Particle tracking in the CALET experiment Paolo Maestro University of Siena & INFN Nicola Mori INFN Sezione di Firenze On behalf of the CALET collaboration PoS (2017) 208 1 35 th ICRC, Busan 2017 Paolo Maestro
Outline Ø Introduction on CALET experiment Ø Tracking procedure: description and implementation Ø Tracking validation and performance with Monte Carlo data Ø Application of the tracking method to flight data Ø Track-based alignment of IMC fibers Ø Data vs. MC comparison 2 35 th ICRC, Busan 2017 Paolo Maestro
CALET payload CGBM (CALET Launched on Aug. 19 th 2015 on FRGF(Flight Releasable Gamma Ray Burst Grapple Fixture) the Japanese H2-B rocket Monitor) ASC (Advanced Emplaced on JEM-EF port#9 Stellar Compass) On Aug. 25 th 2015 GPSR (GPS CAL/CHD Receiver) MDC (Mission Data Controller) CAL/IMC CAL/TASC Port # 9 ・ Mass: 612.8 kg ・ JEM Standard Payload Size 1850 mm (L) × 800 mm (W) × 1000 mm (H) ・ Power Consumption: 507 W （ max ） ・ Telemetry: Medium 600 kbps (6.5GB/day) / Low 50 kbps 3 35 th ICRC, Busan 2017 Paolo Maestro
CALET instrument CGBM HXM x2 CHD-FEC CHD-FEC 7keV-1MeV CHD IMC-FEC IMC-FEC IMC LaBr 3 (Ce) SGM x1 TASC-FEC TASC-FEC 0.1-20MeV TASC BGO CALORIMETER Detector Measure Geometry Readout Plastic Scintillator CHD Charge (Charge Detector) (Z=1-40) 14 paddles × 2 layers (X,Y) PMT+CSA Paddle size: 3.2 × 1 × 45 cm 3 IMC 448 Scifi × 16 layers (X,Y) Tracking (Imaging 64 MAPMT+ ASIC Particle ID 7 W layers (3 X 0 ) Calorimeter) Scifi size: 1 × 1 × 448 mm 3 TASC 16 PWO logs × 12 layers (X,Y) APD/PD + CSA (Total Absorption Energy log size: 1.9 × 2 × 32 cm 3 PMT+CSA Calorimeter) e/p separation (for Trigger) Total thickness: 27 X 0 , ~1.2 λ 4 35 th ICRC, Busan 2017 Paolo Maestro
CALET science goals Science Objectives Observation Targets Energy range Nearby CR sources Electron spectrum 1 GeV – 10 TeV Dark Matter Signatures in e/ γ spectra 10 GeV – 10 TeV p-Fe individual spectra 10 GeV – 10 3 TeV CR Origin and Acceleration Ultra Heavy Ions (26<Z ≤ 40) few GeV/amu Galactic CR B/C sub-Fe/Fe ratios Up to some TeV/n Propagation Solar Physics Electron flux < 10 GeV Gamma-ray Gamma and X-rays 3 keV – 30 MeV Transients 5 35 th ICRC, Busan 2017 Paolo Maestro
Tracking in CALET Ø Tracking needed for: • pointing • definition of geometrical acceptance • computing variables for e/p discrimination • identification of hit paddles in CHD and hit scifi’s in IMC à Particle ID Ø Tracking exploits IMC fine granularity and imaging capability Main challenge: find the primary CR particle track in a large amount of shower particle tracks backscattered from TASC à High hit multiplicity à Multiple track candidates Ø Combinatorial Kalman Filter (CKF) algorithm to provide robust track finding and fitting • Algorithm fed with the Z positions of layers, the X/Y coordinate of the center-of- gravity (COG) of each Scifi’s cluster, and its associated position error σ cog . • The algorithm runs separately for the X and Y views to find the projections of the full 3D track, which is then reconstructed. 6 35 th ICRC, Busan 2017 Paolo Maestro
Combinatorial Kalman Filter • A candidate track is created for each XZ view CR Track possible combination of clusters in the first two layers. • Each track is fitted to a straight line and Layer 0 propagated to the next layer k . • The predicted position at layer k is used to associate new clusters to the track. Layer 1 Candidate track Secondary Track propagators • A new track candidate is created for every cluster on layer k , whose COG lies Layer 2 sufficiently near to the predicted state, i.e. Real hits Spurious hits within σ cog • To cope with possible inefficiencies, Layer 3 Missing hit missing points (“holes”) are allowed in track. Layer 4 • Each “branched” track is fitted. Bad tracks ( χ 2 >10 or no. holes>2) are discarded. Good candidates are propagated to layer k+1 , Layer 5 and procedure is iterated, up to the last Bad track Good track layer. Good track 7 35 th ICRC, Busan 2017 Paolo Maestro
Identification of primary particle track Ø For high-energy shower events, in order to reduce the proliferation of track candidates and processing time, only IMC clusters within a ROI (region-of-interest) are fed to KF algorithm. Ø ROI is defined around the back-projection of the TASC shower axis, which is fitted to the COG of the TASC log clusters in each TASC layer. Ø Among all the candidate tracks found by the KF algorithm in each view, the track with: - the largest energy deposited in the associated clusters; - closer to the core of the shower reconstructed with TASC; is chosen as the primary particle track. 8 35 th ICRC, Busan 2017 Paolo Maestro
Tracking example Carbon event with estimated E>10 TeV ROI Shower axis Primary particle track 9 35 th ICRC, Busan 2017 Paolo Maestro
Tracking example Carbon event with estimated E>10 TeV ROI Shower axis Primary particle track Candidate tracks 10 35 th ICRC, Busan 2017 Paolo Maestro
Tracking validation with MC • Samples of electrons, protons, He, C, Fe nuclei were simulated with FLUKA - Isotropic generation in direction - E -1 energy spectrum. Particle energy range: 10 GeV - 100 TeV for nuclei 5 GeV - 20 TeV for electrons • CALET High-Energy Trigger (HET) modelled in simulation. - Coincidence of signals in last four IMC layers and first TASC layer, with thresholds chosen to ensure >95% efficiency for electrons > 10 GeV • Particle selection used to evaluate tracking performance - HET events - Reconstructed track length in TASC > 27 X 0 - Entrance point in CAL above 5 th IMC layer. 11 35 th ICRC, Busan 2017 Paolo Maestro
Angular resolution (electrons) • KF rec. MC truth PSF x = 0.11° Point-spread function (PSF) is half-width of the interval centered on mean value of the residual distribution Δθ = θ rec – θ MC containing 68.3% of events. PSF y = 0.11° 12 35 th ICRC, Busan 2017 Paolo Maestro
Angular resolution (protons) • KF rec. MC truth PSF x = 0.55° Point-spread function (PSF) is half-width of the interval centered on mean value of the residual distribution Δθ = θ rec – θ MC containing 68.3% of events. PSF y = 0.55° 13 35 th ICRC, Busan 2017 Paolo Maestro
Angular resolution (C nuclei) • KF rec. MC truth PSF x = 0.22° Point-spread function (PSF) is half-width of the interval centered on mean value of the residual distribution Δθ = θ rec – θ MC containing 68.3% of events. PSF y = 0.23° 14 35 th ICRC, Busan 2017 Paolo Maestro
CR arrival direction reconstruction Θ 3D distribution (electrons) • θ x and θ y are combined to determine the parameters ( θ , φ ) of the arrival direction of the particle in the space. • The 3D angular resolution ∆Θ 3D is the PSF of the distribution of the scalar product between Δ Θ 3D the reconstructed and the true MC arrival direction. Electrons Fe nuclei 15 35 th ICRC, Busan 2017 Paolo Maestro
Impact point resolution (electrons) • KF rec. MC truth σ x = 420 µ m Image of rec. CHD impact point σ y = 420 µ m 16 35 th ICRC, Busan 2017 Paolo Maestro
Tracking efficiency Acc. A Acc. B Acc. C Tracking efficiency (Electrons) Tracking efficiency (Protons) Three different acceptance configurations: A : track crossing CHD top and TASC bottom B : track crossing 5 th IMC layer and TASC bottom Tracking efficiency (C nuclei) C : track crossing 5 th IMC layer and with length in TASC > 27 X 0 17 35 th ICRC, Busan 2017 Paolo Maestro
Performance of tracking algorithm Particle PSF x/y Tracking efficiency σ x/y (deg) (mm) @ E > 100 GeV (nuclei) > 10 GeV (electrons) electron 0.1 0.42 95% proton 0.55 1.5 90% He nuclei 0.45 1.2 90% C nuclei 0.22 0.58 95% Fe nuclei 0.09 0.26 95% 18 35 th ICRC, Busan 2017 Paolo Maestro
Recommend
More recommend
