HAPD Status. S. Nishida KEK. Dec 11, st Open Meeting of the SuperKEKB collaboration. HAPD Status. 1st SuperKEKB Meeting 1

Size: px

Start display at page:

Download "HAPD Status. S. Nishida KEK. Dec 11, st Open Meeting of the SuperKEKB collaboration. HAPD Status. 1st SuperKEKB Meeting 1"

Holly Armstrong
5 years ago
Views:

1 S. Nishida KEK 1st Open Meeting of the SuperKEKB collaboration Dec 11,

2 Contents 144ch HAPD Key Issues Summary I. Adachia, R. Dolenecb, K. Harac, T. Iijimac, H. Ikedad, Y. Ishiie, H. Kawaie, S. Korparb,f, Y. Kozakaic, P. Križanb,g, T.Kumitai, E. Kurodai, Y. Miyazawah, S. Nishidaa, I. Nishizawai, S. Ogawaj, R. Pestotnikb, N.Sawafujij, S. Shiizukac, T. Sumiyoshih, M. Tabatae, Y.Uekih akek, bjožef Stefan Institute, cnagoya Univ., djaxa, echiba Univ., funiv. of Maribor, guniv. of Ljubljana, htokyo Univ. of Science, itokyo Metropolitan Univ., jtoho Univ. 2

.. ACC (Aerogel Cherenkov Counter) is used in Belle.

3 Introduction PID has been a key issue in B factory experiments. The situation is the same in future B factory: b d v.s. b s, flavour tagging... ACC (Aerogel Cherenkov Counter) is used in Belle. Threshold type Separation only up to ~ 2GeV in the endcap region As an upgrade of the endcap ACC, we are developing Proximity Focusing Aerogel RICH. ACC Aerogel RICH photodetector aerogel radiator upgrade 3

144ch HAPD 144ch HAPD (Hybrid Avalanche Photo Detector) bi alkali photocathode photon HV 8kV 300V bias bombardment gain ~ O(1000) Total gain ~ 104 105 Pixel APD

4 144ch HAPD 144ch HAPD (Hybrid Avalanche Photo Detector) bi alkali photocathode photon HV 8kV 300V bias bombardment gain ~ O(1000) Total gain ~ Pixel APD avalanche gain ~ O(10) Developed with Hamamatsu Photonics. 144 channels, 4 APD chips (36ch/chip). Effective area 64mm 64mm (65%). Quantum efficiency ~ 22 27% at peak. 73 4

5 Beam Test Aerogel SC MWPC Beam Fuji test beam line Prototype Aerogel RICH with 2 3 array of 144 ch HAPD Readout using 48 ASICs. Clear Cherenkov ring is observed. HAPD ASIC MWPC SC e 200mm Black box 5

Beam Test Focusing aerogel Same index detector Aerogel

resolution gets worse Make overlapped images from each

7 4 layer 3 layer 1 layer 2 layer Single photon resolution

2mrad *focusing *same index Thickness[mm] Sigma[mrad] Npe

6 Beam Test Focusing aerogel Same index detector Aerogel detector Aerogel Thicker radiator gives higher Npe, but resolution gets worse Make overlapped images from each aerogel layer ( focusing ) Npe:5.7 4 layer 3 layer 1 layer 2 layer Single photon resolution σ photon : 13.2mrad *focusing *same index Thickness[mm] Sigma[mrad] Npe Sigma[mrad] Npe Index: n1<n2<n3<n4 Angle resolution per track σ track :5.6mrad Thickness[mm] Thickness[mm] 4σ K/π separation at 4.0 GeV/c 6

7 Key Issues Performance (Number of photons). Tolerance to the Magnetic Field. Readout (front end electronics) Structure/Layout. Long Term Stability. Radiation Tolerance. Delivery (and Budget). 7

Readout Electronics 4 trial productions of prototype ASICs (S01 S04) at VDEC. analog Preamp Shaper VGA digital Comparator Shift Register Successfully readout 1 p.e. signal from HAPDs.

8 Readout Electronics 4 trial productions of prototype ASICs (S01 S04) at VDEC. analog Preamp Shaper VGA digital Comparator Shift Register Successfully readout 1 p.e. signal from HAPDs. Now, we are developping new prototype ASIC (SA01,SA02). Production at MOSIS (TSMC 0.35 m process) Digital part for readout is provided with external FPGA for more flexibility to Super Belle DAQ More channels per chips. Preamp Shaper Comparator 8

Readout Electronic Offset Adjustment coarse SA01 result at the test bench Linearity 200mV fine 63000 p.e. 20mV 1 p.

9 Readout Electronic Offset Adjustment coarse SA01 result at the test bench Linearity 200mV fine p.e. 20mV 1 p.e. signal is already saturated if we apply full HV and bias voltage to HAPD. Need modification in the next version. 9

10 Readout Electronics When we connect SA01 and HAPD, we see larger noise than expected. Probably the noise come from outside or the test board. Trying to reduce noise (possible to detect single photon even with present noise level). test board Plan LED HAPD Designing new ASIC (SA02) w/ 36 channels. New test board to read out 144ch with SA01 and FPGA. Design of compact electronics (boards) for SA02. Still many items to develop, but they are within our reach. Threshold scan 10

11 Test under Magnetic Field Test under Magnetic Field is performed. [I.Adachi, S.Korpar, S.Shiizuka] Confirm the operation under 1.5 T magnetic field. Reduction of the distortion of the electric field in the edge channels. Reduction of the effect of back scattered electrons. Photon Glass Photo cathode Metal Ceramic Metal photo electron APD back scattered electron 11

~ a few photons (Need study for single photon case) Hit distributions of each channel.

12 Test under Magnetic Field 1.5T Count Count 0T Light incidence position (mm) Light incidence position (mm) Light intensity ~ a few photons (Need study for single photon case) Hit distributions of each channel. Performance improvement is seen in the magnetic field. Strange behaviour at 0T in edge channels is fixed at 1.5T. Tail parts get smaller in the magnetic field test. 12

13 Test under Magnetic Field Channel at the edge APD ch. boundary Next channel 0T 1.5T 0T 1.5T Light incidence position (mm) Light incidence position (mm) Under the magnetic field, hit distribution is consistent with the sensitive region of APD (blue arrow). 13

Test under Magnetic Field 0T 40mm 1.5T Light incidence position (mm) Light incidence position (mm) Vary the position of the light spot, monitoring one channel.

14 Test under Magnetic Field 0T 40mm 1.5T Light incidence position (mm) Light incidence position (mm) Vary the position of the light spot, monitoring one channel. Hit exists even when light is far from channels. Calculation: maximum 0T, 1.5T for back scattering. Reduction of the effect of back scattered electrons is confirmed. Remaining contribution is probably due to light AD. 14

15 Lifetime Test Lifetime test (acceleration test) has been performed at HPK. Irradiate photons to HAPD with HV and bias on. Only one sample was tested so far. photo cathode current 200 pa, anode current 10 A (gain 50000) per ch hours. Total cathode (4 mc/cm2). 200 C/cm2. Guess from ACC: 30 mc/cm2/year. Corresponding to 6700 years. Output signal gradually decreased: ~73% after 1500 hours operation Q.E.: 21% 25%??? (maybe measurement error) Avalanche gain: almost no change Bombardment gain decreased by 25% Leakage current increased. Looks O.K. More samples will be tested with less acceleration. 15

test (equivalent to 2 1013 /cm2 neutrons = LHC 10 years) showed no change on gain and QE, but

16 Radiation Tolerance Another key issue for long term stability is radiation tolerance (neutrons) Measurement of ECL group, neutron flux is ~ 1011 /cm2/year According to CMS, proton irradiation test (equivalent to /cm2 neutrons = LHC 10 years) showed no change on gain and QE, but increased leakage current by factor of 400. We have installed 2 2 APD in the endcap region of Belle. 16

17 Radiation Tolerance A Accelerator No serious effects are seen so far. Apparently not enough to guarantee radiation tolerance at Super KEKB. Neutrol irradiation test in future. 17

18 Performance Improvement Number of photo electrons (p.e.): presently 5.7 p.e. for 3cm aerogel. It corresponds to 4.0 K/ separation in a simple calculation. But, in reality, more backgrounds etc. are expected. Need to increase the detected number of p.e. HAPD with better Q.E.? 18

19 Ultra/Super bi alkali PC Photo cathode of present HAPDs is bi alkali, but there exist other photo cathode with better QE. QE Peak QE Bi alkali 25% Super Bi alkali 37% Ultra Bi alkali (nm) 700 HPK introduced a new machine which enables the trial for super/ultra bi alkali photocathode. 19

kinds of aerogel thickness: 20mm, 30mm, 40mm Number of

20 Ultra/Super bi alkali PC Photons reaching the photo cathode Detected photons w/ Ultra bi alkali With three kinds of aerogel thickness: 20mm, 30mm, 40mm Number of detected photons (270nm 650nm): 20mm : 30mm : 40mm = 1 : 1.29 :

21 Ultra/Super bi alkali PC Number of detected photons assuming 30mm aerogel Bi alkali Super Bi alkali Ultra Bi alkali Number of detected photons (270nm 650nm): Bi alkali : Super bi alkali : Ultra bi alkali = 1 : 1.79 :

22 Layout 516 HAPDs 540 HAPDs One issue is the HV (How close can we arrange the HAPD?) According to HPK, the distance to the neighbouring HAPDs might not be a big problem. Need consideration for the distance with the grounded structure (Al) 22

23 Delivery Plan We now have 8 HAPD samples. HPK is now developing new prototype machine for mass production of HAPD. From next year, it will become possible to produce one HAPD per day. We expect ~20 more good samples in this JFY. With this new machine, HPK can develop HAPD with new photo cathodes. 2.5 years for the full production 2009/10 order machines for mass production 2011/ / / /08 parts order delivery ~50 ~300 ~400 23

24 Summary Performance (Number of photons) Not enough now, but improvement is expected with new photo cathodes. Tolerance to the Magnetic Field. O.K. ( single photon detection will be checked ) Readout (front end electronics) ASIC are in development. Structure/Layout. Design started. Need to be careful on HV. Long Term Stability. HPK test shows good results. Will be tested with more samples. Radiation Tolerance. Need to test (no bad news so far). Delivery (and Budget). 2.5 years for production 24

25 Backup 25

Introduction aerogel radiator photodetector

( ) (K) = 23 mrad for 4 GeV/c particle (in

confirmed by a test using 4 4 array of flat

26 Introduction aerogel radiator photodetector Target: 4 /K separation at 1<p<4 GeV/c. ( ) (K) = 23 mrad for 4 GeV/c particle (in case of aerogel radiator with n=1.05). Limited by the space. 200mm The scheme of Aerogel RICH has been confirmed by a test using 4 4 array of flat panel PMT (Hamamatsu H8500) m m However, H8500 cannot be used in the magnetic field. 26

27 HAPD Performance [Y.Miyazawa, N.Sawafuji, Y.Ueki] Pulse Height Distribution Noise level pedestal 1 p.e. Bias voltage dependence of Noise. Measured with HV OFF. Noise level is estimated from the width of pedestal. Full depletion at >50V. Leakage current increases at ~300V. No large chip dependence. Pulse Height Measured with HV ON. Light source: LED. Clear separation between pedestal and 1 p.e. peak!! 27

28 Readout with ASIC Readout system requires: High density front end electronics (100k channels) High gain with very low noise amplifiers Deadtime less readout scheme (Pipeline) We have developed ASICs for the front end electronics. analog Preamp History: 2003 (S01) 2004 (S02) 2005 (S03) 2006 (S04) Shaper 4.93[mm] digital VGA Comparator Shift register for HPD (smaller capacitance, smaller signal) for HAPD (larger capacitance, larger signal) Better larger capacitance smaller capacitance, more robust Test production is done 4 times. These versions are still for evaluation. 28

29 ASIC Specification Shaping time 0.3 ~ 2.0 [ s] Variable gain 1.25 ~ 20 Offset adjustment (±200 mv) Noise [e] [I,Nishizawa, Y.Miyazawa] Specification: Production at VDEC (Tokyo Univ) Process: ROHM CMOS 0.35 [ m] Target Noise Level: pF (HAPD) Std. Input Signal: e #(channel) = 18 [ch/chip] Readout: Pipeline with shift register measured noise level S/N = 10 80pF (HAPD) Power Consumption = 3 [mw/ch] ~ pf Capacitance [pf] 29

LED HAPD ASIC LED off varying ASIC threshold LED on 1 p.e. 2 (or more) p.

30 HAPD Readout with ASIC Distribution of output of ASIC (digital) for 1000 LED clocks. Irradiate 1p.e. level photon to 1 channel. LED HAPD ASIC LED off varying ASIC threshold LED on 1 p.e. 2 (or more) p.e LED off LED on Successfully readout 1 p.e. level signal from HAPD using ASIC for the first time!! 30

31 Test under Magnetic Field B U Setup (@ E Hall) Ushiwaka Black Box Ushiwaka Magnet G 10 magnetic field LED HAPD x y stage 31

32 Beam Test Setup Tracking using 2 MWPC 2 3 array of 144 ch HAPD HV at 7kV. Bias voltage of HAPD is chosen at avalanche gain = 40. (or maximum bias voltage). Readout using 48 ASICs. ASIC offset is adjusted so that the noise is below the threshold. threshold is typically ~ 0.5 p.e. level, but depends on each channel's gain and noise level. Standard aerogel (n = 1.045, thickness = 20mm, transmission length = and many other aerogels. 32

33 Cherenkov Ring Observed Result of quick analysis for a run with 2cm aerogel radiator preliminary N(p.e.) = 4.8 c = 12.4 mrad Clear Cherenkov ring is observed! 4.8 photo electrons per track. Consistent with the old result with Flat panel PMT (6.4 p.e.) if we consider smaller acceptance (84% 65%) and gap btw HAPD. Resolution 12.4 mrad per photon (was 13.6 mrad) Single track resolution 5.7 mrad corresponding to 4.0 K/ separtation. 33

34 Background Old result with flat panel PMT However, background shape is different beam induced hits, but why two peaks? additional structure halo? Beam tends to pass the upper HAPD; the additonal structure is considered to be related to the beam. 34

New ASIC In parallel, we are developing a new version of ASIC [H. Ikeda] Present ASIC is (for evaluation) not enough for readout of large amount of channels.

35 New ASIC In parallel, we are developing a new version of ASIC [H. Ikeda] Present ASIC is (for evaluation) not enough for readout of large amount of channels. Drop out the digital part (shift registers) from the ASIC, so that it becomes more flexible. Digital part will be provided by external FPGA. More channels per chip (18 36 or 72). New ASIC was designed and delivered. Test production at MOSIS. TSMC 0.35 m processes. Better S/N; two kind of comparator; shaping time adjustment by parameter; etc. With 12ch (with 3 3mm2 chip) this time, but with more channels (e.g. 3 8mm2) in future. Analog input Digital output 35

only 12ch out of 36ch (to HAPD) can be connected test pulse hitout (LVDS) monitor new ASIC Plan

36 New ASIC Now, testing Parameter setting O.K. Hit signal with test pulse is readout. Many measurement to be done. only 12ch out of 36ch (to HAPD) can be connected test pulse hitout (LVDS) monitor new ASIC Plan input output Performance measurement. (to VME board) Readout HAPD signal (several channels) Development of an interface board with 3 ASIC + FPGA. Design of new ASIC with more channels. 36

37 Threshold Scan Measure the efficiency of digital output varying the threshold. Estimate the pulse height and noise. (3) (2) pulse height noise (1) 37

38 Layout 38

39 Layout 39

R & D for Aerogel RICH

1 R & D for Aerogel RICH Ichiro Adachi KEK Proto-Collaboration Meeting March 20, 2008 2 1 st Cherenkov Image detected by 3 hybrid avalanche photon detectors from a beam test About 3:00 AM TODAY Clear image