Compact RICH detector Claudia Hhne GSI Darmstadt Status

Compact RICH detector Claudia Höhne, GSI Darmstadt

Status electron simulations (March '07) • all relevant detectors (RICH, TRD, TOF) fully implemented in Cbm. Root • RICH → ring finding, fitting, fake rejection, track matching • TRD → 3 different routines available for calculating e-p separation from statistical analysis of energy loss spectra (likelihood, neural net, Wkn) • TOF • analysis of electron ID performance (efficiency, purity) • full electron ID used for physics analysis (low-mass dileptons, J/y) → "proof of principle": got everything running with reasonable results Claudia Höhne CBM collaboration meeting Dresden, September 2007 2

Improvements since March '07 → make use of available infrastructure and continue software development: • systematic studies: → S. Lebedev • limitations/ relevant parameters for ring finding: hit/ ring density, noise level, magnetic stray field → impact on RICH design • electron ID for different identification cuts, usage of detectors, . . . → S. Das • develop user friendly software in Cbm. Root for electron ID → S. Das • optimization of cuts • simple usage for physics analysis • develop event display of CBM detector → S. Lebedev • apply latest developments to physics analysis, understand remaining background sources → T. Galatyuk, A. Maevskaya, . . • optimization of RICH layout • minimum size for reasonable physics output? (costs!) • maximum Claudia Höhne material budget whichmeeting is tolerable global tracking? CBM collaboration Dresden, for September 2007 3

Compact RICH – motivation • save money! → reduce size of RICH: • restrict acceptance (reduce mirror size, photodetector size) • reduce length L (→ overall size). . . first steps will be presented • Cherenkov radiation for electrons → choose different radiator (Cherenkov angle q) in order to compensate for reduction in L Claudia Höhne CBM collaboration meeting Dresden, September 2007 4

Radiator • as the RICH detector should mainly serve for electron identification, the pion threshold for emission of Cherenkov light should be sufficiently high in momentum radiator n gth pthp [Ge. V/c] lthresh [nm] radiation length [m] handling? N 2 1. 000298 41 5. 6 < 160 304 CH 4 1. 00044 33. 6 4. 7 ~ 145 650 CO 2 1. 00045 33. 3 4. 65 ~ 175 183 CF 4 1. 000488 32 4. 47 < 120 N 2 O 1. 000509 31. 4 4. 37 (toxic) CH 3 OH (methanol) 1. 000546 30. 3 4. 2 flammable C 2 H 6 (ethane) 1. 000706 26. 6 3. 71 Claudia Höhne ~ 160 chemically aggressive 340 CBM collaboration meeting Dresden, September 2007 5

Radiator (II) • CO 2 should still allow sufficient e-p separation up to ~ 10 Ge. V/c ! • increase of number of pion rings! momentum distribution of pions in RICH (standard) CO 2 N 2 p [Ge. V/c] Claudia Höhne CBM collaboration meeting Dresden, September 2007 6

Radiator (III) • keeping approximately the same number of photons, the length can be reduced by 30% • expected radius reduction ~ 13 % assuming the same total resolution (0. 15 cm) the relative resolution is still < 3%! radiator n qc gth pthp [Ge. V/c] L [cm] Ng R [cm]# s. R [cm] s. R [%] N 2 1. 000298 1. 4 41 5. 6 250 22 * 6. 1 0. 14 $ 2. 3 CO 2 1. 00045 1. 72 33. 3 4. 65 176 ~22 & 5. 3 ? * from simulation, H 8500 -03 photodetector # R ~ L tanq $ simulation: R = 6. 16± 0. 14 cm & keep appr. constant → L ~1/(sinq)2 → factor 0. 66 from N 2 to CO 2 Claudia Höhne ~30% reduction! CBM collaboration meeting Dresden, September 2007 7

Radiator (IV) • dispersion similar in both gases Claudia Höhne CBM collaboration meeting Dresden, September 2007 8

Radiator (V) • absorption also starts appr. at the same wavelength • N 2 absorption edge ~ 150 nm • CO 2 absorption edge ~ 175 nm Claudia Höhne CBM collaboration meeting Dresden, September 2007 9

RICH layout standard RICH compact RICH first approach: keep overall same concept for layout and only reduce length mirror tilted by 5° Claudia Höhne CBM collaboration meeting Dresden, September 2007 10

RICH layout standard RICH compact RICH gas N 2 CO 2 radiator length 2. 5 m 1. 76 m full length of RICH * 2. 9 m 2. 1 m end of RICH 4. 5 m 3. 7 m mirror position 4. 1 m 3. 36 m mirror radius 4. 5 m 3 m mirror size 2 x (5. 7 x 2) m 2 ~ 22. 8 m 2 2 x (4. 2 x 1. 4) m 2 ~ 11. 8 m 2 photodetector size ** 2 x (3. 2 x 1. 4) m 2 ~ 9 m 2 2 x (2. 4 x 0. 78) m 2 ~ 3. 7 m 2 # channels ~ 200 k ~ 85 k * start at 1. 6 m behind target ** still reducable on (small) acceptance losses Claudia Höhne CBM collaboration meeting Dresden, September 2007 11

Photodetector plane standard RICH compact RICH • distribution of Cherenkov photons for 1000 Ur. QMD events reduce at least to (3. 2 x 1. 2) m 2 can be reduced further (~15%) → 7. 7 m 2 in total (170 k channels) → 3. 1 m 2 (72 k channels) analysis: cut inner region with r < 130 cm Claudia Höhne analysis: cut inner region with r < 140 cm CBM collaboration meeting Dresden, September 2007 12

Ring projection standard RICH compact RICH ring projection in ¼ of photodetector plane for different (f, q) q – 5, 10, 15, 20, 25, 30 degrees y [cm] f – 0, 20, 40, 60, 80 degrees → bended photodetector plane needed! f . . . next step y [cm] q x [cm] Claudia Höhne CBM collaboration meeting Dresden, September 2007 13

Simulations simulation • Ur. QMD, central Au+Au collisions, 25 AGe. V + 5 embedded e± each • magnetic field: "Active. Field" • standard setup, only RICH geometry replaced reconstruction • all routines for RICH can deal with mirrors/ photodetector planes tilted by arbitrary angles around x-axis • Cbm. Rich. Hit. Producer • Cbm. Rich. Projection. Producer reduces current performance but will be improved • compact RICH: distorted rings in particular in outer part no radius correction yet available no fake ring rejection by additional ring quality checks Claudia Höhne CBM collaboration meeting Dresden, September 2007 14

Numbers. . . • central Au+Au collisions, 25 AGe. V beam energy standard RICH compact RICH rings/event 97 75 . . . from e 68 54 . . . from e (NSTS ≥ 6) 14 13 . . . from p 29 20. 5 NPMT/e-ring * 22 ± 4. 6. . . not joking ; -) * Hamamatsu H 8500 -03 MAPMT Claudia Höhne CBM collaboration meeting Dresden, September 2007 15

Electron acceptance standard RICH compact RICH acceptance = tracks in RICH acceptance/ tracks in STS acceptance • STS acceptance : NSTS + NMVD ≥ 4, use prim. vertex tracks only • RICH acceptance: track projected on photodetector p = 10 Ge. V/c p = 0. 5 Ge. V/c y Claudia Höhne CBM collaboration meeting Dresden, September 2007 y 16

Ring reconstruction efficiency standard RICH compact RICH p [Ge. V/c] Hough Transform p [Ge. V/c] Elastic Net Ring Finder → next slide Claudia Höhne CBM collaboration meeting Dresden, September 2007 17

Ring finding – compact RICH Hough Transform: → losses in outer regions due to distortions! for "good rings" 90% level reached as for standard RICH! → will change with optimized geometry! → finally same performance as for standard RICH expected distorted rings in outer regions! Claudia Höhne CBM collaboration meeting Dresden, September 2007 18

Radius vs momentum standard RICH compact RICH electrons only broad radius distribution for compact RICH due to distortions! → to be improved by bended photodetector plane (and later software corrections) Claudia Höhne CBM collaboration meeting Dresden, September 2007 19

Ring-track distance standard RICH compact RICH distance distribution broader due to ring distortions (affects ring center fit) no fake ring rejection routines yet!! Claudia Höhne CBM collaboration meeting Dresden, September 2007 20

RICH electron identification standard RICH compact RICH lower p threshold for Cherenkov radiation! fake rings!! (Def. : < 60% "true" hits) Claudia Höhne CBM collaboration meeting Dresden, September 2007 21

RICH + TOF + TRD electron identification standard RICH compact RICH • TRD: Neural Net Method, require output values to be > 0. 8 • TOF 2 -dimensinal cut in m 2 vs p plane → similar performance except fake ring rejection! → in particular TRD copes up for (currently) reduced RICH performance! 96% purity of id. e Claudia Höhne 87% purity of id. e CBM collaboration meeting Dresden, September 2007 22

Electron efficiency standard RICH compact RICH efficiency for identifying electrons embedded in Ur. QMD events, normalized to RICH acceptance p [Ge. V/c] • major improvement since spring! (was ~60% for full ID) Claudia Höhne p [Ge. V/c] • efficiency loss due to ring finding • with optimized photodetector plane same performance as for standard RICH expected! CBM collaboration meeting Dresden, September 2007 23

Pion suppression standard RICH compact RICH p [Ge. V/c] → factor 104 p suppression at 80% e-efficiency! Claudia Höhne p [Ge. V/c] pions are fine! improve on fakes and e-efficiency! CBM collaboration meeting Dresden, September 2007 24

Summary (I) • need for redesign of RICH detector after we could prove the feasibility of low-mass dilepton and J/y physics with the "first guess" layout of RICH • first steps towards a compact RICH detector presented: • factor > 2 reduction in mirror size (to be adopted to photodetector) • ~ factor 2. 8 reduction in photodetector size. . . while to first extend keeping the physics performance! → RICH becomes affordable!! • optimize further and finally come up with a mature RICH design !!! • increase activities on RICH R&D in parallel! → activities started on mirror R&D and photodetector module design aiming finally for a small RICH prototype Claudia Höhne CBM collaboration meeting Dresden, September 2007 25

Summary (II) being RICH is wonderful! Claudia Höhne CBM collaboration meeting Dresden, September 2007 26

Radius versus momentum standard RICH compact RICH particles from Ur. QMD only p new HT supresses rings with R < 5 cm → pion rings suppressed Claudia Höhne as expected bad resolution due to distorted rings. . . CBM collaboration meeting Dresden, September 2007 27

Radius standard RICH compact RICH • radius for rings from primary electrons with 1 Ge. V < p < 6 Ge. V → should be gaussian! not gaussian! still distortions in! broad distribution of radii! Claudia Höhne CBM collaboration meeting Dresden, September 2007 28

Distance (II) • mirror tilted by 5° • photodetector tilted by -19° • plot distance for ideal matching stronger ring distortions → increased distances → Cbm. Root can deal with these cases! Claudia Höhne CBM collaboration meeting Dresden, September 2007 29

RICH identified e in TRD standard RICH compact RICH TRD: Neural Net Method, require output values to be > 0. 8 TRD copes up for reduced RICH performance! Claudia Höhne CBM collaboration meeting Dresden, September 2007 30

RICH identified e in TOF standard RICH compact RICH TOF 2 -dimensinal cut in m 2 vs p plane Claudia Höhne CBM collaboration meeting Dresden, September 2007 31

Identified electrons identified electrons/ event from Ur. QMD only • "standard" includes ring radius corrections and fake ring rejection by neural net evaluation → comparable performance except fakes!! RICH only RICH + TRD + TOF compact standard all id. e 2. 03 4. 23 7. 28 true e 1. 76 4. 07 1. 3 1. 1 true p 0. 05 true p 0. 18 0. 17 true p 0. 01 0. 02 fake rings 1. 52 0. 56 fake rings 0. 41 0. 15 purity 56% 80% purity 87% 96% compact standard all id. e 6. 5 9. 1 true e 3. 66 true p Claudia Höhne CBM collaboration meeting Dresden, September 2007 32

Appendix: compact Rich performance number of hits for fake rings → cut! Claudia Höhne CBM collaboration meeting Dresden, September 2007 33

Appendix: compact RICH performance (II) momentum distribution of pions Claudia Höhne CBM collaboration meeting Dresden, September 2007 34

Appendix: compact Rich performance (III) ring-track distance vs mometum true Claudia Höhne CBM collaboration meeting Dresden, September 2007 false 35