1470 IEEE TRANSACTIONS ON NUCLEAR SCIENCE, VOL. 57, NO. 3, JUNE 2010
Low Temperature Scintillation in ZnSe CrystalsIoan Dafinei, Mauro Fasoli, Fernando Ferroni, Eva Mihokova, Filippo Orio, Stefano Pirro, and Anna Vedda
Abstract—The search for new materials which could be usedas scintillating bolometers is a subject of great interest in viewof next-generation experiments dedicated to the study of neu-trinoless double beta decay (DBD). Preliminary measurementsperformed on ZnSe bolometric and scintillation characteristicsgave very promising results for what concerns the amplitude ofthe scintillation signal and the capacity to discriminate between áand â particles. The current work studies the possibility to developa scintillating ZnSe crystal adapted for the use as hybrid (thermaland scintillation) detector operating at very low temperatures� �����, able to discriminate between á and â induced events.The results of optical transmission, X-ray excited steady-stateluminescence and thermo-luminescence measurements performedon ZnSe crystals produced in different conditions are presented.Based on these measurements, the scintillation properties ofundoped ZnSe crystals are carefully described. The correlationbetween these measurements and bolometric/scintillation tests isused in order to fix the production conditions of ZnSe crystalswhich best guarantee scintillation efficiency at very low tempera-tures for possible use in a DBD experiment.
Index Terms—Bolometers, double beta decay, scintillation detec-tors, ZnSe.
I. INTRODUCTION
T HE determination of neutrino mass is one of the mainchallenges in particle physics today. It is currently known
[1]–[3] that neutrinos oscillate between “flavor” eigenstates(electron-type, muon-type, and tau type) and that such os-cillation can only take place if at least one of the states hasfinite mass. However, the absolute scale of the neutrino masscannot be determined from oscillation experiments. Moreimportant, since they have no known charge-like attribute,neutrinos may be the only known fermions which do not havedistinct anti-particles (same mass and spin but opposite electriccharge). Hypothetical fermions of this type are called Majoranaparticles, in contrast to Dirac particles which have distinctantiparticles. The implications of massive neutrinos for modelsbeyond the Standard Model differ for Majorana and Dirac neu-trinos. Studies of the rare radioactive decay process known asdouble beta decay (DBD) offer the possibility to address both ofthese outstanding issues: the discovery of the neutrinoless DBD
Manuscript received July 01, 2009; revised September 04, 2009 and October09, 2009; accepted October 23, 2009. Date of current version June 16, 2010.
I. Dafinei is with the INFN Sezione di Roma, Rome 00185, Italy ( e-mail:[email protected]).
F. Ferroni and F. Orio are with the Università La Sapienza, Dipartimentodi Fisica and INFN Sezione di Roma, Rome 00185, Italy (e-mail: [email protected]; [email protected]).
S. Pirro is with the INFN Sezione di Milano Bicocca, Milan 20126, Italy(e-mail: [email protected]).
Digital Object Identifier 10.1109/TNS.2009.2035914
would be the signature of the Majorana nature of neutrinos andwould allow to find the values of neutrino mass. One of theseveral methods utilized to search for neutrinoless DBD is thebolometric technique in which the heat produced as a result ofthe energy released by stopping the electrons or positrons ismeasured. Usually in such experiments, the detector is also thesource of the double beta decay events [4].
The study of low temperature scintillating materials to beused in rare events physics became of major interest during re-cent years because such materials may offer the solution to builddetectors able to discriminate the rare signal from the back-ground caused by cosmic rays and natural radioactivity (mainlyfrom and radioactive chains). Part of this back-ground can be reduced with suitable shielding of the exper-imental setup, always placed deeply underground. Neverthe-less the inevitable contamination of the materials placed in theimmediate proximity of the detector presents a characteristicalpha-particles background having a continuum energy spec-trum that extends up to several MeV thus covering the neutrino-less DBD region of interest (2500 to 5500 keV). It is exactly thisbackground that can be eliminated by the use of hybrid cryo-genic detectors with double readout (heat and scintillation) inwhich and induced events can be discriminated thanks totheir different quenching factor [5].
The working principle of cryogenic detectors [6] is based onthe fact that at very low temperatures the heat capacity (C) of adielectric crystal varies with temperature (T) following Debye’slaw where is the Debye temperature of the ma-terial. Due to this characteristic, the heat capacity of a crystal attemperatures below 50 mK becomes so small that even a tinyamount of heat input such as that deposited by an elementaryparticle will result in a temperature increase of the crystal, usu-ally measured with special neutron transmutation doped (NTD)Ge thermistors. The production technology of NTD Ge consistsof modifying Ge crystals by neutron exposure. The neutron cap-ture and subsequent and electron capture (EC) decays producep and n doping. The effective doping depends on the neutrondose (of the order: ). A further improvement ofcryogenic detectors can be achieved in the case of a scintillatingmaterial by the simultaneous measurement of the temperaturevariation of the crystal and the light emitted following the en-ergy deposit of an elementary particle.
The aim of this paper is to present the results of preliminarywork aimed at finding the production conditions of ZnSe crys-tals which best guarantee a high scintillation efficiency at verylow temperatures for possible use in a future DBD experiment.The paper is organized as follows: Section II describes the rea-sons for the choice of ZnSe crystal as a possible material to beused in DBD experiments as well as the nature of the ZnSe sam-ples studied in the present work (dimensions and production de-tails). Section III presents the experimental methods and tools
Authorized licensed use limited to: Universita degli Studi di Roma La Sapienza. Downloaded on June 24,2010 at 11:58:28 UTC from IEEE Xplore. Restrictions apply.
DAFINEI et al.: LOW TEMPERATURE SCINTILLATION IN ZNSE CRYSTALS 1471
used to characterize the ZnSe crystal samples and the relatedexperimental results. Conclusions are given in Section IV.
II. MATERIAL REQUIREMENTS AND SAMPLE PREPARATION
A. Overview
In principle, any crystal which contains one of the DBDnuclides and scintillates is suitable for use as scintillatingbolometer for the study of DBD. Nevertheless isotopic abun-dance constraints as well as expected DBD energy spectrumcharacteristics limit the possibilities to only a few materials.ZnSe crystal is one of the best placed among these materialsbecause the expected double beta decay has thetotal kinetic energy value well abovethe limit of natural radioactivity background (2615 keV). Theisotope has a relatively large natural abundance (8.73%)which further places any Se compound in a favorable positionas active material to be used in DBD experiments. Moreover,luminescence in ZnSe is already known [7], [8] and the materialis currently produced in several laboratories and/or industrialunits for use in infrared optics and optoelectronic devices [9].On the other hand, currently still missing is a dedicated study ofthe luminescence properties of ZnSe at very low temperatures
at which scintillating bolometers are operated.Such a study allowing deep understanding of the scintillationprocess in ZnSe is fundamental for the possible use of ZnSecrystals in next-generation DBD experiments.
Our preliminary measurements performed on ZnSe bolo-metric and scintillation characteristics gave very promisingresults for what concern the amplitude of scintillation signaland the capacity to discriminate between alpha and beta par-ticles. Nevertheless quite large spread of characteristics wasnoted even for samples delivered by the same producer, de-pending on crystal growth and possible successive treatmentconditions. This behavior is easy to understand given the factthat luminescence in AII-BVI compounds is a process drivenby the nature and concentration of point defects. The mostcommon defects are vacancies and interstitials which may actas donor (Se vacancies and Zn intersititials ) or deeplevels in the forbidden band and/or acceptors (Zn vacancies
and Se interstitials ). Doping with isovalent (O, S,Te) elements is a very commonly used technique aimed atenhancing the luminescence performance of ZnSe. Supple-mentary local defects are created in this case which introducesupplementary energetic levels within the forbidden band, thusfurther easing radiative recombination of non-equilibrium freecarriers (luminescence).
A dedicated study was needed therefore in order to under-stand the origin of the different low temperature scintillationefficiencies measured for different ZnSe samples and possiblyfind the crystal growth and post-growth treatment conditionswhich guarantee a reliable production of ZnSe crystals suitablefor DBD application.
B. ZnSe Sample Preparation
Eight sets of optically polished ZnSe samples were orderedfor the present study from the “Alkor Technologies” companyof Saint-Petersburg, Russian Federation. Each set of twin sam-
Fig. 1. ZnSe cubic samples ����������� � prepared in different growthand post-growth conditions for this work.
ples was produced under different growth and post-growth con-ditions and consisted of one cubic sampleand four slabs with plan parallel faces. Thenumbering of samples in the following corresponds to the dif-ferent conditions of production (set of twin samples, from 1 to8).
The control of scintillating properties of ZnSe through crystalgrowth conditions is a very difficult task due to very restric-tive properties of the material, mainly the high melting point
with a phase diagram showing a very sharp congruentmelting which allows only small (0.05%) deviations from the1/1 stoichiometric ratio [10], [11]. The existence of a solid-solidphase transition at , the natural tendency of twinning,and the different vapor pressure of components further increasethe difficulty of controlling undoped ZnSe properties only bycrystal growth conditions.
All ZnSe samples used in the present work were grown fromthe same (stoichiometric) raw material of 6N purity. Specialcare was taken in the selection of the raw material used for thegrowth of ZnSe crystals which, for in the case of DBD applica-tion, have to be free of any radio-contaminant (especially fossilradio-isotopes like , , , and their daughters whichcan mimic or stimulate the searched signal in the specific energyrange where DBD signal is expected). Te contamination of ZnSecrystals grown in this work was also specially avoided due to thefact that Te is also a DBD element (both and areDBD active). Differences in samples composition is due there-fore only to different Se content of as-grown crystal ingot alongthe growth axis, a consequence of the higher evaporation coeffi-cient of Zn with respect to Se. The Se excess concentration mayvary from 0.001% to 0.05% between the top and the bottom ofthe as grown ingot. Further differences between sample compo-sition and structure are obtained through post-growth thermaltreatment, made in Se or Zn vapor-enriched atmosphere. Fig. 1provides a picture of the eight cubic samples belonging to theeight sets of samples studied in this work.
III. EXPERIMENTAL TECHNIQUES AND RESULT
Optical transmission measurements at room tempera-ture were performed on the optically polished ZnSe slabs
belonging to the eight sets of samples,using a Perkin Elmer Lambda900 spectrophotometer. As Fig. 2shows, the absorption spectra of the eight samples are differentin the region near the fundamental absorption edge. The differ-ences extend up to 2.15 eV, corresponding to the 475–575 nmwhich explains the different color from pale green to dark redof the samples (see Fig. 1).
Authorized licensed use limited to: Universita degli Studi di Roma La Sapienza. Downloaded on June 24,2010 at 11:58:28 UTC from IEEE Xplore. Restrictions apply.
1472 IEEE TRANSACTIONS ON NUCLEAR SCIENCE, VOL. 57, NO. 3, JUNE 2010
Fig. 2. Optical absoption spectra measured at room temperature on ZnSe slabsof 1 mm width, prepared in the same conditions as samples 1 to 8 given in Fig. 1.
Steady-state X-ray excited radio-luminescence andthermo-luminescence measurements were performed usinga home made measuring set-up consisting of a vacuumchamber Torr) and a cryostat allowing workingtemperatures in the range from 8 K to 320 K. In situ X-rayirradiation with an X-ray tube (Philips 2274) was performedthrough a Beryllium window. The emitted light (both for radioand thermo-luminescence) was measured by a CCD (JobinYvon Spectrum One 3000) coupled to a monochromator (JobinYvon Triax 180). Radio-luminescence spectra presented inthis work were made in identical conditions on crystal samplesof identical geometry and mechanicalquality (optically polished); therefore, it is possible to performa quantitative analysis of the luminescence yield for differentpreparation conditions of the samples.
As Fig. 3 shows, the X-Ray excited spectra observed at roomtemperature have the same structure for all sam-ples studied though the relative intensity of different emissionbands may differ from one sample to another. Besides the exci-tonic emission band at (2.7 eV) clearly seen in theinset of Fig. 3 (logarithmic scale), three main emission bandsare evidenced by a detailed analysis of each spectrum:
Studying the precise microscopic origin of each band is notthe subject of the present work. Nevertheless it is to be stressedthat the 645 nm band (enhanced by Te doping in commonly pro-duced ZnSe:Te scintillators [7]) is to be attributed only tocenters (Zn vacancies). As mentioned above, the Te contamina-tion is to be excluded in our case.
Preliminary low temperature X-ray excited luminescencespectra were measured on sample No. 1 considered as represen-tative, since at room temperature, it clearly features all emission
Fig. 3. X-ray excited luminescence measured at room temperature on ZnSeslabs of 1 mm width, prepared in the same conditions as samples 1 to 8 given inFig. 1. Emission peaks are better evidenced in the inset.
Fig. 4. X-ray excited luminescence measured at low temperatures on ZnSesample No. 1. The sharp luminescence increase below 1.24 eV (above 1000nm) is an artifact related to the spectral correction at the high wavelength limitof the detection range.
bands. The corresponding spectra represented in Fig. 4 havefundamentally the same composition for the entire temperaturerange (10 K to room temperature), there is no supplementaryemission band, i.e., the nature of luminescent centers remainsthe same. As expected from other studies [7], the emissionefficiency in the red region of the spectrum (610 nm band) isconsiderably increased at low temperatures while the infraredemission band at 970 nm remains practically unchanged. Thisbehavior is further confirmed by the relatively good correlationbetween the intensity of 610 nm band at room temperature andthe light yield of samples at very low temperatures illustratedat the end of this section. A follow-on detailed study of lowtemperature X-ray excited luminescence for all eight sets ofZnSe samples is under way. These measurements are expectedto explain the dependence of low temperature scintillationefficiency on ZnSe crystal production conditions and possiblygive further information on the temperature dependence of therelative weight of the emission bands at 610 nm and 645 nmwhich is to be noted on sample No. 1.
The same luminescent centers take part in the thermostimu-lated luminescence (TSL) process. As Fig. 5 shows, TSL spectra
Authorized licensed use limited to: Universita degli Studi di Roma La Sapienza. Downloaded on June 24,2010 at 11:58:28 UTC from IEEE Xplore. Restrictions apply.
DAFINEI et al.: LOW TEMPERATURE SCINTILLATION IN ZNSE CRYSTALS 1473
Fig. 5. Normalized TSL spectra measured in ZnSe sample No.1. The TSLsignal was integrated in two temperature regions (25 K to 100 K and 100 Kto 200 K).
Fig. 6. Measurement setup for low temperature light yield (LY). Each ZnSecrystal is surrounded by identical reflecting walls which guarantee an identicallight collection by the LD which is to be mounted above. Light production effi-ciencies (relative values) of the eight ZnSe cubic samples are given in the tablebelow the picture.
measured for the sample No. 1 are composed of the same emis-sion bands, except the excitonic one. The normalized TSL signalis given for two temperature ranges (from 25 K to 100 K andfrom 100 K to 200 K) in order to put in evidence the low energyshift of the emission in the high temperature integration range.Below 100 K, the carriers recombine practically only at the redemission centers (610 nm and 645 nm) while above 100 K, therecombination at the infrared center (970 nm) becomes predom-inant.
The light production efficiency at very low temperatures (10mK) of the ZnSe crystals was measured with a bolometric lightdetector (LD) consisting of a pure Ge disc absorber (40 mm di-ameter, 1 mm thick, having a thin (600Å) layer depositedon one face) thermally coupled with aNTD-Ge thermistor for which the andare chosen in order to work at temperatures between 9 and 14mK The layer is deposited on the monocristalline Ge discby a special technique in order to extend in the IR region the
Fig. 7. Correlation between low temperature light yield and RT emission at610 nm (sample number is given near each experimental value).
Fig. 8. Correlation between low temperature LY and optical absorption spectraof ZnSe crystal samples obtained with a similar thermal treatment protocol.
optical absorption of the device. The working principle of thelight detector is simple: the light emitted by the ZnSe crystalis fully absorbed in the disc and the correspondingtemperature rise of the pure Ge substrate is measured by theNTD thermistor. A detailed description of the performance ofthis low temperature light detector as well as the experimentalsetup of the detector coupled to a scintillating crystal is givenin [12]. The eight ZnSe cubic sampleswere mounted in the setup (Fig. 6) which guarantees identicalconditions for all samples from the point of view of light collec-tion. Each ZnSe crystal was surrounded by identical reflectingwalls and was placed at the same distance with respect to thelight detector so that a reliable classification of the eight sam-ples from the point of view of their scintillation efficiency ispossible. NTD-Ge thermistors are also mounted on each ZnSecrystals for the collection of the bolometric signal. In this way,each ZnSe cubic sample works as a hybrid (thermal and scintil-lation) detector. Light production efficiencies (relative values)of the eight ZnSe cubic samples are given in the table below thepicture.
Due to the complexity of LY measurement at low temper-atures, this technique cannot be used as routine test of ZnSecrystals certification for DBD use and/or for tuning of thermaltreatment conditions in the R&D phase of crystals production.These activities will be significantly simplified if the correlationbetween low temperature light yield and RT emission at 610 nmmentioned above and illustrated in Fig. 7 will be confirmed byfurther tests.
Authorized licensed use limited to: Universita degli Studi di Roma La Sapienza. Downloaded on June 24,2010 at 11:58:28 UTC from IEEE Xplore. Restrictions apply.
1474 IEEE TRANSACTIONS ON NUCLEAR SCIENCE, VOL. 57, NO. 3, JUNE 2010
More, it is expected that RT transmission spectra of ZnSecrystals prepared with a given thermal treatment protocol will bestrongly correlated with their low temperature LY as illustratedin Fig. 8.
IV. CONCLUSION
Four emission bands were found in the X-ray excited lumi-nescence spectrum of ZnSe crystals produced in different con-ditions from the same (stoechiometric, 6N purity) raw material.Besides a weak band at 460 nm attributed to excitonic recombi-nation, the other bands are situated in the red (610 nm and 645nm) and infrared (970 nm) spectral regions. The post-growthtreatment conditions leading to the best luminescence efficiencyat 10 mK were found to be those corresponding to samples No. 5and 4. The excellent low temperature light yield performance ofsuch samples place the ZnSe crystal among the best suited ma-terials to be used in DBD experiment of next generation. Furtherlow temperature measurements are on the way on ZnSe samplesprepared in different conditions in order to understand the mi-croscopic nature of luminescence mechanism at cryogenic tem-peratures (below 50 mK) and possibly predict the scintillationbehavior at these temperatures from transmission measurementsmade at room temperature.
REFERENCES
[1] Y. Fukuda et al., “Super-kamiokande collaboration,” Phys. Rev. Lett.,vol. 81, p. 1562, 1998.
[2] S. N. Ahamed et al., “SNO collaboration,” Phys. Rev. Lett., vol. 92, p.181301, 2004.
[3] T. Araki et al., “KamLAND collaboration,” Phys. Rev. Lett., vol. 94,p. 081801, 2005.
[4] E. Fiorini, “Searches on neutrino physics with cryogenic detectors,” J.Phys.: Conf. Series, vol. 173, p. 012004, 2009.
[5] S. Pirro, J. W. Beeman, S. Capelli, M. Pavan, E. Previtali, and P.Gorla, “Scintillating double-beta decay bolometers,” Phys. AtomicNuclei, vol. 69, no. 12, pp. 2109–2116, 2006.
[6] E. Fiorini and T. O. Niinikoski, “Low-temperature calorimetry for raredecays,” Nucl. Instrum. Meth. Phys. Res., vol. 224, no. 1–2, pp. 83–88,Jul. 1984.
[7] N. Starzhinkiy et al., “New trends in the development of AIIBVI-basedscintillators,” IEEE Trans. Nucl. Sci., vol. 55, pp. 1542–1546, Jun.2008.
[8] E. Krause et al., “Influence of growth non-stoichiometry on opticalproperties of doped and non-doped ZnSe grown by chemical vapourdeposition,” J. Crystal Growth, vol. 138, pp. 75–80, 1994.
[9] V. D. Ryzhikov et al., “Properties of semiconductor scintillatorsZnSe(Te,O) and integrated scintielectronic radiation detectors basedthereon,” IEEE Trans. Nucl. Sci., vol. 48, pt. 1, pp. 356–359, Jun. 2001.
[10] Bulk Crystal Growth of Electronic, Optical & Optoelectronic Mate-rials, P. Capper, Ed. New York: Wiley, 2005.
[11] Springer Handbook of Electronic and Photonic Materials, S. Kasapand P. Capper, Eds. New York: Springer, 2006.
[12] S. Pirro, C. Arnaboldi, J. W. Beeman, and G. Pessina, “Developmentof bolometric light detectors for double beta decay searches,” Nucl.Instrum. Meth. Phys. Res., vol. A 559, pp. 361–363, 2006.
Authorized licensed use limited to: Universita degli Studi di Roma La Sapienza. Downloaded on June 24,2010 at 11:58:28 UTC from IEEE Xplore. Restrictions apply.
