Seminar BELLE II Particle Identification Detector and readout system Andrej Seljak advisor: Prof. Samo Korpar October 2010
Outline Motivation BELLE experiment and future upgrade plans RICH proximity focusing detector upgrade and readout Contribution overview 2
Motivation Particle Identification (PID) is a crucial aspect of most High Energy Physics (HEP) experiments. In a typical experiment beams collide within the detectors (or a single beam collides with a fixed target) We wish to reconstruct as fully as possible the resulting events, in which many particles emerge from the interaction point. Example 1: B factory Particle identification reduces the fraction of wrong Kp combinations (combinatorial background) by ~5x 3
Motivation Without PID Example 2: HERA-B K+K- invariant mass. With PID K+K- invariant mass (GeV) The inclusive ϕ K+K- decay only becomes visible after particle identification is taken into account. Φ K+K- 4
KEKB electron - positron collider We need B mesons / B factories 8GeV e- KEK-B is electron-positron collider for B physics Located in Tsukuba in Japan Consist of a linear injector and two 3km storage rings 3.5GeV e+ ɤ(4S) B0 B 0 5
The Belle spectrometer Detector μ and KL 14/15 planes RPC+Fe Threshold Cherenkov counter aerogel n=1.015 - n=1.03 3.5 GeV e Silicon micro-strip detector + Calorimeter CsI 18 X0 8 GeV e- Superconductive magnet Central drift chamber with small cells Magnetic field 1.5 T Time of flight 6
Event example Particle decays in the interaction point. Particle momenta is measured in drift chambers by their curvatures in the magnetic field. Magnetic field is perpendicular to the projection plane. Particle Identification Devices identify charged particles 7
Kaon identification requirements Region 3-4 GeV/c not sufficiently covered with the present system 8
Planned upgrade For studying very rare decays of B and D mesons, we need bigger data sample. We have to upgrade: Accelerator KEKB Super-KEKB To increase luminosity Spectrometer Belle Belle II Increase triggering rate Increase reconstruction efficiency We plane to upgrade the accelerator and the spectrometer by 2014. With the upgraded detector we expect to collect 100 times bigger data sample. 9
Requirements for detector upgrade Critical issues at increased luminosity: - Higher background (x20) Radiation damage and fake counts due to noise - Higher event rate ( x40) Better and more efficient DAQ system - New requirements Identification of μ at low momenta b sμμ Possible solutions: - Vertex detector upgrade - Replace the internal part of the drift chamber with silicon detector. - Better particle identification device - Calorimter exchange: CsI (Tl) for pure Csl - Better electronics and computer system 10
Belle II Spectrometer overview Replace or upgrade all components (except magnet and structure) 11
Cherenkov radiation Treshold Phenomena appears when a charged particle travels faster then then the speed of light trough a medium with a refractive index n. v 1 = c n Cherenkov angle angle between particle direction and emitted photons cos C = 1 n Number of emitted photons per unit length per unit energy 2 d N = c sin 2 C = / hc 1 1/ 2 n 2 E dl de h Number of detected photons in the detector: N fot =L =370 /cm ev hc c kv E z E ele E T E sin 2 c E de h 2 N fot =L N 0 sin c E 12
Cherenkov radiation Proximity focusing n = 1.05 at p = 4GeV/c C (pion)~308mrad and C (pion)- C (kaon) ~23mrad we want large Nphot since track = and thin aerogel layer c N 13
Proximity focusing Rich detector sensor selection Requirements for ARICH detector photo sensors: Single photon detection High magnetic field (1.5T) Radiation environment Candidates: For the final detector the HAPD was chosen. HAPD, Hybrid avalanche photodiode MCP-PMT, Multi channel plate photo multiplier SiPM, Silicon Photomultiplier 14
Hybrid avalanche photodiode Principle of operation: Incident photon creates a photoelectron in the bi-alkali photocathode The electron accelerates in the electrostatic feld Bombardment gain ~ 10000 Avalanche gain in photo diode bias ~ 300V Gain in the order of ~ 10 Total gain about 104-105 Technical specifcations: Package 72x72mm2 QE ~ 25% at 400nm Sensitive area 67% 144 pads 15
Test of the proximity focusing RICH detector Proximity focusing RICH tested in 2GeV/c electron beamline The setup consisted of: Two layers of Aerogel Cherenkov radiators n1=1.054 and n2 = 1.065, Thickness d1 = d2 = 20mm Transmission lengths @ 400nm: λ1=47.8mm and λ2=55.2 mm Radiator to sensor distance 200mm Six (2x3) array of HAPDs 16
Readout electronics The 144ch from HAPD are read out with ASICs. Gain Shaper Comparator Asic specififactions: Production at VDEC Process: CMOS 0.35 um Noise Level: 1200e @ 80pF(HAPD) Std. Input Signal: 104-105 e Shaping time 0.25 ~ 1.0 us Variable gain 1.25 ~ 7.5 Individual offset adjustment Analog signal monitoring 17
Readout board design Front end electronic requirements: Very limited space available Adjustment of chip parameters due to irradiation. Design: Single board to read all 144 HAPD channels Digitally controlled offsets and thresholds Temperature control Possibility to monitor the analog signals. 18
New electronics block diagram 19
Readout electronics 20
Belle II Aerogel RICH Full scale view 1.7m from the interaction point in the forward direction Radius min:0.44, max:1.14m The radiator plane consists of hexagonal tiles The detector plane consists of 530 HAPDs modules 92% of the surface covered by HAPDs Expected Raw Data stream @ trigger rate 30kHz is 780Mb/s. 21
Overview of my contribution: I am involved in the following activities: Detector design based on the Monte Carlo simulation Design, production and tests of the prototype Study of Radiation hardness of components Detector construction and commissioning 22
Conclusions Belle spectrometer will be upgraded to collect 100 times bigger data sample. One of the key issues is the upgrade of the Particle identification system. In the foreword direction proximity focusing aerogel RICH will enable efficient kaon identification. The Hybrid Avalanche photodiode will be used as a photon detector. Special front end electronics should fit in very limited space. I am/will be involved in the design, construction and the commissioning of the Aerogel RICH. Thank you for your attention. 23
backup Efficiency and purity in particle identification particle type 1 type 2 Some discriminating value 24
backup Super KEK B upgrade Intercation point More RF power Damping ring Goal:increase the number of interactions by 50x dn/dt = Lσ σ crossection L luminosity KEKB world record L = 1.7x1034/cm2 /s Plan for upgrade 35 L = 8*10 /cm2 /s 25
backup Single photon Cherenkov resolution Optimal Aerogel Tickness 2 c 2 a 2 Ad Two main contributions to single photon resolution (n=1.05, d=1cm) 2 pad size a (~6mm) aerogel thickness d cos c a a= 8mrad 1 12 l d 2 l a d cos c sin c d 4 mrad d= d 1 cm 12 l d 2 Cherenkov angle resolution per track track = c N 26
backup Δ particle and anti-particle Blue: Time decay for Anti-B Red: Time decay for B Difference between measurements B mason decays differently from his antiparticle. 27
backup 28
Preliminary backup System noise analysis under investigation: Charge sensitive amplifier and shaper 4 gain settings 4 peaking constants (250,500,750,1000ns) Transfer function Response 29
Preliminary backup System noise analysis under investigation: 4 gain settings, 4 shaping constants Transfer function Response 30
backup Preliminary System noise analysis under investigation: CR-RC 2nd order filter heaving 40db/decade cut TS Characteristic shaping time (e.g. peaking time) Fi Fv "Form Factors" that are determined by the shape of the pulse. Ci Total capacitance at the input node (detector capacitance + input capacitance of preamplifier + stray capacitance +... ) Typical values of Fi CR-RC shaper Fi, Fv = 0.924 Fv = 0.924 CR-(RC)4 shaper Fi = 0.45 Fv = 1.02 CR-(RC)7 shaper Fi = 0.34 Fv = 1.27 31
backup Signal processing is done on FPGA FPGA has built in shift registers, one per channel. Shifting is done via storage ring downclock frequency. At trigger event, the last 4 values in each register are sent toward DAQ. 32