The Hadron Blind Detector of the PHENIX Experiment

The Hadron Blind Detector of the PHENIX Experiment at RHIC RICH 2010, Cassis, May 2 -7, 2010 Itzhak Tserruya

Outline Ø Motivation Ø Detector Concept Ø Design and Construction Ø Operation Ø Performance Ø Itzhak Tserruya Summary RICH 2010, Cassis, May 2 -7, 2010 2

1. Motivation Itzhak Tserruya RICH 2010, Cassis, May 2 -7, 2010 3

Motivation (I) Electron pairs (or dileptons in general) are unique probes to study the hot and dense matter formed in relativistic heavy ion collisions at RHIC: – best probe for chiral symmetry restoration and in-medium modifications of light vector mesons , ω and – sensitive probe for thermal radiation: QGP qqbar * e+e. HG + - * e+e. Experimental challenge: huge combinatorial background created by e+e- pairs from copiously produced 0 Dalitz decay and conversions. e+ e o e+ e - Itzhak Tserruya combinatorial e+ e - pair RICH 2010, Cassis, May 2 -7, 2010 4

Background sources in PHENIX Ø Main sources contributing to the combinatorial background: 0 e + e- q e+ e- q Ø It often happens that only one electron is detected in PHENIX and the other is lost due to: § limited geometrical acceptance § low p. T particle curling in the magnetic field or not reconstructed. ~12 m q Baseline PHENIX detector: S/B ratio of 1/200 at m = 500 MEV/c 2 Main goal of the HBD: recognize tracks from conversions and 0 Dalitz decays thereby considerably reducing the combinatorial background q

2. Concept Itzhak Tserruya RICH 2010, Cassis, May 2 -7, 2010 7

Upgrade Concept Strategy Exploit the fact that 0 Dalitz decays and conversions have a very small opening angle to identify them. • Create a field free region close to the vertex to preserve the opening angle of close pairs. • • Identify electrons in the field free region • reject close pairs. Hardware realization * HBD in inner region * Inner coil B 0 for r 60 cm Main task of the HBD: distinguish single vs double e-hit Itzhak Tserruya RICH 2010, Cassis, May 2 -7, 2010 8

HBD Concept HBD concept: ♣ windowless Cherenkov detector (L=50 cm) ♣ CF 4 as radiator and detector gas ♣ Proximity focus: detect circular blob not ring UV-photon beam axis hadron 5 cm E Detector element: ♣ Triple GEM stack with pad readout ♣ Cs. I reflective photocathode evaporated on top face of top GEM ♣ RB mode of operation to repel ionization charge CF 4 radiator detector element Itzhak Tserruya RICH 2010, Cassis, May 2 -7, 2010 50 cm ~1 cm 10

Very attractive features… Unprecedented N 0 Bandwidth 6 - 11. 5 e. V (100 -200 nm) N 0 ≈ 800 cm-1 in an ideal detector with no losses Reflective photocathode No photon feedback Hadron blind: Most of the ionization charge in the drift region is repelled away from the GEM stack and collected by the mesh Hexagonal pads with size (a=15. 5 mm area =6. 2 cm 2 ) comparable to Cherenkov blob size (10. 2 cm 2) hadrons: single pad hit, electrons: 2 -3 pads per hit Low granularity ~1000 pads per central arm acceptance Low gain primary charge of 5 -10 e/pad gain of 5 x 103 is enough …but many open questions extensive R&D A. Kozlov et al, NIM A 523 (2004) 345 Z. Fraenkel et al, NIM A 546 (2005) 466,

3. Design and Construction Itzhak Tserruya RICH 2010, Cassis, May 2 -7, 2010 14

The Detector designed and built at the Weizmann Institute Each arm has 12 (23 x 27 cm 2) triple GEM stacks: Two identical arms • Mesh electrode • Top gold plated GEM for Cs. I & two standard Gems All panels made of honeycomb • Pad electrode & FR 4 structure FEEs v Readout plane with 1152 hexagonal pads made of Kapton in one single sheet to serve as gas seal v Low material budget: vessel 0. 92%, gas 0. 54%, electronics ~1. 5% total under 3% X 0. Reado ut v ~ 350 gluing operations per arm plane v Leak rate 0. 12 cc/min (~1 vol /year)! v Side panel Mylar window HV terminals Sealin g Service panel Triple GEM module with mesh grid v Low dead/inactive area of 6% (0. 5 mm tolerance between adjacent modules)

HBD Construction Jigs Jig #1: Glue active panels together and to the PCB cathode board 6 active panels glued together Jig #2: Glue the rest of the panels to complete the HBD box

The people behind the detector

Detector assembly Cs. I evaporation and final detector assembly in clean tent at Stony Brook Cs. I Evaporator and quantum efficiency measurement (on loan from INFN) Can make up to 4 photocathodes in one shot 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) Itzhak Tserruya RICH 2010, Cassis, May 2 -7, 2010 19

Cs. I evaporation facility Absolute QE of Cs. I photocathodes Ø 4 photocathodes produced per shot together with chicklets for QE monitorin Ø For each GEM 3 measurements are taken at 160 nm across X axis Excellent reproducibility. Ø Excellent stability

The SB plant Jason the crew Gabby Greg Matt Ben-II

4. Operation Itzhak Tserruya RICH 2010, Cassis, May 2 -7, 2010 22

HBD – short history § The HBD was commissioned during the 2007 RHIC run § The detector encountered severe high voltage problems which ultimately damaged many GEMs: • Minor GEM sparks induced larger, more damaging sparks due to large stored energy in filter capacitors • Spark in one module would induce sparks in other modules by optical coupling • Problem exacerbated by Le. Croy 1471 N PS that reapplied HV after a trip causing HV spurious HV spikes § The main problem with GEM sparking was due to dust ! § Detector rebuilt (minimum exposure to glove box, Zener diodes, relay box ) § Detector successfully operated in Run-9 (pp collisions) and in Run-10 (Au+Au collisions – still ongoing) each one about 6 m long. Itzhak Tserruya RICH 2010, Cassis, May 2 -7, 2010 Fe 55 x-ray UV lamp 23

Readout and noise Excellent electronic noise in all modules. Typically = 1. 5 ADC counts corresponding to 0. 15 f. C or 0. 2 p. e. at a gain of 5000. Allowed to operate the detector at a low gain of 3000 to 5000. The entire readout chain worked reliably and smoothly (BNL & Columbia Univ) Mean Pedestal rms vs HV Sigma Mean Sigma

Gain determination using scintillation hits Exploit scintillation hits (identified as single pad hits not belonging to tracks) to determine the gain on a pad by pad basis. Forward Bias Scintillation Ionization Gain determination: q Fit scintillation component with an exp fctn: • 1/slope = G. <m> = avrg nr of scintillation photons in a fired pad) Zoom q In pp collisions <m> 1 Reverse Bias q In Au +Au collisions, assuming the nr of scintillation photons per pad follows a Poisson distribution: <m> = Scintillation: unchanged Ionization: suppressed Itzhak Tserruya RICH 2010, Cassis, May 2 -7, 2010 25

Correction for P/T Variation q Gain monitoring: gain is calculated for each module and each run. q Gain variations due to p/T changes, automatically compensated by varying the operating HV in 5 pre-determined for p/T windows. RUNS PROCESSED: 288094 - 288900 T 3 T 2 IR ACCE SS CONFIG CHANGEAPPLY HV CHANGE T 1 ES 3 WN 5 shown above

Adjusting the Reverse Bias using scintillation light • Requires high precision in adjusting the mesh voltage with wrt the GEM stack, to preserve the p. e. collection efficiency • Requires setting ~ 4 KV PS to ~ 5 V precision ( 0. 1 %) • Scan voltage between mesh and top GEM for each module • Method invented in Run-9 Electrons hadrons ES 1 =-10 V Hits per event • Exploit the scintillation that remains unchanged when switching from FB to RB Operating Point Red (+5 V), black (0 V) green(-5 V), blue(-10 V), rest(-15 V and lower) • Simple, fast and precise Pulse height [ADC counts]

Gas. Transmission Continuously monitored using a monochromator system – Recent scan on April 20 Gas flowing in recirculation mode, with scrubbing at 4. 5 lpm Gas transmission stable over the entire run Input gas upper 90% Output gas lower 90%

5. Performance Itzhak Tserruya RICH 2010, Cassis, May 2 -7, 2010 30

Position Resolution q Hadron tracks reconstructed in central arms projected to HBD. • Position resolution: z ≈ ≈ 1. 5 cm Dictated by: • pad (hexagon) size: a = 1. 55 cm 2 a/√ 12 = 0. 9 cm • vertex resolution ~ 1 cm Itzhak Tserruya RICH 2010, Cassis, May 2 -7, 2010 31

Hadron Blindness Hadron suppression illustrated by Hadron rejection factor comparing hadron spectra in FB and RB (same number of central tracks) Pulse height Ø Strong suppression of hadron signal at reverse drift field ØLarger rejection by combining pulse height and cluster size Itzhak Tserruya RICH 2010, Cassis, May 2 -7, 2010 32

Electron - hadron separation in RB q HBD in RB mode q Reconstructed tracks in Central Arms are projected to HBD q Single electrons selected by Dalitz open pairs (m< 150 Me. V/c 2) Hadrons d. E/dx signal in the small (~ 100 mm) region above the top GEM and in the first transfer gap very small wrt electron Cherenkov signal

Single electron detection efficiency 1. Single electron efficiency using a sample of open Dalitz decays ( V cut rejects conversions; not very effective for conversions in the radiator gas ) : 90 % 2. Single electron efficiency derived from the J/psi region ( m > 3. 5 Ge. V/c 2 after background subtraction): = 90. 6 9. 9 %

Single vs double electron separation Fully reconstructed 0 Dalitz pairs (m < 150 Me. V/c 2) in the central arms Matched to HBD into two separate clusters or one single cluster. Double electron response ( 0 Dalitz close pairs) Single electron response ( 0 Dalitz open pairs) ~ 22 p. e. per single electron track ~ 40 p. e. per two electron track Agrees with expected yield taking into account p. e. collection efficiency and transmission loss in the gas Itzhak Tserruya RICH 2010, Cassis, May 2 -7, 2010 35

Combinatorial Background rejection With the performance results shown in the previous slides we should be very close to the design values for the CB rejection. q **Preliminary ** rejection numbers: - matching to HBD - double hit cut - close hit cut - single pad cluster cut 7. 1 2. 2 6. 5 ~2

Summary Ø Ø Ø A novel HBD detector has been developed, installed and is successfully being operated in the PHENIX set-up Analysis of data taken in pp and Au+Au collisions show: v Clear separation between e and h v Expected Hadron rejection factor v Excellent electron detection efficiency v Good separation of single vs double electron hits v Considerable reduction of the combinatorial background Looking forward to the physics Itzhak Tserruya RICH 2010, Cassis, May 2 -7, 2010 37

HBD Team Brookhaven National Lab B. Azmoun, R. Pisani, T. Sakaguchi, A. Sickles, S. Stoll, C. Woody Columbia University (Nevis Labs) C-Y. Chi Stony Brook University Z. Citron, M. Connors, M. Durham, T. Hemmick, J. Kamin, B. Lewis, V. Pantuev, M. Prossl, J. Sun University of California Riverside A. Iordanova, S. Rolnick Weizmann Institute of Science Z. Fraenkel*, A. Kozlov, A. Milov, M. Naglis, Ravinovich, D. Sharma, I. Tserruya * deceased I.

Event display Itzhak Tserruya RICH 2010, Cassis, May 2 -7, 2010 39