The Timing Counter of the MEG experiment: calibration and performance
P. W. Cattaneo, M. De Gerone, S. Dussoni, F. Gatti, M. Rossella, Y. Uchiyama, R. Valle
aa r X i v : . [ phy s i c s . i n s - d e t ] A p r The Timing Counter of the MEG experiment: calibration andperformance
P. W. Cattaneo a ∗ , M. De Gerone b , S. Dussoni b , F. Gatti b , M. Rossella a , Y. Uchiyama c , R. Valle ba INFN Pavia, Via Bassi 6, I-27100, Pavia, Italy. b INFN and Universit´a di Genova, Dipartimento di Fisica, Via Dodecaneso 33, I-16146, Genova, Italy c ICEPP, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-0033, Japan
The MEG detector is designed to test Lepton Flavor Violation in the µ + → e + γ decay down to a BranchingRatio of a few 10 − . The decay topology consists in the coincident emission of a monochromatic photon indirection opposite to a monochromatic positron. A precise measurement of the relative time t e + γ is crucial tosuppress the background. The Timing Counter (TC) is designed to precisely measure the time of arrival of the e + and to provide information to the trigger system. It consists of two sectors up and down stream the decaytarget, each consisting of two layers. The outer one made of scintillating bars and the inner one of scintillatingfibers. Their design criteria and performances are described.
1. The MEG experiment
MEG is an experiment designed to improve sig-nificantly the limits on the Branching Ratio (BR)of the µ + → e + γ decay [1]-[2]. This decay chan-nel is strongly suppressed in the Standard Modelbut is allowed to a measurable level in many al-ternative theories [3].The topology of this decay is very simple, consist-ing of monochromatic e + and γ with a commonspace time origin, with energy close to half the µ + mass, moving in opposite directions. Thesefeatures can be summarized as • E ( e + ) = E sige + ≈ m µ + / • E ( γ ) = E sigγ ≈ m µ + / • T e + γ = T e + − T γ = 0 • v e + = v γ • Θ e + γ = Θ e + − Θ γ = π where v is the origin space vertex, T the origintime, Θ the relative polar angle and E the energy.The main difficulty in this measurement is that, ∗ Corresponding author: [email protected] in order to collect the high statistic requiredto improve the existing BR in a few years, avery high rate of µ + must decay on the target: ≈ Hz is the design value. The Michel decay µ + → e + νν produces e + with an energy up to E sige + , therefore high resolution energy measure-ment is required to reduce the Michel e + contam-ination.The γ background originates from radiative decay(RD) µ + → e + γνν , e + e − annihilation in flightand e + bremsstrahlung. The kinematic limit is E sigγ + and therefore background suppression needshigh resolution measurement of E ( γ ).The γ and e + directions and the e + vertex need tobe measured with high precision to discriminatecombinatorial background from random associa-tion of e + and γ from different µ + decays. Precisemeasurements of T e + and T γ reduce the combina-torial background from different µ + decays.The photon detector is a Liquid Xenon Calorime-ter (LXe) exploiting scintillation light [8] thatprovides a measurement of the photon energy, po-sition and timing.The positron is observed by a spectrometer hav-ing at the core a set of 16 low mass Drift Cham-bers (DCH) embedded in a high magnetic field1( O (1 T )) with a gradient along the z axis to bendthe positron with a radius weakly dependent onthe emission polar angle. The DCH measures the e + momentum and position but is unable to de-liver trigger information and to provide precisetiming [7]. The trigger and timing informationon the e + are delivered by the Timing Counter,that is described in detail in the following.
2. The Timing Counter: the design
The TC is required to cover the solid angle op-posite to the LXe providing high efficiency for e + detection. This requirement is satisfied divid-ing the TC in two modules, called sectors, placedsymmetrically with respect to the decay target.Each sector must provide a precise measurementof the e + timing both at the trigger and anal-ysis levels ( O (100 ps )). Furthermore it providesa measurement of the crossing point both for thetrigger, requiring the e + and γ moving in oppositedirection, and for the analysis, DCH-TC matchand determination of the e + track path length.These requirements are met with a two layer de-tector: the outer one measures the transversecoordinate ( φ with respect to the beam direc-tion) and provides in addition time informationfor trigger and analysis. The inner one measuresthe longitudinal coordinate ( z along the beam di-rection) providing trigger and analysis informa-tion. The outer layer consists of 15 scintillator barslocated along the z-axis at fixed radius ( ≈ cm )in a barrel-like array with 10.5 ◦ gap (see Fig.1).This configuration has high acceptance for µ + → e + γ decay with momentum p e + = 52 . M eV /c and reduced acceptance for Michel events. Thenumber of bars is matched to the number ofDCH and to the trigger requirement for selectingcollinear e + − γ . The bars have a square sectionwith edge 4 . cm and length 80 cm and are readby fine-mesh PMTs adequate for use in high in-tensity magnetic field. The criteria leading to thechoice of these parameters are explained in [4].The signals from the PMTs are processed by aDouble Threshold Discriminator (DTD) with a low threshold to reduce Time Walk effect anda high threshold to remove background events.When the DTD is fired, it delivers a NIM signal.The PMT and the NIM signals are read by theDomino Ring Sampler a custom designed digitizeroperating in MEG at frequency up to 2GHz [5].The full waveforms are stored so that the time canbe extracted offline with optimized algorithms.Figure 1. The MEG timing counter The inner layer is made of 256 5 × mm scin-tillating fibers providing trigger and analysis in-formation on the z coordinate [6] (see Fig.1).Each fiber is read out by APDs (Avalanche PhotoDiode) that allow operations in high intensitytransverse magnetic field with a gain ≈ .The analog signal from the APD are filtered, 16of them are summed and made available to thetrigger. In parallel, the signals are discriminatedand the digital hitmap is made available for theanalysis.
3. The Timing Counter: calibration andperformances
The longitudinal layers started taking data in2006 and have been active during the followingyears. To achieve the design resolution, it turnedout necessary to calibrate carefully the detector.The calibration tools naturally available are thelarge flux of Michel e + from µ + decays andthe cosmic rays. There are two other calibra-tion sources requiring a dedicated set up. A π − beam impinging on a liquid H target pro-duces through charge exchange reaction Dalitzdecays π → γe + e − . Protons of kinetic energy T ≈ M eV from a custom Cockroft-Walton ac-celerator produces on a Boron target the nuclearreaction B ( p, γ ) C .Important calibration items are the Time Walk(TW) terms that account for the amplitude de-pendence of the timing measurement. They aremeasured using e + hitting two or three bars. Thesame events are used to measure the TC time res-olution.Other crucial calibration is the measurement oftiming offset of the different PMTs to allow com-bining results from different part of the TC. Thiscalibration is obtained with the Dalitz or Boronevents, that have two particles emitted at thesame time, using the LXe as reference.The single bar time resolutions with TW correc-tion and offset subtracted are shown in Fig.2.The transverse layer worked only partially dueto problems with the digital readout and excessnoise, that prevented an efficient use of the de-tector. With the data available, it was possibleto verify that the tracks reconstructed from DCHmatch the hits on both transverse and longitudi-nal layers reducing the combinatorial backgroundand delivering improved information on the e + track length.
4. Conclusions
The TC of MEG has performed as high res-olution timing detector according to the expec-tations playing a crucial role in the MEG trig-ger system. The transverse layer met unexpectedproblems with part of the digital readout, but theworking part delivered satisfactory results. Forthe 2010 run, both layers are expected to be fullyoperational.
REFERENCES
1. T. Mori et al., The MEG experiment: searchfor the µ + → e + γ decay at PSI, May 1999 bar_resoEntries 30Mean 15.11RMS 5.801 Bar number0 5 10 15 20 25 ( T ) ( n s ) σ bar_resoEntries 30Mean 15.11RMS 5.801 Figure 2. The single bar time resolution usingtwo-bar coincidence.(http://meg.web.psi.ch/docs/index.html)2. A. Baldini et al., The MEG ex-periment: search for the µ + → e + γγ