean is the center of this interval. The black symbols are the results for σ effective /mean at selected values of μ IND, and the red line is an analytical fit through the results. [1] X.F. Gentit, Nucl. Instr. Meth. A 486 (2002) MC simulation Contribution to EE energy resolution (50 GeV e-) CMS ECAL The plot presents a simulation of the contribution to the EE energy resolution due to the longitudinal non-uniformity in light collection caused by a loss of crystal transparency, expressed through the induced absorption coefficient μ IND. The light collection efficiency EFF(Z, μ IND ) is obtained with LITRANI [1], as described in previous plots. The energy deposition along an EE crystal, E(Z), is simulated with Geant4 for 50 GeV electrons.

Simulation 50 GeV e- CMS ECAL 12 The accumulated charge of photo-electrons at the VPT photo-cathode and the resulting VPT aging VPT(Q) is based on the average behavior of ~10 VPTs tested by CMS ECAL. The degradation of amplitude due to the loss of transparency is calculated as the mean of the distribution described in an earlier slide, where μ IND is the sum of the hadronic and electromagnetically induced absorption, as estimated above. The total degradation is then VPT(Q)x The total degradation of amplitude versus η is plotted for various pairs of instantaneous and integrated luminosity. [1] P.C. Bhat, A.P. Singh, and N.V. Mokhov, CMS NOTE-2009/019 [2] M. Huhtinen et al., Nucl. Instr. Meth. A 545 (2005) [3] R.-Y. Zhu, IEEE Trans. Nucl. Sci. 44 (1997) The plot represents the relative loss of ECAL EE response for 50 GeV electrons as a function of rapidity and for various values of integrated luminosity, as obtained from a simulation. For each set of values of instantaneous luminosity, integrated luminosity and pseudorapidity η, the following parameters are estimated: The fluence of charged hadrons in CMS based on MARS simulations[1], and μ IND for hadron damage [2]. The ionizing dose, based on MARS simulations [1], and μ IND for ionizing damage, assuming a simple model for equilibrium between darkening and recovery [3] and a maximum admitted value of μ IND =2 m -1 at construction.