oxon@mac.comcherenkov telescope array

CTA 報告 113：Schwarzschild-Couder光学系を用いた小・中口径望遠鏡の開発

奥村 曉、朝野 彰、片桐 秀明A、佐藤 雄太、重中 茜A、 田島 宏康、中村 裕樹、山根 暢仁、他 CTA Consortium

名古屋大学 宇宙地球環境研究所 (ISEE)、茨城大理A

2016 年 9 月 22 日　日本物理学会

Cherenkov Telescope Array (CTA)

2

© G. Pérez, IAC, SMM

SouthNorth

LST MST SCT SST

1 km

3 km

Small-Sized Telescope (SST) GCT (SC) ~35@S ASTRI (SC) ~35@S Davis–Cotton ~20@S D ~4 m FOV ~9° E = 5 TeV – 300 TeV

Large-Sized Telescope (LST) 4@North + 4@South D = 23 m FOV = 4.5° E = 20–200 GeV

Medium-Sized Telescope (MST) 15@N + 25@S D = 12 m FOV = 8° E = 100 GeV – 10 TeV

Schwarzschild–Couder Telescope (SCT) 25@S D = 9.6 m FOV = 8° E = 200 GeV – 10 TeV

Array Layout Examples

Schwarzschild–Couder Telescope (SCT)

At first, proposed as Advanced Gamma-ray Imaging System (AGIS)

Now an extension for the CTA South

‣ Will improve the sensitivity in 100 GeV – 10 TeV

‣ Davies–Cotton MST × 25

‣ Schwarzschild–Couder × 25

Challenges

‣ High-quality mirrors and fine alignment

‣ SiPM array and ASIC

Will realize a large FOV of 8° and fine angular resolution of ~0.06°

3

9.6 m

5.4 m

Primary

Secondary

Camera

Ideal PSF at 1°

1 mm

Okumura+ (2016)

28 M. Wood et al. / Astroparticle Physics 72 (2016) 11–31

Fig. 19. Performance of benchmark arrays: M61SC, M61DC, M25DC, L5, and L61. Top left: 68% containment angle of the gamma-ray PSF after applying point-source cuts. Top

right: Gamma-ray effective area after point-source cuts. Middle left: Differential rate of the total cosmic-ray background (protons and electrons) after point-source cuts. Middle

right: Differential rate of protons after point-source cuts. Bottom left: Differential point-source sensitivity for a 50 h observation time. Shown as the solid gray line is the differential

sensitivity of Array I from [18] evaluated with the most sensitive analysis at each energy from the four alternative analyses presented in that work (MPIK, IFAE, SAM, and Paris-MVA).

Bottom right: Differential diffuse-source sensitivity (D = 0.5◦) for a 50 h observation time.

degree. Here it should be noted that the sensitivity advantage of theDC telescopes below 50 GeV under low NSB is lost in case of threetimes increased NSB and that the SC design is better for six timeshigher NSB at all energies.

4.11. Number of telescopes in the array

One of the most important parameters concerning the sensitivityof an IACT array is the number of telescopes. A larger number of tele-scopes increase both the total effective area for triggering and recon-structing gamma-ray showers but also increase the average number

of telescopes that participate in the reconstruction of each shower.Increasing the number of telescopes leads to better point-source sen-sitivity and an improved gamma-ray PSF.

Fig. 18 compares the performance of arrays with between 5 and 61telescopes. We investigate the scaling relation of the improvement insensitivity with increasing number of telescope. In the limit of an infi-nite array the point-source sensitivity should scale with the numberof telescopes as N1/2

tel. However we observe an increase of sensitiv-

ity that is slightly better than the N1/2tel

at all energies. This empha-sizes that in the case of small telescope arrays increasing the numberof telescopes yields larger improvements as compared to the case of

28 M. Wood et al. / Astroparticle Physics 72 (2016) 11–31

Fig. 19. Performance of benchmark arrays: M61SC, M61DC, M25DC, L5, and L61. Top left: 68% containment angle of the gamma-ray PSF after applying point-source cuts. Top

right: Gamma-ray effective area after point-source cuts. Middle left: Differential rate of the total cosmic-ray background (protons and electrons) after point-source cuts. Middle

right: Differential rate of protons after point-source cuts. Bottom left: Differential point-source sensitivity for a 50 h observation time. Shown as the solid gray line is the differential

sensitivity of Array I from [18] evaluated with the most sensitive analysis at each energy from the four alternative analyses presented in that work (MPIK, IFAE, SAM, and Paris-MVA).

Bottom right: Differential diffuse-source sensitivity (D = 0.5◦) for a 50 h observation time.

degree. Here it should be noted that the sensitivity advantage of theDC telescopes below 50 GeV under low NSB is lost in case of threetimes increased NSB and that the SC design is better for six timeshigher NSB at all energies.

4.11. Number of telescopes in the array

One of the most important parameters concerning the sensitivityof an IACT array is the number of telescopes. A larger number of tele-scopes increase both the total effective area for triggering and recon-structing gamma-ray showers but also increase the average number

of telescopes that participate in the reconstruction of each shower.Increasing the number of telescopes leads to better point-source sen-sitivity and an improved gamma-ray PSF.

Fig. 18 compares the performance of arrays with between 5 and 61telescopes. We investigate the scaling relation of the improvement insensitivity with increasing number of telescope. In the limit of an infi-nite array the point-source sensitivity should scale with the numberof telescopes as N1/2

tel. However we observe an increase of sensitiv-

ity that is slightly better than the N1/2tel

at all energies. This empha-sizes that in the case of small telescope arrays increasing the numberof telescopes yields larger improvements as compared to the case of

SCT Advantages

MC study shows SCTs have better angular resolution of arrival direction by a factor of ~1.5 than Davies–Cotton MSTs

~30% better sensitivity for point sources

Less expensive camera pixel cost4

Angular Resolution Point Source Sensitivity

Wood+ (2016)

SCT × 61

DC-MST × 61

SCT × 61

DC-MST × 61

The SCT Camera

The plate scale of 97.5 mm/deg and FOV of 8° require SiPM or MAPMT camera modules

The focal plane is covered with 11,328 SiPM pixels

64 pixels per module are readout by 4 TARGET ASICs

5

~80 cm = 8° FOV

11,328 pixels

SiPM

TARGET 7

Pre-amps, ASICs, FPGA …

The SCT Prototype is under Construction

6

Sep 11, 2016 at the VERITAS site http://cta-psct.physics.ucla.edu/index.html

Primary

Secondary

Camera

Small-Sized Telescope (SST)

7

3 different designs proposed

‣ ASTRI

‣ 1M-SST (Davies–Cotton)

‣ Gamma-ray Cherenkov Telescope (GCT)

GCT uses similar techniques as in SCT ‣ SC optics

‣ TARGET modules and SiPMs

‣ Backplane board

‣ Camera control and DAQ software

The GCT prototype was inaugurated at the Paris Observatory in Dec 2015

Prototype of Gamma-ray Cherenkov Telescope (GCT)

Secondary Mirror

Primary Mirror

Camera

Inauguration of the GCT Prototype (Dec 2015)

8

The GCT Camera Prototype (with MAPMTs)

9

~40 cm

64-ch MAPMT × 32 2,048 pixels

@ Univ. Leicester, UK

@ Paris Observatory, France

Lid

Water chiller

Camera Module

TARGET ASIC × 4 (16 ch per ASIC)

FPGA

Hamamatsu MAPMT

Pre-amp Boards

LED Flasher

The First Cherenkov Light in CTA (Nov 26, 2015)

10

・Very Preliminary Calibration

・Mirrors are not aligned

・Under the Paris sky

40 ns

SiPM-based GCT Camera

32 MAPMTs will be replaced by SiPMs, and the latest TARGET ASICs will be used

A prototype with improved camera mechanics will be built in 2017

Three production telescopes will be at the CTA South in 201811

SiPM-based Camera 3 mm SiPM (16 × 16 pixels)

52 mm

SiPM Glass Window

Coolant

Figure 17: Trigger threshold (a) and noise (b) as a functionof delay between a signal synchronous with the samplingclock and the generation of the input pulse to the ASIC.

Figure 18: Same as Fig. 15, but the input pulse timing issynchronized with the signal sampling clock. The delay be-tween the reference signal and input pulse is 25 ns in thisfigure.

Figure 19: Trigger threshold (a) and noise (b) as a function ofPMTref4 and Thresh (in DAC counts) with analog samplingdisabled. White indicates a region of parameter space wherethe trigger does not function properly.

3.3.3. Trigger performance with sampling disabled585

Because the performance of the trigger system586

is strongly dependent on the analog sampling, we587

characterized the trigger performance with sam-588

pling disabled. Figure 19 shows trigger threshold589

and noise from our scan in PMTref4 and Thresh590

when the sampling is disabled.591

In this configuration, the minimum workable592

threshold is . 5 mV (1.2 p.e.), and the trigger noise593

is .0.5 mV (0.13 p.e.). Figure 20 shows an ef-594

ficiency curve for sampling disabled for which the595

threshold was 4.76 mV. Also in this casfe the mea-596

sured e�ciencies are well fit by an S curve.597

In conclusion, the performance of the trigger cir-598

cuit with sampling disabled meets the desired sen-599

sitivity and noise level. Therefore, two options600

were considered: 1) reducing the interference on the601

ASIC between sampling and triggering circuits (by602

improving isolation and increasing the gain on the603

trigger path), 2) separating data and trigger path604

into two di↵erent ASICs. Option 1) was chosen as605

the most cost e↵ective for the design of TARGET 7,606

and option 2) for subsequent ASIC pairs.607

14

The outcome is comparable to the former measurements: over the measured range of PMTref4 we get lowestthresholds of < 10 mV (2.5 PE) and over the whole range of PMTref4 and Thresh we get a noise of < 2 mV (0.5PE), while the noise increases with larger thresholds. The visible fluctuations of the noise are caused by statisticalfluctuations. It shows that thresholds between 2.5 PE and 40 PE can be set to 0.5 PE precision.

Thus, the desired performance of the trigger ASIC T5TEA is met even at this early stage of our board design.

OUTLOOK

Pulse Amplitude [mV]2 4 6 8 10 12

Trig

ger E

ffici

ency

0

0.2

0.4

0.6

0.8

1

0.018) mV± = (2.802 µ

0.024) mV± = (0.550 σ

FIGURE 8. E�ciency curve of T5TEA measuredwith the newest design of the evaluation board.

Since the presented measurements do not represent final results, fur-ther improvements are expected: on the one hand by tuning all avail-able parameters and on the other hand by improving the design of theevaluation boards. Figure 8 shows an e�ciency curve measured withthe newest design of our evaluation board and the improvements areclearly visible: a lower threshold of 3 mV (3/4 PE) and a lowernoise of 0.6 mV (0.15 PE). Currently ongoing investigations ad-dress the whole readout chain with SiPM, bu↵er and shaper boardand an evaluation board. The next steps in the near future includeproduction and testing of complete TARGET modules and the pro-duction of an entire camera, called GCT, which will be used in theprototype telescope CHEC-S.

ACKNOWLEDGMENTS

We gratefully acknowledge support from the agencies and organiza-tions under Funding Agencies at www.cta-observatory.org.

REFERENCES

[1] L. Tibaldo, J. A. Vandenbroucke, A. M. Albert, S. Funk, T. Kawashima, M. Kraus, A. Okumura, L. Sapozh-nikov, H. Tajima, G. S. Varner, T. Wu, A. Zink, and f. t. CTA consortium, ArXiv e-prints August (2015),arXiv:1508.06296 [astro-ph.IM] .

[2] B. S. Acharya, M. Actis, T. Aghajani, G. Agnetta, J. Aguilar, F. Aharonian, M. Ajello, A. Akhperjanian,M. Alcubierre, J. Aleksic, and et al., Astroparticle Physics 43, 3–18March (2013).

[3] A. De Franco, R. White, D. Allan, T. Armstrong, T. Ashton, A. Balzer, D. Berge, R. Bose, A. M. Brown,J. Buckley, P. M. Chadwick, P. Cooke, G. Cotter, M. K. Daniel, S. Funk, T. Greenshaw, J. Hinton, M. Kraus,J. Lapington, P. Molyneux, P. Moore, S. Nolan, A. Okumura, D. Ross, C. Rulten, J. Schmoll, H. Schoorlem-mer, M. Stephan, P. Sutcli↵e, H. Tajima, J. Thornhill, L. Tibaldo, G. Varner, J. Watson, and A. Zink, ArXive-prints September (2015), arXiv:1509.01480 [astro-ph.IM] .

[4] A. N. Otte, J. Biteau, H. Dickinson, S. Funk, T. Jogler, C. A. Johnson, P. Karn, K. Meagher, H. Naoya,T. Nguyen, A. Okumura, M. Santander, L. Sapozhnikov, A. Stier, H. Tajima, L. Tibaldo, J. Vandenbroucke,S. Wakely, A. Weinstein, D. A. Williams, and f. t. CTA Consortium, ArXiv e-prints September (2015),arXiv:1509.02345 [astro-ph.IM] .

[5] A. Albert, S. Funk, T. Kawashima, M. Murphy, A. Okumura, R. Quagliani, L. Sapozhnikov, H. Tajima,L. Tibaldo, J. Vandenbroucke, G. Varner, and T. Wu, ArXiv e-prints July (2016), arXiv:1607.02443 [astro-ph.IM] .

TARGET ASIC – 16-ch Readout and Trigger

We have been developing TARGET ASICs since 2009

The trigger performance was worse than expected in T1, T5, and T7 (TARGET variants)

Finally satisfactory trigger performance achieved in the latest design (TARGET C & T5TEA)

12

TARGET 5

Albert+ (2016)

T5TEA

15–45 mV Thd

~3 mV Thd (~0.8 p.e.)

Funk+ (2016)

Crosstalk Rate (%)0 10 20 30 40

PDE

(%)

10

20

30

40

50

60

m / 3 mmµRef. 50 m / 3 mmµLCT5 50 m / 6 mmµLCT5 50

m / 3 mmµLCT5V 50 m / 3 mmµLCT5 75 m / 6 mmµLCT5 75

m / 3 mmµLVR 50 m / 6 mmµLVR 50 m / 6 mmµLVR 75

m / 6 mmµFJ 35 m / 6 mmµNUV 30

SiPM Characterization for SCT and SST

SiPM characterization in various temperature/voltage conditions underway

Recent SiPMs have lower crosstalk rate and higher PDE compared to ~2010

PDE ~60% is expected with a lens array13

See 山根暢仁 24aSG-10

HPK LCT5

HPK LVR

HPK ~2010

Summary

We have developed ‣ Schwarzschild–Couder telescopes for CTA Medium and Small-

Sized Telescopes

‣ Cameras with the TARGET ASIC technology and SiPMs

The prototype of GCT (a Small-Sized Telescope design) was inaugurated at the Paris Observatory Dec 2015

The first Cherenkov events ever in CTA were recorded

Further improvements of TARGET, photodetectors, and camera mechanics foreseen

14