Belle/Belle II CsI(T l ) electromagnetic calorimeter Kenkichi Miyabayashi (Nara Women s University, Japan) Calorimeter workshop in ELPH 2015 Mar. 10 th 1
Outline Calorimeter for e + e - collider at region Solution taken at Belle SuperKEKB and Belle II Challenge : beam background immunity Belle II electromagnetic calorimeter Day-1 : CsI(Tl) with waveform sampling readout Forward Endcap upgrade with Pure CsI Summary 2
Calorimeter for e + e - at region Wide dynamic range: 20MeV~8GeV 1/3 of B decays have π 0, most of γ~100mev. Radiative B decays (B K*γ, etc.) γ up to 4GeV Bhabha, e + e - γγ calibration, up to 8GeV High energy resolution σ E /E ~ 4% at 100MeV σ γγ ~ 5MeV/c 2 for π 0 High position resolution σ x : 5~10mm at the incident point 3
Belle CsI(T l ) calorimeter In total, 8736 CsI(Tl) crystals (6624 in Barrel, 1152 in Fwd. Endcap and 960 in Bwd. Endcap) Covering 12<θ<155 in Lab. frame. Inner radius = 1250mm. e- e+ 4
CsI(T l ) with PIN-PD Casing Preamps Available solution in the magnetic field in 80 s-90 s. 30cm long CsI(T l ) =16.1X 0 Belle 2 PIN-PDs Gore-Tex and Aluminum-coated myler wrapping. 8736 CsI counters 5 are used.
Mechanical structure Taken from H. Ikeda s Doctor thesis 6
Snapshot during construction Counters tested by cosmic Putting counters into mechanical structure
Barrel CsI installation 8
Accommodated counters Laser welding : broken. 90 fold wall and screws. 9
Read out Electronics Scintilation light Photodiode CsI crystal Preamp. Inside Belle detector Two preamp. signal are summed at 1st stage of Shaper-QT board. Shaping part (200ns) (1µs) Cal. Trigger Electronics MQT 300A (QtoT) TDC(LeCroy1877S) DAQ Near the Belle detector To achieve 18bit-eq. dynamic range, QtoT conversion and TDC readout with auto-range selection is used. PIN-PD capacitance = 80pF, noise = 1000 electrons, 5000p.e./MeV Equivalent Noise Energy = 0.2MeV. Additional noise coming from pileup. Electronics Hut 10
The energy deposit in i-th counter(e i ) is obtained by; Calibration where TDC i =digitized signal, PED i =pedestal of i-th ch and e i =electronics gain const., obtained by daily elec. calib. run. by minimizing Ei = TDC i - PEDi ei Ci Cosmic rays give C i without beam (initial input). C i absolute calibration; Bhabha, e + e - γγ. # 2 = (Esumj - Eexpj) 2! j " 2 which results in the matrix inversion method.
Cosmic ray Calibration The energy deposit is predictable by cosmic ray MC. Can be done without beam. Gives good initial inputs for Bhabha calibration. Innermost crystals in Endcaps can be calib. only by this method. Suitable to mon. L.O. as a func. of rad. dose. The crystal surrounded by neighboring hits can be calibrated.
Bhabha, γγ Calibration Matrix inversion fairly converges! (except for inner -most crystals in Endcaps) After Bhabha calib. γγ is used to carry out fine tuning; less systematic by material in front. C i (after)/c i (before) g 1.2 1.15 1.1 1.05 1 0.95 0.9 0.85 0.8 2500 2000 1500 1000 500 Crystal ID 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 Cell ID 458.1 / 52 Constant 2326. 35.38 Mean 1.003 0.5460E-04 Sigma 0.5478E-02 0.4997E-04 0 0.8 0.85 0.9 0.95 1 1.05 1.1 1.15 1.2 g
Non-linearity correction by π 0 Bhabha, γγ calib. at highest E. point. Interpolation in low energy region needs verification. M γγ = { E 1 E 2 (1-cosα) } 1/2 π 0 mass peak gives information in low ene. region(< 1GeV), because EM shower is well predicable by sim. 0.99 0.98 0.97 0.96 Non-linearity correction (applied by div. for Exp. Data) 2.0 8.900 / 6 P1 0.8493 P2 0.8631E-01 P3-0.9390E-02 P4-0.1056E-02!"# log10(e in MeV)
Radiation dose and L.O. change 4Gy 1Gy 4Gy 15
Shower reconstruction algorithm Crystal hits below threshold (0.5MeV~2MeV) are dropped by TDC readout. Seed crystal: local max. >10MeV. Recorded hits inside the 5by5 matrix surrounding the seed crystal are clustered. Energy: summing up each hits. Position: energy-weighted center of gravity. Shower leakage and Position systematic are corrected based on simulation. Seed crystal Hits exceeding threshold.
Performance x 10 2 2500 2000 1500 1000 500 25000 20000 15000 10000 5000 All E/E=1.7! 0 0.6 0.8 1 1.2 E/E Forward endcap E/E=1.6! e + e - γγ sample 0 0.6 0.8 1 1.2 E/E x 10 2 5000 4000 3000 2000 1000 6000 5000 4000 3000 2000 1000 Barrel E/E=1.7! 0 0.6 0.8 1 1.2 E/E Backward endcap E/E=2.2! 0 0.6 0.8 1 1.2 E/E x 10 3 1200 1000 800 600 400 200 0 0 in hadronic events m = 4.75 +- 0.08 (MeV/c 2 ) σ γγ =5 2 0.08 0.1 0.12 0.14 0.16 0.18 M (GeV/c 2 ) 17
SuperKEKB e + New beam pipe & bellows Belle II New IR New superconducting / permanent final focusing quads near the IP Replace short dipoles with longer ones (LER) e - Redesign the lattices of HER & LER to squeeze the emittance Low emittance positrons to inject Damping ring Add / modify RF systems for higher beam current Positron source TiN- coated beam pipe with antechambers Low emittance gun Low emittance electrons to inject New positron target / capture section Currents 2, β * y 1/20 w.r.t. KEKB 18 to aim 40 luminosity
Nano-beam collision KEKB SuperKEKB To increase luminosity, small β function is used. To handle hourglass effect, β>size of collision spot, large crossing angle, one bunch behaves as super bunch. 19
Belle II Detector KL and muon detector: Resistive Plate Counter (barrel outer layers) Scintillator + WLSF + MPPC (end-caps, inner 2 barrel layers) EM Calorimeter: CsI(Tl), waveform sampling (baseline) (opt.) Pure CsI for end-caps electron (7GeV) Beryllium beam pipe Particle Identification Time-of-Propagation counter (barrel) Prox. focusing Aerogel RICH (fwd) 2cm diameter Vertex Detector 2 layers DEPFET + 4 layers DSSD Central Drift Chamber He(50%):C2H6(50%), Small cells, long lever arm, fast electronics positron (4GeV) Upgrade to give optimum performance under 20 20 beam background!
Challenge : Beam BG immunity 1GeV,1GHz = 0.16W HER (e- ) LER (e+) LER (4GeV e+) HER (7GeV e-) Rad. Bhabha 0.63 W (eff. 0.98GHz) 0.88W (eff. 0.78GHz) Touschek 0.07 W (0.11GHz) 0.02 W (0.02 GHz) Coulomb 0.07 W (0.10GHz) 0.001W (0.001GHz) Mostly from Rad.Bhabha 21
CsI(T l ) bright, but slow. Scintillation decay time is ~1µs Signal ( V(t) or I(t) ) Signal by physics particle incident t Pileup caused by beam background incident behaves as noise increase Beam background immunity is really desired challenge in SuperKEKB/Belle II environment. 22
Waveform sampling readout Shaper output signal Gate width=100ns t Signal charge Timing (leading/trailing edges) with range information by QtoT converter (MQT300A) Digitized by TDC 1.76MHz, 18bits digitizer, waveform fit to get energy and timing (i.e. Digital Signal Processing) Reduction factors; 7 BG showers 1.5~2 pileup noise t 23
Readout for Belle II 24
Early prototype tested at Belle 1/8 of backward endcap was connected to waveform sample readout prototype, 1fb -1 was taken on (4S) resonance. Timing σ=9ns@100mev Pileup reduction 25
Status of cosmic test Trigger module 26
Typical cosmic event Trigger module is here 10h=2.5M events have been taken continuously. φ θ 27
Pure CsI at Endcap? Photo Pentode (Fine mesh dynode) Pure CsI crystals Because of short scintillation decay time, ~30ns, Pure CsI crystal is almost pileup free. Photo Pentode readout is regarded as a baseline, noise~0.2mev. LAAPD is also being tested. Part of forward endcap impact to physics? CsI(Tl) and PIN-PD are expected to be alive deterioration? 28
Summary Belle CsI(Tl) calorimeter successfully operated for 10 years to bring various physics output. SuperKEKB is aiming 40 luminosity w.r.t. KEKB. Most beam background from Rad. Bhabha, i.e. proportional to luminosity(!) Beam background immunity is a challenge. Use all the existent CsI(T l ) with waveform sampling readout electronics. Cosmic test is on going. Full simulation studies for CsI(Tl) deterioration and PureCsI effectiveness are to be done in FY2015. 29
Electronics arrangement Waveform fit and multiplexing information by FPGA 30