- Email: [email protected]
A new direction sensitive Optical Module to be employed in deep sea neutrino telescopes
ARTICLE IN PRESS Nuclear Instruments and Methods in Physics Research A 595 (2008) 67–69
Contents lists available at ScienceDirect
Nuclear Instruments and Methods in Physics Research A journal homepage: www.elsevier.com/locate/nima
A new direction sensitive Optical Module to be employed in deep sea neutrino telescopes Andrea Bersani Istituto Nazionale di Fisica Nucleare, Sezione di Genova, Via Dodecaneso 33, 16146 Genova, Italy
On behalf of the NEMO Collaboration a r t i c l e in fo
abstract
Available online 8 July 2008
The NEMO (NEutrino Mediterranean Observatory, [I. Amore, [NEMO Collaboration], et al., Int. J. Mod. Phys. A 22 (2007) 3509, arXiv:0709.3991 [astro-ph]]) project studies, within the KM3NeT framework [U.F. Katz, Nucl. Instr. and Meth. A 567 (2006) 457, arXiv:astro-ph/0606068], new technologies for a 3 km -scale neutrino telescope in the Mediterranean Sea. The telescope goal is the investigation of the high energy component of the cosmic neutrino spectrum. The detection of these neutrinos is a promising tool for a better understanding of the mechanisms that originate the extreme energy cosmic rays. Neutrino energy and direction are reconstructed collecting the Cherenkov light produced in water by the muon coming from a neutrino interaction. Two prototypes of a new large area (10 in.) 4-anode photomultiplier, and manufactured by Hamamatsu, on request of the NEMO Collaboration, will be used for the first time to detect the direction of the detected Cherenkov light at the NEMO Capo Passero site. & 2008 Elsevier B.V. All rights reserved.
Keywords: Cosmic rays Neutrino telescope Photon detector
1. Conventional vs. directional Optical Modules Deep sea neutrino telescopes were designed since the late 1970s, with a common structure: an array of vertical structures featuring a series of layers of light sensitive elements, usually called Optical Modules (OM). After the pioneering work of Dumand Collaboration [3], new techniques were developed both for sea water and for polar ice detectors [1–7]. All the developed Optical Modules have a very similar structure. A pressure resistant glass sphere hosts a large emispherical photomultiplier, optically coupled to the sphere using a layer of optical gel or glue. The power supply and front-end electronics are usually included in the sphere and directly connected to the PMT. This setup is well proven, and allows quite a good track reconstruction if the number of hits is sufficiently high, i.e. 10 or more. This means that the track length must be sufficiently large. This is due to the limited density of sensitive elements which can be installed in such kind of detector. A steep reduction of the efficiency under 1 TeV is the observed effect for a cubic kilometer scale detector, instrumented with 104 OMs. The hit multiplicity is partially recovered grouping several OMs in each layer: this is realized, as an example, in ANTARES with 3 OMs looking in different directions and placed at a small
E-mail address: [email protected] 0168-9002/$ - see front matter & 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.nima.2008.07.010
distance one from the other [8]. Nevertheless, the number of OMs is limited by costs and failure rate considerations, and, for low energy tracks, the poor definition of the track position w.r.t. the OM itself, is a strong limitation to the reconstruction capability for the detector as a whole. An important increase in the hit multiplicity can be achieved if in any OM a certain number of sensitive elements is installed. A proposed solution [9] features 12 3 in. PMTs: this means an active area comparable with the 10 in. PMT, but a complete redefinition of readout electronics should be needed. Another solution is under study at CERN [10], involving an HPD with a pixelized anode instead of the PMT as the active element of the OM. We decided instead to ask industry to develop a 4-anode, 10 in. PMT, which has, at least, the same photocathode area as a traditional, equivalent radius PMT. This solution has not the same granularity, but is compatible with the present readout chain with minor improvements. Such a kind of PMT can be coupled to a properly designed light guide, divided in four sectors, with mirrors, which can relate the incoming light direction to the photocathode quadrant which is hit. This is the second main improvement introduced for this OM: the angular acceptance of each quadrant is smaller than that of a single anode PMT. Therefore theposition of the track w.r.t. the OM is measured much better. This improves the track reconstruction. The arrangement of the PMT and the light guide in the glass sphere is shown in Fig. 1.
ARTICLE IN PRESS 68
A. Bersani / Nuclear Instruments and Methods in Physics Research A 595 (2008) 67–69
Table 1 Performances of the four anodic prototype compared to Hamamatsu R7081
Nominal voltage (v) Gain Peak to valley ratio Dark noise (thr. 0.3 pe) (Hz) Transit time spread (ns)
Requirement
R7081
Prototype
o2000
1340
1550
5 107 42 o10 000 o4
5 107 2.8 910 3
5 107 3 1200 4
Q1
22026
20 13359 40
2. Directional OM realization The realization of a directional OM requires the development of completely novel technologies, including:
a 4-anode PMT, the light guides, power supply and voltage divider, frontend electronics.
2.1. The 4-anode photomultiplier
4914
60 Theta (deg)
Fig. 1. A section of a directional OM: if compared to a traditional one, the thick light guide with mirrors is the main difference.
8103
2980 80
1808 1096
100
665 120 403 140
244 148
160 20
40
60
80 100 Phi (deg)
120
140
160
Fig. 2. First anode response as a function of the position of the light source on the photocathode surface (polar coordinates).
The 4-anode PMT was realized by Hamamatsu, starting from their former experience in large area PMTs. Two samples were delivered to us, after some measurement made at factory. The angular acceptance, compared to the one of a traditional PMT, shows a more constant behavior as a function of the incidence angle, with a steep decrease at angles greater than 801: this is mainly due to the larger surface of the first dynodes. A series of measurement were performed at the testing facility of the NEMO group at the INFN Sezione di Catania. We studied the single photoelectron peak, the transit time spread, the gain and the cross-talk of the prototype, to have a complete characterization and make feasible a comparison with previous models. A summary of these measurements is shown in Table 1. Cross-talk between the four quadrants, due to focusing of the photoelectrons on the wrong dynode, is very small, of the order of few %. A map of the photocathode response varying the position of a light source on the PMT surface is shown in Fig. 2. 2.2. The light guide
Fig. 3. A photograph of the light guide, before the installation in the pressureresistant sphere. One quarter is removed from the working position and the ‘‘focusing’’ behavior is evident if looking to the front sector.
A proper plexiglas light guide was realized to couple optically the multianode PMT to the pressure resistant sphere. A cage of mu-metal to screen the Earth magnetic field is foreseen between the light guide and the sphere. The same optical glue used for the NEMO OMs is used to keep the guide in position. The thickness of the guide was optimized to have the maximum separation between the four quadrants. This is achieved with a light guide with thickness comparable with the transverse dimension of each quadrant. In fact, the total effective area of the PMT is almost unchanged, due to Liouville’s theorem, but a very effective focusing is achieved with a guide, acting as a
Winston cone, 10 cm thick. A photograph of the light guide is shown in Fig. 3. The lateral surfaces of the light guide must be reflective to avoid a loss of sensitivity due to dead angles. It was decided to use a very thin foil of reflective material. Several types of reflective films have been tested, some based on aluminized surfaces (for example, aluminized mylar, or HiFi [11]) and others based on multilayer sandwiches of many transparent foils with different refractive indices (for example, 3M Vikuiti ESR Film [12]). The best performances were achieved with the latter, optically coupled to
ARTICLE IN PRESS A. Bersani / Nuclear Instruments and Methods in Physics Research A 595 (2008) 67–69
2.4
3M (adhesive) 3M
2.2
HiFi Aluminum Teflon
2
Aluminated Mylar
0
10
20
30
40
50
60
70
Incidence Angle (deg) Fig. 4. Comparison of the reflectivities of different materials optically coupled to plexiglas.
Effective Areas Ratio
Reflected Power (mW)
Reflectivity - Matching: optical grease 4 3.5 3 2.5 2 1.5 1 0.5 0 -0.5 -1
69
1.8 1.6 1.4 1.2 1 2
2.5
3
3.5
4
4.5
5
5.5
6
6.5
7
Log (E/GeV) Fig. 6. Ratio of the effective areas calculated for a cubic kilometer detector with and without directional OMs: a gain of a factor 2 is predicted at low energy.
other hand, at low energies, the effective area is the sum of effective areas of several small detectors (corresponding to the single vertical structures which make the whole telescope), so the increase in sensitivity is large. These results were obtained with a rough optimization of the reconstruction code. A more complete optimization is underway and we expect further improvements in the detector sensitivity. Fig. 5. The newly developed power supply and voltage divider for the 4-anode PMT.
the light guide with optical glue. A comparison of our measurements is summarized in Fig. 4. 2.3. Power supply In a deep sea neutrino telescope, high voltage is generated directly inside the pressure resistant sphere. For commonly used PMTs, such as Hamamatsu R7081-2, commercial high voltage generators and supply exist, produced, for example, by ISEG Spezialelektronik GmbH. A dedicated power supply was developed and built by Genova division for the 4-anode PMT, it is shown in Fig. 5.
4. Perspectives and acknowledgments A prototype for a directional OM for deep sea neutrino telescopes is being developed in the INFN Sezione di Genova, within the framework of the NEMO Collaboration, as an R&D for KM3NeT. The realization of two fully instrumented OMs is underway, and these OMs will be installed in the Capo Passero site. Reconstruction code optimization is underway, and preliminary results are very encouraging. An increase of the overall detector sensitivity of a factor of two is predicted by Monte Carlo simulations at low energies. I am grateful to Katia Fratini, Mauro Taiuti and Vladimir Kulikovsky for the valuable help. References
3. Final performances comparison The performances of a directional OM were shown in the previous sections: what is more interesting, however, is a comparison between a cubic kilometer detector, instrumented with traditional OMs and one instrumented with directional OMs. Several simulations were done and are underway to determine the improvement which can be achieved using these directional OMs as shown in Fig. 6. The effective area of a cubic kilometer detector increases up to more than a factor of two at low energies (100 GeV) and reaches the ‘‘standard value’’ at energies of the order of 100 TeV. In fact, at high energies the track emits light which can be collected by several towers; so no improvement is possible, since the effective area of the detector is almost equivalent to the geometrical area of the whole detector; on the
[1] I. Amore, [NEMO Collaboration], et al., Int. J. Mod. Phys. A 22 (2007) 3509 arXiv:0709.3991 [astro-ph]. [2] U.F. Katz, Nucl. Instr. and Meth. A 567 (2006) 457 arXiv:astro-ph/0606068. [3] A. Roberts, in: Honolulu 1980, Proceedings, Dumand Signal Processing, pp. 1–5. [4] P. Amram, [ANTARES Collaboration], et al., Nucl. Instr. and Meth. A 484 (2002) 369 arXiv:astro-ph/0112172. [5] E.G. Anassontzis, [NESTOR Collaboration], et al., Nucl. Instr. and Meth. A 479 (2002) 439. [6] K. Hanson, O. Tarasova, [IceCube Collaboration], Nucl. Instr. and Meth. A 567 (2006) 214. [7] R.I. Bagduev, et al., Nucl. Instr. and Meth. A 420 (1999) 138 arXiv:astro-ph/ 9903347. [8] E. Aslanides, [ANTARES Collaboration], et al., arXiv:astro-ph/9907432. [9] P. Kooijman, Nucl. Instr. and Meth. A 567 (2006) 508. [10] A. Braem, C. Joram, J. Seguinot, P. Lavoute, C. Moussant, Nucl. Instr. and Meth. A 570 (2007) 467. [11] HiFi Industrial Film, UK hhttp://www.hififilm.comi. [12] 3M hhttp://www.3m.comi.
Contents lists available at ScienceDirect
Nuclear Instruments and Methods in Physics Research A journal homepage: www.elsevier.com/locate/nima
A new direction sensitive Optical Module to be employed in deep sea neutrino telescopes Andrea Bersani Istituto Nazionale di Fisica Nucleare, Sezione di Genova, Via Dodecaneso 33, 16146 Genova, Italy
On behalf of the NEMO Collaboration a r t i c l e in fo
abstract
Available online 8 July 2008
The NEMO (NEutrino Mediterranean Observatory, [I. Amore, [NEMO Collaboration], et al., Int. J. Mod. Phys. A 22 (2007) 3509, arXiv:0709.3991 [astro-ph]]) project studies, within the KM3NeT framework [U.F. Katz, Nucl. Instr. and Meth. A 567 (2006) 457, arXiv:astro-ph/0606068], new technologies for a 3 km -scale neutrino telescope in the Mediterranean Sea. The telescope goal is the investigation of the high energy component of the cosmic neutrino spectrum. The detection of these neutrinos is a promising tool for a better understanding of the mechanisms that originate the extreme energy cosmic rays. Neutrino energy and direction are reconstructed collecting the Cherenkov light produced in water by the muon coming from a neutrino interaction. Two prototypes of a new large area (10 in.) 4-anode photomultiplier, and manufactured by Hamamatsu, on request of the NEMO Collaboration, will be used for the first time to detect the direction of the detected Cherenkov light at the NEMO Capo Passero site. & 2008 Elsevier B.V. All rights reserved.
Keywords: Cosmic rays Neutrino telescope Photon detector
1. Conventional vs. directional Optical Modules Deep sea neutrino telescopes were designed since the late 1970s, with a common structure: an array of vertical structures featuring a series of layers of light sensitive elements, usually called Optical Modules (OM). After the pioneering work of Dumand Collaboration [3], new techniques were developed both for sea water and for polar ice detectors [1–7]. All the developed Optical Modules have a very similar structure. A pressure resistant glass sphere hosts a large emispherical photomultiplier, optically coupled to the sphere using a layer of optical gel or glue. The power supply and front-end electronics are usually included in the sphere and directly connected to the PMT. This setup is well proven, and allows quite a good track reconstruction if the number of hits is sufficiently high, i.e. 10 or more. This means that the track length must be sufficiently large. This is due to the limited density of sensitive elements which can be installed in such kind of detector. A steep reduction of the efficiency under 1 TeV is the observed effect for a cubic kilometer scale detector, instrumented with 104 OMs. The hit multiplicity is partially recovered grouping several OMs in each layer: this is realized, as an example, in ANTARES with 3 OMs looking in different directions and placed at a small
E-mail address: [email protected] 0168-9002/$ - see front matter & 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.nima.2008.07.010
distance one from the other [8]. Nevertheless, the number of OMs is limited by costs and failure rate considerations, and, for low energy tracks, the poor definition of the track position w.r.t. the OM itself, is a strong limitation to the reconstruction capability for the detector as a whole. An important increase in the hit multiplicity can be achieved if in any OM a certain number of sensitive elements is installed. A proposed solution [9] features 12 3 in. PMTs: this means an active area comparable with the 10 in. PMT, but a complete redefinition of readout electronics should be needed. Another solution is under study at CERN [10], involving an HPD with a pixelized anode instead of the PMT as the active element of the OM. We decided instead to ask industry to develop a 4-anode, 10 in. PMT, which has, at least, the same photocathode area as a traditional, equivalent radius PMT. This solution has not the same granularity, but is compatible with the present readout chain with minor improvements. Such a kind of PMT can be coupled to a properly designed light guide, divided in four sectors, with mirrors, which can relate the incoming light direction to the photocathode quadrant which is hit. This is the second main improvement introduced for this OM: the angular acceptance of each quadrant is smaller than that of a single anode PMT. Therefore theposition of the track w.r.t. the OM is measured much better. This improves the track reconstruction. The arrangement of the PMT and the light guide in the glass sphere is shown in Fig. 1.
ARTICLE IN PRESS 68
A. Bersani / Nuclear Instruments and Methods in Physics Research A 595 (2008) 67–69
Table 1 Performances of the four anodic prototype compared to Hamamatsu R7081
Nominal voltage (v) Gain Peak to valley ratio Dark noise (thr. 0.3 pe) (Hz) Transit time spread (ns)
Requirement
R7081
Prototype
o2000
1340
1550
5 107 42 o10 000 o4
5 107 2.8 910 3
5 107 3 1200 4
Q1
22026
20 13359 40
2. Directional OM realization The realization of a directional OM requires the development of completely novel technologies, including:
a 4-anode PMT, the light guides, power supply and voltage divider, frontend electronics.
2.1. The 4-anode photomultiplier
4914
60 Theta (deg)
Fig. 1. A section of a directional OM: if compared to a traditional one, the thick light guide with mirrors is the main difference.
8103
2980 80
1808 1096
100
665 120 403 140
244 148
160 20
40
60
80 100 Phi (deg)
120
140
160
Fig. 2. First anode response as a function of the position of the light source on the photocathode surface (polar coordinates).
The 4-anode PMT was realized by Hamamatsu, starting from their former experience in large area PMTs. Two samples were delivered to us, after some measurement made at factory. The angular acceptance, compared to the one of a traditional PMT, shows a more constant behavior as a function of the incidence angle, with a steep decrease at angles greater than 801: this is mainly due to the larger surface of the first dynodes. A series of measurement were performed at the testing facility of the NEMO group at the INFN Sezione di Catania. We studied the single photoelectron peak, the transit time spread, the gain and the cross-talk of the prototype, to have a complete characterization and make feasible a comparison with previous models. A summary of these measurements is shown in Table 1. Cross-talk between the four quadrants, due to focusing of the photoelectrons on the wrong dynode, is very small, of the order of few %. A map of the photocathode response varying the position of a light source on the PMT surface is shown in Fig. 2. 2.2. The light guide
Fig. 3. A photograph of the light guide, before the installation in the pressureresistant sphere. One quarter is removed from the working position and the ‘‘focusing’’ behavior is evident if looking to the front sector.
A proper plexiglas light guide was realized to couple optically the multianode PMT to the pressure resistant sphere. A cage of mu-metal to screen the Earth magnetic field is foreseen between the light guide and the sphere. The same optical glue used for the NEMO OMs is used to keep the guide in position. The thickness of the guide was optimized to have the maximum separation between the four quadrants. This is achieved with a light guide with thickness comparable with the transverse dimension of each quadrant. In fact, the total effective area of the PMT is almost unchanged, due to Liouville’s theorem, but a very effective focusing is achieved with a guide, acting as a
Winston cone, 10 cm thick. A photograph of the light guide is shown in Fig. 3. The lateral surfaces of the light guide must be reflective to avoid a loss of sensitivity due to dead angles. It was decided to use a very thin foil of reflective material. Several types of reflective films have been tested, some based on aluminized surfaces (for example, aluminized mylar, or HiFi [11]) and others based on multilayer sandwiches of many transparent foils with different refractive indices (for example, 3M Vikuiti ESR Film [12]). The best performances were achieved with the latter, optically coupled to
ARTICLE IN PRESS A. Bersani / Nuclear Instruments and Methods in Physics Research A 595 (2008) 67–69
2.4
3M (adhesive) 3M
2.2
HiFi Aluminum Teflon
2
Aluminated Mylar
0
10
20
30
40
50
60
70
Incidence Angle (deg) Fig. 4. Comparison of the reflectivities of different materials optically coupled to plexiglas.
Effective Areas Ratio
Reflected Power (mW)
Reflectivity - Matching: optical grease 4 3.5 3 2.5 2 1.5 1 0.5 0 -0.5 -1
69
1.8 1.6 1.4 1.2 1 2
2.5
3
3.5
4
4.5
5
5.5
6
6.5
7
Log (E/GeV) Fig. 6. Ratio of the effective areas calculated for a cubic kilometer detector with and without directional OMs: a gain of a factor 2 is predicted at low energy.
other hand, at low energies, the effective area is the sum of effective areas of several small detectors (corresponding to the single vertical structures which make the whole telescope), so the increase in sensitivity is large. These results were obtained with a rough optimization of the reconstruction code. A more complete optimization is underway and we expect further improvements in the detector sensitivity. Fig. 5. The newly developed power supply and voltage divider for the 4-anode PMT.
the light guide with optical glue. A comparison of our measurements is summarized in Fig. 4. 2.3. Power supply In a deep sea neutrino telescope, high voltage is generated directly inside the pressure resistant sphere. For commonly used PMTs, such as Hamamatsu R7081-2, commercial high voltage generators and supply exist, produced, for example, by ISEG Spezialelektronik GmbH. A dedicated power supply was developed and built by Genova division for the 4-anode PMT, it is shown in Fig. 5.
4. Perspectives and acknowledgments A prototype for a directional OM for deep sea neutrino telescopes is being developed in the INFN Sezione di Genova, within the framework of the NEMO Collaboration, as an R&D for KM3NeT. The realization of two fully instrumented OMs is underway, and these OMs will be installed in the Capo Passero site. Reconstruction code optimization is underway, and preliminary results are very encouraging. An increase of the overall detector sensitivity of a factor of two is predicted by Monte Carlo simulations at low energies. I am grateful to Katia Fratini, Mauro Taiuti and Vladimir Kulikovsky for the valuable help. References
3. Final performances comparison The performances of a directional OM were shown in the previous sections: what is more interesting, however, is a comparison between a cubic kilometer detector, instrumented with traditional OMs and one instrumented with directional OMs. Several simulations were done and are underway to determine the improvement which can be achieved using these directional OMs as shown in Fig. 6. The effective area of a cubic kilometer detector increases up to more than a factor of two at low energies (100 GeV) and reaches the ‘‘standard value’’ at energies of the order of 100 TeV. In fact, at high energies the track emits light which can be collected by several towers; so no improvement is possible, since the effective area of the detector is almost equivalent to the geometrical area of the whole detector; on the
[1] I. Amore, [NEMO Collaboration], et al., Int. J. Mod. Phys. A 22 (2007) 3509 arXiv:0709.3991 [astro-ph]. [2] U.F. Katz, Nucl. Instr. and Meth. A 567 (2006) 457 arXiv:astro-ph/0606068. [3] A. Roberts, in: Honolulu 1980, Proceedings, Dumand Signal Processing, pp. 1–5. [4] P. Amram, [ANTARES Collaboration], et al., Nucl. Instr. and Meth. A 484 (2002) 369 arXiv:astro-ph/0112172. [5] E.G. Anassontzis, [NESTOR Collaboration], et al., Nucl. Instr. and Meth. A 479 (2002) 439. [6] K. Hanson, O. Tarasova, [IceCube Collaboration], Nucl. Instr. and Meth. A 567 (2006) 214. [7] R.I. Bagduev, et al., Nucl. Instr. and Meth. A 420 (1999) 138 arXiv:astro-ph/ 9903347. [8] E. Aslanides, [ANTARES Collaboration], et al., arXiv:astro-ph/9907432. [9] P. Kooijman, Nucl. Instr. and Meth. A 567 (2006) 508. [10] A. Braem, C. Joram, J. Seguinot, P. Lavoute, C. Moussant, Nucl. Instr. and Meth. A 570 (2007) 467. [11] HiFi Industrial Film, UK hhttp://www.hififilm.comi. [12] 3M hhttp://www.3m.comi.