Antony Sarrat CEA Saclay Dapnia Simulation Of a

Antony Sarrat CEA Saclay, Dapnia Simulation Of a TPC For T 2 K Near Detector Using Geant 4 A. Sarrat ILC TPC meeting, DESY, 15/02/06

Introduction • T 2 K near detector will have 3 TPCs for PID and momentum measurement. • A TPC equipped with Micromegas detector was tested at CERN in Nov. 2005 – > Monte Carlo simulation of these tests • Useful for resolution study (vs pad size, magnetic field, gas properties, …) A. Sarrat ILC TPC meeting, DESY, 15/02/06

T 2 K Near Detector Magnet + Side-MRD EM calorimeter Magnet + Side-MRD A. Sarrat ILC TPC meeting, DESY, 15/02/06 FGD TPC EM calorimeter n beam TPC FGD+H 2 O P 0 D Pb-P 0 D EM calorimeter

CERN Setup (HARP TPC) ~1. 5 m drift length ~ 26 cm ~ 154 cm ~ 52 cm Micromegas End-plate ~ 1300 cm 2 instrumented HARP TPC A. Sarrat ILC TPC meeting, DESY, 15/02/06

The Monte Carlo Simulation A. Sarrat ILC TPC meeting, DESY, 15/02/06

Flow Chart • Geant 4 used for the tracking. • Everything else done manually (through Geant user’s routine) A. Sarrat ILC TPC meeting, DESY, 15/02/06

Detector Simulation • Magnet (Fe) B = 0. 0, 0. 2 and 0. 4 T • TPC’s body (Al) • Drift volume (gas) Ar + CF 4 + Iso. C 4 H 10 (3%) (2%) or e. g. Ar + CO 2 A. Sarrat ILC TPC meeting, DESY, 15/02/06

Primary Events • Cosmic spectrum Vertex generated on the top trigger plane - Flat in (x, z) - Fixed y = +100 cm Vertex X vs Z Distribution X Z Primary Spectrum (Ge. V) or source (5. 9 ke. V g) – z = +154 cm – x, y = 0 • 55 Fe A. Sarrat ILC TPC meeting, DESY, 15/02/06 Pm

Track Generation • Use GEANT 4 simulation package to do the tracking inside the TPC gas volume. • Physics processes: – > standard EM processes – > transportation – > delta-rays production Save step information (energy loss and pre-step time-space position) during the tracking. A. Sarrat ILC TPC meeting, DESY, 15/02/06

Ionization • Eloss in a step is used to generate the ionization electrons. • WI = 26 e. V (Sauli) • Step limited to 1 mm (≤ 1 mm) 8 é / step <=> 80 cm-1 A. Sarrat ILC TPC meeting, DESY, 15/02/06 m = 16. 65 Peak ~ 8é / step

Drift And Diffusion From Magboltz: • Drift speed B = 0. 2 T s = 0. 029 Vdrift = 6. 5 cm. ms-1 • Diffusion coefficient s. L = sz = 290 mm. cm-1/2 s. T = 310 mm. cm-1/2 (0. 0 T) 240 mm. cm-1/2 (0. 2 T) 160 mm. cm-1/2 (0. 4 T) Use 2 D Gaussian for sx & sy A. Sarrat ILC TPC meeting, DESY, 15/02/06

Track After Drift Vertex Y X A. Sarrat Z ILC TPC meeting, DESY, 15/02/06

Digitization Transverse (x, y) view • Use the same layout than during data taking at CERN 8*8 mm 2 pads with 0. 1 mm pitch and 4 mm staggering. • 2 detectors, 1024 pads each (768 active pads per detector during data taking because of electronic/DAQ troubles) A. Sarrat ILC TPC meeting, DESY, 15/02/06

Electronic Signal Charge (ADC) • 4 th order semi-gaussian filter: (X e-X)4 with X = (T - T 0) / t and t = 190 ns • Gain fluctuations with an exponential distribution. A. Sarrat ILC TPC meeting, DESY, 15/02/06 Time (tick=100 ns) G = 6000 Gain

Digitized Event Q < 160 Q > 300 Q > 800 (ADC) A. Sarrat ILC TPC meeting, DESY, 15/02/06

Noise Simulation Electronic noise: Gaussian with: m = 0, s ~ 2. 0 ADC unit (same as in data) With 2048 pixels and 400 time bins per pixel: ~ 800 k noise generated / event Save only the noise above 4 s = 8. 0 ADC (in bins without signal) A. Sarrat ILC TPC meeting, DESY, 15/02/06

MC Event With Noise Hits cut @ 5 s threshold Q < 160 Q > 300 Q > 800 (ADC) cut @ 4 s A. Sarrat ILC TPC meeting, DESY, 15/02/06

Output • Text file – Primary info. (X, Y, Z) and (Px, Py, Pz) – Number of pads with hit – For each pad with at least one hit: • Number of hits • Time and charge of each hit • Root file with histograms for debugging A. Sarrat ILC TPC meeting, DESY, 15/02/06

Data / MC Comparison A. Sarrat ILC TPC meeting, DESY, 15/02/06

Gain Tuning Using Run with 55 Fe radioactive source 55 Fe Simulation of 5. 89 ke. V photons Larger s in data: channel to channel fluctuations ? A. Sarrat Run ILC TPC meeting, DESY, 15/02/06

Number of Pad Fraction Data B = 0. 2 T 75% 2 pads 75% 50% 25% 20 MC B = 0. 2 T 50% 1 pad 25% 50 100 (cm) 20 50 100 (cm) Fraction of number of pad per cluster vs drift length A. Sarrat ILC TPC meeting, DESY, 15/02/06

Resolution vs Z at B=0. 2 T A. Sarrat ILC TPC meeting, DESY, 15/02/06

Resolution vs Z at B=0 A. Sarrat ILC TPC meeting, DESY, 15/02/06

Resolution vs Z at B=0. 4 T A. Sarrat ILC TPC meeting, DESY, 15/02/06

Conclusion / To Do • A Monte Carlo simulation of a TPC was written for T 2 K near detector, using Geant 4 package. • The method and C++ code are simple. • Work needs to be done to understand some distributions that do not agree with data • Comparison with data gives good agreement for resolution versus drift length A. Sarrat ILC TPC meeting, DESY, 15/02/06

Transverse Diffusion B = 0. 2 T s = 0. 024 A. Sarrat B = 0. 2 T s = 0. 024 ILC TPC meeting, DESY, 15/02/06

Transverse Diffusion B = 0. 4 T s = 0. 016 Distribution of sx/y / √(drift length) A. Sarrat ILC TPC meeting, DESY, 15/02/06

Comparison Data/MC A. Sarrat ILC TPC meeting, DESY, 15/02/06

Estimated Track Width Data B = 0. 2 T s ~ 2. 8 mm MC B = 0. 2 T s ~ 2. 5 mm z = 1 m Estimated track width vs time A. Sarrat ILC TPC meeting, DESY, 15/02/06