Construction and expected performance of the Hadron Blind Detector for PHENIX experiment at RHIC Alexander Milov (for the PHENIX HBD group) XIX International conference on Ulterarelativistic Nucleus-Nucleus Collisions, Shanghai, China Alexander Milov QM 2006, Shanghai Nov 15, 2006
People in this project v Weizmann Institute of Science (Israel) A. Dubey, Z. Fraenkel, A. Kozlov, M. Naglis, I. Ravinovich, D. Sharma, L. Shekhtman (on leave from BINP), I. Tserruya (project leader) v Stony Brook University (USA) W. Anderson, A. Drees, M. Durham, T. Hemmick, R. Hutter, B. Jacak, J. Kamin v Brookhaven National Lab (USA) B. Azmoun, A. Milov, R. Pisani, T. Sakaguchi, S. Stoll, C. Woody (Physics) J. Harder, P. O’Connor, V. Radeka, B. Yu (Instrumentation Division) A. Sickles, v Columbia University, Nevis Labs (USA) C-Y. Chi v University of Tokyo (Japan) T. Gunji, H. Hamagaki, M. Inuzuka, T. Isobe, S. X. Oda, K. Ozawa, S. Saito Y. Morino, v RIKEN (Japan) S. Yokkaichi v Waseda University (Japan) Y. Yamaguchi v KEK (Japan) S. Sawada Alexander Milov QM 2006, Shanghai Nov 15, 2006 2
Why di-electrons? v Effects of chiral symmetry restoration manifest themselves in terms of in-medium modifications of the line shapes of low mass vector mesons (e. g. , mass shifts, spectral broadening) Part of the p+p run no bkg. subtraction v Lepton pairs are unique probes because they provide direct information undistorted by further interactions. Entire Au. Au run • ρ (m = 770 Me. V τ ~ 1. 3 fm/c) e+e • ω (m = 782 Me. V τ ~ 20 fm/c) e+e • φ (m =1020 Me. V τ ~ 40 fm/c) e+e- Alexander Milov QM 2006, Shanghai Nov 15, 2006 3
Background sources v Main source of the background due to external and internal conversions of the photons coming from π0. π0 e+ eπ0 e+ e- e+ e po e+ e “combinatorial pairs” v The goal is to reduce the background by a factor of 100 S/B ~ 1/500 v Distinct pattern of the background producing decays: small inv. mass small opening angle Irreducible charm background signal v A rejection factor of >90% on a close pair will reduce the background to an acceptable level. Alexander Milov total background QM 2006, Shanghai charm signal Nov 15, 2006 4
The detector concept v Proximity Focused Windowless Cherenkov Detector B≈0 signal electron v Radiator gas = Working gas Primary choice pure CF 4 n = 1. 00062 ( =28 ) L = 50 cm Blind to π0 with p. T<4 Ge. V/c v Radiating particles produce blobs on an image plane qp Cherenkov blobs e- partner positron needed for rejection air opening angle e+ (θmax = cos-1(1/n)~36 mrad Blob diameter ~ 3. 6 cm) v To preserve the pair opening angle θpair the magnetic field is turned off (compensated) in the detector v Background processes produce 2 close blobs and single electrons only 1 Alexander Milov ~1 m v Image plane: Cs. I photocathode on top of tiple GEM stack used for electron amplification separated by 90% transparent mesh from the main volume QM 2006, Shanghai Nov 15, 2006 5
Challenges & Solutions v The space where B can be compensated is limited to ~50 cm but the number of p. e. must be high enough to allow for effective amplitude analysis of overlapping and distorted blobs. ü Match the Cs. I Q. E. ~70% @ 10 e. V and pure CF 4 bandwidth (6 -11. 5 e. V) to get unprecedented N 0 ≈840 cm-1 (x 6 larger than any e/π RICH ever built!) v The detector has to let all ionizing particles through without seeing them, but pick up single photoelectrons. ü Make Cs. I + GEMs into a new type of semitransparent photocathode such that it a) is sensitive to the ionization reaching its surface from Cherenkov light b) electric field drives MIP ionization back into the gas volume v The detector must be thin to produce little own background but leak tight to keep water away from absorbing UV light. ü Windowless design (CF 4 without quencher = gaseous radiator = detector gas). üCombine functions of the detector structural elements (pad plane = gas seal) Alexander Milov QM 2006, Shanghai Nov 15, 2006 6
The Image plane v Start with a GEM v Put a photocathode on top HV v Electron from Cherenkov light goes into the hole and multiplies v Use more GEMs for larger signal v Pick up the signal on pads v And why is it Hadron Blind? v Mesh with a reverse bias drifts ionization away from multiplication area v Sensitive to UV and blind to traversing ionizing particles Alexander Milov QM 2006, Shanghai Nov 15, 2006 7
The Detector is designed and built at the Weizmann Institute v The detector fits under 3%X 0 and it is leak tight to keep water out 0. 12 cc/min (~1 volume per year)! FEEs Side panel Reado ut plane Mylar window Honeycomb panels v Readout plane with 1152 hex. pads is made of Kapton in a single sheet to serve as a gas seal v Each side has 12 (23 x 27 cm 2) triple GEM Detectors stacks: Mesh electrode Top gold plated GEM for Cs. I Two standard GEMs pads HV terminals Sealin g Service panel Alexander Milov Triple GEM module with mesh grid QM 2006, Shanghai Nov 15, 2006 8
Detector elements v Detector construction involves ~350 gluing operations per box v Dead areas are minimized by stretching GEM foils on a 5 mm frames and a support in the middle. v GEM positioning elements are produced with 0. 5 mm mechanical tolerance. Alexander Milov QM 2006, Shanghai Nov 15, 2006 9
Detector assembly “Clean Tent” a. k. a. “The Battle Field of Stony Brook” Cs. I Evaporator and quantum efficiency measurement (on loan from INFN) Laminar Flow Table for GEM assembly High Vacuum GEM storage 6 men-post glove box, continuous gas recirculation & heating O 2 < 5 ppm H 2 O < 10 ppm Class 10 -100 ( N < 0. 5 mm particles/m 3) Alexander Milov QM 2006, Shanghai Nov 15, 2006 10
Photocathode production v Cs. I evaporation station was given on loan to Stony Brook from INFN/ISS Rome Thank you Franco Garibaldi & Italian team! v Produces 4 photocathodes per shot 240 – 450 nm of Cs. I @ 2 nm/sec Vacuum drops to 10 -5 Torr and then to 10 -7 Torr (water out of the structure). Contaminants measured with RGA v Photocathode Q. E. is measured “in situ” from in 165 -200 nm wavelength range over entire area v Photocathodes transported to glove box without exposure to air v 4 small “chicklets” at same control Alexander Milov QM 2006, Shanghai evaporated time for full QE (120 -200 nm) Nov 15, 2006 11
Some of the production steps GEMs pre-installed for evaporation Photocathode installation chain: removal from transfer box, gain test, installation into the HBD. First module installed in HBD West Alexander Milov QM 2006, Shanghai Nov 15, 2006 12
The GEM stacks v GEMs produced at CERN Tested for 500 V in air @ CERN Framed & tested @ WIS for gain uniformity Tested at SUNYSB prior to installation Gain uniformity between 5% and 20% v GEM statistics 133 produced (85 standard, 48 Au plated) 65 standard, 37 Au plated passed all tests 48 standard, 24 Au plated installed GEMs combined into stacks are matched to minimize gain variation over the entire detector 5% v All GEMs pumped for many days under 10 -6 Torr prior to installation into detector Alexander Milov QM 2006, Shanghai Nov 15, 2006 13
GEM gain stability v During gain mapping, a single pad is irradiated with a 8 k. Hz 55 Fe source for ~20 min. Then all other pads are measured (~1. 5 h) and the source is returned to the starting pad. Secondary rise v Gain is observed to initially rise and then reach a plateau. Rise can be from few % to almost a factor of 2. 1. 5 Initial Rise v Further study show that the gain increase is rate dependent (10 -30%) v This does not impose a problem for GEM operation at PHENIX GEMs will reach operating plateau in a few hours Rates are lower then during mapping Alexander Milov QM 2006, Shanghai Nov 15, 2006 14
Photocathode quality v Q. E. needs to distinguish a single electron from a pair. 36 72 Number of photoelectrons v Absolute Q. E. must be continuously controlled and preserved. At the production stage During transportation and installation During physics data taking 27 cm Flat position dependence v At the production stage the Q. E. is as high as measured in R&D stage and uniform Alexander Milov QM 2006, Shanghai Nov 15, 2006 15
Gas transparency Monochromator (120 -200 nm) is a part of the HBD gas system Movable mirror D 2 lamp H 2 O & O 2 must be kept at the few ppm level to avoid absorption in the gas Turbopump Lamp Monitor Gas Cell Monitor Measure photocathode current of Cs. I PMTs Heaters are installed on each detector to drive out water from GEMs and sides of detector vessel Alexander Milov QM 2006, Shanghai Nov 15, 2006 16
Full scale prototype test Tested in PHENIX with p-p collisions at RHIC April-June ‘ 06 Pulse height, reverse bias v Full scale detector prototype: 1 GEM + Cs. I stack module installed in the volume 68 readout channels full readout chain v Pure CF 4 gas system v LVL 2 triggers to enrich e-sample electrons hadrons e/π rejection ~85% at εe ~90 % Cluster size, reverse bias MIP electrons hadrons Forward Bias+Landau Reverse Bias Alexander Milov QM 2006, Shanghai Nov 15, 2006 17
Now HBD West (front side) Installed 9/4/06 Alexander Milov HBD East (back side) Installed 10/19/06 QM 2006, Shanghai Nov 15, 2006 18
Summary v The HBD will provide a unique capability for PHENIX to measure low mass electron pairs in heavy ion collisions at RHIC v This detector incorporates several new technologies (GEMs, Cs. I photocathodes, operation in pure CF 4, windowless design) to achieve unprecedented performance in photon detection N 0~840 cm-1 v The operating requirements are very demanding in terms of leak tightness and gas purity, but we feel they can be achieved v Tests with the full scale prototype were very encouraging and demonstrated the hadron blindness properties of the detector. v The final detector is now installed in PHENIX and ready for commissioning and data taking during the upcoming run at RHIC Alexander Milov QM 2006, Shanghai Nov 15, 2006 19
BACKUPS Alexander Milov QM 2006, Shanghai Nov 15, 2006 20
Challenges & Solutions v The space where B can be compensated is limited to ~50 cm but the number of p. e. must be high enough to allow for effective amplitude analysis of overlapping and distorted blobs. ü Match the Cs. I Q. E. ~70% @ 10 e. V and pure CF 4 bandwidth (6 -11. 5 e. V) to get unprecedented N 0 ≈840 cm-1 (x 6 larger than any e/π RICH ever built!) v The detector has to let all ionizing particles through without seeing them, but pick up single photoelectrons. ü Make Cs. I + GEMs into a new type of semitransparent photocathode, which a) does not have usual losses for such type of photocathode b) allows multi-stage multiplication to follow it. v The detector must be thin to produce little own background but leak tight to keep water away from absorbing UV light. ü Go to windowless design by using CF 4 without quenching gas both as a radiator and working gas due to the fact that GEMs have no photon feedback Alexander Milov QM 2006, Shanghai Nov 15, 2006 21
PHENIX now e+ e- e- e+ ~12 m Alexander Milov QM 2006, Shanghai Nov 15, 2006 22
HBD parameters Acceptance nominal location (r=5 cm) retracted location (r=22 cm) | | ≤ 0. 45, =135 o | | ≤ 0. 36, =110 o GEM size ( , z) Number of detector modules per arm Frame Hexagonal pad size Number of pads per arm Dead area within central arm acceptance Radiation length (central arm acceptance) Weight per arm (including accessories) 23 x 27 cm 2 12 W: 5 mm T: 0. 3 mm a = 15. 6 mm 1152 6% box: 0. 92%, gas: 0. 54% <10 kg Alexander Milov QM 2006, Shanghai Nov 15, 2006 23
Readout chain 15 mm Differential output 19 mm Preamp (BNL IO 1195) 2304 channels total Noise on the bench looks very goo Gaussian w/o long tails 3 s cut < 1% hit probability Alexander Milov QM 2006, Shanghai Nov 15, 2006 24
Run Plan q Run 7 (Dec ‘ 06 – June ’ 07) q ~ 4 weeks commissioning with Au x Au beams at s. NN = 200 Ge. V q 10 weeks data taking with Au x Au at s. NN = 200 Ge. V q 10 weeks data taking with polarized p-p beams at s = 200 Ge. V q Run 8 (Fall ’ 07 – Summer ’ 08) • 15 weeks d-Au at s. NN = 200 Ge. V • 10 weeks polarized p-p at s = 200 Ge. V q Run 9 (Fall ’ 08 – Summer ’ 09) • 10 -15 weeks heavy ions (different energies and possibly species) • 15 -10 weeks polarized p-p at s = 500 Ge. V (including commissioning) Alexander Milov QM 2006, Shanghai Nov 15, 2006 25
Photocathode and gas. Ø Photocathode: ü Cs. I is an obvious choice. ü We are using INFN built evaporator, currently at Stony Brook to do this project. § High area, Ø Gas CF 4 (was not really known): § High vacuum, ü Has high electron extraction § In-situ Q. E. control, probability § Zero exposure to open air. ü Has avalanche self quenching mechanism Ø Gas CF 4 (well known): ü Transparent up to 11. 5 e. V, makes perfect match to Cs. I ü Is a good detector gas. Alexander Milov QM 2006, Shanghai Nov 15, 2006 26
The design. Made of 2 units with R~60 cm, the volume is filled with CF 4 magnetic field is turned off Electrons emit Cherenkov light is registered by 12 photo-detectors in each unit Signal is read out by 94 pads in each unit, pad size ~ size of a circle Accumulating ~36 photoelectrons from each primary electron, while most other operational RICHes have ~15 or less. High statistics allows to separate 2 close electrons even if their signals overlay! 36 72 Number of photoelectrons Alexander Milov QM 2006, Shanghai Nov 15, 2006 27
Event display (simulation). Alexander Milov QM 2006, Shanghai Nov 15, 2006 28
Background sources? Ø In the decays contributing to the background: ü π 0 e + e- γ ü π 0 γ γ e + e- γ Ø Only one electron is detected in PHENIX and another is lost Ø To cut the background we need a new detector such that: ~12 m Alexander Milov ü It sees only electrons ü Located at the origin ü It does not produce its own background (is thin) ü… ü… ü… QM 2006, Shanghai Nov 15, 2006 29
What does it look like • All raw materials (FR 4 sheets, honeycomb, HV resistors, HV connectors) ordered and most of them in house • Detector box design fully completed • Jig design underway • Small parts (insert, pins, screws, HV holders. . ) in the shops • Detector construction to start Nov. 1 st • PCB design almost complete • Detailed construction schedule foresees shipment of boxes to SUNY in January 2006. Alexander Milov QM 2006, Shanghai Nov 15, 2006 30
Mechanical parts and PCB final design. Quick MC shows no difference with standard cells Alexander Milov QM 2006, Shanghai Entrance window frames are ready, the window itself to be tight between them Nov 15, 2006 31
Скачать презентацию Construction and expected performance of the Hadron Blind
6b9b40daece3e63598d83372e63fff34.ppt