- Email: [email protected]
The Majorana Demonstrator calibration system
Nuclear Inst. and Methods in Physics Research, A 872 (2017) 16–22
Contents lists available at ScienceDirect
Nuclear Inst. and Methods in Physics Research, A journal homepage: www.elsevier.com/locate/nima
The Majorana Demonstrator calibration system N. Abgrall a , I.J. Arnquist b , F.T. Avignone III c,d , A.S. Barabash e , F.E. Bertrand d , M. Boswell f , A.W. Bradley a , V. Brudanin g , M. Busch h,i , M. Buuck j , T.S. Caldwell k,i , C.D. Christofferson l , P.-H. Chu f , C. Cuesta j , J.A. Detwiler j , C. Dunagan l , Yu. Efremenko m , H. Ejiri n , S.R. Elliott f , Z. Fu j , V.M. Gehman f , T. Gilliss k,i , G.K. Giovanetti p , J. Goett f , M.P. Green o,i,d , J. Gruszko j , I.S. Guinn j , V.E. Guiseppe c , C.R. Haufe k,i , R. Henning k,i , E.W. Hoppe b , M.A. Howe k,i , B.R. Jasinski q , K.J. Keeter r , M.F. Kidd s , S.I. Konovalov e , R.T. Kouzes b , A.M. Lopez m , J. MacMullin k,i , R.D. Martin t , R. Massarczyk f, *, S.J. Meijer k,i , S. Mertens a , J.L. Orrell b , C. O’Shaughnessy k,i , A.W.P. Poon a , D.C. Radford d , J. Rager k,i , A.L. Reine k,i , K. Rielage f , R.G.H. Robertson j , B. Shanks k,i , M. Shirchenko g , A.M. Suriano l , D. Tedeschi c , J.E. Trimble k,i , R.L. Varner d , S. Vasilyev g , K. Vetter a,1 , K. Vorren k,i , B.R. White f , J.F. Wilkerson k,i,d , C. Wiseman c , W. Xu q , C.-H. Yu d , V. Yumatov e , I. Zhitnikov g , B.X. Zhu f a
Nuclear Science Division, Lawrence Berkeley National Laboratory, Berkeley, CA, USA Pacific Northwest National Laboratory, Richland, WA, USA c Department of Physics and Astronomy, University of South Carolina, Columbia, SC, USA d Oak Ridge National Laboratory, Oak Ridge, TN, USA e National Research Center ‘‘Kurchatov Institute’’ Institute for Theoretical and Experimental Physics, Moscow, Russia f Los Alamos National Laboratory, Los Alamos, NM, USA g Joint Institute for Nuclear Research, Dubna, Russia h Department of Physics, Duke University, Durham, NC, USA i Triangle Universities Nuclear Laboratory, Durham, NC, USA j Center for Experimental Nuclear Physics and Astrophysics, and Department of Physics, University of Washington, Seattle, WA, USA k Department of Physics and Astronomy, University of North Carolina, Chapel Hill, NC, USA l South Dakota School of Mines and Technology, Rapid City, SD, USA m Department of Physics and Astronomy, University of Tennessee, Knoxville, TN, USA n Research Center for Nuclear Physics and Department of Physics, Osaka University, Ibaraki, Osaka, Japan o Department of Physics, North Carolina State University, Raleigh, NC, USA p Department of Physics, Princeton University, Princeton, NJ, USA q Department of Physics, University of South Dakota, Vermillion, SD, USA r Department of Physics, Black Hills State University, Spearfish, SD, USA s Tennessee Tech University, Cookeville, TN, USA t Department of Physics, Engineering Physics and Astronomy, Queen’s University, Kingston, ON, Canada b
a r t i c l e
i n f o
Keywords: Neutrinoless double-beta decay Germanium detector Majorana Detector calibration
a b s t r a c t The Majorana Collaboration is searching for the neutrinoless double-beta decay of the nucleus 76 Ge. The Majorana Demonstrator is an array of germanium detectors deployed with the aim of implementing background reduction techniques suitable for a 1-ton 76 Ge-based search. The ultra low-background conditions require regular calibrations to verify proper function of the detectors. Radioactive line sources can be deployed around the cryostats containing the detectors for regular energy calibrations. When measuring in low-background mode, these line sources have to be stored outside the shielding so they do not contribute to the background. The deployment and the retraction of the source are designed to be controlled by the data acquisition system and do not require any direct human interaction. In this paper, we detail the design requirements and implementation
* Corresponding author. 1
E-mail address: [email protected] (R. Massarczyk). Department of Nuclear Engineering, University of California, Berkeley, CA, USA.
http://dx.doi.org/10.1016/j.nima.2017.08.005 Received 7 February 2017; Received in revised form 2 August 2017; Accepted 3 August 2017 Available online 8 August 2017 0168-9002/Published by Elsevier B.V.
N. Abgrall et al.
Nuclear Inst. and Methods in Physics Research, A 872 (2017) 16–22
of the calibration apparatus, which provides the event rates needed to define the pulse-shape cuts and energy calibration used in the final analysis as well as data that can be compared to simulations. Published by Elsevier B.V.
The choice of radioactive source material is determined by the physics goals of the Majorana Demonstrator. For this experiment, the dynamic range of the germanium detector and the attached read-out chain goes from sub keV up to 3 MeV in high gain channels and from about 10 keV to 10 MeV in low gain channels. Our region of interest is a 3-keV window at the 𝑄-value. Therefore, a 228 Th source is suitable for energy calibration using regions below and above the ROI. The source is used to provide a reliable energy calibration for low energies using Pb X-rays around 80 keV as well as at higher energies using the 208 Tl line at 2614 keV. The intensity and relative location of the full-energy peaks used for our regular calibration can be seen in Table 1. Due to the shielding of the copper cryostat, the X-rays themselves can only be seen by combining a number of calibration sets. This additional information is used as a cross check of the calibration curve at very low energies, so that the Majorana Demonstrator is, for example, able to perform searches for dark matter at low energies [11]. For the highest energy 𝛾-ray, we aim for an uncertainty in the number of events of less then 1%, so that fits of the full-energy peak shape allow an accurate determination of the energy. Adequate statistics during one calibration are determined by the number of events in the 228 Th decay chain peaks. Its 1.9-yr half-life is sufficient to provide enough statistics even through the end of MJD operations to periodically evaluate pulseshape cut efficiency and detector stability. It is expected that over the life-time of the Majorana Demonstrator the system will operate in a stable and reliable manner, so that the period between calibrations can be increased to a bi-weekly or monthly schedule. With this increased period the calibration time can then be adjusted in length to reflect the decay of the source. Since multiple energy depositions in a single detector (multi-site events) can sum to energies in our region of interest, we are making use of the excellent pulse shape discrimination abilities of PPC detectors to identify single and multi-site events [12–14]. Hence an additional requirement is the provision of enough data to test the efficiency of pulse-shape discrimination cuts in the final analysis. In addition to basic physics requirements, we are technically limited to a total event rate that can be handled by the front end electronics without excessive pile-up. The characteristic slow charge collection of point contact detectors – which makes pulse-shape discrimination possible – in conjunction with the resistive feedback of our front end, limits the allowed event rate to roughly 100 Hz per detector. Detectors on the outermost circumference of the array lie closer to any source deployed around the vacuum cryostat and they see a higher rate than those in the center. Therefore, a compromise must be struck between source activity, the position and extent of the source, and the time allotted for calibration runs. The modular design of the MJD cryostats in combination with the general construction approach have led us to adopt a line source that is deployed through the Demonstrator’s Pb and Cu shielding in a helical track that surrounds each cryostat module, see Figs. 1 and 2. Details of the low-radioactivity design can be found in Section 3.2. In such a design, the radioactive material is distributed along a cylindrical lineshaped container which can be moved along a track. The results of a Geant4 [15] simulation were used to determine the optimum source geometry and activity. The length of the radioactive source container itself was fixed by the manufacturer. Within the simulation framework MaGe [16,17] the source activity and the distance of the track to the cryostat were varied. Different choices of pitch angle for a source along a helical path around the cryostat were considered as well. Since the radial distance between cryostat and the calibration track tubing could not vary appreciably due to the close geometry of the
1. Introduction Neutrinoless double-beta decay (0𝜈𝛽𝛽) is a hypothesized yet unobserved second order process not permitted by the Standard Model. Such a second order weak process would violate lepton number conservation [1,2]. Completed searches to date and first results from running experiments [3–7] indicate half lives extending beyond 1025 years. If this process were to be observed experimentally, it would signal the Majorana nature of the neutrinos and indicate a violation of the lepton number. Through sphaleron processes in the early universe, the violation of baryon number would then result, a necessary condition and explanation for the present-day matter excess over antimatter in the universe [8]. The Majorana Demonstrator (MJD) [9] is a research and development effort aimed at deploying novel background reduction techniques. Such techniques are needed to build a 1-ton experiment with a projected background rate after analysis cuts of less than 1 count/(ROI ton⋅ year) at the Q-value of the 0𝜈𝛽𝛽 decay, 2039 keV. The width of the Region-ofInterest (ROI) is 3.1 keV [10]. This effort is accomplished by fielding an array of highly enriched p-type point contact (PPC) Ge detectors underground at the 4850 ft level of the Sanford Underground Research Facility with special attention to the radio-purity of materials in the environment surrounding the Ge detectors. The array is divided into two cryostats, each containing 7 strings of 4 or 5 detectors. Each of the two cryostats has its own vacuum and cooling system. These independent assemblies are referred to as Modules 1 and 2. Twenty nine kg of the detectors are made of material that is enriched to >87% in 76 Ge and 15 kg from natural germanium. In total 58 detector units are installed of which 35 contain enriched material. The detector masses vary between 0.5 and 1.0 kg and their sizes range from 3 to 6.5 cm in height and around 3–4 cm in radius. In this paper, we describe the design requirements and implementation of a system for calibrating the detectors in energy and providing high statistics samples for pulse-shape analysis.
2. Requirements and design The calibration system must provide events from a known radioactive source to each detector in the array in about an hour long data set. This ensures that regular calibrations do not significantly reduce the amount of live time needed for the physics program of the Demonstrator. The current MJD commissioning and data-taking plan foresees calibrations on a weekly basis. It allows observation of the performance of the pulse-shape analysis, the stability of the energy scale, gain, and energy resolution over certain periods of time. The gain of Ge detectors is generally stable, and can be monitored closely via use of an external pulser. We also monitor the low- and high voltage supplies of the detectors as well as temperatures in the electronic systems and the shield. So far no large fluctuations combined with gain shifts have been observed and the system is stable. However, periodic events like nitrogen fills can induce noise or small temperature drifts within the system. Since the whole system is very complex and uses a lot of non-standard electronics due to the radio-purity requirement, the weekly calibrations are used to cross check performance and ensure high-quality data taking. 17
N. Abgrall et al.
Nuclear Inst. and Methods in Physics Research, A 872 (2017) 16–22
Table 1 Overview on most important gamma lines used for MJD calibration. The energies of the single (SE) and double-escape (DE) peak of the 2614-keV line are given in italics. Since their line shape differs from the shape of a full-energy deposit they are not used for energy calibration. But both escape lines are used as a test for single-site and multi-site events. The given intensities are source intensities without detection efficiency due to the MJDdetector geometry and shielding. Energy (keV)
Isotope
Intensity per decay
238.63 240.99 277.36 300.09 583.19 727.33 785.37 860.56 1592.53 (DE) 2103.53 (SE) 2614.53
212
0.433 0.041 0.023 0.032 0.304 0.065 0.011 0.044
Pb 224 Ra 208 Tl 212 Pb 208 Tl 212 Bi 212 Bi 208 Tl 208 Tl 208 Tl 208 Tl
A sketch of the whole assembly is shown in Fig. 3. The radioactive source containers are placed within an outer Teflon™ container. The wall thickness of the outer containment is 0.25 mm and the total diameter is 4 mm. For positioning purposes (see Section 3.3), a series of 5-mm long NdFeB grade N42 magnetic slugs are embedded within the outer tubing, one pair at the front end of the source, and a second pair with a different spacing between the magnets at the trailing end, see Fig. 3. The positions of these magnets is sensed by Hall-effect devices located outside on the track. The position of the source containers and magnets are fixed within the whole source assembly using epoxy. Empty spaces between the radioactive parts and the two magnet pairs are filled with additional Teflon tubing of slightly smaller diameter to stabilize the assembly. The total as-built length of a line source assembly is about 4.7 m. This includes two one-meter long pieces with radioactive material and 2.7 m needed to push and pull the whole source assembly through the shield.
0.356
3.2. Implementation in the low-radioactivity design
overall design, the focus was set on the other two parameters. Results indicate that shielding and symmetry considerations produce roughly equivalent rates in the external detectors, and the driving consideration becomes the event rate in the central strings. Though the solid-angle exposure to the source is consistent for all detectors of an array, there is a greater shielding effect to the top and bottom center detectors. Variations in detector dimensions across the range of our available detector sizes had a negligible effect on event rate. For this analysis the double-escape peak of the 2614-keV line in 208 Tl is used which is part of the decay chain of 228 Th. To validate the pulse-shape cut efficiency on a regular basis, an-hour long calibration run should contain at least 400 events in this doubleescape peak (as a proxy for the single site analogue for 0𝜈𝛽𝛽). With the maximum rate in the outer detectors of around 100 Hz, the outcome of the simulation suggests a thorium source of a total activity of ∼10 kBq is needed so that the sufficient number of events can be reached in a calibration data set of roughly one hour in length. To understand the behavior and the relative number of single-site events in the Compton shoulder to full energy events, a 60 Co source should be available in the commissioning phase and can be exchanged between the two MJD modules at their respective commissioning times. The Cobalt source provides the advantage of a simple decay pattern with only two major 𝛾-ray energies so that Compton events can be easily identified. In Section 3.1 we discuss the implementation requirements of the line source within the Majorana geometry.
The MJD shield consists of several layers of different materials. Electro-formed copper (5-cm thick), commercial copper (5-cm thick), lead (45-cm thick) and poly (30-cm thick) complete the shielding and protect the detectors against radiation of natural origin [9]. An overview on how the track for the calibration sources passes these layers is given in Fig. 1. A radon exclusion box covers the inner three layers and is purged at all times with boiled-off nitrogen. Around this box, two layers of plastic scintillator are used as an active muon veto. The calibration source has to pass through all of these layers; its support structure is located outside the full shielding behind the poly wall, see Fig. 1. A line source has to be pushed and pulled through the shielding on a defined path without destroying the radon purge. Therefore, a track is needed that forms a closed volume passing through the different layers of shielding and the veto panels. The line source assembly moves in a purged track around the cryostat (35.6-cm diameter). This track is made of 1–2 in. wide polytetrafluoroethylene (PTFE) tubing. The materials used for the track and source assembly have been leached in agreement with the MJD cleanliness procedures [18]. During the cleaning steps, a leach removes surface contamination and insures against the case of abrasions of the tube material that no radioactive contamination remains close to the detectors. Parts of the material were assayed after leaching in the MJD assay program. The chosen PTFE3 has a very small natural radioactivity with mass fractions below 3.1⋅ 10−12 for 238 U and around 1.5⋅ 10−12 for 232 Th. Inside the shielding, a nitrogen purge line is attached at the inner end to the track and is actively purging the whole track at all times minimizing radon intrusion. The purge line has a smaller diameter than the source track in order to reduce material around the detectors. Its attachment to source track is done by a coupling made of electro-formed copper and can be seen in Fig. 4. To ensure a closed volume when not calibrating, the end of the track laying outside the shielding is closed by a pneumatic gate valve when no source is deployed. This valve is installed on the plate indicated in Fig. 5. The closed track volume, in combination with the active purge, prevents lab air from back-filling the calibration track. For control and safety reasons, two light ‘‘emitter and detector’’ pairs are placed on the outside of the gate valve. One pair is used to check the status of the gate valve. A closed valve prevents infrared light from reaching the emitter, thus a non-detection of light indicates a closedvalve. The second pair is used to verify that the radioactive line source is fully retracted behind the valve in its storage position. A source out of storage position lies in front of the infrared detector, so that no light or less light then usual is detected. A detection of light indicates that the source is fully retracted and it is safe to close the valve. This system is
3. Mechanical implementation 3.1. Radioactive line sources Each line source consists of a radioactively-doped epoxy injected into a 3-mm-diameter tube that is sealed at both ends, produced to our custom specifications by Eckert & Ziegler Analytics, Inc.2 The source activities are conveniently tuned at production; once cured, it can be treated as a sealed source within an containment of 0.1-mm wall thickness. Deviations from homogeneity along the length of the source are below 3%. Four 228 Th calibration sources were prepared in 1-m lengths, each with an integrated activity of 5.18± 0.30 kBq. A single 60 Co source for building the pulse-shape analysis (PSA) library was prepared with an integrated activity of 6.3± 0.30 kBq over a 2 m length. The length was chosen so that the container encapsulating the radioactive material is long enough to encircle one cryostat twice. In the final configuration, we fabricated 3 line-source assemblies: one that consists of only the 60 Co source and two assemblies each consisting of a pair of 228 Th sources with a total activity of 10.36 kBq. 2
3
http://www.ezag.com/home/—Rod Line sources. 18
http://www.coleparmer.com.
N. Abgrall et al.
Nuclear Inst. and Methods in Physics Research, A 872 (2017) 16–22
Fig. 1. Drawing of a Majorana module assembly with the calibration system at final position in the Demonstrator design including the vacuum system, nitrogen supply as well as the shielding layers. Copper shielding is shown in brown, lead bricks in dark gray and the poly shield in purple. Not all veto panels are shown for better visibility. The loops of the track as well as the calibration track through the shield and the inner loops around the cryostats are visible in cyan. The calibration controls are located on the table next to the poly-shield under the cover next to the loops of the outer storage track. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
As mentioned this emitter–detector system works as a safety back-up. In the event of a power outage the system is not able to move and the source stays in position. The gate valve has to be actively supplied with pressurized air and power to change status so that a loss of air pressure or power would only results in a steady state; it would not close and cut a source. A logic device4 is placed in the power supply of the gate valve. The device lets power pass to the gate valve after it receives positive feedback from the Arduino checking the source status with the second infrared sensor pair. In the case of a broken sensor pair, no light is detected resulting in the feedback that the source is deployed and neglecting the request of closing the valve. At all times the outer part of the system behind the poly wall is easily accessible for trouble shooting and the end of the line source assembly can always be used to pull out a source by hand, see Fig. 1.
Fig. 2. Close-up of the drawing in Fig. 1 which shows the relative location of the detector array in one module, the surrounding calibration track, and the cross-arm connection to vacuum and cooling systems towards the left. The detector positions can be seen within the cryostat around which the line source assembly wraps twice. At the inner end of the 1/2-inch track, the smaller 1/4-inch purge line is coupled. The shielding surrounding the detectors and through which the calibration track is guided is removed in this drawing for better visibility.
3.3. Source positioning The source assembly is guided on the outside between two drive rollers, see Fig. 6. One of the drive rollers is rotated by a motor. The second drive roller passively rotates on an axle below the first one. The axle is pulled by two springs against the actively rotating drive roller so that the line source is gently clamped between the two drive rollers. This ensures a constant grip of the drive rollers on the line source so that the source can be pushed and pulled reliably over hundreds of deployments and retractions. Inside the inner copper shielding, the track is wrapped in a helical shape around the cryostat. The tubing is held in place by clean electroformed copper brackets which are attached to the cryostat. Fig. 4 shows the as-built configuration of the track around one module before it was moved into the shielding. For a deployed source the part containing the radioactive material wraps twice around the cryostat allowing a simultaneous calibration of all detectors, in contrast to the calibration system of Gerda [19]. There, three collimated sources can be lowered into the cryostat containing the detectors and calibrate the array one layer of detectors at a time.
Fig. 3. Sketch of the line source assembly. The outer Teflon™ container (blue) has a wall thickness of 0.25 mm. One pair of NdFeB magnets (black) is located at both ends of the line source. Two of the 1-m long container with 228 Th-material from Eckert & Ziegler (red) with radioactive material are located next to each other. Empty space is filled with Teflon™ material of thinner diameter to stabilize the source (gray). The left end of the source would be pointing towards the detectors and is the part the encircles one cryostat. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
used as a back-up to the magnet detection system. An improper closing of the valve could damage the thin-walled line source and in the worst case lead to a contamination of radioactive material.
4
19
http://dataprobe.com/ibootbar/.
N. Abgrall et al.
Nuclear Inst. and Methods in Physics Research, A 872 (2017) 16–22
Fig. 5. Schematic overview of the sensors installed at the angle plate of the calibration system. The inner calibration track would be situated on the right side of the image.
Fig. 4. Inner calibration track around Module 1. The 1/4-inch-purge line is attached to the 1/2-inch calibration track.
Since the detector rates at Majorana are heavily dependent on how far into the helical track a source is deployed the system should reproduce its previous position as precisely as possible to have comparable data sets. To reproduce the source position reliably, we have developed a positioning system using Hall effect switches.5 Three magnetic sensors, see Fig. 5, are connected to an Arduino UNO.6 These sensors register a passing magnet in a moving line source via the Hall effect. By reading their signal continuously, the arrangement of three sensors outside and two magnets at the front or the back end of the assembly allows us to determine the status of the source and reproduce its deployed position within a few millimeters. A special sequence in the readout of the magnetic sensors gives feedback to the experiment control if the source is deployed, in deployment, retracted, or in retraction. If not deployed, sources are stored in mirror tracks behind the last magnet sensor at the outer end of the calibration system plate. This mirror track, made of the same PTFE tubing as the other parts of the track, is wound similarly to the track inside the shield, thereby ‘‘mirroring’’ the shape of the deployed source as it is stored. It is possible to access this part of the calibration system and to exchange the whole mirror track with the source inside if another calibration source needs to be used. In the final configuration the 228 Th is installed as the standard source and can be exchanged with the 60 Co line source if additional commissioning is needed. 3.4. System controls The sensors attached to the Arduino controller and the motor can be controlled individually via the data acquisition (DAQ) software ORCA [20] in an expert mode which allows debugging and tests. Nonexpert users can use an interface with reduced functionality on the DAQ machines. This interface allows routine calibration operations. The reduced interface controls for example the run-bit settings which indicate run-type and lab conditions, e.g. a calibration run with Thorium. This interface also takes care of deployment, calibration time, and retraction of the sources and allows operators an easy remote control of the system.
Fig. 6. Radioactive line source between the two drive rollers. Two magnet sensors are placed in front of the drive roller (green boxes, left side in the upper figure). One magnet sensor (blue box) is located between the drive roller and the gate valve on the right. In the lower figure can be seen the springs that pull the idler pulley against the drive pulley above. In both figures the shielding and the cryostat are on the right side of the figures. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
through the lead shield to the detectors. While going through the outer layers of lead, the track on which the line source assembly moves is curved. The outer mirror track was designed so that the source would not lie in this line-of-sight when stored. It is coiled vertically so that the source in storage is located below the cross-arm ensuring that a calibration source in storage does not contribute to the background budget (<3 counts/(ROI ton⋅ year)) of the Majorana Demonstrator [9].
3.5. Interference and background contribution The copper cross-arm of the cryostat delivers the pumping and cooling power to the cryostat, but also provides a direct line-of-sight 5 6
Infineon TLE4905L E6433. https://www.arduino.cc/en/Main/Products. 20
N. Abgrall et al.
Nuclear Inst. and Methods in Physics Research, A 872 (2017) 16–22
An upper estimate of 1.3⋅ 10−3 counts/(ROI ton⋅ year) is indicated by simulations [21]. Tests with stronger button sources at certain locations behind the shield have verified that a source in the outer mirror track does not contribute to the background. We considered the possibility of removable contamination from sources being carried into or left inside the tube next to the cryostat. The mirror tracks and the cover of the system act as a barrier so that direct contact to the sources due to ongoing lab work is avoided and the line sources stay clean. In addition, the source assemblies, parts of the track, and drive rollers are wiped on a regular basis and have not shown abrasion or any signs of dirt. During the last year of running we have not found any increase of backgrounds attributable to the calibration source. Fig. 7. Average detector energy spectrum of Module 1. The spectrum is averaged over the number of detectors and taken with a calibration time of a one hour-long data set. The spectra are shown for the detectors biased at the time of the measurement. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
4. Commissioning and performance During commissioning, a stress test of the system was performed. Calibration sources were deployed and retracted about a hundred times without incident. Assuming a weekly calibration over the expected 5year lifetime of the Majorana Demonstrator the number of operational cycles is of the same order. The software was tested and different failure or mishandling modes were investigated and mitigated. Fig. 7 is taken from a data set from 2015 in which the first module was situated inside the shielding. In this data set, detectors in Module 1 were calibrated using line sources that provide a spectrum with several peaks at different energies. Fig. 8 shows the integral count rate for energies between 20 and 4000 keV. The average count rate for calibrations with the 60 Co and the 228 Th sources is around 40 Hz and 30 Hz, respectively. This fulfills the 100-Hz upper count rate limit. The figure shows that several detectors closer to the line source path have higher rates, as expected. Detectors on the inside or further away from the path have slightly lower rates. However, the built geometry achieved the goal to have a balanced count rate within the array. Fig. 9 shows the spectra for the detector with the highest count rate and for the detector with the lowest count rate. A deployed source around one module can be seen by detectors in the second cryostat. The recorded spectra in these detectors can be used to test stability. However, the count rate is not sufficient for a full calibration. For a full calibration data set of one module, a dedicated deployment of the corresponding source is necessary. In addition, it is possible to deploy both sources at the same time for high-rate data tests. Calibration measurements are also used as a reference for our simulation campaign. The simulations using the Geant4-based framework MaGe [16,17] are compared to measurements in order to validate the implementation of the geometry. Various parameters like the heightratio between the full-energy peaks, the full-energy peak to Compton continuum ratio and multiplicity distributions are affected by the positioning of the source relative to the detectors as well as by the amount of material located between source and detectors. The simulation creates gamma-rays along a continuous source of 2-m length in the deployed position. The simulation geometry includes the track, the full copper cryostat, structural material, shielding, and a thin ‘‘dead layer’’. Each detector has this thin layer of inactive (or partially active) material, the thickness of which affects the size of the step function under a fullenergy peak due to small energy losses of incoming gamma-rays. The adjustment of this layer size in the simulations to realistic values is necessary to get a similar spectral shape in experiment and simulation. A first analysis showed that MaGe is able to describe the measured spectra in calibration runs, as shown in Fig. 10. The simulation is in very good agreement and can be used to build up a pulse-shape library, to study the efficiency of cuts and the background.
Fig. 8. Integral count rate per detector in the high gain channel between 20 and 4000 keV for the two installed line sources. The count rate is given for the detectors used for the first data set of Module 1. Only data points for detectors for which a cobalt and a thorium measurement are available are shown in this plot. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
Fig. 9. Single-detector energy spectrum of Module 1 for the detector with the highest count rate and the detector with the lowest count rate. The count rates refer to the data set that is used to calculate the numbers in Fig. 8. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
5. Summary A calibration system has been designed for the Majorana Demonstrator. The system fulfills special requirements in terms of cleanliness, stable positioning and reproducibility. The strength of the radioactive source was adjusted to the special needs of the data stream and ensures stable data taking. Two systems, one for each cryostat in the shield, were built. Both systems are clones and can be used independently or in tandem. The systems deploy radioactive sources from outside the 21
N. Abgrall et al.
Nuclear Inst. and Methods in Physics Research, A 872 (2017) 16–22
from the Russian Foundation for Basic Research, grant No. 15-02-02919. We acknowledge the support of the US Department of Energy through the LANL/LDRD Program. This research used resources of the Oak Ridge Leadership Computing Facility, which is a DOE Office of Science User Facility supported under Contract DE-AC05-00OR22725. This research used resources of the National Energy Research Scientific Computing Center, a DOE Office of Science User Facility supported under Contract No. DE-AC02-05CH11231. We thank our hosts and colleagues at the Sanford Underground Research Facility for their support. References [1] F.T. Avignone III, et al., Double beta decay, Majorana neutrinos, and neutrino mass, Rev. Modern Phys. 80 (2008) 481–516 URL http://link.aps.org/doi/10.1103/ RevModPhys.80.481. [2] J.D. Vergados, et al., Theory of neutrinoless double-beta decay, Rep. Progr. Phys. 75 (10) (2012) 106301 URL http://stacks.iop.org/0034-4885/75/i=10/a= 106301. [3] H.V. Klapdor-Kleingrothaus, I.V. Krivosheina, The evidence for observation of 0𝜈𝛽𝛽 decay: the identification of 0𝜈𝛽𝛽 events from the full spectra, Modern Phys. Lett. A 21 (20) (2006) 1547–1566 URL https://doi.org/10.1142/S0217732306020937. [4] EXO-200-Collaboration, Search for Majorana neutrinos with the first two years of EXO-200 data, Nature (ISSN: 0028-0836) 510 (7504) (2014) 229–234 URL http: //dx.doi.org/10.1038/nature13432. [5] K. Alfonso, et al., CUORE Collaboration Collaboration, Search for Neutrinoless Double-Beta Decay of 130 Te with CUORE-0, Phys. Rev. Lett. 115 (2015) 102502 URL https://link.aps.org/doi/10.1103/PhysRevLett.115.102502. [6] M. Agostini, et al., Search of neutrinoless double beta decay with the Gerda Experiment, Nucl. Part. Phys. Proc. (ISSN: 2405-6014) 273–275 (2016) 1876–1882 URL http://www.sciencedirect.com/science/article/pii/S2405601415007920 37th International Conference on High Energy Physics (ICHEP). [7] A. Gando, et al., Search for Majorana neutrinos near the inverted mass hierarchy region with KamLAND-Zen, Phys. Rev. Lett. 117 (2016) 082503 URL http://link. aps.org/doi/10.1103/PhysRevLett.117.082503. [8] A.G. Cohen, D.B. Kaplan, A.E. Nelson, Progress in electroweak baryogenesis, Annu. Rev. Nucl. Part. Sci. 43 (1993) 27–70. [9] N. Abgrall, et al., The Majorana demonstrator neutrinoless double-beta decay experiment, Adv. High Energy Phys. 2014 (2014) 1–18 URL http://dx.doi.org/10.1155/ 2014/365432. [10] S. Elliott, et al., Initial results from the MAJORANA DEMONSTRATOR, arxiv 1610 (2016) 01210. URL https://arxiv.org/abs/1610.01210. [11] N. Abgrall, et al., New limits on bosonic dark matter, solar axions, pauli exclusion principle violation, and electron decay from the Majorana Demonstrator, Phys. Rev. Lett. 118 (2017) 161801 URL https://link.aps.org/doi/10.1103/PhysRevLett.118. 161801. [12] S. Elliott, V. Gehman, K. Kazkaz, D.-M. Mei, A. Young, Pulse shape analysis in segmented detectors as a technique for background reduction in Ge double-beta decay experiments, Nucl. Instrum. Methods Phys. Res. Sect. A 558 (2006) 504–510 URL http://www.sciencedirect.com/science/article/pii/S0168900205023867. [13] R. Cooper, et al., A pulse shape analysis technique for the Majorana experiment, Nucl. Instrum. Methods Phys. Res. Sect. A (ISSN: 0168-9002) 629 (1) (2011) 303–310 URL http://www.sciencedirect.com/science/article/pii/S0168900210024915. [14] S. Mertens, et al., Majorana collaboration’s experience with germanium detectors, J. Phys. Conf. Ser. 606 (1) (2015) 012005 URL http://stacks.iop.org/1742-6596/606/ i=1/a=012005. [15] S. Agostinelli, et al., Geant4 a simulation toolkit, Nucl. Instrum. Methods Phys. Res. Sect. A (ISSN: 0168-9002) 506 (3) (2003) 250–303 URL http://www.sciencedirect. com/science/article/pii/S0168900203013688. [16] M. Bauer, et al., MaGe: a Monte Carlo framework for the Gerda and Majorana double beta decay experiments, J. Phys. Conf. Ser. 39 (1) (2006) 362 URL http: //stacks.iop.org/1742-6596/39/i=1/a=097. [17] M. Boswell, et al., MaGe - a Geant4-based Monte Carlo Application Framework for Low-background Germanium Experiments, 2010. ArXiv Preprint Nucl-Ex URL https://arxiv.org/abs/1011.3827v1. [18] N. Abgrall, et al., The Majorana demonstrator radioassay program, Nucl. Instrum. Methods Phys. Res. Sect. A (ISSN: 0168-9002) 828 (2016) 22–36 URL http://www. sciencedirect.com/science/article/pii/S0168900216302832. [19] L. Baudis, et al., Monte Carlo studies and optimization for the calibration system of the Gerda experiment, Nucl. Instrum. Methods A 729 (2013) 557–564 URL http://www.sciencedirect.com/science/article/pii/S0168900213011273. [20] M.A. Howe, et al., Sudbury neutrino observatory neutral current detector acquisition software overview, IEEE Trans. Nucl. Sci. 51 (2004) 878–883. [21] J. MacMullin, Validation of the Background Model for the MAJORANA DEMONSTRATOR, Ph.D. thesis, University of North Carolina, 2015.
Fig. 10. (top) Comparison of a 228 Th line source simulation using MaGe and a measurement of Module 1. The simulated distribution was normalized by matching the integrals of both curves in the range from 2595 keV to 2635 keV. (bottom) Residual between simulation and experiment for the 1-keV binning (blue) averaged over a 10-keV wide window (red). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
shielding to a position next to the detectors so that energy calibrations and other studies are possible. In contrast to other systems, our line source approach can calibrate the entire array at the same time without moving the source. The results can be used to study the agreement of observed energy distributions and multiplicities in experiment and simulation. Regular calibrations allow the study of the stability of data cleaning cuts over a long time. The design ensures a sufficient count rate in all detectors of the array so that one calibration run can be used for all detectors of a module. Also, it has been shown that the calibration system around one module can easily be copied, which is important for the scalability of such a modular approach when going to ton-scale experiments. Both systems are operated remotely and the controls are implemented in the DAQ system of the Majorana Demonstrator. While the exact design depends on the future experimental conditions, the MJD calibration system shows that it is possible to construct a system using easily accessible materials, sensors and readout electronics while still holding to all cleanliness and low-background limits. Acknowledgments This material is based upon work supported by the US Department of Energy, Office of Science, Office of Nuclear Physics under Award Numbers DE-AC02-05CH11231, DE-AC52-06NA25396, DE-FG02-97ER41041, DE-FG02-97ER41033, DE-FG02-97ER41042, DESC0012612, DE-FG02-10ER41715, DE-SC0010254, and DE-FG0297ER41020. We acknowledge support from the Particle Astrophysics Program and Nuclear Physics Program of the National Science Foundation through grant numbers PHY-0919270, PHY-1003940, 0855314, PHY-1202950, MRI 0923142 and 1003399. We acknowledge support
22
Contents lists available at ScienceDirect
Nuclear Inst. and Methods in Physics Research, A journal homepage: www.elsevier.com/locate/nima
The Majorana Demonstrator calibration system N. Abgrall a , I.J. Arnquist b , F.T. Avignone III c,d , A.S. Barabash e , F.E. Bertrand d , M. Boswell f , A.W. Bradley a , V. Brudanin g , M. Busch h,i , M. Buuck j , T.S. Caldwell k,i , C.D. Christofferson l , P.-H. Chu f , C. Cuesta j , J.A. Detwiler j , C. Dunagan l , Yu. Efremenko m , H. Ejiri n , S.R. Elliott f , Z. Fu j , V.M. Gehman f , T. Gilliss k,i , G.K. Giovanetti p , J. Goett f , M.P. Green o,i,d , J. Gruszko j , I.S. Guinn j , V.E. Guiseppe c , C.R. Haufe k,i , R. Henning k,i , E.W. Hoppe b , M.A. Howe k,i , B.R. Jasinski q , K.J. Keeter r , M.F. Kidd s , S.I. Konovalov e , R.T. Kouzes b , A.M. Lopez m , J. MacMullin k,i , R.D. Martin t , R. Massarczyk f, *, S.J. Meijer k,i , S. Mertens a , J.L. Orrell b , C. O’Shaughnessy k,i , A.W.P. Poon a , D.C. Radford d , J. Rager k,i , A.L. Reine k,i , K. Rielage f , R.G.H. Robertson j , B. Shanks k,i , M. Shirchenko g , A.M. Suriano l , D. Tedeschi c , J.E. Trimble k,i , R.L. Varner d , S. Vasilyev g , K. Vetter a,1 , K. Vorren k,i , B.R. White f , J.F. Wilkerson k,i,d , C. Wiseman c , W. Xu q , C.-H. Yu d , V. Yumatov e , I. Zhitnikov g , B.X. Zhu f a
Nuclear Science Division, Lawrence Berkeley National Laboratory, Berkeley, CA, USA Pacific Northwest National Laboratory, Richland, WA, USA c Department of Physics and Astronomy, University of South Carolina, Columbia, SC, USA d Oak Ridge National Laboratory, Oak Ridge, TN, USA e National Research Center ‘‘Kurchatov Institute’’ Institute for Theoretical and Experimental Physics, Moscow, Russia f Los Alamos National Laboratory, Los Alamos, NM, USA g Joint Institute for Nuclear Research, Dubna, Russia h Department of Physics, Duke University, Durham, NC, USA i Triangle Universities Nuclear Laboratory, Durham, NC, USA j Center for Experimental Nuclear Physics and Astrophysics, and Department of Physics, University of Washington, Seattle, WA, USA k Department of Physics and Astronomy, University of North Carolina, Chapel Hill, NC, USA l South Dakota School of Mines and Technology, Rapid City, SD, USA m Department of Physics and Astronomy, University of Tennessee, Knoxville, TN, USA n Research Center for Nuclear Physics and Department of Physics, Osaka University, Ibaraki, Osaka, Japan o Department of Physics, North Carolina State University, Raleigh, NC, USA p Department of Physics, Princeton University, Princeton, NJ, USA q Department of Physics, University of South Dakota, Vermillion, SD, USA r Department of Physics, Black Hills State University, Spearfish, SD, USA s Tennessee Tech University, Cookeville, TN, USA t Department of Physics, Engineering Physics and Astronomy, Queen’s University, Kingston, ON, Canada b
a r t i c l e
i n f o
Keywords: Neutrinoless double-beta decay Germanium detector Majorana Detector calibration
a b s t r a c t The Majorana Collaboration is searching for the neutrinoless double-beta decay of the nucleus 76 Ge. The Majorana Demonstrator is an array of germanium detectors deployed with the aim of implementing background reduction techniques suitable for a 1-ton 76 Ge-based search. The ultra low-background conditions require regular calibrations to verify proper function of the detectors. Radioactive line sources can be deployed around the cryostats containing the detectors for regular energy calibrations. When measuring in low-background mode, these line sources have to be stored outside the shielding so they do not contribute to the background. The deployment and the retraction of the source are designed to be controlled by the data acquisition system and do not require any direct human interaction. In this paper, we detail the design requirements and implementation
* Corresponding author. 1
E-mail address: [email protected] (R. Massarczyk). Department of Nuclear Engineering, University of California, Berkeley, CA, USA.
http://dx.doi.org/10.1016/j.nima.2017.08.005 Received 7 February 2017; Received in revised form 2 August 2017; Accepted 3 August 2017 Available online 8 August 2017 0168-9002/Published by Elsevier B.V.
N. Abgrall et al.
Nuclear Inst. and Methods in Physics Research, A 872 (2017) 16–22
of the calibration apparatus, which provides the event rates needed to define the pulse-shape cuts and energy calibration used in the final analysis as well as data that can be compared to simulations. Published by Elsevier B.V.
The choice of radioactive source material is determined by the physics goals of the Majorana Demonstrator. For this experiment, the dynamic range of the germanium detector and the attached read-out chain goes from sub keV up to 3 MeV in high gain channels and from about 10 keV to 10 MeV in low gain channels. Our region of interest is a 3-keV window at the 𝑄-value. Therefore, a 228 Th source is suitable for energy calibration using regions below and above the ROI. The source is used to provide a reliable energy calibration for low energies using Pb X-rays around 80 keV as well as at higher energies using the 208 Tl line at 2614 keV. The intensity and relative location of the full-energy peaks used for our regular calibration can be seen in Table 1. Due to the shielding of the copper cryostat, the X-rays themselves can only be seen by combining a number of calibration sets. This additional information is used as a cross check of the calibration curve at very low energies, so that the Majorana Demonstrator is, for example, able to perform searches for dark matter at low energies [11]. For the highest energy 𝛾-ray, we aim for an uncertainty in the number of events of less then 1%, so that fits of the full-energy peak shape allow an accurate determination of the energy. Adequate statistics during one calibration are determined by the number of events in the 228 Th decay chain peaks. Its 1.9-yr half-life is sufficient to provide enough statistics even through the end of MJD operations to periodically evaluate pulseshape cut efficiency and detector stability. It is expected that over the life-time of the Majorana Demonstrator the system will operate in a stable and reliable manner, so that the period between calibrations can be increased to a bi-weekly or monthly schedule. With this increased period the calibration time can then be adjusted in length to reflect the decay of the source. Since multiple energy depositions in a single detector (multi-site events) can sum to energies in our region of interest, we are making use of the excellent pulse shape discrimination abilities of PPC detectors to identify single and multi-site events [12–14]. Hence an additional requirement is the provision of enough data to test the efficiency of pulse-shape discrimination cuts in the final analysis. In addition to basic physics requirements, we are technically limited to a total event rate that can be handled by the front end electronics without excessive pile-up. The characteristic slow charge collection of point contact detectors – which makes pulse-shape discrimination possible – in conjunction with the resistive feedback of our front end, limits the allowed event rate to roughly 100 Hz per detector. Detectors on the outermost circumference of the array lie closer to any source deployed around the vacuum cryostat and they see a higher rate than those in the center. Therefore, a compromise must be struck between source activity, the position and extent of the source, and the time allotted for calibration runs. The modular design of the MJD cryostats in combination with the general construction approach have led us to adopt a line source that is deployed through the Demonstrator’s Pb and Cu shielding in a helical track that surrounds each cryostat module, see Figs. 1 and 2. Details of the low-radioactivity design can be found in Section 3.2. In such a design, the radioactive material is distributed along a cylindrical lineshaped container which can be moved along a track. The results of a Geant4 [15] simulation were used to determine the optimum source geometry and activity. The length of the radioactive source container itself was fixed by the manufacturer. Within the simulation framework MaGe [16,17] the source activity and the distance of the track to the cryostat were varied. Different choices of pitch angle for a source along a helical path around the cryostat were considered as well. Since the radial distance between cryostat and the calibration track tubing could not vary appreciably due to the close geometry of the
1. Introduction Neutrinoless double-beta decay (0𝜈𝛽𝛽) is a hypothesized yet unobserved second order process not permitted by the Standard Model. Such a second order weak process would violate lepton number conservation [1,2]. Completed searches to date and first results from running experiments [3–7] indicate half lives extending beyond 1025 years. If this process were to be observed experimentally, it would signal the Majorana nature of the neutrinos and indicate a violation of the lepton number. Through sphaleron processes in the early universe, the violation of baryon number would then result, a necessary condition and explanation for the present-day matter excess over antimatter in the universe [8]. The Majorana Demonstrator (MJD) [9] is a research and development effort aimed at deploying novel background reduction techniques. Such techniques are needed to build a 1-ton experiment with a projected background rate after analysis cuts of less than 1 count/(ROI ton⋅ year) at the Q-value of the 0𝜈𝛽𝛽 decay, 2039 keV. The width of the Region-ofInterest (ROI) is 3.1 keV [10]. This effort is accomplished by fielding an array of highly enriched p-type point contact (PPC) Ge detectors underground at the 4850 ft level of the Sanford Underground Research Facility with special attention to the radio-purity of materials in the environment surrounding the Ge detectors. The array is divided into two cryostats, each containing 7 strings of 4 or 5 detectors. Each of the two cryostats has its own vacuum and cooling system. These independent assemblies are referred to as Modules 1 and 2. Twenty nine kg of the detectors are made of material that is enriched to >87% in 76 Ge and 15 kg from natural germanium. In total 58 detector units are installed of which 35 contain enriched material. The detector masses vary between 0.5 and 1.0 kg and their sizes range from 3 to 6.5 cm in height and around 3–4 cm in radius. In this paper, we describe the design requirements and implementation of a system for calibrating the detectors in energy and providing high statistics samples for pulse-shape analysis.
2. Requirements and design The calibration system must provide events from a known radioactive source to each detector in the array in about an hour long data set. This ensures that regular calibrations do not significantly reduce the amount of live time needed for the physics program of the Demonstrator. The current MJD commissioning and data-taking plan foresees calibrations on a weekly basis. It allows observation of the performance of the pulse-shape analysis, the stability of the energy scale, gain, and energy resolution over certain periods of time. The gain of Ge detectors is generally stable, and can be monitored closely via use of an external pulser. We also monitor the low- and high voltage supplies of the detectors as well as temperatures in the electronic systems and the shield. So far no large fluctuations combined with gain shifts have been observed and the system is stable. However, periodic events like nitrogen fills can induce noise or small temperature drifts within the system. Since the whole system is very complex and uses a lot of non-standard electronics due to the radio-purity requirement, the weekly calibrations are used to cross check performance and ensure high-quality data taking. 17
N. Abgrall et al.
Nuclear Inst. and Methods in Physics Research, A 872 (2017) 16–22
Table 1 Overview on most important gamma lines used for MJD calibration. The energies of the single (SE) and double-escape (DE) peak of the 2614-keV line are given in italics. Since their line shape differs from the shape of a full-energy deposit they are not used for energy calibration. But both escape lines are used as a test for single-site and multi-site events. The given intensities are source intensities without detection efficiency due to the MJDdetector geometry and shielding. Energy (keV)
Isotope
Intensity per decay
238.63 240.99 277.36 300.09 583.19 727.33 785.37 860.56 1592.53 (DE) 2103.53 (SE) 2614.53
212
0.433 0.041 0.023 0.032 0.304 0.065 0.011 0.044
Pb 224 Ra 208 Tl 212 Pb 208 Tl 212 Bi 212 Bi 208 Tl 208 Tl 208 Tl 208 Tl
A sketch of the whole assembly is shown in Fig. 3. The radioactive source containers are placed within an outer Teflon™ container. The wall thickness of the outer containment is 0.25 mm and the total diameter is 4 mm. For positioning purposes (see Section 3.3), a series of 5-mm long NdFeB grade N42 magnetic slugs are embedded within the outer tubing, one pair at the front end of the source, and a second pair with a different spacing between the magnets at the trailing end, see Fig. 3. The positions of these magnets is sensed by Hall-effect devices located outside on the track. The position of the source containers and magnets are fixed within the whole source assembly using epoxy. Empty spaces between the radioactive parts and the two magnet pairs are filled with additional Teflon tubing of slightly smaller diameter to stabilize the assembly. The total as-built length of a line source assembly is about 4.7 m. This includes two one-meter long pieces with radioactive material and 2.7 m needed to push and pull the whole source assembly through the shield.
0.356
3.2. Implementation in the low-radioactivity design
overall design, the focus was set on the other two parameters. Results indicate that shielding and symmetry considerations produce roughly equivalent rates in the external detectors, and the driving consideration becomes the event rate in the central strings. Though the solid-angle exposure to the source is consistent for all detectors of an array, there is a greater shielding effect to the top and bottom center detectors. Variations in detector dimensions across the range of our available detector sizes had a negligible effect on event rate. For this analysis the double-escape peak of the 2614-keV line in 208 Tl is used which is part of the decay chain of 228 Th. To validate the pulse-shape cut efficiency on a regular basis, an-hour long calibration run should contain at least 400 events in this doubleescape peak (as a proxy for the single site analogue for 0𝜈𝛽𝛽). With the maximum rate in the outer detectors of around 100 Hz, the outcome of the simulation suggests a thorium source of a total activity of ∼10 kBq is needed so that the sufficient number of events can be reached in a calibration data set of roughly one hour in length. To understand the behavior and the relative number of single-site events in the Compton shoulder to full energy events, a 60 Co source should be available in the commissioning phase and can be exchanged between the two MJD modules at their respective commissioning times. The Cobalt source provides the advantage of a simple decay pattern with only two major 𝛾-ray energies so that Compton events can be easily identified. In Section 3.1 we discuss the implementation requirements of the line source within the Majorana geometry.
The MJD shield consists of several layers of different materials. Electro-formed copper (5-cm thick), commercial copper (5-cm thick), lead (45-cm thick) and poly (30-cm thick) complete the shielding and protect the detectors against radiation of natural origin [9]. An overview on how the track for the calibration sources passes these layers is given in Fig. 1. A radon exclusion box covers the inner three layers and is purged at all times with boiled-off nitrogen. Around this box, two layers of plastic scintillator are used as an active muon veto. The calibration source has to pass through all of these layers; its support structure is located outside the full shielding behind the poly wall, see Fig. 1. A line source has to be pushed and pulled through the shielding on a defined path without destroying the radon purge. Therefore, a track is needed that forms a closed volume passing through the different layers of shielding and the veto panels. The line source assembly moves in a purged track around the cryostat (35.6-cm diameter). This track is made of 1–2 in. wide polytetrafluoroethylene (PTFE) tubing. The materials used for the track and source assembly have been leached in agreement with the MJD cleanliness procedures [18]. During the cleaning steps, a leach removes surface contamination and insures against the case of abrasions of the tube material that no radioactive contamination remains close to the detectors. Parts of the material were assayed after leaching in the MJD assay program. The chosen PTFE3 has a very small natural radioactivity with mass fractions below 3.1⋅ 10−12 for 238 U and around 1.5⋅ 10−12 for 232 Th. Inside the shielding, a nitrogen purge line is attached at the inner end to the track and is actively purging the whole track at all times minimizing radon intrusion. The purge line has a smaller diameter than the source track in order to reduce material around the detectors. Its attachment to source track is done by a coupling made of electro-formed copper and can be seen in Fig. 4. To ensure a closed volume when not calibrating, the end of the track laying outside the shielding is closed by a pneumatic gate valve when no source is deployed. This valve is installed on the plate indicated in Fig. 5. The closed track volume, in combination with the active purge, prevents lab air from back-filling the calibration track. For control and safety reasons, two light ‘‘emitter and detector’’ pairs are placed on the outside of the gate valve. One pair is used to check the status of the gate valve. A closed valve prevents infrared light from reaching the emitter, thus a non-detection of light indicates a closedvalve. The second pair is used to verify that the radioactive line source is fully retracted behind the valve in its storage position. A source out of storage position lies in front of the infrared detector, so that no light or less light then usual is detected. A detection of light indicates that the source is fully retracted and it is safe to close the valve. This system is
3. Mechanical implementation 3.1. Radioactive line sources Each line source consists of a radioactively-doped epoxy injected into a 3-mm-diameter tube that is sealed at both ends, produced to our custom specifications by Eckert & Ziegler Analytics, Inc.2 The source activities are conveniently tuned at production; once cured, it can be treated as a sealed source within an containment of 0.1-mm wall thickness. Deviations from homogeneity along the length of the source are below 3%. Four 228 Th calibration sources were prepared in 1-m lengths, each with an integrated activity of 5.18± 0.30 kBq. A single 60 Co source for building the pulse-shape analysis (PSA) library was prepared with an integrated activity of 6.3± 0.30 kBq over a 2 m length. The length was chosen so that the container encapsulating the radioactive material is long enough to encircle one cryostat twice. In the final configuration, we fabricated 3 line-source assemblies: one that consists of only the 60 Co source and two assemblies each consisting of a pair of 228 Th sources with a total activity of 10.36 kBq. 2
3
http://www.ezag.com/home/—Rod Line sources. 18
http://www.coleparmer.com.
N. Abgrall et al.
Nuclear Inst. and Methods in Physics Research, A 872 (2017) 16–22
Fig. 1. Drawing of a Majorana module assembly with the calibration system at final position in the Demonstrator design including the vacuum system, nitrogen supply as well as the shielding layers. Copper shielding is shown in brown, lead bricks in dark gray and the poly shield in purple. Not all veto panels are shown for better visibility. The loops of the track as well as the calibration track through the shield and the inner loops around the cryostats are visible in cyan. The calibration controls are located on the table next to the poly-shield under the cover next to the loops of the outer storage track. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
As mentioned this emitter–detector system works as a safety back-up. In the event of a power outage the system is not able to move and the source stays in position. The gate valve has to be actively supplied with pressurized air and power to change status so that a loss of air pressure or power would only results in a steady state; it would not close and cut a source. A logic device4 is placed in the power supply of the gate valve. The device lets power pass to the gate valve after it receives positive feedback from the Arduino checking the source status with the second infrared sensor pair. In the case of a broken sensor pair, no light is detected resulting in the feedback that the source is deployed and neglecting the request of closing the valve. At all times the outer part of the system behind the poly wall is easily accessible for trouble shooting and the end of the line source assembly can always be used to pull out a source by hand, see Fig. 1.
Fig. 2. Close-up of the drawing in Fig. 1 which shows the relative location of the detector array in one module, the surrounding calibration track, and the cross-arm connection to vacuum and cooling systems towards the left. The detector positions can be seen within the cryostat around which the line source assembly wraps twice. At the inner end of the 1/2-inch track, the smaller 1/4-inch purge line is coupled. The shielding surrounding the detectors and through which the calibration track is guided is removed in this drawing for better visibility.
3.3. Source positioning The source assembly is guided on the outside between two drive rollers, see Fig. 6. One of the drive rollers is rotated by a motor. The second drive roller passively rotates on an axle below the first one. The axle is pulled by two springs against the actively rotating drive roller so that the line source is gently clamped between the two drive rollers. This ensures a constant grip of the drive rollers on the line source so that the source can be pushed and pulled reliably over hundreds of deployments and retractions. Inside the inner copper shielding, the track is wrapped in a helical shape around the cryostat. The tubing is held in place by clean electroformed copper brackets which are attached to the cryostat. Fig. 4 shows the as-built configuration of the track around one module before it was moved into the shielding. For a deployed source the part containing the radioactive material wraps twice around the cryostat allowing a simultaneous calibration of all detectors, in contrast to the calibration system of Gerda [19]. There, three collimated sources can be lowered into the cryostat containing the detectors and calibrate the array one layer of detectors at a time.
Fig. 3. Sketch of the line source assembly. The outer Teflon™ container (blue) has a wall thickness of 0.25 mm. One pair of NdFeB magnets (black) is located at both ends of the line source. Two of the 1-m long container with 228 Th-material from Eckert & Ziegler (red) with radioactive material are located next to each other. Empty space is filled with Teflon™ material of thinner diameter to stabilize the source (gray). The left end of the source would be pointing towards the detectors and is the part the encircles one cryostat. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
used as a back-up to the magnet detection system. An improper closing of the valve could damage the thin-walled line source and in the worst case lead to a contamination of radioactive material.
4
19
http://dataprobe.com/ibootbar/.
N. Abgrall et al.
Nuclear Inst. and Methods in Physics Research, A 872 (2017) 16–22
Fig. 5. Schematic overview of the sensors installed at the angle plate of the calibration system. The inner calibration track would be situated on the right side of the image.
Fig. 4. Inner calibration track around Module 1. The 1/4-inch-purge line is attached to the 1/2-inch calibration track.
Since the detector rates at Majorana are heavily dependent on how far into the helical track a source is deployed the system should reproduce its previous position as precisely as possible to have comparable data sets. To reproduce the source position reliably, we have developed a positioning system using Hall effect switches.5 Three magnetic sensors, see Fig. 5, are connected to an Arduino UNO.6 These sensors register a passing magnet in a moving line source via the Hall effect. By reading their signal continuously, the arrangement of three sensors outside and two magnets at the front or the back end of the assembly allows us to determine the status of the source and reproduce its deployed position within a few millimeters. A special sequence in the readout of the magnetic sensors gives feedback to the experiment control if the source is deployed, in deployment, retracted, or in retraction. If not deployed, sources are stored in mirror tracks behind the last magnet sensor at the outer end of the calibration system plate. This mirror track, made of the same PTFE tubing as the other parts of the track, is wound similarly to the track inside the shield, thereby ‘‘mirroring’’ the shape of the deployed source as it is stored. It is possible to access this part of the calibration system and to exchange the whole mirror track with the source inside if another calibration source needs to be used. In the final configuration the 228 Th is installed as the standard source and can be exchanged with the 60 Co line source if additional commissioning is needed. 3.4. System controls The sensors attached to the Arduino controller and the motor can be controlled individually via the data acquisition (DAQ) software ORCA [20] in an expert mode which allows debugging and tests. Nonexpert users can use an interface with reduced functionality on the DAQ machines. This interface allows routine calibration operations. The reduced interface controls for example the run-bit settings which indicate run-type and lab conditions, e.g. a calibration run with Thorium. This interface also takes care of deployment, calibration time, and retraction of the sources and allows operators an easy remote control of the system.
Fig. 6. Radioactive line source between the two drive rollers. Two magnet sensors are placed in front of the drive roller (green boxes, left side in the upper figure). One magnet sensor (blue box) is located between the drive roller and the gate valve on the right. In the lower figure can be seen the springs that pull the idler pulley against the drive pulley above. In both figures the shielding and the cryostat are on the right side of the figures. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
through the lead shield to the detectors. While going through the outer layers of lead, the track on which the line source assembly moves is curved. The outer mirror track was designed so that the source would not lie in this line-of-sight when stored. It is coiled vertically so that the source in storage is located below the cross-arm ensuring that a calibration source in storage does not contribute to the background budget (<3 counts/(ROI ton⋅ year)) of the Majorana Demonstrator [9].
3.5. Interference and background contribution The copper cross-arm of the cryostat delivers the pumping and cooling power to the cryostat, but also provides a direct line-of-sight 5 6
Infineon TLE4905L E6433. https://www.arduino.cc/en/Main/Products. 20
N. Abgrall et al.
Nuclear Inst. and Methods in Physics Research, A 872 (2017) 16–22
An upper estimate of 1.3⋅ 10−3 counts/(ROI ton⋅ year) is indicated by simulations [21]. Tests with stronger button sources at certain locations behind the shield have verified that a source in the outer mirror track does not contribute to the background. We considered the possibility of removable contamination from sources being carried into or left inside the tube next to the cryostat. The mirror tracks and the cover of the system act as a barrier so that direct contact to the sources due to ongoing lab work is avoided and the line sources stay clean. In addition, the source assemblies, parts of the track, and drive rollers are wiped on a regular basis and have not shown abrasion or any signs of dirt. During the last year of running we have not found any increase of backgrounds attributable to the calibration source. Fig. 7. Average detector energy spectrum of Module 1. The spectrum is averaged over the number of detectors and taken with a calibration time of a one hour-long data set. The spectra are shown for the detectors biased at the time of the measurement. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
4. Commissioning and performance During commissioning, a stress test of the system was performed. Calibration sources were deployed and retracted about a hundred times without incident. Assuming a weekly calibration over the expected 5year lifetime of the Majorana Demonstrator the number of operational cycles is of the same order. The software was tested and different failure or mishandling modes were investigated and mitigated. Fig. 7 is taken from a data set from 2015 in which the first module was situated inside the shielding. In this data set, detectors in Module 1 were calibrated using line sources that provide a spectrum with several peaks at different energies. Fig. 8 shows the integral count rate for energies between 20 and 4000 keV. The average count rate for calibrations with the 60 Co and the 228 Th sources is around 40 Hz and 30 Hz, respectively. This fulfills the 100-Hz upper count rate limit. The figure shows that several detectors closer to the line source path have higher rates, as expected. Detectors on the inside or further away from the path have slightly lower rates. However, the built geometry achieved the goal to have a balanced count rate within the array. Fig. 9 shows the spectra for the detector with the highest count rate and for the detector with the lowest count rate. A deployed source around one module can be seen by detectors in the second cryostat. The recorded spectra in these detectors can be used to test stability. However, the count rate is not sufficient for a full calibration. For a full calibration data set of one module, a dedicated deployment of the corresponding source is necessary. In addition, it is possible to deploy both sources at the same time for high-rate data tests. Calibration measurements are also used as a reference for our simulation campaign. The simulations using the Geant4-based framework MaGe [16,17] are compared to measurements in order to validate the implementation of the geometry. Various parameters like the heightratio between the full-energy peaks, the full-energy peak to Compton continuum ratio and multiplicity distributions are affected by the positioning of the source relative to the detectors as well as by the amount of material located between source and detectors. The simulation creates gamma-rays along a continuous source of 2-m length in the deployed position. The simulation geometry includes the track, the full copper cryostat, structural material, shielding, and a thin ‘‘dead layer’’. Each detector has this thin layer of inactive (or partially active) material, the thickness of which affects the size of the step function under a fullenergy peak due to small energy losses of incoming gamma-rays. The adjustment of this layer size in the simulations to realistic values is necessary to get a similar spectral shape in experiment and simulation. A first analysis showed that MaGe is able to describe the measured spectra in calibration runs, as shown in Fig. 10. The simulation is in very good agreement and can be used to build up a pulse-shape library, to study the efficiency of cuts and the background.
Fig. 8. Integral count rate per detector in the high gain channel between 20 and 4000 keV for the two installed line sources. The count rate is given for the detectors used for the first data set of Module 1. Only data points for detectors for which a cobalt and a thorium measurement are available are shown in this plot. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
Fig. 9. Single-detector energy spectrum of Module 1 for the detector with the highest count rate and the detector with the lowest count rate. The count rates refer to the data set that is used to calculate the numbers in Fig. 8. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
5. Summary A calibration system has been designed for the Majorana Demonstrator. The system fulfills special requirements in terms of cleanliness, stable positioning and reproducibility. The strength of the radioactive source was adjusted to the special needs of the data stream and ensures stable data taking. Two systems, one for each cryostat in the shield, were built. Both systems are clones and can be used independently or in tandem. The systems deploy radioactive sources from outside the 21
N. Abgrall et al.
Nuclear Inst. and Methods in Physics Research, A 872 (2017) 16–22
from the Russian Foundation for Basic Research, grant No. 15-02-02919. We acknowledge the support of the US Department of Energy through the LANL/LDRD Program. This research used resources of the Oak Ridge Leadership Computing Facility, which is a DOE Office of Science User Facility supported under Contract DE-AC05-00OR22725. This research used resources of the National Energy Research Scientific Computing Center, a DOE Office of Science User Facility supported under Contract No. DE-AC02-05CH11231. We thank our hosts and colleagues at the Sanford Underground Research Facility for their support. References [1] F.T. Avignone III, et al., Double beta decay, Majorana neutrinos, and neutrino mass, Rev. Modern Phys. 80 (2008) 481–516 URL http://link.aps.org/doi/10.1103/ RevModPhys.80.481. [2] J.D. Vergados, et al., Theory of neutrinoless double-beta decay, Rep. Progr. Phys. 75 (10) (2012) 106301 URL http://stacks.iop.org/0034-4885/75/i=10/a= 106301. [3] H.V. Klapdor-Kleingrothaus, I.V. Krivosheina, The evidence for observation of 0𝜈𝛽𝛽 decay: the identification of 0𝜈𝛽𝛽 events from the full spectra, Modern Phys. Lett. A 21 (20) (2006) 1547–1566 URL https://doi.org/10.1142/S0217732306020937. [4] EXO-200-Collaboration, Search for Majorana neutrinos with the first two years of EXO-200 data, Nature (ISSN: 0028-0836) 510 (7504) (2014) 229–234 URL http: //dx.doi.org/10.1038/nature13432. [5] K. Alfonso, et al., CUORE Collaboration Collaboration, Search for Neutrinoless Double-Beta Decay of 130 Te with CUORE-0, Phys. Rev. Lett. 115 (2015) 102502 URL https://link.aps.org/doi/10.1103/PhysRevLett.115.102502. [6] M. Agostini, et al., Search of neutrinoless double beta decay with the Gerda Experiment, Nucl. Part. Phys. Proc. (ISSN: 2405-6014) 273–275 (2016) 1876–1882 URL http://www.sciencedirect.com/science/article/pii/S2405601415007920 37th International Conference on High Energy Physics (ICHEP). [7] A. Gando, et al., Search for Majorana neutrinos near the inverted mass hierarchy region with KamLAND-Zen, Phys. Rev. Lett. 117 (2016) 082503 URL http://link. aps.org/doi/10.1103/PhysRevLett.117.082503. [8] A.G. Cohen, D.B. Kaplan, A.E. Nelson, Progress in electroweak baryogenesis, Annu. Rev. Nucl. Part. Sci. 43 (1993) 27–70. [9] N. Abgrall, et al., The Majorana demonstrator neutrinoless double-beta decay experiment, Adv. High Energy Phys. 2014 (2014) 1–18 URL http://dx.doi.org/10.1155/ 2014/365432. [10] S. Elliott, et al., Initial results from the MAJORANA DEMONSTRATOR, arxiv 1610 (2016) 01210. URL https://arxiv.org/abs/1610.01210. [11] N. Abgrall, et al., New limits on bosonic dark matter, solar axions, pauli exclusion principle violation, and electron decay from the Majorana Demonstrator, Phys. Rev. Lett. 118 (2017) 161801 URL https://link.aps.org/doi/10.1103/PhysRevLett.118. 161801. [12] S. Elliott, V. Gehman, K. Kazkaz, D.-M. Mei, A. Young, Pulse shape analysis in segmented detectors as a technique for background reduction in Ge double-beta decay experiments, Nucl. Instrum. Methods Phys. Res. Sect. A 558 (2006) 504–510 URL http://www.sciencedirect.com/science/article/pii/S0168900205023867. [13] R. Cooper, et al., A pulse shape analysis technique for the Majorana experiment, Nucl. Instrum. Methods Phys. Res. Sect. A (ISSN: 0168-9002) 629 (1) (2011) 303–310 URL http://www.sciencedirect.com/science/article/pii/S0168900210024915. [14] S. Mertens, et al., Majorana collaboration’s experience with germanium detectors, J. Phys. Conf. Ser. 606 (1) (2015) 012005 URL http://stacks.iop.org/1742-6596/606/ i=1/a=012005. [15] S. Agostinelli, et al., Geant4 a simulation toolkit, Nucl. Instrum. Methods Phys. Res. Sect. A (ISSN: 0168-9002) 506 (3) (2003) 250–303 URL http://www.sciencedirect. com/science/article/pii/S0168900203013688. [16] M. Bauer, et al., MaGe: a Monte Carlo framework for the Gerda and Majorana double beta decay experiments, J. Phys. Conf. Ser. 39 (1) (2006) 362 URL http: //stacks.iop.org/1742-6596/39/i=1/a=097. [17] M. Boswell, et al., MaGe - a Geant4-based Monte Carlo Application Framework for Low-background Germanium Experiments, 2010. ArXiv Preprint Nucl-Ex URL https://arxiv.org/abs/1011.3827v1. [18] N. Abgrall, et al., The Majorana demonstrator radioassay program, Nucl. Instrum. Methods Phys. Res. Sect. A (ISSN: 0168-9002) 828 (2016) 22–36 URL http://www. sciencedirect.com/science/article/pii/S0168900216302832. [19] L. Baudis, et al., Monte Carlo studies and optimization for the calibration system of the Gerda experiment, Nucl. Instrum. Methods A 729 (2013) 557–564 URL http://www.sciencedirect.com/science/article/pii/S0168900213011273. [20] M.A. Howe, et al., Sudbury neutrino observatory neutral current detector acquisition software overview, IEEE Trans. Nucl. Sci. 51 (2004) 878–883. [21] J. MacMullin, Validation of the Background Model for the MAJORANA DEMONSTRATOR, Ph.D. thesis, University of North Carolina, 2015.
Fig. 10. (top) Comparison of a 228 Th line source simulation using MaGe and a measurement of Module 1. The simulated distribution was normalized by matching the integrals of both curves in the range from 2595 keV to 2635 keV. (bottom) Residual between simulation and experiment for the 1-keV binning (blue) averaged over a 10-keV wide window (red). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
shielding to a position next to the detectors so that energy calibrations and other studies are possible. In contrast to other systems, our line source approach can calibrate the entire array at the same time without moving the source. The results can be used to study the agreement of observed energy distributions and multiplicities in experiment and simulation. Regular calibrations allow the study of the stability of data cleaning cuts over a long time. The design ensures a sufficient count rate in all detectors of the array so that one calibration run can be used for all detectors of a module. Also, it has been shown that the calibration system around one module can easily be copied, which is important for the scalability of such a modular approach when going to ton-scale experiments. Both systems are operated remotely and the controls are implemented in the DAQ system of the Majorana Demonstrator. While the exact design depends on the future experimental conditions, the MJD calibration system shows that it is possible to construct a system using easily accessible materials, sensors and readout electronics while still holding to all cleanliness and low-background limits. Acknowledgments This material is based upon work supported by the US Department of Energy, Office of Science, Office of Nuclear Physics under Award Numbers DE-AC02-05CH11231, DE-AC52-06NA25396, DE-FG02-97ER41041, DE-FG02-97ER41033, DE-FG02-97ER41042, DESC0012612, DE-FG02-10ER41715, DE-SC0010254, and DE-FG0297ER41020. We acknowledge support from the Particle Astrophysics Program and Nuclear Physics Program of the National Science Foundation through grant numbers PHY-0919270, PHY-1003940, 0855314, PHY-1202950, MRI 0923142 and 1003399. We acknowledge support
22