LXe Beam Test Result CEX beam test 2004
LXe Beam Test Result CEX beam test 2004 Cryogenic Equipment Preparation Status Liquid Xenon Photon Detector Group 1/41
Charge Exchange Beam Test at pi. E 5 l New PMTs R 9288 TB higher QE and better performance under high BG ¡ Resolutions to be improved ¡ l l l New calibration alpha sources New refrigerator with higher cooling power TEST at pi. E 5 beam line ¡ l Analysis framework ¡ l Gain experience ROME in online (offline also) analyses Waveform data obtained with DRS prototype boards 2
PMT Development Summary 1 st generation R 6041 Q 2 nd generation R 9288 TB 3 rd generation R 9288 ZA 228 in the LP (2003 CEX and TERAS) 127 in the LP (2004 CEX) 111 In the LP (2004 CEX) Not used yet in the LP Rb-Sc-Sb Mn layer to keep surface resistance at low temp. K-Sc-Sb Al strip to fit with the dynode pattern to keep surface resistance at low temp. K-Sc-Sb Al strip density is doubled. 4% loss of the effective area. 1 st compact version QE~4 -6% Under high rate background, PMT output reduced by 10 -20% with a time constant of order of 10 min. Higher QE ~12 -14% Good performance in high rate BG Still slight reduction of output in very high BG Higher QE~12 -14% Much better performance in very high BG 3
Alpha sources on wires 4 tungsten wires plated with Au (50 micron f) l Po attached on the wires, 2 active points per wire l ¡ ¡ ~40 Bq per point on 2 wires at the rear side ~130 Bq per point on 2 wires at the front side Active points are coated with Au (200 -400Å) l Fixed on the wall with spring. l Alpha sources on the walls were removed l wire LED gamma 4
New Refrigerator (PC 150 W) MEG 1 st spin-off l Technology transferred to a manufacturer, Iwatani Co. Ltd l Performance obtained at Iwatani l ¡ ¡ ¡ 189 W @165 K 6. 7 k. W compressor 4 Hz operation 5
CEX Elementary process p-p p 0 n p 0(28 Me. V/c) g g 54. 9 Me. V < E(g) < 82. 9 Me. V • FWHM = 1. 3 Me. V q Eg Eg p 0 Eg • Requiring q > 175 o FWHM = 0. 3 Me. V 54. 9 Me. V 170 o Requiring q>170 o 82. 9 Me. V 1. 3 Me. V for q>170 o 0. 3 Me. V for q>175 o q Eg 6
Beam Test Setup H 2 target+degrader LYSO Eff ~14% LP Eff(S 1 x. LP)~88% Na. I S 1 beam 7
Beam Condition l Profile at the target (with a pill counter) ¡ ¡ l Vertical 13. 2 mm Horizontal 9. 9 mm Pion rates (w/o separator) 1. 8 m. A and 4 cm Target E. ¡ ¡ Slits 80: 2. 07 x 108 п -/sec Slits 100: 3. 95 x 108 п - /sec Optimization of degrader thickness 20 mm + 3. 3 mm x n 8 Profile at S 1, 2 mm/bin
Operation Status l l Thanks to a new refrigerator we succeeded to operate the detector (almost) without using LN 2 except for power break and recovery. New pressure reducer also helped this while precooling and liquefaction. Circulation/purification continued during DAQ. History September ¡ 18~21 Pre-cooling (72 hrs) • 21~24 Liquefaction (79 hrs) • 24 Circulation start (~30 cc/min) • 24 Electronics setup October l DAQ started l 25 DRS boards installed l 29 Recovery of xenon • ¡ 9
Data set Gain ADC gate Beam intensity event# * Low - middle 32 + 29** k high - low - middle 48 k high - low 55 k middle 110 + 44** k high - low 77 k middle 85 k high 47 k 400 nsec High 600 nsec 400 nsec Normal 600 nsec l And Waveform data… 10
Analysis Result Calibration Energy Timing 1 st look on waveform Data 11/41
Alpha data l l One of the rear wires found to be slipped Weighted position average surround wires due to shadow effect. Reconstructed Position is far from wires Wire (50 μm ϕ) Alpha 40 μm Po half-life=138 days 12
Source Position Reconstruction l. The two wires on the front face are a little displaced LXe GXe 13
Alpha data analysis Nphe[0] for top-left alpha with alpha emission angle selection Center of the PMT-0 14
LXe/MC, absorption length evaluation 4 front sources Applying the QEs determined in GXe (-75˚C) 15
Q. E. evaluation with alpha events in liquid Q. E. evaluation using alpha data in the liquid is also possible. Higher light yield Expected better evaluation if xenon is pure! R 9288 R 6041 Data #8528 normal gain front 4 alphas MC reflection on quartz on no absorption scattering length : 45 cm for 175 nm 16
Energy Reconstruction Cut-based Qsum Analysis Linear Fit Analysis 17/41
Cut-based Qsum analysis Event Selection l Cut-based Qsum analysis Analyze only central events to compare with the previous result ¡ ¡ ¡ |Xrec|, |Yrec|<2 cm 70 Me. V < ENa. I+ELYSO < 105 Me. V Sigma 2 > 40 (discard events if shallow) l Sigma 2: broadness of the event measured by using front face PMTs depth parameter Exenon[nph] 83 Me. V to Xe 55 Me. V to Xe MC 18
Cut-based Qsum analysis Correction and selection efficiency 83 Me. V 55 Me. V Before depth correction 55 k 8129 15026 1750 3018 1362 # of events no cut 55 Me. V selection with the other gamma position selection depth selection After depth correction with a linear function 260 k # of events In 55 Me. V peak 15 k 78 % 19
Cut-based Qsum analysis Energy Resolution CEX 2003 s=1. 53% 55 Me. V CEX 2004 FWHM = 4. 5 ± 0. 3 = 1. 23 ± 0. 09 % FWHM=4. 8 % 83 Me. V s=1. 16 ± 0. 06% FWHM = 5. 0 ± 0. 6 σ = 1. 00± 0. 08 % FWHM=5. 2% 20
Linear Fit analysis 55 Me. V event selection l In general it is possible to obtain higher efficiency with the linear fit analysis Y (cm) Correlation with Na. I/Lyso 83 Me. V in LXe 55 Me. V in LXe X (cm) Small displacement (~ 0. 5 cm) 21
Linear Fit analysis Energy (Linear Fit) and Qsum reconstruction No selection, 600 k events Na. I+sat cut, 83 k events Na. I cut, 144 k events Black: Linear Fit Red: QSUM Na. I+sat+coll cut, 54 k events Linear Fit trained using MC including Fresnel reflection; used Q. E. determined with six sources. No large differences changing Q. E. set. The Linear Fit works better. Na. I cut: 70 Me. V<QNAI<100 Me. V Coll. cut: (X 2 + Y 2)1/2 < 4. 75 cm 22
Energy vs. Depth Correction along X & Y Linear Fit analysis E (Me. V)) E (Me. V)) Red: all events; Green: no saturated No Need Anymore We observed a slight position dependence of the reconstructed Energy. Z (cm) Remove ADC saturated events is equivalent to a depth cut. It can be corrected by using a parabolic interpolation. 23
Linear Fit analysis Reconstructed Energy (updated) 55 Me. V 83 Me. V Saturation & Na. I cut + R<1. 5 cm FWHM = 5. 6 % FWHM = 4. 8 % Correction (X&Y) effect 0. 3 % 24
Position dependence of energy resolution 25
Timing Analysis Intrinsic, L-R analysis Absolute, Xe-LYSO 26/41
The algorithm l l p- T = TDC - Tref TDC correction for time-walk and position g S 1 Na. I l l And correction for position TL, TR by weighted average of Ti g LP LYSO t. LP - t. LYSO i=r. m. s. of Ti cut on Qi> 50 pe TL Left Right l <T> = (TL TR)/2 g TR 27
L-R analysis Intrinsic resolution, L-R analysis • Position and Tref corrections applied • Applied cuts: Old data New data • |x|< 5 cm, |y|<5 cm • ELYSO+ENa. I >20 Me. V • RF bunch and TDC sat. • Study of vs Npe • = 65 ps @ 35000 pe • = 39 ps @100000 pe • QE still to be applied 28
Xe- LYSO analysis Absolute resolution, Time reference (LYSO) (TLYSO(R) -TLYSO(L))/2 PMT 1 PMT 2 =64 psec • LYSO PMT 1 & 2 • Coorected for x-coord. (not for y) LYSO Corrections applied for time walk (negligible at high energy deposit) slit • slit gamma with 1 cm slit 29
Xe- LYSO analysis Absolute timing, Xe-LYSO analysis 55 Me. V normal gain 110 psec high gain 103 psec High Normal gain LYSO Beam L-R depth reso. 110 64 61 = 65 = 56 33 psec 103 64 61 = 53 = 43 31 psec 30
1 st look on the waveform data 31/41
DRS Setup l LP Front Face DRS 0 DRS 1 • DRS inputs • LP: central 12 PMTs • LYSO: 2 anode signals for each DRS chip as time reference Two DRS chips were available. ¡ ¡ ¡ 10 ch/chip (8 for data and 2 for calibration) in total 16 for data 2. 5 GHz sampling (400 ps/sample) 1024 sampling cells Readout 40 MHz 12 bit Free running domino wave stopped by trigger from LP • DRS chip calibration • Spike structure left even after calibration, which will be fixed by reprogramming FPGA on the board. Xe(g) 32
Simple Waveform Fitting l Simple function with exponential rise and decay can be nicely fitted to the xenon waveform. (and also LYSO waveform) l Other Fitting functions ¡ Gaussian tail l ¡ CR-RCn shaping l ¡ V(t)=A(exp(-((t-t 0)/τrise)2)exp(-((t-t 0)/τdecay)2)) V(t)=A((t-t 0)/τdecay)n exp((t-t 0)/τdecay) Averaged waveform l template Xenon τrise=7. 0 nsec τdecay=35 nsec 33
a/g separation & LYSO timing l Alpha events are clearly discriminated from gamma events. ¡ LYSO time resolution is similar to that obtained with TDC. Pulse shape discrimination Time constant l This does not highly depend on the fitting procedure. LYSO time resolution γ α 34 Pulse height [m. V]
Averaged Waveform l An averaged waveform can be used ¡ ¡ ¡ for fitting as a template for simulating pileup for testing analysis algorithm etc. Average The measured waveforms are averaged after synchronizing them with T 0 Use the “template” for fitting! l l Pulse shape seems to be fairly constant for the gamma event. -40 m. V -160 m. V -1200 m. V 35
Simulation of Pileup Events Overlapping pulses are simulated using averaged waveform to test rejection algorithm. l Real baseline data obtained by the DRSs is used. l Npe 1=2000 phe Npe 2=1000 phe (3000 phe is typical for 50 Me. V gamma) ΔT=-30 nsec ΔT=+60 nsec 36
Trial of Pileup Rejection It seems easy to break up overlapping pulses >10 ns apart from each other. l Rejection power is being investigated for different sets of (Npe 1, Npe 2) and ΔT. l Original Npe 1=2000 phe Npe 2=1000 phe ΔT=-10 nsec Differential ΔT=-15 nsec easy ΔT=-5 nsec ? Difficult but not impossible ΔT=+15 nsec easy 37
Cryogenic Equipment Preparation Status 38/41
PC 150 W performance at Iwatani at PSI New PT(190 W) and KEK original (65 W) l Condition: ¡ ¡ 6. 7 k. W(60 Hz) 4 Hz Twater=20 C (Iwatani 2003. 12) 6. 0 k. W(50 Hz) 4 Hz Twater>30 C (PSI 2004. 7) Calorimeter operation without LN 2 at PSI(Sep. to Oct. 2004) 42 -day operation without degradation in cooling performance 39
Current status/schedule of liquid-phase purification test l l l 17/Jan wire installation & closing the cryostat 24/Jan setup in Pi. E 5 -13/Feb evacuation 7 -20/Feb liq. N 2 piping 14/Feb-13/Mar liquefaction and test 14/Mar recovery Purifier cartridge Liquid pump • New calibration wires with higher intensity • 9 Me. V gamma from Nickel LP top flange 40 xenon
End of Slide 41/41
The algorithm TDC correction for time-walk and position (point-like approx) l vertex reco. by weighted average of PMTs (new QE set, see Fabrizio Cei’s talk) l TL, TR by weighted average of Ti i=r. m. s. of Ti cut on Qi> 50 pe l <T> = (TL TR)/2 42
The algorithm T 9 F 20 = (290 5) ps = (345 5) ps Side PMTs are less sensitive to z-fluctuations than Front PMTs 43
TLXe - TLYSO Global non-linear corrections for g-vertex ( 50 ps) mainly due to: • scale compression (operated by PMT average) • finite shower size 44
Beam spot on target Beam profile • H = 13. 2 mm • V = 9. 9 mm (as measured by Peter) p = 62. 3 ps 45
- Slides: 45