1 HBD Run7 Report T.K. Hemmick for the HBD crew PHENIX Collaboration Meeting, Boulder, July 12, 2007.

Slides:



Advertisements
Similar presentations
Prototype Progress… The Stony Brook Crew & Bob Azmoun, Craig Woody.
Advertisements

HBD Meeting 4/25/06 SB and BNL Crew. 4/25/06 B. Azmoun 2 Installation of HBD into PHENIX Transported HBD from USB under gas (CF4) flow (maintained ~ 1Torr.
Benji Lewis APS April Meeting Hadron Blind Detector implementation during PHENIX Run-10 Benji Lewis On behalf of the PHENIX Collaboration.
1 HBD Update Itzhak Tserruya DC meeting, April 9, 2008 April 9, 2008.
Bulk Micromegas Our Micromegas detectors are fabricated using the Bulk technology The fabrication consists in the lamination of a steel woven mesh and.
Preliminary Quintuple GEM CsI Ring Imaging Cherenkov Detector S.A. Zajac*, T.K. Hemmick*, K. Dehmelt*, T.E. Videbaek*, & M.A. Blatnik** *Stony Brook University.
1 Stony Brook Update Getting Serious about HBD East T.K. Hemmick for the Tent Crew.
Slide 1 Diamonds in Flash Steve Schnetzer Rd42 Collaboration Meeting May 14.
01/07/08I. Ravinovich1 HBD Update Ilia Ravinovich DC Meeting, January 9, 2008.
HBD Operation in Run 9 Thomas K Hemmick for the HBD Crew.
The Transverse detector is made of an array of 256 scintillating fibers coupled to Avalanche PhotoDiodes (APD). The small size of the fibers (5X5mm) results.
1 HBD Construction status Itzhak Tserruya DC meeting, BNL, February, 2006 Plans for Run-6 Box construction GEM assembly and testing Status of urgent items.
HBD Meeting 3/14/06 B. Azmoun BNL Testing and Preparation of HBD Full Scale Prototype for Upcoming Engineering Run Puzzling Behavior of GEM’s To Do List.
9/18/2015I. Ravinovich1 HBD performance in Run-9 I. Ravinovich WIS.
1 HBD Commissioning (II) Itzhak Tserruya HBD group meeting November 28, 2006 Progress from October 3 to November 28, 2006.
I. Ravinovich Di-electron measurements with the Hadron Blind Detector in the PHENIX experiment at RHIC Ilia Ravinovich for the PHENIX Collaboration Weizmann.
Cube Measurements Tent Crew. Scintillation BNL 241 Am Semi- collimated  Spectralon Diffuse UV Reflector SBD  -Trigger Scint. Light Poisson.
Current Status of Hadron Analysis Introduction Hadron PID by PHENIX-TOF  Current status of charged hadron PID  CGL and track projection point on TOF.
Preliminary Results from a Trial Beam Test of the Small HBD Prototype at LEGS Bob Azmoun BNL HBD Working Group Meeting May 10, 2005.
1 Stony Brook Update: A bit more on Negative Ions T.K. Hemmick for the Tent Crew.
1 Stony Brook Update: Keep It Clean T.K. Hemmick for the Tent Crew.
1 HBD Commissioning Itzhak Tserruya DC meeting, BNL, April 11, 2007.
1 Thomas K Hemmick May 17 th, The Gas Electron Multiplier: GEM Thomas K. Hemmick Stony Brook University.
Itzhak Tserruya IEEE 2007, October 29, 2007, Hawaii Itzhak Tserruya Weizmann Institute of Science, Rehovot, Israel for the HBD group: for the HBD group:
1 HBD Update: DC Meeting June 2006 Thomas K Hemmick for the HBD Crew…
12/12/09 MPGD 神戸大学 1 Yusuke Komatsu A B. Azmoun B, C. Woody B, K. Ozawa A University of Tokyo A,Brook Haven National Lab. B.
1 HBD update Itzhak Tserruya DC Upgrades meeting, January 14, 2005 NIM paper II: Generic R&D ~ completed Full scale prototype construction Pending issues.
1 HBD Towards Run-9 Itzhak Tserruya H/L meeting, April 3, 2008 April 3, 2008.
The 21st International Conference on Ultrarelativistic nucleus-nucleus collisions, March 30 – April 4, Knoxville, TN Results from cosmics and First LHC.
1 Stony Brook Update T.K. Hemmick for the Tent Crew.
1 Sasha Milov DC meeting... March 7, 2007 HBD status Sasha Milov March 7, 2007.
1 TPC R&D proposal from the WIS group I. Ravinovich WIS I. Ravinovich12/15/2015.
M. Muniruzzaman University of California Riverside For PHENIX Collaboration Reconstruction of  Mesons in K + K - Channel for Au-Au Collisions at  s NN.
6-Aug-02Itzhak Tserruya PHENIX Upgrade mini-Workshop1 Boris Khachaturov, Alexander Kozlov, Ilia Ravinovich and Itzhak Tserruya Weizmann Institute, Israel.
HBD Status Craig Woody BNL DC Meeting October 3, 2007.
Itzhak Tserruya, BNL, May13, HBD R&D Update: Demonstration of Hadron Blindness A. Kozlov, I. Ravinovich, L. Shekhtman and I. Tserruya Weizmann Institute,
HBD Status Report C. Woody For the HBD Crew DC Meeting June 13, 2007.
HBD Status for Run 10 C.Woody For the HBD Group Collaboration Meeting January 15, 2010.
Future Possibilities for Measuring Low Mass Lepton Pairs in Christine Aidala for the Collaboration Quark Matter 2002, Nantes.
STAR J/  Trigger in dA Manuel Calderon for the Heavy-Flavor Group Trigger Workshop at BNL October 21, 2002.
Triple GEM Beam Test within PHENIX DC Upgrades Meeting 4/14/04 B. Azmoun PHENIX Brookhaven National Lab.
1 Sasha Milov DC meeting Sept 12, 2007 HBD status Sasha Milov Sept 12, 2007.
1 HBD Update Back to the Future T.K. Hemmick for the HBD crew.
Update on the HBD Craig Woody BNL DC Meeting May 11, 2005.
Collection of Photoelectrons from a CsI Photocathode in Triple GEM Detectors Craig Woody Brookhaven National Lab B.Azmoun 1, A Caccavano 1, Z.Citron 2,
STAR Collaboration Meeting, BNL – march 2003 Alexandre A. P. Suaide Wayne State University Slide 1 EMC Update Update on EMC –Hardware installed and current.
STAR Analysis Meeting, BNL – oct 2002 Alexandre A. P. Suaide Wayne State University Slide 1 EMC update Status of EMC analysis –Calibration –Transverse.
1 Stony Brook Update: Another Frustrating week... However difference is isolated: CF 4 T.K. Hemmick for the Tent Crew.
F Don Lincoln, Fermilab f Fermilab/Boeing Test Results for HiSTE-VI Don Lincoln Fermi National Accelerator Laboratory.
2/18/2016I. Ravinovich1 HBD Run-10 data analysis I. Ravinovich WIS.
Plans to Test HBD Prototype in Run 6 Craig Woody BNL DC Meeting March 8, 2006.
1 HBD R&D: Update Itzhak Tserruya (for A. Kozlov, I. Ravinovich and L. Shekhtman) Weizmann Institute, Rehovot DC meeting Feb. 14, 2003.
HBD Status Report from Run 9 Craig Woody BNL DC Meeting July 8, 2009.
HBD Report Craig Woody BNL DC Meeting January 7, 2009.
1 HBD Commissioning Itzhak Tserruya DC meeting, BNL December 13, 2006 Progress from October 3 to November 28, 2006.
Collection of Photoelectrons from a CsI Photocathode in Triple GEM Detectors C. Woody B.Azmuon 1, A Caccavano 1, Z.Citron 2, M.Durham 2, T.Hemmick 2, J.Kamin.
January 13, 2004A. Cherlin1 Preliminary results from the 2000 run of CERES on low-mass e + e - pair production in Pb-Au collisions at 158 A GeV A. Cherlin.
1 HBD Update Itzhak Tserruya DC meeting, May 7, 2008 May7, 2008.
Preparations to Install the HBD for Run 6 Craig Woody BNL PHENIX Weekly Meeting January 26, 2006.
Measurements of low mass e + e - pairs in p+p and Au+Au collisions with the HBD upgrade of the PHENIX detector Mihael Makek Weizmann Institute of Science.
Initial Performance of the PHENIX Hadron Blind Detector at RHIC Craig Woody Brookhaven National Lab For the PHENIX Collaboration N IEEE Nuclear.
P HOTON Y IELD DUE TO S CINTILLATION IN CF4 Bob Azmoun, Craig Woody ( BNL ) Nikolai Smirnov ( Yale University )
R&D on Hadron Blind detector, recent results Issues addressed: - gain limits in CF 4 with heavily ionizing particles - operation.
PHENIX J/  Measurements at  s = 200A GeV Wei Xie UC. RiverSide For PHENIX Collaboration.
6/29/2016I. Ravinovich1 HBD status in Run-10 I. Ravinovich WIS.
Quark Matter 2002, July 18-24, Nantes, France Dimuon Production from Au-Au Collisions at Ming Xiong Liu Los Alamos National Laboratory (for the PHENIX.
Ilia Ravinovich for the PHENIX Collaboration
Saikat Biswas, A. Abuhoza, U. Frankenfeld, C. Garabatos,
Charles F. Maguire Vanderbilt University
Presentation transcript:

1 HBD Run7 Report T.K. Hemmick for the HBD crew PHENIX Collaboration Meeting, Boulder, July 12, 2007

2 The 3 main properties of the HBD 1. As expected 2. OK. 3. Encouraging. Needs more work. 2. Electron detection efficiency1. Electron - hadron separation 3. N 0 or single/double electron separation

3 Outline  Detector operation  HV problem  Electronics  Photocathodes  Recirculation gas system  Gas transmission monitor system  Detector performance  Position resolution  Gain  e-h separation  Electron detection efficiency  Single vs. double electron separation  New GEM production

4 Operation

5 HV problem  Main cause of massive trips: Normal GEM spark would cause Lecroy HV to trip Mesh also trips leaving large stored energy on filter capacitor As GEM voltage goes to zero, large  V develops across gap between mesh and top GEM, ultimately resulting in a large spark Light from the mesh  GEM spark induces sparks in other GEMs leading to a massive trip.  Visual inspection of HBD West confirmed this picture (see next slide)  The massive trips did a lot of irreversible damage, destabilizing the entire detector and requiring reconditioning.  The problem was exacerbated by: The over-current protection that transformed genuine sparks into massive trips The attempts to reach 3800V (which corresponds to a gain of 10 4 according to our lab measurements but was way too high for the actual gain at the IR). The operation of modules with large negative dI.  Work in progress: What causes the non-linear V vs. I behavior of damaged GEMs? Can it be cured by ultra-sound cleaning or by air gun? It is interesting that the non-linear behavior in one GEM disappeared when dismounting it from the vessel for inspection.

6 Visual Inspection “Pock Marks”: Four patches (one per quadrant). These match locations of GEM-mesh sparks. These DO draw the non-linear currents.

7 Mesh  GEM sparks solutions Good SolutionBetter Solution  During the run: Install resistors in the mesh chain to provide a path to ground for the stored charge to go in case of trip.  Other (better?) solutions under consideration:  install Zener diodes to limit the voltage between the mesh and the top GEM to less than 300V.  Use a CAEN PS, as used during the entire R&D period at the WI, that provides a path to ground whenever a trip condition occurs.

8 After the fix  The trip frequency dropped dramatically after adding the resistors to the mesh chain.  The frequency increases during week 10 and 12 are correlated to bad weather. We definitely need to adjust the operating conditions to changes in atmospheric pressure. Trip frequency per week and per module  ES3 was a perfect module (10,10, 10) till its top GEM got damaged by a massive trip.  The damaged top GEM remained quite stable during the rest of the run  All other operating modules had at least one damaged GEM. Most of them remained stable till the end of the run.

9 Readout, noise and event size nThe entire readout chain worked reliably and smoothly during the entire duration of the run. nElectronic noise was excellent in all modules. Typically  = 1.5 ADC counts corresponding to 0.15 fC or 0.2 p.e. at a gain of nDuring the run, 2 modules (WN1 and WS3) developed an anomalous high noise at HV greater than ~3000 V nSimple zero suppression was implemented early in the run: sample(8) – sample(0) > n.  n However, event size remained large due to the scintillation light that generates genuine signals. n A new readout scheme was developed and implemented around mid-May that reduced the event size by a factor of 2 without any loss of information. Normal module One of the two noisy modules

10 Photocathodes  Relative QE scan of three top GEMs removed from HBD West.  They show full QE on all three modules (like brand new!!) and no evidence of any QE change, in spite of the relatively high levels of water in the detector for several months.  The flash lamp monitoring at the IR showed these three GEMs as having 100%, 25% and 150% of the expected QE, respectively.

11 Recirculation gas system  The recirculation gas system worked very well.  The H 2 O degassing/diffusion rate was too high. To reduce absorption it was necessary to increase the flow rate through the detector. With small changes it was possible to increase the gas flow by a factor of 2 above the design value.  With a flow of 5 lpm, the absorption is estimated to be 10-15%.  Modifications of the gas system are planned to allow increase of the gas flow by at least another factor of 2 in run8.  For several months we had a large disagreement between the the H 2 O sensor of the Panametrics monitor (giving readings of about 10 ppm ) and the transmission measurements (corresponding to about 80 ppm of water). Explanation: the Panametrics sensor is damaged by CF4, leading to too low ppm values of water.  A new sensor, installed toward the end of the run, showed also signs of getting damaged (a decrease of the ppm reading not corroborated by the transmission monitor).  Panametrics is studying the case.  A Kahn hygrometer, supposedly more robust with aggressive gases like CF4, has been ordered.

12 Gas transmission monitor system Typical transmission measurement at a flow of 5 lpm Gas in W in bypassHBD-E out

13 Gas transmission summary With a flow of 5 lpm we still loose some 10-15% of the p.e.. We plan to further increase the gas flow in run8 to reduce the losses.

14 Detector Performance (So far looked at “peripheral” events with up to 50 tracks in the central arms)

15 Tracking & position resolution Run ES4 at 3600V FB Position resolution:  z ≈   ≈ 1 cm Dictated by pad size: hexagon a = 1.55 cm (2a/√12 = 0.9 cm) Hadrons selected in central arm: Vertex +/- 20 cm < 50 tracks 3  matching to PC3 and EMCal n0 < 0 EMC energy < 0.5 Projected onto HBD:  Z in HBD +/- 2 cm  in HBD +/- 25 mrad

16 Hadron Blindness & e-h separation Electron - hadron separation (RB)Hadron rejection factor Hadron suppression illustrated by comparing hadron spectra in FB and RB (same number of central tracks) Results very similar to those obtained in previous lab or beam tests

17 Electron detection efficiency  Identify e in central arm using RICH and EMCal  Project central arm track to HBD  Relative e-detection efficiency in HBD obtained by varying the charge threshold of the closest (matched) pad EN3 G ≈ 3300 (several runs at “nominal” voltage) All modules ≈ 6600 (Run “nominal” voltage + 100V) ~4 p.e.  Start to loose efficiency at pad threshold larger than about 4p.e. Is it a real efficiency loss or are we loosing efficiency for electrons converted in the gas near the GEMs? Needs further study.

18 Electron analysis n RB runs # & with +100V, ~13.7M mb events, after +/- 20 cm vertex cut ~9.3M mb events n Used “peripheral” events with < 50 DC+PC1 tracks in both arms n High quality tracks 31 or 63 3  matching cuts to EMC n 0 > 2, disp < 5,    npe < 10 n 0.6 < E/p < 1.2 n n HBD analysis steps: 1. read all HBD pads in the event 2. apply gain equilibration 3. build the clusters with A > 10 ADC counts 4. project the electron track to HBD and look for a closest pad with A > 50 ADC counts 5. Apply the matching cuts for corresponding cluster in |  Z| < 2 cm and |  < 30 mrad

19 Single vs double electron separation  For a better comparison should restrict the East arm mass spectrum to the same acceptance as the HBD.  HBD response is a mixture of single, double electrons and a “continuum” (from pad threshold to double electron response) due to conversions in CF 4.  Hint of single vs. double hit separation?  Nr of p.e. (18?) is on the low side but not too far from the theoretical limit (36 p.e.) considering the known losses:  pad threshold ~10%  reverse bias ~15-20%  gas transmission ~15%  Need at least a factor of 10 more statistics to start looking at rejection. Data production is underway. e + e - pair spectrum in the central East arm Inclusive Matched to HBD-E Pulse height of matched hits in HBD-E 18 p.e.

20 Various means… nExpected separation between single and double electron hits for various n pe values: uincludes Poisson statistics with mean = n pe ueach photo-electron has been smeared with an exponential of mean one N pe =

21 Revisiting the single p.e. detection efficiency Is electron attachment in CF 4 affecting the single p.e. detection efficiency? Width of 55 Fe peaks in CF 4 (  /mean = 15%) are consistent with statistics of 109 e’s convoluted with exponential gain of each e. Electron scattering cross-sections in Ar and CF 4 : elastic momentum transfer (σ m ), vibrational excitation (σ ν4, σ ν3, σ νind ), electron attachment (σ a ), dissociation (σ d ), excitation (σ exc ), and ionization (σ ion ).  Detection efficiency reaches saturation at V GEM1 > 480V and E trans > 2.7 kV/cm.  The absolute detection efficiency in CH 4 was measured to be ~1 at a gain of 20.  Do the similar rates measured with CF 4 provide a proof of full detection efficiency also in CF 4 ?

22 Gain

23 Identification of scintillation hits: Single pad hits (not belonging to tracks)  Exploit the scintillation light to derive the gain of the detector In FB mode we see two components:  a fast exponential distribution due to scintillation  a slow exponential distribution due to charged particles As expected, in RB mode:  the scintillation signal survives entirely  the charged particle signal is strongly suppressed Zoom

24 Gain derived from the scintillation (I) Scintillation hit identification: Use peripheral events defined by the number of central tracks Use single pad hits which do not belong to any track Gain determination: Fit the range (10-50) ADC channels with an exponential function 1/slope increase with event multiplicity Gain = 1/slope (where = avrg nr of scintillation photons in a fired pad) Assuming the nr of scintillation photons per pad follows a Poisson distribution: A fired pad measures: = P(0) = probability to have no hit in a pad = =

25 Gain derived from the scintillation (II)  P(0) is not measured  Determine the probability P(0,th) of not firing a pad for a given threshold and extrapolate to zero threshold P(0,th) = 1 – [nr of fired pads (A>th)] / [total nr of pads]* * The large pads are excluded in this analysis i.e. the total nr of pads = 93 or 94.

26 It does work for all modules… Run HV: nominal … and different HV setting = 5.9 = 5.7 Run HV: nominal +100 = 9.5 = 10.9 = 5.3

27 Absolute gain  ADC calibration used here: 10 ADC ch = 1 fC  The gains are still high compared to NIM paper :  EN3: HV= 3580V  V GEM = 481 V Gain [NIM]* = HV= 3480V  V GEM = 468 V Gain [NIM]* = 660 *G = exp[0.066  V – ] dI [%]

28 Gain derived from mip signal (I) Procedure:  Runs (“nominal” voltage) and run (“nominal” + 100V)  Events with less than 50 tracks in the E arm  Select hadrons tracks in the E arm Vertex +/- 20 cm < 50 tracks matching to PC3 and EMCal n0 < 0 EMC energy < 0.5  Project & match to HBD   = ±0.03 rad  z = ± 2 cm  Subtract random matching  Fit pulse height with Landau distribution (in the range ADC > 30 and 75 for the two runs).  Use the mean of the fit to define the gain

29 Gain derived from mip signal (II) Red: all cluster sizes Blue: single pad cluster Filled symbols: high gain run Open symbols: low gain run

30  mip absolute gain derived assuming a primary charge of 18e in a 1.5 mm drift gap (as measured in KEK).  The gain derived from scintillation is very reliable.  The gain values derived from scintillation and mip signal are consistent within 30%  but the gains measured at the IR are a factor of ~4 higher than measured in the lab.  The origin of this puzzle is not understood.  Long term effect of gain saturation in the GEMs? Comparison of gain derived from mip and scintillation signal

31 More on scintillation Shades in PISA: height 2.5cm, spacing 2.5cm x 2.5cm  The scintillation hits provide a very convenient and reliable mean to determine the detector gain.  However, it seems that we see much more scintillation hits than expected, primarily due to tracks crossing the detector and not originating from the vertex (not yet fully quantified).  As a consequence, the occupancy in central Au-Au collisions is high and affects the pattern recognition (clusters are huge).  Double approach:  Install shades to reduce the scintillation light. MC studies underway to optimize the geometry and quantify the reduction of scintillation light seen by the CsI photocathode as well as the loss of Cherenkov light.  develop a new cluster algorithm to cope with the scintillation background.

32 New GEM production Towards run8

33 Status of GEM production at WIS n 25 Au plated and 30 standard GEMs were ordered from CERN n A first batch of 20 GEMs (10 standard + 10 Au plated) was received at WI on May 15. A second and final shipment completing the order was received at the end of June. n Production proceeds at a nominal rate of 1 GEM per day and an average of 5 GEMs per week. n Priority is given to the production of Au plated GEMs. n The production includes the following quality tests (same tests as last year): V test in air (clean room) of the foil before framing V test after framing V test after soldering SMD resistors V test in CF 4 in the lab 5. Gain mapping in Ar/CO 2 at 484V n So far 18 GEMs have been fully tested and are in the database. n A box with 24 GEMs -15 standard (12 from last year spares + 3 from this year) and 9 Au plated (3 + 6)- has already been shipped to SB. n A second box should be ready by the end of July.

34 Work at BNL n A new HV board under construction for test at BNL n Increase gas flow of recirculation gas system  Remeasure the scintillation yield in CF 4

35 Cleaning the glovebox and the HBD vessel is top priority before production starts  Laminar table (ONLY place where GEMs are handled in air) is better than class 1.  The tent is quite clean w/ typically class  Glovebox should be better than class 100 as it was before. This requires: u Opening it; Clean everything; Close it. u New HEPA filters? (probably a good idea) u Test empty u Test Full u Added additional filtering…tested yesterday at Class 20 and falling. n HBD vessel should be clean at least at the same level. The preferred solution: u empty it and keep the GEMs in HVAC to preserve QE. u Allows interior of HBD to be cleaned as well. u Still allows later testing with old GEMs. u Allows tests of things like Ultrasound cleaning of Cu GEMs that draw non-linear currents. u Allows test fits of Blinds

36 Production at SB  Sequence of operations and tests foreseen during the production at SB:  Full HV operation in CF 4 of each triple stack inside the test box before assembly into the HBD vessel.  Assembly inside HBD W vessel  Close HBD W vessel once a few modules (3 to 4) are installed and perform full HV operation in CF 4. This is a time consuming step. Requires closing the HBD, flushing gas for several days before doing the test and reopening the vessel to complete the module assembly. But it is considered to be a crucial test.  Complete module installation inside the vessel.  Close HBD W, take out of glove box, flush with CF 4 and perform a full detector HV test before shipping to BNL.  Repeat same sequence on HBD E.  Production ready to start

37 Things are ready at Stony Brook Evaporator working. Test shot done recently. QE measurement system working LEGS gas system up and running in recirculation mode with increase flow and safety system Many new students arriving: Thomas Videbaek, Leila Makdisi, Cristina Chu, …