Compact PCI Bus. Fig. 2. The block diagram of the electronics. The requirements on the electronics are: (1) Dynamic range: must respond linearly to r lepton ...
Modern Physics Letters A J Vol. 19, Nos. 13-16 (2004) 1117-1124 © World Scientific Publishing Company
n
. . . . . _ . .... ^*i World w w w . w o r l Scientific dscientlfi,com
A
mma m
P E V COSMIC N E U T R I N O S FROM T H E M O U N T A I N S
P. Y E H e , H. ATHAR C , N. LA BARBERA 6 , S. B O U A I S S l \ O. CATALANO 6 , G. CUSUMANO 6 , Y. C H P , W.-S. HOU e , Y. B. HSIUNG e , C.-C. HSU e , M. A. HUANG-'7, J. G. LEARNED*?, G.-L. LIN d , T. MINEO 6 , B. SACCO 6 , M. SASAKP, J.-G. SHIU e , J.-J. TSENG", K. UENO e , F. VANNUCCl' 1 , Y. VELIKZHANIN e , M.-Z. WANG e a
Institute of Physics, Academia Sinica, Taipei, Taiwan 115 IFCAI, Consiglio Nazionale delle Ricerche, Palermo, Italy 0 Physics Division, National Center for Theoretical Sciences, Hsinchu, Taiwan 300 d Institute of Physics, National Chiao-Tung University, Hsinchu, Taiwan 300 e Department of Physics, National Taiwan University, Taipei, Taiwan 106 f General Education Center, National United University, Miao Li, Taiwan 360 9 Department of Physics and Astronomy, University of Hawaii at Manoa, Honolulu, Hawaii 96822, U.S.A. h LPNHE, University of Paris VI and VII, Paris, France 1 Insititute for Cosmic Ray Research, University of Tokyo, Kashiwa 277-8582, Japan b
(The NuTel Collaboration) The NuTel collaboration is building a wide field-of-view Cerenkov telescope to be installed on a mountain site for observing near horizontal air showers emerging from another mountain. Cosmic tau neutrinos is the primary source of such showers. This technique will be realized for the first time in the vT energy range of 2 PeV to 1000 PeV. The telescope has enough sensitivity to observe cosmic neutrinos from sources like Active Galactic Nuclei and Galactic Center assuming fluxes from current theoretical models. Keywords: Active galactic nuclei; Earth skimming; tau neutrino appearance; extensive air shower; Cerenkov telescope. PACS Nos.: 14.60.Lm, 95.55.Vj, 95.85.Ry, 98.54.Cm
1. Introduction The origin of ultra-high energy cosmic rays (UHECR) is still a great puzzle1. Bottom-up theories propose that they originate from energetic processes in Active Galactic Nuclei (AGN) or Gamma Ray Bursts (GRB). The hadron component of such energetic processes could interact with accreting materials to produce neutrinos through the decay of charged pions. On the other hand, top-down theories like to suggest that UHECR are decay products of topological defects or heavy relic particles. According to these theories, there are more neutrinos than gamma rays and protons 2 . Measurement of the neutrino flux at and below the ankle region (i.e. < 3 x 1018 eV) provides a good discriminator between the two scenarios. Studies of small scale anisotropy of cosmic rays near the ankle region reveal a small but significant excess of events near galactic center (GC) 3 ' 4 ' 5 ' 6 , where a su1117
1118
P. Yeh et al.
permassive black hole with mass 2.6 x 1O6M0 is known to exist 7 . These anisotropics could be signals of potential source or acceleration site of cosmic rays in galactic center. The neutrino flux from GC would be higher than prediction if it is a point source of neutrinos 8 . Unlike charged cosmic rays, neutrinos do not bend in magnetic field and provide a pointing capability to the production site. Compared to photons which suffer attenuation during propagation, the weakly-interacting neutrinos are complementary probes to the universe. However, the difficulty of neutrino observation has always been the weak interaction which directly translates to rare signals. Recent results on atmospheric neutrinos add an interesting twist to cosmological neutrino detection. Super-Kamiokande data strongly suggest that muon neutrinos oscillate into tau neutrinos9. Since cosmic neutrinos are predominantly produced via pion decays, one does not expect much directly produced cosmic vT flux10. However, with the maximum oscillation seemingly taking place, the fluxes Tiy^ : !F(yT) is expected to reach 1 : 1 for distant observers. Detecting a r lepton descendent on Earth would not only probe AGN/GRB/GC mechanisms, but would also constitute a z/T-appearance experiment. All neutrino experiments need a large target volume to convert incoming neutrinos into detectable objects like charged leptons or proton recoil for observation. Traditional design is surrounding the target volume by detectors. The target has to be large enough to have sensitivity. Targets such as lake water (Baikal 11 ), sea water (ANTARES 12 , NESTOR 13 ) and south pole ice (AMANDA 14 , IceCube 15 ) have been used. To shield the detectors from backgrounds from cosmic rays and other sources, such experiments are located deep under the Earth surface. Although the target is free, the vast amount of sensors surrounding it drives the cost high. The idea of using Earth 1 6 ' 1 7 or mountains 18 ' 19 as target and use atmosphere as a calorimeter for shower development has been a new trend. AMANDA and Auger 20 are using horizontal showers to observe high energy neutrinos. The NuTel collaboration 21 is specifically using the mountain idea to build a wide field-of-view Cerenkov telescope on a mountain to observe air showers emerging from another mountain or the Earth. The target mountain also function as a shield to block cosmic rays and star lights. Among all flavors of neutrinos, this technique is sensitive to vT because only the r lepton has a reasonable probability to escape the mountain and initiate extensive air showers. A preliminary feasibility study showed that this technique performs best in the vT energy range of 1 PeV to 1000 PeV 22 ' 23 , and the most plausible source in this energy range are AGNs 24 , which produces an integral r lepton flux approximately 7.5 k m - 2 sr _ 1 yr _ 1 . The candidate site for the observatory is Mt. Hualalai on the Big Island of Hawaii facing Mauna Loa with a valley of 30 - 40 km wide. The galactic center passes the field of view several times a year. In this paper we describe the design of the optical system, electronics and data acquisition system of the telescope.
PEV Cosmic Neutrinos from the Mountains
Mirror diameter F number Field of View segments Image Intensifier input window magnification factor MAMPT gain Number of pixels Focal surface pixel size Transportation efficiency Fig. 1.
1119
1.8 m 0.74 50° x 50° 17 40 cmx40 cm 10 4 x 104 (typical) 1024 0.5° 50% (typical)
Left: a schematic view of the optical unit. Right: parameters of the optical unit.
2. The NuTel Apparatus The Cerenkov photons produced by particles in an extensive horizontal air shower arrive at a telescope with a typical angular span of 1° to 2° and time spread of 10 ns to 100 ns. A fast single photon sensor with angular resolution of 0.5° per pixel can capture a course-grained image of the shower. We use multi-anode photomultiplier tubes (MAPMTs) to achieve this. The major background during observation is expected to be from random ambient photons and cosmic ray showers. While the cosmic ray showers can be rejected from its arrival direction away from mountain, the night sky background (NSB) must be reduced in the online system. The NSB has been measured elsewhere 25 ' 26 to be as 2.0 x 103 photons/m 2 /sr/ns in the relevant wavelength band of 300 nm 600 nm, and was confirmed by our own measurement on the LuLin mountain in central Taiwan. The night mountain background (NMB) is about 10 times smaller. These backgrounds can be greatly reduced by requiring pixel clusters. The NuTel concept was conceived with a wide field-of-view Fresnel lens similar to the EUSO design 19 . It is recently changed to a Baker-Nunn optical system developed by Ashra 27 for a better matching on timing and availability. The NSB generates a photoelectron rate per pixel around 16 MHz. The online electronics can reduce it to 100 Hz or so by requiring pixel clusters. An identical telescope nearby looking at the same field of view is used to reduce the random background to a level comparable to the expected cosmic neutrino flux. This stereo design also has advantages in reconstruction of the air showers. The optical unit of one telescope contains a segmented mirror, a corrector lens, an image intensifier (II), optical fibers and multi-anode preamplifiers (MAMPT), where the optical path up to the image intensifier is the Ashra design. The segmented mirror reflects Cerenkov photons to the focal surface, where the II with matching input surface collects the photons and intensifies the image. Optical fibers coupled to the output phosphorescent screen of the II are used to transport the image to the input window of MAPMTs. See figure 1 for the schematics.
1120
P. Yeh et al. 16 channels (1 channel shown)
Preamplifier Board
32 channels
x64
handle 512 channels
Data Collection Module
x32
Single Board Computer
x2
MAPMT x16 Trigger I PGA
V1 64 channels Signal-Sharing Board
1 InlerUri*
Compact PCI Bus
Fig. 2.
The block diagram of the electronics.
The requirements on the electronics are: (1) Dynamic range: must respond linearly to r lepton showers in the energy range of 2 PeV - 1000 PeV. (2) Speed: must be fast enough to handle the NSB (~ 16 MHz per pixel) and record the Cerenkov signals. (3) Photon sensitivity: must have low noise level to detect single Cerenkov photon. (4) Uniformity: the response of each pixel to Cerenkov photons must be the same or properly compensated to show the true image. (5) Synchronization: must be able to synchronize telescopes within one digitization clock. We have designed the electronics by using as many commodity as possible to keep the cost down and the development time short. We use low-cost amplifiers, 10bit pipeline ADCs, low-cost FPGAs, and sharing of high voltage outputs. Another feature of the design is that there is no analog discriminator: all trigger decisions are in digital space. Figure 2 shows the schematics of the electronics chain: MAPMTs, preamplifier boards, data collection modules (DCM) and single board computers (SBC). A signal sharing board accepts all 64 channels of an MAPMT and couple each channel to 2 other signal-sharing channels resistively to split the charge. The signalsharing channels are about 2.1° away in field of view to reduce the probability of being shined by the same shower. They provide charge information when the main channel is saturated. A preliminary measurement on the first working channel shows that it extends the dynamic range by a factor of 3. The 16-channel charge-sensitive preamplifier boards amplifies and shapes the signal with a time constant of 387 ns to match the 40 MHz clock of DCM. With this relatively long time constant and the high background rate from NSB, the shaped pulses will pile up. The output voltage will reach a DC level of approximately 50 mV for a steady NSB of nominal flux. That's a tolerable 2.5% of the dynamic range of 2 V. Onboard the DCM there are 32 10-bit pipeline ADCs, 4 readout FPGAs and
PEV Cosmic Neutrinos from the Mountains
1121
1 trigger FPGA to process signals from 32 preamplifier channels. The number of photoelectrons at each clock is determined with a simple double correlation logic in the readout FPGA: Ni = Ai — exp(—25/387)Aj_i where the exponential factor is a convenient 15/16. Such a logic allows us to operate the telescope in a harsh background condition. The readout FPGA applies 3 programmable thresholds (Low, High, Very High) on Ni to produce a bitmask. All 3 thresholds are adjustable during observation. The events that passes VH threshold has the highest priority during processing. The trigger FPGA applies configurable cluster trigger algorithms on the bitmasks of 32 channels. The resulting trigger bit is sent to the next DCM to be OR-ed with its trigger bit. This operation is repeated in a daisy chain fashion until the OR-ed trigger bit reaches the terminal DCM in the crate, otherwise known as the Master DCM. The daisy chain is pipelined to achieve low or even zero deadtime. The data acquisition system (DAQ) is composed of compact PCI crates, each with one single board computer (SBC) and 16 DCM boards. One SBC collects event data from 16 DCMs in the same crate through the 33 MHz 32-bit wide compact PCI bus. the data are read when the event buffers of DCMs are half-full to maximize data transfer rate. The CPU on SBC assembles the event and apply more sophisticated trigger algorithms to further reduce data volume. The data is finally gathered onto one SBC and written to storage. The SBC runs a Linux a operating system to coordinate resources and tasks. Although not a real time system, the interrupt latency of Linux is measured to be less than 20 /xs, sufficient for the DAQ task. Full event reconstruction will be done in SBC during day time when the telescopes are turned off. 3. Trigger and Reconstruction Figure 3 shows the image for signal generated by Corsika28 air shower simulation and night sky backgrounds taken in a 25 ns exposure time. The signature of an air shower is pixel cluster. The online trigger FPGA of the DCMs first identifies a seed pixel above a high threshold H, then examines the neighboring pixels. There are 2 major types of algorithms: threshold-based and sum-based. The threshold-based algorithm requires M neighboring pixels to pass a low threshold L, while the sum-based algorithm requires the total number of photoelectrons in the neighboring pixels to exceed the threshold S. The trigger rate of different algorithms are shown in figure 3. NuTel has 2 telescopes about 10 m apart for coincidence and triangulation. The optical axes of telescopes are parallel to the normal direction of the telescope plane. In the Corsika simulation, ir~ showers are injected from different distances to the telescope plane to form the signal sample for reconstruction study. Simulation shows that the system can detect Cerenkov photons RS 1 km away from a 1 PeV hadronic Linux is a trademark of Linus Torvalds.
1122
P. Yeh et al.
N
~10
r
6
|io i-
2
^T~" iSi, .„ i v K
™ Single Pixel 9.2) Duo Trigger
1 10
-2 •
Seed Threshold H N
d
^m6:
• % " "
tr 10 H
2
10
:,._„X
X
_D>
!
T — H = p.e. -••• H = 2 p.e. H =;
^ - - '• \ X 2
0
2
4
6 8 10 Additional p.e.
Fig. 3. Upper left: an image from a 1 PeV pion shower travelling through 20 km of air near the telescope. Pions are the dominant decay product of tau lepton decays. Lower left: an image from random photon background with a flux of 2000 p h / m 2 / n s / s r . Upper right: NSB trigger rates as a function of seed threshold H for 3 different algorithms. Lower right: NSB trigger rates as a function of number of photoelectrons in the neighboring pixels for the sum trigger algorithm.
shower. A shower has 6 parameters: energy E, shower maximum position (x,y,z), and direction of shower axis (61,62) where 6i is the angle between shower axis and the direction from shower maximum to «-th telescope. The reconstruction of the shower is done by minimizing a chi-square that represents the difference of observed number of photoelectrons on each pixel rij and expected number of photoelectrons f(E,x,y,z,61,62). The function / is obtained from air shower simulation. The reconstructed energy resolution is 40%, and the angular resolution is about 1.0°. 4. Expected Sensitivity The sensitivity of the NuTel telescope is determined based on the flux reachable with 2 years of observation time: Nsigna,i = / dEv dET „(E) < 4.7 x 102 eV/cm 2 /s/sr in the energy range of 2 PeV - 1 EeV. Using a theoretical vT flux30 we found that the prototype NuTel system can observe 0.5 AGN neutrino events per calendar year.
5. Conclusion The NuTel collaboration is building a neutrino telescope to observe Cerenkov photons from air showers produced by r leptons emerging from a mountain or Earth surface. Such a shower is a distinct signature of cosmic r neutrinos, that makes NuTel a r-appearance neutrino experiment. NuTel will use Ashra's mirror and image intensifier design. This speeds up the development of the telescope tremendously, while providing timely feedback, electronics and operating experience to the Ashra collaboration. The NuTel telescope can function as a pilot for Ashra. The prototype NuTel system will be integrated and moved to Hawaii for field test and observation in 2004. An upgrade in field of view and number of channels is forseen if the prototype works well. By the end of 2006, the NuTel prototype will extend the energy reach of cosmic tau neutrinos by 2 orders of magnitude compared to current experiments. The sensitivity on cosmic tau neutrino flux with 2 years of observation is estimated to be E2(pI/(E) < 4.7 x 10 2 eV/cm 2 /s/sr. An upgrade would provide more stringent upper limit on the diffuse flux of cosmic tau neutrinos.
Acknowledgments This work is funded by Ministry of Education of Taiwan under grant MOE 89-NFA01-1-4-2.
1124 P. Yeh et al.
References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30.
J. G. Learned and K. Mannheim, Ann. Rev. Nucl. Part. Sci. 50, 679 (2000). P. Bhattacharjee and G. Sigl, Phys. Rept. 327, 109 (2000). e-print: astro-ph/9811011. M. A. Huang, Ph.D. thesis, University of Utah, 1996. D. J. Bird et al., Astrophys. J. 511, 739 (1999). N. Hayashida et al., Astropart. Phys. 10, 303 (1999). e-print: astro-ph/9807045. J. A. Bellido, R. W. Clay, B. R. Dawson, and M. Johnston-Hollitt, Astropart. Phys. 15, 167 (2001). e-print: astro-ph/0009039. R. Schoedel et al, Nature 419, 694 (2002). K. Takahashi and S. Nagataki, e-print: astro-ph/0108507. S. Fukuda et al, Phys. Rev. Lett. 85, 3999 (2000). e-print: hep-ex/0009001. H. Athar and G.-L. Lin, Astropart. Phys. 19, 569 (2003). e-print: hep-ph/0108204. C. Spiering et al, Nucl. Phys. Proc. Suppl. 14B, 51 (1990). L. Moscoso, Nucl. Phys. Proc. Suppl. 77, 492 (1999). e-print: hep-ex/9809020. F. Ameli et al, Nucl. Instrum. Meth. A423, 146 (1999). F. Halzen et al, Nucl. Phys. Proc. Suppl. 77, 474 (1999). e-print: hep-ex/9809025. J. Ahrens et al, Nucl. Phys. Proc. Suppl. 118, 388 (2003). e-print: astro-ph/0209556. G. Domokos and S. Kovesi-Domokos, e-print: hep-ph/9801362. J. L. Feng, P. Fisher, F. Wilczek, and T. M. Yu, Phys. Rev. Lett. 88, 161102 (2002). e-print: hep-ph/0105067. D. Fargion, A. Aiello, and R. Conversano, e-print: astro-ph/9906450. F. Vannucci, in NATO Advanced Research Workshop on Cosmic Radation: From Astronomy to Particle Physics (Oujda, Morocco, 2001, submitted for publication). C. Hojvat, , presented at 32nd Rencontres de Moriond: High-Energy Phenomena in Astrophysics, Les Arcs, France, 18-25 Jan 1997. NuTel Collaboration, (2002). [Online] http://www.nutel.org/. G. W. S. Hou and M. A. Huang, Nucl. Phys. Proc. Suppl. 118, 425 (2003). e-print: astro-ph/0204145. M. H. Huang, J. J. Tseng, and G. L. Lin, in International Cosmic Ray Conference (Tsukuba, Japan, 2003). O. E. Kalashev, V. A. Kuzmin, D. V. Semikoz, and G. Sigl, Phys. Rev. D66, 063004 (2002). e-print: hep-ph/0205050. G. Hermann, C. Kohler, T. Kutter, and W. Hofmann, e-print: astro-ph/9508028. S. Preuss, G. Hermann, W. Hofmann, and A. Kohnle, Nucl. Instrum. Meth. A481, 229 (2002). e-print: astro-ph/0107120. M. Sasaki, Y. Asaoka, and M. Jobashi, Nucl. Instrum. Meth. A501, 359 (2003). D. Heck et al, Report FZKA 6019, Forschungszentrum Karlsruhe (2001). G. J. Feldman and R. D. Cousins, Phys. Rev. D57, 3873 (1998). e-print: physics/9711021. D. V. Semikoz and G. Sigl, e-print: hep-ph/0309328.
