1. Very-fast TOF Workshop, U of Chicago, Nov.18.2005.; Takayoshi Ohshima TIMING PROPERTIES OF MCP-PMT DEVICES -  <10 psec TOF Counter - T. Ohshima (Nagoya U.) 1. TOP counter and TOF counter 2. R&D of MCP-PMT’s 3. TOF counter ■ New Approaches ■ Beam test (1) ■ Beam test (2) ■ Know-how Footnote: MCP-PMT にもとずく 10 psec TOF counter R&D の報告である。

2. Very-fast TOF Workshop, U of Chicago, Nov.18.2005.; Takayoshi Ohshima  Hybrid Avalanche Photo-Diode S. Matsui et al, NIM A463 (2001) 220-226  Fine-mesh multi-anode PMT M. Hirose et al, NIM A460 (2001) 326-335  Linear array multi-anode PMT Y. Enari et al, NIM A547 (2005) 490-503  Micro-Channel Plate PMT M. Akatsu et al, NIM A528 (2004) 763-775 1.HPK10 (3809U-50-25X), 2.HPK6 (3809U-50-11X) 3.BINP (multialkali) (GaAs extended) 4. Burle (85001-501) Photon device for TOP counter M. Akatsu et al, NIM A440 (2000) 124-135; T. Ohshima ICFA Instr. Bull. 20 (2000) 2 < 10 ps TOF counter TOP counter の photon detector の開発が動機。要請する性能を満たすもの⇒ PMT, HAPD & MCP へ。この過程で MCP-PMT の光時間分解能を活用し TOF counetr を発想。 回路を除くと 5 ps の分解能をすでに得る。 By doing R&D on these issues, most of them are now in satisfaction. In the course of R&D studies, we come across an idea to have less 10 ps TOF counter. Quantum Efficiency Collection Efficiency photocathode material cathode – 1 st MCP gap vacuum Structure of MCP-PMT Cross-talk Multi-anode structure Ion-feedback layer Position resolution ~ 1 mm Single photon sensitive high detection efficiency Fast timing TTS < 50 ps Operational under 1.5 T Long life-time Rate dependence 1. TOP counter and TOF counter

3. Very-fast TOF Workshop, U of Chicago, Nov.18.2005.; Takayoshi Ohshima 2. R&D of MCP-PMT’s (Single photon pulses) (Transite Time Spread) TTS ● Multi-anode linear-array PMT (L16 & L24) 70-80; 120 ps ● Hybrid Avalanche Photo-Diode 150 ps ● Micro-Channel-Plate PMT 30-40 ps MCP(HPK10 3809U-50-25X) HAPD(HPK R7110U-07) L24(HPK R6135-L24X) L16(HPK R5900-L16) PMT(HPK H7195) 1 ns/div Footnote: これまで開発研究した光検出器の１光子に対する信号と測定 TTS 。信号の立ち上がりの速さを比較せよ。

4. Very-fast TOF Workshop, U of Chicago, Nov.18.2005.; Takayoshi Ohshima Fluctuations of 1.TTS 2.Decay-time (T d  TTS) 3.Light-path (T γ  TTS) 4.N γ Points for TOF counter Footnote: TOF 精度を決める要因。⇒ MCP-PMY/Cherenkov/path/# photons （時間広がりがなく、多量の光子が＝１．＆４．） Photo-statistics 1/  N γ is varied only at Td, T γ << TS. 1.30-40 ps (MCP-PMT), 70-80 ps (L16) 2.Cherenkov light 3.Normal incidence (a timing spread due to quartz thickness = 1-2 ps for 1 cm quartz.) ⇒  = (30x2–30) ps /1cm/(  12N γ ) = 9 ps/  N γ / 1 cm) 4. 50 detected photons/1 cm quartz For short path, no chromaticity effect.  = 30-40 ps/  50 = 5-6 ps quartz: n=1.47;  =45 o (for GeV/c particles) time photon signals 一 二 三 四 五 六 七

5. Very-fast TOF Workshop, U of Chicago, Nov.18.2005.; Takayoshi Ohshima Footnote: TOF 精度を決める要因。⇒回路系の精度（実測値＝ 7-9 psec ）。ビームテストでは att.&amp は不要。 矩形波で測定 測定回路の寄与＝ 8.8psec Divider HPK C5594: bandwidth=50 kHz-1.5 GHz gain=36dB (@0.1 GHz) NF = 5 dB Divider divider HUBER+SUHNER SMA cable MULTIFLEX MF 141 ： Impedance=50 ohm Operating frequency= 18 GHz Capacitance=95 pF/m Time delay= 4.7 ns/m Attenuation= a f(GHz)^1/2 + b f(GHz) (a=0.37320), (b=0.02790) 2. R&D of MCP-PMT’s (Test circuit) using single photons from a light-pulser 壱 弐 参

6. Very-fast TOF Workshop, U of Chicago, Nov.18.2005.; Takayoshi Ohshima NIM A528 (2004) 763-775, by M. Akatsu et al, “MCP-PMT timing property for single photons” MCP(Micro-Channel Plate) チャンネル径 Footnote: 開発研究の MCP-PMT 性能比較。 multialkali 2. R&D of MCP-PMT’s (MCP-PMT’s)

7. Very-fast TOF Workshop, U of Chicago, Nov.18.2005.; Takayoshi Ohshima Footnote: ADC & TDC spectra (pedestal = 100count)  =46ps HPK10 R3809U-50-25X BINP N4963 （１光子照射,HV:3.2kV ) Gain=10 6  =34ps (pedestal = 100count) Gain=3x10 6 Single photon peak 2. R&D of MCP-PMT’s (ADC spectra)

8. Very-fast TOF Workshop, U of Chicago, Nov.18.2005.; Takayoshi Ohshima 2. R&D of MCP-PMT’s (Gain vs TTS)

9. Very-fast TOF Workshop, U of Chicago, Nov.18.2005.; Takayoshi Ohshima 2. R&D of MCP-PMT’s (TTS vs B)

10. Very-fast TOF Workshop, U of Chicago, Nov.18.2005.; Takayoshi Ohshima TTS=46 ps, N  = 200/ 4 cm quartz,   = 46/  200 = 3 ps ⇒  expected = 9 ps including circuit fluctuation of 9 ps.  observed = 10.6 ps ⇒ With different TTS [ L16(TTS=80 ps) & MCP(TTS=46 ps) ] and similar N γ ’s,  observed = 11-12 ps is attained,where the circuit fluctuations (7-9 ps) dominate the ambiguity. 3. TOF counter (TOF by HPK10) 3. TOF counter (TOF by HPK10) Footnote: HPK10 TOF のビームテスト。期待値＝９ ｐｓ ｖｓ。測定値＝１０．６ ｐｓ。 Cerenkov radiator Since the light-pulser’s jitter yields an essential contribution on the measurement,

11. Very-fast TOF Workshop, U of Chicago, Nov.18.2005.; Takayoshi Ohshima 3. TOF counter (TOF － PMT w/o Radiator) NIM A547 (2005) 490, Y.Enari et al, Cross-Talk of a Multi-Anode PMT and Attainment of a s sim 10 ps TOF counter Footnote: HPK10 を単独でビーム照射。分解能＝１３．６ psec 。 By hitting an MCP-PMT directly by charged beam, TOF resolution of  = 13.6 ps was attained. HPK10(TTS=46 ps)

12. Very-fast TOF Workshop, U of Chicago, Nov.18.2005.; Takayoshi Ohshima 3. TOF counter (TOF-PMT w/o Radiator (continue)) ● window thickness = 4 mm ⇒ N γ expected = 25 photons vs. N γ detected = 50 photons Inspection:  0 = [ 13.6 2 – 9 2 ] 1/2 = 10 ps = 46 ps/  21 (photons) 21 photons vs. 25 / 50 photons ⇒ Timing of photons from the 1 st MCP plate is 100 ps earlier than those from photo-cathode, but its gain would be lower so that effective # of photons would be less than 25. ⇒ Yield of 25 photons is really from the MCP? Footnote: 実測値＝１３．６ psec 。 GaAs photo-cathode にすれば photon 数は２倍、分解能＝１０ psec が期待できる。 MCP-PMT 自体が高分解能 TOF counter として働く。 MCP had a 4 mm-thick quartz window, so that about 20-25 detectable photo-electrons were expected while we observed about twice. At the time, it was inferred that the extra photons more than the expectation might be yielded by MCP layer itself. However, the measured and readout system resolutions of 13.6 ps and 9 ps indicate the intrinsic resolution of the MPC be 10 ps, which corresponds about detectable 20 photo-electrons. Where these extra photo-electrons come from is a mystery. Anyway, MCP itself provides 10 ps resolution.

13. Very-fast TOF Workshop, U of Chicago, Nov.18.2005.; Takayoshi Ohshima In a case the most photons produced at the window, equipping thicker window, 10 mm, would improve TOF resolution better than 10 ps. 3. TOF counter (TOF-PMT w/o Radiator (continue)) 25x10/4=60 photons, 46 ps/  60 =6 ps;  =  6 2 +7 2 = 9 ps Circuit error Footnote: MCP-PMT(HPK10) の window を 10 mm の quartz とする。また、回路系の分解能＝９ → ７ psec とする。 そうすると、 MCP-PMT 単独で分解能＝９ psec が期待できる。ただし、 GaAsP photo-cathode を想定していない。 また、 HPK10(TTS= ４６ psec ）でなく、 HPK6 （ TTS=30 psec ）ならば window は５．６ mm でよい。 総体として、回路系の分解能を改良することが最重要。 MCP-PMT (TTS=46 ps) When MCP has a thick quartz window, say, 10 mm, then 60 photo-electrons and 6 ps resolution are expected. Including readout system uncertainty, suppose to be it 7 ps, results in 9 ps accuracy in total. If MCP having better TTS and better circuit are prepared, the resolution will be improved.

14. Very-fast TOF Workshop, U of Chicago, Nov.18.2005.; Takayoshi Ohshima 3. TOF counter (10 ps TOF-Counter) MCP-PMT (TTS=30 ps) Footnote: 10 mm の quartz 輻射体を設ける。 Photon 数は６０（ quartz から）と５０（ PMT から）であり、回路系の分解能＝９ → ７ psec とする。 その結果、分解能＝８ psec が期待できる。 MCP-PMT を HPK ６ (TTS= ３０ psec ）でなく HPK １０（ TTS= ４６ psec ）とすると分解能＝８－９ psec が期待できる。 ただし、 GaAsP photo-cathode を想定していない。 particle Quartz (10mm) Nγ=60(60+50) photons  =30ps/  60(110) =3-4 ps with 7 ps circuit error  = 8 ps Or, put 10 mm-thick quartz in front of MCP, for instance, with 30 ps TTS. 3-4 ps intrinsic resolution is attained. Readout system uncertainty would dominate the resolution of 8 ps.

15. Very-fast TOF Workshop, U of Chicago, Nov.18.2005.; Takayoshi Ohshima 3.4 TOF counter (2 nd BEAM-TEST: 5 ps TOF Beam-Test) Aims (1)Study of TOF resolution using SPC (Becker & Hickl GmbH’s) Time-Correlated Single Photon Counting Modules (SPC-134): - channel resolution = 813 fs - electrical time resolution = 4 ps RMS - repetition rates upto 200 MHz (2) Study of extra photons (from MCP itself?) 回路系の分解能 7-8 ps 。これが分解能を決めている。 SPC （分解能４ｐｓ）を当面使用して、 MCP-PMT の分解能を study 。 This SPC includes CFD, TAC, ADC, and MCA (Micro Channel Analyzer). Up to here, the attained resolution was limited mostly by the uncertainty of readout circuit..

16. Very-fast TOF Workshop, U of Chicago, Nov.18.2005.; Takayoshi Ohshima SET-UP LOGIC CIRCUIT The thickness of quartz radiator is varied. We don ’ t need any other readout electronics for MCP ’ s; only the common stop signal is prepared by scintillation counters. - cable: SMA, BNC - discri: 300 MHz - SPC-134: 0.86/count (CFD-TAC-ADC) - AMP: 50 k-1.5 GHz - ATTN: < 18 GHz - power splitter: Using HPK6 (TTS=30 ps) with 3 mm-thick window instead HPK10.

17. Very-fast TOF Workshop, U of Chicago, Nov.18.2005.; Takayoshi Ohshima For SINGLE PHOTONS ADC, TDC and  t (~30 ps) raw signals GAIN, TTS and CE vs HV  t for single photons (spc used) 2 nd BEAM-TEST: “5 ps TOF Beam-Test” (cont.) (CAMAC) Pulser (single photon) による測定

18. Very-fast TOF Workshop, U of Chicago, Nov.18.2005.; Takayoshi Ohshima 2 nd BEAM-TEST: “5 ps TOF Beam-Test” (cont.) For 3 GeV/c PIONS Circuit resolution (  t = 4.1 ps ) TOF w/o radiator (  t = 7.7 ps ) TOF w radiator (  t = 6.2 ps ) Beam による測定。 SPC の分解能。 No radiator & 10 mm crystal 。 6.2(ps) 2 – 4.1 (ps) 2 = 4.7 (ps) 2

19. Very-fast TOF Workshop, U of Chicago, Nov.18.2005.; Takayoshi Ohshima  t vs RADIATOR THICKNESS ■ N  vs RADIATOR THICKNESS ADC distribution of MCP-plate alone Although the number of the photo-electrons increases by using thicker quartz, the resolution gradually deteriorates. It is because the uncertainty of the light path due to the quartz thickness. Almost 1 photo-electrons is seen on an average. The extra photo-electrons are not produced at MCP, it might be at the MCP window.

20. Very-fast TOF Workshop, U of Chicago, Nov.18.2005.; Takayoshi Ohshima 3.3 TOF counter (Know-How: Window materials) チェレンコフ光子数 Footnote: quartz, borosilicate window で検出できる光の波長特性が変わる。輻射体が短い場合は chromaticity も効かないので、 quartz がよい。 Borosilicate 1mm Quartz 3-4 mm Quartz/ Borosilicate In order to have larger number of photo-electrons, a consideration of the window material is important. Photon yields iare a few times different between HPK(3-4mm quatrz) and BINP(1mm Borosi).

21. Very-fast TOF Workshop, U of Chicago, Nov.18.2005.; Takayoshi Ohshima 3.3 TOF counter (Know-How ： Photocathode materials) Footnote: Bi-alkali はほうけい酸ガラスで、 multi-alkali は quartz か？ Not only QE but also -range depend on material. Cherenk ov ∝ 1 /  The choice of photo-cathode material is quite essential. GaAsP indicates much higher QE and wider sensitive frequency range. BINP serves Blue extended GaAs window, which also has a good property. Suitable choice of the material would improved TOF resolution by enlarging the number of photo-electrons. In order to further improve the resolution, we need more photo-electrons. Using thicker radiator rather deteriorates the resolution. Our detector equips already 10 mm-thick quartz. How to increase the number of photo-electrons?

22. Very-fast TOF Workshop, U of Chicago, Nov.18.2005.; Takayoshi Ohshima LIFT-TIME No time to talk

23. Very-fast TOF Workshop, U of Chicago, Nov.18.2005.; Takayoshi Ohshima 4. Summary By R&D We have developed MCP-PMT’s which satisfies the most of our requirements.  TTS = 30 ps &  t = 5 ps is obtained by a beam test. R&D of MCP-PMT is now focused on GaAsP photocathode & Lifetime improvement R&D of readout circuits is focused on Highly stable CFD & TDC