http://crossmark.crossref.org/dialog/?doi=10.1140/epjc/s10052-015-3657-5&domain=pdf
Web End = http://crossmark.crossref.org/dialog/?doi=10.1140/epjc/s10052-015-3657-5&domain=pdf
Web End = http://crossmark.crossref.org/dialog/?doi=10.1140/epjc/s10052-015-3657-5&domain=pdf
Web End = Eur. Phys. J. C (2015) 75:546DOI 10.1140/epjc/s10052-015-3657-5
Regular Article - Experimental Physics
http://crossmark.crossref.org/dialog/?doi=10.1140/epjc/s10052-015-3657-5&domain=pdf
Web End = http://crossmark.crossref.org/dialog/?doi=10.1140/epjc/s10052-015-3657-5&domain=pdf
Web End = http://crossmark.crossref.org/dialog/?doi=10.1140/epjc/s10052-015-3657-5&domain=pdf
Web End = http://crossmark.crossref.org/dialog/?doi=10.1140/epjc/s10052-015-3657-5&domain=pdf
Web End = http://crossmark.crossref.org/dialog/?doi=10.1140/epjc/s10052-015-3657-5&domain=pdf
Web End = http://crossmark.crossref.org/dialog/?doi=10.1140/epjc/s10052-015-3657-5&domain=pdf
Web End = http://crossmark.crossref.org/dialog/?doi=10.1140/epjc/s10052-015-3657-5&domain=pdf
Web End = Lowering the radioactivity of the photomultiplier tubes for the XENON1T dark matter experiment
XENON Collaboration1,a
E. Aprile2, F. Agostini3,17, M. Alfonsi4,10, L. Arazi5, K. Arisaka6, F. Arneodo7, M. Auger8, C. Balan9, P. Barrow8,L. Baudis8,b, B. Bauermeister10, A. Behrens8, P. Beltrame5,c, A. Brown11, E. Brown12,14, S. Bruenner13, G. Bruno3,14,R. Budnik5, L. Btikofer15, J. M. R. Cardoso9, D. Coderre15, A. P. Colijn4, H. Contreras2, J. P. Cussonneau16,M. P. Decowski4, A. Di Giovanni7, E. Duchovni5, S. Fattori10, A. D. Ferella3, A. Fieguth14, W. Fulgione3,M. Galloway8, M. Garbini17, C. Geis10, L. W. Goetzke2, C. Grignon10, E. Gross5, W. Hampel13, R. Itay5, F. Kaether13,G. Kessler8, A. Kish8, H. Landsman5, R. F. Lang11, M. Le Calloch16, D. Lellouch5, L. Levinson5, C. Levy12,14,S. Lindemann13, M. Lindner13, J. A. M. Lopes9,d, A. Lyashenko6, S. Macmullin11, T. Marrodn Undagoitia13,e,J. Masbou16, F. V. Massoli17, D. Mayani8, A. J. Melgarejo Fernandez2, Y. Meng6, M. Messina2, B. Miguez18,A. Molinario18, M. Murra14, J. Naganoma19, U. Oberlack10, S. E. A. Orrigo9,f, P. Pakarha8, E. Pantic6,R. Persiani16,17, F. Piastra8, J. Pienaar11, G. Plante2, N. Priel5, L. Rauch13, S. Reichard11, C. Reuter11,A. Rizzo2, S. Rosendahl14, J. M. F. dos Santos9, G. Sartorelli17, S. Schindler10, J. Schreiner13, M. Schumann15,L. Scotto Lavina16, M. Selvi17, P. Shagin19, H. Simgen13, A. Teymourian6, D. Thers16, A. Tiseni4, G. Trinchero18,C. Tunnell4, O. Vitells5, R. Wall19, H. Wang6, M. Weber13, C. Weinheimer14, M. Laubenstein3
1 Laboratori Nazionali del Gran Sasso, Assergi, Italy
2 Physics Department, Columbia University, New York, NY, USA
3 INFN-Laboratori Nazionali del Gran Sasso and Gran Sasso Science Institute, LAquila, Italy
4 Nikhef and the University of Amsterdam, Science Park, Amsterdam, Netherlands
5 Department of Particle Physics and Astrophysics, Weizmann Institute of Science, Rehovot, Israel
6 Physics and Astronomy Department, University of California, Los Angeles, CA, USA
7 New York University Abu Dhabi, Abu Dhabi, United Arab Emirates
8 Physik-Institut, University of Zurich, Zurich, Switzerland
9 Department of Physics, University of Coimbra, Coimbra, Portugal
10 Institut fr Physik & Exzellenzcluster PRISMA, Johannes Gutenberg-Universitt Mainz, Mainz, Germany
11 Department of Physics and Astronomy, Purdue University, West Lafayette, IN, USA
12 Department of Physics, Applied Physics and Astronomy, Rensselaer Polytechnic Institute, Troy, NY, USA
13 Max-Planck-Institut fr Kernphysik, Heidelberg, Germany
14 Institut fr Kernphysik, Wilhelms-Universitt Mnster, Mnster, Germany
15 Albert Einstein Center for Fundamental Physics, University of Bern, Bern, Switzerland
16 Subatech, Ecole des Mines de Nantes, CNRS/In2p3, Universit de Nantes, Nantes, France
17 Department of Physics and Astrophysics, University of Bologna and INFN-Bologna, Bologna, Italy
18 INFN-Torino and Osservatorio Astrosico di Torino, Turin, Italy
19 Department of Physics and Astronomy, Rice University, Houston, TX, USAReceived: 27 March 2015 / Accepted: 4 September 2015 / Published online: 23 November 2015 The Author(s) 2015. This article is published with open access at Springerlink.com
ae-mail: mailto:[email protected]
Web End [email protected]
b e-mail: mailto:[email protected]
Web End [email protected]
c Present address: University of Edinburgh, Edinburgh, UK
d Also at Coimbra Engineering Institute, Coimbra, Portugal
e e-mail: mailto:[email protected]
Web End [email protected]
f Present address: IFIC, CSIC-Universidad de Valencia, Valencia, Spain
Abstract The low-background, VUV-sensitive 3-inch diameter photomultiplier tube R11410 has been developed by Hamamatsu for dark matter direct detection experiments using liquid xenon as the target material. We present the results from the joint effort between the XENON collaboration and the Hamamatsu company to produce a highly radio-pure photosensor (version R11410-21) for the XENON1T dark matter experiment. After introducing the photosensor and its components, we show the methods and results of the radioactive contamination measurements of the individual
123
546 Page 2 of 10 Eur. Phys. J. C (2015) 75 :546
materials employed in the photomultiplier production. We then discuss the adopted strategies to reduce the radioactivity of the various PMT versions. Finally, we detail the results from screening 286 tubes with ultra-low background germanium detectors, as well as their implications for the expected electronic and nuclear recoil background of the XENON1T experiment.
1 Introduction
Among the various experimental methods for the direct detection of dark matter particles, liquid-xenon (LXe) time projection chambers have demonstrated the highest sensitivities over the past years[14]. The XENON100 experiment[5] did not nd any evidence for dark matter and published the worlds best upper limits on spin-independent[6] and spin-dependent[7] couplings of weakly interacting massive particles (WIMPs) to nucleons and neutrons, respectively, for WIMP masses above 10GeV/c2. Recently, the
LUX experiment has conrmed and improved upon these results, reaching an upper limit on the spin-independent WIMP-nucleon cross section of 7.6 1046 cm2 at a WIMP
mass of 33GeV/c2 [8].
To signicantly increase the experimental sensitivities, the XENON collaboration is building the XENON1T experiment[9]. The construction at the Laboratori Nazionali del Gran Sasso (LNGS) in Italy started in autumn 2013 and will continue until summer 2015. Detector commissioning, and a rst science run are expected for fall and late 2015, respectively, while the design dark matter sensitivity will be reached after two years of operation. With 35 times more target mass, and a background goal 100 times lower than that of XENON100[10,11], the XENON1T sensitivity to spin-independent WIMP-nucleon cross sections is expected to be 21047 cm2 at a WIMP mass around 40GeV/c2. This ambi
tious goal demands an ultra-low background in the central detector region, implying a very low radioactivity of all detector construction materials, in particular those in the vicinity of the xenon target. Among these, the photosensors, that detect the primary and secondary scintillation light produced in a WIMP-nucleus collision, are the most complex, multi-material components. Photomultiplier tubes (PMTs) were shown to cause a major contribution to the electronic recoil background from detector materials in the XENON100[10] and LUX[12] experiments. Other current and planned experiments based on xenon as a detection medium for rare-event searches are XMASS[13], PandaX[14], LZ[15], DARWIN[1618], NEXT[19] and EXO[20].
Many of these projects employ PMTs for signal readout, therefore the availability of ultra-low background, high-quantum efciency sensors, capable to operate stably for long periods at low temperatures is of crucial importance. PandaX,
XENON1T, LZ and possibly DARWIN use or plan to use a new, 3-inch diameter tube, R11410, recently developed by Hamamatsu[21]. The tube was extensively tested at room temperature and in LXe[22,23], and rst measurements of the radioactivity of an early PMT version were presented in[2325].
In this paper we present the most recent version, named R11410-21, which has been developed by Hamamatsu together with the XENON collaboration to produce a tube that meets the strict background requirements of the XENON1T experiment. In Sect.2, we summarise the general characteristics of the PMT. In Sect.3, we introduce the measurement techniques that were employed to evaluate the radioactive contamination of individual PMT components and of the PMT as a whole. In Sect.4, we present the main screening results for the construction materials and the rst PMT production batches. Finally, in Sect.5, we interpret the results and discuss their impact on the expected backgrounds of the XENON1T experiment.
2 The R11410 photosensor
The R11410 photomultiplier is a 3-inch diameter tube produced by Hamamatsu and used in xenon-based dark matter and double beta decay detectors. It operates stably at typical temperatures and pressures in a LXe detector, around
100 C and 2atm., respectively[23]. Apart from a greatly reduced intrinsic radioactivity level, as we will show in the following sections, a major advantage is its high quantum efciency (QE) at the xenon scintillation wavelength of 175nm. Hamamatsu has measured a mean value of QE =
35 % for the tubes delivered for XENON1T with a few tubes having a QE as high as 40%. For selected tubes of an earlier version, the QE has been measured by some of us at various temperatures[26] conrming the values reported by the company. Along with 90% electron collection efciency[22], the tube ensures a high detection efciency for VUV scintillation photons produced by particle interactions in xenon.
The R11410 photomultiplier has a VUV-transparent quartz window and a low-temperature bialkali photocathode deposited on it. A 12-dynode electron-multiplication system provides a signal amplication of about 3.5106 at 1500 V
operating voltage. The peak-to-valley ratio is at least 2, showing a good separation of the single photoelectron signal from the noise spectrum[23]. Inside the tube, the dynodes are insulated using L-shaped quartz plates. The body of the PMT is about 115mm long[27], and it is made out of a Kovar alloy, most of which has a very low cobalt content (cobalt-free Kovar). On the back side of the PMT, the stem uses ceramic material to insulate the connections to the individual dynodes. The individual components are listed in Table 2.
123
Eur. Phys. J. C (2015) 75 :546 Page 3 of 10 546
Fig. 1 A schematic gure of the R11410 PMT (not to scale), showing its main components.The numbers correspond to those in Table 2, the colours to those in Fig. 5
The tube has been intensively tested to demonstrate its stable behaviour during long-term operation in LXe and in the presence of an electric eld[23]. At 100 C, this PMT shows a
low dark count rate of about 80Hz at 1600 V above an area
of the single photoelectron (PE) peak of 1/3PE. The corresponding single photoelectron resolution is between 35 and 40%[23]. In Fig.1 we show a sketch of the R11410 PMT, displaying its main components that were introduced in this section.
3 Techniques and measurements
We now briey describe the techniques used to measure the radioactive contamination of the PMTs and of the various PMT components. The most relevant isotopes are those of the natural uranium and thorium decay chains. In addition to the gamma rays, spontaneous ssion and (, n) reactions can produce fast neutrons in a given material. These neutrons can cause nuclear recoils in the LXe, similar to the expected interactions of WIMPs, and thus present a serious source of background. We also study the 60Co, 40K and 137Cs isotopes, among others, because their gamma-emission contributes to the electron-recoil background of the experiment. This contribution can, however, be reduced by signal/background discrimination in xenon time projection chambers that measure both the light and charge signals produced in a particle interaction in the active detector volume.
3.1 Germanium spectroscopy
Gamma-ray spectroscopy is a standard method to screen and select materials for rare-event searches. Germanium spectrometers in dedicated low-background shields combine a high energy resolution with a very low intrinsic background.
We have employed some of the worlds most sensitive high-purity germanium (HPGe) spectrometers, the GeMPI[28 30] and Gator[31] detectors. They are located at the LNGS laboratory, under an average overburden of 3600m.w.e., where the muon ux is reduced to 1m2 h1. The detec
tors are coaxial, p-type HPGe crystals, with masses around2.22.3kg, housed in electro-rened copper cryostats. The cryostats are surrounded by shields made of low-activity copper, lead and polyethylene, and are housed in air-tight boxes which are permanently ushed with pure nitrogen gas to suppress the inux of 222Rn. The sample chambers have dimensions of (25 25 30)cm3, allowing the placement
of rather large samples around the germanium detectors. Sample handling chambers equipped with airlocks and glove boxes are located above the shields, with compartments that allow for the storage of samples prior to their measurement and thus for the decay of 222Rn, 220Rn and their progenies. Before introducing the material samples into the airlock, their surface is cleaned. The samples are degreased with an acid soap and rinsed with deionized water in an ultrasonic bath. Finally, the surfaces are cleaned with ethanol. The PMTs are also cleaned before the measurements, their surface is wiped with ethanol.
Sealed calibration sources can be fed into the inner chambers in order to calibrate the energy scale and to determine the energy resolution. The relative efciency of the Ge spectrometers is around 100%. In -spectroscopy, the quoted efciency is dened relative to a 7.62cm (diameter) 7.62 cm (height) NaI(Tl) crystal, for the 1.33MeV
60Co photo-absorption peak at a source-detector distance of 25cm[32]. The integral counting rate in the energy region 1002700keV is in the range 610 events/h. They can detect sample activities down to 10Bq/kg for 226Ra and
228Th/228Ra. To determine the concentration of a radionuclide in a given sample, the most prominent gamma lines
123
546 Page 4 of 10 Eur. Phys. J. C (2015) 75 :546
Table 1 Main material components used across the different R11410 PMT versions. (1)For the production of the R11410-21 tube, that features the same design as the R11410-20 version, very low-background
materials were selected to fulll the demanding background requirements of the XENON1T experiment
Part R11410 R11410-10 R11410-20 R11410-21(1) R11410-30
Stem Glass Ceramic Ceramic Ceramic Sapphire
Body Kovar alloy Kovar alloy Co-free Kovar Co-free Kovar Co-free Kovar
Insulator Ceramic Quartz Quartz Quartz Quartz
Al seal Std. purity Std. purity High purity High purity High purity
in the spectrum are analysed after subtraction of the background spectrum. The detection efciencies for each line are determined in Monte Carlo simulations which are run individually for each type of sample and the full detector-sample geometry. As a concrete example, the efciencies for most gamma lines for a batch of 15 PMTs is around 0.50.8%. The screening time is typically 23weeks for each of these results. Upper limits are calculated as described in [33] and given at 90%C.L. Further details about the analysis procedure can be found in [31].
3.2 Glow-discharge mass spectrometry
To detect trace impurities from solid samples, glow-discharge mass spectrometry (GDMS) was used. The measurements presented here were carried out by the Evans Analytical Group (EAG)[34] on behalf of the XENON collaboration. The samples are exposed to a gas discharge or plasma atom-isation/ionisation source. Argon is used as discharge gas to sputter the atoms out from the sample. The ion beam is then directed into a system of focusing lenses and cones, which nally inject the ions into a quadrupole mass spectrometer. The uncertainty of the result is between 20 and 30% for all measurements in Al, Cu and Fe based samples (aluminium, stainless steel, Kovar), and 3050% for all other samples (quartz, sapphire and ceramic). The detection limit is better than 109 g/g (ppb) for conductors and an order of magnitude higher for non-conductors. This method detects traces of uranium and thorium in a sample, which are then translated into contaminations on 238U and 232Th while germanium counting mainly measures the 226Ra and 228Th/228Ra isotopes of the U and Th chains in a given material. Assuming secular equilibrium in the decay series, the conversion factors from units of Bq/kg to concentrations by weight, more common in mass spectrometry, are[35]: 1Bq 238 U/kg [hatwide]
=81109gU/g
(81ppb U) and 1Bq 232Th/kg [hatwide]
=246109 gTh/g (246ppb
Th). The GDMS measurement is particularly relevant for the early part of the 238U chain, which can be barely detected via germanium spectroscopy due to the low energy and low intensity of the emitted gamma lines.
3.3 Overview of the PMT screening measurements
The radioactivity of the materials employed to manufacture the R11410 tube has been gradually reduced from one version to the next, with the goal of minimizing the overall radioactivity of the nal product. The rst R11410 version used a glass stem, standard Kovar alloy for the body, and ceramic as a dynode insulator. The seal between the body and the quartz window was made with aluminium of standard (std.) purity. As this tube version still had a fairly high radioactivity level[24], the subsequent versions were targeted at replacing the main components with new materials such as Co-free Kovar alloy for the body of the tube, quartz instead of ceramic as an insulator, and high-purity aluminium for the seal. We note that the second PMT version was originally delivered as R11410-MOD and later renamed to R11410-10. The R11410-21 version employs low-background materials selected by us (see Sect.4.1) for those PMTs to be operated in XENON1T only. In an attempt to further reduce the radioactivity of the tube, a stem made out of sapphire has been proposed by Hamamatsu, R11410-30. It will not be employed in XENON1T, for the overall background reduction would be insignicant, as we detail in Sect.5. The modications that occurred between the different PMT versions are summarised in Table1.
4 Results
In this section, we rst present the measurements of the intrinsic radioactive contamination of the individual phototube construction materials. These are followed by the germanium screening results for the fully operational PMTs, and a comparison with other PMT types operated in rare event searches based on liquid xenon.
4.1 Material samples
A total of 12 material samples, most of these to be used for the R11410-21 PMT construction, have been measured. Table2 compiles a list of the studied materials, the sample masses used for germanium screening, as well as the mass of each material employed for manufacturing of one photomultiplier tube. The resulting mass for one tube is 190.5g, while the
123
Eur. Phys. J. C (2015) 75 :546 Page 5 of 10 546
Table 2 List of measured material samples including the mass of each component per PMT. (1)The sapphire (sample 12) is an alternative material to the ceramic (sample 9) for the stem of a possible future PMT version
No. Sample Mass forGe-screening [g]
Mass per PMT [g]
1 Quartz: faceplate (PMT window)
1179 30
2 Aluminium: pure Al for sealing
515 0.6
3 Kovar: Co-free body 500 78
4 Stainless steel sheets: electrode disk
555 8.2
5 Stainless steel sheets: dynodes
510 7.2
6 Stainless steel: shield 519 4
7 Quartz: L-shaped insulation
838 14.4
8 Kovar: ange of faceplate
525 18
9a Ceramic: stem 597 16
9b Ceramic: stem 498 16
10 Kovar: ange of ceramic stem
511 14
11 Getter (10 pieces) 0.58 5.810212 Sapphire (for
R11410-30)(1)
243 16
measured mass of a PMT is 1891g. Thus, both numbers
are in agreement within the uncertainties.
For most components, the mass of the screened material was around or above 500g, with the exception of the quartz, sapphire and getter samples. For the analysis with the GDMS technique, typically O(10)g of each material were needed.
The results of the germanium and mass spectrometry measurements for the isotopes of the uranium and thorium chains are summarised in Table3. Upper limits are given at 90% condence level, and the quoted errors include statistical and systematic errors. The item numbers correspond to the materials as listed in Table2.
We remark that the presented results on 238U are obtained by inspecting the 92.4 and 92.8keV lines from 234Th and the 1MeV line from 234mPa. However, one can also infer the 238U activity from the measured activity for 235U, as we can assume their natural abundance (where the activities are related as A235 0.05 A238). In general, the analysis of the
235U lines provide a higher sensitivity. Secular equilibrium between various parts of the natural decay chains may be broken in processed materials, as the half-lives of 238U and
228Th and of some of their daughters are long, and in most cases equilibrium was not re-established. Nonetheless, the results of both techniques, HPGe screening and GDMS, are compatible with one another. For the ceramic and sapphire
samples, only germanium screening results are reported due to large systematic uncertainties of the GDMS method for this type of material. The observed contamination with 40K,
60Co and 137Cs of each sample is shown in Table4.
In addition to the 10 getter pieces, a getter extracted directly from a PMT was measured. As only one piece was available, the obtained upper limits for most isotopes are much above the sensitivity of those for other materials. Nonetheless, a contamination with 40K clearly above the one measured for the 10 getter samples is observed. Further investigations are ongoing. The discrepancy is likely due to processes that occur during the PMT assembly, and the implications are discussed in Sect.5.
The sum of the measured contamination of all PMT parts, for a given isotope, is shown at the bottom of each table. We calculate a total upper limit (L) using either the measured contamination level, if available, or the upper limit from the most sensitive method, and a total detection (D) using only measured contaminations. These sums determine the range for the total expected radioactivity budget of a PMT, to be compared with the screening of the whole PMT, as presented in the next section. In Fig.2 (top), we show as an example the spectrum of the ceramic sample, together with the background spectrum of the HPGe detector.
From all screened materials, the ceramic has the highest contribution to the total radioactivity of the tube, in particular regarding the 238U component, which is at the level of2.5mBq/PMT.
4.2 Measurements of PMT batches
All XENON1T tubes are screened before their installation into the detector in order to verify the low radioactivity values of each production batch. We have screened one batch of R11410-20 PMTs and 14 batches of R11410-21 PMTs with the Gator detector. Two batches of 4 PMTs were screened with a GeMPI detector. The results of both analyses, performed independently, agree with one another within the quoted errors. In the measurements performed with the GeMPI detector, the four tubes were placed with the PMT cathode facing the Ge crystal. Due to this conguration, the estimation of the detection efciency determined by Monte Carlo simulations is affected by the location of the radioactive contamination in the tube. The systematic uncertainty has been evaluated to be up to 30%. For the measure
ments with Gator, low-background polytetrauoroethylene (PTFE) support structures were fabricated to ensure the safe and reproducible handling of all PMTs during the screening process. In Fig.3 we show a schematic view of 15 tubes in the Ge detector screening chamber. This is the maximum number of tubes that can be measured simultaneously in the PTFE holders, due to space constraints. Figure4 shows the sensi-
123
546 Page 6 of 10 Eur. Phys. J. C (2015) 75 :546
Table 3 Results of the germanium (Ge) screening and the mass spectroscopy (GDMS) in mBq/PMT for isotopes in the 238U, 235U and 232Th
chains. The sample numbers correspond to those in Fig.1, Table 2 and Fig.5. For the GDMS measurement, only results for 238U and 232Th are available. The ceramic material (sample 9) was measured twice, with
different HPGe detectors and analyzed by different groups, as a cross check of the sensitivity. The total upper limit (L) is calculated using the measured contamination level or the inferred upper limit while the total detection (D) considers only detected values. Sample 12 (alternative material to ceramic) is not considered in the sums listed in the table
Sample Method 238U 226Ra 228Ra (232Th) 228Th 235U
no. (detector)
1 Ge <0.33 3.6(6) 102 <1.2102 <1.1102 <1.1102
1 GDMS <1.8 <0.63
2 Ge <2.8102 <7.2104 <6.6104 6.6(3) 104 <5.51042 GDMS <0.15102 <2.5104
3 Ge <9.2 <0.26 <0.36 <0.34 <0.17
3 GDMS <9.5102 <3.2103 4 Ge <0.90 <2.5102 <4.3102 <3.3102 <1.91024 GDMS 5.0 102 0.68 102
5 Ge <0.53 2.7(6) 102 <9.4103 <9.4103 <1.11025 GDMS <0.17 <15103
6 Ge <0.37 1.3(2) 102 8(4) 103 8(4) 103 <5.21026 GDMS <3.4102 <17103
7 Ge <0.2 2.9(3) 102 <1.1102 <0.7102 <0.6102
7 GDMS <0.88 <0.3
8 Ge <0.79 3.7(7) 102 <1.8102 <1.8102 <1.81028 GDMS 4.4 102 0.98 102
9a Ge 2.4(4) 0.26(2) 0.23(3) 0.11(2) 0.11(2)
9b Ge 3(1) 0.30(3) 0.21(3) 0.11(2) 0.11(3)
10 Ge <0.65 <8.3103 <3102 <7.5103 <1.710210 GDMS <1.7102 <6103
11 Ge <0.7 0.035(4) <2102 <2.9102 0.018(3)
Total L Ge <16 <0.75 <0.74 <0.67 <0.44
Total D Ge 2.7 0.46 0.24 0.12 0.13
Total L GDMS <3.1 <1
Total D GDMS 0.09 0.02
Sample 12 Ge <0.72 0.72(4) 0.13(3) 0.08(2) <3.4102
tivity for various isotopes as a function of time, assuming that 15 PMTs are installed in the detector chamber.
For most relevant isotopes to our study, a sensitivity below 1mBq/PMT can be achieved after about 418days of measuring time. An example spectrum is shown in Fig.2 (bottom), along with the background spectrum of the Ge spectrometer, including the PTFE PMT holder. The screening results for 238U, 226Ra, 235U, 228Ra, 228Th, 40K, 60Co and
110mAg for each batch are shown in Table 5. As the presented upper limits on 238U are obtained from inspecting the 234Th and 234mPa decays, we note here that the inferred
238U activity from the measured activity for 235U is about (81)mBq/PMT.
The results for 40K are lower for the detector which measured only 4 tubes at once (marked as (1)) compared to the batches measured with Gator. However, considering the 30%
systematic error due to the efciency uncertainty for this detector, the values become compatible across the detectors. Gamma lines from the anthropogenic isotope 110mAg (T1/2 = 249.8 d) have also been identied in the spectrum.
The contamination has been located by two separate measurements with a HPGe detector as being present in the silver brazing between the ceramic stem and the leads and the Kovar ange of the PMT. The screening results of two samples provided by Hamamatsu with masses of 50 and 160g, respectively, are consistent with the 110mAg activities observed in the individual PMT batches.
Comparing the total activities from Tables3 and 4 with the results from the PMT batch screening in Table5, we see that these are consistent for all inspected isotopes, with the exception of 40K and 60Co. We will discuss these in Sect.5.
123
Eur. Phys. J. C (2015) 75 :546 Page 7 of 10 546
Table 4 Results of the germanium screening, in mBq/PMT, for 40K,
60Co and 137Cs for the main raw materials of the R11410-21 tube. The totals L and D are calculated as in the previous table
Sample no. 40K 60Co 137Cs
1 <8.1 102 <4.5 103 <4.8 1032 <5.7 103 <1.2 104 <2.3 1043 <0.99 7(2) 102 <0.104 <6.4 102 7.2(6) 102 <8.0 1035 6(3) 102 6(3) 103 <7.9 1036 <3.2 102 2(1) 103 <2.6 1037 7(3) 102 <2.3 103 <1.6 1038 7(4) 102 0.26(2) 9(4) 1039 1.1(2) <2102 <21029 1.6(2) <1.6102 <1.2 102 10 7(4) 102 0.22(2) 1.3(6) 102 11 8(3) 102 <4.2 103 <4.2 103
Total L <3.1 <0.66 <0.16
Total D 1.9 0.61 0.021
Sample 12 0.14(6) <8.8 103 <1.5 102
Fig. 3 Schematic view of 15 R11410-21 PMTs in the Gator detector chamber, held by custom-made, low-background PTFE support structure. The germanium detector in its cryostat is seen in green in the centre
Energy [keV]
-1
day
-1
-6
10
U
Ceramic (Al Background
O
2
)
3
Pb
Th
Pb
Ac
K
Tl
Bi
Bi
Cs Bi
Bi
Bi
Tl
Bi
Counts keV
-7
10
10
10
-8
-9
500 1000 1500 2000 2500
-1
day
-1
-6
10
15x R11410-21 Background
U
Ag
(e
e
)
Mn
Pb
Co
Bi
Cs
Ac
Ag
Bi
Co Co
Ag
K
Tl
Counts keV
-7
10
10
-9
10
-8
500 1000 1500 2000 2500
Energy [keV]
Fig. 2 (Top) Spectrum of the ceramic sample (no.9, red) along with the detector background (black). (Bottom) Spectrum of 15 Hamamatsu R11410-21 PMTs measured in July 2013 (red). The background spec-
trum of the Gator detector together with the PTFE support structure for the PMTs is also shown (black)
123
546 Page 8 of 10 Eur. Phys. J. C (2015) 75 :546
In Table 6, we show the screening results normalized by active PMT photocathode area, and compare these with previous versions and other PMT types used in liquid xenon experiments. We remark that the R11410-10 activities
Measuring time [days]
Sensitivity [mBq/PMT]
238
Ra
226
U
235
Ra
228
Th
228
U
2
10
10
1
-1
10 2 4 6 8 10 12 14 16 18 20
Fig. 4 Sensitivity for various isotopes as a function of measuring time, assuming that 15 PMTs are placed inside the detector chamber
obtained by the LUX collaboration [25] are slightly lower than those measured by the PandaX [14] and the XENON collaborations for the same version of the tube. The R8520 sensor is a 1-inch square tube employed in the XENON100 experiment[5] and the R8778 is a 2-inch cylindrical tube operated in the LUX detector[8].
The radioactivity per photocathode area of the XENON1T phototube, R11410-21, has been lowered for all isotopes compared with previous versions, and in particular with the one of the R8520 tubes employed in the XENON100 experiment.
5 Discussion and impact on the XENON1T background
The contributions of the individual materials to the radioactivity budget of the R11410-21 PMT are shown in Fig.5. Each row shows the relative contribution of these materials to the 238U, 226Ra, 228Ra, 228Th, 40K, 60Co and 137Cs levels, in those cases when an actual detection was possible.
Table 5 Screening results for several R11410 PMT batches using a HPGe detector. The number of measured tubes and the measuring time, t, in days of each measurement are shown in the rst columns. The
137Cs upper limit is in the range <0.20.3mBq/PMT for all measured batches. (1)These three sets of 4 PMTs were screened with a different detector
PMT version Batch no. t [d] Activity [mBq/PMT] (no. of units)
238U 226Ra 235U 228Ra 228Th 40K 60Co 110m Ag
v-20 (10) 0 15 <18 <0.82 <0.79 0.9(3) 0.9(2) 12(2) 1.3(2) <0.21
v-21 (10) 1 26 <18 0.4(1) 0.5(1) <1.1 0.4(1) 12(2) 0.7(1) 0.89(1)
v-21 (16) 2 15 <16 0.5(1) 0.29(9) <0.85 <0.61 13(2) 0.79(8) 1.1(2)
v-21 (4)(1) 2b 6 <19 0.7(2) <1.0 <0.7 0.3(1) 11(3) 0.7(2) 0.9(2)
v-21 (15) 3 11 <20 <0.82 <0.52 <1.1 0.5(2) 13(2) 0.73(9) 0.51(7)
v-21 (15) 4 22 <13 0.5(1) 0.35(9) 0.4(1) 0.4(1) 12(2) 0.73(9) 0.21(4)
v-21 (15) 5 16 <17 0.6(1) <0.57 <0.93 <0.62 14(2) 0.63(7) 0.22(6)
v-21 (11) 6 23 <15 0.6(1) <0.55 <0.77 0.7(1) 14(2) 0.71(7) 0.23(4)
v-21 (4)(1) 6b 39 <11 0.5(1) <0.30 0.3(1) 0.3(1) 8(1) 0.9(1) 0.24(4)
v-21 (11) 7 23 <19 1.0(1) 0.4(1) <0.77 0.7(1) 15(2) 1.0(1) 0.22(6)
v-21 (15) 8 14 <20 0.9(2) <0.85 0.7(2) 1.0(2) 20(3) 1.2(1) <0.32
v-21 (4)(1) 8b 36 <8 0.7(1) <0.36 0.3(1) 0.2(1) 10(1) 1.4(2) 0.17(4)
v-21 (15) 9 20 <14 0.57(9) <0.44 <0.79 0.5(1) 13(2) 0.81(8) 0.53(6)
v-21 (15) 10 26 <15 0.45(7) <0.44 0.5(1) 0.45(8) 13(2) 0.87(8) 0.60(6)
v-21 (15) 11 12 <10 0.5(2) <0.47 <1.17 0.6(1) 12(2) 0.77(9) 0.59(7)
v-21 (15) 12 18 <10 <0.71 <0.45 0.7(2) 0.7(1) 11(2) 0.78(8) 0.71(7)
v-21 (15) 13 34 <10 0.50(6) 0.38(8) 0.6(1) 0.50(7) 12(1) 0.82(7) 0.73(6)
v-21 (15) 14 21 <16 0.53(8) <0.41 <0.82 0.5(1) 14(2) 0.81(8) 0.63(6)
v-21 (15) 15 17 <15 0.54(9) <0.50 <0.74 0.6(1) 12(2) 0.81(8) 0.66(7)
v-21 (15) 16 19 <10 0.55(9) <0.38 0.8(2) 0.5(1) 10(1) 0.78(8) 0.52(6)
v-21 (12) 17 11 <17 0.6(2) <0.63 <1.1 0.7(1) 11(2) 0.76(9) 0.65(8)
v-21 (14) 18 13 <16 0.5(1) <0.59 0.6(2) 0.6(1) 11(2) 0.85(9) 0.62(7)
v-21 (14) 19 12 <14 0.6(1) <0.46 <1.1 0.6(1) 12(2) 0.76(9) 0.56(7)
123
Eur. Phys. J. C (2015) 75 :546 Page 9 of 10 546
Table 6 Screening results normalised per effective photocathode area of the PMT. These data are compared with results from older R11410 versions, as well as from other PMT types employed in liquid xenon detectors. The R8520 PMT is employed in XENON100[5] and PandaX
[14], the R8778 PMT is used in the LUX detector[8], and the R11410-10 version in the PandaX experiment[14]. The effective photocathode areas are 4.2cm2, 15.9cm2 and 32.2cm2 for the R8520, R8778 and R11410 tubes, respectively
PMT type Normalized activity [mBq/cm2] Ref.
238U 226Ra 228Th 235U 40K 60Co
R11410-21 <0.4 0.016(3) 0.012(3) 0.011(3) 0.37(6) 0.023(3) This work
R11410-20 <0.56 <0.03 0.028(6) <0.025 0.37(6) 0.040(6) This work
R11410-10 <3.0 <0.075 <0.08 <0.13 0.4(1) 0.11(2) [24]
R11410-10 (PandaX) <0.02 <0.02 0.04(4) 0.5(3) 0.11(1) [14]
R11410-10 (LUX) <0.19 <0.013 <0.009 <0.26 0.063(6) [25]
R11410 1.6(6) 0.19(2) 0.09(2) 0.10(2) 1.6(3) 0.26(2) [24]
R8778 (LUX) <1.4 0.59(4) 0.17(2) 4.1(1) 0.160(6) [25]
R8520 <0.33 0.029(2) 0.026(2) 0.009(2) 1.8(2) 0.13(1) [24]
1) Quartz: faceplate (PMT window)2) Aluminum: sealing3) Kovar: Co-free body4) Stainless steel: electrode disk5) Stainless steel: dynodes6) Stainless steel: shield7) Quartz: L-shaped insulation8) Kovar: flange of faceplate9) Ceramic: stem10) Kovar: flange of ceramic stem11) Getter
137
60
Cs
Co
40
228
K
Th
228
226
Ra
Ra
238
U
0 10 20 30 40 50 60 70 80 90 100
Relative contribution [%]
Fig. 5 Contribution of individual raw materials of the R11410-21 phototube to the total contamination in 238U, 226Ra, 228Ra, 228Th, 40K, 60Co
and 137Cs. The component numbers correspond to those in Fig.1 and Table 2
Inferred upper limits on the various isotopes are thus not used to calculate the relative contributions.
The main contribution to the isotopes in the 238U and
232Th chains clearly comes from the ceramic stem (no.9) of the PMT. In an attempt to reduce the 238U and 232Th activities, sapphire (no.12) was considered as a ceramic replacement. Our measurements show a reduction in 238U by more than a factor of three, which is, however, counteracted by a
226Ra level almost three times higher. The 228Ra contamination is reduced by less than a factor of two and the 228Th activity is barely lower than in the case of ceramic (see also Table3). Monte Carlo simulations using the sapphire screening results show that the neutron-induced background from the PMTs is reduced by less than 10%[36]. Therefore, such a replacement would not imply a signicant improvement in the overall background.
The isotope 137Cs has been detected only in the two Kovar alloy anges, namely the ange to the PMT faceplate (no.8) and the one to the stem (no.10) but its overall activity is very low, at the level of 10Bq/PMT.
In all Kovar alloy (no.3, 8, 10) and stainless steel (no.4, 5, 6) samples, 60Co has been detected, the main contribution coming from the Kovar anges in the PMT faceplate and stem (no.8 and 10). The total measured 60Co contamination in the tubes is at the level of 0.8mBq/PMT, while the sum of the components yields a range 0.610.66mBq/PMT. The discrepancy could be due to the metallic connections inside the tube that were not screened. A further reduction in the 60Co content could be achieved by substituting the material of the anges, which is Kovar alloy, with Co-free Kovar.
The main contributor to 40K is the ceramic stem. However, out of the 13mBq of 40K detected per tube, only about1.4mBq are located in the ceramic (no.9). Even taking into account the upper limits on this isotope from all other measured samples, a total upper limit on the 40K contamination of <3.1mBq results. This is signicantly lower than the contamination of the nal tube. We speculate that the additional
40K contamination is in the unmeasured cathode material as, in general, bialkali contain potassium. This assumption
123
546 Page 10 of 10 Eur. Phys. J. C (2015) 75 :546
is supported by the measurement of the individual getter taken from one phototube, which shows a higher concentration of 40K, namely 2.6 0.9mBq[37], compared to the
10 getter samples. The contamination of the getter could have happened during the deposition process of the photocathode material.
The average radioactive contaminations per PMT, or upper limits when no positive detection was available, were used in a full GEANT4-based Monte Carlo simulation of XENON1T to quantify the contribution to the background of the experiment. The simulation takes into account a detailed detector geometry, with the 248 R11410-21 PMTs placed at their correct positions in the TPC, above and below the liquid xenon target. The total electronic-recoil background from single-site gamma interactions in the medium is (0.014 0.002)events/y in a central, ducial volume con
taining 1 tonne of LXe. This value is calculated for a dark matter search region of 212keV electronic recoil energy and an assumed background discrimination level of 99.75%.
Although a contamination from the anthropogenic 110mAg has been found in most of the screened PMT batches, the expected contribution to the background of XENON1T by the start of the science run is negligible. This contribution makes 0.36% of the total electronic recoil background, without even considering the decay of 110mAg. It will be further reduced due to the relatively short half live of 249.8days of
110mAg.
The nuclear-recoil background from neutron interactions has also been investigated. We follow our procedure described in [11] to calculate the neutron spectra and yields, which are based on the SOURCES-4A code [38]. The estimated rate, taking into account the activity in each PMT material and the subsequent neutron production rates and spectra, is 0.060 0.015events/y in a 1 tonne ducial vol
ume in the 550keV region for nuclear-recoil energies with an acceptance of 50%[36]. The overall background goal of XENON1T is <1event for an exposure of 2ty. Thus, we
can safely conclude that the PMTs will not limit the sensitivity of the experiment.
Acknowledgments We gratefully acknowledge support from NSF, DOE, SNF, UZH, FCT, Region des Pays de la Loire, STCSM, BMBF, MPG, FOM, the Weizmann Institute of Science, EMG, INFN, and the ITN Invisibles (Marie Curie Actions, PITN-GA-2011-289442). We thank the technical and engineering personnel of Hamamatsu Photonics Co. for the fruitful cooperation in producing the ultra-pure photomultiplier tube. We are grateful to LNGS for hosting and supporting the XENON project.
Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/
Web End =http://creativecomm http://creativecommons.org/licenses/by/4.0/
Web End =ons.org/licenses/by/4.0/ ), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.
Funded by SCOAP3.
References
1. L. Baudis, Phys. Dark Univ. 1, 94 (2012). http://arxiv.org/abs/1211.7222
Web End =arXiv:1211.7222 2. L. Baudis, Phys. Dark Univ. 4, 50 (2014). http://arxiv.org/abs/1408.4371
Web End =arXiv:1408.4371 3. M. Schumann, EPJ Web Conf. 96, 01027 (2015). http://arxiv.org/abs/1501.01200
Web End =arXiv:1501.01200 4. T. Marrodn Undagoitia, PoS TIPP 2014, 011 (2014)5. E. Aprile et al., XENON100 Collaboration. Astropart. Phys. 35, 573 (2012). http://arxiv.org/abs/1107.2155
Web End =arXiv:1107.2155
6. E. Aprile et al., XENON100 Collaboration. Phys. Rev. Lett. 109, 181301 (2012). http://arxiv.org/abs/1207.5988
Web End =arXiv:1207.5988
7. E. Aprile et al., XENON100 Collaboration. Phys. Rev. Lett. 111, 021301 (2013). http://arxiv.org/abs/1301.6620
Web End =arXiv:1301.6620
8. D.S. Akerib et al., LUX Collaboration. Phys. Rev. Lett. 112, 091303 (2014). http://arxiv.org/abs/1310.8214
Web End =arXiv:1310.8214
9. E. Aprile et al., XENON1T Collaboration. http://arxiv.org/abs/1206.6288
Web End =arXiv:1206.6288 10. E. Aprile et al., XENON100 Collaboration. Phys. Rev. D 83 (2011) 082001 [Erratum-ibid. D 85 (2012) 029904]. http://arxiv.org/abs/1101.3866
Web End =arXiv:1101.3866
11. E. Aprile et al., XENON100 Collaboration. J. Phys. G 40, 115201 (2013). http://arxiv.org/abs/1306.2303
Web End =arXiv:1306.2303
12. D.S. Akerib et al., LUX Collaboration. Nucl. Instrum. Methods A 704, 111 (2013). http://arxiv.org/abs/1211.3788
Web End =arXiv:1211.3788
13. K. Abe et al., XMASS Collaboration. Phys. Lett. B 719, 78 (2013)14. X. Cao et al., Sci. China Phys. Mech. Astron. 57, 1476 (2014). http://arxiv.org/abs/1405.2882
Web End =arXiv:1405.2882
15. D.C. Malling, D.S. Akerib, H.M. Araujo et al., http://arxiv.org/abs/1110.0103
Web End =arXiv:1110.0103 16. L. Baudis, PoS IDM 2010, 122 (2011). http://arxiv.org/abs/1012.4764
Web End =arXiv:1012.4764 17. L. Baudis, DARWIN Consortium. J. Phys. Conf. Ser. 375, 012028 (2012). http://arxiv.org/abs/1201.2402
Web End =arXiv:1201.2402
18. M. Schumann, L. Baudis, L. Btikofer, A. Kish, M. Selvi., http://arxiv.org/abs/1506.08309
Web End =arXiv:1506.08309
19. D. Lorca et al., NEXT Collaboration. Nucl. Instrum. Methods A 718, 387 (2013). http://arxiv.org/abs/1210.4020
Web End =arXiv:1210.4020
20. M. Auger et al., JINST 7, P05010 (2012). http://arxiv.org/abs/1202.2192
Web End =arXiv:1202.2192 21. Hamamatsu, http://www.hamamatsu.com/
Web End =http://www.hamamatsu.com/ 22. K. Lung et al., Nucl. Instrum. Methods A 696, 32 (2012). http://arxiv.org/abs/1202.2628
Web End =arXiv:1202.2628
23. L. Baudis, A. Behrens, A. Ferella, A. Kish, T. Marrodn Undagoitia, D. Mayani, M. Schumann, JINST 8, 04026 (2013). http://arxiv.org/abs/1303.0226
Web End =arXiv:1303.0226
24. E. Aprile et al., XENON100 Collaboration. Astropart. Phys. 35, 43 (2011). http://arxiv.org/abs/1103.5831
Web End =arXiv:1103.5831
25. D.S. Akerib et al., Nucl. Instrum. Methods A 703, 1 (2013). http://arxiv.org/abs/1205.2272
Web End =arXiv:1205.2272
26. A. Lyashenko et al., JINST 9 11, P11021 (2014). http://arxiv.org/abs/1410.3890
Web End =arXiv:1410.3890 27. Hamamatsu, Datasheet R11410-21. http://www.hamamatsu.com/
Web End =http://www.hamamatsu.com/ 28. N. Neder, G. Heusser, M. Laubenstein, Appl. Radiat. Isot. 53, 19 (2000)
29. G. Heusser, M. Laubenstein, N. Neder, Proc. Intern. Conf. Isot. Environ. (2004)
30. D. Budjas, W. Hampel, M. Heisel, G. Heusser, M. Keillor, M. Laubenstein, W. Maneschg, G. Rugel et al., http://arxiv.org/abs/0812.0768
Web End =arXiv:0812.0768 31. L. Baudis, A.D. Ferella, A. Askin et al., JINST 6, P08010 (2011). http://arxiv.org/abs/1103.2125
Web End =arXiv:1103.2125
32. G.F. Knoll, Radiation Detection and Measurement (Wiley & Sons, New York, 2000)
33. C. Hurtgen, S. Jerome, M. Woods, Appl. Radiat. Isot. 53, 45 (2000)34. Evans Analytical Group, http://www.eaglabs.com/mc/glow-discharge-mass-spectrometry.html
Web End =http://www.eaglabs.com/mc/ http://www.eaglabs.com/mc/glow-discharge-mass-spectrometry.html
Web End =glow-discharge-mass-spectrometry.html
35. G. Heusser, Ann. Rev. Nucl. Part. Sci. 45, 543 (1995)36. E. Aprile et al., XENON1T Collaboration, to be published37. G. Heusser et al., http://arxiv.org/abs/1507.03319
Web End =arXiv:1507.03319 38. W.B. Wilson et al., SOURCES-4A. Technical Report. LA13539MS
123
You have requested "on-the-fly" machine translation of selected content from our databases. This functionality is provided solely for your convenience and is in no way intended to replace human translation. Show full disclaimer
Neither ProQuest nor its licensors make any representations or warranties with respect to the translations. The translations are automatically generated "AS IS" and "AS AVAILABLE" and are not retained in our systems. PROQUEST AND ITS LICENSORS SPECIFICALLY DISCLAIM ANY AND ALL EXPRESS OR IMPLIED WARRANTIES, INCLUDING WITHOUT LIMITATION, ANY WARRANTIES FOR AVAILABILITY, ACCURACY, TIMELINESS, COMPLETENESS, NON-INFRINGMENT, MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. Your use of the translations is subject to all use restrictions contained in your Electronic Products License Agreement and by using the translation functionality you agree to forgo any and all claims against ProQuest or its licensors for your use of the translation functionality and any output derived there from. Hide full disclaimer
SIF and Springer-Verlag Berlin Heidelberg 2015
Abstract
The low-background, VUV-sensitive 3-inch diameter photomultiplier tube R11410 has been developed by Hamamatsu for dark matter direct detection experiments using liquid xenon as the target material. We present the results from the joint effort between the XENON collaboration and the Hamamatsu company to produce a highly radio-pure photosensor (version R11410-21) for the XENON1T dark matter experiment. After introducing the photosensor and its components, we show the methods and results of the radioactive contamination measurements of the individual materials employed in the photomultiplier production. We then discuss the adopted strategies to reduce the radioactivity of the various PMT versions. Finally, we detail the results from screening 286 tubes with ultra-low background germanium detectors, as well as their implications for the expected electronic and nuclear recoil background of the XENON1T experiment.
You have requested "on-the-fly" machine translation of selected content from our databases. This functionality is provided solely for your convenience and is in no way intended to replace human translation. Show full disclaimer
Neither ProQuest nor its licensors make any representations or warranties with respect to the translations. The translations are automatically generated "AS IS" and "AS AVAILABLE" and are not retained in our systems. PROQUEST AND ITS LICENSORS SPECIFICALLY DISCLAIM ANY AND ALL EXPRESS OR IMPLIED WARRANTIES, INCLUDING WITHOUT LIMITATION, ANY WARRANTIES FOR AVAILABILITY, ACCURACY, TIMELINESS, COMPLETENESS, NON-INFRINGMENT, MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. Your use of the translations is subject to all use restrictions contained in your Electronic Products License Agreement and by using the translation functionality you agree to forgo any and all claims against ProQuest or its licensors for your use of the translation functionality and any output derived there from. Hide full disclaimer