Klaus Fhl PANDA Cerenkov workshop in Glasgow 11

Klaus Föhl, PANDA Cerenkov workshop in Glasgow, 11 May 2006 Disc DIRC

. . . the intended agenda. . . • • • quick orientation for non-pandas brief particle ID motivation Cherenkov radiation flypast lightguides and simulations photo readout and B-field Plexiglass? Temperature! To. P Test Experiments. . .

the current GSI Gesellschaft für Schwerionenforschung

the new FAIR SIS 100/300 planning as of 2004 Facility for Antiproton and Ion Research

Antiprotons at FAIR SIS 100/300 Panda HESR planning as of 2004 1 Ge. V/c – 15 Ge. V/c

PANDA Side View Pbar AND A Anti. Proton ANihilations at DArmstadt

Particle ID in PANDA

Particle ID in PANDA 22 degrees 5 degrees

Particle ID & Kinematics pp i. e. charmonium production D+ K - + + - + + K - K + + pp KK T=5, 10, 15 Ge. V/c - + K+ even or K- - + + + distinguish and K (K and p). . . if mass known, particle identified need to measure two quantities: d. E/dx energy momentum (tracking in magnetic field) velocity (Cherenkov Radiation) pp DD D K T=6. 6 Ge. V/c For what channels do we not have this factor 2 -3 reduction?

Cerenkov Radiation <1 =1 Cerenkov angle depends on particle speed the cone gives a ring image on a detector plane material with a different dispersion chromatic dispersion prism: correcting dispersion lens: turning angle into position parallel light paths

DIRC: Ba. Bar-type versus Disc 4 -fold direction ambiguity angle and edges crucial 2 -fold ambiguity in disc, lifted at readout only parallel surfaces required

Solid Angle onto flat surface conserving angles and circles 45 90 degrees

Light transmitted in DISC conserving angles and circles 45 90 degrees

Colour fringes on rings conserving angles and circles 45 90 degrees

coordinates measured at rim 45 90 degrees

3 -prong event in DISC 45 90 degrees

Li. F focussing is better than 1 mm over the entire line chosen as focal plane polynomial coefficients: c 2= -3. 0/(60^2) c 3= -0. 5/(60^3) c 4= -0. 1/(60^4)

Location Changes

Location Changes

Location Changes

Lightguide-Designs polynomial coefficients: c 2= -3. 0/(60^2) c 3= -0. 5/(60^3) c 4= -0. 1/(60^4) focussing is better than 1 mm over the entire line chosen as focal plane possibly difficult design requirements: 1) vertical focal plane (normal to B-field) 2) short focal plane (high dispersion deg/mm) polynomial coefficients: c 2= -5. 4/(60^2) c 3= -0. 9/(60^3) c 4= -0. 5/(60^4)

Status of simple Disc Simulations – perfect surfaces – proper directions • recent improvements – true 3 D – analysis of pixel hits • in the pipeline – angular straggling -important for (e, ) and ( , ) – further optimising – include upstream tracking (necessary? ) • NOT: – – – no diffraction no polarisation no background (particles and photons) no maximum likelihood analysis not free of minor approximations (KISS)

status of simulations vertex provided position provided all from DISC data 64 lightguides (no pixels) 128 (no pixels) Y Y L R L A A IN E M R I unpixelised focal plane nondispersive materials L no chromatic correction E R P fluctuations numerical artefact - it’s on the “to do” list. . .

• further optimisation • resolution scaling with pixels • resolution not scaling with pixel size (momentum resolution) 4 ~ (pixel number * quantum efficiency)

Solenoid and Yoke Environment Yoke Solenoid Housing

Photon Detectors • • • phototubes APDs channel plate phototubes optical fibres and external phototubes HPDs with magnetic imaging

Position-sensitive Phototubes H 8500 H 9500 B-field probably too strong R 3292 10 cm

Light guide or fibre readout? Yoke determination

HPD with magnetic imaging Klaus Föhl 2 -June-2004

E B epho toca thod e a c i l is de s uf possibly higher quantum efficiency in reflective photocathode or geometry Detect trip S icon Sil B e. E de pho toca fused silica

Temperature • • cold solenoid, cold EMC maybe coolde APDs Si. O 2, Li. F different expansion coefficients dew, condensation on surfaces

Radiation Countermeasures? Yoke what radiation fields? do we need radiation shielding? will PB act: -as absorber -or as converter?

Plexiglass as Cerenkov radiator? • transmission – Si. O 2 300 -600 nm N 0/mm=14 – plexi 400 -600 nm N 0/mm= 7 • radiation hardness – Ba. Bar “Spectrosil” proven – plexiglass “hamm wer doa” not proven • but: radiation length X 0 three times larger – 36 cm versus 12 cm (40. 5 g/cm 2 vs 26 g/cm 2) more photons per X 0 less chromatic dispersion no UV-grade material necessary (glass, glue, PMT) – focussing optics probably ok for thicker radiator – availability? time stability? radiation hardness? maybe not such a stupid idea higher lower dispersion

Time-of-Propagation in a dispersive medium Light propagation speed perpendicular to Cherenkov-light-emitting particle track: =300 nm photon is 6% slower than 600 nm fused silica (aka quartz) 6% 2% larger Cherenkov angle – 2% shorter path 4% time difference ( =600 nm is “faster”) difference equivalent to =0. 04 for 120 cm radial distance To. P=8. 3 ns (400 nm) 0. 33 ns spread in arrival time

To. P in DISC – some thoughs. . . • chromatic time correction – do not see how (I see no space for red light to run extra length) (unless photon detector timing can be made colour-dependent) • disc not self-timing “GPS altitude problem” • external time reference should be 100 ps/sqrt(N) • if time reference from target vertex factor 2 better overall situation equivalent to 4. 5 metres TOF • >>50*multiplicity pixels needed • multiple hits can be separated if spaced apart

Towards Test Experiments • Radiator slab (fused silica, plexiglass) • Focussing lightguide – Edinburgh workshop: • perspex: ok • quartz: we are happy to try (difficulties anticipated) • photon readout • DAQ

Conclusions?

Conclusions?

Material Testing transmission and total internal reflection of a fused silica sample (G. Schepers and C. Schwarz, GSI)

Outline • • • FAIR international accelerator facility Particle ID – the physics requirements Cerenkov Radiation DIRC in PANDA Detector performance Conclusions and Outlook working on Cerenkov detectors for PANDA: Edinburgh, GSI, Erlangen, Gießen, Dubna, Jülich, Vienna, Cracow, Glasgow

centre hole K K over 64 s each recta dete 20% (dens 8% r fuse 10 m figure of merit N N(ideal) = N 0 x 1 geometric trans N(detected) = 8 K threshold Pion-Kaon-Separation

Conclusions • optical properties of this design are good enough • performance depends on number of pixels • • optical test bench phototubes + electronics operational detector slice testbeam experiments

Side View 2000 mm 1500 mm 10 mm fused silica plate (density 2. 2 g/cm 3 , 8% radiation length) plate radius 1500 mm , detection plane radius 2000 mm wavelength range 300 -600 nm, detection efficiency 20% figure of merit N 0 = 152 cm -1 N(ideal) = N 0 x 1 cm x sin 2( ) = 82 N(detected) = 82 x 0. 61 = 50 geometry transmittance

Photon Lines in space particle vertices target point

Lensing cylinder lense N. B. to be compared with 10 mm pixel height spread over prism width

Chromatic Correction higher dispersion glass spread =300 nm to 600 nm

Lensing cylinder lense N. B. to be compared with 10 mm pixel height spread over prism width

Chromatic Correction higher dispersion glass spread =300 nm to 600 nm

Chromatic Correction higher dispersion glass + spread =300 nm to 600 nm effective pixel height

Po ve ynt cto ing r Cherenkov radiation wa ve c fro nt

Po ve ynti ct ng or Cherenkov radiation in a dispersive medium wa ve fr c on t

Cherenkov radiation in a dispersive medium fused silica (aka quartz) 6% 2%

Momentum Thresholds p K K K aerogel n=1. 05 p total internal reflection limit fused silica n=1. 47 p n=1. 47

tracks in Solenoid field solenoid field taken to be homogenous within the real field shape the particles are better aligned with the field lines

fused silica Li. F side view front view B. Morosov, P. Vlasov et al. December 2004 Li. F fused silica

concurrent optimisation goals adjusting polynomial coefficients (c 2 fixed, c 3 and c 4 so far used only) to find a mirror shape that provides overall acceptable focussing along a straight line (easier to instrument)

conserving angles and circles

fused silica side view completely within medium all total reflection compact design all solid material flat focal plane focussing is better than 1 mm over the entire line chosen as focal plane polynomial coefficients: c 2= 1/1200 c 3= -0. 5/(60^3) c 4= -0. 1/(60^4)

Particle ID in PANDA If particle mass is known, the particle is identified. 22 degrees 5 degrees For particle ID, two quantities are required: d. E/dx energy momentum (trackingininmagneticfield) velocity (Cherenkov. Radiation)

briefly on Barrel-DIRC time-of-propagation version Klaus Föhl, FAIR-Panda-PID-meeting, 5/12/2005

incident particle at 45 degrees =0. 95 =1 Cherenkov radiation in a dispersive medium fused silica slab 3 m long =0. 99 =600 nm =300 nm correction 1 =300 nm correction 2

Cherenkov radiation in a dispersive medium fused silica (aka quartz) 6% 2% reduce wavelength range to improve sensitivity

dispersion correction needs to cover entire angular range of incident particles

dispersion correction no correction improving over the entire angular range

my conclusions Barrel-DIRC • photon group velocity in standard PMT timing is enough consider to cut out <400 nm dispersive medium • photon detector number set photons/pixel << 1 most stringent requirement by statistics • dispersive correction not configuration angle-dependent useless for the barrel covering all relevant angles • reference timing provided no external timing required to analyse barrel DIRC data by first arriving photons