Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://rtm-cs.sinp.msu.ru/docs/borexino/0012030.pdf Äàòà èçìåíåíèÿ: Thu Jan 24 13:54:42 2008 Äàòà èíäåêñèðîâàíèÿ: Mon Oct 1 20:25:12 2012 Êîäèðîâêà:

SCIENCE AND TECHNOLOGY OF BOREXINO: A REAL TIME DETECTOR FOR LOW ENERGY SOLAR NEUTRINOS
Borexino Collaboration
G. Alimontik, C. Arpesellaa, H. Backb, M. Balataa, T . Beaun, G. Bellinik&, J . Benzigerq, S. Bonettik, A. Brigattik, B. Caccianigak, L. Cadonatir, F. Calapricer, G. Ceccheto, M. Cheni, A. DeBario, E. DeHaasr, H. de Kerretn, O. Donghia, M. Deutschd, F. Eliseip, A. Etenkol, F. von Feilitzschf, R. Fernholzr, R. Forda, B. Freudigerh, A. Garagiolak, C. Galbiatir, F. Gattig, S. Gazzanaa, M. Giammarchik, D. Giugnik, A. Golubchikovk, A. Gorettik, C. Griebf, C. Hagnerf, T . Hagnerf, W. Hampelh, E. Hardingr, F. Hartmannk, R. von Hentigf, H. Hessf, G. Heusserh, A. Iannir, P. Inzanik, S. Kidnerr, J . Kikoh, T . Kirstenh, G. Korgak*, G. Korschinekf, D. Krynn, V. Lagomarsinog, P. LaMarche#, M. Laubensteina, F. Loeserr, P. Lombardik, S. Magnik, S. Malvezzik, J . Maneirak**, I. Mannoc, G. Manuziog, F. Masettip, U. Mazzucatop, E. Meronik, P. Musicog, H. Nederh, M. Nefff, S. Nisia, L. Oberauerf, M. Obolenskyn, M. Pallavicinig, L. Pappk*, L. Perassok, A. Pocarr, R. Raghavanm, G. Ranuccik, W. Raua, A. Razetog, E. Resconig, T . Riedelf, A. Sabelnikovl, P. Saggesek, C. Salvog, R. Scardaonik, S. Schoenertf***, K. Schuhbeckf, H. Seidelf, T . Shuttr, H. Simgenh, A. Sonnenscheinr, O. Smirnove, A. Sotnikove, M. Skorokhvatovl, S. Sukhotinl, R. T artagliaa, G. T esterag, R. Vogelaarb, S. Vitaleg, M. Woj cikj, O. Zaimidorogae, Y. Zakharovh+. a) Gran Sasso National Laboratories (LNGS), Assergi, Italy b) Virginia Polytechnic Institute and State University, Blacksburg VA, USA c) Research Institute for Particle and Nuclear Physics, Budapest, Hungary d) Massachusetts Institute of Technology, Cambridge MA, USA e) Joint Institute for Nuclear Research, Dubna, Russia f) Technical University Munich, Garching, Germany g) Physics Department of the University and INFN, Genoa, Italy h) Max Planck Institute for Nuclear Physics, Heidelberg, Germany i) Queen's University, Kingston, Canada j) Jagellonian University, Krakow, Poland k) Physics Department of the University and INFN, Milan, Italy l) Kurchatov Institute, Moscow, Russia m) Bell Laboratories, Lucent Technologies, Murray Hill NJ, USA n) College de France, Paris, France o) Physics Department of the University and INFN, Pavia, Italy p) Chemistry Department of the University and INFN, Perugia, Italy q) Department of Chemical Engineering, Princeton University, Princeton NJ, USA r) Department of Physics, Princeton University, Princeton NJ, USA
& #

Spokesperson Project manager * On leave from (c) ** On leave from: Department of Physics, University of Lisbon, Lisbon, Portugal *** Now at (h) + Permanent address: Institute of Nuclear Research, Moscow, Russia

submitted to 'Astroparticle P hysics', Novemb er 27, 2000


2

ABSTRACT Borexino, a real-time device for low energy neutrino spectroscopy is nearing completion of construction in the underground laboratories at Gran Sasso, Italy (LNGS). The experiment's goal is the direct measurement of the flux of 7Be solar neutrinos of all flavors via neutrino-electron scattering in an ultra-pure scintillation liquid. Seeded by a series of innovations which were brought to fruition by large scale operation of a 4-ton test detector at LNGS, a new technology has been developed for Borexino. It enables sub-MeV solar neutrino spectroscopy for the first time. This paper describes the design of Borexino, the various facilities essential to its operation, its spectroscopic and background suppression capabilities and a prognosis of the impact of its results towards resolving the solar neutrino problem. Borexino will also address several other frontier questions in particle physics, astrophysics and geophysics.

1. T h e New S o l a r Neu t ri n o Pr ob l e m In 1968, Ray Davis pioneered an ex periment in the Homestake mine (USA) to look directly into the center of the S un and to probe the thermonuclear fusion reactions that ge nerate solar energ y by observing neutrinos ( ) from t h ese react i ons 1 . Not only di d he accom p l i s h t h i s go al but he al so di scovered a si g n i f i cant and unex p l a i n ed shortfa ll of the sola r flux compared to astrophysical predictions. Since then, the science and technology of solar de te c tion 2,3 as well as the sophistication of models of the sola r inte rior 4,5 and their predictions of flux es have advanced remarkably. The first ge neration detectors Kamiokande 6 (based on -e scattering ) and Gallex 7 and Sa ge 8 (based on e c a p ture in 71 Ga) t h at fol l o wed t h e Hom e st ake det ect or 9 (based on e c a p ture in 37 Cl) have ex tended observations over the full spectrum of solar 's . The flux de fic its pe rsist ove r the e n tire spe c t ra l ra ng e a nd still re ma in a puzzle . On the other hand, advances in helioseismolog y leave little room for substantial departures from the Standard Solar Model (SSM) and its predictions of flux es. 5 Consequently the basic thrust of these data - the new solar problem - is now focused more sharply on the itself and on its non-standard properties, such as nonze ro ma ss, fla vor mix ing10 , m a g n et i c m o m e nt , decay, and ot hers. A non-standard implies physics beyond the Standard Model of elementary particles. F o r this reason, the second g e neration solar neutrino ex periments are now a t the c u tting e d ge of pa rtic le physic s : Supe rka mioka nde (SK) 11 , the Sudbury Neutrino Observatory (SNO) 12 , and B o rex i no. F u rther, the focus is shifting to the low e n e r gy pa rt of the sola r ne utrino spe c t rum < 1 Me V. In re a l time , this ra ng e is accessible ex clusively to Borex i no. The Borex i no detector is now nearing completion of construction. In particular, it will ex plore the 7 B e neutrinos. The pa ra digm shift in this fie l d wa s g e ne ra te d from the inc o mpa tibility be twe e n the results of Gallex / Sag e on the one hand and Kamiokande/SK on the other. These e x pe rime nts broug ht ne w se nsitivitie s for sola r detection. The g a llium ex periments se t a low e n e r gy thre shold to obse r ve the inte g r a t e d flux from the proton-proton (pp) reactions, from 7 Be-decay, and from other reactions in the S un. Kamiokande/S K ha ve yi e l de d re a l -time da ta on the hig h e n e r g y flux from a sing le solar source, t h e decay of 8 B . The sign al observed in Ga, 50% of t h e ex pect ed m a gn i t ude 13 , is practically ex hausted by the flux from the p+p reactions, a 'm ust-see' flux nearly independent of solar models if neutrinos are mass-less particles as described by the


3

standard model of electro-weak interactions. On this basis, the Ga sig n als would imply a ne a r tota l a b se nc e of 7 B e in the Sun. The compelling consequence is a near absence of 8 B as wel l , si nce t h at can be g e nerat e d onl y from t h e react i on 7 Be+p 8 B. In st ark cont rast , a subst a nt i a l fl ux from decayi n g 8 B (e ve n if with a de fic it of 50%) is observed by Kamiokande/SK. The sig n al rates at low and hig h energ i es are thus internally inconsistent with the sequential logi c of energy production in the Sun, with or without invoking modification of the SSM. 14,15,16 This is re fe rre d to a s the 7 Be- 8 B problem. The validity of the SSM is strong ly supported by helioseismolog i cal observations. 5 The m o st l i k el y ex pl anat i on for t h e m i ssi ng flux lies in non-standard physics and, in particular, in non-z e ro m a sses t h at enabl e flavor conversion. The main observable consequence is an energy dependence of the flux deficits. If this were observed ex perimentally, it would unambig uously prove neutrino flavor oscillations. Indeed, several physical scenarios of flavor conversion fit all presently a v a ila ble e x pe rime nta l da ta a nd still le a v e the sola r mode ls inta c t . With the s e developments, a sharpened solar -problem sets well-defined directives to the nex t ge neration ex periments. The key questions are: W h at is the flux at the Earth from 7 B e in the Sun? and: How can one demonstrate solar flavor conversion directly? Bo rex i n o 17 i s desi g n ed t o m easure t h e 7 Be- induc e d inte ra c tion ra te a nd to answer the central questions phrased above. Early laboratory research on radiopuritie s of sc intilla tion liquids 18 ga ve initial hints that the backg r ound from natural ra dioa c tivity a t low e n e r gi e s ma y be suffic i e n tly c ontrolla ble to ma ke re a l -time spectroscopy of sub-MeV solar 's fe a s ible . Ex plic it de monstra tion of tha t possibility ha d the n to c o me from the 4-ton te st de te c t or CTF (Counting Te st F a c ility). The pl an of t h i s paper i s as fol l o ws: In S ec. 2, we sket ch t h e physi cal foundations for solar emission, flavor conversion, the likelihood rang es for nonstandard parameters, basic aspects of -e - scattering , and the sign al detection process in B o rex i no. As the first real-time ex periment for low energy solar spectroscopy, B o rex i no is based on a new technolog y, the foundations of which are briefly covered in Sec. 3. Nex t we describe the B o rex i no detector and several a n c illa ry fa c ilitie s a s we ll a s the ope ra tiona l de ta ils of the de te c t or (Se c . 4). The response of B o rex i no to various non-standard scenari o s i s t h en t r eat ed (S ec. 5). W e conclude with remarks on the outlook for the impact of B o rex i no on critical topics in non-solar science in particular (Sec. 6) and in non-standard phenomenology in ge neral (Sec. 7). 2. Ph ysic al Fo undat i ons 2.1 The Solar Neutrino Spectrum The basis of solar energy production is the fusion of 4 protons to form an -pa r tic le , a process that releases 26.73 MeV of energ y. The 'P P chain' of reactions that achieves this end (Fig .1) is initiated by the 'pp' reaction p + p d + e + + e . The pp and its variant 'p ep' reaction, p + e - + p d + e are dri v en by t h e weak i n t e ract i on and control the slow rate of hydrog en burning in the Sun to last altog e ther > 9 Gyr. The ne x t ste p s to the te rmina tion to -part i c l e s occur vi a 3 branches, P P I (86%), P P II ( 1 4 % ) a n d P P III ( 1.6 â 10 -4 ). Neutrinos are emitted in each of these branches. The flux es from the pp and pep reactions in PPI are considered 's tandard candle'


4

flux es since, as initiating reactions, they are relatively independent of the rest of the chain in a quasi-static solar interior and thus, they do not strong ly depend on details of the SSM. The flux es are basically determined by the observed solar luminosity and by t h e m ean energ y rel ease i n t h e 4p conversion. Contrary to this, the e fl ux es from t h e el ect ron-capt u re decay of 7 B e in PPII a nd e s pe c i a lly from the + decay of 8 B i n P P III d e p e n d s t r o n gl y o n t h e d e t a i l s o f t h e s o l a r m o d e l , p a r t i c u l a r l y, t h e cent r al t e m p erat ure i n t h e S un. 19 The CNO c yc l e , while le ss importa nt tha n the PP cycle, still adds non-neg lig ible contributions of flux es from the + -decays of 13 N, 15 O and 17 F. The Sun thus e mits a ric h spe c t rum of ne utrinos whose me a s ure m e n t offe rs a complete and detailed probe of the PP chain. The most abundant component is the pp- continuum at (0-0.42) MeV. The nex t feature is the mono-energe tic line at 0.862 MeV from 7 B e (which also produces a weak line at 0.384 MeV under the ppcontinuum) followed by another line at 1.44 MeV from the 'p ep' reaction. Underlyi ng these features are weak continua from the CNO reactions (0-1.7) MeV. This low energy part of t h e spect rum covers t h e dom i n ant P P I and P P II l i nks. The decay of 8 B i n P P III c o n t r i b u t e s t h e s m a l l e s t f l u x , 10 -4 of the pp flux , in a continuum up to 14 MeV. F i nally, 's wi t h hi g h er energ i es up t o 20 Me V ma y be e mitte d from the poorly characteriz ed 3 He + p ('hep') reaction. It is clear from F i g.1 that the key link i n t h e P P chai n i s P P II, a l i nk accessi bl e onl y vi a 7 Be- 's . Thus, B o rex i no is a key probe of the roles of astrophysics and physics in the sola r problem. The S S M fl ux es from the various e sources are g i ven in Table 1 (see F i g . 2 for the solar spect rum 20 ). Unc e r ta intie s in the s e pre d ic tions de pe nd on the sta g e of t h e react i on i n t h e P P - chai n and on t h e associ at ed nucl ear cross-sect i ons. The (1 ) e rror e s tima te s of the the o re tic a l flux e s a r e (pp) 1%, ( 7 Be) 9% and ( 8 B) 20%. The re la tive l y la rge ( 8 B) ari s es from t h e acut e dependence on t h e sola r te mpe r a t ure (~T 20 ) and from the low precision ( 20%) of t h e m easured crosssection of the production reaction 7 Be(p , ) 8 B. 2.2 Neutrino Flavor Conversion A direct consequence of a non-deg e nerate mass spectrum of neutrinos and flavor mix ing is flavor conversion. 21 During transport in matter or vacuum, solar 's t h a t we re inhe re ntly e mitte d a s pure e can be converted spontaneously to other flavors, µ or . The l a t t e r are ei t h er not det ect ed at al l (as i n charg e d current based detectors), or observed with reduced (factor 5) cross-sections (as in neutral current based detection). F l avor conversion can thus create solar sig n a l de fic its in a na tura l way. In a two flavor model with a 2 â 2 matrix that transforms a doublet of mass e i ge nsta te s 1 and 2 with ma sse s m 1 and m 2 t o fl avor st at es e and x via the mix ing angle , the survival probability of the e-flavor can be written in terms of m2 = m12- m 2 2 and sin 2 (2 ). P a tterns of e-flavor survival vs. energy are shown in F i gs . 3 and 4 for two distinct physical scenarios. In vacuum, conversion manifests itself as flavor


5

Table 1: Predicted fluxes arriving at the Earth and their theoretical uncertainties5 for the various solar sources.

Source p-p pep hep
7

Energy [MeV] 0.42 1.44 18.8 0.86 (90%), 0.38 (10%) 15 1.20 1.70

Flux [106 cm-2 s-1] 59 400 ± 600 139 ± 1.4 0.0021 4 800 ± 430 5.15 + 0.98 - 0.72 605 532
+ 115 - 79 + 117 - 80

Be B N O

8

13 15


6

Figure 1: The nuclear reaction sequence of the pp-chain inside the Sun


7
20

Figure 2: The solar neutrino spectrum, from Ref.


8

osc illa tion with a wa ve le ng th (osc illa tion le ngth) = 4 E/ m 2 . F i g. 3 shows the inte re sting re g ime whe n R, the Ea rth-Sun ba se line - the 'j ust-so' osc illa tions. 22 Conversion can also occur as a resonant process in solar matter (the MSW effect ). 23 This flavor conversion is energ y dependent, especially pronounced at lowe r energi es (see F i gu re 4). The effect offers a convincing ex planation for the apparent paradox of the 7 B e - and 8 B- sig n a l s. F i g u re 4 shows the surviva l proba bilitie s for electron neutrinos after flavor conversion for three distinct islands of ( m 2 , ) that happen to offer solutions that are consistent with all ex perimental solar neutrino data acquired so far ('SMA', 'LMA', 'LOW ') . 24 The basic observable of flavor conversion of solar 's is the e -fla vor surviva l proba bility, me a s ure d by a flux 'd isa ppe a r a n c e ' re la tive to the SSM va lue s for ( 7 Be, 8 B ) . 'A ppearance effects' are also possible as direct proofs of conversion. The energy dependence creat es charact eri s t i c di st ort i ons of t h e st andard weak-i nt eract i on spect ral shapes of continua, in particular that from 8 B. Ti m e vari at i ons of t h e si gn al are observabl e i n cert a i n scenari o s. In vacuum , t h e pha se of the fla vor osc illa tion c h a n g e s slowly a s the Ea rth-Sun ba se line va rie s in the eccentric Earth orbit, but nevertheless this produces larg e chang e s in the flavor survival of a mono-energe tic line <1 MeV, such as that from 7 B e . This re sults in a seasonal signal variation 25 (in a ddition to the sma ll [ 7%] variation ex pected just from the R -2 e ffe c t ). Va ria b ility e v e n on a day/night basis i s possi bl e (especi al l y for the 7 Be line flux ) because the converted flavor arriving at the Earth can be reconverted to e-flavor in passing throug h the Earth matter, thus enhancing the nig h t time sign a l 26 (see Fi g . 4). Al l t h ese effect s on t h e 7 Be- l i n e and on t h e l o w energ y part of the 8 B spectrum are observable in B o rex i no (Sec. 5). Matter and vacuum conversion occur in distinct reg i ons of the m 2 -sin 2 (2 ) ma p. Va c uum osc illa tions re quire la rg e mix ing [sin 2 (2 ) 1] tha t se ts the osc illa tion amplitudes; and tiny mass differences, m 2 10 - 11 to 10 -9 tha t re sult in e x pe rime nta lly interesting oscillation periods. Matter conversion dominates for m 2 from 10 -8 to 10 -4 (eV/c 2 ) 2 e v e n for sma ll mix ing , with sin 2 (2 ) ra ng ing from ve ry sma ll mix ing angl es ( 10 -3 ) to the ma x imum va lue of 1. Solar ex periments thus probe larg e param e t e r spaces i n key reg i m e s not accessi bl e t o any ot her approach. Gl obal anal ys i s of t h e dat a from al l t h e present ex peri m e nt s al ready const r ai ns the pa ra me te rs to a fe w isla nds in the m 2 -sin 2 (2 ) maps 27 for the MS W effect (F ig. 5) a nd for va c uum osc illa tions (F ig . 6). The e -fla vor surviva l proba bility plots in F i gu re s 3 a nd 4 re fe r to just the s e isla nds in the param e t e r space. 2.3 Neutrino Electron Scattering Solar neutrino detection in B o rex i no is based on -e - scattering using the well-known technique of liquid scintillation (LS) spectroscopy. Scintillation photons ge nerated in a l a rge m a ss of arom at i c orga ni c l i qui d are det ect ed. They yi el d a m easure of t h e energy and of the spatial positions of the events in the detector. Photon emission in the LS is isotropic , thus, the inhe re nt dire c tiona lity of the -e scattering process cannot be utilized for deriving the directionality of the source.


9

Figure 3: Flavor survival vs. energy of neutrinos in vacuum oscillations for m2 = 6.5â10-11(eV/c2)2 and sin2(2) = 0.75.


10

Figure 4: The e survival probability due to neutrino flavor conversion, calculated for the three different MSW solutions: SMA (Small Mixing Angle), LMA (Large Mixing Angle), and LOW (Low probability, low mass). From Ref.24 For updated explicit best fit parameters see caption of Fig. 5. At night time, flavor regeneration in the Earth can enhance the rate. Shown are yearly average survival probabilities: total (solid lines); day-time (dotted lines); night-time (dashed lines). The 7Be- line at 862 keV is almost completely suppressed in the SMA solution, as indicated by experimental data.


11

Figure 5: Plot of the allowed MSW solutions (99% C.L.) in the m2-sin2(2) parameter space as deduced from the results of the Homestake, Superkamiokande, Gallex and Sage experiments. The best fit parameters are: SMA (Small Mixing Angle), with m2 = 5.4â10-6(eV/c2)2 and sin2(2) = 5.5â10-3 LMA (Large Mixing Angle), with m2 = 1.8â10-5(eV/c2)2 and sin2(2) = 0.76 LOW (Low probability, low mass), with m2 = 7.9â10-8(eV/c2)2 and sin2(2) = 0.96. Only the total event rates have been used in the calculations.27 Bahcall and Pinsonneault 1998 = Ref.5


12

Figure 6: Plot of the allowed solutions in the m2-sin2(2) parameter space for vacuum oscillations between active neutrinos as deduced from the results of the Homestake, Superkamiokande, Gallex and Sage experiments. Only the total event rates have been used in the calculations.27 Bahcall and Pinsonneault 1998 = Ref.5


13

Neutrino-electron scattering is a purely weak process with a precisely known crosssection. 28 F o r the mono-energ e tic 7 Be- line at 0.862 MeV, the profile of the recoil electrons is a unique 'f lat box ' with a spectral (Compton) edg e at 0.66 MeV, de te rmine d by the line energy. The recoi l profi l e and i t s edg e energ y are t hus spect roscopi c si g n at ures for t h e 7 Be- sig n al in B o rex i no. In principle, the shape of the profile is slig htly diffe re nt for e and µ . Unfortunately, the difference is probably too small to be useful in practice. Even thoug h the recoil profile ex tends from ze ro to the e d ge , in pra c tic e the optimum 7 Be- sig n al window in B o rex i no is (0.25-0.8) MeV, set mainly by backg r ound considerations ( 14 C). In this window, the nomina l sc a tte ring c r oss-se c tion is 5 â 10 -4 4 cm 2 (at 1 MeV). Thus, for a 100-ton detector and ( 7 Be) as i n Tabl e 1, an event rat e of 18,000/yr is ex pected, compared with a typic a l 100/yr rate in the first g e neration ex periments. B o rex i no is thus a high rate solar detector. Hard Core Signal Limit: The e -e scat t e ri ng i s dri v en by t h e charg e d (C C ) and neut ral (NC ) weak current s, whereas µ and scat t e ri ng occurs onl y vi a t h e NC . The effective cross-section thus depends on the flavor. In the 7 Be- sign al window, the e v e n t ra te due to the NC inte ra c tion a l one is 20% of the tota l ra te (the NC/CC si gn al fract i on at t h ese l o w energ i es i s 50% hi g h er t h an at 8 B- energ i es). Thus, even for complete conversion of the 7 Be- e (l i k el y accordi n g t o t h e Ga resul t s ), B o rex i no will record a pure NC sig n al from just the converted µ, . The NC component thus sets a substantial 'h ard core' lowe r limit on the sola r sign a l in B o rex i no and effectively makes the ex periment se nsitiv e to all flav o rs . Sterile Neutrinos: The discussion so far tacitly assumes conversion to normal 'act i v e' 's . In principle, conversion is possible to 's terile' species i.e., neutrinos of the wrong he lic ity with inte ra c tion stre ng ths re duc e d to va nishing le ve ls by the fa c t or (m /E ) 2 , so that they are effectively decoupled from weak processes. In this case, for complete conversion, the NC scattering is absent, thus the rate could drop even be low the 'h a r d-c o re ' limit of the -e sign a l for a c tive ne utrinos, a ll the wa y to null l e vel s . S u ch a resul t coul d be t h e si g n at ure for t h e presence of st eri l e 's . In fact, this sign ature is unique to 7 Be s o l a r 's because: (i ) A 7 Be- si gn al bel o w S S M -l evel s by a fact or 4 cannot be caused by astrophysical reasons, that being ruled out by the conflict in the 8 B- 7 B e sig n als. 27 (ii) The gl obal results of present ex periments cannot accommodate a reduction by a fact or >3 for any ot her feat ure i n t h e spect rum ex cept t h e 7 B e -line . 27 Thus a detector based on -e - scat t e ri ng speci fi c t o t h e 7 B e flux , such as B o rex i no, has access to the window below the hard core sig n al limit for a unique ide n tific ation of ste r ile 's if the substantially more string ent backg r ound requirements can be met.


3. Te c hnologic a l Fo undat i ons The el ect ron recoi l si gn al i n Borex i no from 7 Be- 's occurs in the energ y window (0-0.66) MeV at a rate of 0.1-0.5 events per day and per ton of targ et material. Obse rva tion of suc h ra re low e n e r g y e v e n ts ha s ne ve r be e n a tte mpte d in re a l time because of the large un-shieldable backg r ound of the detector medium itself, arising from ubiquitous na tura lly ra dioa c tive c onta mina n ts suc h a s 238 U and 232 Th. Radiations from these (and other) activities produce events at vastly hig h er rates in


14

the e n tire ra ng e up to 5 Me V e v e n in so-c a lle d hig h purity ma te ria l s c e r tifie d to ppt (i.e . pa rts-pe r-trillion) U/Th c onte n t. The only re a l -time obse r va tions of sola r 's todate have therefore been made in energ y reg imes above 6 MeV. 238 U g e nerates a backg r ound rate of 10 5 Hz/g U. With this ra te , a sig n a l/noise 1 implie s a c onta mina tion limit of 10 -1 6 g U /g . At the start of planning for B o rex i no, suc h puritie s ha d ne ve r be e n de monstra t e d e v e n in la bora t ory sa mple s, le t a l one on the multi-100ton scale. No chemical insig h t was available to assess whether such puritie s a r e re a liza b le e v e n the o re tic a lly, no c h e mic a l proc e ss ha s be e n de monstra t e d to achieve purification to such levels, nor were any analytical methods known for determining such ultra-trace amounts. The high est purity ye t observed, in the case of wa te r, fe ll short by orde rs of ma g n itude . While it wa s sug g e ste d tha t liquids le ss polar than water mig ht be further purifiable, no measured data either precluded or promised the above purity levels. B ackg r ound from ex ternal sources is also severe, thoug h removable in principle by massive shielding of the detection volume. In this respect, however, JRRG SUHFHGHQWV Z HUH DYDLODEOH IURP ZDWHU ÙHUHQNRY GHWHFWRUV VXFK D V .DPLRNDQGH They showed that the ex ternal backgr ound can be suppressed by defining a fiducial detection volume in the core of a massive tank of pure water larg e enoug h to provide a shie ld thic kne ss of 5m. To adopt this g e neral approach to low energ y VSHFWURVFRS\ VLJQDO O XPLQRVLW\ K LJKHU WKDQ W KDW RI W KH ÙHUHQNRY HIIHFW L V UHTXLUHG $ natural choice is the well known technique of liquid scintillation that offers 50 time s la rg e r sig n a l luminosity a nd a llows e v e n t se nsitivity down to 50 keV. Such a strategy still involves two open problems: (i ) Containment of the LS in the shield liquid (buffer) sets severe constraints. (ii) Althoug h widely used in many areas in physics and biosciences, liquid scintillation had never been applied on the multi-100ton scale for spectroscopy at 50 keV. Thus, a new technolog y had to be developed for observing low energ y solar 's in re a l -time . 3.1 Radiopurity 3.1.1 Chemistry The sources of radioactivity in the B o rex i no LS stem from: ( i ) wi del y preval ent primordial radioac tiv itie s: 40 K, 238 U and 232 Th and t h ei r decay products 226 Ra and 210 Pb; ( ii ) am bi ent noble gas activities: 222 Rn (radon) and 85 Kr; and ( iii ) c o smoge nic ac tiv itie s: 14 C, 7 B e and 3 H (tritium). 14 C i s a speci al case t h at chal l e ng es t h e cent r al concept of Borex i no because i t s backgr ound critically interferes with the sig n al and it is inseparable from the orga nic liquid sc intilla tor. All the othe r impuritie s a r e c h e mic a lly se pa ra ble , the que stion being the degr ee of removal. They can be g r ouped chemically as metals (U, Th, K, Ra, Pb, and B e ) and noble ga ses (Rn and Kr). Tritium is not critical because of its l o w decay energy. Radiocarbon: The -spectrum of 14 C ( E ma x =0.156 MeV) overlaps the sig n al window of pp- 's (0-0.25) MeV and, because of the limited energy resolution in liquid sc intilla tors, se riously ta ils into the sig n a l profile of 7 Be- 's (0-0.66) MeV. At the start of B o rex i no, the most sensitive limit of 14 C in a n y ma te ria l wa s


15

C/ 12 C < 1 0 -1 5 which produces a 14 C r a te 2 â 10 7 /d,ton in an org a nic liquid, ruling out de te c t a b ility of the muc h we a k e r sig n a l s from pp- 's ( 2/d,ton) and from 7 Be- 's ( 0.5/d,ton). The latter can still be usefully measured in a window above 0.25 MeV if 14 C/ 12 C i s at l east as l o w as 10 -1 8 ( 10 -6 of the 14 C in modern carbon). In the B o rex i no framework, this leaves no option but an org a nic material naturally 'free' of 14 C . The onl y m eans t o approach t h i s g o al i s t h e st ri ct use of petrochemical org a nics. The underg round residence of such material for millions of years removes the origina l 14 C, leaving, at worst, only small amounts of fresh 14 C from possible neutron reactions underg round. These premises were encourag ed by tests on isotopically enri ched nat u ral g a s, 29 sel ect ed as a pet r ochem i cal prox y. Measured rat i o s 14 C/ 12 C <10 -1 8 se t the sta g e for te sts of the a c t ua l sc intilla tor pe troc he mic a l s in the CTF (se e Sec. 3.2.3). Me tallic Radioimpuritie s: The purity reg imes of Borex i no are unprecedented (e.g., [ K ] 10 -1 4 g/ g, [ U , T h ] 10 -1 6 g/ g, [ R a ] 10 -2 2 g/ g, [ 7 Be] 10 -2 6 g / g ) and i n accessi bl e to the most sensitive analytical methods that offer at best, ppt sensitivities in favorable cases. F o r this reason we ultimately constructed a Counting Test F acility to ex plore low level backgr ounds, particularly for U and Th (see Sec. 3.2). B e fore the CTF was constructed, however, a number of basic studies were made that provided insigh t on scintillator materials and some of their backgr ound problems. The studies also sug g e sted methods for removing observed radioactive impurities, methods ultima te ly e m ploye d in the CTF . We me ntion brie fly the s e studie s on the sc intilla tor solve n t (initia lly TMB , la te r PC) a nd the wa ve le ng th shifte r (PPO). (i ) S o l v ent st udi es (TMB and P C ): Di rect m easurem ent s of U and Th i n t h e sc intilla tor solve n ts c ould a c h ie ve le ve ls of 1 ppt by mass spectrometry. Efforts at c onc e n tra ting impuritie s by a n a l ys is of re sidue s a f te r distilla tion le d to limits of 10 -1 5 g/ g for U a nd Th in tri-me thyl bora t e , a n e a r ly possibility for the sc intilla tor solvent. W h ile these studies did not reach the required sensitivities they did show that such solvents are quite pure as received, hinting at the possibility of the hig h er purities required for B o rex i no. Radioactive tracer studies were also performed in which one tracks impurities ra the r tha n me a s uring the i r a b solute c onte n t. 17,18,30 In ste a d of ta rg e ting impurity content with ina d e qua te a n a l ytic a l se nsitivitie s a nd unc e r ta in bla nk va lue s , one tra c k s impurity transfe r in processes of purification, handling and storag e. Starting from liquid sc intilla tor ma te ria l s c onta i ning U, Th, Pb, Ra a nd K a t high le ve ls typical of terrestrial abundances ( 0.1 to 1ppm) and spiked with tracers for these e l e m e n ts, ordina ry proc e sse s suc h a s distilla tion re move d the impuritie s in the distillate by about 10 orders of mag n itude. Useful efficiencies were obtained for water ex traction. In addition, tracer work led to the discovery of a new purification technique. La boratory tests showed that silica or alumina columns, normally used only for re moving optic a l impuritie s, a r e a l so powe rful de vic e s for ra diopurification. They removed impurities that had been initially in the liquid, and did not cont am i n at e t h e sol v ent s i n so doi ng . In t h e case of 7 B e with the most string e n t specifications (see above), pseudocumene (PC, chosen for B o rex i no) was activated dire c tly by protons a nd ne utrons (a s in c o smog e n ic a c tiva tion) a nd distille d, removing the 7 B e a c tivity by a t le a s t thre e orde rs of ma g n itude . 30,31 Wa te r e x tra c tion also produced useful B e purification. 31
14


16

(ii) W a velength shifter studies on PPO:[ 2,5-(C 6 H 5 ) 2 (C 3 HNO) = 2,5diphenylox a z o le] : Commercial PPO was found to contain ppm concentrations of K that were readily det ect ed by neut ron act i v at i on anal ys i s . S i nce t h e P P O i s m i x ed wi t h P C at concentrations of 10 -3 , the K impurity in the PPO would be at a level of 10 -9 in the scintillator, or about 5 orders of mag n itude hig h er than the requirement of 10 -1 4 g/ g. Purification studies showed that the K could be readily removed from PPO by a wa te r e x tra c tion proc e ss. Na me ly, the PPO is dissolve d in PC a t high c onc e n tra tion a nd the mix ture is the n vig o rously c onta c t e d with hig h purity wa te r. During mix ing, t h e K i s t r ansferred t o t h e wat e r, because of i t s hi g h er affi ni t y for m o re pol ar solutions. Afte r mix ing , the wa te r a nd the PC-PPO liquids pha se se pa ra te (the y a r e immisc i ble ) , a nd the K sta ys with the wa te r, le a v ing the PC-PPO solution purifie d. R e peat ed appl i cat i on of t h e procedure reduces t h e K cont ent i n t h e P P O t o an undetectable level. However, the detection sensitivity limit for K in the scintillator has not reached l e vel s bel o w 10 -1 2 g/ g , neither by neutron activation measurement of PPO nor by direct spectroscopic measurements in the CTF . The purity requirement for K i s t hus not di rect l y confi r m e d, but i t i s ex pect ed t o be m e t because of t h e hi g h efficiency of water ex traction for removal of K. These studies led to the development of a wa te r e x tra c tion c o lumn in the purific a tion syste m s for the CTF a nd for B o rex i no. The principal conclusion reached in the work above with several candidate sc intilla tor liquids is tha t me ta llic impuritie s ha ve a va nishing solubility in non-polar orga nic liquids. The e x tre m e puritie s c a n be ma inta ine d inde finite ly by re mova l of me ta llic impuritie s by ordina ry c h e mic a l proc e sse s suc h a s distilla tion a nd wa te r- or solid column ex traction. Li quid scintillators, in g e neral, thus offer a practical route to low energy neutrino spectroscopy. Radon and Krypton: Radioactive noble ga ses are mobile and chemically inert. They can be segr egated using their volatility but not by chemical reactions. 222 Rn (3.8d) emanates from Ra-decay in the 238 U chai n. The decay of 222 Rn in the scintillator creates most of the radiation of the U chain and in addition, leaves a deposit of its end product, 22yr 210 Pb 210 B i that produces backg r ound just in the 7 Be- sig n al window. 85 Kr (10.85yr) is prevalent in the atmosphere and also produces backgr ound in the sign al window. Thus, minimizing air leaks and Rn emanation sources i n al l part s of t h e det ect or are cl earl y m a j o r eng i neeri n g i ssues wi t h several facet s (see bel o w). 3.1.2 Analytics Active Spectroscopic Tags: Solar neutrinos are detected in B o rex i no throug h their scattered electron events, without any distinct sig n ature. However, a real-time identification is possible for several key types of backgr ound events that are caused by ra dioimpuritie s in the sc intilla tor. 17 The t a g s are hi g h l y speci fi c t ool s for backgr ound spectroscopy and play a crucial role in: (i ) the active suppression of most of the intrinsic background in the signal window; a n d (ii) in the on-line qua ntific a tion of impuritie s. Untagge d parts of the backgr ound from decay chain seg m ents associated with the de te c t e d impurity c a n be re c onstruc te d a nd sta tistic a lly c u t from the obse r ve d spe c t rum to the e x te nt tha t the y a r e in se c u la r e quilibrium with the dire c tly ide n tifie d species. The tag s /statistical cuts can suppress typically 90% of the backg r ound from U/Th and Kr-decays. The tags are of two kinds: pulse shape discrimination (PSD) of


17

-part i c l e s and del a ye d coi n ci dences (DC ) of correl a t e d decays of i m puri t y a c tivitie s. PSD: For every decay of 238 U or 232 Th, there is a sequence of 8 or 6 -pa r tic le s e mitte d, re spe c tive l y, of e n e r g y (4-9)Me V . The y produc e sc intilla tion sig n a l s quenched by factors of 10 to 15 relative to - or -signa ls, so tha t the y a ppe a r in the 0.25-0.8 MeV energ y window. Thus, a major backg r ound ag ainst the 7 Be- induced e - -signa l is from U/Th -part i c l e s. However, t h e si g n al quenchi ng m echani s m al so re sults in c h a r a c t e r istic a lly long sc intilla tion time s tha t in turn le a d to pulse sha p e s that differ from the much faster e - / pulses. PSD can tag -events with 90-99% e ffic i e n c y, de pe nding on the c o mposition of the sc intilla tor a nd on the -energy. 32 Del a yed C o i n ci dences (D C ): Key components in the decay chains of U/Th and in the -decay of 85 Kr a r e e mitte d a s time -c orre la te d c o inc i de nc e pa irs whic h c a n be ta gge d with high spe c i fic ity. 17 The - coi n ci dences i n 212 B i and 214 B i and the - pa ir in 85 Kr qua ntify the pre s e n c e of the long -live d pa re nt impuritie s. The 212 Bi t a g a ssa ys the e n tire se g m e n t following 228 Th in the 232 Th chai n whi l e 214 B i ta g s the 226 Ra ( 222 Rn) - 214 Po se g m e n t of the 238 U chai n. The Kr and 214 B i ta g s a r e pa rtic ula r ly useful in tracing the orig ins of persistent backg r ound from the g a ses, in particular, at m o spheri c l eaks of Kr and R n and i n t e rnal source m a t e ri al s t h at em anat e R n. T r ace 238 U determination: The 238 U- 234 Th- 234 Pa-decays in the initial segm ent in the U chain do not offer DC tags or conveniently detectable -pa r tic le s. Equilibrium of 238 U with the rest of its chain cannot be assumed a priori , since purification effi ci enci es are chem i cal l y speci fi c. Thus, t h e speci fi c det e rm i n at i on of U i n t h e scintillator is vital, at least sample-wise by off-line methods. The standard methods (10 -1 0 -10 -1 2 g/ g sensitivities) fall far short of B o rex i no levels. A major innovation was made to ex tend the sensitivity of U-neutron activation analys is (NAA) by >6 orders of magn itude by a new technique, IS AN (Isomer Spectroscopy of Activated Nucl ei ). 33 The method is based on detecting impurities by activating them by a mode t h at creat es nucl ear i s om ers at l east i n one of t h e i s ot opes of t h e i m puri t y. Is om eri c decays can be speci fi cal l y det ect ed by DC t a g s . IS AN can be real i z ed i n pract i ce, pa rtic ula r ly for U: the re a c tion 238 U+n 239 U 239 Np ( DC ) occurs naturally in normal neutron activation of U, needing only technical modifications of NAA count i n g for t r ace det e rm i n at i on of U. 17 The technique is now developed for standard analys is of scintillation solvents with a trace sensitivity of 10 -1 6 to 10 -1 7 g/ g. 33,34 3.1.3 Optical properties While not directly related to radioactive impurities, one should note the importance of non-ra dioa c tive impuritie s tha t c a n a ffe c t the optic a l c l a r ity of the solve n t. Suc h impuritie s, whic h a r e org a nic in na ture , re duc e the a tte nua tion le ng th of lig ht pa ssing through the sc intilla tor, c ontra ry to the re quire me nts for a la rg e liquid sc intilla tor a s in B o rex i no. 35 Degr adat i on of t h e sol v ent can occur wi t h ex posure t o ai r and ex posure to certain metal surfaces. Distillation of the solvent was found to remove the impuritie s a nd re store the optic a l qua lity of the sc intilla tor. 36


18

3.2 Counting Test Facility 3.2.1 CTF, a prototype for Borexino A 'C ounting Te st F a c ility' (CTF ) wa s c onstruc te d a nd insta lle d in Ha ll C of the Gra n Sasso La boratory. The g o als of CTF were: (i ) to a c h ie ve ultra - purity in the sc intilla tor on a ma ssive sc a l e ; (ii) to actually measure the prevalent radiopurity; (iii) to pe rform spe c t rosc opy a t ve ry low e n e r g i e s in a multi-ton de te c t or; (i v ) to develop technical solutions for a host of eng i neering issues involved in achieving the above. CTF is a large-scale test detector that is nonetheless modest in siz e relative to B o rex i no. A mass in the 4 ton rang e was set by the need to make the prevailing sc intilla tor ra dioimpuritie s me a s ura b le via DC ta gge d e v e n ts, while a wa te r shie ld thickness of 4.5m was needed to not disturb this measurement by ex ternal radiation. The primary go al of CTF was to develop solutions directly applicable to operational issue s for B o re x i no; but a t the sa me time the r e wa s the long -ra ng e g o a l of pe rforming qua lity c ontrol during B o re x i no ope ra tions. Detailed reports on the CTF have been published. 37,38,39,40 In this se c tion we outline only the ma in fe a t ure s a nd re sults. As a simplifie d sc a l e d ve rsion of the B o rex i no detector, a 4m 3 volume of liquid sc intilla tor is c onta i ne d in a 2m dia m e t e r transparent nylon vessel (inner vessel, IV ) mounted at the center of an open structure that supports 100 phototubes (PMT) 41 whic h de te c t the sc intilla tion sign a l s. The PMT's are coupled to optical concentrators viewing the IV with 20% optical coverage . The whol e syst em i s pl aced wi t h i n a cyl i ndri cal t a nk (11m â 10m) that contains 1000 tons of ultra-pure water which provides a shielding of 4m ag ai nst ne utrons origina ting from the roc k a nd a g a i nst e x te rna l -rays from the P M T's and othe r de te c t or ma te ria l s. The CTF a l so inc l ude s syste m s for wa te r a nd sc intilla tor purification that consist of units for water ex traction, vacuum distillation, silicag el c o lumn tre a tme nt, a nd for nitroge n spa r g i ng . The e n tire de te c t or is built on c l e a n room construction standards. A principal part of the CTF prog ram is the development of technical solutions for achi e vi ng l a rge scal e LS ul t r a-puri t y. Al l part s of t h e det ect or cont act i n g t h e LS inc l uding stora g e , ha ndling , purific a tion a nd de te c t or filling - me t e x a c ting sta nda rds to counter two basic modes of contamination: (i ) ga se ous ra dioa c tivity (e ma na te d Rn from Ra in ma te ria l s a nd le a k a g e of Rn, Kr and 14 CO 2 from outside); and (ii) dust a nd pa rtic ula t e s (with spe c i fic a c tivitie s 10 10 time s hig h e r tha n B o re x i no tole ra nc e s ). The forme r re quire s tha t the e n tire de te c t or a nd a n c illa ry syste m s be engi neered to hig h g a s-tig htness to rig i dly ex clude ex ternal air. The Rn backg r ound i s i t s el f rel a t i v el y short l i v ed, but deposi t e d P b -Bi daught er act i v i t y on t h e surfaces i s l ong-l i v ed. Because of t h e l a t t e r, R n cont rol i s necessary not onl y i n t h e det ect or and its operations but during materials production and assembly above gr ound. A rig i d quality control program using underg round Ge spectrometers and radon emanation tests screened all construction materials for radiopurity. The PMT g l ass was a special type with ra dioa c tivity 10 times lower than normal. The surfaces of the detector subsys tems and of the plants, especially those contacting the LS were smoothed (to 0.6-0.8 µm) and pickled/passivated or electro polished for effective cleaning of particulates.


19

3.2.2 Major CTF Systems The Liquid Sc intillator (L S): The sc intilla tor use d for the ma jor pa rt of te sts in the CTF is pseudocumene (P C), widely used because of many desirable properties, e.g . high spe c i fic sc intilla tion output ( 12000 photons/MeV), long lig ht transmission le ngths (typ ic a lly 7m) a nd c o mme rc ia l a v a ila bility in kiloton qua ntitie s. In orde r to shift the e mission wa ve le ng th to 380nm, to achieve long transmission lengths and to better match the peak of the photosensitivity of the PMT's, a fluor has to be added to the sc intilla tor. 35,40,42 The key issue to achieving both larg e transmission leng th and hig h photoelectron yi eld is auto absorption and re-emission of the flour. A research program on flours resulted in the choice of PPO at a concentration of 1.5 g/ L scintillator. Some 5 tons of the LS were transported underg round within 24-48 hours of production at the petrochemical refinery, thus minimizing cosmog enic production of 7 B e during tra n sporta tion a t se a le ve l. The PC wa s produc e d in the Enic he m pla n t at S a rroch (S ardi ni a). Test s were al so carri ed out wi t h an al t e rnat e LS sol v ent , phenyl x yl yl et hane (PXE), with c o mpa r a b le sc intilla tion a nd optic a l prope rtie s but a spe c i fic gr a v ity of 0.992, providing near-neutral buoyancy with a water shield. The flash point is 150 ° C , com p ared t o t h at of P C at 45 ° C. Sc intillator Containme n t Ve sse l (I V): The design of CTF and B o rex i no calls for a hi gh l y t r ansparent cont ai nm ent vessel for t h e LS i n t h e core of a m a ssi ve vol um e of shield liquid. Though this is simple in concept, in practice there are severe constraints on the desig n and materials. The material must be chosen for its streng th, optical properties, radiopurity and chemical compatibility with the LS and shield liquids ('buffe r' ). The me c h a n ic a l de sign must insure ve sse l inte gr ity a nd a llow feasible and cost effective construction. The key desig n innovation here is the choice of a thin me mbra ne c onta i nme n t ve sse l, ra the r tha n a rig i d she ll. 43 As one of the few c hoic e s a v a ila ble for ma x imum c o mpa tibility with the va rious c onstra i nts, thin (0.5mm) nylon film wa s se le c t e d a s the ve sse l ma te ria l . In a ddition to the ne e d for high bulk radiopurity, one must also face the reality that dust and aerosol-borne Rn daughters can be deposited on the nylon surface during film manufacture or balloon fa bric a tion with the possibility of c onta mina tion of the LS ove r time . 43 Ex trusion of the nylon as well as construction of the IV were carried out under clean room conditions, with special efforts to reduce the plateout of radon daug hters. The density of PC is 0.88g /cm 3 , we ll be low tha t of wa te r. Ne ve rthe le ss, the c onc e p t of membrane containment of the PC LS in water was found to be mechanically sound at the 4-ton level and the CTF vessel was operated safely for two years. Wa te r Purific ation: This syste m c onsists of a filling pla n t a nd a re -c irc u la tion loop. 44 Pre -filte re d (0.1 µm) water passes throug h a reverse osmosis unit, a continuous de-ioniz a tion unit and finally is sparge d by counter current flow of nitrogen in a stripping column. Sc intillator Purific ation: The CTF sc intilla tor purific a tion syste m is de sign e d to perform pre-purification of the scintillator and its components and on-line purific a tion a f te r the sc intilla tor is insta lle d in the de te c t or. Pa rtic ula r ly importa nt is the removal the impurities that are known to be present in the fluor and in the


20

quencher materials. W e have two major plants for both, 'off-line ' (pre- or batch-) purification and for 'on-line ' purific a tion 36,45 : (i ) The PPO-fluor and DMP-quencher are pre-purified in concentrated PC solutions using the solid c o lumn a nd ba tc h purific a tion pla n t. It a llows ba tc h ope ra tions for water ex traction, nitrog en deg a ssing and spargi ng, as well as ultra-purification and filtra tion by me a n s of c h e mic a l pa c k ing s . The pla n t ha s be e n g e ne ra lly te ste d during CTF ope ra tions a nd use d to pre p a r e the sc intilla tor mix ture s thus-fa r re porte d. It c a n a l so ha ndle on-line purific a tion of post-loa d e d sc intilla tor solutions a s we ll a s PC (fluor free liquid). (ii) The second plant consists of sub-micron filtration, water ex traction (of ionic species), vacuum distillation in a six plate system and nitrogen stripping (of noble ga ses such as R n and Kr, wat e r, and act i v e g a ses l i k e O 2 and CO 2 ). This pla n t ha ndle s purific a tion of PC solutions in both on a nd off-line mode s. The distilla tion subsys tem is used on PC (fluor-free) liquid and has the additional feature to remove the fluor from the scintillator. The water ex traction feature allows for continuous online wa te r e x tra c tion of the post-loa d e d sc intilla tor solutions. 3.2.3 Results The CTF ha s se t ne w mile stone s in a t le a s t two re spe c t s: (i ) It is the largest nuclear detector ever built with sensitivity down to the sub-100 ke V spe c t ra l re gime ; a nd (ii) It has the lowest backg r ound at 0.03 events/kg,keV,yr in the (0.25-2.5)MeV energy window, even lower in rate and in the energ y reg ime as the presently best detectors for double beta decay. De ta ile d re sults from the CTF a r e summa rize d e l se whe r e . 37,40 He re we outline only the results relevant to the new technolog y for B o rex i no. Signal Spectroscopy: A si gn al yi el d of 300 photoelectrons/MeV was measured in the CTF sc intilla tion de te c tion syste m , pe rmitting for the first time , nuc le a r spe c t rosc opy down to a thre shold of < 50 ke V in a multi-ton nuc le a r de te c t or. F i g . 7 shows the -spectrum of 14 C ( E ma x = 156 keV) in the reg i on (20-350) keV. Indeed, the spectrum is precise enoug h to test the weak-interaction shape of the spectrum. 38 The re la tive timing of the PMT hits a llowe d a n e v e n t position re solution of 12cm (1 ) for the 214 Po -line (at the quenched energ y of 750 keV). A PSD efficiency of 97% was achieved for 214 Po -ide ntific a tion with a loss of 2.5% of -events (F ig . 8). Ne w e ffe c t s in the tra n sport of sc intilla tion lig ht in ma ssive de te c t ors due to overlapping emission/absorption bands were identified and their effect on pulse he ight a nd timing wa s studie d . 40 Impurity Spectroscopy and Radiopurity: Events as sing les and as tag g e d by DC le d to the ide n tific a tion of the lowe st c onta mina tion le ve ls of the princ i pa l impuritie s ever recorded i n any m a t e ri al . From t h e 14 C spectrum above, the rate leads to 14 C/ 12 C = (1.94 ± 0.09) â 10 -1 8 38 , close to the desig n g o al of B o rex i no and establishing qua ntita tive l y its ba sic de sig n pre mise . The 226 Ra c onte n t in the LS wa s de te rmine d by the 214 Bi - 214 P o DC tag s which m easure t h e R n seg m ent of t h e U chai n. The 238 U c onte n t in the LS wa s de te rmine d by off-line IS AN-NAA measurements of the 238 U itse l f. The lowe st a c h ie ve d ra te for the former was 1.5 events/d in the CTF , corresponding to (3.5 ± 1.3) â 10 -1 6 g/ g o f 238 U-equivalent. This is an upper limit since part of the sig n al was due to Rn from


21

Figure 7: 14C -spectrum between 20 keV and 350 keV measured in the CTF. The endpoint energy of 14C is 156 keV.


22

Figure 8: - separation in the CTF based on pulse shape differences demonstrated with coincident pulses from 214Bi ( -continuum up to 3 MeV) and from 214Po ( , quenched to 751 keV). The parameter plotted on the ordinate is (anti)correlated to the relative proportion of charge contained in the steep initial part of the signal (first 48 ns from a total of 500 ns). For detail, see.37


23

Figure 9: Energy distribution of the events tagged as 85Kr-candidates in the CTF. The distribution is compatible with the expected -spectrum (endpoint energy 173 keV).


24

Figure 10: Detection of neutrons emitted by muon interactions. Left: Delay time distribution between the muon interaction and the neutron capture. Right: Energy spectrum of gammas after neutron capture on protons. Recognition of such events is required for the discrimination of the cosmic ray muon induced residual background.


25

the wa te r tha t migra te d through the nylon ve sse l into the LS . The Th-equivalent contamination was identified by the 212 Bi - 212 Po D C ta g s observed at the lowest rate of (0.31 ± 0.10)counts/d corresponding to (4.4 ± 1.4) â 10 -1 6 gT h/g. These values were observed in the PC as recei ved f r om t h e ref i n ery without any purification (with only the addition of 1.5g /L of purified scintillator fluors). IS AN-NAA analys is (performed at a later stag e after operating the distilla tion c yc l e ) g a ve < 2 â 10 -1 6 g 238 U/g 34 , limited by the method rather than by the sa mple 's U-concentration. Radon and Krypton: One of the most important contributions of the CTF to B o rex i no technology is the ex perience acquired on the radon problem, both in the sc intilla tor itse l f a s we ll a s e x te rna l to the sc intilla tor, ma inly in the wa te r shie ld. In fusion of radon into the scintillator occurred episodically in large scale during various operations of liquid handling as well as at persistent low levels, possibly via pe rme a tion throug h the nylon ve sse l a nd during the spa r g i ng of the sc intilla tor with large amounts of nitrogen. At the time of these ex periments, the methods used to produce R n -free ni t r ogen were not ye t suffi ci ent l y effect i v e. Meanwhi l e t h i s has be e n dra s tic a lly improve d to a re c o rd purity le ve l of 0.5 µBq / m 3 (see Sec. 4.2). The leakage of Rn in the atmospheric air could be distingu ished by detection of the accom p anyi ng 85 Kr vi a i t s DC t a g . Fi g . 9 shows t h e t a g g e d spect rum of l eakag e t r aces of 85 Kr. 46 On-Line Purific ation of the LS: All the insta lle d on-line syste m s we re ope ra te d successfully but revealed the inherent possibility of radon infusion, stressing the necessity of Rn removal by nitrog en sparg i ng following the purification process. These operations also showed the removability of accumulated Rn by-products 210 Po a nd the da ug hte r a c tivitie s of B i a nd Pb. Se ve ra l c yc l e s of wa te r e x tra c tion re move d the s e impuritie s sa tisfa c t orily (by a t le a s t â 20). 36,38 The perform ance of t h e solid column purification was evaluated in the CTF campaign with PXE (doped with ter-phenyl and bis-MS B fluors) by (now improved) off-line IS AN-NAA techniques. The re sults showe d the a c h ie va bility of ultima te ultra - purity in a n a r oma tic solve n t with the re sults:[ 238 U] <1 â 10 -1 7 g/ g a n d [ 232 Th] < 2 â 10 -1 6 g / g , com p ared t o [U ]= ( 2 . 7 ± 1.5) â 10 -1 4 g/ g and [ T h] =(3.2 ± 1.6) â 10 -1 4 g / g initia lly obse r ve d in PXE. 34 Purity of Shield Wa ter: The wat e r produced by t h e C T F pl ant achi e ved t h e de sign go a l s with the following impurity re sults a s me a s ure d by IC P ma ss spectrometry (U/Th: Is pra), by emanation techniques (Ra, Rn: MPIK Heidelberg ) and by chromatogr aphy (K: B e ll La bs): [ U ,Th] 10 -1 4 g/ g; [ K ] 10 -1 2 g/ g. T h e U , T h a n d R n values are 4 and 6 orders of magn itude respectively, lower than for the raw Gran Sasso water. Measurements of Ra (long-lived parent of Rn) ga ve <20mB q /ton initially and <2mB q/ton after a recent upg rading of the assay system. 44,47,48 Cosmic Ray Inte rac tions: In terfering backgr ound due to passage of cosmic ray muons in the cave was studied mainly by two methods: identification via the PMT 'hit pa tte rn ' of t h e event , and pul se shape anal ys i s . The dat a showed t h at as m u ch as 2% of suc h e v e n ts c ould be mista ke n for low e n e r g y e v e n ts in the sc intilla tor. 37 More information was obtained from an ex ternal muon veto detector with a relatively small coverag e which sampled the effect of muons in producing events at low energi es correlated with neutrons produced by spallation. F i g. 10 shows the characteristic 2.2 MeV -rays from proton capture of diffusing neutrons in correlation with the muon sig n al. These measurements directly determined the rate of


26

cosmic ray spallation backg r ound on-line at low energi es for the first time as 0.3/d,ton of material (see Sec . 4.4.3). The results stressed the need for identifyi ng muon sign als with an efficiency of 99% and provided the main thrust for desig n ing an efficient muon veto for B o rex i no. C T F-II: Several questions raised during operations in the CTF are being re visite d in a n upgra de d CTF - II fa c ility de sig n e d to substa ntia lly re duc e the operating sing les rates encountered in CTF - I, due mainly to a hig h radon concent r at i on i n t h e shi e l d wat e r. The chi e f upg rades are a nyl on screen bet w een t h e sc intilla tor ve sse l a nd the PMT 's and a larg er (but still partial) muon veto by upward looking PMT 's arrang ed over the entire bottom floor of the detector. CTF - II is pla nne d a s the ma in qua lity c ontrol fa c ility for the LS a nd for the ope ra ting syste m s of B o rex i no. 4. T h e B o rexi n o Detector 4 .1 Detector System s B o rex i no is being constructed in Hall C of the Gran Sasso underg round laboratory (LNGS) in It aly (F ig . 11) with an overburden of 3500 meter water equivalent (m.w.e.). The suppression of the the cosmic muon flux is roug hly six orders of magn itude, down to a value of 1.1/m 2 ,h. The archi t ect ure of t h e det ect or (Fi g . 12) basi cal l y ai m s at reduci n g t h e -ray backgr ound at the central fiducial volume (F V) to a level well below the minimum solar sign al via g r aded shielding by various thickness of increasingl y radio-pure ma te ria l s a s the F V is a pproa c h e d. The fina l la ye r of shie lding is the oute r pa rt of the scintillation liquid itself (the active buffer) contained in a membrane balloon inner vessel (IV). S o ft ware sel ect i on of event s i s rest ri ct ed t o t h e FV, creat i n g i n effect , a 'wa ll-le ss ' counting device with backg r ound limited mostly by the scintillator radioactivity. The IV is immersed in a hig h purity non-scintillating (inactive buffer) liquid contained by an outer vessel made of stainless steel (OV) which also supports the PMT a rra y. The whole a rra ng e m e n t is imme rse d in turn, in hig h -purity wa te r contained in the main tank. This g r aded shield architecture provides a total shield thickness of 5 m.w.e. which suppresses the ex ternal -ray and neutron backgr ounds by some 10 to 11 orders of magn itude. The central part of the detector is the active mass of LS . The scintillation lig ht is viewed by 2200 PMT 's. Outward looking tubes RQ WKH VXUIDFH RI W KH V WHHO V SKHUH DFW D V D PXRQ YHWR G HWHFWRU X VLQJ W KH ÙHUHQNRY light produced by muons crossing the outer water buffer. Liquid Scintillator and Inner Vessel: The choi ce for t h e LS i n Borex i no i s P C with 0.17% PPO added as in the CTF . The scintillation pulses in PC are relatively fast with a decay time of 3.5ns for -particles; however, the effective decay time is broadened to 5ns because of absorption/re-emission phenomena as the photons propag a te over long path leng ths in the detector. 40,42 The LS i s capabl e of ex cel l e nt PSD a s te ste d in the CTF . The tota l ma ss of LS is 290 tons contained in an IV 8.5m in diameter. The software defined F V diameter is nominally 6m, containing 100 tons of LS a c ting a s the sola r targ et and as detector. A larg er fiducial volume may be possible with the low ex ternal backg r ound ex pected in the present desig n. The IV , made of 0.1mm thick transparent nylon, is tethered by a system of strings. 49 The


27

Figure 11: Layout of Hall C in the LNGS.


28

Figure 12: Sketch of the Borex ino detector. About 300 tons of liquid scintillator are shielded by 1040 tons of a transparent buffer liquid. The scintillation light is viewed by 2200 PMT's. Reconstruction of the position of point-like events allows the determination of a 100 ton fiducial inner mass - the solar neutrino target. Outward looking tubes on the VWHHO VSKHUH VXUIDFH DFW DV PXRQ YHWR GHWHFWRU 7KH\ XVH WKH ÙHUHQNRY OLght produced by muons that intersect the outer water buffer. The latter serves also as a shield against external radiation.


29

string tensions will be monitored. A second nylon enclosure placed near the PMT shell minimizes inflow of radon and other impurities that may diffuse into the vicinity of the IV from the outer parts of the detector and produce -ray backg r ound in the F V . Outer Vessel and Inactive Buffer: The OV is a stainless steel sphere 13.7m in diameter and serves two purposes: (i ) it is the mechanical structure that supports the buoyancy force of the lig hter sc intilla tor a nd buffe r c o mpa r e d to wa te r, a nd (ii) it is the support base for the PMT 's. The OV contains a transparent buffer liquid having a mass of 1040 tons and a de nsity ve ry c l ose to tha t of the LS in the IV . The ba sic a im in the de sign is a nonscintillating buffer liquid that achieves a buoyancy condition as close to neutral as possible to minimize mechanical stresses on the IV balloon. The choice is PC but without the PPO fluors. However, even in the absence of fluors the PC itself produc e s sc intilla tions in the ultra v iole t. This is a pote n tia l proble m sinc e the PMT 's, whi c h are t h e m o st radi oact i v e part of t h e det ect or, si t i n t h e buffer. Because of t h e possibility of imperfect position reconstruction, some of these events may appear in the fiducial volume, producing a backgr ound in the sign al window. Thus, a 'quenching' compound (dimethyl phthalate, DMP) 50 is added to the buffer PC at a concentration of 5g /L to suppress the buffer fluorescence without affecting the tra n smission of the IV sc intilla tion lig ht. The DMP is a l so c h e mic a lly c o mpa tible with the ma te ria l s of the OV a nd PMT 's. An a lte rna t e buffe r de sig n tha t wa s c onside r e d is ba se d on a wa te r shie ld (a s in the CTF ) using PXE a s the sc intilla tor in the IV . The de nsity of PXE is only 0.5% off that of water at typical operating temperatures. The density match could, if necessary, be improved by small amounts of suitable additives. The scintillation and optic a l prope rtie s of PXE a r e ve ry simila r to those of PC. Phototubes, Light Concentrators and Signal Quality: The OV sphere supports 2200, 20cm ETL PMT 's 41 installed inside the sphere, connected by feed-throug h 's across the OV wall to a sing le cable outside the sphere carrying both sig n al and hig h voltage . The back-end sealing s of the PMT 's have been desig n ed to be compatible for operation in PC or water buffer. The PMT 's have been sel ect ed for l o w radi oact i v i t y gl ass, low dark pulse rate, low after pulse rate, and a (1 ) tra n sit time spre a d of 1ns, much smaller than the scintillation pulse-width. About 1800 of these tubes are equipped with light concentrators 51 t o enhance t h e opt i cal coverag e , ex pect ed t o be effect i v el y 30%. Sig n als from the remaining 400 PMT 's without light concentrators will be used for distingu ishing muon tracks in the buffer and point like events in the scintillator (see below). The shape of the concentrators is desig n ed to produce a uniform response to every event in the IV with a misalignment tolerance of 5 degr ees. The P M T and concent r at or syst em i s rat e d t o produce an average si gn al efficiency of 400 photo-electrons/MeV that translates to an energ y resolution of 5% (1 ) at 660 keV. The resolution is suitable for observing the basic spectral sig n ature of the Compton like e d g e of the re c o il e l e c t ron profile from 7 B e solar neutrinos. The system position resolution, determined mainly by the scintillation rise time and by the photoelectron yi eld, is ex pected to be <10cm (1 ) i n each ort hogonal di rect i on at 1 MeV.


30

Muon Veto: Even though the underg round location suppresses the cosmic muons, the CTF tests showed that muons that intersect the detector produce a prompt backgr ound of low energ y events in the sign al window. 39 These event s ari s e m o st l y from muons that intersect the non- DFWLYH EXIIHU UHJLRQ DQG SURGXFH ÙHUHQNRY - and scintillation photons. Despite the presence of a quencher in the PC buffer, high levels of photons g e nerated by the muons are not completely suppressed. To reduce this backg r ound to <1 event/d, an efficient muon veto system has been desig n ed. 52 The B o rex i no muon veto consists of an 'outer detector ' of 210 PMT 's on the outside of the OV sphere and an 'inte rna l de te c t or ' with 400 PMT 's without concentrators. The IRUPHU DFWV DV Z DWHU ÙHU enkov detector for intersecting muons. The latter accepts photons basically from all directions (unlike sign al PMT 's wi t h concent r at ors t h at accept photons mainly from the active scintillator reg i on). B y comparing the total pulse heig ht of both classes of PMT 's, muon tracks in the non-active buffer regi on can be separated from point-like events in the active scintillator. In addition, the different time pattern created by muon events in the inner detector compared to point like -signals can be used to separate backg r ound events. The overall desig n has been optimized in order to establish a redundant muon veto system to suppress this backgr ound by a factor of 10 4 . External Tank: Overall containment of the detector is provided by an ex ternal ta nk 18 me te r in dia m e t e r a nd 17 me te r hig h, fille d with wa te r a s shie ld a g a i nst ex t e rnal - UD\V D QG QHXWURQ UDGLDWLRQ ,Q D GGLWLRQ WKH ZDWHU DFWV DV D ÙHUHQNRY medium for the muon veto system described above. An ex ternal plant (used in the CTF ) feeds ultra-pure water to the tank. This ensures high optical transparency for ÙHUHQNRY RSHUDWLRQ 4.2 Ancillary Plants Scintillation Fluid Management: An elaborate system is under construction to store, handle and purify some 300 tons of PC scintillator and 1000 tons of PC buffer liquid. The main components of the system are: four storag e tanks, four purification systems, detector filling systems, and the two detector systems of CTF and B o rex i no. In addition, there is an interconnection system for distributing the 300 tons of sc intilla tor from the stora ge ta nks into the B o re x i no de te c t or a nd/or the CTF via or bypassing the purification systems. The system is made of electro polished stainless ste e l plumbing a nd high qua lity va lve s a nd fitting s (He le a k ra te < 1 0 -8 mbar L/ s), consistent with high purity chemical methods and the ex clusion of radon and kryp ton. On-line purification of the LS is based on four systems: g a s removal, water ex traction, distillation, and solid column chromatog r aphy to remove non-g a seous contaminants. Ultra-pure nitrogen is used to scrub Rn or Kr. Cosmog enic 7 Be i s produced during sea-level ex posure to cosmic rays in the sign ificantly long periods between large-scale production and storag e underg round. It creates a critical backgr ound orders of magn itude larg er than the sign al. One task of the on-site distilla tion fa c ility (possibly a l so of the purific a tion c o lumn) is to re move the cosm ogeni c 7 B e contamination; the other is g e neral purification for removal of optic a l impuritie s a nd othe r ra dioa c tive c onta mina n ts. Wa te r Purific ation Sy ste m : Some 2000 tons of ultra-pure water must be produced and m a i n t a i n ed t o serve as t h e out er shi e l d and as a hi g h l y t r ansparent


31

ÙHUHQNRY PHGLXP $QRWKHU tons are required for operating the CTF . In addition, water is used for cleaning the detector and the scintillator fluid handling syst em and for ul t r a-puri f i cat i on processes such as wat e r ex t r act i on for t h e scintillator. The system has a production capacity of 2 tons/h. The performance of the system is summariz e d in Table 2. Nitrogen Plant: The nitrog en plant consists of three liquid nitrog en storag e tanks, 6m 3 each, t w o at m o spheri c evaporat ors and a wat e r bat h el ect ri c heat er t o produce N 2 g a s of up to 250m 3 /h. Pure nitrog e n g a s is use d in g r e a t qua ntitie s in ha ndling the liquid syste m s, pe rforming the fina l N 2 sparg i ng st ep i n t h e LS purification and in saturating the scintillator with nitrog en for ox yg en-freedom (necessary to maintain scintillation efficiency). In order to reach the desired Rn purity in the sc intillator (order of 1 µBq / m 3 ), the nitroge n must c onta i n le ss tha n 1 µBq R n 222 /m 3 N 2 . At the time , this wa s a fa c t or 100 beyond state-of-the-art boiloff techniques. Technolog y was thus developed for large scale ultra-purification of N 2 . High purity nitrog en (1 µBq / m 3 due to 222 R n ) i s produced by charcoal col u m n purification of the liquid N 2 prior to evaporation. Up to 100 m 3 /h of hig h purity N 2 c a n be supplie d via e l e c t ro polishe d line s e quippe d with high qua lity va lve s . Qua lity control of these ga ses is done after Rn concentration on charcoal with miniaturiz e d radon proportional counters. The monitoring se nsitivity i s a record 0.5 µBq / m 3 for 222 Rn. 48,53 4.3 Operational Elem ents 4.3.1 Signal Processing and Data Acquisition The basi c observabl es for t h e i d ent i f i cat i on of e v e n ts in B o re x i no a r e the tota l energy released in the scintillator, as measured by the number of photons emitted, and the time distribution of these photons. The electronic sig n al processing scheme shown in F i g . 13 is de sig n e d to a c h ie ve the timing prope rtie s ne e d e d for a va rie t y of key tasks: reconstruction of the event position, PSD of - and - types of events, and the ide n tific a tion of a va rie t y of de la ye d c o inc i de nc e ta g s with a wide ra ng e of time bases.
238

U

10

-3

10

-6

10

-7

Contamination
226 232

Raw water 3·10 10 10
-1

Borexino design 10 10
-6 -6 -6

Achieved 10 10
-6

Ra Th K

-3 -3

-7 -6 -6

40 222

5·10 10

<2·10 3·10

Rn

10

-6

Table 2: Summary of water radiopurity purification performance. Activities are given in Bq/kg.


32

Figure 13: Block diagram of the Borexino electronics layout.


33

The data acquisition (DAQ) is based on the Li nux operating system. The DAQ softwa re is e n tire l y c u stom ma de , with e x te nsive use of multi-ta sking te c hnique s. User i n t e rfaces are al l based on W E B t echni ques. The si g n al from P M T 's is AC coupled to a front-end card 54 that performs noise filtering, pre-amplification, shaping and integr ation of the input sign al. It provides both a linear response used for time measurement and a voltage sign al proportional to the total charg e . Each front-end board provides also an analog sum of 12 linear output sig n als that ex tend the dynamic rang e of the system to 30 MeV by means of a fl ash ADC syst em . The out put s of t h e front end cards are sent t o a speci al l y design ed VME slave card that performs the analog to dig ital conversion of the charg e sign al (8 bits resolution), measures the time of the linear sign al with 0.4ns resolution, computes the sum of recorded hits in a time window of 60ns (used for trigge ring) and stores the whole information in a dual port random access memory. The outer muon tubes are read with a different front end system that performs a charg e to time conversion of each sig n al after a linear pre-amplification. The converted sig n al is then sent to time-dig itizers which are read by their own processor. The trig g e r can be g e nerated both by the internal as well as by the outer muon detector. W e require at least N PMT hits occurring in a time window of 60ns in the inne r de te c t or to g e ne ra te a trig g e r. The trig g e r thre shold N, de te rmine d by the ope ra ting sing le s ra te , will ultima te ly de pe nd on the 14 C contamination, ex pected to be the largest contributor to the sing les rate. In the inner detector, N i s ex pect ed t o be 15-20, corresponding to (40-50) keV energ y deposition. 4.3.2 Calibration and Monitoring Even wi t h an ul t r a-pure LS (10 -1 6 g / g ) that leads to very low backg r ound, det e rm i n at i on of t h e det ect i on effi ci enci es i s vi t a l . For ex am pl e, t h e - separation effi ci ency rem a i n s cri t i cal especi al l y wi t h i n t h e window since -pa r tic le s of 4 Me V a r e 10 time s more pre v a l e n t tha n the sign a l . With a n a n tic ipa t e d se pa ra tion e ffic i e n c y of 90%, any energy or positional dependence on the separation would strong ly impa c t the spe c t ra l inte rpre ta tion a nd the infe rre d flux . In addition, the application of a number of software cuts, a basic one being the determination of the F V on a continuous basis, require the determination of precise efficiencies. An accurate calibration of the detection system is thus of paramount importance for quantitative precision, stability and reliability of the spectroscopy. The calibration prog ram 55 c ove rs the e n e r g y a nd the position se nsitivity of the de te c t or using a c tive ta gs of trace impurities in the LS as well as ex ternal point sources inserted periodically into the tank. Laser Monitor: The pulse timing and the ga in of each individual phototube of the inner detector are calibrated by a laser system. Photons from this source are distributed to all PMT 's vi a t h i n quart z fi bers connect ed t o t h e opt i cal concent r at ors. The light yi eld corresponds to sing le photoelectrons as in real event s . The out er muon veto detector will be calibrated by a set of blue lig ht-emitting diodes that are mounted on the inside wall of the outer tank. Their ga ins match typical photon yi elds RI ÙHUHQNRY HYHQWV Native Dispersed Radioactive Sources: The energ y response can be continuously monitored using the internal trace radioactivities native to the LS as wel l as sel ect ed i n sert ed sources. The spect roscopi c feat ures of t h e sources t o g e t h er


34

cover t h e ent i r e rang e of energ i es up t o 5 Me V tha t is of inte re st to most of the physics questions addressed by B o rex i no. A monitor for low energ y short- and long te rm sta b ility is offe re d by the re la tive l y high ra te -spectrum of 14 C (0-156) keV. Even at a concent r at i on of 10 -1 7 g / g U and Th, the - DC tag s can be observed at a rate of 100 events/month and the mono-energ e tic -part i c l e s at an even hi g h er rat e . The s e fe a t ure s c a n be utilize d a s long te rm monitors of the e n e r g y re solution a nd spe c t ra l sta b ility while the spa tia l c o inc i de nc e s of the DC ta g s c a n be use d to monitor the operating position resolution. F i nally a calibration point at a high er energy is offered by the neutron capture -rays at 2.2 MeV which can be tag g e d in delaye d coincidence trigge red by the muon that produces the neutron. The CTF showed that these occur at a rate of 0.3 neutrons/ton,d , thus producing 100 neutron capture events/d in B o rex i no. Positioned Point Sources: A series of point calibration sources is being design ed which can be positioned throughout the detector volume. F o r high energ y 's, a 222 Rn-source will be used at a g i ven position, supplementing the dispersed internal Rn radiations. W e envision for the lower energy -rays a 7 B e source (478 keV) and for pure 's a 32 P source (1709 keV endpoint). F o r low energ y -pa r tic le s, e ithe r 232 Th (4.0 MeV) or 238 U are consi d ered. For l i n e- / sources 113 Sn and 137 Cs ma y be suita ble . Calibration of the Neutrino Response: In order to provide a direct demonstration of the overall response of the detector, calibration by means of a man-made sub-MeV -sourc e with a c tivitie s in the Me g a c u rie ra ng e is fore se e n. S u ch a fi nal source t e st i s i m port a nt especi al l y i f t h e m easured sol a r sig n a l is null or close to the backgr ound level. Plans include both electron-neutrino ( 51 C r) and el ect ron-ant i n eut r i no ( 90 Sr) sources. A tunnel has been installed just below the e x te rna l ste e l ta nk for inse rting a nd re trie ving the he a v ily shie lde d sourc e c onta i ne rs on st eel t r acks. 4.3.3 Event Simulation Neutrino-induced events and backgr ound have been studied by a simulation code, st ruct ured i n t h ree seg m ent s . In t h e fi rst one 56 a ll the pa rtic le s produc e d in a i n t e ract i on or i n radi oact i v e decay are g e nerat e d wi t h a det a i l e d m odel i n g of t h e el ect rom a g n et i c shower produced by t h e el ect rons or -rays . Thi s l eads t o t h e energ yspace characteriz ation of these events. A second seg m ent tracks the lig ht emitted in the scintillator. It takes into account absorption, re-emission and scattering in the scintillator and in the buffer, as well as reflections on surfaces. A different version of this tra c k ing ha s a l so be e n se t up using the GEANT4 c ode . 57 In both codes the lig ht col l ect ed on t h e P M T 's is converted into time and charg e sig n als. The third seg m ent, the 'reconstruction code ', pe rforms the inve rse proc e ss in whic h, sta r ting from the PMT 's sign als, the space-time coordinates of the events are evaluated by max imum likelihood using a probability density function. The typical values obtained with this code for the (1 ) space and energy resol u t i ons i n t h e si m u l a t i on of an el ect ron of 1 MeV i n t h e cent e r of t h e IV are 8cm and 50 keV. These simulations have been suc c e ssfully te ste d in the CTF da ta a n a l ys is.


35

4.4 External Background The principal concern in B o rex i no is the non-shieldable internal backgr ound, due to the ra dioimpuritie s of the LS itse l f. Ne ve rthe le ss, a n importa nt role in the de te c t or design is also played by the need to suppress the backgr ound from ex ternal sources, such as the construction materials and the surrounding rocks, as well as the cosmic ray interactions in the underg round environment. This impacts the architectural desi gn and dem a nds ex t r aordi n ary care i n t h e sel ect i on of m a t e ri al s of t h e det ect or by characterization of radiopurities. We briefly outline the methods used for material qua lity c ontrol, the on-line dia gnostic s a v a ila ble for de te rmina tion of the ove ra ll ex ternal backg r ound, and some details on the effects of cosmic ray interactions. 4.4.1 Radiopurity of Detector Materials The basic architecture of B o rex i no calls for hig h purity construction materials at every laye r of the detector. Quality control of radiopurity specifications of every ma te ria l , from the ste e l in the e x te rna l ta nk to the sc intilla tor in the IV a s we ll a s the materials of all liquid handling systems, has been performed by various counting methods. The major techniques were: Rn emanation measurements, Ge spectroscopy, mass spectroscopy and NAA, with progressively high er sensitivities but less wider applicability. Table 3 shows the B o rex i no desig n radiopurities compared with what has been achieved in CTF . Sa mple s of N 2 , air and water were routinely assaye d for R n with Lu cas cells 58 and later with miniature proportional counters. The key aspect of this technique is the c onc e n tra tion proc e dure of Rn with low bla nk va lue s from la rge sa mple s. 47,48,53 Rn emanation of detector components and materials is routinely tested at a sensitivity level of 50 µBq 222 Rn. 59 In related work, the permeation of Rn throug h various ma te ria l s, e s pe c i a lly the nylon film of the LS ve sse l wa s te ste d. This re ve a l e d ma jor e ffe c t s vita l to the de sign of the IV . 60,49 La rg e vol um e Ge spect rom e t e rs wi t h R n free counting and special shielding assaye d hundreds of samples (up to kg mass) for U/Th a t a se nsitivity of 10 -1 0 g/ g. Most of the bulk material in the g e neral construction were screened by this method, setting up an ex tensive database on bulk material radiopurity. 61,62 Inductively coupled mass spectrometry (ICPMS) is applicable to U,Th de te c tion with a typic a l se nsitivity of 10 -1 2 g / g , rising however to 10 -1 5 g/ g f o r wa te r in pa rtic ula r . 63 This technique has been applied to assay water samples, nylon m a t e ri al (aft er concent r at i on by chem i cal ashi ng or di ge st i on) and LS m a t e ri al (aft er c onc e n tra tion by wa te r e x tra c tion). B y fa r the most se nsitive te c hnique a pplic a b le to key impurity species, e.g. specific determination of 238 U in the 10 -1 7 g/ g r e gi m e , NAA-IS AN was used to decide critical questions in the LS purity. In formation on the purity of 40 K in PPO, the scintillation fluor, comes only from NAA. B e sides U,Th and K, other long lived naturally occurring nuclides can also be measured with this technique. 41 4.4.2 On-line Measurement of External Background The ma jor -ray and neutron backg r ound at the F V arises from the rock environment and the outer parts of the detector (OV steel sphere, PMT assemblies, the liquid buffer and IV walls). The radiopurity (as well as Rn emanation) of all these materials


36

Table 3: Requirements on the radiopurity of detector materials for Borex ino and values measured in the CTF

Material Stainless steel External water PM Scintillator Scintillator

Borexino design 10 10
-9

CT F achieved 10 10
-9

Unit g/g of T h,U equiv. g/g of T h,U equiv. g/g of T h,U equiv. g/g of T h,U equiv.
14

-1 0 -8

-1 4 -8

10 10 10

10 10 10

-1 6 -1 8

-1 6 -1 8

C/12C


37

is known so that the -ray flux can be simulated precisely at any part of the detector. 56,61 The dom i n ant source i s t h e P M T array, especi al l y t h e P M T gl ass, even with low activity g l ass. The 3.25m buffer liquid shields the F V from the PMT backgr ound. In this, 1.25 m of the shielding is active and the combined effect of passive shielding plus active rejection of ex ternal backg r ound energ y deposition in the active reg i on achieves 7-8 orders of mag n itude in backg r ound reduction, to <0.1 event/d. The ex ternal backgr ound produces the larg est fraction of events in the active buffe r zone in the sc intilla tor volume , thus, the i r ma gn itude a nd ra dia l de pe nde nc e can be used to separate the ex ternal component as well as to estimate the -ray l eakage i n t o t h e FV. Fi g. 14 shows t h e resul t of a Mont e-C a rl o cal cul a t i on of t h e ex ternal backg r ound for a lifetime of 10 days . It depicts the energy spectrum vs. the reconst r uct e d radi al di st ance of t h e event . No event s are seen i n t h e window in the F V . B ackg r ound events beg i n to appear as the energ y and the radial distance increases towards the edge of the F V , affecting , e.g ., the window of the solar pep sign al at (0.9-1.5) MeV. The problem is central for measuring the 8 B- flux in the regi on of (3-6) MeV, measurable only in B o rex i no. 4.4.3 Cosmogenic Radioactivity The effect of muons in B o rex i no is twofold: (i ) production of prompt events in the window; and (ii) muon induced radioactivity by various reactions. 52,64 The muon veto described above is the main defense ag ainst (i ). Veto effect i v eness agai nst (ii) depends basically on the lifetimes of the induced radioactivities. The dead time incurred by vetoes using a muon sig n al is limited to a few seconds. Thus, activities with lifetimes shorter than this rang e can be vetoed by the muon detector. The absence of isotopes heavier than 12,13 C (and some 16,17,18 O in the fluor) is a key factor that limits the list of possible activities. A fundamental da tum is the CTF re sult of 0.3 neutrons/ton,d correlated with muons, that sets a limiting rate on all possible spallation reactions. This translates to some 100 spallations per day in the B o rex i no F V . The main question is what fraction of these events produce long-lived radioactivity that eludes the muon veto. To obtain more data on this point, the interaction of high energy muons on carbon and on liquid sc intilla tors wa s studie d in a n e x pe rime nt a t the SPS muon be a m a t CERN. 64 Production cross-sections for possible radionuclides from the interaction of 100 and 190 GeV muons on 12 C were measured. The possible ra dioa c tivitie s from a C ta rg e t too long -live d to be ve toe d by the muon sign al are 11 Be, 11 C, 10 C and 7 Be. 11 B e is re la tive l y ha rmle ss sinc e its production cross-section is very small. 10,11 C are both + emitters so that their decay spectra are offset by 1.02 MeV out of the B e - sig n al window (0.25-0.8) MeV by c a l orime tric summing of the a nnihila tion ra dia tions with the + spect rum . The production of 10,11 C from 12 C re le a s e s ne utrons whic h will be c a p ture d by protons after a typical delay of several 100 µs to emit a tag g a ble 2.2 MeV -ray. 10 C with a me a n life time of =28s could possibly be tag g e d reasonably near the located si t e of t h e n-capt u re. The case of 11 C with =1765s may be less straig htforward. 7 Be


38

Figure 14: Monte Carlo simulation of the ex ternal background, corresponding to 10 days of data acquisition: Plotted is the energy [MeV] vs. the distance from the center of the IV [cm]. The (empty) box at the lower left corresponds to the energy-position window of interest for 7Be-e in Borex ino (the innermost 2.25m are truncated). The shade code scale on the right is the log10 of the number of events.


39

presents direct interference since its 478 keV -ray produces backg r ound in the sign al window itself. However, the event rate in the F V at saturation after nearly a year is ex pected to be only 0.4 counts/d. 5. R e sponse t o N o n- St andar d Solar N e ut r i nos As described in Sec. 2, the results of the present solar ex periments are gl obally c o mpa tible with se ve ra l conversion scenarios. On a quantitative basis set by the SSM flux es, the possible pa ra me te rs a r e re stric t e d to a fe w isla nds in the m 2 sin 2 (2 ) space (see Figu res 5,6). The urg e nt need is to find specific solar-model independent effects that can prove flavor conversion. Two such effects for the high energy (>6 MeV) 8 B- si gn al - a devi at i on from t h e Ferm i spect ral shape and a day/ nigh t sig n al variation - have been sought in high precision SK data after a lifetime of >1000 days . Neither effect has been observed. 65,11 The re ma ining hig h energy test, a difference in the CC and NC sig n al rates, is being pursued by SNO. B o rex i no will unveil a new real-time energ y window to search for conversion effect s, vi z . t h e l o w energy solar spect rum focusi ng on t h e 7 B e -ne u trinos. With the evidence for conversion coming from the 7 Be- 8 B proble m ra the r tha n from mode ldependent flux deficits, strong effects at low energ y are predicted, and thus a distinct sign ature in B o rex i no should be seen. The basic data ex pected from B o rex i no are the ma gn itude of the -e scattering sig n al rate due to mono-energ e tic 7 Be solar 's and a possible time de pe nde nc e of this ra te ove r da y/ nig h t or se a s ona l pe riods. Also, the spect ral shape of t h e 8 B -e scattering sign al in the lower energy (3-6) MeV regime may be measured. 7 Be Neutrino Signal Rates: With 100 tons of the PC scintillator targ et, the SSM and a standard predict, in the energ y window (250-800) keV, a sig n al rate of 55/d, 80% of which is due to 7 B e -neutrinos. F i g . 15 shows a simulated sig n al spectrum observable in B o rex i no and the backg r ound ex pected for 10 -1 6 g U /g . B a sed on the class of radiopurity and tag efficiencies demonstrated in the CTF , such radiopurity level is achievable in B o rex i no. If the solar e is fully converted to other active fla vors, the ra te drops to a ha rd c o re minimum of 11/d, arising only from the NC interaction. W ith reasonable precision, this low sig n al would strong ly sug g e st (if not prove) flavor conversion rather than an astrophysical effect since such a rate for 7 Be e l e c t ron ne utrinos is inc o mpa tible with the obse r ve d 8 B- si gn al rat e . Evi d ence for non-standard 's would become sig n ificantly strong er if the B o rex i no sig n al drops further, below the hard core value, in the event of full conversion to sterile neutrinos. B o rex i no is sensitive to this scenario if the backgr ound can be reduced sign ificantly below that in F i g . 15. In that case, just an upper limit that rules out the hard-core sign al rate is sufficient to prove conversion to sterile 's unambiguously. F o r inte rme d iate ra te s be twe e n the ha rd c o re a nd the SSM limits, infe re nc e of fla vor conversion is less definitive. The major conversion scenarios are the MSW effect and vacuum oscillations, w h i c h present four different parameter regimes as discussed in Sec. 2 and summariz e d in F i gu res 4-6. F i g u re 16 is a plot of iso-depression lines of the ex pected sign al due to


40

Figure 15: Monte Carlo simulation of signals (in arbitrary units) and background. Plotted are the 7Be- signal expected from the SSM (dotted-dashed), the signal from all other neutrino sources together (dotted), the background (dashed), and the sum spectrum from all events (solid line). The background is calculated for 10-16 gU/g, an / discrimination of 90%, and 1 statistical cuts.


41

B e neutrinos in the B o rex i no detector in the m 2 -sin 2 (2 ) plane. As for the other neutrino types and for the absolute rates ex pected in B o rex i no, they are listed in Table 4 for the SSM and for the different MSW solutions. The MSW effect predicts sign ific a n tly diffe re nt ra te s for the sma ll mix ing (SMA) a nd the la rg e mix ing (LMA) regi ons. In the SMA, the 7 Be- 's are nearly fully converted, leading to a sig n al rate close to the hard core value. Such a result would strong ly sug g e st flavor conversion. In the LM A, the sig n a l is typic a lly re duc e d to 50% of t h e S S M val u e. In t h i s case conclusions on flavor conversion can be reached only in conjunction with other ex periments. In the LO W reg i on, the rate is similar as for LM A, but a drastic day/ nigh t effect can be detected (see below). Temporal Variation of the 7 Be Neutrino Signal: Unlike the ra te s, te mpora l variations of the sig n al are always decisive observables for flavor conversion. In two of t h e vi abl e scenari o s, 's are reconverted after leaving the Sun either (i ) by vacuum osc illa tions on the wa y to the Ea rth or (ii) by matter rege neration in passage throug h the Ea rth. In both c a s e s , the re sult is a time va ria tion of the sig n a l: (i ) The va c uum osc illa tion e ffe c t de pe nds on the va ria tion of the Ea rth-Sun dista n c e in the Ea rth 's eccentric orbit, thus the time variation is seasonal. This scenario is va lid for ma x ima l mix ing a nd ve ry sma ll ma ss pa ra me te rs, m 2 10 -9 to 10 -1 1 (eV/c 2 ) 2 . F i g. 17 shows the daily sign al in its seasonal dependence for two values of m 2 tha t a r e c onsiste nt with da ta from SK. 11 (ii) The passage through the Earth occurs nig h tly, thus, MSW reg e neration of the electron flavor in the Earth matter enhances the sig n al during the nig h t and produces a day/ nigh t effect (DNE) of the sig n al. The mag n itude of the DNE depends on the pa th le ngth throug h the Ea rth, the pe ne tra tion towa rds the c o re (with inc r e a s ing density of the Earth matter) and for statistics, the length of nigh t hours over the year. In addition, the reg e neration and thus the DNE depends also on the ge ographical latitude of LNGS as well as on the tilt of the Earth ax is relative to the Sun, i.e. the day of the year. The effect thus chang e s throug hout the year in a pattern charact eri s t i c of t h e g e og raphi cal and of t h e parameters. Earth reg e neration effects are particularly enhanced in the LO W reg i on. Borex i no i s wel l posi t i oned t o ex pl ore t h i s because t h e l a rg est vari at i on occurs j u st at t h e energy of 7 Be- 's. This can be seen in F i g . 4 which displays the DNE for different MSW scenarios. The mono-energ e tic feature of the 7 B e -line is pa rtic ula r ly suita ble for t e st i n g because t h e DNE i s not frag m ent e d over a wi de energ y rang e. 17,66 Low Energy Spectral Shape of 8 B Neutrinos: A g e neral charact eri s t i c of t h e energy dependence of flavor conversion by the MSW effect is a g r adual decrease in fl avor survi v al - and st rong er observabl e effect s - t o wards l o w energi es. The distortion of the spectral shape of the 8 B sig n a l (re la tive to tha t from a F e rmi sha p e of the spect rum ) t hus i n creases t o wards t h e l o w end of t h e spect rum . The hi g h HQHUJ\ SDUW R I WKH VSHFWUXP ! 0 H9 K DV E HHQ P HDVXUHG L Q ZDWHU ÙHUHQNRY GHWHFWRUV Kamiokande, SK and now in SNO. In this reg i on, so far, spectral deviations appear sma ll, if a n y. It is the r e f ore importa nt to e x te nd the da ta to lowe r e n e r g i e s 49 , the only foreseeable chance for which lies in B o rex i no because of its low energy sensitivity although the relatively small targ et mass compared to SK is an inherent limitation. The ex pect ed rat e s for 8 B- 's i n Borex i no for t h e di fferent scenari o s are i n cl uded i n Table 4.
7


42

Table 4: Solar neutrino counting rates ex pected via -e scattering per day in Borex ino in three windows of the recoil electron energy for 4 scenarios: -SSM, the standard solar model -LMA, the large mixing angle solution -SMA, the small mixing angle solution, -LOW, the region of a day/night effect. The rates are calculated for the characteristic parameters quoted in the caption of Fig. 5, for the mean Sun-Earth distance, for 113m3 fiducial volume, and for PC as scintillator. No radiative corrections are included. The case of vacuum oscillations is not shown in the Table. In this case the rate would be subject to strong seasonal variation. Hence, real-time detection of the mono-energetic 7Be-'s enables a unique test of vacuum oscillations.

0.25 - 0.80 MeV

p-p
7

0.22 43.3 2.0 4.0 5.5 0.07 0.08 55.2 1.43 0.13 1.80 0.02 0.10 3.48 0.45

0.15 24.4 0.95 2.27 2.86 0.03 0.03 30.7 0.68 0.07 0.86 0.01 0.04 1.66 0.17

0.08 9.20 0.39 0.87 1.12 0.01 0.04 11.7 0.28 0.03 0.35 0.00 0.05 0.71 0.22

0.13 22.8 1.03 2.13 2.86 0.03 0.04 29.0 0.73 0.07 0.92 0.01 0.05 1.78 0.23

Be

p-e-p
13 15

N O F

17 8

B Sum

0.80 - 1.50 MeV

p-e-p
13 15

N O F

17 8

B Sum

1.50 - 5.50 MeV

8

B


43

Figure 16: Is o-depression lines in the m 2 -sin 2 (2 ) pl ane for t h e ex pect ed m easurabl e sig n al due to 7 B e neutrinos in the B o rex i no detector. Plotted are contours for 25, 35, 50, 70 and 90 percent of the full SSM sign al due to 862 keV 7 Be- 's, respectively. The NC contribution from µ - e - scattering is included. Averag ed over day and nig h t and over the year. F o r absolute rates, see Table 4.


44

Figure 17: Annual variation of the daily 7Be- produced Borexino counting rate due to vacuum oscillations in combination with the eccentricity of the Earth's orbit. The dashed upper curve shows the no-oscillation signal with its purely geometrical variation ~R-2 (7%). Neutrino oscillations have particularly distinct effects on the mono-energetic 7Be-line. Two oscillation mass parameters have been selected to illustrate how accurately m2 can be determined: (i) solid line, m2 = 4.2â10-10(eV/c2)2. This is the best fit for the Superkamiokande data. (ii) dashed line, m2 = 3.2â10-10(eV/c2)2, only marginally lower. Full mixing was assumed in both cases. The broadening of the 7Be-line due to the temperature in the solar center as well as the radial 7Be distribution in the solar core have been taken into account.


45

6. N o n- solar N e ut r i no Sc ie nc e W h ile the basic objective of B o rex i no is the direct observation and measurement of the sola r 7 Be- flux , the fa c ility c a n be a pplie d to a broa d ra ng e of frontie r que stions in particle physics, astrophysics and g e ophysics. 67 The unique low e n e r gy se nsitivity and ultra-low backg r ound in B o rex i no bring new capabilities to attack problems in these fields. Much of this research can be undertaken simultaneously with solar observations. In particular, antineutrino ( e ) spectroscopy can be performed simultaneously with a distinct tag independent of solar- spectroscopy. Other re se a r c h ne e d s a dditiona l fa c ilitie s suc h a s a r tific ia l sources or foreig n targ ets in the sc intilla tor. As in the c a s e of sola r 's, the CTF is we ll suite d for te sting the feasibility of novel methods to attack some of the questions. 6.1 Antineutrino Science Detection of Low e flux e s : B y far the best method to detect e i s t h e cl assi c R e i n es re a c tion of c a p ture by protons in the sc intilla tor liquid : e + p n + e + . The positron sign al energy (kinetic energ y + 1.02 MeV annihilation energ y) E = E( e ) ­ Q where the threshold energ y Q = 1.8 MeV. Thus, even at threshold E( e )= Q, this reaction produces an easily detectable sig n al at 1.02 MeV. The e ta g is ma de possible by a delaye d coincidence of the positron and by a 2.2 MeV -ra y e mitte d by capture of the neutron on a proton after a delay of 200 µs caused by the moderation of the neutron to thermal energ i es. The tag suppresses backgr ound by a factor 100, t hus, t h e act i v e buffer i s not cruci a l . In favorabl e ci rcum st ances, t h e ent i r e scintillator mass of 300 tons may be utilized. One of the few sources of correlated backg r ound is muon induced activities that emit -neut r on cascades. However, al l suc h c a s e s ha ve life time s <1s. Thus they can be vetoed by the muon sign al. Overall, a si gn al rat e as l o w as 1 e event per year and 300 tons appears measurable. The inte re sting sourc e s of e are supernovae, the Sun, the Earth and nuclear power react ors, whi c h can be di st i n g u i s hed from each ot her m a i n l y by t h e charact eri s t i cal l y different energy spectra of the sign als. The sensitivity limit (1 event/yr) corresponds to a flux of 5 â 10 2 e /c m 2 ,s or 10 -4 â ( 8 B ) . B o rex i no is thus one of the most se nsitive e de te c t ors e v e r de sign e d a nd it is ta ilore d to se ve ra l proble m s in pa rtic le physics, ge ophysics and astrophysics, all of which predict low e flux es. Solar Antineutrinos: The SSM predicts no e emission from the Sun. Nonstandard physi cs however, can creat e e by some conversion mechanisms. One such model, still viable, is based on an off-diag onal or 't ransition' ma g n e tic mome nt ( µtr ) a llowe d the o re tic a lly for a Majorana neutrino. The best present limit on µtr , based on astrophysical g r ounds, is 3 â 10 -1 2 µB (B ohr mag n eton). 68 La boratory limits a r e le ss stringe n t by typic a lly 2 orde rs of ma g n itude . 68 The inte ra c tion of µtr with solar magn etic fields may produce a spin-flavor precession, i.e. a spin-flip and a flavor conversion resulting in µ or a n tine u trinos. 69 The optimal conditions apply to 8 B 's in the LO W ma ss a r e a of the MSW ma p, i.e . m 2 10 -7 (eV/c 2 ) 2 and nearma x ima l mix ing . So fa r, this e ffe c t c a u se s only sola r e di sappearance. However, an inte re sting c onse que nc e is the appearance of solar el ect ron a n tine u trinos a t the Earth. B ecause of the max imal mix ing inherent in the model, the resulting µ/ antineutrinos can convert to e simply by va c uum osc illa tion. 17,66,70 The ma g n itude of this e flux depends on M 2 with M = µtr â B (B is the sola r ma g n e tic fie l d). The


46

c u rre nt e x pe rime nta l limit 71 on the flux of e wi t h energ y >7 MeV i s <5 â 10 4 /c m 2 , which corresponds to an event rate of <110 e /yr in B o rex i no. This already sets a limit of M< 7 â 10 -1 1 µB kGauss. At a sensitivity of 1 event/yr, B o rex i no can reach a limit M < 1 0 -1 2 µB kGauss. Antine u trinos from the Earth's Inte rior: A key fact or i n t h e concept u al foundations of g e ophysical models of the Earth 's inte rior is the ra diog e n ic he a t from t h e decay of U and Th i n t h e Eart h 's crust, currently believed to be 40% of the tota l heat flow of 40TW . B o rex i no can make a basic contribution to testing these models by detecting e emitted by these nuclides in the range of (1.8-3.3) MeV. Three speci fi c aspect s m a ke such a m easurem ent val u abl e 72 : (i ) One obt ai ns a 'whole Earth ' view of the crustal U/Th abundance instead of sampling by laborious field work; (ii) The U and Th abundances can be separately measured; (iii) by c o mbining da ta from B o re x i no (Eura s ia n pla t e ) with those from the KAMLAND detector of similar design planned in J a pan (interface of Asian and oceani c crust s ), t h e rel a t i v e di st ri but i on of U/ Th i n t h e cont i n ent a l and oceani c crust s may be probed. Depending on the g e ophysical model, e rates between 10 and 30 event s / yr can be ex pect ed i n Borex i no. 72 Long-baseline e from European Reactors : The flux of e from astrophysical sources (including the Earth) are essentially model-dependent. Nuclear power reactors emit e with a known flux and spectral shape with energi es up to 8 MeV. B o rex i no is sensitive to this flux from power reactors situated all over Europe at an average basel i n e di st ance of 750km. There are no reactors in It aly itself. The multireact or e fl ux produces an accurat e l y ( 5%) predictable sign al in B o rex i no of 30 e v e n ts/yr. The we ll de fine d se t-up offe re d by this c o mbina tion ma ke s a n ide a l t e rrest ri al l ong-basel i n e ex peri m e nt for a m odel -free search for vacuum osc illa tions. 73 The se nsitivity for osc illa tions is in the ra ng e m 2 [1 0 -3 -10 -5 ]( e V /c 2 ) 2 , filling the ga p between the results of the CHOOZ ex periment and solar neutrinos. The European react or-Borex i no com b i n at i on i s t hus al so sui t a bl e t o probe osc illa tions in the LM A-MSW area, as it is currently planned for the Kamland detector. 6.2 Neutrinos and Antineutrinos from Type-II Supernovae The occurrence of a supernova burst creates a flux of a ll type s in the e n e r g y ra ng e of 10 's of MeV. They occur as a short pulse that lasts for several seconds. Two fa c t ors ma ke its de te c tion in B o re x i no inte re sting . 74 As the first ma jor de te c t or with liquid sc intilla tor ope ra ting a t low e n e r g i e s , B o re x i no ma y be unique de spite the rel a t i v el y l o w det ect or m a ss because: (i ) low energ i es in the F e rmi-Dirac spectrum of the supernova 's can be probed; (ii) e c a n be de te c t e d with hig h se nsitivity a nd (iii) Supernova neutrino interactions with 12 C in the liquid sc intilla tor offe r a unique tool for a CC/NC analys is of the flux by c o mbining the re sults from the re a c tions 12 C( e ,e - ) 12 N, 12 C( e ,e + ) 12 B and 12 & x x ) 12 C*(15.1 MeV). The sign a l s a r e e s tima te d to be 80 e - event s and 30 events from reactions on 12 C for a 3 â 10 53 erg burst at a distance of 10kpc. The principal results of interest to the supernova mechanism as well as to physi cs are:


47

(i ) the flux es of different flavors via the NC/CC data; (ii) the e spect rum ; (iii) limits on the ma ss via the diffe re nc e in time s of a rriva l be twe e n the CC a nd NC sign als; and (i v ) possible ne utrino osc illa tion sig n a l s. 6.3 Neutrino Physics with Megacurie Sources Man-made sources of neutrinos from intense radioactive 51 Cr sources have been used in the Ga ex periments to subject the radiochemical detectors to rig o rous throug hput tests of all procedures and components. 75 Gallex has ex panded this approach to achi e ve a re al c a libration of the de te c t or by c o mple me nting the sourc e e x pe rime nt with a hig h -sta tistic s 71 As doping ex periment. 76 The appl i cat i on of a Meg acuri e neutrino source is also foreseen for B o rex i no to demonstrate the performance of the de te c t or. This is e s pe c i a lly va lua b le in the e v e n t of a ve ry sma ll or null sig n a l (e .g. with complete conversion to active or sterile neutrinos). In addition, as a detector based on -e scattering , B o rex i no offers the interesting opportunity to probe the scattering mechanism to the level of precision sufficient to uncover new physics. The recoi l el ect ron profi l e i n -e scat t e ri ng devi at es at l o w energ i es from t h e norm a l weak-i nt eract i on m e di at ed scat t e ri ng i f t h e neut ri no carri es a st at i c m a g n et i c m o m e nt µ . As a massive detector with uniquely low backg r ound and low energ y sensitivity, Borex i no offers a new means for searching for µ using a r tific ia l ne utrino sources 17,77 such as 90 Sr - 90 Y (whic h e mits antine u trinos ) and 51 Cr whic h e mits neutrinos. These ideas have recently been updated 78 for the g e ometry of Borex i no with consideration of the non-removable backg r ound of solar neutrino scattering for various scenari o s. In t h e case of a 51 Cr source, t h e i s ot ope enri ched source material is potentially available and it offers the advantag e of a mono-energ e tic line at l o wer energi es t h an 90 Sr, althoug h the shorter lifetime is a drawback in this application. The sensitivity on µ a tta ina b le in B o re x i no 17 for a 1 MCi source is µ 5 â 10 -1 1 µB for 51 Cr and 3 â 10 -1 1 µB for 90 Sr. The la tte r se nsitivity c ould be e nha nc e d by the technical feasibility of multi-MCi sources. A tunnel facility for inserting and withdrawing sources under the B o rex i no detector is already built into the detector. 6.4 Double Beta Decay in 136 Xe The phenomenon of -decay offers the only means ye t known for identifyi ng the Majorana nature of the neutrino. The search for -decay has been traditionally carried out with targ et masses in the tens of kg scale. B o rex i no offers an opportunity for e nha nc ing this se a r c h on the multi-ton sc a l e 17,79 by e x ploiting the hig h solubility of noble g a ses in liquid scintillators without affecting radiopurity or scintillator effi ci ency. A favorabl e candi dat e i s 136 Xe. Tests could be planned first in the CTF II on the scale of 10kg of enriched 136 Xe for assessing sig n al quality, backg r ound etc. If successful, the measurements could be ex tended in B o rex i no itself with 2 tons of 136 Xe.


48

7. Conclusions and Outlook The 'Solar Neutrino Problem ' remains unsolved but restated more sharply than ever in te rms of the 7 Be- 8 B problem. Apparent paradox es that cut throug h caveats with c o mpe lling simplic ity ha ve a r ise n be fore in we a k inte ra c tion physic s a nd le d to scientific revolutions. An apparent violation of energ y conservation led to the postulation of the neutrino itself. The - puz z l e led to parity violation and a helical neutrino. The 7 Be- 8 B proble m sig n a l s a nothe r bre a k throug h, this time towa rds a flavor mix ed, massive neutrino, with emphasis on e - µ mix ing. The sa me phenomenon, neutrino oscillations, is strong ly indicated by the SK observation of µde fic its re la tive to e in cosmic ray produced atmospheric neutrinos that have penet r at ed t h e Eart h 80 , indicative of µ- mix ing . How will the s e de ve lopme n ts ta ke sha p e in the ne x t fe w ye a r s a nd wha t role will B o re x i no pla y? Pointers to the immediate future can be g a ug ed from the ong oing accumulation of SK solar neutrino data. The ex periment has so far produced no 'smoking g u n ' for solar neutrino conversion despite the admirable precision of its data. Limits on spe c t ra l de via tions in the 8 B spectrum and the day/ nig h t effect continue to point to a lack of evidence for any of the conversion scenarios, MSW or vacuum oscillations. In the ne x t ste p the ne wly sta r te d SNO will re e x a mine the 8 B spect ral shape for sm al l distortions, with hig h e r se nsitivity via inve rse -decay. The rat i o of NC / C C si g n al s in SNO can establish conversion in almost all scenarios, but cannot g i ve any decisive result if conversion occurs to sterile neutrinos. The uncertainties in the astrophysical models and in the nuclear cross-section of 7 B e+p do not rule out smaller SSM 8 B- flux es, smaller deficits thus, smaller conversion effects at 8 B e n e r gi e s to sta r t with. This sc e n a r io is in fa c t c onsiste nt with the present SK results. On the other hand, the very ex istence of the 7 Be- 8 B fl u x paradox is a strong indication of flavor conversion in the sub-MeV energ y rang e. B o rex i no will provide the first opportunity to ex plore this reg ime with two kinds of da ta : (i ) the value of the 7 Be- flux , and (ii) 'appearance ' e ffe c t s suc h a s time va ria tion of the sola r sign a l . In scenari o s where fl avor survi v al vari es sl owl y wi t h energ y (such as S M A, LM A), the B o re x i no sig n a l is time inde pe nde nt (be s ide s the trivia l se a s ona l 7% R -2 variation) and is a measure of the 7 Be- flux that depends, in principle, on the solar model. However, if the measured sig n al is decidedly at or below the NC 'hard core ' limit, such a result would be a strong indication for conversion to sterile neutrinos. The va c uum osc illa tion sc e n a r io pre d ic ts a ra pid va ria tion of the surviva l probability with the energ y and with the distance from the Sun. This effect is hardly det ect abl e i n S K and i n S NO, ham p ered by an averag e over t h e wi de rang e of energi es i n t h e 8 B- spectrum. B o rex i no, on the other hand, looks for a monoenerge tic neutrino sig n al and it can thus decouple the energy dependence from the pe riodic a l c h a n g e of the Sun-Ea rth dista n c e . B o re x i no ha s a unique a b ility to probe the va c uum osc illa tion sc e n a r io, throug h the de te c tion of a time va ria tion in the sign a l tha t is sig n ific a n tly la rg e r tha n the R -2 effect . The time sc a l e of the va ria tions ra ng e from a se a s ona l to monthly a nd e v e n da ily ba sis for va c uum osc illa tions with m 2 10 - 11 up to 10 -8 (eV/c 2 ) 2 . At somewhat high er m 2 values, 10 -7 (eV/c 2 ) 2 , t h e LO W m a ss MS W scenari o produces a day/ ni g h t e ffe c t . Suc h time va ria tions a r e 'smoking g uns ' for flavor conversion rega rdless


49

whether neutrinos convert to active or sterile species. In either of the above two possibilitie s, B o re x i no would pla y a de c i sive a nd indispe n sa ble role in de monstra ting neutrino flavor conversion, rega rdless of the type of non-standard neutrino that ge ne ra te s it in na ture . A c knowle d ge me nt s Many individuals besides the authors have contributed to the development of the B o rex i no project. W e wish to acknowledge particularly the following who made sign ificant contributions in early stag es: R. Cereseto, N. Darnton, A. F a lg iani, T. Goldbrunner, J . J o chum, M. J ohnson, A. Manco, A. Nostro, S. Pakvasa, M. Pa rodi, A. Pe rotti, G. Pie r i, A. Pre d a , P. Ra gh a v a n, P. Rothsc hild, P. Ulluc c i a nd me mbe r s of IR MM for the me a s ure m e n ts ma de in Ge e l . The c o lla bora tion wa nts to thank the La boratori Nazionali del Gran Sasso [ LNGS] and A. B e ttini for continuous help and support. We sincerely thank the funding bodies: Is tituto Nazionale di F i sica Nucl eare [ INFN] , and Mi ni st ero del l a Uni v ersi t a' e del l a R i cerca S c i e nt i f i ca e Tecnolog ica [ M URST] (Italy); IN 2P3 (F rance); B undesministerium f Ý r B ildung , W i ssenschaft, F o rschung und Technologie [ B MB F ] , Deutsche F o rschungsg emeinschaft [ D F G ] and Max - Planck- Gesellschaft [ M PG] (Germany); the National Science F oundation and B e ll La boratories (USA); the Natural Sciences and Engi neering Research Council of Canada, and ag encies in Hung ary, Poland and Russia for their ge nerous support of this project.


50

Ref eren ces
R. Davis Jr. et al., Phys. Rev. Lett. 20 (1968) 1205 T. A. Kirsten, Rev. Mod. Phys. 71 (1999) 1213 3 J. N. Bahcall et al., (Editors), Solar Neutrinos: The First Thirty Years. Frontiers in Physics, Perseus Press (1995) 4 J. N. Bahcall and M. H. Pinsonneault, Rev. Mod. Phys. 64 (1992) 885; Rev. Mod. Phys. 67 (1995) 781 5 J. N. Bahcall et al., Phys. Lett. B 433 (1998) 1 6 Kamiokande Collaboration, K. S. Hirata et al., Phys. Rev. Lett. 65 (1990) 1297 ; Y. Fukuda et al., Phys. Rev. Lett. 77 (1996) 1683 7 Gallex Collaboration, P. Anselmann et al., Phys. Lett. B285 (1992) 376; Gallex Collaboration, W. Hampel et al., Phys. Lett. B447 (1999) 127 8 Sage Collaboration, A. I. Abazov et al., Phys. Rev. Lett. 67 (1991) 3332; Sage Collaboration, J. N. Abdurashitov et al., Phys. Rev. C60 (1999) 055801 9 B. T. Cleveland et al., Astrophys. J. 496 (1998) 505 10 N. Hata and P. Langacker, Phys. Rev. D49 (1994) 3622; V. Berezinsky et al., Phys. Lett. B365 (1996) 185 11 Superkamiokande Collaboration, Y. Fukuda et al., Phys. Rev. Lett. 81 (1998) 1158 ; Phys. Rev. Lett. 82 (1999) 1810 ; Phys. Rev. Lett. 82 (1999) 2430. 12 SNO Collaboration, J. Boger et al., Nucl. Instr. Meth. Phys. Res. A449 (2000) 172 13 GNO Collaboration, M. Altmann et al., Phys. Lett. B490 (2000) 16 14 R. S. Raghavan, Science 267 (1995) 45 15 T. A. Kirsten, Ann. New York Acad. Sci. 759 (1995) 1; Nuovo Cimento C19 (1996) 821 16 V. Castellani et al., Phys. Rep 281 (1997) 309 17 Borexino Collaboration, C. Arpesella et al., Borexino at Gran Sasso: Proposal for a real time detector for low energy solar neutrinos, G. Bellini, M. Campanella, D. Giugni, and R. S. Raghavan (Eds.) (INFN Milano) (1991); see also: R. S. Raghavan et al., Design Concepts for Borex. ATT Bell Labs Memo 11131-880114-02 (1988) 37 18 R. S. Raghavan, ATT Bell Labs Technical Report 11121-901129-63 (1990) 19 J. N. Bahcall, Neutrino Astrophysics, Cambridge University Press (1989) 20 J. N. Bahcall, from the website http://www.sns.ias.edu/~jnb/ 21 Z. Maki, M. Nakagawa, S. Sakata, Progr. Theor. Phys. 28 (1962) 247; B. Pontecorvo, JETP 53 (1967) 1717 22 S. L. Glashow and L. Krauss, Phys. Lett. B190 (1987) 197 23 S. P. Mikheyev and A. Y. Smirnov, Sov. J. Nucl. Phys. 42 (1985) 913 ; L. Wolfenstein, Phys. Rev. D17 (1978) 2369 24 J. N. Bahcall and P.I. Krastev, Phys. Rev. C56 (1997) 2839 25 L. Wolfenstein, Phys. Rev. D17 (1978) 2369 S. Pakvasa and J. Pantaleone, Phys. Rev. Lett. 65 (1990) 2479 26 M. Cribier et al., Phys. Lett. B188 (1987) 168; A. J. Baltz and J. Weneser, Phys. Rev. D37 (1988) 3364 27 J. N. Bahcall et al., Phys. Rev. D58 (1998) 96016; J. N. Bahcall et al., Phys. Lett. B477 (2000) 401; G. L. Fogli et al., Phys. Rev. D62 (2000) 013002; M. C. Gonzalez-Garcia et al., Nucl. Phys. B 573 (2000) 3 28 G. t'Hooft, Phys. Lett. B37 (1971) 195 29 R. S. Raghavan and R. Beukens, Annual Report, Isotrace Laboratory, University of Toronto (1992) 30 A. DeBari et al., Radiopurity of the Borexino Scintillator, ATT Bell Labs Memo 11121-921015-37 (1992) 31 R. B. Vogelaar et al., Nucl. Instr. Meth. Phys. Res. 372 (1996) 59 32 G. Ranucci et al., Nucl. Instr. Meth. A350 (1994) 358; A354 (1995) 389; A412 (1998) 374 33 T. Goldbrunner et al., J. Rad. Nucl. Chem., 216 (1997) 293 34 R. S. Raghavan (in ref.17); R. von Hentig, Ph. D. Thesis, Technical University Munich (1999) 35 F. Masetti et al., J. Luminescence 68 (1996) 15 36 J. B. Benziger et al., Nucl. Instr. Meth. A417 (1998) 278 37 Borexino Collaboration, G. Alimonti et al., Nucl. Instr. Meth. A406 (1998) 411 38 Borexino Collaboration, G. Alimonti et al., Phys. Lett. B422 (1998) 349 39 Borexino Collaboration, G. Alimonti et al., Astroparticle Phys. 8 (1998) 141 40 Borexino Collaboration, G. Alimonti et al., Nucl. Instr.Meth. A440 (2000) 360 41 G. Ranucci et al., Nucl. Instr. Meth., A333 (1993) 553 42 M. Johnson et al., Nucl. Instr. Meth. A414 (1998) 459; M. Johnson, PhD Thesis, Princeton University (1998) 43 Borexino Collaboration, J. B. Benziger et al., A Proposal for the Borexino Solar Neutrino Experiment, Princeton University Report (1996) 44 M. Balata et al., Nucl. Instr. Meth. A370 (1996) 605; M. G. Giammarchi et al., Ultrapure Water 13 (1996) 59 45 F. X. Hartmann, Borexino Collaboration, Proceed.of the 4th International Solar Neutrinos Conference, W.Hampel (ed.), published by: MPI fÝr Kernphysik, Heidelberg, Germany (1997) 202 46 C. Galbiati, Ph. D. Thesis, University of Milan (1999) 47 H. Simgen, Diploma Thesis, Heidelberg University (2000)
2 1


51
48 49

W. Rau, Ph. D. Thesis, Heidelberg University (1999) L. Cadonati, PhD Thesis, Princeton University (2000) 50 M. Chen et al., Nucl. Instr. Meth. A420 (1999) 189 51 L. Oberauer, Proceed. Vth SFB 375 Ringberg Workshop, Results-Projects-Perspectives, Tegernsee, Germany, Ed. D. Thomas (1998) 13 52 Borexino Collaboration, M. Deutsch et al., Proposal to NSF (1996); L. Oberauer and S. Schoenert, Proceed. IVth SFB 375 Ringberg Workshop, Ed. M. Altmann et al. (1997) 33; M. Chen et al.,(Borexino Collaboration), Proposal to NSF (1998) 53 G. Heusser et al., Appl. Rad. Isot. 52 (2000) 691 54 V. Lagomarsino and G. Testera, Nucl. Instr. Meth. A430 (1999) 435 55 J. Maneira, Proceed. 2nd Intern. Workshop, New Worlds in Astroparticle Physics, Faro, Portugal, Ed.A. Mourao, M. Pimenta, P. Sa (1998) 131 56 S. Bonetti et al., Nucl. Instr. Meth. A329 (1993) 57 For details see: http://www.info.cern.ch/asd/geant4/description.html 58 M. Wojcik and W. Wlazlo, Nucl. Instr. Meth. A345 (1994) 351 59 W. Rau and G. Heusser, Appl. Rad. Isot. 53 (2000) 371 60 M. Wojcik et al., Nucl. Instr. Meth. A449 (2000) 158 61 C. Arpesella, Appl. Rad. Isot. 4 (1996) 991 62 H. Neder et al., Appl. Rad. Isot. 53 (2000) 191 63 P. R. Trincherini et al., J. Trace Microprobe Techniques, 10 (1992) 295 64 T. Hagner et al., Astroparticle Phys. 14 (2000) 33; B. Heisinger et al., TUM/SFB 375/335 (1997) 65 Y. Suzuki, Proc. Neutrino 2000 (Sudbury), Nucl. Phys. B (Proceed. Suppl.), to be published. Meanwhile see: http://nu2000.sno.laurentian.ca/Y.Suzuki/index.html 66 R. S. Raghavan et al., Phys. Rev. D44 (1991) 3786 67 Borexino Collaboration, Physics Beyond Solar Neutrinos, Addendum to Borexino Proposal (1999) 68 G. Raffelt, Phys. Rev. D39 (1989) 2066; L. Oberauer et al., Astroparticle Phys. 1 (1993) 377 69 E. Akhmedov, Sov. J. Nucl. Phys. 47 (1987) 689; C. S. Lim and W. Marciano, Phys. Rev. D37 (1988) 1368 70 C. S. Lim et al., Phys. Lett, B243 (1990) 389 71 M. Aglietta et al., JETP Lett. 63 (1996) 791 72 R. S. Raghavan et al., Phys. Rev. Lett. 80 (1998) 635; Rostschild et al., Geophys. Res. Lett. 25 (1998) 1083 73 S. Schoenert, Nucl. Phys. B (Proceed. Suppl.), 70 (1999) 195 74 L. Cadonati et al., Astroparticle Phys. Nov. 2000 (subm.) 75 Gallex Collaboration, W. Hampel et al., Phys. Lett. B420 (1998) 114 76 Gallex Collaboration, W. Hampel et al., Phys. Lett. B436 (1998) 158 77 R. S. Raghavan, Proc. 25th Int. Conf. High Energy Phys., Singapore 1990, Ed. K. K. Phua and Y. Yamaguchi, (Phys. Soc. Japan) 1 (1991) 482 78 A. Ianni, D. Montanino, Astroparticle Phys. 10 (1999) 331 79 R. S. Raghavan, Phys. Rev. Lett. 72 (1994) 1411; B. Caccianiga and M. Giammarchi, Astroparticle Phys. 14 (2000) 15. 80 Super-Kamiokande Collaboration, Y. Fukuda et al., Phys. Rev. Lett. 82 (1999) 2644 ; Phys. Lett. B467 (1999) 185 ; T. Kajita and Y. Totsuka , Rev. Mod. Phys. (2000) submitted.