PhD thesis: Simulation & Reconstruction for the PANDA Barrel DIRC Official name: Open charm analysis tools Supervisor: Prof. Klaus Peters Maria Patsyuk

PANDA GSI MeetingMaria Patsyuk2 Overview 1. PID for PANDA (PID requirements for DIRC) 2. Challenge of the PANDA Barrel DIRC 3. PANDA DIRC design — baseline and options 4. Simulation and reconstruction 5. Summary

PANDA GSI MeetingMaria Patsyuk3 PANDA Important detector requirement: good hadron PID – π/K separation Benchmark channels:

PANDA GSI MeetingMaria Patsyuk4 Hadron PANDA 1. Based on BABAR experience – DIRC (Detection of Internally Reflected Cherenkov light) detector, developed and successfully used for the first time at SLAC, 8+ years of stable and reliable operation 2. Advantages of DIRC detectors: - thin in radial direction and in radiation length = less material in front of the calorimeter and the smaller radius of the calorimeter - fast and tolerant to background - robust and stable detector operation 3. PANDA: two DIRC detectors - Barrel DIRC – similar to BABAR DIRC with several improvements - endcap Disc DIRC - Institutions involved: Dubna, Edinburgh, Glasgow, Erlangen, Gießen, GSI, Mainz, Vienna

PANDA GSI MeetingMaria Patsyuk5 Basics of DIRC detectors ● Charged particle with propagating in a medium with refractive index n emits Cherenkov photons on a cone with opening angle ● ● For media with n > √2 there are always some totally internal reflected photons ● High quality rectangular radiator shape is required to conserve the magnitude of the Cherenkov angle during the total internal reflections of propagating photons ● Photons exit radiator bars through focusing elements into expansion volume and are imaged on photon detector array Mirror DIRC delivers information about the position of the detected photon (x,y) and the time (t) – is intrinsically a 3-D device

PANDA GSI MeetingMaria Patsyuk6 Basics of the PANDA DIRC ● Radiator bars are made of Synthetic Fused Silica with n ≈ 1.47 ● Ultimately DIRC PID provides likelihood for a particular particle to be e, μ, π, K, p based on the hit pattern ● PANDA particle identification requirement: ≥ 3 standard deviations π/K separation in the momentum range 0.5 GeV/c – 3.5 GeV/c ● 3 σ π/K separation at 3.5 GeV/c means ≤ 3 mrad Cherenkov angle resolution Expected PID performance example, based on the initial design and simplified fast simulation BABAR DIRC, taken as the initial version of the PANDA Barrel DIRC

PANDA GSI MeetingMaria Patsyuk7 Required DIRC resolution ● Approach based on the PID TAG report ● Measure of PID quality – separation power (selection procedure is not fixed) ● The most challenging PID for DIRC – K/π at high momentum ● Parametrized detector response (no particle tracking) ● Gaussian detector signal (mean, sigma) for K and π in each point of the phase space (p, θ) ● Mis-id map (p, θ) ● Benchmark channels (used in PID TAG report) → overlay signals with mis-id maps and calculate the average mis-id. Gaus signal, gaus bkg → mis-id = 1 - efficiency

PANDA GSI MeetingMaria Patsyuk8 Approach in detail Basic parameters used (for the simple geometry): ● SPR = 18 mrad ● PD total efficiency = 8.8 % ● Tracking resolution = 1 mrad (was not considered in the PID TAG report) ● N photons is parametrized (λ1 = 290nm, λ2 = 580nm) (λ1 = 280 nm, λ2 = 330 nm, in the PID TAG report the photon yield is about 3 times smaller than expected) Most crutial region P = 3 GeV/c, blue line – PandaRoot, red – parametrization

PANDA GSI MeetingMaria Patsyuk9 Mis-id for the benchmark channels Separation power map (simple DIRC geometry, no focusing) Mis-id map (simple DIRC geometry, no focusing) Efficiancy map (simple DIRC geometry, no focusing)

PANDA GSI MeetingMaria Patsyuk10 Mis-id for the benchmark channels Mis-id 0.5%

PANDA GSI MeetingMaria Patsyuk11 PANDA Barrel DIRC ● PANDA DIRC is based on BABAR DIRC with several important improvements: - compact expansion volume (EV) - focusing system - fast photon timing ● 80 radiator bars (synthetic fused silica) in 16 bar boxes, each bar 17 mm x 33 mm x 2500 mm ● Barrel radius 46 cm ● 30 cm deep expansion volume filled with oil ● k channels of MCP-PMTs, ● time resolution ~100 ps ● To achieve design resolution and reduce ● the detector cost, many parameters are to be ● optimized and a number of design ● options are being studied by: 1. Simulation 2. Test experiments Barrel DIRC in simulation

PANDA GSI MeetingMaria Patsyuk12 Performance study: Simulation - detailed detector simulation using Geant - the baseline design as well as a number of design options are implemented Critical parameters: 1. width and thickness of the radiator 2. focusing system: lens or mirror 3. shape of the expansion volume The prism, the lens and the focusing mirrors in the simulation

PANDA GSI MeetingMaria Patsyuk13 Performance study: Simulation To evaluate the performance of different designs a proven BABAR-type reconstruction was used. The Cherenkov angle is determined for each detected photon by comparing the direction of the particle track (taken from other detector systems) and the direction of the detected photon, approximated using the pixel and the bar positions (taken from the look-up table) → single photon resolution of the detector. Apply this procedure to all photons from the same track → peak at the right value + combinatorial background

PANDA GSI MeetingMaria Patsyuk14 map of single photon resolution for one bar box, 3 GeV muons No focusing, fused silica bars are directly attached to the expansion volume (EV) Simple estimation of single photon Cherenkov angle resolution mrad bar 1 bar 2 bar 3 bar 4 bar 5 Performance study: Simulation

PANDA GSI MeetingMaria Patsyuk15 Performance study: Simulation A map of number of detected photons per charged track A map of Cherenkov angle resolution per track assuming ideal tracking and perfect bar shape bar 1 bar 2 bar 3 bar 4 bar 5 bar 1 bar 2 bar 3 bar 4 bar 5 Required track resolution – better than 3 mrad !

PANDA GSI MeetingMaria Patsyuk16 Summary & Outlook ● The hadron PID for PANDA in the barrel region will be performed by the Cherenkov detector of the novel DIRC-type ● PANDA Barrel DIRC is based on the BABAR DIRC with some important improvements: fast photon timing, compact expansion volume and focusing system ● The required DIRC resolution is being studied ● In order to choose the best realization of these improvements and optimize the design in terms of detector performance and cost the design options are being studied in detailed detector simulation

PANDA GSI MeetingMaria Patsyuk17 DIRC beauty A pretty kaleidoscope effect typical for DIRC optics. Looking into the bar from the end towards the front where optical grease was used to couple bar to the window, after the bar was pulled back from the window, some grease residue left on the bar