Письма В Эчая Particles and Nuclei, Letters

XJ02A0050

ПИСЬМА В ЭЧАЯ

PARTICLES AND NUCLEI, LETTERS

6[109]-2001

116 116 Status of Cd Double p Decay Study with CdWO4Scintillators • New Limits on 2(3 Processes in 40Ca and 46Ca •

by Using Low Radioactive CaF2(Eu) Crystal Scintillators The Single State Dominance in 2v[ip-Decay transitions • to Excited 0+ and 2+ Final States Present Status of the MONOLITH Project • Technique of Neutrino-Induced Muon Detection • on the Earth Surface High-Sensitive Spectrometer of Fast Neutrons • and the Results of Fast Neutron Background Flux Measurements at the Gallium-Germanium Solar Neutrino Experiment (SAGE) New Experimental Limits on the Electron Stability G and Excitation of Nuclear Levels in 23Na, 127I and 129Xe Induced by the Electron Decay on the Atomic Shell Element-Loaded Organic Scintillators for Neutron • and Neutrino Physics A Summary of NANPino 2000 •

Издательский отдел ОИЯИ ДУБНА

JINR Publishing Department DUBNA РЕДАКЦИОННЫЙ СОВЕТ

Д. В. Ширков — председатель A. Н. Сисакян — зам. председателя П. И. Зарубин — зам. председателя С. Г. Стецеико — ученый секретарь

B. А. Бедняков А. А. Бельков И. В. Богуславский М. К. Гайдаров Ю. П. Гангрский Ш. Гмуца A. Е. Дорохов C. Дубничка B. И. Журавлев И.Звара — члены совета П. С. Исаев И. Натканец C. Олариу Е. Б. Плеханов A. Г. Попеко B. В. Сиколенко Т. А. Стриж Г. Н. Тимошенко Д. И. Хубуа

EDITORIAL BOARD

D. V. Shirkov — Chairman А. N. Sissakian — Vice-Chairman P. I. Zarubin — Vice-Chairman S. G. Stetsenko — Scientific Secretary

V. A. Bednyakov A. A. Belkov I. V. Boguslavsky М. К. Gaidarov Yu. P. Gangrsky S. Gmuca A. E. Dorokhov S. Dubnicka V. I. Zhuravlev I. Zvara — Members of the Board P. S.Isaev I. Natkaniec S. Olariu E. B. Plekhanov A. G. Popeko V. V. Sikolenko T. A. Strizh G. N. Timoshenko D. I. Khubua

6[109]-2001

ПИСЬМА О ФИЗИКЕ ЭЛЕМЕНТАРНЫХ ЧАСТИЦ И АТОМНОГО ЯДРА

PHYSICS OF PARTICLES AND NUCLEI, LETTERS

Дубна 2001 В журнале «Письма о физике элементарных частиц и атомного ядра», кратко — «Письма в ЭЧАЯ», публикуются статьи, содержащие результаты ориги- нальных теоретических, экспериментальных, научно-технических, методических и прикладных исследований. Содержание публикуемых статей определяется тема- тикой научных исследований в ОИЯИ: теоретическая физика, физика элементар- ных частиц, релятивистская ядерная физика, физика атомного ядра и связанные вопросы общей физики, нейтронная физика, физика конденсированных сред, физика низких температур и криогенная техника, физика и техника ускорителей, методика физического эксперимента, компьютерные технологии в физике, прикладные работы по всем перечисленным разделам физики, включая радиобио- логию, экологию и ядерную медицину. Журнал зарегистрирован в Министерстве Российской Федерации по делам печати, телерадиовещания и средств массовых коммуникаций. Издателем журна- ла является Объединенный институт ядерных исследований. Журнал выходит шесть раз в год.

The journal PHYSICS of PARTICLES and NUCLEI, LETTERS, brief name PARTICLES and NUCLEI, LETTERS, publishes the articles with results of the original theoretical, experimental, scientific-technical, methodical and applied research. Subject-matter of articles covers the principal fields of research at JINR: theoretical physics, elementary particle physics, relativistic nuclear physics, nuclear physics and related problems in other branches of physics, neutron physics, condensed matter physics, physics and technique at low temperature, physics and technique of accelerators, physical experimental instruments and methods, physical computer experiments, applied research in these branches of physics and radiology, ecology and nuclear medicine. : The journal is registered in the Ministry of the Russian Federation for Press and is published bimonthly by the Joint Institute for Nuclear Research.

References to the articles of the PARTICLES and NUCLEI LETTERS should contain: — names and initials of authors, — title of journal, — year of publication, — publication index, — page number.

For example: Alexeev G.D. II Part, and Nucl., Lett. 2000. No.5[102]. P.5. PROCEEDINGS OF THE INTERNATIONAL WORKSHOP

«NON-ACCELERATOR NEW PHYSICS IN NEUTRINO OBSERVATIONS»

Dubna, July 19-22, 2000

Edited by V.A.Bednyakov

NANPino-2000

ТРУДЫ МЕЖДУНАРОДНОЙ КОНФЕРЕНЦИИ «НЕУСКОРИТЕЛЬНАЯ НОВАЯ ФИЗИКА В НЕЙТРИННЫХ НАБЛЮДАЕМЫХ»

Дубна, 19-22 июля 2000 г.

Научный редактор В.А.Бедняков Topics: • Neutrino as a key to physics beyond the Standard Model, present status and future prospects; • Solar and atmospheric neutrino anomalies as manifestations of non- zero neutrino masses and mixings; • Reactor and accelerator neutrino experiments; • Studies of the neutrino properties in underground experiments; • Neutrino astrophysics and cosmology; • Experimental facilities for neutrino observations.

Local Organizing Committee: V.A.Bednyakov, V.B.Brudanin, S.G.Kovalenko, V.E.Kovalenko, A.A.Smolnikov, E.A.Yakushev

NANPino-2000 is sponsored by the Joint Institute for Nuclear Research, the Russian Foun- dation for Basic Research, the Institute for Nuclear Research of the Russian Academy of Sciences and UNESCO.

NANPino-2000

Тематика: • нейтрино как ключ к физике за рамками стандартной модели, современное состояние и перспективы; • солнечные и атмосферные нейтринные аномалии как проявление ненулевых нейтринных масс и смешиваний; • реакторные и ускорительные нейтринные эксперименты; • исследование свойств нейтрино в подземных лабораториях; • нейтринная астрофизика и космология; • экспериментальное оборудование для исследования физики нейтрино.

Местный организационный комитет: В.А.Бедняков, В.Б.Бруданин, С.Г.Коваленко, В.Э.Коваленко, А.А.Смольников, Е.А.Якушев

Конференция проведена при поддержке Объединенного института ядерных ис- следований, Российского фонда фундаментальных исследований, Института ядерных исследований РАН и ЮНЕСКО. СОДЕРЖАНИЕ CONTENTS

P. G. Bizzeti, F. A. Danevich, T. F. Fazzini, A. Sh. Georgadze, V. V. Kobychev, B. N. Kropivyansky, P. R. Maurenzig, A. S. Nikolaiko, O. A. Ponkratenko, V. I. Tretyak, S. Yu. Zdesenko, Yu. G. Zdesenko m n6 Status of Cd Double /3 Decay Study with CdWO4ScintilIators П. Г. Биззети, Ф. А. Даневич, Т. Ф. Фаззини, А. Ш. Георгадзе, В. В. Кобычев, Б. Н. Кропивянский, П. Р. Маурензиг, А. С. Николайко, О. А. Понкратенко, В. И. Третьяк, С. Ю. Здесенко, Ю. Г. Здесенко ll6 U6 Исследование 2/3-распада Cd с помощью сцинтилляторов CdWO4 . .

P. Belli, R. Bernabei, С. J. Dai, F. Grianti, H. L. He, G. Ignesti, A. Incicchitti, H. H. Kuang, J. M. Ma, F. Montecchia, O. A. Ponkratenko, D. Prosperi, V. I. Tretyak, Yu. G. Zdesenko New Limits on 2/3 Processes in 40Ca and 46Ca by Using Low Radioactive CaF2 (Eu) Crystal Scintillators П. Белли, Р. Бернаби, С. Д. Дай, Ф. Грианти, X. Л. Хе, Г. Игнести, А. Инсичитги, X. X. Куанг, Д. М. Ма, Ф. Монтечиа, О. А. Понкратенко, Д. Проспери, B. И. Третьяк, Ю. Г. Здесенко Новые экспериментальные пределы на 2/3-распады 40Саи 46Са, полученные с использованием низкорадиоактивных

кристаллических сцинтилляторов CaF2 (Eu) 18

S. V. Semenov, F. Simkovic, P. Domin The Single State Dominance in 2v/?/?-Decay Transitions to Excited 0+ and 2+ Final States C. В. Семенов, Ф. Шимковиц, П. Домин Доминантность основного состояния промежуточного ядра в 2уДб-переходах в возбужденные 0+ и 2+ конечные состояния 26

A. A. Petrukhin Present Status of the MONOLITH Project А. А. Петрухин Современное состояние проекта МОНОЛИТ 32 V. M. Aynutdinov, V. V. Kindin, К. G. Kompaniets, A. A. Pertukhin, D. A. Room, V. V. Shutenko, A. V. Stepanov, 1.1. Yashin Technique of Neutrino-Induced Muon Detection on the Earth Surface В. М. Айнутдинов, В. В. Киндин, К. Г. Компаниец, А. А. Петрухин, Д. А. Роом, В. В. Шутенко, А. В. Степанов, И. И. Яшин Методика регистрации мюонов от нейтрино на поверхности Земли 43

J. N. Abdurashitov, V. N. Gavrin, A. V. Kalikhov, V. L. Matushko, A. A. Shikhin, V. E. Yanits, O. S. Zaborskaia High-Sensitive Spectrometer of Fast Neutrons and the Results of Fast Neutron Background Flux Measurements at the Gallium-Germanium Solar Neutrino Experiment (SAGE) Д. Н. Абдурашитов, В. Н. Гаврин, А. В. Калихов, В. Л. Матушко, А. А. Шихин, В. Е. Янц, О. С. Заборская Высокочувствительный спектрометр быстрых нейтронов и результаты измерения фона быстрых нейтронов в галлий-германиевом нейтринном эксперименте (SAGE) ...... 53

P. Belli, R. Bernabei, С. J. Dai, H. L. He, G. Ignesti, A. Incicchitti, H. H. Kuang, J. M. Ma, F. Montecchia, О. A. Ponkratenko, D. Prosperi, V. I. Tretyak, Yu. G. Zdesenko New Experimental Limits on the Electron Stability and Excitation of Nuclear Levels in " Na, m I and ш Хе Induced by the Electron Decay on the Atomic Shell П. Белли, P. Бернаби, С. Д. Дай, X. Л. Хе, Г. Игнести, А. Инсичитти, X. X. Куанг, Д. М. Ма, Ф. Монтечиа, О. А. Понкратенко, Д. Проспери, В. И. Третьяк, Ю. Г. Здесенко Новые экспериментальные пределы на стабильность электрона и возбуждение ядерных уровней в 23 Na, ш I и m Хе, вызванных распадом электрона на атомной оболочке 58

V. В. Brudanin, V. I. Bregadze, N. A. Gundorin, D. V. Filossofov, О. I. Kochetov, I. B. Nemtchenok, A. A. Smolnikov, S. I. Vasiliev Element-Loaded Organic Scintillators for Neutron and Neutrino Physics В. Б. Бруданин, В. И. Брегадзе, Н. А. Гундорин, Д. В. Философов, О. И. Кочетов, И. Б. Немченок, А. А. Смольников, С. И. Васильев Элементосодержащие органические сцинтилляторы для нейтронной и нейтринной физики 69

A Summary of NANPino 2000 (J. D. Vergados) Итоги NANPino 2000 (Дж. Д. Вергадос) 78 Ill1 XJ0200051

Письма в ЭЧАЯ. 2001. №6[109] Particles and Nuclei, Letters. 2001. No.6[109]

УДК539.165 : : ; :;

STATUS OF 116Cd DOUBLE 0 DECAY STUDY WITH 116 CdWO4 SCINTILLATORS P. G.BizzetiК F.A.Danevicha, T.F.Fazzinib, A.Sh. GeorgadzeQ, V. V.Kobycheva, B.N.Kropivyanskya, P.R. Maurenzigb, A. S.Nikolaikoa, O. A. Ponkratenkoa , V. I.Tretyak a, S. Yu. Zdesenko a, Yu. G. Zdesenko a : . ° Institute for Nuclear Research, MSP 03680 Kiev, Ukraine 6 Dip. di Fisica, Universita di Firenze and INFN, 50125 Firenze, Italy -

A new set-up with four enriched 116CdWC>4 scintillators with the mass of 0.34 kg is presented.

The measured half-life for 2v2p decay as well as the Tx/i limits for different modes of neutrinoless 2/? decay of 116Cd are reported. Представлена новая экспериментальная установка, состоящая из четырёх сцинтилляторов 46CdWC>4 с массой 0,34 кг. Приведены экспериментальные пределы для 2i/2/3- и безнейтринной мод распада 116Cd.

INTRODUCTION

Neutrinoless (0v) double /3 decay, is forbidden in the Standard Model (SM) since it violates lepton number (L) conservation. However many extensions of the SM incorporate L violating interactions and thus could lead to the 0v2(3 decay [1,2]. Currently, besides conventional neutrino {u) exchange mechanism, there are many other possibilities to trigger this process [2]. Therefore, at present 0V2/? decay is considered as a powerful test of new physical effects beyond the SM, and even the absence of this process would help to restrict or harrow this wide choice of theoretical models. ' ••••:. With the aim to enlarge the number of 2(3 decay candidate nuclides studied at a sensitivity comparable with that: for 76Ge [3,4] and 136Xe [5] (neutrino mass limit of 0.5-2 eV), cadmium tungstate crystal scintillators, enriched in 116Cd to 83%, were developed and exploited in u6Cd research [6,7]. The measurements were carried out in the Solotvina Underground Laboratory in a salt mine 430 m underground (~ 1000 m w.e.) [8]. In the first phase of 116 the experiment only one CdWO4 crystal (121 g) was used. The background rate in the energy range 2.7-2.9 MeV(Q2/3 = 2805 keV [9]) was equal to « o:6 cotints/y- kg • keV. With 116 22 19175 h statistics the half-life limit for Oi/2/З decay of Cd was set as T1/2(0v) > 3.2-10 у

(9.0% C.L.), which corresponds to the restriction on the neutrino mass mv < 3.9 eV [7]. Limits on 0i/2/? decay with emission of one (Ml) or two (M2) Majorons were obtained, too: 21 20 T1/2(0vMl) > 1.2-.10 у and Т1/2(0Щ2) > 2.6-:10 у (90% C.L.) [10]. In the present, paper new and advanced results of 116Cd research obtained with the help of an upgraded apparatus are described. , 8 Bizzeti P. G. et al.

116 1. NEW SET-UP WITH FOUR CdWO4 DETECTORS

116 1.1. Set-up and Measurements. In the new apparatus four enriched CdWO4 crystals (total mass 339 g) are viewed by the PMT (EMI9390) through one light-guide 10 cm in diameter and 55 cm long, which is glued of two parts: quartz 25 cm long and plastic 116 scintillator 30 cm long. The CdWO4 crystals are surrounded by an active shield made of

15 natural CdWO4 scintillators with total mass of 20.6 kg. The veto crystals are viewed — by a low background PMT (017 cm) — through an active plastic light-guide (017x49 cm).

In turn the whole array of CdWO4 counters is placed inside an additional active shield made of polystyrene-based plastic scintillator with dimensions 40x40x95 cm. Together with both 116 active light-guides a complete 4тг active shield of the main CdWO4 detectors is provided. The outer passive shield consists Counts/20 keV of HP copper (thickness 3-6 cm),

105- lead (22.5-30 cm) and polyethylene V p'ucd.e^siekev (16 cm). Two plastic scintillators

4 (120x130x3 cm) are installed above 10 - \ 137Cs,662keV Y the passive shield to provide a cosmic ft I 3 40 muons veto. The set-up is isolated care- V\ Y K, 1461 ceV w - fully against air penetration. All mate- io2- rials used in the installation were pre- u6 viously tested and selected for low ra- 2p Cd, Q2fs = 2805 keV 10- l dioactive impurities in order to reduce their contributions to background. 1 Twiiii p.i 1 • The data acquisition system is based J HWUj „ . on two IBM personal computers (PC) 1 ю- < I. 1l\ l IIЛI "(IIliTlI f \i 1 and a CAMAC crate with electronic 1000 2000 3000 4000 5000 units. For each event the amplitude of a AkeV signal, its arrival time and the additional tags (the coincidence between different Fig. 1. Background spectrum of u6CdWO4 detectors detectors; the signal of radio-noise de- (339 g) measured in the set-up with four enriched crys- tection system; triggers for light emit- tals during 4629 h (solid histogram). The old data ob- ting diode (LED) and pulse shape digi- tained with one I16CdWO crystal (121 g; 19986 h) is 4 tizer) are stored on the hard disc of the shown for comparison (thin histogram; the data are nor- first computer. The second computer malized to 4629 h and mass of the new detector). The records the pulse shape (in 2048 chan- background components used for fit in the energy region n6 nels with a 50 ns channel's width) of 900-2900 keV: a) 2v2/3 decay of Cd (fit value is 116 the CdWO scintillators in the en- T (2v) = 2.6(l)-1019 y); b) 40K inside the 116CdWO 4 1/2 4 ergy range 0.25-5 MeV. detector (activity value from the fit is 0.8(2) mBq-kg"1); 40 The energy scale and resolution c) K in the shielding CdWO4 crystals (fit value is 1 226 232 of the main detector were determined 2.1(3) mBq'kg- ); d) Ra and Th contamination in the measurements with different of PMTs 7 sources (22Na, 40K, 60Co, 137Cs, 241 The energy dependence of the resolution can be expressed as FWHM(keV) = \/-226 + 16.6.E + 6.42 • 10-3£2, where energy E is in keV. Also, the relative light yield for a particles as compared with that for electrons (a/{3 ratio) and energy resolution were U6 116 Status of Cd Double /? Decay Study with CdWO4 Scintillators 9 measured with a source (241Am) and corrected by using the time-amplitude analysis (see 6 below) as following: a//3 = 0.15(1) +7Л0~ Еа and FWHMa(keV) = О.ШЕа-фа is in keV). The routine calibration is carried out with a 207Bi and 232Th 7 sources. The dead time of the spectrometer and data acquisition is monitored permanently with the help of an LED optically connected to the main PMT. The actual dead time value is « 4.2 %. The background spectrum measured during 4629 h with four 116CdWO4 crystals is given in Fig. 1, where the old data obtained with one 116CdWO4 crystal of 121 g are also shown for comparison. In the energy region 2.5-3.2 MeV the background rate is reduced to a value of 0.03 counts/y • kg • keV. 1.2. Time-Amplitude Analysis of the Data. The energy and arrival time of each event can be used for the analysis and selection of some decay chains in 232Th, 235U and 238U families. As an example we consider here in detail the time-amplitude analysis of the following sequence

Counts/20keV ... . . •....•

500 1000 1500 2000 2500 3000 Energy of 1 st events, keV Counts/20 keV

500 1000 1500 2000 2500 3000 Energy of 2nd events, keV Counts/10 ms

228Th activity 38(3) nBq • kg"

0.2 0.4 0.6 0.8 Time interval detween 1 and 2 events, s

Fig. 2. The energy spectra of the first (a) and second (b) a particles from the Rn 216 Po 212 Pb chain selected by time-amplitude analysis from 116CdWO4 data. Their equivalent energies in the /З/7 energy scale are near 5 times smaller because the relative light yield for a particles as compared with that for electrons (a//3 ratio) is « 0.2. c) Time distribution between the first and second events together with exponential fit (T1/2 = 0.15(1) s, while the table value is T1/2 = 0.145(2) s [16]}

232 220 216 of a decays from Th family: Rn (Qa '= 6.40 MeV, T1/2 = 55.6 s) -> Po (Qa = 212 6.91 MeV, T1/2 = 0.145 s) -4 Pb. The events in the energy region 0.7-1.8 MeV were used 10 Bizzeti P. G. et al. as triggers. Then all events (within 0.9-1.9 MeV) following the triggers in the time interval 10-1000 ms were selected. The spectra of the 220Rn and 216Po a decays obtained in this way from data — as well as the distribution of the time intervals between the first and second events — are presented in Fig. 2. Using these results and taking into account the efficiency of the time-amplitude analysis and the number of accidental coincidences, the determined 228 232 116 1 activity of Th ( Th family) inside the CdWO4 crystals is as low as 38(3) /xBq -kg" . The same technique was applied to the sequence of a decays from the 235U family: 211 238 (2i9Rn _„ 2i5Po (Ti/2 = i;78 ms) -> Pb) and sequence of /3 and a decays from the U 214 214 210 1 family: Bi -* Po (T1/2 = 164.3 (is) -> Pb. Activity of 5.5(14) /xBq-kg- for the 227Ac (the 235U family) and limit < 5 /iBq-kg"1 for the 226Ra chain (238U family) in the 116 232 235 238 CdWO4 crystals were set. Finally, all couples of events found for Th, U and U families were eliminated from the measured data. 1.3. Pulse-Shape Discrimination. Due to different shapes of scintillation signal for various kinds of sources (a particles, protons, 7 quanta and cosmic muons were investigated), the pulse-shape (PS) discrimination method based on the optimal digital filter [11] was developed and clear discrimination between 7 rays (electrons) and a particles was achieved [12].

Counts/20 keV 102 J

1500 2000 2500 3000 3500 4000 E, keV

116 Fig. 3. a) Initial spectrum of CdWO4 crystals (339 g, 4629 h) in anticoincidence with shielding detectors without pulse-shape discrimination; b) PS selected /З/7 events (see text); c) the difference between spectra in Figs. 3,a and 3,b together with the fit by the response function for 212Bi —* 212Po 208 228 116 -> Pb decay chain. The fit value is 37(4) ^Bq-kg~l for Th activity inside CdWO4 crystals 116 116 Status o/ Cd Double /3 Decay Study with CdWO4 Scintillators 11

The pulse, shapes of enriched crystals were measured for a particles with an 241 Am source and for 7 rays with 60Co, 137Cs, 207Bi and 232Th sources in the special calibration runs. The numerical characteristics of the pulse shape (shape indicator, SI, see for more details Ref. 12) are well described by a Gaussian functions, whose mean values and standard deviations aa and 1 cr7 have a slight energy dependence . It allows us to determine the efficiency of the PS event selection for the different chosen intervals of SI values. The PS selection technique ensures the very important possibility to discriminate «illegal» events: double pulses, a events, etc., and thus to suppress background. . •••..-. Since the shape indicator characterizes the full signal, it is also useful to examine the pulse front edge. For example, it was found that at least 99% of «pure» 7 events (measured with calibration 232Th source) satisfy the following restriction on pulse rise time: At fas) < 12 1.24 - 0.5E-, + 0.078Я , where Е7 is dimensionless variable expressed in MeV. Hence, this filter was applied to the background data, and all events, which do not pass the test, were excluded from the residual P/j spectrum. , ;.•'-•. The results of PS analysis of the >• data are presented in Fig. 3. The Counts/20 keV initial (without PS selection) spec- 500- U6 trum of the CdWO4 scintillators : у not rejected by PS analysis in the energy region 1.2-4 MeV — 400- collected in anticoincidence with active shield — is depicted in Fig. 3, a, while the spectrum after PS selec-' 300- tion of the /3/7 events, whose SI lies in the interval SI — 3.0ст < SI 7 7 200- < SI7 +, 2.4• from chains in Th and U families. The total a activity Q 1/2 116 1 208 Pb cannot be time resolved in the of the CdWO4 crystals is derived as 1.4(3) mBqkg"

CdWO4 scintillator and will result in one event. To determine the residual activity of 228Th in the crystals, the response function 116 212 212 208 of CdWO4 detectors for the Bi -4 Po ->• Pb chain was simulated with the help of

5 'For the 7-s (300-3200 keV) SI7= 18.09 - (4.5-I0- B7), cr7= 2.61 - (A.l-XQ-^E-,) + 707/JB7, while for 4 the a particles (4000-6000 keV) Sl«= 29.0;

GEANT3.21 code [13] and event generator DECAY4 [14]. The simulated function is shown in Fig. 3, c. The high energy part of the experimental spectrum is well reproduced (x2 = 1.3) by the expected response for 212Bi —» 212Po —» 208Pb decays1. Corresponding activity 228 116 of Th inside the CdWO4 crystals, deduced from the fit in the 1.9-3.7 MeV energy region, is 37(4) fiBq kg"1, that is in good agreement with the value determined by the time- amplitude analysis of the chain 220Rn —> 216Po -> 212Pb. Besides, the front edge analysis of 80 events with the energy 2.0-4.2 MeV (SI > SI7 + 2.54ст7; At > 0.2 fi$) was fulfilled and the half-life derived from the average time delay between the first and second part of the 212 signal is T1/2 = 0.31(6) ps, in agreement with the Po table value T1/2 = 0.299(2) /is [15]. Figure 4 represents the spectrum after PS selection of the background events, whose SI lies in the interval SI7 + 2.4cr7 < SI < SIa + 2.4<7Q (яа 90 % of a events). The obtained distribution with maximum at 0.95 MeV is well reproduced by the model, which includes a particles from chains in 232Th and 238U families (under assumption that secular radioactive 116 equilibriums are broken). The total a activity of the CdWO4 crystals deduced from Fig. 4 is 1.4(3) mBq-kg"1.

2. RESULTS AND DISCUSSION

2.1. Two-Neutrino Double Beta Decay of 116Cd. To determine the half-life of two- neutrino 20 decay of 116Cd, the background was simulated by the Monte Carlo method. In addition to 116Cd 2i/20 decay distribution, only three components shown in Fig. 1 were used to build up the background model: 40K contamination

of the enriched and natural CdWO4 scintillators and external 7 background caused by 232Th and 238U contamination of the PMTs2. This simple background model describes experimental data in the energy interval 900- 2900 keV reasonably well (x2 =1.3) and gives the following results: the activities of 40K inside the enriched and natural CdWC>4 crystals are equal 1000 1500 2000 2500 1 E,keV to 0.8(2) and 2.1(3) mBq-kg- , respectively; the half-life of two-neutrino 116 Fig. 5. The 2i/2@ decay Kurie plot and its fit by the 2/? decay of Cd is Tl/2(2v) = straight line in 1100-2400 keV region 2.6(1) • 1019 y. Taking advantage of the high statistics in our experiment (approximately 3600 events of 116Cd two-neutrino 2/3 decay are

'The rest of spectrum below 1.9 MeV (Fig. 3, c) can be explained as high energy tail of the PS selected a particles (see Fig. 4). 2The radioactive impurities of all PMTs used in the installation were previously measured by R&D low background set-up as (0.4-2.2) Bq/PMT and (0.1-0.2) Bq/PMT for 226Ra and 228Th activity, respectively [10]. n6 116 Status of Cd Double /3 Decay Study with CdWO4 Scintillators 13

contained within the interval 900-2900 keV), we can prove our model with the help of experimental 2i/2/3 decay Kurie plot: K{e) = [S(e)/{(e4 + 10e3 + 40e2 + бОе + ЗО^}]1/5, where S is the number of events with the energy e (in electron mass units) in the experimental spectrum after background subtraction. For the real 2u2/3 decay events such a Kurie

plot should be the straight line K{e) ~ (Q2p — £•), where Q2p is the 2/9 energy release. The experimental Kurie plot (Fig. 5) is well fitted in the region 1.1-2.4 MeV by the straight line

with Q2p = 2790(87) keV (table value is Q2/? = 2805(4) keV). Taking into account the energy resolution of the detector, fit in the energy region 1.2-2.8 MeV yields a very similar 19 value Q2/3 = 2779(52) keV and half-life corresponding to T1/2(2i/) = 2.5(3)-10 y; thus justifying our assumption that experimental data in the region above 1.2 MeV are related mainly with 116Cd two-neutrino 2/5 decay.

Table 1. Different origins of the systematical uncertainties and their contributions to the half-life value of 116Cd two-neutrino 2/3 decay ••*' Origin of the systematical error Value range Contribution to 19 T1/2(2i/) value, 10 y Life measuring time 96±* % . + 0.05,-0.2 Efficiency of PS analysis 98lg.% + 0.05,-0.3 Detection efficiency of 2^2/3 decay 96±4 % ±0.1 (GEANT model uncertainty)

9oSr_?oY impurity in n6CdWO4 < 0.17 mBq•kg"1 + 0.5 23 m lie 1 1 Pa impurity in CdWO4 < 0.19 mBq•kg" + 0.3

To estimate systematical uncertainties of the measured half-life, different origins of errors were taken into account, whose contributions are listed in the Table. The final value is equal to:

19 : T1/2(2b) = 2.6±0.1(stat.)J:°;I(syst.)-10 y. _ ' ' " ;

19 • Our result is in agreement with those measured earlier (Ti/2(2i/) = 2.6loJ " Ю у [16] 19 and Г1/2(2г/) = 2.7+o4(stat.)+og(syst.)• 10 у [6]) and disagrees to some extent with the 19 value Tl/2(2v) = 3.75 ± 0.35(stat.) ±0.21(syst.) • 10 у from Ref. i7'. 2.2. New Limits for 0i/2fi Decay of 116Cd. The high energy part of the experimental u6 spectrum of the CdWO4 crystals measured in anticoincidence with the shielding detectors and after the time-amplitude and pulse-shape selection is shown in Fig. 6. The peak of Ог/2/З decay is absent, thus from the data we obtain a lower limit of the half- 23 116 life: lim T1/2 = ln2iV?7t/lim5, where N = 4.66.- 10 is the number of Cd nuclei; t is the measuring time (t = 4629 h); ц is the total detection efficiency for Ог/2/З decay, and lim,5

'Note, that in [17] the quite small detection efficiency (1.73 %) was calculated by the Monte Carlo method without experimental test, thus perhaps systematical errorcould be higher than the quoted value. ' 14 Bizzeti P. G. et al. is the number of events in the peak which can be excluded with a given confidence level. The value of the detection efficiency 77мс = 0.83 was calculated by the DECAY4 and GEANT3.21 codes, while the efficiency of the PS analysis TJPS = 0.98 was determined as described above, thus the total efficiency ц = т?МС • 7?ps = 0.81. To estimate limS energy interval 2.6- 3.1 MeV (containing 91 % of Oi/2/З peak, where there is only one measured event) was considered. The expected background in the same energy region is 3.2^^;J counts (1.9 ± 0.7 2 events from PMT contamination; 0.4±0.1 events from 2v2/3 distribution; 0.9* ) 9 counts from mentioned 212Bi -+ 212Po -»•208 Pb chain). Following the PDG recommendation [18,19] we can derive from these numbers the excluded limit as lim S = 1.8(0.5) with 90 % (68 %) C.L., which leads to half-life limits for neutrinoless 2/3 decay of u6Cd:

23 Г1/2(0г/2/3) > 0.7(2.5) • 10 y, 90 % (68 %) C.L.

Using calculations [20], one can Counts/20 keV obtain restrictions on the neutrino mass and right-handed admixtures in the weak interaction: m < 3.0 eV, 2 v 10 77 < 3.9-10~8, A < 3.4• 10-6 at 90% C.L., and neglecting right-handed

contribution mv < 2.6(1.4) eV at 10 - 90 % (68 %) C.L. On the basis of calculations [17] we get a similar result: т„ < 2.4(1.3) eV at 90% (68%) C.L. In accordance with Ref. 21 the value of the .R-parity violating parameter of minimal SUSY standard Ov20Ml 1 model is restricted by our T1/2 limit ю- to e < 8.8(6.4) • 10~4 at 90 % (68 %) 1000 2000 3000 4000 £,keV C.L. (calculations [22] give more stringent restrictions: e < 3.4(2.4) • 4 Fig. 6. Part of experimental spectrum of the 116CdWO4 ИГ ). detectors measured during 4629 h (histogram) together with Excited levels of u6Sn with 19 the fit from 2i/20 contribution (T1/2 = 2.6 • 10 y). The E\ew < Qip can be also populated smooth curves Of2/?M1 and 0i>2/?M2 are excluded with in Ог/2/З decay of 116Cd. In this case 90 % C.L. distributions of Oi/Ml and 0г/М2 decay of116 Cd one or several 7 quanta, conversion 21 20 + with Г1/2 = 3.7 • 10 у and T1/2 = 5.9 • 10 y, respec-electrons and/or e e~ pairs will be tively. In the insert the expected peak from 01/2/3 decay emitted in a deexcitation process, in with 7\/2(0t/) = 1.0 • 1022 у is shown together with the addition to two electrons emitted in excluded (90 % C.L.) distribution (solid histogram) with 2/3 decay. The full absorption of all 22 T1/2(0u) = 7.0 • 10 у emitted particles should result in the peak with E = Q2/3. Calculated with the help of GEANT3.21 and DECAY4 codes full peak efficiencies for Ог/2/З decay to the first 116 and second excited levels of Sn (2f with Elev = 1294 keV and Of with E\ev = 1757 keV) are: 77(2^") = 0.14 and r?(Of) = 0.07. These numbers and the value of lim 5 = 1.8(0.5) with 90 % (68 %) C.L. (determined for the g.s. —> g.s. transition) give the following restrictions on half-lives of 116Cd neutrinoless 2/3 decay to excited levels of 116Sn: 116 116 Status of Cd Double /3 Decay Study with CdWO4 Scintillators 15

22 T1/2(g.s.-* 2+) > 1.3(4.8)•Ю y, • 90%(68%)C.L.,,;.: ... , 22 V T1/2(g.s.-t Of) ^ 0.7(2-4)-10 y, 90%(68%)C.L. , /. • .;

To obtain half-life limits for 0v2(3 decay with one (two) Majoron(s) emission the.measured spectrum was.fitted in the energy region 1.6-2.8 MeV for OfMl mode (1.6-2.6 MeV for 0i/M2) by using only three theoretical, distributions: 7 background from measured PMT-s. contamination (22CRa and 232Th chains) and two-neutrino.2/3 decay of 116Cd, as background, and 0^2/? decay with one (two) Majoron(s) emission, as effect. With this simple model the X2 value was equal to 1.1 both for Oz/Ml and 0^M2 fits. As a result, the number of events under a theoretical O^Ml curve was determined,as 9 ± 21, giving no statistical evidence for the effect. It leads to an upper limit of 41(26) events at 90% (68 %) C.L., that together with an efficiency value 77 = 0.905 corresponds to the half-life limit: ,, •,

21 Tl/2(0uMl) > 3.7(5.9) • 10 y, 90% (68%) C.L.

A similar procedure for 0i/2/3 decay with two Majorons emission gives:

20 Ti/2(Oi/M2) > 5.9(9.4) -10 y, 90%(68%)C.L.

Both the present half-life limits are more stringent than those established in our previous measurement during 19986 h [10] and in the NEMO experiment [17]. ••••.• The probability of neutrinoless 2/3 decay with Majoron emission can, be expressed as: {2\/2(0fMl)} = (5м)2 |NME|2 G, where {дм) is the effective Majoron-neutrino coupling constant, NME is the nuclear matrix element and G is the kinematical factor. Using our result 21 : T1/2(0i/Ml) > 3.7(5.9)-10 у and values of G and NME calculated in the QRPA model with proton-neutron pairing [23] we obtain <7м < 12(9.5) • 10~5 (дм < 6.5(5.4) • 10~5 On the basis of calculation [17]) with 90% (68%) C.L., which is one of the best restriction up-to-date obtained in the direct 2/3 decay experiments [1].

CONCLUSION

The new set-up with four 116CdWO4 crystals (339 g) is running since October 1998 in the Solotvina Underground Laboratory. Improved passive shield, new active shield made of fifteen CdWO4 crystals (total mass 20.6 kg), as well as time-amplitude and pulse-shape analysis of the data result in the reduction of the background rate in the 2.5-3.2 MeV region to 0.03 counts/y • kg • keV: For 4629 h of the exposition the half-life for 2v2/3 decay1 of 116Cd 19 is measured as T1/2(2u) = 2.6±0.1(stat.)+o:I(syst.)- 10 y. The T1/2 limits for neutrinoless 116 23 20 decay of Cd are set at T1/2 > 0.7.(2.5) • 10 у at 90% (68%) C.L. for transition to ground state of 116Sn, while for decays to the first 2f and second 0+ excited levels of 116Sn at 22 22 T1/2 > 1.3(4.8) • 10 у and > 0.7(2.4) • 10 у with 90 % (68 %) C.L., respectively. For 0^2/3 decay with emission of one or two Majorons, the limits are Ti^iOvMl) > 3.7(5.8) • 1021 у 20 and T1/2(0uM2) > 5.9(9.4) - 10 у at 90% (68%) C.L. Restrictions on the value of the neutrino mass, right-handed admixtures in the weak interaction, and the neutrino-Majoron coupling constant are derived as: т„ < 2.6(1.4) eV, 77 < 3.9 • 10~8, Л < 3.4 • 10~6, and дм < 12(9.5) • 10~5 at 90 % (68 %) C.L., respectively [24]. 16 Bizzeti P. G. et al.

116 In August 1999 one of our CdWO4 crystals was annealed at high temperature, and 116 its light output has increased by w 13%. The PMT of the main CdWO4 detectors was changed by a special low background EMI tube with the RbCs photocathode, whose spectral response better fits the CdWC>4 scintillation light. As a result, the spectrometric parameters of the detector were improved. In particular, the energy resolution of the main detector is now 11.4 % at 1064 keV and 8.6 % at 2615 keV (comparing with those before this upgrading: 14.5 % and 11 %). Besides, the PS discrimination ability of the detector was improved, too. It is expected that after approximately 5 years of measurements the half-life limit 23 Ti/2(0i/2/?) >4-10 у will be reached which corresponds to mu < 1.2 eV. The bounds on neutrinoless 20 decay with Majorons emission and 20 transitions to the excited levels of 116Sn would be improved, too. The present research was supported in part by the Ukraine Fundamental Researches Foun- dation (Grant F5/1930-98).

REFERENCES

1. Мое М., Vogel P. II Ann. Rev. Nucl. Part. Sci. 1994. V.44. P. 247; Tretyak V. I., Zdesenko Yu. G. II At. Data Nucl. Data Tables. 1995. V.61. P.43. 2. Klapdor-Kleingrothaus H. V. II Int. J. Mod. Phys. A. 1998. V. 13. P. 3953; Suhonen J., CivitareseO. II Phys. Rep. 1998. V. 300. P. 123; FaesslerA., Simkovic F.J. //Phys. G: Nucl. Part. Phys. 1998. V.24. P.2139. 3. Baudis L. et al. II Phys. Rev. Lett. 1999. V.83. P. 41. 4. Aalseth С E. et al. II Phys. Rev. С 1999. V. 59. P. 2108. 5. Luescher R. et al. II Phys. Lett. B. 1998. V.434. P. 407. 6. Danevich F. A. et al. II Phys. Lett. B. 1995. V. 344. P. 72; Georgadze A. Sh. et al. II Phys. At. Nucl. 1995. V.58. P. 1093. 7. Danevich F.A. et al. II Nucl. Phys. B. (Proc. Suppl.). 1999. V.70. P. 246. 8. Zdesenko Yu. G. et al. II Proc. of the 2 Intern. Symp. Underground Phys., Baksan Valley, 1987. M., 1988. P. 291. 9. Audi G., Wapstra A. H. II Nucl. Phys. A. 1995. V. 595. P. 409. 10. Danevich F. A. et al. II Nucl. Phys. A. 1998. V. 643. P. 317. 11. Gatti E., Martini F. II De Nuclear Electronics 2, IAEA. Vienna, 1962. P. 265. 12. Fazzini T. et al. II Nucl. Instr. Meth. A. 1998. V.410. P. 213. 13. GEANT. CERN Program Library Long Write-up W5013. CERN, 1994. 14. Ponkratenko O.A. et al. II Proc. of the Intern. Conf. йп Non-Accelerator New Physics NANP'99, Dubna, June 28-July 3, 1999. To be published in «Phys. At. Nucl.». 15. Firestone R. B. Table of Isotopes / Ed. by V. S. Shirley. 8th ed. N.Y., 1996. 16. Ejiri H. et al. II J. Phys. Soc. Japan. 1995. V. 64. P. 339. 17. Arnold R. et al. II Z. Phys. С 1996. V. 72. P. 239. 18. Particle Data Group. Review of Particle Physics // Eur. Phys. J. С 1998. V. 3. P. 1. 116 116 Status of Cd Double /3 Decay Study with CdWO4 Scintillators 17

19. Feldman G.J., Cousins R. D. II Phys. Rev. D. 1998. V.57. P. 3873. 20. Staudt A. et al. If Europhys. Lett. 1990. V. 13. P. 31. 21. Hirsch M. et al. II Phys. Rev. D. 1996. V.53. P. 1329. 22. Faessler A. et al. II Phys. Rev. D. 1998. V. 58. P. 115004. 23. Hirsch M. et al. II Phys. Lett. B. 1996. V.372. P. 8. 24. Danevich F. A. et al. II Phys. Rev. С 2000. V. 62. P. 045501. XJ0200052

Письма в ЭЧАЯ. 2001. №6[109] Particles and Nuclei, Letters. 2001. No.6[109]

УДК 539.165 NEW LIMITS ON 2(3 PROCESSES IN40 Ca AND 46Ga

BY USING LOW RADIOACTIVE CaF2(Eu) CRYSTAL SCINTILLATORS P. Belli0-, R. Bernabeia, С J. Dai1, F, Grianti0, H. L. Heb, G. Ignestid, A. Incicchittid, H. H. Kuangb, J. M. Mab, F. Montecchiaa, 0. A. Ponkratenkoe, D. Prosperid, V. I. Tretyak6, Yu. G. Zdesenkoe aDip. di Fisica, Universita' di Roma «Tor Vergata» and INFN, sez. Roma2,1-00133 Rome, Italy bIHEP, Chinese Academy, P.O. Box 918/3, Beijing 100039, China cIstituto di Fisica, Universita di Urbino and INFN, sez. Bologna, Italy dDip. di Fisica, Universita' di Roma «La Sapienza» and INFN, sez. Roma, 1-00185 Rome, Italy eInstitute for Nuclear Research, MSP 03680 Kiev, Ukraine

The development of highly radiopure CaF2(Eu) crystal scintillators has been performed aiming at a substantial sensitivity enhancement of the 2/3 decay investigation and of the search for dark matter particles with spin-dependent (SD) interaction. The results of CaF2(Eu) background measurements and simulation are presented. New and highly improved Ti/2 limits on the 2/3 decay of 46Ca and the double electron capture of 40Ca are obtained.

Совершенствование сверхчистых кристаллических сцинтилляторов CaF2(Eu) велось с целью повышения чувствительности регистрации процессов 2/3-распада и поиска частиц темной мате- рии со спинзависимым взаимодействием. Представлены результаты моделирования и измерения фона сцинтилляторов CaF2(Eu), а также новые экспериментальные пределы для 2/3-распада 46Са и двойного электронного захвата 40Са.

INTRODUCTION

The investigations of the neutrinoless (0i/) double /3 decay (which violates the lepton number conservation) is a powerfull tool to search for information leading to possible new physics beyond the standard model [1,2]. Moreover, at present, the Ог/2/З decay is also considered a powerful test for different extensions of the standard model (including several SUSY models), which could offer not only complementary but — in some cases — competitive and superior results than other running or forthcoming accelerator and nonaccelerator experiments [2-4]. The most sensitive results in the search for the 2/3 decay have been obtained by the so-called «active source» technique; in this case, the detector, which contains the 2/?-decay candidate nuclei, serves itself as source [1]. This approach provides much higher detection efficiency and density of the candidate nuclei than the «passive source» technique, where a thin source is separated from the detector. For instance, we can compare the 100 % efficiency quoted for Ge semiconductor detectors in Ref. 5 with the 4% efficiency of the NEMO2 tracking detector searching for the 2i>2/? decay of 100Mo [9] and of 116Cd [10]. As examples of the results achieved with the «active source» technique, we recall the 25 76 highest half-life limit T1/2 > 1.6 • 10 у (т„ < 0.5 eV) established for Ge (enriched HP New Limits on 2/3 Processes 19

22 23 Ge detectors) [5] and the T1/2 limits in the range 3-10 -4-10 у (т„ < 2-4 eV) obtained for 136Xe (high pressure Xe TPC) [6], for .130Te (low-temperature bolometers) [7] and for 116 116 Cd (enriched CdWO4 scintillators) [8]. Simple considerations can easily demonstrate that progress in the 2/3 decay research is directly connected with the development of. «active» detectors. Suitable detectors for 2/3 decay. studies are the well-known calcium fluoride scintillators [11,12]. In fact, they have a relatively large light output (~ 60 % of Nal(Tl)) and offer excellent operational characteristics (nonhygroscopicity, inertness, insolubility, good mechanical properties, etc.) when used as /3 particles or low-energy radiation detectors [12-14]. In particular, they can be usefully utilized to search for the 2/3 decays of 46Ca (natural abun- 48 dance: 6 = 0.004%; Quo = 990.4 keV) and of Ca (5 = 0.187%; Q0f} = 4272 keV) 40 besides the double electron capture of Ca (6 = 96.941%; Q2EC - Ш.8 keV). The first experiment of this kind regarded the 2/3 decay of 48Ca [15];let us also mention the latest ones on the 2/3 decay of 48Ca [16], of 46Ca [17] and on the double electron capture of 40Ca [18]. Moreover, the CaF2(Eu) detectors have been considered as promising detectors for the direct search of Dark Matter particles with SD coupling [18,19]. All the considerations mentioned above motivated the R&D program for the development of radiopure calcium fluoride detectors carried out by the DAMA collaboration with the Bicron company [17,18']; the aim was to enhance the sensitivity of searches for 2/3 decay and dark matter particles with SD interaction. We present here the new step of development performed in the framework of this R&D program and devoted mainly to the study of the CaF2(Eu) background and to the understanding of its origins. New results on 2/3 decay processes are presented. More details on the experiment can be found in [20].

: ; ' ' • • • ' 1. MEASUREMENTS '

Two CaF2(Eu) crystals grown by the Bicron company were used in the present work. Each crystal has 3" diameter by 1" length (370 g mass). These crystals are labelled in the following as Bicroh-1 and Bicron-2. ... . The background measurements have been performed in the. Gran Sasso underground laboratory inside the R&D DAMA set-up described in Ref. 18 for the Bicron-1 test. Here we only recall the main features of the apparatus for the measurements performed with Bicron-2. The crystal was coupled to a low backgound PMT EMI9265B53/FL through a 10 cm long TETRASIL-B light guide; it was wrapped by a teflon diffuser (~ 100 /лп thick) as also the light guide. The detector and the PMT were inside about 10 cm of low radioactive copper shaped in order to contain them tight. We note that in this measurement no Nal(Tl) anti- Compton detector was used as instead done in the Bicrori-1 data taking'[18]. The detector and the ~ 10 cm Cu envelop were closed inside a sealed low-radioactive Cu box, in one turn surrounded by a shield made of 10 cm of low-radioactive copper and 15 cm of low radioactive lead followed by 1.5 mm Cd foil and about 4/10 cm of polyethylene/paraffin. To avoid any contact with environmental air (which can contain Radon in trace), the Cu box was continuously flushed with high purity (HP) nitrogen gas (long stored deep underground) and maintained at about 1 mbar overpressure. Finally, the whole shield was sealed in a plexiglass box also maintained in HP nitrogen atmosphere. : . 20 Belli P. et al.

counts/8 keV counts/8 keV

ю -

500 1000 1500 2000 2500 3000 3500 4000 400 600 800 1000 1200 1400 £,keV £,keV Fig. 1. Measured background spectrum of the Bicron-2 crystal for the energy region 300-4000 keV. The three peaks at about 0.81, 1.01 and 1.15 MeV can be ascribed to a particles from 232Th and 238U chains internal impurities of the crystal, the fitting of the a peaks is shown (b)

The rejection of the residual noise just Bicron-2 above the software energy threshold (4 keV here)1 profits by the different timing structure between the noise (PMT fast signals with decay times of the order of tens ns) and the CaF2(Eu) «scintillation» (signals with decay times of the 10 12 order of 940 ns) pulses, whose shapes were E, keV recorded over 3125 ns by a Lecroy Transient digitizer. The residual noise can be rejected Fig. 2. Measured background spectra of the from the production data by applying software Bicron-2 and Bicron-1 crystals for the energy cuts; the corresponding software cut efficiency region 4-20 keV for each energy bin can be evaluated by applying the same cuts to the source data (collected in the same experimental conditions and energy range). The background spectrum of the Bicron-2 crystal was measured in this apparatus during 1906.3 h in the three energy regions: 4-20, 40-230 and 200-4000 keV. Data previously collected with Bicron-1 were also available for analysis; the measuring time was 631.4 h for the energy interval 4-24 keV and 260.9 h for the energy region 200-3300 keV [18]. During the measurement, the energy scale and resolution of the detector were calibrated 109 137 22 periodically with different sources: Cd (£7 = 22 keV); Cs (Ey = 32; 662 keV); Na

'We recall that in any kind of experiment requiring an energy threshold at keV level, software cuts or — sometimes — hardware procedures near threshold have to be applied, although generally the corresponding procedures are not explicitly quantified by the authors. New Limits on 2/3 Processes 21

60 (£7 = 511; 1275 keV) and Co (£7 = 1173; 1332 keV). In addition, background peaks (.E-y = 128 and 169 keV) from 152Eu contamination in the Bicron-2 crystal were used for calibration in the energy region 40-230 keV. • The measured background spectrum of Bicron-2 crystal in the energy region 300-4000 keV is shown in Fig. 1, while the low energy part (4-20 keV) and the middle energy (40^100 keV) are given in Figs. 2 and 3, respectively. Three peaks at the energies of about 0.81, 1.01 and 1.15 MeV are present; considering the typical a/ft ratio for the light output of the 232 238 CaF2(Eu) crystals, these peaks can be attributed to a particles due to Th and U internal impurities. Moreover, in the lower energy region, three clear peaks at energies ~ 47, ~ 129 and ~ 169 keV are also present; as was shown by simulation (see below), such peaks could 152 be explained by Eu (Ti/2 = 13 y) activity created in the CaF2(Eu) crystal by the neutron activation at Earth surface. counts/4 keV counts/0.5 keV 2500

2000 -

1500 -

1000 -

500

0 50 100 150 200 250 E,'keV,

Fig. 3. Experimental spectrum of the Bicron-2 crystal up to the energy of 400 keV. The three peaks at 152 •~ 47, ~ 129 and ~. 169.keV can be ascribed to Eu (T1/2 = 13 y) activity created in the CaF2(Eu) crystal by the neutron activation at Earth surface. The model distribution from 152Ец contamination is also shown (see text for details)

2. BACKGROUND SIMULATION AND ESTIMATE OF THE 2(3 DECAY HALF-LIFE LIMITS The locations and amounts of the radioactive contaminations have been estimated by simulating the background spectra of the CaF2(Eu) crystal by the GEANT3.21 package [22] and the event generator DECAY4 [23], which describes the initial kinematics of the events. The considered background model includes external (from the 40K, 232Th and 238U contaminations in the PMT) and internal (from 40K, 152Eu, 232Th and 238U intrinsic impurities 22 Belli P. et al. in the CaF2(Eu) crystal) contributions; it was supposed that equilibrium in U and Th chains may be broken. The fit of the experimental spectrum in the energy interval 40-2700 keV following this model gives the activities of the possible impurities in PMT and in crystal; these are shown in Table 1. As evident, the equilibrium in U and Th chains for the CaF2(Eu) contaminations is broken, as often in detectors.

Table i. Radioactive impurities of the PMT and CaF2(Eu) crystal (mBq/kg)

232 238 152 40 Source of Th-chain U-chain Eu K background 228 226 232Th Ra 238 у 230Th Ra 2iopb PMT < 100 400(200) 4(2) • 104

CaF2(Eu) 1.2(5) 11(5) 0.5(2) 4.5(8) 2.4(10) 0.5(2) 7.5(20) 5(2) Bicron-2

The fit in the energy interval 300-1500 keV is shown together with the experimental data in the insert of Fig. 1, where a reasonable agreement is achieved. Also the energy dependence of the a/P ratio for the CaF2(Eu) crystal was obtained: a/13 = 0.130 + 0.017 Ea, Ea being the energy of a particle in MeV. The experimental spectrum up to the energy of 400 keV is given in Fig. 3. There is also shown the model distribution from 152Eu contamination, which consists of four peaks at the energies: a) ~ 7 keV (L-shell binding energy of Sm); b) ~ 47 keV (.ftT-shell binding energy of Sm); с) ~ 128.8 keV (sum of L-shell binding energy with the most intensive ~ 121.8 keV 7 ray or conversion electron from 152Eu decay); d) ~ 168.8 keV (sum of fiT-shell binding energy with the ~ 121.8 keV 7 ray). The absence in the data of the ~ 7 keV peak (see also Fig. 2) could be explained by the known cross sections uncertainties (±10%) in the low energy region 40-120 keV, which were used for simulation. The estimated activity of the 152Eu in the Bicron-2 crystal — considering the higher energy peaks — is 7.5(20) mBq • kg":. Such activity could be reached after approximately five months of activation of the CaF2(Eu) crystal (in our case the content of Eu is ~ 0.5 % [24]) at the Earth surface, where the thermal neutron flux is about 5 • 10~~3 neutrons/cm2/s. The background model described above was used to estimate half-life limits for the 40 46 double EC capture of Ca (Q2EC = 193.8 keV) and the neutrinoless double 0 decay of Ca (Qpp = 990.4 keV). The number of candidate nuclei in each CaF2(Eu) crystal is equal to 2.766 • 1024 and 1.141 • 1020 for 40Ca and 46Ca, respectively. The double EC capture of 40Ca has to be followed by the emission of two X rays with 3.2 keV energy each and it is expected that an inner bremsstrahlung photon will carry off the energy released in the neutrinoless process. Efficiencies for the different double /3 decay processes were calculated by using the GEANT3.21 package [22] and the event generator DECAY4 [23]. The simulated response functions of CaF2(Eu) crystals and the calculated values of efficiencies for the different 2/3 decay processes were used together with the corresponding background rates, the measuring times and the numbers of 40Ca and 46Ca nuclei in order to estimate half-life limits for the effects under investigation. These limits are summarized in Table 2 and were obtained in two different ways. Firstly, they have been obtained by using New Limits on 20 Processes 23

Table 2. Half-life limits for different 2/3 processes of 40Ca and 46Ca

Tj/2 limit, у T1/2 limit, у T1/2 limit, у . 2/3 process (one a approach): : '. (x2 fit procedure) (X2 fit procedure) (Peak energy) •-•68%-C.L. с 68% (90%) C.L. 68 % (90 %) C.L. Bicron-2 Bicron-1 Bicron-2 Bicron-1 (Combined from Bicron-1+Bicron-2) 40 21 2 21 0г^2ЕС of Ca 4.4 • 10 2.7 : 10 ' 4,1(2.5) • 10 2,4(1.4) • 1021 4.9(3.0) • 1021 •. (6.4 keV) 0^2EC of 40Ca 5.5 • 1020 3.4(1.8) • 1020 3.4(1.8) • 1020 (193.8 keV) 40 21 2i/2EC of Ca 8.8 • 1021 5.3 • 1021 8.2(5.0) • Ю21 4.8(2.8) • 1021 9.8(5.9) • 10 (6.4 keV) 46 0i/2/Jof Ca 1.8 -10" 9.1 • 1016 1.5(0.9) • 1017 8Д4.9) 1016 1.7(1.0) - 1017 (990 keV)

the so-called «one a approach», in which the counts/0.5 keV excluded number of signal events is determined simply as the square root of the number of background counts in a given energy window AE. Notwithstanding its simplicity, this method gives the right scale of the

experiment's sensitivity. Moreover, the IY/2 limits were improved by using the standard procedure of the least square fit of the experimental data by the sum of the background

model and of the: simulated response function 60 - of the detector for the 2/3 decay process under investigation. From such a fit, the number of 40 - events in the expected signal peak (and the 20 - uncertainty) were determined, usually giving no statistical evidence for the effect. Then using these values, the excluded — at 68 % (90 %) C.L. — numbers of events for the different 2/3 decay processes were calculated in accordance with the Particle Data Group pro- Fig. 4. The 4-20 keV experimental energy spec- cedure [25]. trum of the Bicron-2 crystal and the fitting curve, where the excluded — at 90 % C.L. — peak from As an example, part of the experimental 40 21 0i/2EC of Ca with half-life 2.o • 10 у is also spectrum of the crystal Bicron-2 in the energy shown interval 4-20 keV and the fitting curve are shown in Fig. 4, where the excluded — at 90 % C.L. peak from the 0г/2ЕС decay of 40Ca 21 with half-life 2.5 • 10 у is also shown. • • • .;• ,.; In the last column of Table 2 the half-life limits obtained by combining the results for both crystals: Bicron-2 and Bicron-1 [18] are shown. In this case the procedure was as following. The numbers of events in the expected peak — determined for each crystal by X2 fit as described above — were simply added: S = S\ +.S2, while their error bars were 24 Belli P. et al.

2 2 1/2 added quadratically: a(S) = [cr {Si) + cr (52)] . The values of S and a{S) were used to calculate the combined excluded numbers of events for different 2/3 decay processes. The T1/2 limit obtained for Ог/2/З decay of 46Ca is higher than the best result previously available [17]. Restriction for the two neutrino 2EC of 40Ca is also improved from value 4.6-1021 у [18] to ~ 1022 y. It should be stressed that the last one is the highest half-life limit obtained up-to-date for the 2/3+ decay processes and, in particular, for the double electron capture (see reviews [1] and the last work [26] for comparison). The same is true for the 40 neutrinoless 2EC of Ca, which T\/2 limit is established for the first time.

CONCLUSIONS

46 40 The new and highly improved Tx/2 limits on the 2/3 decay processes of Ca and Ca achieved here have surely demonstrated the feasibility and perspectives of developed CaF2(Eu) scintillators for high sensitive investigations of the 2/3 decay. The use of these scintillators in the search for dark matter particles with SD coupling is also another interesting possibility. Further efforts are in progress.

REFERENCES

1. Мое M., Vogel P. II Ann. Rev. Nucl. Part. Sci. 1994. V. 44. P. 247; Tretyak V.I., Zdesenko Yu. G. II At. Data Nucl. Data Tables. 1995. V.61. P. 43; Suhonen J., Civitarese O. II Phys. Rep. 1998. V.300. P. 123. 2. Klapdor-Kleingrothaus H. V. et al. II J. Phys. G: Nucl. Part. Phys. 1998. V.24. P. 483. 3. Bednyakov V.A. et al. II Mod. Phys. Lett. A. 1997. V. 12. P. 233. 4. Hirsch M., Klapdor-Kleingrothaus H. V. II Prog. Part. Nucl. Phys. 1998. V.40. P. 323. 5. Baudis L. et al. II Phys. Rev. Lett. 1999. V. 83. P. 41. 6. Luescher R. et al. II Phys. Lett. B. 1998. V.434. P. 407. 7. Alessandrello A. et al. II Phys. Lett. B. 1998. V.433. P. 156. 8. Danevich et al. F. A. II Phys. Rev. С 2000. V. 63. P.045501. 9. Dassie D. et al. II Phys. Rev. D. 1995. V.51. P. 2090. 10. Arnold R. et al. II Z. Phys. С 1996. V.72. P. 239. 11. Knoll G.F. Radiation Detection and Measurement. 2nd ed. N.Y., 1989. 12. Colmenares С et al. II Nucl. Instr. Meth. 1974. V. 114. P. 277. 13. Campbell M. et al. II Nucl. Instr. Meth. 1976. V. 137. P. 235. 14. PouliotJ. etal. //Nucl. Instr. Meth. 1988. V.270. P.69. 15. Der Mateosian K, Holdhaber M. II Phys. Rev. 1966. V. 146. P. 810. 16. You K. et al. II Phys. Lett. B. 1991. V. 265. P. 53. 17. Bernabei R. et al. II Nuovo. Cim. A. 1998. V. 111. P.347. 18. Bernabei R. et al. II Astropart. Phys. 1997. V.7. P. 73. 19. Bacci С et al. II Astropart. Phys. 1994. V.2. P. 117. New Limits on 2/3 Processes 25

20. Belli P. et al. II Nucl. Phys. B. 1999. V.563. P. 97. 21. Ellis J., Flores R. A. II Phys. Lett. B. 1991. V. 263. P. 259. 22. Brun R. et al. CERN Program Library Long Write-Up W5013. CERN, 1994. 23. Ponkratenko O.A. et al. II Phys. Atom. Nucl. 2000. V.63. P. 1282. 24. Krus D. Bicron Co. Private communication. : 25. Particle Data Group: Review of Particle Physics // Phys. Rev. D. 1996. V.54. P. 1. 26. Belli P. et al. II Astropart. Phys. 1999. V. 10. P. 115. XJ0200053

Письма в ЭЧАЯ. 2001. №6[109] Particles and Nuclei, Letters. 2001. No.6[109)

УДК 539.165

THE SINGLE STATE DOMINANCE IN 2i//?/?-DECAY TRANSITIONS TO EXCITED 0+ AND 2+ FINAL STATES S. V. Semenova, F. Simkovicb, P. Dominb "Russian Research Centre «Kurchatov Institute», Moscow, Russia bDepartment of Nuclear Physics, Comenius University, Bratislava, Slovakia

A single state dominance theoretical analysis of the two-neutrino double beta decay (2u/3[3 decay) transitions have been carried out for nuclear systems with A = 110,114,116 and 128. The energy denominators of the perturbation theory have been considered exactly. New results for the 2i//3f3-uecay transitions to the 0+ and 2+ excited states of the final nucleus are presented. A possibility of detecting these modes experimentally is addressed. Проведен теоретический анализ двухнейтринного двойного бета-распада (21//?/?-распад) в пред- положении доминантности основного состояния промежуточного ядра. Исследовались ядерные си- стемы с А = 110,114,116 и 128. Энергетические знаменатели теории возмущений учитывались точно. Получены новые результаты для 21/ДО-переходов в возбужденные 0+- и 2+-состояния ко- нечных ядер. Обсуждается возможность экспериментального детектирования этих процессов.

The two-neutrino double beta decay (2vf3(3 decay) remains at the forefront of nuclear physics [1-3]. The established 2i//3/?-decay half-lifes for ground state to ground state transition for a couple of isotopes constrain nuclear theory and stimulate its further development. Additional experimental information about the 2i/f30 decay and related processes is of great interest [3-6]. The attention of experimentalists is paid to the transitions to the 0+ and 2+ states of the final nucleus. The detection of these transitions has the advantage of additional experimental signature: By deexcitation of these nuclear states one or two gamma- quanta with strictly fixed energies are emitted. It is worth to notice that there is a first positive evidence for 2i/fi0 decay of 100Mo for transition to the Of excited state of 100Ru with Tf/a = 6.111:1 • Ю20 у [4-7]. With present low-background detectors there is a chance to detect the 2i/(3f3 decay to the excited 0+ and 2+ states of the final nucleus at the level of 1021-1022 years. In view of this fact the theoretical calculations of 2i//3/3-decay half-lifes are highly required [8]. The aim of this contribution is to present theoretical predictions for some 2i//3/3-decay transitions to excited 0+ and 2+ states. The 2i//3/3 decay process may occur in second order in standard theory of weak interaction. Thus the problem of the calculation of the 2i//3/?-decay matrix elements consists in construction of the full set of virtual intermediate nuclear states of the double-odd nucleus. This is a complex task, which continues to be challenging for the specialists of different nuclear models. There are few 2i//?/3-decay nuclear systems where the spin-parity of the ground state of the intermediate nucleus is 1+. Some times ago it was suggested that the 2i//?/3-decay nuclear matrix elements governing these processes could be dominated by single transition through this 1+ intermediate state (the so-called Single State Dominance Hypothesis (SSD)) [9,10]. The Single State Dominance 27

This idea is supported by recent SSD calculations [10-12]. However; due to inaccurate experimental determination of both 2i^/3/?-decay half-lifes and log/tEc values for the electron capture it is not possible to decide whether SSD is realized or not. Recently, it was shown that by measuring differential decay rates the SSD hypothesis can be confirmed or ruled out [13] already by ongoing NEMO III experiment [14], By assuming the SSD for the 2z//?/?-decay half-life we can write

where the dimensionless integral over lepton phase space is

T+l rT+2-ei pT+2-ei-e2 /

*F{Zhei)F{Zhe2)p^p2e2b>\u\D{K,L). (2) Here, Wi, pi arid Ei (i = 1,2) are energies of antineutrinos, momenta and energies of electrons in units of the mass of electron me, respectively, шг = T + 2 — e± — e2 — OJ\ and the kinetic energy of leptons in final state is T = (Ei — Ef —2me)/me = Qpp. Ei (Ef) is the energy of the initial (final) nuclear state. F(Zf,e) and дл denote the relativistic Coulomb factor [1,2] and the vector axial coupling constant, respectively. The form of the factor D(K,L) depends on the spin and parity of the final nuclear state. We have

D(K, L) = K2 + L2 + KL if Л = 0+, (3) D(K,L) = (K-L)2 if JJ = 2+.

The К and L factors are built of the energy denominators of perturbation theory

R= . •..- . 1 : '. • .,••, 1

L = • ——; :—•' H —

-e2- with fix = Ei - (Ei + Ef)/2. E\ is the energy of the ground state of the intermediate nucleus with Jn = 1+. The beta transition amplitudes

M\ - (i+s. ii г л on/ щ = {jj у г у i+s;> (5) can be calculated in the framework of various nuclear models [1-3] or deduced from log/t values of electron capture and single beta decay processes as follows:

(6) 28 Semenov S. V., Simkovic F., Domin P.

Table 1. The calculated г^ДО-decay haif-lifes within SSD hypothesis with exact (T^~ь*—exc)exc)) \ and ( аррг> approximated (Г1 /Г ) К and L factors

rp(2u-exc) Nucleus Transition Mi log/tp- T\/2

+ 20 20 o -* o+s. 3.913 3.703 4.66 1.40 • 10 1.31 • 10 f)"b . О > 25 25 nopd 2.627 3.060 5.528 5.56 • 10 4.64 • 10

26 26 log /4EC = 4.08 П~Н . 0 1.031 2.262 6.8 2.57 • 10 2.54 • 10 31 0+ - 2+ 1.025 2.259 7.39 4.28 • 1031 4.08 • 10 m + + 25 Cd 1.051 3.366 4.473 1.26 • 1025 1.25 • 10 o -og s. log /4ЕС = 4.9

+ 19 19 o ->• o+s. 5.489 3.664 4.662 1.32 • 10 1.14- 10 116 24 Cd 0+^2+ 2.958 2.399 5.85 1.09 • 1025 7.30 • 10

23 23 log /<ЕС = 4.39 0+-.0+ 2.051 1.945 5.88 2.01 • 10 1.88 • 10 28 0+ - 2+ 1.356 1.598 6.31 2.92 • 1028 2.45 • 10 29 0+ - 2+ 1.134 1.487 6.4 1.86 • 1029 1.62 • 10

+ 25 25 o -»o+s. 1.697 3.298 6.08 1.25 • 10 1.23 • 10 32 log/tEC = 5.05' 0+-2+ 0.830 2.865 6:38 1.00 • 1033 9.84 • 10 where D = (27r3ln2)/(G|m^) = 6146.7 s. In this case the exact form of the 2i//?/?-decay half-life within the SSD is

( 2 xc) 8тг T1 , p (o+->j;) =

= 2.978 • 1014 У- (7) Hexc(T,JJ)

Here, Ac = h/mec is the Compton wave length of the electron. We stress that the half-life T^ in Eq. (7) depends only on two unknown experimental quantities, which are log /£EC and log ftp-. We note that the half-life in Eq. (7) does not depend explicitly on Gp and дл- There is a full cancellation of these factors appearing in Eq.(l) and in expressions for M\, М[ in Eq. (6) through £>-factor. In the previous SSD calculations [10,11] the Jf and L factors in Eq. (4) were replaced with their approximate values by assuming the energies of the outgoing leptons to be equal, i.e.,

D(K,L) _ 12 if 3} (8) 4(ei -e f + D(K,L) 2 - if JJ* it 6

In this way the dependence of D(K, L) on lepton energies is ignored. This approximation is commonly used in the 2i//3/?-decay calculations including contributions from a large number of intermediate nuclear states as it allows one to factorize the nuclear part and the integration The Single State Dominance 29

over the phase space of outgoing leptons [1-3]. However, within the SSD there is no need for doing it. It was found that this approximation leads to a significant overestimation of 2i/00- deeay half-life in the case of 100M6 [13]. There is an interest to examine this approximation also for other nuclear systems. By assuming Eq. (8) the integral over phase space in Eq. (2) can be written as

О pT+\ рТ+2-ei а г Я ее (Г,0+) = A/ dej de2F{Zhe{)x

T+2-ei , ......

ds2F(Zf,e1)x (9)

We note that factors #appr(T,0+) and ЯаРРг(Т,2+) are related with the common G(2l/>(0+) (2u and G )(2+) factors [1,2] as follows: ... • ;

23 а г 1 +) = 2.414- 10- ^Я РР (Г,0+) у" , .:.•;• :. '" /л f\\

^fr(T, 2+) = 7.243 • 10-23^ЯаРРг(Т, 2+) у"»; (2) = ^Mf

We have used G/з = 1.149 • 10~5 GeV~2 and QA = 1.25. Within the above approximation the О^ДО-decay half-life takes the form

The calculated 2i//?/?-decay half-lifes of 110Pd, 114Gd, 116Gd, and 128Te for transitions to some of the lowest 0+ and 2+ states of the final nucleus are listed in the Table. By glancing the results in the Table we see that the 2f/?/?-decay half-lifes for transitions to the excited states are significantly above the level of 1023 years. The only exception is the transition 0+ -• Of in the case of 2v00 decay of 116Cd. The evaluated half-life Tfy « 2.0 • 1023 favor this rare transition to be observed in the near future 2i>/?/3-decay experiments. We notice that for A ='• 116 system the transition to 2+ state is strongly suppressed in comparison with the transitions to Of state even in spite of the fact it is favored by larger Qpp value. From the Table it follows that the 110Pd, u4Cd, 116Cd, and 128Te isotopes are not good candidates for near future experimental study of 2i//3/3-decay transition to the 2+ final state. We remark that the 0+ —+ 2+ transition has been not observed yet. We note that the SSD half-lifes in the Table for ground state to ground state transition differ slightly from those presented in Ref. 12. It is because in the present calculations we used Gp = 1.149 • lO1"5 GeV2 [15] instead of GFCOSBC in order to account for radiative corrections [16], which are included in expressions for log ft values. In addition, in Ref. 12 we employed D(K,L) = 3(K + L)2/4 [I] instead of Eq. (8) what introduced inaccuracy of the order of a small factor (K - L)2. 30 Semenov S. V., Simkovic F., Domin P.

In our previous SSD calculation of 2vP/3 decay of 100Mo we have found that the exact consideration of the energy denominators plays an, important role [13], The numerical study of this effect has shown that for the 0+ —>• Of transition the corresponding half-life time is corrected only by 20 percent. However, for 0+ —* 2+ transition this effect has been found to be large (factor 2-3). For the nuclear systems discussed in the Table this effect is smaller. It is because the difference between the energies of 1+ ground state of the intermediate nucleus (.Ei) and the 0+ ground state of the parent nucleus (Ei) is larger for these nuclear systems in comparison with A = 100 one. Nevertheless, one can see that the exact consideration of denominators entering the expressions for К and L in Eq. (4) leads always to a smaller value of 2i//3/?-decay half-life. The relevant difference of ground state energies for studied nuclear systems are as follows:

110 114 El-Ei = 0.368 MeV ( Pd), 0.933 MeV ( Cd), Ei-Ek = -0.042 MeV (116Cd), 0.748 MeV (128Te), (12) Ег-Ei = -0.342 MeV (100Mo).

It is obvious that more experimental information about 2i//3/?-decay transitions to excited states is needed. It is especially important in connection with preparation of neutrinoless double beta (0i//3/? decay) experiments measuring the transition to excited states, which perhaps will be able to improve the existing limits on different lepton number violating effective parameters (effective Majorana neutrino mass, parameters of right-handed currents, Я-parity violating supersymmetry, etc.). In preparation of both 2г/ДО- and О^ДО-decay experiments the theoretical predictions are very useful. Till now there were missing information especially about 2i//3/?-decay transitions to 2+ excited states. In our contribution we presented exact SSD 2i//?/?-decay half-lifes for transitions to both ground and excited final states for A = 110,114,116 and 128 nuclear systems. At present, it is not clear whether the SSD is realized or not. The chance to shed more light on this problem has the ongoing NEMO III experiment [13]. But even, if the SSD being only very approximate, it is expecting to give at least the correct order of the magnitude of 2i//?/3-decay half-lifes of interest. In summary, the 2i//?/?-decay transitions to excited 0+ and 2+ final states have been studied in the framework of the Single State Dominance Hypothesis without any approximation. The calculated half-lifes for 2i//?/?-decay of U0Pd, 114Cd, 116Cd, and 128Te are presented. We have found that perhaps the 2z//?/?-decay 0+ —> Of transition of n6Cd is a good candidate for future experimental study. . The authors are grateful to A. S. Barabash, Yu. V. Gaponov and V. V, Khruschey for useful discussions. We acknowledge a partial support from the Grant Agency of the Czech Republic, Grant No. 202/98/1216.

REFERENCES

1. Doi M., Kotani Т., Takasugi E. II Prog. Theor. Phys. (Suppl.). 1985. V.83. P. 1. 2. Haxton W.C., StephensonG.J. II Prog. Part. NucL Phys. 1984. V. 12. P. 409. 3. Faessler A., Simkovic F. 111. Phys. G. 1998. V.24. P.2139.

4. Barabash A. S. et al. II Phys. Lett. B. 1995. V. 345. P. 408. The Single State Dominance 31

5. Barabash A. S. II Czech. J. Phys. 2000. V. 50. P. 447.

6. Barabash A.S. et al. II Phys. Lett. B. 1995. V. 345. P.408; De Braeckeleer L. et al. II Yad. Fiz. 2000. V. 63. P. 1288.

7. Barabash A.S. II Czech. J. Phys. 1998. V.48. P. 155; Piquemal F. II Yad. Fiz. 2000, V. 63. P. 1296.

8. Aunola M., Suhonen J. II Nucl. Phys. A. 1996. V. 602. P. 133. ..

9. Abad J. et al. II Ann. Fis. A. 1984. V. 80. P. 9.

10. Griffiths A., Vogel P. II Phys. Rev. C. 1992. V.46. P. 181.

11. Civitarese O., Suhonen J. II Phys. Rev. С 1998. V. 58. P. 1535; Nucl. Phys. A. 1999. V.653. P.321.

12. Semenov S. V. et al: II Yad. Fiz. 2000. V. 63. P. 1271.

13. Simkovic F., Domin P., Semenov S. V. nucl-th/0006084. 2000.

14. Barabash A.S. (NEMO III collaboration). Privite communication.

15. Hardy J. С et al. II Nucl. Phys. A. 1990. V.509. P. 429.

16. Gaponov Yu. V. II Proc. of the III Intern. Symp: «Weak and Electromagnetic Interactions in Nuclei», (WEIN92). Singapore, 1993. P. 87. : XJ0200054

Письма в ЭЧАЯ. 2001. №6[109] Particles and Nuclei, Letters. 2001. No.6[109]

УДК 539.126 PRESENT STATUS OF THE MONOLITH PROJECT A. A. Petrukhin Moscow State Engineering Physics Institute (Technical University), Moscow For the MONOLITH Collaboration'

MONOLITH is a proposed massive (34 kt) magnetized tracking calorimeter at the Gran Sasso laboratory in Italy, optimized for the detection of atmospheric muon neutrinos. The main goal is to establish (or reject) the neutrino oscillation hypothesis through an explicit observation of the full first oscillation swing. The Am2 sensitivity range for this measurement comfortably covers the complete Super-Kamiokande allowed region. Other measurements include studies of matter effects, the NC up/down ratio, the vjv ratio, the study of cosmic ray muons in the multi-TeV range, and auxiliary measurements from the CERN to Gran Sasso neutrino beam.. Depending on approval, data taking with the part of the detector could start towards the end of 2004.

МОНОЛИТ — проект массивного (34 кт) трекового калориметра из намагниченной стали для лаборатории Гран-Сассо в Италии — оптимизирован для регистрации атмосферных мюонных нейтрино. Его основная цель — подтвердить (или опровергнуть) гипотезу нейтринных осцилля- ции путем прямого наблюдения первого осцилляционного минимума. Область чувствительности этих измерений по разности квадрата масс полностью перекрывает область, разрешенную экспе- риментом Супер-Камиоканде. Другие планируемые эксперименты включают: изучение эффектов влияния вещества, отношения вверх/вниз для нейтральных токов, отношения потоков антиней- трино/нейтрино, изучение мюонов космических лучей в мульти-ТэВ-ной области энергий, экспе- рименты с пучком нейтрино из ЦЕРНа. В случае одобрения проекта, набор данных может быть начат в конце 2004 года.

INTRODUCTION

The question whether neutrinos are massive, and hence the question of the existence of neutrino oscillations, is currently one of the main unsettled challenges in physics. All experiments measuring the flux of solar neutrinos observe a deficit compared to the prediction of solar models. The ratio of muon to electron events observed in atmospheric neutrino interactions is measured by most experiments to be less than expected from models of cosmic ray propagation through the atmosphere. The measurement of the up/down asymmetry of this ratio by the Super-Kamiokande collaboration is generally considered to be the strongest evidence for neutrino oscillations so far. Furthermore, possible oscillation signals in the v^—Ve and v^-Ve channels were observed by the LSND experiment. Finally, massive neutrinos could play an important role in the missing dark matter problem if at least one mass eigenstate lies in the eV range. All these observations make the study of neutrino oscillations a very worthwhile endeavor.

'For the list of the Collaboration see [2]. Present Status of the MONOLITH Project 33

While the cumulative evidence for neutrino oscillations is very striking, a definitive proof that the observed anomalies are actually due to neutrino oscillations is still missing. The current observations of atmospheric neutrinos are consistent with the hypothesis of maximal Vn oscillations, but do not yet exclude some alternative unconventional explanations [1]. The main physics goal of the MONOLITH experiment [2, 3] is to establish the occurrence of neutrino oscillations in atmospheric neutrinos through the explicit observation of the full first oscillation swing in v^ disappearance, and to investigate and presumably exclude alternative explanations. This also yields a significantly improved measurement of the oscillation parameters with respect to previous measurements. The strong magnetic field, adopted in the detector design to extend the sensitivity to the full parameter range allowed by current experiments, is a novel feature among atmospheric neutrino detectors. The charge and momentum measurement of muons from charged current (CC) events also allows unique systematic studies of the atmospheric neutrino flux, and the search for potential matter effects in neutrino oscillations. . The measurement of the oscillation pattern can be usefully supplemented by measurements in the CERN to Gran Sasso neutrino beam. Provided that the neutrino oscillation hypothesis is confirmed, another goal of the experiment is to further investigate the nature of these oscillations. Depending on the oscillation

parameters, oscillations into active (VT) or sterile (us) neutrinos can be distinguished through their different effects on the up/down ratio of neutral current (NC)-like events, and/or through the presence or absence of matter effects yielding a distortion of the observed oscillation pattern as a function of energy and/or muon charge. A particularly interesting option in the context of the standard 3-neurrino scenario would be the measurement of the sign of Am2

via an MSW resonance in Vp—ve oscillations, along the same line as currently studied for Neutrino Factory beams [4]. Due to its ability of in situ measurement of the energy of every muon in the multi-TeV range, MONOLITH will also be a unique facility for pioneer investigations of cosmic ray muons in the unexplored 100 TeV energy region. The results of these studies should give information which is relevant for the solution of the problem of the knee in the cosmic ray energy spectrum. For this experiment we have designed a detector which has been optimized for the detection of atmospheric neutrinos (with additional substantial sensitivity to the neutrino beam from CERN) and can achieve these physics goals.

1. EXPERIMENTAL SET-UP

To explicitly detect an oscillation pattern in the LIE spectrum of atmospheric muon neutrinos, the energy E and direction в of the incoming neutrino have to be measured in each event. The latter can be estimated, in the simplest experimental approach, from the direction of the muon produced from the u^ charged-current interaction, The neutrino energy E can be obtained by means of energy measurements of the muon and of the hadrons produced in the interaction. In order to make the oscillation pattern detectable, the ratio of the neutrino path-length to its energy LIE have to be measured with a FWHM error smaller than half of the modulation period. The energy and angular resolutions of the detector or, more generally, the experimental approach are constrained by this condition. 34 Petrukhin A. A.

14.5 m

13.5 m -~

30.0 m

Fig. 1. Schematic view of the MONOLITH Detector. The arrangement of the magnetic field is also shown

In addition to these requirements on the LIE resolution, the experimental apparatus should guarantee the identification of the neutrino flight direction. In case the interaction vertex is not identified, this condition turns out to be very stringent and requires the identification of the muon flight direction with high efficiency. Different techniques based on the increase of curvature along the track in the magnetic field, on multiple scattering along the track or on time-of-flight measurements, can be used. The latter technique is more effective and allows almost perfect up/down discrimination of the relevant z^-CC events for timing resolution of the order of 2 ns. A precise time-of-flight along the muon track will be also of utmost importance in the rejection of the cosmic muon background in the sample of partially contained i/^-CC events. A large modular structure has been chosen for the detector (Fig. 1). One module consists of a stack of 125 horizontal 8 cm thick iron planes with a surface area of 15.0 x 14.5 m, interleaved with 2.2 cm gaps housing sensitive elements. The height of the detector is thus 13.1 m including antiseismic mechanical supports. The magnetic field configuration is also shown in Fig. 1; iron plates are magnetized at a magnetic induction of « 1.3 T. The detector consists of two modules. Optionally, the downstream module could be complemented by an end-cap of vertical planes to improve the performance for noncontained muons from the CNGS beam (see section 2.4). The total mass of the detector exceeds 34 kt. The sensitive elements provide two coordinates with a pitch of 2.8 cm, and a time resolution better than 2 ns. The full detector is 30.0 m long, 14.5 m wide and 13.1 m high, which is consistent with construction and operation in the Gran Sasso Hall С It is also possible to configure the detector for installation in Hall B. With the envisaged modular structure, we anticipate that the first module could be ready for operation 4 years after experiment approval. 1.1. Glass Spark Counters. Taking into account the overall dimensions of the apparatus (125 planes 30.0 m long and 14.5 m wide) the total active detector area is ~ 54000 m2. The large active area requires a low cost detector, suitable for a fast mass production. Present Status of the MONOLITH Project 35

For these reasons Glass Spark Counters (GSC) [5, 6] have been chosen as active elements. They are derived from resistive-plate chambers by substituting the bakelite by commercial float glass of high resistivity; for this reason they are sometimes called Glass RPC. Moreover they provide a timing of the order of 1-2 ns, and therefore excellent up-down discrimination of muons. The GSC is a gaseous detector composed of two parallel electrodes made of glass with a volume resistivity of about 1012 ficm. The two electrodes, 2 mm thick, are kept 2 mm apart by means of suitable spacers. The gap between the two glass electrodes defines the gas volume where the particle detection occurs. Under particular gas mixture and electric field configurations (typically ~ 48% Argon + ~ 4% isobutane + ~ 48% R134A and ~ 4 -r- 5 kV • mm"1) the detector is operated in spark mode. Typical signal amplitudes of the order of 100 -е- 200 mV/50 п are observed, corresponding to a charge of about 100 -r- 200 pC (depending on the intensity of the applied electric field). The GSCs will be equipped with X and Y pick-up strip electrodes, both with a pitch of ~ 3 cm (2.8 cm), mounted on the opposite sides of the detector planes. The pulses are induced on the strips and then discriminated by the electronics directly connected to the strip ends. The electronics should also provide a fast discriminated signal for timing purposes. This type of readout system will provide a three-dimensional tracking of particles inside the apparatus with a time resolution of ~ 1 ns and a spatial accuracy of ~ 1 cm on both views. 1.2. The Trigger System. The GSC signals are read out digitally. In addition, the signals of 16 or more channels are summed and sent to a TDC via a discriminator. These discriminator signals are also available for trigger purposes. As every spark provides both an X and Y coordinate, the processing of only one coordinate is sufficient for triggering. This leads to about 9000 trigger channels for the flat cable design. A VME trigger board processes groups of 120 of these trigger channels with a programmable local trigger logic. This results in 76 VME boards for all trigger channels. The logic allows one to trigger on certain track patterns, such as 2 or 3 hit layers out of 5 consecutive layers. This results in a trigger efficiency near 100 % for the relevant atmospheric neutrino events (energy > 1.5 GeV). Through a combination of signals from different trigger boards on an additional logic level, a global multiplicity trigger is also possible. The signals of the veto counters are read out by TDC's and are also available for trigger purposes. The expected trigger rate (mainly originating from cosmic ray muons and random coincidences of radioactive decays) is expected to be much less than 1 Hz. The trigger initiates the readout of the digital and TDC information, which is collected and further processed via VME CPU modules. The complete events are then fed into an online computer, which performs a preliminary event reconstruction. At this «2nd level trigger» stage, events can be filtered and/or distributed to several output streams. 1.3. The DAQ System. The MONOLITH DAQ system will be VME based. DAQ electronics will be located on the top of the detector, along the hall axis in 4 points (spaced .~ 7.5 m), two for each detector module. Since VME crates and their read-out are the major DAQ cost, a big effort has to be done to minimize their number: STAS (control and readout) units serving 32 digital chains will be developed (the current commercial STASs serve 8 digital chains) and 64 chs/unit TDC modules will be used. In this way every DAQ station will provide the readout of 16 digital chains and 32 TDC channels for all the 125 detector layers. The 60 STAS and TDC modules needed will be 36 Petrukhin A. A. stored in 6 crates (3 9U and 3 6U). Every crate will serve 40 layers. In every DAQ point 2 VME crates devoted to trigger and slow control systems are foreseen. Front-end data read-out will be performed using Motorola MVME CPUs. To minimize soft-ware cost, the Linux option, as operating system of VME embedded CPUs, will be investigated. Diskless CPU booting, event building and monitoring will be performed using a Linux PC for each DAQ station.

2. NEUTRINO OSCILLATIONS

2.1. Why Atmospheric Neutrinos? Atmospheric neutrino experiments offer several advantages over currently operational or planned long baseline neutrino beam programs.

• A very large L/E range (from about 1 to 105 km • GeV"1; a typical long baseline beam covers only one or two orders of magnitude). Therefore, a very large range of oscillation parameters can be studied simultaneously.

• Two identical sources for a single detector: a near (downgoing neutrinos) and a far (upgoing neutrinos) one.

• For some of the measurements, e. g., the confirmation of the oscillation pattern, there is currently no alternative to atmospheric neutrino detectors if the atmospheric Am2 is low. The pattern measurement is competitive even at high Am2.

• During the next decade large matter effects with high energy neutrinos can only be observed in atmospheric neutrino experiments, since the current long baseline distances of 250 and 730 km are too short for a significant effect. Matter effects already yield discrimination between the pure 2-flavour vti—vr and v^ —Sterile oscillation scenarios in Super-Kamiokande and MACRO. Adding muon charge discrimination in future large mass detectors allows the search for MSW-like resonances in subdominant contributions to 3 or more flavour oscillations. In particular, some sensitivity to the determination of the sign of the atmospheric Am2 and to complicated hybrid oscillation scenarios can be obtained.

Future new atmospheric neutrino experiments are therefore an important complement to current and future long baseline neutrino programs. However, none of the experiments which have yielded indications for neutrino oscillations have so far succeeded to measure an actual sinusoidal oscillation pattern. Figure 2 shows the L/E distribution published by Super-Kamiokande [7] compared to the expectation for neutrino oscillations and to a functional form suggested by a recent neutrino decay model [8]. Once the detector resolution is taken into account, the two hypotheses are essentially indistinguishable [8]. Even though the current evidence is very suggestive of neutrino oscillations, a more precise measurement of the oscillation pattern is the only way to actually prove the oscillation hypothesis for atmospheric neutrinos. The crucial issue here is to prove that muon neutrinos do not only disappear, but actually reappear at some larger L/E. Present Status of the MONOLITH Project 37 О 1.4

10 102 103 104 LIE, km • GeV"1

Fig. 2. a) L/E distribution from Super-Kamiokande [7] compared to the best fit oscillation hypothesis (continous line), and to a parametrization corresponding to the neutrino decay model of Ref. 8 (dashed line). The oscillations are smoothed out by detector resolution, b) L/E distribution to be expected from MONOLITH for Am2 = 3 • 103 eV2 compared to the best fit oscillation hypothesis (oscillating line) and to the corresponding best fit of the neutrino decay model of Ref.8 (smooth threshold effect)

2.2. Distinction of u^ -+ vr vs. v^ -* vs. If the current indications for three independent Am2 are confirmed, the only way out is the introduction of at least a fourth neutrino. Taking into account the LEP results [9] on the number of neutrinos, any extra neutrinos must be either very massive (га„ > Mz/2) or sterile with respect to weak interactions (e. g., a right-handed neutrino or left-handed antineutrino). Present experiments are only now starting to distinguish oscillations with sterile neutrinos from standard flavour oscillations for either atmospheric or solar neutrinos. Significant v^—Vg oscillation contributions are therefore allowed in both cases. Since the existence of one or more light sterile neutrinos would be evidence for new physics, proving or disproving the Vp—vT oscillation hypothesis for atmospheric neutrinos is a crucial issue. Furthermore, it would exclude or confirm a large class of neutrino oscillation models (see, e.g., [10] and references therein). The observation of r appearance in long baseline beams would be the most direct evidence for v\i—vT oscillations, but potentially leaves some loopholes for the interpretation of the atmospheric neutrino results. These loopholes would be closed by a direct v^ —» vT vs. v^ —* i/s distinction in atmospheric neutrino experiments. The MONOLITH experiment can exploit the spirit of both approaches to improve on these measurements. The first technique, based on r appearance, consists in measuring the up/down asymmetry of «NC-like» events (muon-less events) as a function of the visible 2 2 2 energy. For Am < 10~ eV , oscillations of v^ into vT would in fact result in an excess of muon-less events produced by upward neutrinos with respect to muon-less downward, since charged-current vx interactions would contribute to the muon-less event sample, due to the large r branching ratio into muon-less channels. Moreover, due to threshold effect 38 Petrukhin A. A. on r production, this excess would be important at high energy. Oscillations into a sterile neutrino would instead result in a depletion of upward muon-less events. Discrimination between v^ —> vT and i/^ —> i/s is thus obtained from a study of the asymmetry of upward to downward muon-less events. 2.3. Three Flavour Oscillations. Even in the absence of sterile neutrinos, matter effects are present in the case of a small contribution from v^—v^ oscillations at the «atmospheric» Am2. The CHOOZ limit [11] constrains the size of this contribution, and the expected effects are mostly small. However, if the v^—v,, mixing angle would be close to the CHOOZ limit, the corresponding MSW resonance might be observable as a localized utl rate suppression either in v^ or in v^, yielding a measurement of the sign of Am2. This possibility is currently being investigated further. Another possible exception could be the observation of a large neutrino/antineutrino asymmetry predicted by tri-maximal mixing models [12], which are currently not yet completely excluded. Furthermore, fits of the solar neutrino results leaving out Homestake allow the solar Am2 to be as high as a few 10~4 eV2. If the mixing is nonmaximal, differences in the v and v rates could again be observed. 2.4. Physics with CNGS Neutrino Beam. In addition to atmospheric neutrinos, for which it has been optimized, MONOLITH will also detect more than 100000 events (CC + NC) per year from the CERN to Gran Sasso (CNGS) neutrino beam, which is scheduled to start operations in 2005. Beam neutrinos arrive at an angle of 4° (upward) from the horizontal direction, almost parallel to the «long» axis of the MONOLITH detector. Despite its optimized design for atmospheric neutrinos, the MONOLITH detector in its uniformly horizontal option can re- construct « 46 % of the CC events with optimal energy resolution. In terms of events, this result makes MONOLITH the detector which will monitor the CNGS beam with the highest statistical precision and good energy resolution (pi 41000 events/year expected to be reconstructed with energy resolution < 22 %). This huge sample can complement the MONOLITH atmospheric neutrino analysis of the L/E distribution in the region 20-200 km • GeV"1. The remaining part of the data is reconstructed with less precision, but can still be used for flux monitoring purposes.

3. COSMIC RAY MUON STUDIES

MONOLITH is also well suited for the study of very high energy cosmic ray muons. Using a so-called «pair meter» technique [13], the energy of each muon in the multi-TeV range can be directly measured. The resulting measurement of the cosmic ray muon spectrum yields the opportunity to test potential extensions of the standard model which would affect the shape of this spectrum. 3.1. Muon Energy Spectrum. The spectrum of primary cosmic ray particles measured by means of extensive air shower (EAS) observations is well described by a power law over many decades in energy. However, it exhibits a well-known change in slope (the so-called «knee» ) between 1015-1016 eV (1-10 PeV). In nucleon-nucleon collisions, PeV energies in the laboratory frame correspond to centre-of-mass energies in the TeV region, where new physics and new (heavy) particles are predicted by various theoretical models. Above 2 TeV, these energies are out of reach of the existing accelerators. Present Status of the MONOLITH Project 39

There are two main possibilities for the explanation of the knee: a. change of the primary spectrum (and/or composition), or the appearance of new processes in very high energy interactions. Currently, most investigators support the first point of view. However, the second option remains possible. To explain the knee in this way, one should assume that the primary spectrum is not changed, but the part of the primary energy (above the knee) is carried away by the particles which are not or incompletely detected by EAS arrays. This missing energy AE may be estimated as [14]:

Д7/72 (i) E where 71 and 72 are the power indices of the primary spectrum and of EAS spectrum above the knee; E\ and E2 are primary particle energy and detected EAS energy; Д7 = 72 — 71, and EQ is the knee position. Among the known particles only neutrinos and very high energy muons can carry this missing energy. Neutrinos remain undetected, and usual EAS set-ups do not measure the muon energy but only estimate their number. In turn, these leptons could originate from the decay of new particles in the several TeV mass region or from a new state of matter (supersymmetry, compositeness, technicolor, quark-gluon plasma, superstrings, etc.). To have the necessary impact, such particles or states should be produced with a large cross section (of the order of tens of mb). As was pointed out recently [14], a good possibility to find such new particles in cosmic rays is to perform a direct measurement of the muon energy spectrum in the region of 100 TeV and higher. The expected fluxes of very high energy muons which are needed for the explanation of the knee within the frame of the above model are shown in Fig. 3 together with the rate of standard cos- 1 2 3 mic ray muons (originated from тг, К decays). In 10° 10 10 10 10" Surface muon energy, TeV 3 years the MONOLITH detector will allow to reg- ister about 100 events with surface muon energies Fig. 3. Expected integral spectra of muons exceeding 100 TeV for conventional muon production mechanisms, whereas about 2-3 times higher from various processes crossing MONO- statistics may be expected in the case of the appear- LITH in 3 years: 1 — muons from 7r, К decays; 2 — «prompt» muons for a ratio ance of the «new» muoris. At 1 PeV, the expected 3 rate could be increased by more than an order of •R^/TT = 10 ; 3,4 — lower and upper esti- magnitude. This drastic change of the muon energy mates for muons from new VHE sources to spectrum around 100 TeV would be an excellent explain the «knee» (see text) signature of new processes. In the case of a positive result, muon energy spectrum studies with MONOLITH may lead to the discovery of new physical processes of muon generation in the TeV (centre.^of-mass) energy region. On the other hand, the absence of an excessive muon flux in this energy range would reinforce the evidence in favour of a cosmophysical origin of the knee. -..-• 40 Petrukhin A. A.

3.2. The Pair-Meter Technique for TeV Muons. The size and the structure of the MONOLITH detector allow one to utilize a new method of muon spectrometry in the TeV energy range. The technique is based on the energy dependence of the cross section of direct electron-positron pair production by muons. In passing through a thick layer of matter, high-energy muons produce secondary cascade showers, mainly via electron-positron pair production process. Measurements of the number and energies of these cascades allow one to evaluate the muon energy. The most important advantage of the technique in comparison with other existing methods of muon spectrometry is the absence of upper limitation on measurable muon energy: the energy resolution is not deteriorated with increasing of particle energy. The relative energy measurement error is determined mainly by the total target thickness T expressed in radiation lengths:

(2) where a = 1/137 is the fine structure constant. The most informative region of energy transfers from the view-point of muon energy reconstruction is e - (10~3-10~2) E^; thus, to perform measurements in the energy range E^ ~ 10 TeV and higher, it is sufficient to detect secondary electromagnetic cascades with energies £ ~ 10 GeV. Due to the fact that the technique is based on the observation of multiple interactions of muon in the set-up material, it is rather tolerant with respect to individ- 120 ual cascade energy measurement errors, and ,; i/. may be implemented with a detector of a 100 relatively simple structure. IPeV 80 •4 r The thickness of the target in MONO- LITH (over 500 Xo) provides the possi- '.i • bility to measure individual muon energies, 40 30 TeV .• and hence to obtain direct measurements of 10 TeV . 20 the energy and angular distributions of the detected particles. Examples of simulated events for several muon energies are given 0 100 200 300 400 Strip number in Fig. 4. On the average, a 10 TeV muon will generate (via electron pair production) Fig. 4. Muons with different energies in the MONO- 5 cascades with energies greater than 10 LITH digital calorimeter. 120 layers of 8 cm steel GeV and about 20 cascades above 1 GeV. and 3 cm wide strips have been assumed in the sim- For a 100 TeV muon, the typical number ulation of the detector response of cascades will already be 5, 20, and 50 for cascade energy thresholds 100, 10, and 1 GeV respectively, which will allow one to estimate the muon energy with about 50 % accuracy (and approximately log-Gaussian distribution of errors). Such accuracy is quite sufficient for muon energy spectrum studies. The MONOLITH detector operated as a pair meter will have several thousand (!) times larger effective acceptance than the biggest magnetic spectrometers [15, 16], the results of which are still used as a standard of high energy muon spectrum measurements. Present Status of the MONOLITH Project 41

CONCLUSION •=•-.•:,••'.,/, : ; •.• • :• ,• • •.-, ;. •

We showed the feasibility of a34kt magnetized iron detector which is able to

• measure the neutrino bscillatioh pattern in atmospheric neutrinos, therefore proving the oscillation hypothesis. Auxiliary beam measurements are being studied;

• significantly improve the measurement of Am2 and sin2#, covering the full allowed

. range independent of how it-mightevolve in the.future; ; ; : . ' ;.:. ..;...

• improve the discrimination between the f^—Vr and v^—vs oscillation hypotheses. De- pending on the oscillation parameters, this could be achieved through the measurement of the up/down ratio of NC-like events, through the presence, or absence of a distortion of the oscillation pattern as a function of energy by matter effects, or through ihe (non)observation of an asymmetry in the oscillation patterns for neutrinos and antineutrinos;

• with some luck, measure charge-dependent matter effects in i/^—ue oscillations, and therefore the sign of Am2, i. e., the hierarchy of the mass pattern.

• obtain the first measurement of the cosmic ray muon energy spectrum and angular distribution around 100 TeV, and significantly improve the knowledge of muon production mechanisms and the primary spectrum and composition in the 100 TeV-10 PeV region.

Many of these measurements are unique to this detector, and therefore complementary to other planned neutrino physics programmes. They can be achieved on a relatively short timescale, at a cost which is dominated by the required detector mass. Furthermore, a detector of this kind fits into even more ambitious long-term programmes for neutrino factories.

REFERENCES

1. See e.g. PakvasaS. II Invited talk at the 8th Intern. Workshop on Neutrino Telescopes, Venice, Italy, Feb. 23-26, 1999; hep-ph/9905426, and references therein. 2. MONOLITH Proposal, LNGS P26/2000, CERN/SPSC 2000-031, SPSC/M657. August 15, 2000. http://castore.mi.infn.it/ /sim monolith/proposal/

3. PicciP., Pietropaolo F. Atmospheric Neutrino Oscillations Experiments. ICGF RAP. INT. 344/1997. Torino, 1997; CERN preprint SCAN-9710037.

4. BuenoA., CampanelliM., RubbiaA. hep-ph/0005007. May 2000.

5. Bencivenni G. et al. II Nucl. Instr. Meth. A. 1994. V.345. P. 456.

6. Gustavino С et al. INFN/TC-99/14; Nucl. Instr. Meth. (accepted).

7. Fukuda Y. et al. (Super-Kamiokande Collaboration) II Phys. Rev. Lett. 1998. V. 81. P. 1562.

8. Barger V. et al. //Phys. Rev. Lett. 1999. V.82. P.2640.

9. Particle Data Group II Eur. Phys. Journ. С 1998. V.3. P. 1. 42 Petrukhin A. A.

10. BilenkyS.M. et al. hep-ph/9906251.

11. Apollonio M. et al. (CHOOZ Collaboration) II Phys. Lett. B. 1999. V.466. P.415.

12. Harrison P. F., Perkins D. H, Scott W. G. II Phys. Lett. B. 1999. V.458. P. 79.

13. Kokoulin R.P., Petrukhin A. A. /I Nucl. Instr. Meth. A. 1988. V.263. P. 468; Sov. J. Part. Nucl. 1990. V. 21. P. 332.

14. Petrukhin A.A. II Talk at the Xlth Rencontres de Blois «Frontiers of Matter», Blois, France, June 28-July 3, 1999; preprint МЕРЫ, No. 006-99. 1999.

15. Allkofer O.C. et al. II Nucl. Phys, B. 1985. V. 259. P. 1.

16. Matsuno S. et al. II Phys. Rev. D. 1984. V.29. P. 1. XJ0200055

Письма в ЭЧАЯ. 2001. №6[109] Particles and Nuclei, Letters. 2001. No.6[109]

УДК 539.126

TECHNIQUE OF NEUTRINO-INDUCED MUON DETECTION ON THE EARTH SURFACE V. M. Aynutdinov, V. V. Kindin, K. G. Kompaniets, A. A. Petrukhin, D. A. Room, V. V. Shutenko, A. V. Stepanov, 1.1. Yashin Moscow State Engineering Physics Institute (Technical University), Moscow

Methods of the rejection of atmospheric muon background for cosmic ray neutrino detection in a ground level Cherenkov water detector are described. The background rejection factor on the level 1010 is reached, and thereby a possibility to detect neutrino-induced muons on the Earth surface is shown. Описана методика подавления фона атмосферных мюонов при регистрации нейтрино космиче- ских лучей в наземном черенковском водном детекторе. Получен фактор режекции фона на уровне 10 миллиардов, что подтверждает возможность регистрации мюонов, генерированных нейтрино, на поверхности Земли.

INTRODUCTION

То decrease the background of cosmic ray muons sophisticated neutrino telescopes are usually placed deep underground or underwater. At the same time, not less complex arrays are constructed on the Earth surface for studies of EAS and other components of cosmic rays. Therefore the idea to combine these types of detectors is very attractive. The main problem is related with very large background on the surface for rare neutrino events detection. Theoretical considerations of this problem show that even on the Earth surface the interval of zenith angles exists, in which the flux of muons induced by neutrinos from the bottom hemisphere is larger than the flux of backward scattered near-horizontal atmospheric muons and of products of their interactions [1-3]. However, the grave technical problems connected with the necessity of selection of rare neutrino events from various kinds of background events still remain. Therefore measuring system of the set-up must provide the reliability of distinction of particle motion directions on the level better than 1010. In principle, Cherenkov Water Detectors (CWD) can provide such rejection factor by using the directionality of Cherenkov light. Besides, CWD of large area and volume can be constructed relatively easily. However, significant fluctuations of photomultiplier response at small fluxes of Cherenkov light, and hence the probabilistic nature of CWD response, create some difficulties in the reliable track reconstruction. In the present work the technique of neutrino events selection at large atmospheric muon background is described. This technique was elaborated and tested by using experimental data collected with Cherenkov water detector NEVOD during experimental runs in 1996-1997. 44 Aynutdinov V. M. et al.

1. EXPERIMENTAL ARRANGEMENT

NEVOD is the Cherenkov water detector on the Earth surface. General view of the first part of the set-up is given in Fig. 1. Crosses in the figure represent the quasispherical modules (QSM) which form nearly cubic spatial lattice with inner fiducial volume 6 x 6 x 7.5 m. Detector description is presented in paper [4]. Hereafter only three basic systems of the set-up are described: QSM, triggering system and calibration telescope system (CTS). The parameters of these systems are especially important for the analysis of rare events. Each quasi-spherical measuring module consists of six PMTs with flat photocathodes with 15 cm diameter. PMTs are directed along rectangular coordinate axes. The main Fig. 1. Schematic view of NEVOD-91 set-up with feature of QSM is the possibility to deter- 91 QSM and 2 x 32 scintillation counters mine the Cherenkov light direction on the basis of amplitude analysis only (without time- of-flight technique). On the top and on the bottom of the tank, 2 x 32 scintillation counters are located, which form 1024 narrow angle telescopes. Telescope system allows one to select atmospheric muons in zenith angle interval from 0° up to 45°. The use of this telescope system gives the possibility to determine the NEVOD response for the passage of a single muon, and to calibrate the measuring modules. Two-level trigger system was used in the set-up. In each module, the coincidence of two adjacent PMTs within 80 ns time gate is required. The higher level: trigger is.based on the coincidences of signals between several modules, the number and location of which may be varied. In the present analy-. sis, events selected with the following conditions were used: nd and nu correspond to coincidences of ^ n quasi-spherical modules with hit downward- and upward-looking Ю1 4 6 PMT, respectively; t x b is the telescope trig^ Coincidence multiplicity ger, i. e., coincidences of any top (t) scintillation counter with any bottom one (b). Trigger d allows one to select events from the bottom Fig. 2. Dependence of trigger rate on coincidence hemisphere: Cherenkov radiation is detected multiplicity: • — trigger с; А — trigger и; Т — by down-looking PMT. Organization of the triggerd trigger и is the same as d trigger (symmet- rical triggers). It gives the possibility to,use down-going atmospheric muons for estimation of d trigger efficiency. The dependence of basic trigger rates on the coincidence multiplicity is presented in Fig. 2. For the comparison, Technique of Neutrino-Induced Muon 45

с trigger rate (coincidences of signals from any triggered QSM) is also shown. This trigger allows to accept events from any direction. The main operating time of the set-up includes the threshold coincidence multiplicity equal to 7. Under this condition, d trigger rate was about 5-6 Hz, and и trigger rate about 2 KHz. The rejection factor for downward going atmospheric muons on the trigger level was close to 300 (it was estimated as the ratio of counting rates for 7u and Id triggers).

2. CALIBRATION EXPERIMENTS

The principle of u-induced events selection is based on the comparison of Cherenkov water detector response for downward and upward going muons taking into account the set-up symmetry. Downward going atmospheric muons were selected with scintillation telescope trigger and и trigger. Data acquisition system provided the possibility to record each telescope event (t x b) during all the period of exposition. Set-up calibration with г* trigger was conducted in two modes of operation. In 1996, the basic exposition with d trigger was switched off during such calibration. In 1997, collection of calibration data was realized without breakdown of the exposition. In this case, for reduction of и trigger rate from 2 KHz to an acceptable level, one event from every 4096 was selected on the hardware trigger level. The full set of calibration data contains about 2 • 107 telescope events and ~ 107 и trigger events detected during the expositions of 1996-1997. 2.1. Response for Vertical Muons. Us- ing scintillation telescope trigger, the re- «p sponse for near vertical muons (0°—30°) was j measured. An example of such event is shown in Fig. 3. It is possible to introduce the following parameters, which characterize JVO=11 the single muon response: 7V0 — the total N number of triggered modules, which allows one to separate muons and cascade showers; |MTJ = 3 4* / Ч~ Nd — the number of down-looking PMT |АЛу = О which in the case of near vertical down- Zav = 3-9 m going muons detection can be triggered, e. g., by dark noise, knock-on electrons and scattered light; ANz — difference of the num- f"9 1 ""1 ber of QSM which show opposite directions / of Cherenkov light in vertical; ANx and efa ANy — differences between the numbers of QSM which show opposite directions along Fi8- 3- Vertical muon event selected by calibration the set-up axes in horizontal plane; Zav — av- telescope. Hit phototubes are marked by the circles erage vertical coordinate of triggered QSMs. In Fig. 3, these parameters are shown for the displayed event. The average values of these parameters and typical ranges of their variation (about 90 % of the events) for telescope muons are given in Table 1. 46 Aynutdinov V. M. et al.

Table 1. Parameters of NEVOD response for near vertical telescope muons Parameter Symbol Average Limits No. triggered QSM No ~ 12 7-18

No. down-looking PMT Nd ~1 0-2 о Difference of numbers of AiVz 4-14 Triggered QSM indicating AiVx 0 ±4

Opposite directions ANY 0 ±4 Average Z-coordinate, m Zav 4.1 3.2-4.7

2.2. Reliability of Muon Motion Direction Reconstruction. From the point of view of selection of rare events from the bottom hemisphere in conditions of large background, the important characteristic of the set-up is the reliability of reconstruction of particle motion direction. In the first turn, this parameter depends on reliability of estimation of light direction by means of measuring modules. The reliability of this estimation is determined by detection system properties (PMT dark noise, cross-talks, etc.), by characteristics of water (absorption and scattering lengths), and by probability of the production of secondary particles by muon. Quantitatively, the reconstruction reliability is directly related with probabilities of various PMT number combinations of module PMT hits. For estimation of these probabilities, muons selected 1 © by scintillation telescope system were used. Only events with large fraction.of hit PMTs Muon track that can detect direct Cherenkov light from the track (more 70 %) were used for analysis. This Fig. 4. Scheme of module calibration (top view) selection allows one to suppress multiparticle events. Experimental probabilities of various combinations of hit PMT of the measuring module are presented in Table 2 for vertical muon calibration (see Fig. 4). Muon track was located at the distance equal to 1.25 m from the module centre. PMT No. 6 and No. 5 (not seen in Fig. 4) are directed up and down, correspondingly. Only PMT

Table 2. Probabilities of various combinations of hit PMT in triggered QSM Combinations with 6-1 6-2-1; 6-3-1 6-2; 6-4 6-2-3; 6-4-3 6-3 up-looking PMT (No. 6) Average azimuth angle 0° 45°; 315° 90°; 270° 135°; 225° 180° Experiment 0.81 3.0-10"2 0.8 • Ю-2 -, 0.2-lO"3 0.4-10-2 2 3 Simulation 0.80 3.1 • 10"2 1.0 • Ю- 0.2 • 10- 0.5 • 10"2

Combinations with 5-1 5-2-1; 5-3-1 5-2; 5-4 5-2-3; 5-4-3 5-3 down-looking PMT (No. 5) Experiment 3.0; Ю-2 0.9 • 10~3 0.8 • 10-3 1.3 • КГ5 0.6 -10~3 Simulation 2.9-10-2 1.1.-КГ3. 1.0-lO"3 0.4 -КГ5 0.7 • 10"3 Technique of Neutrino-Induced Muon 47

5-, 4- 315 315 3- 2- 1- lgJV 0- 1- 2- 3- 4- 225 225 135 5-

Fig. 5. Azimuth diagrams of measuring module response: a) combinations with up-looking hit PMT; b) with down-looking hit PMT

No. 1 and No. 6 can detect direct Cherenkov light. Each combination of triggered PMT corresponds to a different average azimuth angle indicated by a measuring module. Examples of averaged azimuth diagrams of the module response are presented in Fig. 5. The results of module response simulation for identical conditions are presented in Table 2, too. Simulation data are in a reasonable agreement with the experiment. On the basis of calibration experiments and j ,™ simulation data, it is possible to calculate the probability of observation of a given response as a function of track location Р(в, £) = P -5 - (response | 6, £), where в is zenith angle, £ is \ the vector of other track parameters. Examples -10- i. of P(9, £) averaged over £ are presented in e~o° :... [9-30° Fig. 6 for two calibration muon events with в -15- about 0° and 30°. For nearly isotropic muon flux this proba- -20 - bility function P(9, £) could be used for maximum likelihood estimation of track parameters. •>S • However, in conditions of the heavily changing 0 20 40 60 80 100 120 140 160 180 flux (the order of 1010 in the case of up-going muon reconstruction), plain likelihood tech- Fig. 6. Examples of averaged probability func- nique can lead to a serious error of zenith angle tions P(0) calculated for a vertical muon and an estimate. To prevent such mistakes, the con- inclined one (arb. units) ditional probability F(9 \ response) that takes into account the realistic zenith angle distribution of muons was used for individual event analysis:

dF(fl) J (1) d cos в 48 Aynutdinov V. M. et al. where Ifi(0) is the intensity of muons near the Earth surface (mainly atmospheric muons up to 90°, scattered muons from 90° to 120°, neutrino-induced muons for angles more than 120°). The position в of maximum of the conditional probability distribution is determined both by the angular dependence of muon flux and by P(response \в, £). Taking into account the detector symmetry, calibration data were used for investigations of conditional probability distributions for upward going muons. For this purpose, calibration events (7u trigger with some additional cuts discussed in section 3.2) were turned-over (mirror events), and distributions F(6 | response) were calculated. Dis- tribution of the value of в corresponding to the maximum of F(6 \ response) for 100 mirror events is presented in Fig. 7. One 0 20 can see that more than 90 % selected mirror events are reconstructed as up-going muons. It evidences for the possibility of selec- Fig. 7. Distribution of в for turned-over (mirror) tion of real up-going muons in the NEVOD events set-up.

3. EXPERIMENTAL STUDIES

For experimental study of possibilities to detect neutrino-induced muons on the Earth surface in Cherenkov water detectors, series of experiments was conducted with NEVOD set-up in 1996-1997. The total net operation time was more than 2 • 103 h., 3.4 • 107 events were recorded with Id trigger. Total number of atmospheric muons that could be detected with trigger lu during this period was about 1010. Obtained experimental material permits to investigate the main sources of neutrino-induced event imitations and to analyze the possibilities of neutrino detection. The experimental data treatment included the following main stages: data cleaning, event selection on the basis of simple fast criteria, and subsequent consideration of selected events. The final objective of the analysis was the search of neutrino event candidates. 3.1. Experimental Data Cleaning. Preliminary data cleaning was conducted during experimental runs and data base accumulation. The runs which contained malfunctions in measuring system operation were excluded. The main objective of the off-line data cleaning was the exclusion of separate unreliable measuring modules from the analysis. Taking into account the necessity of selection of rare events, measuring module was excluded if probability of a false hit of any its PMT was more than 10 %. Estimation of these probabilities was made on the basis of results of scintillation telescope calibration. As an example, time dependence of the number of excluded modules is presented in Fig. 8 for one of the periods of the exposition. Additionally, this analysis allows one to exclude the periods of exposition with abnormally large number of unreliable QSM. Technique of Neutrino-Induced Muon 49

3.2. Fast Criteria for Event Selection. The parameters listed in Table 1 (Section 2.1) can be used for preliminary upward-going muon selection. At the choice of critical values of these parameters it is necessary to take into account that narrow interval of parameter values results in a higher rejection factor but lower efficiency of useful event selection, and vice versa. As a compromise, the criteria presented in Table 3 were chosen for upward-going muon selection. Additional pa- о rameter JV (the number of triggered QSM 10.02.97 02.03.97 22.03.97 out 20.02.97 12.03.97 01.04.97 in external planes of the set-up looking out of Date the fiducial volume) was introduced to exclude the events, which are generated outside of the Fig. 8. The number of modules excluded from the sensitive volume, mostly in the areas along the analysis vs. time water tank. The corresponding background rejection factors JR are given in Table 3, too. The combined rejection factor for this preliminary selection is about 105.

Table 3. Criteria for upward-going muon selection Parameter Range /л Comments No 8-20 Exclude noise and showers,

Nd <2 ~30 and indicate particle

ANZ <-8 motion direction

\ANX\ ^4 Select muons crossing

\ANY\ <4 ~80 central part of the fiducial iVout ^2 volume >3.7 ~40 Suppresses near-horizontal muons Combined rejection factor ~105

As a result, 328 events (of 3.4 • 107) remained after selection with these criteria. The probability to lose useful events was about 65 % (i. e., 35 % selection efficiency). This efficiency was estimated both with scintillation telescope events (35 %) and events selected with lu trigger (34 %). 3.3. Analysis of Selected Events. The classification of selected events was performed on the basis of the track reconstruction made under assumption that single muons were detected. Two main parameters obtained with this procedure were used for the analysis: • в — reconstructed zenith angle of the track; • r — the ratio of the number of hit PMT which can detect direct Cherenkov light from the reconstructed track to the number of all triggered PMT. Distribution of events in the values of these parameters is presented in Table 4. 50 Aynutdinov V. M. et al.

Table 4. Classification of selected events 9 r Number of events 1996 1997 Total

Any <0.7 46 78 124 < 120° >0.7 66 126 192 ^ 120° >0.7 4 8 12

- a 2 io-'

% ю-3 I — g io-5 - 1 a. g 10"7 IT

9 io- ~ I 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 20 40 60 80 100 120 140 160 180 0 20 40 60 80 100 120 140 160 180 e e

Fig. 9. Examples of F(0 | response) for two of selected events

About 40% of selected events have the value of r less than 0.7. At the same time, for real muons 95 % of events satisfy the condition of r > 0.7 for 7u trigger, and 93 % for telescope trigger. The low value of r indicates 2 - -|-|- that this group of events does not correspond to single muons. Most part of these events may 1 - be interpreted as muon groups, showers, and random muon coincidences. III HI 1 1 1 1 1 The second group of events (somewhat 0 20 40 60 80 100 120 140 160 180 less than 60 %) contains mainly near-horizontal Zenith angle muons. Only the third group (12 events) can contain Fig. 10. Distribution of 12 selected events in neutrino candidates and is interesting for the zenith angle estimation subsequent analysis. For all events of this group the conditional probabilities of various zenith angles at a fixed response F(9 \ response) were calculated (under the same assumption that they are single muons). The zenith angle distribution of muons near the Earth surface was taken into account for this estimations (see Eq. (1)). Typical examples of 9 distributions are presented in Fig. 9. Eleven events have a character of distribution similar to that shown in Fig. 9, a (two peaks, global maximum near 90°). These events can be interpreted as near- Technique of Neutrino-Induced Muon 51 horizontal atmospheric muons. Only one event has maximum of probability distribution in the range of zenith angle greater than 120° (Fig. 9,b). The distribution of zenith angles 0, which correspond to the maximum of F(6 | response), is shown in Fig. 10 for all 12 selected events. Only one event can be regarded as neutrino event candidate — zenith angle estimation is about 150° (see Fig. 9, b).

4. DISCUSSION

In the frame of the present analysis, selection of the upward-going muons included the following steps: — primary experimental data were selected with Id trigger which provided the background rejection factor about 300; — fast criteria of event selection elaborated on the basis of the analysis of single muon response taking into account the set-up up-down symmetry were applied (rejection factor ~ 105); — the individual analysis of events was conducted on the basis of track reconstruction and calculations of the conditional в distributions for the given response (rejection factor ~ 300). As a result of the analysis of experimental data, one candidate for neutrino-induced muon was selected. The experimental estimation of the number of atmospheric muons that crossed fiducial volume of detector during the time of the exposition was about 1010. So, the rejection factor of atmospheric muons on the level 1010 has been reached with Cherenkov water detector NEVOD. Full set of criteria for neutrino candidates provides the efficiency of useful event selection equal to 30 %, which is determined by 35 % efficiency of fast criteria, 95 % efficiency of track reconstruction and about 90 % efficiency of probabilistic analysis (Fig. 7). Systematic uncertainty of the efficiency estimation can arise from some up-down asymmetry of the NEVOD set-up. Firstly, the water layer under the measuring modules lattice is slightly more than the water layer over it. Secondly, because of sediments on photocathodes of up-looking

Fig. 11. Possible interpretations of the neutrino event candidate: a) front view; b) side view 52 Aynutdinov V. M. et al.

PMT during long-time experiment, the sensitivity of down-looking PMT is somewhat higher on the average. It is necessary to underline, that both these factors lead to the increase of the efficiency of useful event selection. There is a question: is the selected neutrino candidate the real neutrino event? In this connection it is important to note that all the criteria of event selection are necessary but not sufficient conditions for interpretation of the event as neutrino-induced muon. In particular, zenith angle probability distributions for events were calculated without taking into account correlations between measuring module responses. Besides, a priori information was used that all events were single muons. However, more complicated events also could be detected during the exposition and could pass the selection criteria. As an example, in Fig. 11 two possible interpretations of the selected neutrino candidate are presented: neutrino-induced muon from the bottom (Fig. 11, a) and pair of near-horizontal muons (Fig. 11, b).

CONCLUSION

The analysis of upward-going muon imitations showed that the most significant source of the background is represented by near-horizontal muons. The described technique of the events selection has provided the suppression of atmospheric muon background with a rejection factor on the level of 1010. Thus, the possibility to detect high energy neutrinos with Cherenkov water detectors on the Earth surface is shown. The relatively low efficiency of useful event selection obtained in the present configuration of the detector (about 30 %) was caused by a small thickness of the fiducial volume (6 m). The efficiency can be significantly improved with the increase of the set-up sizes. Acknowledgements. Authors are grateful to R. P. Kokoulin for discussions, which helped us to shape our understanding of u-induced events selection procedure described here, and to all people who took part in Cherenkdv water detector NEVOD creation and operation. The research is performed with a financial support of the Ministry of Education, the Ministry for Science and Technology of the Russian Federation (Project NEVOD, reg. no. 01-63), and RFBR (grant 99-02-18177) .

REFERENCES

1. Elbert J. W. et al. II Europhys. Lett. 1991. V. 14. P. 181.

2. Aynutdinov V.M. et al. Neutrino Telescopes / Ed. by M.Baldo Ceolin. Venezia, 1996. P. 429.

3. Aynutdinov V. M. et al. II Astropart. Phys. 2000. V. 14. P. 49.

4. Aynutdinov V. M. et al. II Proc. of the 24 ICRC. Roma, 1995. V. 1. P. 1076; 1060; 1072. XJ0200056

Письма в ЭЧАЯ. 2001. №6[109] Particles and Nuclei, Letters. 2001. No.6[109]

УДК 539.16 HIGH-SENSITIVE SPECTROMETER OF FAST NEUTRONS AND THE RESULTS OF FAST NEUTRON BACKGROUND FLUX MEASUREMENTS AT THE GALLIUM-GERMANIUM SOLAR NEUTRINO EXPERIMENT (SAGE) J. N. Abdurashitov, V. N. Gavrin, A. V. Kalikhov, V. L. Matushko, A. A. Shikhin, V. E. Yants, O. S. Zaborskaia Institute for Nuclear Research, Russian Academy of Sciences, Moscow

The principle of operation, design, registration system and main characteristics of a fast neutron spectrometer are described. The spectrometer is intended for direct measurements of ultra-low fluxes of fast neutrons. It is sensitive to neutron fluxes of 10~7cm~2 -s"1 and lower. The detection efficiency of fast neutrons with simultaneous energy measurement was determined from Monte-Carlo simulation to be equal to 0.11 ± 0.01. The background counting rate in the detector corresponds to a neutron flux of (6.5 ±2.1) • 10~7cm~2 -s"1 in the range 1.0-11.0 MeV. The natural neutron flux from the surrounding mine rock at the depth of 4600 meters of water equivalent was measured to be (7.3±2.4)10~7cm~2-s~1 in the interval 1.0-11.0 MeV. The flux of fast neutrons in the SAGE main room was measured to be 2.3 • 10"7cm"2 • s"1 in 1.0-11.0 MeV energy range.

Описываются конструкция, принцип действия, система регистрации и основные характе- ристики спектрометра быстрых нейтронов. Спектрометр предназначен для измерения ультра- слабых потоков быстрых нейтронов. Его чувствительность к нейтронным потокам составляет 10~7см~2 -с"1. Эффективность регистрации быстрых нейтронов с одновременным измере- нием энергии была определена моделированием с использованием метода Монте-Карло и со- ставляет 0,11 ± 0,01. Фоновая скорость счета в детекторе соответствует потоку нейтронов (6,5±2,2)-10~7см~2-с~1 в области 1,0-11,0 МэВ. Измеренный нейтронный поток от окружающей породы на глубине 4600 м в. э. составляет (7,3 ± 2,4) • 10~7см~2 • с"1 в области 1,0-11,0 МэВ. В главном помещении галлий-германиевого нейтринного эксперимента измеренная величина потока быстрых нейтронов не превышает величину 2,3 • 10~7см~2 • с"1 в области 1,0-11,0 МэВ.

INTRODUCTION

It is well known that fast neutrons from the surrounding rocks are one of the background sources for an underground experiments, as the solar neutrino flux registration and search for double beta decay. The sources of the fast neutrons are (a,n) reactions on the light elements (C, O, F, Na, Mg, Al, Si). Measurements of fast neutron flux in the laboratory of the Gallium-Germanium Solar Neutrino Telescope (GGNT) and in an unshielded room at the same depth have been performed using a special high-sensitive spectrometer with registration system based on a fast two-channel digital oscilloscope. The laboratory is located under Mt. Andyrchy (Northern Caucasus Mountains, Russia) in a tunnel that penetrates 3.5 km into a mountain, at a depth of 4600 m of water equivalent. 54 Abdurashitov J. N. et al.

1. DETECTOR STRUCTURE AND OPERATION PRINCIPLE

The neutron spectrometer was designed taking into account requirements for minimizing the background from 7 rays and random coincidences. It is a calorimeter based on a liquid organic scintillator-thermalizer viewed by photomultipliers (PMTs) with 3He proportional counters (NCs), uniformly distributed through the scintillator volume [1].

Fast neutrons with En > 1 MeV entering the scintillator are decelerated down to thermal energy, and diffuse in the detector volume until they are either captured in a neutron counter or captured by scintillator protons or leave the detector. The amplitude of the light scintil- lations from recoil protons, which are produced during neutron thermalisation, is on average proportional to the initial neutron energy. About 19 of all proportional counters provide a. «neutron label» of an event. Such a technique allows us to suppress the external 7-ray background significantly. Low-level signal electronics for the spectrometer were designed with signal-tp-noise ratio optimization and full pulse shape analysis required for efficient rejection of background events. A data acquisition and processing system is based on a fast (100 MHz) two-channel PC/AT interfaced digital oscilloscope. The acquisition software was written in the С programming language. To simplify the structure of the apparatus, signals from all PMTs and NCs are multiplexed into independent channels called «PMT channel» and «NC channel» respectively. A signal from the NC channel triggers the data acquisition system. The full waveform of events in the PMT and NC channels are registered independently inside selected time intervals before and after the trigger, they are called «prehistory» and «history» accordingly. These time intervals can be adjusted on the basis of calibration measurements. The last generation of the data acquisition system is described in detail in [2].

2. MEASUREMENTS Three series of measurements were performed with different background conditions. In the first series, the fast neutron flux in one of the additional rooms of GGNT was measured. To suppress the 7-ray background the detector was shielded by an 8 cm thick lead shield. In the second series, the spectrometer without any shielding was situated in the main room of GGNT, which is shielded with special low-background concrete and steel. In the third series, the internal background of the detector was investigated. A shielding of 30 cm of borated polyethylene and 35 cm of water was used. To calibrate the PMT channel, a 60Co 7 source was used [3]. The energy of the centre of the Compton edge was assumed to be equal to 1 MeV on the electron energy scale, which corresponds to ~ 4 MeV of a neutron energy scale. A Pu-Be source was used to calibrate the NC channel [3]. The spectrum produced by the Pu-Be source has a specific shape due to a wall effect, which distorts the counter event spectrum. In spite of this distortion, the range of energies observed for true neutron events is less narrow compared to the broad background spectrum produced by internal alphas. We used only the events from the «neutron window» coincident with PMT signals in order to suppress the internal background of the detector. The delay time is a specific feature of the detector and depends on the detector design. The acquisition system allows us to measure directly the delay time for neutron events. Such measurements were carried out using a Pu-Be source. A fitting procedure leads to a time constant of Tj/2 = 55 fis. High-Sensitive Spectrometer of Fast Neutrons 55

22.5 Efficiency 20.0 - lMeV 17.5 15.0 3MeV 12.5 5 MeV- § 10.0 7MeV о 7.5 9 MeV о 12MeV 5.0 15MeV 12.5 0.0 0 12 3 4 5 6 7 8 9 10 4 6 8 10 12 14 16 £,MeV Eo, MeV

Fig. 1. Responce functions

Fig. 2. Detector efficiency: triangles — s(Eo); circles — without threshold of registration influence

Decays of radioactive isotopes of Bi and Po, which can take place in the helium counter walls, have been considered as the main sources of the significant internal background. Beta decay of 214Bi can fire the scintillator, followed by a delayed capture a signal from Po decay in helium counters:

214 214 Bi(e) -^ Po(a) (1) thus imitating an actual neutron event. The delay time distribution of the events obtained for the series results in a time constant Ti/2 = 164 [is. It confirms our assumption about the possible origin of the detector internal background. ; The detection efficiency depends in a complicated manner on the response function of the detector. The response function for an infinite organic scintillator, calculated using Monte- Carlo simulation, is shown in Fig. 1. The neutron's thermalization process in the detector of the actual geometry was studied also. As a result we have obtained that the total detection efficiency can be expressed as composition of three contributions:

= e - eoui(En)), (2) where e « 19 % is the efficiency of registration of thermalized neutrons by 3He-counters; £out(En) is the probability for a neutron to leave the detector; eth{En) is caused by the registration threshold. Thus the dependence of efficiency on initial neutron energy has a form that is shown in Fig. 2. Unfolding of obtained spectra was not performed, the mean value of efficiency in 1-7 MeV energy range £tot = 11 % was used to calculate the neutron flux.

3. RESULTS OF THE MEASUREMENTS

The conditions of the performed measurements and the values of the calculated rates of counting are shown in the Table, for counting rates: Щ — in the PMT channel; HNC — in the NC channel; .RNW — in the «neutron window» and ЯсОГ — for the correlated events. 56 Abdurashitov J. N. et al.

Table 1. Fast neutrons background flux measurement (conditions of measurements and rates of Counting) .Rtot = H.N + Rrand + Rbkg, Rcor = Rtot - Rrand = RN + Rbkg

Conditions ^

H20 21 96.4 ±0.8 29.5 ±0.5 1.25 ±0.4 0.07 ±0.001 1.25 ±0.40 11.36 ±3.78 Mine Rock 140 ±0.4 106.4 ±0.3 46.6 ±0.2 2.96 ±0.13 0.29 ±0.17 1.42 ±0.45 12.91 ±4.29 SAGE 512±4 74.2 ±0.2 25.8 ±0.1 1.93 ±0.12 1.11 ±0.17 -0.43 ±0.45 -3.91 ±4.11

We consider three contributions to the experimentally measured counting rate Rtot: the random coincidence rate i?rand> the internal background counting rate i?bkg> and «neutron» counting rate Rn, so that

Rn — Rtot — -Rbkg — (3)

We assume that the total background 7 spectrum has the same shape as the random, coincidences spectrum. To obtain the random coincidences spectrum the total spectrum of the background gammas has been normalized to a calculated random coincidences rate. The maximum rate random coincidences in the case of absolutely independent signals in the PMT and NC channels can be calculated in the following way:

•"rand — '7'ni-"'> V*)

where r7 is the 7 rate; r% is the neutron counter rate in the determined energy window; At is the time window. Applying the subtraction procedure as described, the spectra for each series of measurements were obtained. Figures 3, a, b and 3, c, d give the details of the measured and

• I i ( 1 1 '1 I 1 1 1' t' 1 . , , i , '. ,

- f

• III , I • , 1 , I -. 1 t 1 i 1 . ( 1 0.5 1.0 1.5 2.0 2.5 0.0 0.5 1.0 2.0 2.5 £,MeV

Fig. 3. Fast neutrons amplitude distributions for Mine Rock measurements (electron scale) (o, b) and in the Main SAGE Room (electron scale) (c, d): a,c) solid line — total number; dashed line •— random coinciences; b, d) solid line — correlated event; dashed line — background of the detector High-Sensitive Spectrometer of Fast Neutrons 57 calculated spectra. Taking into account the detection efficiency of 0.11 ±0.01 and detector full square of 6267.5 cm2, the value of neutron fluxes was obtained as (7.3 ± 2.4) • 10~7cm~2 • s~x for measurements in an additional room of GGNT, < 2.3 • 10~7cm~2 • s~1 for measurements in the main room of GGNT and (6.5 ±2.1) • 10~7cm""2 • s~x for measurements in a water shield.

CONCLUSIONS

The main results of the measurements can be summarized as follows. (I) Using the fast neutron high-sensitive spectrometer neutron energy distributions with a resolution of 60 % are measured. The sensitivity of the detector is estimated as 10~7cm~2 • s"1. The registration threshold was determined as 1 MeV. (II) The efficiency dependence on initial neutron energy was calculated by Monte-Carlo simulations. In calculations of the neutron flux we used the efficiency value of 0.11 ± 0.01. (Ill) The internal background of the detector was measured in a shield of water and borated polyethylene. The background counting rate in the detector corresponds to a neutron flux of (6.5 ± 2.1) • 10~7cm~2 • s"1 in the range 1.0-11.0 MeV. Analysis of the time distribution showed that the origin of the detector background is 214Bi- 214Po decay in the walls of the helium counters. (IV) The natural neutron flux from the surrounding mine rock at a depth of 4600 meters of water equivalent was measured to be (7.3 ± 2.4) • 10"7cm-2 • s"1 in 1.0-11.0 MeV. The flux of fast neutrons in the SAGE main room was measured to be < 2.3 • 10~7cm~2 • s~l in 1.0-11.0 MeV energy range. (V) Based on theoretical calculations [3] one can show that the neutron contribution to effect, measured by SAGE, is negligible [4].

Acknowledgments. We thank J. S. Nico for intensive cooperation. We acknowledge the financial support of the Russian Foundation of Basic Research (project 00-02 26711), initiative group «Young Scientists of Russia» and Department of Sciences of Russia. This research was made possible in part by grant of RFBR No. 98-02 16962.

REFERENCES

1. Abdurashitov J.N. et al. II Instr. and Exp. Tech. 1997. V.40/6. P. 741.

2. Abdurashitov J.N. et al. http://www.xxx.lanl.gov/physics/9908035.

3. Abdurashitov J. N. et al. http://www.xxx.lanl.gov/physics/0001007; Phys. of At. Nuclei. 2000. V. 63, No. 7 (in press).

4. Kornoukhov V. N. II Ph. D. thesis. Institute for Nuclear Research of the Russian Academy of Sciences. M., 1998 (in Russian). 5. Abdurashitov J.N. et al. II Phys. Rev. С 1999. V.60. P.055801. XJ0200057

Письма в ЭЧАЯ. 2001. №6[109] Particles and Nuclei, Letters. 2001. No.6[109]

УДК 539.184

NEW EXPERIMENTAL LIMITS ON THE ELECTRON STABILITY AND EXCITATION OF NUCLEAR LEVELS IN 23Na, 127I AND 129Xe INDUCED BY THE ELECTRON DECAY ON THE ATOMIC SHELL P. Belli", R. Bernabeia, С J.Dai!1, H. L. Heb, G. Ignesti0, A. Incicchittf, H. H. Kuangb, J. M. Mab, F. Montecchiaa, O. A. Ponkratenkod, D. Prosper?, V. I. Tretyatl,Yu.G.Zdesenkod "Dip. di Fisica, Universita' di Roma «Tor Vergata» and INFN, sez. Roma2, 1-00133 Rome, Italy ь IHEP, Chinese Academy, P.O. Box 918/3, Beijing 100039, China : c Dip. di Fisica, Universita' di Roma «La Sapienza» and INFN, sez. Roma, I-OO185 Rome, Italy d Institute for Nuclear Research, MSP 03680 Kiev, Ukraine

The background measurements have been performed in the Gran Sasso National Laboratory of INFN with the help of the large mass highly radiopure ~ 100 kg DAMA Nal(Tl) set-up (34866 kg • day . statistics) and the ~ 6.5 kg liquid Xenon DAMA scintillator set-up (2257.7 kg day). New life time limits on the charge nonconserving (CNC) electron decays have been established: r(e~ —* veyeue) > 24 26 4.2(2.4)-10 yandr(e" -• ue-y) > 3.4(2.0) • 10 у at 68% (90%) C.L. Life time limits on the CNC electron capture with nuclear levels excitation of 23Na, 127I, and 129Xe are also established; they are in the range r > 1.5 • 1023—4 • 1024 y. All limits are at least few times higher than the ones previously available......

В подземной лаборатории Гран-Сассо были проведены измерения фона с использованием уста- новок DAMA Nal(Tl) с массой ~ 100 кг (статистика 34866 кг-сут) и 6,5 кг жидкого ксенона (2257,7 кг • сут). Установлены новые экспериментальные пределы на время жизни электрона: 24 26 т(е~ -^ ueVeue) > 4,2(2,4) • 10 лет и т(е~ -^ vei) > 3,4(2,0) • 10 лет на 68% (90%) уровне достоверности. Установлены пределы на время жизни электрона с возбуждением ядерных уровней в 23Na, 1271 и 129Хе: т > 1,5 • 1023—4 • 1024 лет. Все полученные пределы в несколько раз выше установленных ранее.

INTRODUCTION

Since electron is the lightest electrically charged particle, the stability of the electron implies the conservation of electric charge. In the framework of the standard quantum electrodynamics, the charge conservation is a direct consequence (Weinberg theorem [1]) of massless photons which are imposed by the fundamental underlying principle of gauge invariance. Nevertheless, the possibility that the electric charge conservation may be broken in future unified gauge theories and the related implications have been intensively discussed in literature [2-7]. Although no self-consistent theory describing electric charge nonconservation has been yet constructed (see for details reviews [6] and refs. therein), many efforts have been devoted to test this fundamental feature of the nature in direct experiments [8-19] since the early search by Feinberg and Goldhaber in 1959 [8]. New Experimental Limits on the Electron Stability 59

The idea of the pioneering experiment [8] was to use a Nal(Tl) scintillator and to look for the X ray and Auger electrons cascade, which would follow the decay of а К electron in a Iodine atom (energy release is 33.2 keV). This approach — named «disappearance» approach — is sensitive to all the decay modes giving decay particles which escape the

detector without depositing energy (for example: e~ —> veveve). Another approach, sensitive

to the e~ —> i/e7 decay mode, searches for 255.5 keV 7 quantum; in this case electron decays in the surrounding materials will contribute as well as the ones inside the detector. All the results available in literature — for both types of experiments — are summarized in Table 1. The best limits on the mean life of the electron previously established were: in the 24 25 «disappearance» channel те > 1.3 • 10 у [19J and for e~ —> vej decay re > 3.7(2.1) • 10 у at 68% (90%) C.L. [17]'.

Table 1. Experimental limits on the electron fife time at 68 % (90 %) C.L. Detector Volume, cm"3 Limit on Limit on Year [ref]

Te(e~ —> veueve), у Te(e~ -»• Vei), у Nal(Tl) 1287 1.0-1018 1.0-1019 1959 [8] 21 Nal(Tl) 348 2.0 -10 4.0-1022 1965 [9] Ge(Li) 66 5.3-1021 — 1975 [10] 22 Nal(Tl) 1539 2.0 -10 3.5 • 1023 1979 [11] Ge(Li) 130 2.0 -1022 3.0 -1023 1983 [12] Hp-Ge 135 — 1.5(1.1)-1025 1986 [13] Hp-Ge 3-140 2.7(1.7)-1023 — 1991 [14] NaI(Tl) 17 10570 1.2-1023 — 1992 [15] Hp-Ge 591 — 2.4(1.2)-1025 1993 [16] Hp-Ge 48+2•209 4.3(2.6) • 1023 3.7(2.1)-1025 1995 [17] LXe 2000 1.5-1023 2.0(1.0)-1025 1996 [18] Hp-Ge 132 1.3-1024 — 1998 [19] Nal(Tl) 9 • 2643 4.2(2.4) • 1024 — 1999 [20] & this work LXe 2000 — 3.4(2.0) • 1026 2000 [21] & this work

Another method of searching for the disappearance of electrons on atomic shells involving nuclear levels excitation was proposed and realized at first by Holjevic et al. [23]. The idea is to consider the possible influence of the electron decay on the atomic nucleus. The exploited process, analogous to an electron capture, does not change the nucleon charge but leaves the nucleus in an excited state: {A, Z) + e~ —> {A, Z)* + ve. Possible mechanisms of such CNC processes were considered in Refs. 23, 24, where their advantages for the CNC quest, involving the CNC nuclear excitation through both the weak boson and photon mediating

'it has been shown that the decay channel e —» i/e7 may be affected — in the case of explicit violation of the electric charge conservation due to the presence in the Lagrangian of a CNC term — by the catastrophic emission of a huge amount of longitudinal bremsstrahlung photons with very small energies; thus the decay of an electron will not be accompanied by the 255.5 keV 7 rays [6]. On the contrary, the filling of the shell after the electron disappearance would occur before the emission of soft photons and will not be affected by them (see [17] and refs. therein). Therefore the «disappearance» т limit is considered more «safe» and model independent, as credited by the Particle Data Group [22]. ' 60 Belli P. et al. processes, have been pointed out. The CNC electron capture can feed the excited states of the nucleus with energies Eexc up to me(? — Ев (Ев is the binding energy of the electron). In the de-excitation process the nucleus returns to the ground state emitting one or more 7 quanta and conversion electrons which could be observed by a suitable detector. It is supposed that CNC excitations feed preferably the lowest levels with difference in spin between ground and excited states AJ — 0,1 and that К electrons are most probably involved in the process, being the closest to the nucleus. Results of previous experiments to search for the CNC electron capture are summarized in Table 2.

Table 2. Experimental lifetime limits on the CNC electron capture involving nuclear levels excitation of23Na, 127I, and 129Xe

Nucleus, Efficiency Excluded area Lifetime limits г, у •Eexc V of effect S [25,26] & this work [23] [24] 90 % C.L. 90% C.L. 90% C.L. 68 % C.L. 23Na 440.0 keV 0.60 3009 1.5 • 1023

57.6 keV 0.99 3149 2.4 • 1023 2.1 • 1021 5.8 • 1022 202.9 keV 0.89 3418 2.0 • 1023 1.9 • 1021 5.6 • 1022 375.0 keV 0.72 3061 1.8 • 1023 2.4 • 1021 21 418.0 keV 0.61 2974 1.6 • 1023 2.4 • 10 129 Xe 39.6 keV 0.99 18.5 1.1 • 1024 236.1 keV 0.97 5.5 3.7 • 1024 318.2 keV 0.65 6.1 2.2 • 1024 321.7 keV 0.67 5.6 2.5 • 1024 411.5 keV 0.50 4.6 2.3 • 1024

This paper describes the new improved limits on the electron instabilities which were obtained as a «by-product» results of the DAMA data taking with two low-background detectors operating in the Gran Sasso National Laboratory of INFN: DAMA Nal(TI) scintillator with mass near 100 kg [27,28] and 6.5 kg DAMA liquid Xe (LXe) scintillator enriched in I29Xe at 99.5% [18,29], mainly dedicated to the particle Dark Matter direct search. The results of present work were previously published in Refs.20, 21, 25, 26 where more details on measurements and data processing can be found.

1. EXPERIMENTAL SET-UPS AND MEASUREMENTS

1.1. Measurements with Nal(Tl) Detectors. The detailed description of the highly radiopure ~ 100 kg DAMA set-up and its performances are discussed in Ref. 28. Here we briefly recall the main features of this apparatus. The data were collected with nine 9.70 kg Nal(Tl) crystal scintillators enclosed in radiopure Cu housings. Each detector has two 10 cm long tetrasil-B light guides directly coupled to New Experimental Limits on the Electron Stability 61 the opposite sides of the bare crystal. Two photomultipliers (PMT) EMI9265-B53/FL work in coincidence and collect light at single photoelectron threshold, while 2 keV is the software energy threshold [27,28]. The detectors are enclosed in a low radioactive Gu box inside a low radioactive shield made of 10 cm Cu and 15 cm Pb; the last is surrounded by 1.5 mm Cd foils and about 10 cm of polyethylene. A high purity (HP) Nitrogen atmosphere is maintained inside the Cu box by a continuous flux of HP Nitrogen gas from bottles stored underground for a long time. The whole shield is wrapped in Supronyl and maintained also in the HP Nitrogen atmosphere. The installation is subjected to air conditioning (the maximum level of temperature variation is < 0.2 °C) to avoid any influence of the temperature on the light yield of the crystals, on the PMT's spectral sensitivity and gain and on the stability of the electronics. This allows one to keep constant the energy scale, the energy resolution and the energy threshold of the detectors, as verified also by the continuous monitoring of the stability parameters and by the routine energy scale calibrations. A pulse shape analysis was used to reject the residual noise by exploiting the different time structure of the PMT noise (fast pulses with decay time of order of tens ns) and scintillation signals (decay time of order of hundreds ns). In the calibration measurements the typical energy resolution is a/E = 7.5 % at 59.5 keV. The knowledge of the energy scale was assured by periodical calibration with 241Am source and by monitoring the position and resolution of the 210Pb peak (46.5 keV) present at level of few cpd • kg""1 in the measured energy distributions. This peak is mainly due to a surface contamination by environmental Rn occurred during the first period of the crystals storage deep underground. The standard deviation of the position of this peak estimated for all nine detectors during about 180 days — without applying any correction — does not exceed 1.2% [27,28]; thus the effect is negligible considering the continuous calibration monitoring and correction. In conclusion, owing to the mentioned procedures, the energy scale, the energy resolution and the energy threshold of the detectors are well established. 1.2. Measurements with Liquid Xenon Detector. The LXe DAMA set-up (~ 6.5 kg — i. e., ~ 2 1 —• of liquid Xenon scintillator) and its performance has been published in Refs. 29 and only the main features of the detector are described here. The gas used is Kr-free Xenon enriched in 129Xe at 99.5 %. The U/Th contamination of 129Xe does not exceed w 2 ppt at 90 % C.L. The vessel for the LXe is made of OFHC low radioactivity copper (< 100 /zBq-kg"1 for U/Th and < 310 /nBq-kg"1 for K). The scintillation light collection is assured by three EMI PMTs with MgF2 windows, working in coincidence. Their measured quantum efficiency —• for normal incidence — ranges between 18 and 32% at the LXe scintillation wavelength (175 nm) with a flat behaviour around this value. The PMTs collect the scintillation light through three windows (3" in diameter) made of special cultured crystal quartz (total transmission of the LXe ultraviolet scintillation light is « 80%, including the reflection losses). A low radioactivity Cu shield inside the thermo-insulation vacuum cell surrounds the PMTs; then, 2 cm of steel (insulation vessel thickness), 5-10 cm of low radioactivity Cu, 15 cm of low radioactivity Pb, и 1 mm of Cd and « 10 cm of polyethylene are used as outer hard shielding. The environmental Rn near the external insulation vessel of the detector is removed by continuously flushing high purity (HP) Nitrogen gas (from bottles stored underground for a long time) inside a sealed Supronyl envelope, which wraps the whole shield. Each PMT is connected with a low noise preamplifier. For every event the following data are stored: (i) amplitudes of each PMT pulse and (ii) amplitude and shape of the sum pulse 62 Belli P. et al.

(recorded by a Lecroy transient digitizer). The energy dependence of the detector resolution was measured [29] and can be expressed as following: a/E = 0.056 + 1.19/\/Ж where a is in keV. Some other information can be found in Ref. 29.

2. DATA ANALYSIS AND DISCUSSION OF THE RESULTS

2.1. Decay e —> veveve. The idea of the present work is to use the distinguished features of the DAMA Nal(Tl) set-up to look for signals from X ray and Auger electron cascade, which would follow the decay not only of ъ. К electron (energy released 33.2 keV) but also of a L electron (energy release of about 5 keV) in a Iodine atom. Each Iodine atom contains 8 electrons on L shell (two electrons on LI, two on LI, and four on L3 subshell), while only 2 are available on К shell. Thus, the possibility to investigate the energy region corresponding to L-shell electron decays will increase the source strength by a factor 4 with respect to the standard procedure searching for X-electron decay. The study of the L-electron decay is possible here owing to the low energy threshold and the low background rate of the DAMA set-up. The statistics considered in the present analysis is 19511 kg-day (DAMA/NaI-1 & 2 running periods) [27]. The 2-20 keV energy distributions of each detector can be found in Refs. 27. Since the behaviours of these distributions in the energy region of interest here are not very different, the cumulative energy distribution (Fig. 1) can be used for the electron lifetime estimate.

cpd/kg/keV 2.2

3.75 4 4.25 4.5 4.75 5.75 5.5 5.75 £,keV Fig. 1. Cumulative experimental energy distribution measured by all the detectors in the region of interest for the process searched for; the statistics is 19511 kg day. The dotted line represents the result of a fit given by the sum of a linear function (simplified background model suitable for the present purposes) and of the sum of the three gaussians associated to the process searched for; this last contribution requires only one free parameter (see text)

The possible decay of L electrons in Iodine atoms inside the Nal(Tl) detectors would be visible as a peak at the energy of about 5 keV (5.19 keV for LI shell, 4.85 keV for L2 shell and 4.56 keV for LZ shell [30]) with a/E corresponding to the detector energy resolution. The absence of such a peak in the collected data is evident in Fig. 1. Thus, the experimental spectrum can be used to determine the upper limit of the electron life time using the formula: r = (r]Nt)/S, where TJ is the detection efficiency; N is the number of electrons on L shell New Experimental Limits on the Electron Stability 63 of Iodine atoms; t is the measuring time and S is the number of events due to the effect searched for and excluded with given C.L. The cascade of low energy X rays and Auger electrons with the same energy of about 5 keV will be absorbed in a large Nal(Tl) crystal giving an efficiency 77 = 1. Nine 9.70 kg detectors include 3.51 • 1026 Nal molecules, that correspond to 2.81 • 1027 electrons on L shell of Iodine atoms. Thus, the total Nt is equal to 1.72-1027 electrons • y. As the simplest estimate of the excluded number of events 5 we can accept the standard statistical deviation of the total number of events in the 3.5-6.0 keV energy region. The latter is a very sensitive interval which offers a practically symmetric window centered around the centroid of the 3 peaks and including 66% (?7window) of the total area. The value 5 = (5w)/rjWin 3.6(2.2) • 10 у with 68 % (90 %) C.L. Then, with the aim to make the estimation of S more accurate, the experimental energy distribution in the interval 3.5-6.0 keV was fitted by the sum of two functions: the background and the effect being searched for. As simplified background model, suitable for the present purposes, the linear function has been assumed there. The effect has been represented by the sum of three gaussians, centred at 4.56, 4.85, and 5.19 keV respectively, and with energy resolutions scaled here according to: a/E oc l/л/Ё. The amplitudes of the gaussians have been normalized for two electrons on LI, two electrons on L2, and four electrons on L3 shell (requiring, therefore, only one free parameter for the effect amplitude). From the fit the amplitude of the effect was found to be (—0.0029 ± 0.0240) epd-kg"1, giving no statistical evidence for it (x2/d.o.f. was 1.2). Using these values the upper limit on the events number S was calculated according to the Particle Data Group procedure [31]. In fact, from the amplitude of the effect given by the fit, the lower limit 0.02118 (0.03663) cpd-kg"1 at 68% 24 (90%) C.L. can be estimated, giving: S < 413(715) and r(e~ —• veVeve) > 4.2(2.4) • 10 у at 68 % (90 %) C.L. This result is near 3 times higher than the best limit previously established in the experiment with HP-Ge detectors, where the «disappearance» of Ge i^T-shell electrons was studied [19]. The searches for «disappearance» of electrons on the atomic shells are also related with the experimental quest for the violation of the Pauli exclusion principle (PEP). The transition of electrons to fully filled L shell — process usually forbidden by PEP — will result in an energy release equal to the binding energy of electron on L-shell. From an experimental point of view, both processes are undistinguishable in Nal(Tl) detector; thus the established limit on re could be regarded also as a limit on the probability of the PEP violation.

2.2. Decay e~ —*• vel- The idea of the present work is to search for 7 rays (with the energy of « 255 keV) which could accompany the possible decay of any electron in the LXe scintillator (and in its surroundings) by analyzing the energy distribution collected during about 347 days. The exact value of the total energy deposited in the detector depends on the place where the electron decay occurs. If it happened outside the detector, 7 quantum with initial energy E1 = (mec^ — Ев)/2 can hit the detector and release some energy in it (Ев is the binding energy of the electron on the corresponding atomic shell). If electron decays inside the detector, the additional energy release from X rays and Auger electrons following the atomic de-excitation (with the total energy of Ев) should be also taken into account. 64 Belli P. et al.

The experimental spectrum of the LXe scintillator in the energy region 40-500 keV with total statistics of 2257.7 kg • day is shown in Fig. 2, where the absence of the peak searched for around « 255 keV is evident. Therefore these data can be used to set the bound on the probability of the electron decay in the considered channel. Counts/20 keV To estimate the lifetime limit r, we use the formula: r = tJ2i{r]iNi)/S, where щ is the efficiency to detect « 255 keV 7 quanta from the ith medium (Xe detector and Cu vessel; consecutive layers of steel, Cu and Pb shield were found to be negligible); Ni is the number of atomic electrons in the corresponding ith medium; t is the measuring time; and 5 is the number of the effect's events, which can be excluded with the given confidence leyel on the basis of the experimental data. The гц values and the response functions of the LXe scintillator were evaluated through the Monte Carlo simulation with the help of GEANT3.21 package [32]. The 450 Doppler broadening of the measured 7 lines due to the electrons' movement in different atomic shells was also taken into account [16]. Doppler-broadened Fig. 2. Energy spectrum measured line shape is represented by the sum of 17 Gaussians by the LXe scintillator in the en- for the case of Xe atoms and of 10 Gaussians for Cu ergy region 40—500 keV (total statistics: atoms; however, considering the energy resolution 2257.7 kg-day). In the inset the part of of the LXe scintillator, its total response function to the spectrum in the energy region of inter- the effect searched for is very close to a Gaussian est is shown in linear scale together with centred at « 255 keV with FWHM equal to 80 keV. the fitting curve (dashed line) and excluded The calculated values of the detection efficiency are 26 peak (r = 2.0 • 10 y) 77xe = 85 % and т)Си = Ю %. The S value has been determined in two ways. First, it has been evaluated by using the so-called «one a approach» in which the excluded number of the effect's events is estimated simply as the square root of the number of background counts in a suitably chosen energy window AE. Notwithstanding its simplicity, this method gives the right scale of the sensitivity of the experiment. For instance, 34 counts are present within the interval 205-325 keV which contains 90% of the expected peak; thus, the square-root estimate gives S < 6.5 events. Using this S value, the numbers of electrons in the LXe detector and Cu vessel 27 27 (ЛГхе = 1-64 • 10 ; NCu = 4.67 • 10 ), the measuring time and the calculated efficiencies, we obtain the limit r > 2.7 • 1026 y. Furthermore, the S value was determined by using the standard least squares procedure, fitting the experimental energy distribution in the neighborhood of the peak searched for by the sum of a simplified background model (exponent) and of the effect's peak (represented by the simulated response function of the detector as described above). From the fit, the peak's area equal to -2.4 ± 6.3 counts (X2/d.o.f. = 0.67) has been obtained, giving no evidence for the effect. Then, the number of the effect's events, which can be excluded at 90% (68%) C.L. is calculated [22] as 9.0(5.2), giving the limit r > 2.0(3.4) • 1026 у at 90% (68%) C.L. The fitting curve and the excluded peak are shown in the inset of Fig. 2. The present bound on the e~ — New Experimental Limits on the Electron Stability 65 decay channel is one order of magnitude higher than the best limit (2.1 • 1025 у at 90 % C.L.) previously obtained for this decay mode [17]. In accordance with the results of Ref. 33, Counts/iOkeV the transition probability for the electron decay e~ —+ vej can be written in the c 2 2 form: A™ = е е^(а/32тг)(тес /П)[1 + 2 2 2 (me/m7) ] ~ £^7(a/327r)(mec //i)(me/m7) , where a is the fine structure constant, and the

£е„7 gives a measure of the charge nonconservation. From this equation we can find numer- 60 80 100 120 2 25 ically e evi{me/m-,f < 5.6 • 10' /те1/1 (ге„7 is in years), thus our experimental limit leads to 2 2 51 the bound £ „7(тпе/т7) < 2.8(1.6) • ИГ at 90 % (68 %) C.L. The latter expression, combined with the best laboratory limit on the photon mass 16 m7 < 2 • 10~ eV [34] yields the restrictions 94 П.ПП.ПГ 4,7 < 4.3(2.5) • 10- at 90% (68%) C.L. 300 400 From the other hand, we can use the established те„7 limit to find the bound on the photon mass. The relation between ге„7 and Fig. 3. Energy spectrum measured by m7 was found, for example in the framework the LXe scintillator in the energy re- of the SU(5) model [35] as following: rei/7 a gion 40-500 keV with total statistics of 25 2 10~ (mz/m7) y, where mz = 91.2 GeV is 823.1 kg-day. In the inset the low energy the mass of the Z bozon. Using this relation and part is shown in linear scale together with 26 our value те1П > 3.4 • 10 y, we can receive the fitting curve and excluded peak (r = 15 24 г 29 m7 < 1.6-10" eV. 1.1 • 10 y) for the first excited level of Xe 2.3. Nuclear Levels Excitation Due to the Electron Decay. The data accumulated with the DAM A Nal(Tl) and LXe detectors were used also to establish the limits on the probability of the process in which an electron disappears from the atomic shell and the nucleus is left in an excited state. Such process is analogous to the usual electron capture but does not change the nucleus' charge:

The CNC electron capture can feed the excited states of the nucleus with energies Eexc 2 up to mec — Ев (Ев is the binding energy of the electron in the considered atomic shell). It is supposed that CNC excitation feeds preferably the lowest levels with difference in spin between ground and excited state Д J = 0,1, and that К electrons (being the closest to the nucleus) most probably are involved in the process. In the de-excitation the nucleus returns to the ground state emitting 7 quanta and conversion electrons. X rays and Auger electrons emitted in the relaxation of the atomic shell should be also taken into account. The results of the previous experiments [23,24], performed with Nal(Tl) detectors, are given in Table 2. In this paper the first investigation of the CNC electron capture involving nuclear levels excitation of 129Xe is presented; the statistics considered here is 823.1 kg • day [26]. 129 Five levels of Xe could be excited due to the studied process with Eexc = 39.6; 236.1; 318.2; 321.7 and 411.5 keV [30]. Taking into account the binding energy of the Xe К atomic shell (Eg = 34.6 keV), the energies of the possible peaks in the background spectra 66 Belli P. et al. should be: 74.2; 236.1'; 352.8, 356.3, and 446.1 keV. The experimental spectrum of the LXe scintillator in the energy region 40—500 keV is shown in Fig. 3, where the absence of these peaks is evident. Thus limits can be set for the probabilities of CNC nuclear excitations of 129Xe nuclei. We estimate the life time limits r using the formula r = (rjNt)/S, where 77 is the detection efficiency, N is the number of electrons on К shell of Xe atoms; t is the measuring time, and S is the number of the effect's events, which can be excluded with a given C.L. To calculate the 77 values, de-excitation processes in 129Xe nuclei inside the LXe scintillator and the response function of the detector were simulated with the help of GEANT3.21 package [32]; the code DECAY4 [36] was used for description of the initial event's kinematics. Calculated efficiencies are varied from 0.99 for Eexc — 39.6 keV to 0.50 for Eexc = 411.5 keV (Table 2). The S values were determined in two ways. Firstly, by using the so-called «one a approach», in which the excluded number of the effect's events is estimated simply as square root of the number of background counts in a suitably chosen energy window AE. For instance, in the measured spectrum within the energy interval 45- 103 keV (it contains 95 % of expected 74.2 keV peak area) there are 129 counts; thus, the square root estimate gives S < 11.4 events. Using this value S, total number of К electrons in the LXe detector (N = 6.0 • 1025), measuring time and calculated efficiency, we obtain the limit т > 1.7-1024 у (68 % C.L.) for the 74.2 keV peak. The results for other peaks are within r > (3-7) • 1024 у at 68% C.L. Further, S values were determined by using the standard least squares procedure, where the experimental energy distribution in the neighborhood of the peak searched for was fitted by the sum of background (exponential behaviour for the first peak and a straight line for the others) and signals peak being sought. As the last one the response function of the detector was simulated by a gaussian with the proper width. For example, the obtained area for the first peak (74.2 keV) is —11 ± 15 counts (x2/d.o.f. value is 1.3), thus giving no evidence for the signal. Then, the number of the signal events, which can be excluded with 90% (68%) C.L. were calculated [22] as 18.5 (10.4). It gives the limit 24 r > 1.1(2.0) • 10 у at 90% (68%) C.L. for the first (Eexc = 39.6 keV) excited level of 129Xe. The excluded number of signal events for other levels obtained by a similar procedure and the corresponding т limits are shown in Table 2. For illustration the fitting curve and excluded peak for the first excited level is depicted in the inset of Fig. 3.

The data accumulated with the DAMA Nal(Tl) detectors were also analyzed to set the limits on the CNC nuclear excitations in 23Na and 127I nuclei; the statistics used was 34866 kg • day [25]. The obtained results are summarized in Table 2. Other details can be found in Ref. 25. Possible mechanisms of the CNC electron capture were discussed in [23-26]. Such process could include both the weak boson and photon exchange. Comparing the CNC electron capture with the standard electron capture and with the standard internal conversion process and using the obtained lifetime limits, the bounds on the CNC admixtures in the weak 2 26 2 42 interactions were found: e w < 2.2 • 10~ and e < 1.3 • 10" at 90% C.L. The detailed discussion is given in Refs. 25, 26.

'Because the second excited level is long-lived (

Eexc = 236.1 keV. New Experimental Limits on the Electron Stability 67

CONCLUSION

Using the the low energy threshold and the low background rate of the ss 100 kg DAMA Nal(Tl) set-up, the electron stability has been studied by looking for the signal from X ray and Auger electron cascade which would follow the decay («disappearance») of any L electron of 24 one of the Iodine atoms. The obtained lifetime limit r(e~ —» veVeve) > 4.2(2.4) • 10 у at 68 % (90 %) C.L. is few times higher than the best limit previously achieved by considering the If-shell electrons in Ge detectors.

The highest limit on the lifetime of the electron for the decay e~ —+ i/e7 was established with the super-low background DAMA set-up with the liquid Xe scintillator: r(e~ —» ue^) > 3.4(2.0) • 1026 у at 68 % (90 %) C.L. Lifetime limits on the CNC electron capture with nuclear levels excitation of 23Na, 127I, and 129Xe were obtained; they are in the range т > 1.5 • 1023—4 • 1024 y. All limits are at least few times higher than the ones previously available. The severe restrictions on the CNC admixtures in the weak interactions and on the photon 26 2 42 26 2 94 mass were derived: e^ < 2.2 •ПГ , £ < 1.3-lO" , feiv < 1.1-10- , е „7 < 2.5-ИГ , 15 and m7 < 1.6-ИГ eV [20,21,25,26].

REFERENCES . .

1. Weinberg S.I/ Phys. Rev. B. 1964. V. 135. P. 1049. . 2. Okun L В., Zeldovich Ya. B. II Phys. Lett. B. 1978. V. 78. P. 597. 3. Voloshin M. В., Okun L. B. II JETP Lett. 1978. V. 32. P. 145. 4. Mohapatra R.N. //Phys. Rev. Lett. 1987. V.59. P. 1510. 5. Okun L.B. Leptons and Quarks. Amsterdam, 1982. P. 181. 6. Okun L.B. II Sov. Phys. Usp. 1989. V.32. P.543; Comments Nucl. Part. Phys. 1989. V. 19. P.99; Phys. Rev. D. 1992. V. 45, No. 11. P. VI. 10. ; 7. Rubakov V. A. II Part, and Nuclei, Letters. 2001. No. 3[106]. P. 106. 8. Feinberg G., Goldhaber M. II Proc. Nat. Acad. Sci, USA. 1959. V. 45. P. 1301. 9. Мое M. K., Reines F. II Phys. Rev. B. 1965. V. 140. P. 992.

10. Steinberg R. I. et al. II Phys. Rev. D. 1975. V. 12. P. 2582. 11. Koval'chuk E. L. et al. II JETP Lett. 1979. V. 29. P. 145. 12. Bellotti E. et al. II Phys. Lett. B. 1983. V. 142. P. 435. 13. Avignone F. T. Ill et al. II Phys. Rev. D. 1986. V. 34. P. 97. 14. Reusser D. et al. II Phys. Lett. B. 1991. V.255. P. 143.

15. Ejiri H. et al. II Phys. Lett. B. 1992. V.282. P. 281.

16. Balysh A. et al. II Phys. Lett. B. 1993. V. 298. P. 278. 68 Belli P. et al.

17. Aharonov Y. et al. II Phys. Lett. B. 1995. V.353. P. 168; Phys. Rev. D. 1995. V.52. P. 3785.

18. Belli P. et al. II Astropart. Phys. 1996. V. 5. P. 217.

19. Klimenko A. A. et al. II Yad. Fiz. 1998. V. 61. P. 1229 (in Russian).

20. Belli P. et al. II Phys. Lett. B. 1999. V. 460. P. 236.

21. Belli P. et al. II Phys. Rev. D. 2000. V. 61. P. 117301.

22. Particle Data Group. Review of Particle Physics // Phys. Rev. D. 1996. V.54. P. 1.

23. Holjevk S. et al. II Phys. Rev. С 1987. V.35. P. 341.

24. Ejiri H. et al. II Phys. Rev. С 1991. V.44. P. 502.

25. Belli P. etal. //Phys. Rev. С 1999. V.60. P.065501.

26. Belli P. etal. //Phys. Lett. B. 1999. V.465. P.315.

27. Bernabei R. etal. //Phys. Lett. B. 1996. V.389. P.757; 1997. V.408. P.439; 1998. V.424. P. 195; 1999. V.450. P. 448; 2000. V.480. P. 23; Nuovo Cim. A. 1999. V. 112. P. 1541; Belli P. et al. II Phys. Rev. D. 1999. V.61. P. 023512. •

28. Bernabei R. et al. II Nuovo Cim. A. 1999. V. 112. P. 545.

29. Belli P. etal. //Nuovo Cim. С 1996. V. 19. P.537; Phys. Lett. B. 1996. V.387. P.222; Bernabei R. et al. II Phys. Lett. B. 1998. V.436. P. 379.

30. Firestone R. B. Table of Isotopes. 8th ed. N.Y.: John Wiley & Sons, 1996.

31. Feldman G.J., Cousins R. D. II Phys. Rev. D. 1998. V.57. P.3873.

32. GEANT. CERN Program Library Long Write-up W5013. CERN, 1994.

33. Suzuki M. II Phys. Rev. D. 1988. V.38. P. 1544.

34. Lakes R. II Phys. Rev. Lett. 1998. V. 80. P. 1826.

35. Huang J. С II J. Phys. G. 1987. V. 13. P. 273.

36. Ponkratenko O.A. et al. II Phys. At. Nucl. 2000. V. 63. P. 1282. XJ0200058

Письма в ЭЧАЯ. 2001. №6[109] Particles and Nuclei, Letters. 2001. No.6[109]

УДК 539.1.074.8

ELEMENT-LOADED ORGANIC SCINTILLATORS FOR NEUTRON AND NEUTRINO PHYSICS V. B. Brudanirf, V. I. Bregadze0, N. A. Gundorirf, D. V. Filossofova, O. I. Kochetova, I. B. Nemtchenokf1, A. A. Smolnikovah "Joint Institute for Nuclear Research, Dubna ^Institute for Nuclear Research RAS, Moscow cInstitute of Organoelement Compounds RAS, Moscow

New types of element-loaded (B and Gd) organic scintillators for neutron detection and neutrino experiments have been synthesized recently at JINR. Their optical, spectral, scintillation and radiopurity characteristics are presented and discussed. It is shown that the 5 % B-loaded scintillator has a light output as much as 70 % relative to the unloaded one. The same characteristics for the 3 -X Gd-loaded sample is equal to 51 %. Transparency and other properties of the produced scintillators did not change at normal conditions for at least one year. В работе сообщается о новых типах алементосодержащих (В и Gd) органических сцинтил- ляторов, полученных в ОИЯИ. Обсуждаются их оптические, спектральные и сцинтилляционные свойства. Показано, что световыход пластмассового сцинтиллятора, содержащего 5 % бора, со- ставляет 70 % от световыхода аналогичного по геометрии сцинтиллятора, не загруженного бором. Аналогичная характеристика для пластмассового сцинтиллятора, содержащего 3 % гадолиния, со- ставляет 51%. Свойства новых сцинтилляторов стабильны в обычных условиях хранения как минимум в течение одного года.

INTRODUCTION

Element-loaded organic scintillators (plastic and liquid) offer some potential advantages as detectors for neutrons and neutrinos. This has led to widespread use of developed materials both in solar and atmospheric neutrino, neutrino oscillations, neutrinoless double beta decay future experiments and for precise neutron spectroscopy with high level of sensitivity as well as for other approaches both in basic and applied physics. Modern neutrino physics became a large-scale operation, requiring large detectors to intercept the dispersed particles. New-look large (kilotons) neutrino detectors could be mounted using state-of-the art detector technologies and must be placed in facilities with very low level of inner neutron and gamma- ray background. Wide application of element-loaded organic scintillators in running and future nonaccelerator experiments is connected with relatively high cross sections of specially loaded isotopes for neutrino and neutron reactions. Such types of scintillators also offer flexible configuration with potential to discriminate between signals from gamma-ray and neutron or neutrino events. Some possible applications of element-loaded organic scintillators are presented in Table 1. Recently new types of element-loaded (B and Gd) organic scintillators for neutron detection and neutrino oscillation experiments have been synthesized at JINR. Their optical, 70 Brudanin V. В. et al.

Table 1. Applications of element-loaded organic scintillators

Elements or isotopes Applications 6Li, 10B, 113Cd, 155Gd, 157Gd Neutron detectors, searching for neutrinos oscillations

i76Ybj i60Gd; iooMo> 37C1 Detection of solar neutrinos Pb Detection of astrophysics neutrinos 19F, 73Ge Searching for Dark matter 150Nd, 160Gd, W0Mo, 130Te, 82Se Searching for double /3 decay Pb, Sn, W, Hg, Bi High energy physics

spectral, scintillation and radiopurity characteristics are presented and discussed here. Trans- parency and other properties of the produced scintillators did not changed at normal conditions for at least one year. Such stability in time is very important for long-term experiments. High efficiency for thermal neutron registration achieved for produced samples makes it possible to use such scintillators as a part of neutron high-sensitive spectrometers. The possibility to built large mass cost-effective detectors on the base of other element-loaded (Li, F, Pb, W, Mo, Nd, Yb) organic scintillators with high radiopurity is discussed as possible application for future nonaccelerator experiments. Application of such materials can also be useful for neutron flux monitoring in nuclear power stations, nuclear reactors, storage life of long lived radioactive trash, and for measuring of the environment radioactivity inside civilian buildings.

1. B-LOADED PLASTIC SCINTILLATORS Low energy neutrons (< 100 keV) cannot be detected by proton recoil in usual (unloaded) plastic scintillator since they do not generate enough light. Scintillators are therefore loaded with various additives (for example, В and Gd, see Table 2) so that the neutrons can be detected by exoergic capture reactions.

Table 2. Properties of В and Gd as targets for thermal neutrons detection and searching for neutrinos oscillations Isotope Natural abundance, % Reaction Cross section,b Signatures 10g 19.6 (n,a) 3.8 • 103 a(1.74 MeV) + 7Li(0.84 MeV) 155 Gd 14.7 (n,-y) 6.1 • 10" 7 burst: up to 8 MeV 157 5 Gd 15.7 (n,l) 2.6 • 10 7 burst: up to 8 MeV

Up to now widely used organic scintillators for thermal neutron detection were boron- loaded materials. It is connected with high (n, a) cross section of 10B as well as relatively high natural abundance of it (19.6%). Thermal neutrons may subsequently be captured in boron-loaded scintillator by 10B, which has a thermal cross section of 3838 b. The Q-value of 10B(n,o;)7Li reaction is 2.79 MeV with 2.31 MeV going to the charged particles (1.48 MeV for a) along with a 480 keV 7 ray (94% of the time): Element-Loaded Organic Scintillators 71

n + 10B —> 7Li* + a + 2.31 MeV

—>7Li*+7 (480 keV). Several boron compounds have been considered for use in plastic scintillators [1-8]. As a result of our investigation o-carborane (Fig. 1) has been selected as one of the most perspective chemical compounds with mass fraction of boron as much as 75 %. First samples of polystyrene-based plastic scintillators containing 0.38 %, 0.75 %, 2 % and 5 % mass fraction of boron have been pro- Fig. 1. Chemical struc- duced by using a method of thermal polymerisation. p-Terphenyl and ture of o-carborane: POPOP have been used as activator and waveshifter. • — carbon atoms; The results of investigations of the optical, spectral, scintillation, о — boron atoms and neutron capture characteristics showed good perspectives of such scintillator. It was shown that 5 % B-loaded scintillator has a light output as much as 70 % relative to the unloaded one and is very effective for thermal neutron detection. Some characteristics of the obtained B-loaded scintillator samples (30 mm in diameter, 9 mm thickness) are presented in Table 3 and in Fig. 2.

Table 3. Properties of B-loaded plastic scintillators Characteristics 1Vlass fraction of boron, % 0 0.38 0.75 2.00 5.00 Density, g/cm3 1.05 1.05 1.05 1.05 1.05 Index of refraction 1.575 1.575 1.575 1.575 1.540 Number of В atoms per cm3, 1022 0 0.02 0.04 0.12 0.28 Number of H atoms per cm3, 1022 4.85 4.85 4.85 4.86 4.86 Number of С atoms per cm3, 1022 4.85 4.83 4.82 4.75 4.58 Amax of luminescence, nm 421 421 421 422 423 Transmission (A^S), % 86.8 86.3 86.9 85.7 84.8 Light output, % 100 97 88 78 70 Thermal neutrons (E < 0.5 eV) detection efficiency, % 0 3 5 13 22

Influence of the boron concentration in scintillators on their light output has been carefully investigated with using a set of a, /3, j sources (see Table 4). Using of different types of radiation makes it possible to study surface and inner uniformi- ties of the scintillation medium as well as energy resolution and a//3 quenching factor (relative energy conversion coefficient). It was obtained that a/(3 value is equal to (0.082 ±0.008) for the energy Ea = 5 MeV. Light output of the boron-loaded scintillator samples as a function of В mass fraction is presented in Fig. 3. To investigate thermal neutron detection efficiencies of the B-loaded samples both Pu- Be neutron source with moderators and time-off-flight technique with using the Pulse Fast Neutron Reactor (IBR-30, JINR) have been applied. Scheme of measurements with the Pu-Be source set-up is given in Fig. 4. The spectra accumulated with Pu-Be source for all B-loaded samples are presented in Fig. 5. 72 Brudanin V. В. et al.

Relative intensity 1.00

: 5% в

oo 400 450 500 550 600 300 350 400 450 500 550 600 Wavelength, nm Wavelength, ran Fig. 2. Photoluminescence (a) and transmission (b) (relative to air) spectra of the B-loaded plastic scintillators samples

Table 4. Calibration sources used for light output measurements

Type of source Isotope Energy, MeV a 148Gd+241Am 3.18 + 5.49 a 244Cm 5.80

207Bj P ek = 0.976 137 0 Cs ek = 0.625 137 1 Cs 0.662 60 7 Co 1.17+1.33

100 Pu-Be neutron Plastic source scintillator sample CH(B) Pb

0 1 2.3 4 5 Mass fraction of B,%

Fig. 3. Light output of the boron-loaded scintillator samples vs В mass fraction Fig. 4. Scheme of measurements with the Pu-Be source set-up

One can see that the net area under the neutron induced a peak depends on boron's contamination. To estimate thermal neutron detection efficiencies (Table 3 and Fig. 6) the net a peak areas obtained after subtraction of the baseline (unloaded scintillator) were used. As it followed from data presented in Table 3, the efficiency for thermal neutron detection is achieved as high as 22 % just for the small size control sample with 5 % of B. Element-Loaded Organic Scintillators 73

5%B

12 3 4 5 6 Number of 10B atoms per cm3, 1020 Fig. 5. Spectra of a particles induced by thermal neutron capture in 10B Fig. 6. Dependence of thermal neutron detection efficiency upon the 10B concentration (for the samples 30 mm in diameter, 9 mm thickness). Detection efficiency of the unloaded sample is assumed to be equal to zero

Owing to this fact, a prototype of thermal neutron detector (70 mm in diameter, 70 mm thickness) on a base of the 5 % B-loaded plastic scintillator has been produced recently. This pilot detector has very suitable scintillation characteristics. Namely, it has 70 % light output relative to a control unloaded scintillator, good uniformity of light collection, and 12% of energy resolution (at 1 MeV) in comparison with 11 % for the unloaded one. Investigation of the thermal neutron capture efficiency is in progress now by using both neutron sources and neutron beam. Our special interest was to measure radioactive contaminations in the produced scintillators aimed to test their ability to use in low background nonaccelerator experiments. For such purposes a radiopurity of the pilot B-loaded detector has been investigated in the underground low background laboratory of the Baskan Neutrino Observatory (BNO INR RAS). As a result only limits on the 232Th, 238U, and 40K contamination at the level of < 10~9 g/g were obtained that gives a good chance for using such scintillators in very low background set-ups.

2. Gd-LOADED PLASTIC SCINTILLATORS

Synthesis of gadolinium-loaded plastic scintillators is considered as one of other perspective directions of such investigations. Scintillators contained gadolinium are very interested for thermal neutron detection too, because this element has two isotopes with high enough natural isotope abundance: 155Gd (14.7%) and 157Gd (15.7%) as well as very high thermal neutron capture cross-sections 155Gd (6.1 • 104 b) and 157Gd (2.6 • 105 b). Scintillators with more than 0.5% Gd have not yet been produced [9-11]. It has been shown that the solubility of the complex of gadolinium nitrate with hexamethylphosphotri- amide in methylmetacrylate can be achieved as much as 30 %. This compound has been used in present work as Gd-contained additive. Total chemical composition used for Gd- scintillator synthesis on the base of polymethylmethacrylate is the following: naphthalene, 2.5-diphenyloxazole (PPO) and POPOP. The obtained results showed us a high probability of successful synthesis of gadolinium-loaded plastic scintillators. 74 Brudanin V. В. et al.

Table 5. Properties of Gd-loaded plastic scintillators

Characteristics Mass fraction of Gd, % 0 1 2 3 Density, g/cm3 1.172 1.182 1.195 1.204 Index of refraction 1.480 1.475 1.466 1.457 Number of Gd atoms per cm3, 1022 0 0.0047 0.0095 0.0143 Number of H atoms per cm3, 1022 5.41 5.39 5.38 5.35 Number of С atoms per cm3, 1022 3.82 3.73 3.64 3.53 Amax of luminescence, nm 424 424 425 425 Transmission (Л„^), % 82.8 80.7 78.2 75.2 Light output, % 100 79 62 51 Thermal neutrons (E < 0.5 eV) detection efficiency, % 0 9 10.5 12.5

Relative intensity 1.00

1 % Gd 2 % Gd

0.00 350 400 450 500 550 600 400 450 500 550 600 Wavelength, nm Wavelength, nm Fig. 7. Photoluminescence (a) and transmission (b) (relative to air) spectra of the Gd-loaded plastic scintillators samples

Counts 200 It 150 з%Gd 2% Gd 100 % J 1% Gd 50 3 „- Unloaded 0 0.0 0.5 1.0 1.5 2.0 2.5 80 100 120 140 160 180 200 Mass fraction of Gd, %

Fig. 8. Light output of the Gd-loaded plastic scintillator samples vs Gd mass fraction Fig. 9. Spectra of gammas induced by thermal neutron capture 155Gd and 157Gd (for the samples 30 mm in diameter, 10 mm thickness) Element-Loaded Organic Scintillators 75

Samples of plastic scintillators containing 1 %, 2 % and 3 % of gadolinium have been produced recently. It was shown that 3 % Gd-loaded scintillator has a light output as much as 51 % relative to the unloaded one. The characteristics of the produced Gd-loaded scintillator samples (30 mm in diameter, 10 mm thickness) are presented in Table 5 and in Figs. 7, 8, 9, and 10. To measure light output and thermal neutron detection efficiency the same methods as for the B-loaded samples were used. The difference is that for the case of the Gd-loaded scin- 0.0 0.1 0.2 0.3 0.4 0.5 Total numver of 155Gd and 157Gd atoms tillators a total efficiency depends in higher degree per cm3, 1020 on a sample size due to smaller gamma detection efficiency in comparison with an alpha one. Fig. 10. Dependence of total (thermal neutron and gamma) detection efficiency on the 3. Gd-LOADED LIQUID SCINTILLATORS 155 Gd and 157Gd concentration for the Gd- WITH HIGH FLASH POINT loaded samples (for the samples 30 mm in diameter, 10 mm thickness). Detection ef- Liquid scintillators containing Gd are also ficiency of the unloaded sample is assumed needed for several planned solar neutrino and neu- to be equal to zero trino oscillation experiments. One of the main requirements for such scintillators is their high value of a flash point. So, our next attempt was to find a set of multicomponent organic solvents, which can satisfy above-mentioned require- ment. As a result the composition of gadolinium nitrate with a methylnaphthalene (flash point 82 °C), tributylphosphate (flash'point 193 °C), and 2-(-biphenylyl)-5-phenyloxazole (BPO) has been chosen for such a purpose. The samples of liquid scintillators with concentration of gadolinium 17.3, 51.3, 61.8, and 85:9 mg • ml"1 have been produced and tested. The characteristics of the produced Gd-loaded liquid scintillator samples are presented in Table 6 and in Figs. 11, 12.

Table 6. Properties of Gd-loaded liquid scintillators Characteristics (Concentration of Gc , mg • ml l 0 17.3 51.3 61.8 85.9 Boiling point, °C 245 245 245 245 245 Flash point, °C > 82 > 82 > 82 > 82 > 82 Density, g/cm3 0.99 1.02 1.07 1.09 1.12 Index of refraction 1.500 1.503 1.508 1.515 1.520 Number of Gd atoms per cm3, 1022 0 0.0066 0.0196 0.0263 0.0328 Amax of luminescence, nm 395 397 398 398 400 Transmission* (A|X)> % 47.5 54 54.5 53.6 55.8 Light output, % 100 67 41 37 26

'Quartz cell, length — 5 cm, relative to air. 76 Brudanin V. В. et al.

Relative intensity o/ t.00 Unloaded .— b 85•9mg •ml"1 Gd

0.00 i 1 i 300 350 400 450 500 550 600 400 450 500 550 600 Wavelength, nm i Wavelength, nm

Fig. 11. Photoluminescence (a) and visible range transmission (£>) spectra (quartz cell, length — 5 cm, relative to air) of the liquid Gd-loaded scintillators

CONCLUSION

Recent achievements of chemistry and technology of "organoelement and polymeric compounds and complexes have been involved to produce new types of boron- and gadolinium-loaded organic scintillators. The developed methods allowed us to load in plastic scintillators as much as 5 % В and 3 % Gd practically without losses of their quality. Transparency and other properties of the produced scintillators did not change 20 40 60 80 at normal conditions for at least one year. Concentration of Gd, mg ml-1 The Gd-loaded liquid scintillator with relatively high (> 82 °C) flash point and contami- Fig. 12. Light output of the Gd-loaded liquid nation of Gd up to 85 mg-ml"1 has been pro- scintillators vs concentration of Gd duced, too. High efficiency for thermal neutron registration achieved for produced samples makes it possible to use such scintillators as a part of high sensitive neutron spectrometers. It is shown that the level of measured radiopurity of the produced scintillators is low enough for using in large mass cost-effective detectors for future nonaccelerator experiments.

Acknowledgements. We would like to thank A. I. Akatov, V. V. Fariseeva, R. S. Galperina and N. M. Nevzorova for their participation in this work and technical assistance. We are very grateful to L. B.Bezrukov, N. Giokaris and A. N. Sissakian for their interest to this work and usefull discussion.

REFERENCES

1. Bellinger L.M., Thomas G.E. II Rev. Sci. Instr. 1957. V.28, No. 7. P. 489.

2. Baroni E. E. et al. II Nucl. Electronics. 1962. V. 1. P. 131 (in Russian). Element-Loaded Organic Scintillators 77

3. Pikelner L.B. et al. II РТЕ. 1963. No. 2. P. 48 (in Russian).

4. Popov А. В., Yazvickiy Yu. S. II РТЕ. 1963. No. 4. P. 70 (in Russian).

5. Anisimova G.I. et al. II РТЕ. 1969. No. 1. P.49 (in Russian).

6. Greenwood L. R., Cheleew N.R. II Rev. Sci. Instr. 1979. V.50, No.4. R466.

7. Drake D. M., Feldman W. C, Hurlbut С II Nucl. Instr. Meth. 1986. V. 247. P. 576.

8. Wang S. С et al. II Nucl. Instr. Meth. A. 1999. V.432. P. 111.

9. CzirrJ. B. Gd-Loaded Plastic Scintillator// Nutl. Instr. Meth. 1973. V. 108. P. 613.

10. Aleshin V. I. et al. II РТЕ. 1977. No. 4. P. 68 (in Russian).

11. Gabriel T.A., Lillie R.A., Childs R.L. II Nucl. Instr. Meth. A. 1984. V.221. P. 474. Письма в ЭЧАЯ. 2001. №6[109] Particles and Nuclei, Letters. 2001. No.6[109]

A SUMMARY OF NANPINO 2000

I have been asked by the organizers .to give a short conference summary. I happily accepted, but I now realize that this task is almost impossible to accomplish on such a short notice. Furthermore, in the short time left.till the formal closing of the conference it is impossible to be fair to do all the important contributions to this conference. As you can imagine, my task becomes even harder since I did not keep any notes. You see that m this electronic era the organizers.spoiled us with their innovation to put all the transparencies in the web. So inevitably I will be very brief, restricting myself to some general remarks. I will begin by thanking the members of the local committee for their heroic efforts in making this conference such a success. We are all aware of all the hurdships they had to overcome. I like to stress once more that the subject of this workshop was very well chosen, since it is in tune with the current trends in particle and nuclear physics. In spite of the emphasis of the workshop on nonaccelerator new physics (NANP), it has become clear during this conference that experiments in the NANP area in fact nicely complement the accelerator experiments. The Super-Kamiokande data, which have been interpreted as a definite signal of neutrino oscillations, and the baryon asymmetry in the Universe are two indications of physics beyond the standard model. It is clear that these offer encouragement to all experiments searching for the ultimate theory of nature. In any case the famous philosopher Plato recognized 2500 years ago that the pursuit of ideas cannot always take place in the sunshine. Occasionally one may have to retreat to the basement to accomplish this task. Much more often, of course, one may have to go to cellar to make ideas work. Now I will focus on two new interesting theoretical physics propositions. Prof. Nath discussed the phenomenological implications of modern formulations of high-energy physics in extra dimensions, which yield an effective Planck mass in the TeV region. He showed that in the framework of such theories one can naturally obtain a light neutrino Dirac mass matrix. This way one can get light sterile neutrinos, which will provide a neat interpretation of all three existing neutrino oscillation experiments (atmospheric, solar and NLSD). This Dirac mass of course, does not allow neutrinoless double beta decay. When combined, however, with flavor violating interactions it may lead to pseudo-Dirac neutrinos with perhaps detectable neutrinoless double beta decay rates. Prof. Rubakoff proposed a scenario, which can lead to appearance or disappearance of electrons in our Universe. Admittedly his presentation was qualitative. Neither the resulting rates have reliably been computed nor a characteristic experimental signature has been provided. On the experimental side, we heard a number of interesting presentations manifesting the heroic efforts of the experimentalists to reduce the backgrounds plaguing these very interesting exotic, but low counting rate, processes. It is interesting to witness that the experience gained in one such experimental endeavor — neutrino oscillations, neutrinoless double beta decay and dark matter search — provides a clear insight to the others. We also heard about neutrino A summary of NANPino 2000 79 factories. I believe that, if such factories become feasible, there will be even stronger links between accelerator and nonaccelerator physics. Anyway neutrinoless double beta decay and dark matter searches will still be bona fide NANP. Finally, I cannot avoid thinking that this workshop has given us the opportunity to think about the plight of the Russian and Former Soviet Union scientists. It is important to help these scientists sustain the high scientific level they have achieved in the past. In other words to help them survive till the sun rises again. We can do it in many ways. First by giving these scientists, especially the ones who are at the beginning of their careers, the opportunity to work in the west. We should make sure, of course, that this does not evolve into a mere brain drain. Second by supporting the megascience experiments taking place on Russian soil (Dubna, Baksan, Baikal). Third by submitting joint research proposals to funding agencies in the west. For this to be effective we should increase the available funds for such projects. We should also make sure that the bulk of the money reaches the scientists themselves. I have been the Greek delegate to the INCO committee of the European Union, which deals with grants to non-member states. I know from experience that the political and business pressure from both sides is to allow procedures, which permit the milking the cow at various administrative and bureaucratic levels. But in such schemes, by the time the cow gets to the basement of the researcher, very little milk is left. We should also avoid listening to the sirens calling for setting up superficially glorious projects like environmental «research», information society, etc., even when they have no high quality research component. The presumed purpose is to help these countries solve immediate problems. I like to advocate the view that, if the Russian scientists survive during the present period of their seven lean cows, they can later solve their problems themselves. In the above spirit I am looking forward to the next workshop, which should not be named NANPino but be given a more appropriate name, like NANPSuper. Before closing let us all again thank the organizers for giving us the opportunity of being here in this laboratory today.

J. D. Vergados

Theoretical Physics Division,

University of Ioannina,

GR-45110, Greece ПРАВИЛА ОФОРМЛЕНИЯ АВТОРСКОГО ОРИГИНАЛА СТАТЬИ В ЖУРНАЛ «ПИСЬМА В ЭЧАЯ»

1. Оригинал статьи предоставляется автором в двух экземплярах. Необходимо предо- ставить дискету с текстовым файлом в формате LaTex-2e (не следует вводить свои макро- команды) и файлами рисунков, таблиц, подрисуночных подписей. Оригинал должен включать все необходимые элементы статьи, иметь сквозную нумерацию страниц и быть подписан всеми авторами. 2. На первой странице статьи указывается индекс УДК, название статьи, инициалы и фамилии авторов на двух языках (русском и английском), место работы. Затем следует аннотация на русском и английском языках, включающая характери- стику основной темы, цели работы и ее результаты. В аннотации указывают, что нового несет в себе данная статья. Аннотация должна быть краткой, 5-8 строк. 3. В формулах все буквы латинского алфавита, обозначающие физические величины, набирают светлым курсивным шрифтом (Е, V,mn др.). Векторы следует набирать прямым полужирным шрифтом, без стрелок сверху. Шрифтом прямого светлого начертания набирают следующие обозначения: чисел по- добия (Ar, Re и др.); функций (sin, arcsin, sh и др.); условных математических сокращений (max, min, opt, const, idem, lim, lg, In, log, det, exp) и др. Латинские буквы в индексах набирают строчным курсивом, кроме сокращений, в

том числе и от фамилий (их набирают шрифтом прямого начертания): fabc, С,^ <3уик, ^efr (l> P)> xmx • Буквы русского алфавита в индексах используют, когда отсутствуют стан- дартизованные международные индексы, и набирают строчным шрифтом прямого начер-

тания: С/ф — фазное напряжение; Рв — мощность возбуждения. Символы химических элементов набирают шрифтом прямого светлого начертания: Cl, Fe. 4. Рисунки должны быть четкими и качественными. Желательно, чтобы файл рисунка был подготовлен в формате .eps (Encapsulated PostScript). He рекомендуется предостав- лять цветные рисунки. Рисунки должны быть одного масштаба, их максимальная ширина 13,5 см, максимальная высота 19,5 см (с учетом подписи). Размеры небольших рисунков не должны превышать 7 см по ширине. Не следует загромождать рисунок ненужными деталями: надписи выносятся в под- пись, а на рисунке заменяются арабскими цифрами или буквами. Если рисунок предостав- лен на отдельном листе, то необходимо указать фамилии авторов, название статьи и номер рисунка. 5. Библиографические ссылки приводят в конце статьи в порядке их упоминания в тек- сте под рубрикой «Список литературы». Ссылки на неопубликованные работы не допус- каются. Ниже следуют примеры оформления: • книги: Кокорева Л.В. Проектирование банков данных. М.: Наука, 1998. 241 с. • статьи из сборника: БыстрицкийВ.М. и др. Исследование температурной зависимости скорости образования мезо- молекул с/ф в газообразном дейтерии // Мезоны в веществе: Тр. Междунар. симпоз. Дубна, 1977. С. 199-205. • статей из журналов: Афанасьев Ю.В. и др. Лазерное инициирование термоядерной реакции в неоднородных сфериче- ских мишенях//Письма в ЖЭТФ. 1975. Т. 21, вып. 2. С. 150-155. Barbashov B.M., Pestov I.B. On Spinor Representations in the Weyl Gauge Theory // Mod. Phys. Lett. A. 1997. V. 12, No. 26. P. 1957-1968. • препринта и сообщения: Тяпкин А.А. Экспериментальные указания о существовании тахионов, полученные при исследо- вании черенковского излучения: Препринт ОИЯИ Д1-99-292. Дубна, 1999. 4 с. Muzychka Al. Yu., Pokotilovski Yu.N., Geltenbort P. Search for an Anomalous Transmission of Ultra- cold Neutrons Through Metal Foils: JINR Commun. E3-98-18. Dubna, 1998. 10 p.

Более подробные правила оформления Вы найдете на странице издательского отдела ОИЯИ: wwwl.jinr.ru Письма о физике элементарных частиц и атомного ядра № 6[109]-2001

Учредитель и издатель — Объединенный институт ядерных исследований.

Журнал зарегистрирован в Министерстве РФ по делам печати, телерадиовещания и средств массовых коммуникаций.

Регистрационное свидетельство ПИ № 77-1338 от 10.12.99.

Редактор Э. В. Ивашкевич Технический редактор Е. Н. Водоватова

ЛР № 020579 от 26.06.97. Подписано в печать 21.01.2002. Формат 70x100/16. Бумага офсетная. Гарнитура Тайме. Печать офсетная. Усл. печ. л. 6,45. Уч.-изд. л. 9,37. Тираж 700. Заказ 53068. Цена 11 р. 64 к.

Издательский отдел Объединенного института ядерных исследований 141980, Дубна Московской области, ул. Жолио-Кюри, 6.