avatar University of Washington Inc Services


  • Page 1


  • Page 2

    ANNUAL REPORT Center for Experimental Nuclear Physics and Astrophysics University of Washington April 1, 2017 Sponsored in part by the United States Department of Energy under Grant #DE-FG02-97ER41020.

  • Page 3

    This report was prepared as an account of work sponsored in part by the United States Government. Neither the United States nor the United States Department of Energy, nor any of their employees, makes any warranty, expressed or implied or assumes any legal liability or responsibility for accuracy, completeness or usefulness of any information, apparatus, product or process disclosed, or represents that its use would not infringe on privately-owned rights. Cover design by Gary Holman. The images show students, postdocs and staff working on one of the many research projects at CENPA.

  • Page 4

    UW CENPA Annual Report 2016-2017 April 2017 i INTRODUCTION The Center for Experimental Nuclear Physics and Astrophysics, CENPA, was established in 1998 at the University of Washington as the institutional home for a broad program of research in nuclear physics and related fields. Research activities are conducted locally and at remote sites. The research program emphasis is fundamental symmetries and neutrinos. In neutrino physics, CENPA is the lead US institution in the KATRIN tritium beta decay experiment, the site for experimental work on Project 8, and a collaborating institution in the MAJORANA 76 Ge double beta decay experiment. The Muon Physics group has developed the MuSun experiment to measure muon capture in deuterium at the Paul Scherrer Institute in Switzerland. The group has a leadership role in the new project to measure the anomalous magnetic moment of the muon at Fermilab to even higher precision than it is presently known from the collaboration’s previous work at Brookhaven. The fundamental symmetries program also includes “in-house” research on the search for a static electric dipole moment in 199 Hg, and an experiment using the local tandem Van de Graaff accelerator to measure the electron-neutrino correlation and Fierz interference in 6 He decay. In addition to the research directly supported by DOE’s Office of Nuclear Physics through the CENPA core grant, other important programs are located at CENPA, forming a broader intellectual center with valuable synergies. The “Gravity” group, as it is known, carries out with both DOE and NSF support studies of the weak and strong Equivalence Principles, fundamental precepts of General Relativity, as well as searches for non-Newtonian weak forces such as are predicted by theories with extra dimensions. In addition, they participate in advanced LIGO. The DOE Office of High Energy Physics supports a unique experiment, the ADMX axion search. An unusual spin-off from basic nuclear physics is our program on nanopore sequencing, supported by NIH. Notable The LIGO Scientific Collaboration, which UW is a part of, was awarded the 2016 Special Breakthrough Prize in Fundamental Physics and the 2016 Gruber Cosmology Prize for the detection of gravitational waves from two black holes colliding over a billion light years away. The UW contribution to Advanced LIGO has been in characterizing charge and pressure noise on the LIGO test masses and in low-frequency seismic isolation using high-precision ground-rotation sensors (NSF support). Transitions CENPA Director, Hamish Robertson, is now Professor Emeritus. Peter Doe has assumed the role of US spokesperson for the KATRIN experiment. Fortunately Hamish continues his unique contributions to the KATRIN, Project 8 and TRIMS experiments. David Hertzog has been appointed Interim Director. Associate Director, Diana Parno, has assumed the position of Assistant Research Professor at Carnegie Mellon University. Her legendary organizational talent and comradeship will be greatly missed. Fortunately Diana will continue her physics input to the KATRIN project and connections to CENPA, having established an active CMU group. Gary Holman has been appointed Acting Associate Director. Research Engineer Hannah LeTourneau, who supported

  • Page 5

    ii the helium liquifier, left to go to graduate school. Research Engineer Joben Pdersen is now a full-time staff member working on the accelerator and ion source. Postdoc Clara Cuesta has taken a new position as AIDA2020 postdoctoral fellow at CIEMAT in Madrid, Spain. Postdoc Ana Malagon left the ADMX group in December 2016. She is now in private industry in the Seattle area. Postdocs Walter Pettus and Mathieu Guigue joined the lab this past year. Walter has been working on Project 8 and Mathieu on ADMX. Graduate students Christian Boutan (ADMX), Michael Murray (MuSun), and Tim Winch- ester (SNO) completed their theses this past year. Christian will begin a position at Pacific Northwest National Laboratory. Graduate student Julieta Gruszko was awarded a prestigious MIT Pappalardo Fellowship and Graduate student Matthias Smith was awarded a prestigious INFN Fellowship. Julieta will move to Boston and Matthias to Pisa, Italy to start their new positions in early fall, after completing their theses. Finally, Master’s students Kerkira Stock- ton defended her thesis with the Gravity group this March and Diana Thompson defended her thesis in October 2016. Highlights • In October, KATRIN achieved a major milestone with Director Robertson among those pressing the “First Light” button to send electrons down the entire length of the KATRIN apparatus. • As tritium operation approaches for KATRIN, major upgrades to the detector veto hardware and DAQ rate capability have been implemented that will better meet and extend the physics reach of the experiment. • For newly proposed high-rate measurements with KATRIN, such as a keV-scale sterile neutrino search and an extensive calibration scan of the source properties with a high- intensity e-gun, a new signal shaping filter logic was developed at CENPA and then implemented into the FPGA of the detector readout system. An initial test with an e-gun shows good agreement in performance with the design predictions. • The KATRIN collaboration adopted the UW-group design of overall analysis structure to integrate detector data, slow-control readings, simulation and statistical analysis, as well as our design for quality assurance and data blinding. The first version of the software suite was implemented at CENPA and is being evaluated by the collaboration. The UW group periodically provides training sessions on KATRIN data analysis to the collaboration. • The MAJORANA DEMONSTRATOR (MJD) Collaboration has successfully completed construction of the second of the two modules of enriched Ge detectors, and both modules are now running in a completed shield at the Sanford Underground Research Facility in Lead, SD. • Using data from MJD, we published in PRL a search for low-energy signals from pseu- doscalar and vector Dark Matter and other exotic phenomena

  • Page 6

    UW CENPA Annual Report 2016-2017 April 2017 iii • The MJD Collaboration released preliminary results at Neutrino 2016 and Neutrino Telescopes 2017 showing achievement within uncertainty of reaching the primary back- ground goal. The Collaboration is joining forces with their European counterparts, GERDA, to achieve the highest-sensitivity neutrinoless double-beta decay search to date, and to mount a ton-scale experiment, dubbed LEGEND. • Project 8 has successfully commissioned a Cyclotron Radiation Emission Spectroscopy (CRES) cell with increased gas volume and better signal-to-noise ratio, compatible with first tritium operation. The vacuum manifold to safely handle the tritium gas was built and passed review by the radiation safety committee. The understanding of the rich CRES frequency spectrum is well advanced. • A new analysis of data from the 3 phases of SNO that makes use of a 30% larger fiducial volume and 18% larger live time with the aid of a new event fitter has been completed. The hep neutrino flux is found to be non-zero at more than 95% CL, with a most probable value about 3 times larger than the theoretical prediction. • We finished analyzing and published our determination of a gamma branch in 22 Na aimed at resolving a puzzle, but the puzzle remains after our measurement. The analysis did reveal an interesting story regarding the role of collective motion in 22 Na. • We have two achievements related to our little-a experiment with laser-trapped 6 He. We published a careful analysis of the position determination properties of our MCP detector showing 8-µm precision, an important step towards the determination of the β − ν correlation. We also finished our first physics paper, presenting data on the charge distribution of Li ions, a solid benchmark for atomic theory calculations of similar problems. • We have put together a collaboration and written a proposal for applying the CRES technique developed by Project 8 to measure the 6 He beta spectrum. The analysis of systematics uncertainties indicates this could be the most sensitive way of searching for chirality-flipping currents ever proposed. • The Muon g − 2 project is 97% complete. Beam commissioning has started and experi- ment commissioning is schedule for June, 2017. Substantive UW contributions include: muon storage modeling; NMR probes, electronics and DAQ; shimming the precision magnetic field; and, the construction of the full calorimeter system. We are presently designing and installing a set of novel beam imaging detectors to help steer the muons into the storage ring. Six UW Ph.D. students are involved. • Our g − 2 calorimeter subgroup led a third test beam experiment at SLAC, which featured the final designs of the calibration system, calorimeters, readout electronics, waveform digitizers, DAQ, GPU farm, and offline analysis framework. A technical paper is in preparation. • The UW field group played a leading role in shimming the Storage Ring magnet field to a uniformity that exceeds by a factor of 3 that realized in the BNL E821 experiment. We also developed the instrumentation and measured the radial and longitudinal magnetic field components in the storage ring.

  • Page 7

    iv • Our MuSun experiment concluded its data taking on µd capture with a final run at PSI that was focussed on measuring specific systematic uncertainties. With that, our full statistics has been achieved and the necessary corresponding test measurements have been made. • After reporting a new upper limit on the Hg EDM in 2016, our group has embarked upon an upgrade to the experimental apparatus to both increase its sensitivity and reduce the source of the dominant systematic error. • The ADMX Gen2 experiment — located at CENPA — began data taking operations with unprecedented sensitivity to axion dark matter. • The UW LIGO team members successfully installed a second ground-rotation-sensor at the LIGO Hanford Observatory improving its robustness against wind and consequently improving duty cycle for the second observation run. As always, we encourage outside applications for the use of our facilities. As a conve- nient reference for potential users, the table on the following page lists the capabilities of our accelerators. For further information, please contact Gary Holman, Acting Associate Director (holman@uw.edu) or Eric Smith, Research Engineer (esmith66@u.washington.edu) CENPA, Box 354290, University of Washington, Seattle, WA 98195; (206) 543 4080. Further information is also available on our web page: http://www.npl.washington.edu. We close this introduction with a reminder that the articles in this report describe work in progress and are not to be regarded as publications nor to be quoted without permission of the authors. In each article the names of the investigators are listed alphabetically, with the primary author underlined in the case of multiple authors, to whom inquiries should be addressed. David Hertzog, Interim Director Gary Holman, Acting Associate Director and Editor Daniel Salvat and Walter Pettus, Technical Editors Victoria Clarkson, Assistant Editor

  • Page 8

    UW CENPA Annual Report 2016-2017 April 2017 v TANDEM VAN DE GRAAFF ACCELERATOR Our tandem accelerator facility is centered around a High Voltage Engineering Corporation Model FN purchased in 1966 with NSF funds, with operation funded primarily by the U.S. Department of Energy. See W. G. Weitkamp and F. H. Schmidt, “The University of Washing- ton Three Stage Van de Graaff Accelerator,” Nucl. Instrum. Methods 122, 65 (1974). The tandem was adapted in 1996 to an (optional) terminal ion source and a non-inclined tube #3, which enables the accelerator to produce high intensity beams of hydrogen and helium isotopes at energies from 100 keV to 7.5 MeV. Some Available Energy Analyzed Beams Ion Max. Current Max. Energy Ion Source (particle µA) (MeV) 1H or 2 H 50 18 DEIS or 860 3 He or 4 He 2 27 Double Charge-Exchange Source 3 He or 4 He 30 7.5 Tandem Terminal Source 6 Li or 7 Li 1 36 860 11 B 5 54 860 12 C or 13 C 10 63 860 ∗14 N 1 63 DEIS or 860 16 O or 18 O 10 72 DEIS or 860 F 10 72 DEIS or 860 ∗ Ca 0.5 99 860 Ni 0.2 99 860 I 0.001 108 860 *Negative ion is the hydride, dihydride, or trihydride. Several additional ion species are available including the following: Mg, Al, Si, P, S, Cl, Fe, Cu, Ge, Se, Br and Ag. Less common isotopes are generated from enriched material. We recently have been producing the positive ion beams of the noble gases He, Ne, Ar, and Kr at ion source energies from 10 keV to 100 keV for implantation, in particular the rare isotopes 21 Ne and 36 Ar. We have also produced a separated beam of 15-MeV 8 B at 6 particles/second.

  • Page 9

  • Page 10

    UW CENPA Annual Report 2016-2017 April 2017 vii Contents INTRODUCTION i 1 Neutrino Research 1 KATRIN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1.1 KATRIN status . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1.2 Focal-plane-detector system operation and upgrades . . . . . . . . . . . . . . 2 1.3 Study of the control of the inter-spectrometer Penning trap . . . . . . . . . . 5 1.4 Main spectrometer background studies . . . . . . . . . . . . . . . . . . . . . 6 1.5 Electron detection for time-of-flight operation . . . . . . . . . . . . . . . . . 7 1.6 Firmware upgrade for high-rate measurements . . . . . . . . . . . . . . . . . 8 1.7 Progress on veto upgrade . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 1.8 Data quality for the KATRIN experiment . . . . . . . . . . . . . . . . . . . . 14 TRIMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 1.9 Status of the Tritium Recoil-Ion Mass Spectrometer (TRIMS) . . . . . . . . . 15 1.10 The tritium and 83m Kr gas-handling system for the TRIMS experiment . . . 16 MAJORANA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 1.11 Overview of the MAJORANA DEMONSTRATOR . . . . . . . . . . . . . . . . . 17 1.12 Summary of recent results from the MAJORANA DEMONSTRATOR . . . . . . 18 1.13 Alpha particle discrimination and TUBE . . . . . . . . . . . . . . . . . . . . 21 1.14 Searching for 2νββ to excited states . . . . . . . . . . . . . . . . . . . . . . . 24 Project 8 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 1.15 Status of the Project 8 neutrino-mass experiment . . . . . . . . . . . . . . . 26 1.16 Development of an atomic source for Project 8 . . . . . . . . . . . . . . . . . 30 SNO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 1.17 SNO and the solar-neutrino reaction hep . . . . . . . . . . . . . . . . . . . . 32

  • Page 11

    viii COHERENT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 1.18 The COHERENT experiment . . . . . . . . . . . . . . . . . . . . . . . . . . 34 2 Non-accelerator-based tests of fundamental symmetries 36 Torsion-balance experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 2.1 New limits on ultra-light axionic dark matter . . . . . . . . . . . . . . . . . . 36 2.2 Progress on ground-rotation sensors for LIGO . . . . . . . . . . . . . . . . . 38 2.3 Wedge-pendulum experiment update . . . . . . . . . . . . . . . . . . . . . . 39 2.4 Rotating equivalence-principle test . . . . . . . . . . . . . . . . . . . . . . . . 40 2.5 Active vibration-isolation for torsion-balance experiments . . . . . . . . . . . 41 2.6 A new experimental search for equivalence-principle violating dark matter . 42 Other tests of fundamental symmetries . . . . . . . . . . . . . . . . . . . . . 44 2.7 The mercury electric-dipole-moment experiment . . . . . . . . . . . . . . . . 44 3 Accelerator-based physics 45 3.1 The 2+ + 1 → 31 gamma width in 22 Na and second-class currents . . . . . . . . . 45 3.2 Overview of the 6 He experiments . . . . . . . . . . . . . . . . . . . . . . . . . 47 3.3 Improvements of the 6 He laser-trapping efficiency . . . . . . . . . . . . . . . . 49 3.4 Image calibrations to determine the stability and systematics of the MOT position . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52 3.5 Position-dependent timing response of the MCP detector . . . . . . . . . . . 57 3.6 High-voltage monitoring . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 3.7 Recoil-ion charge-state distribution in 6 He β-decay . . . . . . . . . . . . . . . 61 3.8 Toward measurement of the 6 He β-spectrum with Cyclotron Radiation Emis- sion Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62 4 Precision muon physics 65 4.1 Overview of the muon physics program . . . . . . . . . . . . . . . . . . . . . 65

  • Page 12

    UW CENPA Annual Report 2016-2017 April 2017 ix MuSun . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67 4.2 Muon capture on deuterium, the MuSun experiment: Overview . . . . . . . . 67 4.3 Correcting for Muon-Catalyzed Fusion in the TPC . . . . . . . . . . . . . . . 69 4.4 Background subtraction techniques developed for analysis of R2014 dataset . 71 4.5 MuSun systematic effects run campaign . . . . . . . . . . . . . . . . . . . . . 73 4.6 Constraining stops in TPC wall materials via neutron detection . . . . . . . 75 4.7 Monte Carlo framework and studies . . . . . . . . . . . . . . . . . . . . . . . 77 g−2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79 4.8 Overview of the g − 2 experiment . . . . . . . . . . . . . . . . . . . . . . . . 79 4.9 Calorimeter status and SLAC run . . . . . . . . . . . . . . . . . . . . . . . . 83 4.10 Building a calorimeter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 4.11 Nearline and offline data-analysis framework . . . . . . . . . . . . . . . . . . . 86 4.12 Data quality monitoring . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89 4.13 Calorimeter algorithms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92 4.14 Highlights of the SLAC 2016 data analysis . . . . . . . . . . . . . . . . . . . . 95 4.15 Muon beam injection and storage studies . . . . . . . . . . . . . . . . . . . . 100 4.16 Inflector beam monitoring system . . . . . . . . . . . . . . . . . . . . . . . . 101 4.17 Magnetic field overview - the shimming story . . . . . . . . . . . . . . . . . . 104 4.18 NMR hardware . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108 4.19 Magnetic field data acquisition and analysis . . . . . . . . . . . . . . . . . . . 110 4.20 Radial and longitudinal field measurements in the g − 2 storage ring magnet . 114 5 Axion searches 116 ADMX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116 5.1 Overview of ADMX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116 5.2 Higher Frequency Axion Searches with Sidecar . . . . . . . . . . . . . . . . . 120 5.3 Signal Models for Axion Cavity Searches . . . . . . . . . . . . . . . . . . . . . 121

  • Page 13

    x 6 Education 124 6.1 Use of CENPA facilities in education and course work at UW . . . . . . . . 124 6.2 Student training . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124 6.3 Accelerator-based lab class in nuclear physics . . . . . . . . . . . . . . . . . . 128 7 Facilities 130 7.1 Facilities overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130 7.2 Van de Graaff accelerator and ion-source operations and development . . . . 131 7.3 Laboratory computer systems . . . . . . . . . . . . . . . . . . . . . . . . . . 134 7.4 Electronic equipment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135 7.5 Instrument Shop . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137 7.6 Student shop . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141 7.7 Building maintenance, repairs, and upgrades . . . . . . . . . . . . . . . . . . 142 7.8 Laboratory Safety . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144 8 CENPA Personnel 146 8.1 Faculty . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146 8.2 CENPA External Advisory Committee . . . . . . . . . . . . . . . . . . . . . . 147 8.3 Postdoctoral Research Associates . . . . . . . . . . . . . . . . . . . . . . . . 147 8.4 Predoctoral Research Associates . . . . . . . . . . . . . . . . . . . . . . . . . 147 8.5 University of Washington graduates taking research credit . . . . . . . . . . . 148 8.6 University of Washington undergraduates taking research credit . . . . . . . . 148 8.7 NSF Research Experience for Undergraduates participants . . . . . . . . . . . 149 8.8 Visiting students taking research credit . . . . . . . . . . . . . . . . . . . . . . 149 8.9 Professional staff . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149 8.10 Technical staff . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150 8.11 Administrative staff . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150 8.12 Part-time staff and student helpers . . . . . . . . . . . . . . . . . . . . . . . 150

  • Page 14

    UW CENPA Annual Report 2016-2017 April 2017 xi 8.13 Visitors & Volunteers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150 9 Publications 151 9.1 Published papers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151 9.2 Invited talks at conferences . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156 9.3 Abstracts and contributed talks . . . . . . . . . . . . . . . . . . . . . . . . . . 159 9.4 Papers submitted or to be published . . . . . . . . . . . . . . . . . . . . . . . 162 9.5 Reports and white papers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163 9.6 Master’s degrees granted . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164 9.7 Ph.D. degrees granted . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164

  • Page 15

  • Page 16

    UW CENPA Annual Report 2016-2017 April 2017 1 1 Neutrino Research KATRIN 1.1 KATRIN status J. F. Amsbaugh, J. Barrett∗ , A. Beglarian† , T. Bergmann† , L. I. Bodine, T. H. Burritt, P. J. Doe, S. Enomoto, N. Fong, J. A. Formaggio∗ , F. M. Fränkle† , F. Harms† , L. Kippenbrock, A. Kopmann† , E. L. Martin, A. Müller† , N. S. Oblath∗ , R. Ostertag† , D. S. Parno‡ , D. A. Peterson, A. W. P. Poon§ , R. G. H. Robertson, A. Seher† , D. Tcherniakhovski† , T. D. Van Wechel, K. J. Wierman¶ , J. F. Wilkerson¶ , and S. Wüstling† By making a precise measure of the end point of the electron energy spectrum from tritium beta decay, KATRIN will probe the neutrino mass to a planned sensitivity of 200 meV. The experimental technique, drawing on the experiences of several earlier experiments, uses a windowless, gaseous tritium source and two electrostatic spectrometers which measure the electron energy. Initially proposed in 2001, all major components, shown in Fig. 1.1-1, are now in place and in the final stages of commissioning. ctor Dete eter trom spec Main ter ome pectr Pres genic Cryo system al ping ren Pum Diffe system ping Pum e ms ourc Tri u Rear m syste oils on c nsa mpe ld co c fie s agne eter M 70 m Figure 1.1-1. Arrangement of the principal components of the KATRIN apparatus. ∗ Massachusetts Institute of Technology, Cambridge, MA. † Karlsruhe Institute of Technology, Karlsruhe, Germany. ‡ Presently at Carnegie Mellon University, Pittsburgh, PA. § Lawrence Berkeley National Laboratory, Berkeley, CA. ¶ University of North Carolina, Chapel Hill, NC.

  • Page 17

    2 There have been two data-taking campaigns over the past year: June - August, which focused on studying spectrometer backgrounds and October - December 2016, during which the entire KATRIN beam line was operated for the first time and the focus was on alignment of the beam line components. The two spectrometers were operated in unison for the first time during the autumn cam- paign. The pre-spectrometer operates at a potential several hundred volts below that of the main spectrometer in order to prevent the intense flux of lower energy beta electrons from entering and overwhelming the main analyzing spectrometer. This potential difference, in combination with the magnetic flux tube, results in a Penning trap between the two spec- trometers. Electrons stored in this trap produce backgrounds through ionization interactions and intermittent discharges. To empty this trap in a controlled fashion, Penning wipers peri- odically sweep through the trap. The efficacy of this technique to control the trap is reported below. The background in the main spectrometer was a factor of approximately 40 higher than the design goal. After lengthy investigation, the contribution of possible sources has now been quantified and it has been demonstrated that the primary background results from Rydberg atoms distributed throughout the volume of the spectrometer. The origin of this background and the plans to control it are presented below. On 14th October, with the entire beam line evacuated and the source and transport magnets energized, ‘First Light’ operation was possible, whereby electrons from a gun placed upstream of the source traveled the entire 70-meter length of the apparatus. This enabled the individual components of KATRIN to be aligned, centering the unobstructed flux tube onto the focal plane detector. The magnetic flux tube passes through cylindrical electrodes positioned along the electron transport system. These electrodes are designed to prevent ions in the source from entering and contaminating the spectrometers. The electron gun, which was also capable of producing ions, demonstrated that the electrodes successfully exclude the ions from the spectrometers, which is a necessary requirement for tritium operation. 1.2 Focal-plane-detector system operation and upgrades J. F. Amsbaugh, J. Barrett∗ , A. Beglarian† , T. Bergmann† , L. I. Bodine, T. H. Burritt, P. J. Doe, S. Enomoto, N. Fong, J. A. Formaggio∗ , F. M. Fränkle† , F. Harms† , L. Kippenbrock, A. Kopmann† , E. L. Martin, A. Müller† , N. S. Oblath∗ , R. Ostertag† , D. S. Parno‡ , D. A. Peterson, A. W. P. Poon§ , R. G. H. Robertson, A. Seher† , D. Tcherniakhovski† , T. D. Van Wechel, K. J. Wierman¶ , J. F. Wilkerson¶ , and S. Wüstling† The Focal Plane Detector (FPD) system saw approximately 4 months of 24/7 operation during this year’s commissioning of the source, transport system, and spectrometers. Over ∗ Massachusetts Institute of Technology, Cambridge, MA. † Karlsruhe Institute of Technology, Karlsruhe, Germany. ‡ Presently at Carnegie Mellon University, Pittsburgh, PA. § Lawrence Berkeley National Laboratory, Berkeley, CA. ¶ University of North Carolina, Chapel Hill, NC.

  • Page 18

    UW CENPA Annual Report 2016-2017 April 2017 3 the remaining 8 months, maintenance and upgrades of the FPD were performed as described below. • Magnet maintenance: The detector magnet has been operating reliably since 2009. As the magnet must function reliably for another 6 years, the magnet manufacturer replaced the cold head and performed required maintenance. The magnet has since been successfully operated at the nominal field of 3.6 T in conjunction with the 6 T pinch magnet. The current feedthrough plugs on the vacuum cryostat of both the two- year old pinch magnet and the nine-year old detector magnet suffered from overheating. Design issues have been identified, repairs were carried out, and design improvements are being investigated. This is particularly important since a total of eight of these magnets are employed in the KATRIN beam line. • Detector resolution: It has been observed during recent data runs that the FPD energy resolution has degraded since initial commissioning measurements at CENPA and KIT several years ago. To trace the origin of this deterioration, an effort was made to analyze 241 Am calibration runs regularly taken during detector operation in order to construct a detailed time evolution of the energy resolution. Although stable during the course of previous measurement phases, the temperature-corrected energy resolution appears to have worsened during specific maintenance breaks in which invasive work was done on the FPD system, including the removal of the detector wafer from the system. A separate noise analysis indicates that a source of parallel resistance, likely located near the detector wafer itself, is responsible for the resolution degradation. To enable further studies of the energy resolution, there is a plan to upgrade a test stand at KIT to allow for additional testing of detector components apart from the FPD system. • PAE potential : The design criteria of the post acceleration electrode (PAE) is to provide up to 30-kV energy to charged particles of either polarity. Since its initial testing in 2010, the PAE cannot exceed approximately 10 kV without suffering high voltage breakdown. This breakdown is independent of the polarity, or the presence of the 3.6-T magnetic field. Using a mock-up of the PAE at UW, visiting KIT students Agnes Seher and Raphael Ostertag identified the problem as being a ‘triple junction’ resulting from the contact of a conductor (metal foil electrodes) with an insulator (a quartz tube) in the presence of the PAE vacuum. Removing the electrodes from the quartz tubes and performing a glow discharge conditioning of the system allowed 30-kV potentials without breakdown. These tests were not conducted with magnetic fields. The foils were placed over the quartz to remove a Penning trap at the PAE. Removing the foil electrodes will be carried out at KIT to determine if high potentials may be reached without the Penning trap being a problem. • Pulcinella tests: Pulcinella is a titanium disk that can be inserted into the flux tube and illuminated with UV light, providing photoelectrons whose energy can be varied according to the potential applied to the disk. By attaching a femto-ammeter to the disk, the photoelectron current measured by the ammeter can be compared to the

  • Page 19

    4 electron flux recorded by the detector. The disk may also be used as a ‘fire wall’ to protect the sensitive detector from unexpected, high current events occurring upstream. If the system is acceptably quiet, then the disk can be retracted, exposing the detector. This was tested with both electrons and ions and Pulcinella was found to be a sensitive monitor that may be used as a bellwether for activities upstream of the detector. • DAQ system upgrade: The Data Acquisition System (DAQ) was designed to operate at a maximum rate of ∼ 150 kHz, or about 1 kHz per pixel. This is more than adequate for the region of interest where the rate is not expected to exceed a few mHz. However, to minimize the amount of time spent calibrating the detector and to allow it to be used for investigating the lower energy regions of the spectrum for evidence of sterile neutrinos, it must be capable of higher rate operation without suffering the energy distortions and event counting inaccuracies resulting from event pile-up at high rates. This was achieved by incorporating a second layer of bipolar shaping into the FPGA. The initial results obtained using an electron gun are in good agreement with simulations, indicating that the DAQ can operate at 2 MHz without suffering problems due to event pile-up. These results are described in detail below. • Veto upgrade: The new, simpler and more robust veto system has been operated in its entirety. Final installation of the scintillator end caps will be completed during the next maintenance period. The performance of the system is described below. • Alignment and shadowing: Using the electron gun upstream of the tritium source, the 90-mm diameter sensitive area of the detector was aligned to 1.8 mm of the center of the flux tube. During this exercise it was discovered that the lower segment of the detector appeared to be shadowed. This shadowing was located to a region between the two spectrometers and was initially attributed to the flapper of the valve between the two spectrometers not being fully open. This was found to be not the case. By running the electron gun at a fixed potential of 100 V and applying an adjustable potential to the valve, it was found that only the low energy electrons were being removed, suggesting that the shadowing was due to an electrostatic charging in the region of the valve. Investigations continue, but it is not thought that this will influence the neutrino mass sensitivity. • Summary and future: The FPD continues to perform well for commissioning the KATRIN hardware. The ongoing upgrades and maintenance schedule is expected to meet the needs of tritium running and to allow for initial investigation of additional physics.

  • Page 20

    UW CENPA Annual Report 2016-2017 April 2017 5 1.3 Study of the control of the inter-spectrometer Penning trap M. Fedkevych∗ , F. Fränkle† , L. Kippenbrock, D. S. Parno‡ , and P. C. -O. Ranitzsch∗ A Penning trap is formed by the electromagnetic conditions in the region connecting the KATRIN pre- and main spectrometers. Here, the combination of a strong magnetic field and a positive electric potential well results in the trapping of negative particles. Any low-energy electrons produced between the spectrometers will be confined to the Penning trap, but these electrons can ionize residual gas, producing additional trapped electrons as well as positive ions. Because the main spectrometer is operated at a negative high voltage, the positive ions can be accelerated into the vessel. Here, the positive ions have a probability to ionize residual gas, thereby producing ionization electrons that can reach the detector with energies indistinguishable from the signal beta electrons. Thus, if left undisturbed, the Penning trap will result in the production of an unacceptably high rate of background electrons. The creation of the Penning trap cannot be avoided as long as both spectrometer vessels are operated at negative high voltage (near -18 kV), which is the designed mode of operation for KATRIN1 . A mechanical means of emptying the Penning trap has been designed by collaborators at the University of Münster. In 2016, this device was installed on-site in the inter-spectrometer valve, which was built at CENPA. Known as the “Penning wiper”, the apparatus consists of a metal rod that can be periodically swept through the magnetic flux tube in order to remove trapped electrons. Three of these Penning wipers were installed in the system. During standard KATRIN operation, it is planned to insert a Penning wiper into the flux tube at regular intervals. However, the optimal frequency of this motion must be determined experimentally. During First Light measurements in Nov. and Dec. 2016, the inter-spectrometer Penning trap was investigated for the first time. With the main spectrometer set near to its nominal voltage (-18.6 kV), the pre-spectrometer voltage was incrementally ramped while the detector rate was monitored. When the measured detector rate surpassed a pre-defined value, the Penning wiper was automatically inserted in the flux tube. Rate spikes consistent with a filled Penning trap (Penning “discharges”) were observed when the pre-spectrometer was ramped to -2 kV (when operating the pre-spectrometer magnets at 80% of their nominal field). At this voltage, several frequency settings for the Penning wiper were also tested. No setting was able to prevent the production of a Penning discharge, but the movement of the wiper into the flux tube was generally able to quench a Penning discharge already in progress. Additionally, the wiper was also operated in a static configuration in which the wiper was fixed within the flux tube. This mode of operation is not ideal since the wiper obscures a sizable fraction of the flux tube and results in a reduction of the number of usable detector pixels for the neutrino mass analysis. However, measurements taken with the static Pen- ∗ University of Münster, Münster, Germany. † Karlsruhe Institute of Technology, Karlsruhe, Germany. ‡ Presently at Carnegie Mellon University, Pittsburgh, PA. 1 M. Prall et al., New Journal of Physics, 14, 073054 (2012).

  • Page 21

    6 ning wiper showed no Penning discharges when both spectrometers were operated near their nominal voltage settings. Since the measurements last year, a few upgrades have been made to the system. The position of the Penning wiper has been adjusted, since run data indicated that the wiper did not previously reach the center of the flux tube. This modification should help to ensure that all trapped electrons will be emptied by the movement of the Penning wiper. Additionally, upgrades to the electronics will allow for the simultaneous operation of all three Penning wipers. Further measurements with the Penning wiper are planned for the next data-taking phase: additional operating conditions to be studied include the effect of an increased mag- netic field in the beam line and a lower residual gas pressure (after the planned bake-out of the spectrometer vessels). In preparation for future tests, we have begun to implement routines related to ion transport and scattering in Kassiopeia, the KATRIN simulation code. We have also performed preliminary simulations of the trapped electrons in this region with various pre-spectrometer potential settings. 1.4 Main spectrometer background studies G. Drexlin∗ , F. Fränkle∗ , L. Kippenbrock, M. Kleesiek∗ , D. S. Parno† , and N. Trost∗ A background level of less than 10 millicounts per second is required in order for KATRIN to reach its design sensitivity1 . However, the observed background rate from the main spectrom- eter is over an order of magnitude larger than this value. A number of potential background sources have been studied to discover the origin of this background rate. One background of concern is due to the passage of cosmic ray muons through the walls of the main spectrometer vessel. Simulations indicate that over 35,000 muons pass through the main spectrometer every second2 . Muons produce secondary electrons when traveling through the steel walls of the vessel. A combination of magnetic and electrostatic shielding has been implemented to limit the effect of secondary electrons emitted from the surface, but only during commissioning measurements could the effectiveness of the shielding be determined. By looking at the correlation and coincidence between muon events (detected via nearby plastic scintillator panels) and FPD electron events, no measurable effect on the background rate could attributed to the cosmic ray muon flux with standard KATRIN field settings. A paper describing the muon background for KATRIN is currently in the final stages of preparation. The mechanism thought to be responsible for the majority of the observed background in the main spectrometer is due to the implantation of 210 Pb in the vessel walls3 . This ∗ Karlsruhe Institute of Technology, Karlsruhe, Germany. † Presently at Carnegie Mellon University, Pittsburgh, PA. 1 J. Angrik et al., “KATRIN Design Report”, 2005. 2 B. Leiber, “Investigations of background due to secondary electron emission in the KATRIN-experiment”, PhD thesis, KIT, 2014. 3 F. Harms, “Characterization and Minimization of Background Processes in the KATRIN Main Spec- trometer”, PhD thesis, KIT, 2015.

  • Page 22

    UW CENPA Annual Report 2016-2017 April 2017 7 isotope is a daughter of 222 Rn, which is present in ambient air, to which the interior of the main spectrometer was exposed for some time. The decay chain of 210 Pb includes the alpha decay of 210 Po, which can sputter atoms from the vessel walls. Specifically, this decay seems to cause the expulsion of hydrogen Rydberg atoms from the surface. Rydberg atoms are highly excited but neutral atoms. Their neutrality allows them to bypass the magnetic and electrostatic shielding that blocks most charged particles originating from the vessel walls. Upon entering the volume of the main spectrometer, Rydberg atoms can be easily ionized, even by blackbody radiation, thereby resulting in the production of low-energy ionization electrons that are indistinguishable from signal β electrons. To verify that the previously mentioned process is responsible for the observed background rate, a measurement was performed by collaborators at KIT in December 2016 in which the Rydberg background was temporarily enhanced using an isotope of lead with a much shorter half-life than that of 210 Pb. A 228 Th source was connected to the main spectrometer in order to expose the vessel to 220 Rn, which causes the implantation of 212 Pb in the spectrometer walls, eventually leading to the alpha decay of 212 Po on the surface. Assuming that the observed background comes from Rydberg atoms produced following the alpha decays on the surface, one expects to see a sizable increase in the background rate. Indeed, measurements showed an larger background rate, and, more importantly, it was observed that the increased background decayed away at a rate consistent with the half-life of 212 Pb. This experimental test, along with other measurements of the spectrometer background, confirms the above- mentioned background model. The effect of the high background rate on the sensitivity of KATRIN can be largely miti- gated by reducing the size of the magnetic flux tube used during neutrino mass measurements and measuring a larger portion of the β electron spectrum near the endpoint. Additionally, there is a plan during the next measurement phase to test the use of a UV light source to irradiate the spectrometer surface and thereby reduce the available hydrogen that can be excited into Rydberg states. 1.5 Electron detection for time-of-flight operation T. H. Burritt, P. J. Doe, E. L. Martin, D. A. Peterson, R. G. H. Robertson, and T. D. Van Wechel Implementing a time-of-flight measurement in the KATRIN experiment could allow measure- ment of the energy spectrum above each retarding potential, instead of only measuring the total rate of electrons above each retarding potential. This could significantly reduce the data collection time required to attain the same statistical uncertainty1 . In time-of-flight mode the energy of each electron is determined from the time it takes to pass through the main spectrometer. As flight time is mostly determined by the slow movement through the analyzing plane, where the remaining transverse momentum of the electron results in reduced energy resolution similar to high pass filter mode, the energy 1 Nicholas Steinbrink et al., New J. Phys. 15 113020 (2013).

  • Page 23

    8 resolution in time of flight mode could be nearly as good as the normal high-pass filter mode. The stop signal for time of flight is already available from electron impact on the focal plane detector, but a start signal is still required. The best place for the start signal is between the pre-spectrometer and main spectrometer. Multiple methods of detecting the passage of an electron have been explored, but none have been found suitable so far. A resonant cavity structure was tested that would exchange energy with the passing electrons through interaction with the cavity’s electric field, though the signal-to-noise ratio was too small even averaging millions of events. The signal from the cavities was a change in cavity oscillation amplitude, but it could be either an increase or decrease depending on cavity phase on arrival of the electron. As the method of trying to draw the signal from the noise was straight averaging of the signal, the signal to noise ratio was only improved as the fourth power of the number of events. An alternate signal filtering scheme was devised that would produce a unipolar signal by sampling at quarter waves of the carrier frequency and taking the sum of the square of the differences half periods apart. A new front end amplifier using pseudomorphic high-electron- mobility transistors (pHEMTs) is being explored. The amplifier is still in the design stage and needs additional refinement; simulations still put the signal to noise ratio at around 10-3 . A method of increasing cavity quality by using positive feedback is also being explored as a possible method of improving performance. Electrons trapped between the main spectrometer and pre-spectrometer would cause back- grounds too large for time-of-flight mode to be implemented. A removal method using a wiper that sweeps across the beam line was constructed and installed in the valve between the pre- spectrometer and main spectrometer. An alternative use of a passing electron detector, which would not require single electron sensitivity, may be to detect filling of the penning trap and trigger-wiper activation. 1.6 Firmware upgrade for high-rate measurements T. Bergmann∗ , S. Enomoto, and D. Tcherniakhovski∗ In the KATRIN neutrino-mass measurement, the rate of tritium β electrons at the detector is expected to be less than 1 cps (corresponds to ∼10 mcps/pixel), and the detector system is optimized for this low rate. For example, the charge-integrating preamplifiers are of the CR discharging type (as opposed to the “reset” type) with a long time constant of 1 ms, and the pulse shaping time is in the range of 1 to 10 µs. If two electrons arrive within the discharging time scale (1 ms), the second one will be affected by the tail of the first one, resulting in a lower estimated energy (tail pile-up). If two electrons arrive within the shaping-time time scale (1 to 10 µs), they will make only one pulse after shaping, resulting in a lower summed- energy and inaccurate electron counting (peak pile-up). If more than two electrons arrive within these time-scales, the behavior is further complicated. ∗ Karlsruhe Institute of Technology, Karlsruhe, Germany.

  • Page 24

    UW CENPA Annual Report 2016-2017 April 2017 9 Although the expected electron rate for the normal tritium runs is small, the need for high- rate operation has emerged. Calibrations with an e-gun, such as for the tritium gas density measurement and energy-loss measurement, directly benefit from high-rate measurements, and currently measurements at >50 kcps/pixel are proposed. To facilitate a new program to search for keV-scale sterile neutrinos with KATRIN, the apparatus will operate at very low retarding potentials, and the rate at the detector will reach ∼ 1 Mcps or more. Fig. 1.6-1 shows energy spectra of 18 keV electrons from an e-gun at two different per- pixel rates. At a low rate (left; 4 kcps), the energy peak is clear at 18 keV (there is a tiny peak at 36 keV from double-electron bunch from the e-gun) and electron counting is straightforward, although there is some constant contamination of noise events at low energy. At a high rate (right; 150 kcps), tail pile-up broadens and shifts the peak down to ∼14 keV, causing detection loss by pushing events below threahold. Peak pile-up causes a plateau- shaped spectrum between ∼14 keV and ∼32 keV for double-electron events. The overlap of single-electron events and double-electron events at around 14 keV makes electron counting difficult. Another plateau from triple-electron events is also visible above 35 keV (which actually extends from ∼14 keV to ∼40 keV, according to DRIPS simulation1 ). 4.2 kcps 150 kcps UnipolarEnergy 50000 UnipolarEnergy 1600 Entries: 18555 Entries: 811670 Mean: 18.317 ± 0.02775 Mean: 17.009 ± 0.0076526 RMS: 3.7796 ± 0.019622 45000 RMS: 6.8907 ± 0.0054112 1400 gaus gaus chi2/ndf: 14.4 / 9 40000 chi2/ndf: 50.632 / 12 Constant: 1541.3 ± 18.04 Constant: 44990 ± 87.73 1200 Mean: 18.094 ± 0.015618 Mean: 13.893 ± 0.0033746 Sigma: 1.2893 ± 0.020976 35000 Sigma: 1.6643 ± 0.0047932 1000 30000 800 25000 20000 600 15000 400 10000 200 5000 0 0 0 5 10 15 20 25 30 35 40 45 50 55 60 0 5 10 15 20 25 30 35 40 45 50 55 60 Recorded Energy (keV) Recorded Energy (keV) 16 Mar 2017 21:12, Energy-Orca32582.root 16 Mar 2017 21:09, Energy-Orca32588.root Figure 1.6-1. Energy spectra for 18 keV electrons from an e-gun at a rate of 4 kcps (left) and 150 kcps (right). Pile-up effects are visible in the high rate spectrum. The pole-zero cancellation technique, a common remedy for tail pile-up, cannot be used for this case because of multiple A/C couplings between the preamp and the ADC. Using two shapers in parallel, one with a short shaping time and one with a long (nominal) shaping time, is a common construct to reject peak pile-up; however, this is not feasible as our shaping time is limited by the ADC sampling interval and signal-to-noise ratio. On top of that, making a major modification to the current system was not acceptable given that the total detector system had already been thoroughly tested and the KATRIN tritium measurement had been scheduled in two years, as of 2013 when this upgrade was discussed. We solved these problems by developing a new bipolar shaping (trapezoidal) filter. The filter introduces another differentiation stage (a trapezoidal filter is a kind of differentiation filter) after standard pulse shaping which is “unipolar” in the sense that common CR-(RC)n filters produce gaussian-like pulse shapes, and the trapezoidal filter, which the KATRIN detector system implements in the FPGA, provides a trapezoidal shape, both of which are 1 CENPA Annual Report, University of Washington (2013) p. 13.

  • Page 25

    10 unipolar (pulses grow only in a more positive direction). Adding another differential stage makes the unipolar pulse bipolar, where one signal makes one positive pulse followed by one negative pulse. These bipolar pulses do not accumulate (because of vanishing net sum) and therefore no tail pile-up occurs. The time between the positive and negative components of the bipolar pulse depends upon the pulse rise-time, and is sensitive to the presence of multiple signals within one bipolar event. This can be used to resolve peak pile-up. In addition, the implemen- tation of bipolar shaping is purely an addition of another shaping stage and it will not affect the existing system in principle. The downside for this is inflated noise at low rate, which is ∼40% higher in equivalent noise charge (ENC) compared to the optimal unipolar shaping for the same shaping time. Further, compensation for ballistic loss is not straightforward compared to trapezoidal shaping. Fig. 1.6-2 shows the energy spectra with bipolar shaping for the same data as in Fig. 1.6- 1. The upgraded system can record both unipolar-shaped and bipolar-shaped energies for every event. Although the peak at low rate is wider (as predicted), the peak position remains unchanged at high rate. For low rates, we can simply use the unipolar-shaped energy for better energy resolution. 4.2 kcps 150 kcps BipolarEnergy BipolarEnergy Entries: 18585 Entries: 833100 1200 Mean: 18.487 ± 0.027697 40000 Mean: 20.86 ± 0.0072684 RMS: 3.7755 ± 0.019585 RMS: 6.6301 ± 0.0051396 gaus gaus chi2/ndf: 44.138 / 13 35000 chi2/ndf: 1339.4 / 18 1000 Constant: 1115.9 ± 13.185 Constant: 36619 ± 67.201 Mean: 18.196 ± 0.024036 Mean: 18.124 ± 0.0039614 Sigma: 1.8459 ± 0.034219 30000 Sigma: 2.1659 ± 0.0051912 800 25000 600 20000 15000 400 10000 200 5000 0 0 0 5 10 15 20 25 30 35 40 45 50 55 60 0 5 10 15 20 25 30 35 40 45 50 55 60 Recorded Energy (keV) Recorded Energy (keV) 16 Mar 2017 21:14, Energy-Orca32582.root 16 Mar 2017 21:11, Energy-Orca32588.root Figure 1.6-2. Bipolar-shaped energy spectra for 18 keV electrons from an e-gun at 4 kcps (left) and 150 kcps (right), showing that the peak position remains unchanged at high rate. Fig. 1.6-3 shows a two-dimensional plot of energy and peak-to-valley-time (i.e. the time between the bipolar positive and negative pulses) for 18 keV electrons at 150 kcps (the same dataset as in Fig. 1.6-2 right). The double-electron plateau structure in the energy spectrum (projection to the horizontal axis; equivalent to Fig. 1.6-2, right plot) between 18 keV and 36 keV has a distinguishably large peak-to-valley time at the energy of single-electron peak, ∼18 keV, enabling us to resolve the overlap in the energy spectrum. The well-separated cluster of triple-electron events is also visible on the right side. Overall, the new bipolar shaping filter behaves exactly as designed, and accurate electron counting is possible with the bipolar filter even above 100 kcps per pixel. Agreement with DRIPS simulation prediction (not shown here) is quite good and our next step is to use the simulation to develop an analysis method for electron counting.

  • Page 26

    UW CENPA Annual Report 2016-2017 April 2017 11 100 104 Bipolar Peak-to-Valley Time (CLK) 90 80 103 70 60 50 102 40 30 10 20 10 0 1 0 10 20 30 40 50 60 Bipolar Energy (keV) Figure 1.6-3. Peak pile-up separation using bipolar peak-to-valley time, for 18 keV electrons at 150 kcps from an e-gun. The color scale indicates the number of DAQ events per bin for the entire run. 1.7 Progress on veto upgrade T. H. Burritt, P. J. Doe, S. Enomoto, N. Fong, A. Müller∗ , D. A. Peterson, and T. D. Van Wechel The cosmic muon veto for the KATRIN FPD is being upgraded to take advantage of new silicon photo-multiplier (SiPM) models and recent technical advancements from other exper- iments. The current KATRIN veto yields two to three detected photons per cosmic muon per SiPM, while other experiments typically report much higher numbers such as 20 detected photons per muon per SiPM, even with similar detector construction. Due to the high dark rate of SiPMs at ∼100 kHz, 99% of our recorded data is currently dominated by accidental coincidences from dark noise, even with precise control of the SiPM temperature with Peltier coolers in a box with carefully regulated dry nitrogen flow. An upgrade to a high light-yield system will not only reduce the data size by ∼ 90% but will also improve the system’s ro- bustness, stability, and ease of operation by eliminating the cooling system and associated temperature regulation. It will then reduce systematic effects associated with the instability. The design of the new veto system and the results from prototype tests were reported last year1 . Based on the demonstrated superior performance (more than an order of magnitude higher light-yield compared to the existing veto), stability (no cooling necessary, stable against ∗ Karlsruhe Institute of Technology, Karlsruhe, Germany. 1 CENPA Annual Report, University of Washington (2016) p. 6.

  • Page 27

    12 operating condition fluctuations) and robustness (the fibers, the most fragile part of the system, are replaceable even after installation), we produced 12 stave panels, 3 end-cap panels, 6 front-end electronics modules, and 2 controller modules, as well as FPGA firmware and embedded software for control, monitoring, and automated calibration. The system was shipped to KIT in June 2016, then assembled and installed into the detector system. Fig. 1.7-1 shows the installation of the panels and electronics. Figure 1.7-1. Installation of the new veto panels and electronics. Prior to installation, all assembled panels were tested and characterized. Fig. 1.7-2 shows the SiPM hit rates for various coincidence conditions with two stacked stave panels. Neglect- ing accidental coincidence events, forming coincidences among different SiPMs eliminates SiPM dark hits, as SiPM dark hits are local to each SiPM. If coincidences are formed for SiPMs on different panels, it also eliminates panel-contained events, typically due to envi- ronmental γ-rays, and remaining events are only due to through-going cosmic muons (plus

  • Page 28

    UW CENPA Annual Report 2016-2017 April 2017 13 the neglected accidentals). As can be seen in Fig. 1.7-2, an ADC threshold of 200 will select almost all cosmic muon events while the contamination from accidental coincidence of SiPM dark hits and environmental gammas can be strongly suppressed. Figure 1.7-2. SiPM hit rates for various coincidence settings. Neglecting accidentals, red is by cosmic muons, the difference between blue and red is from environmental gammas, and the difference between black and blue is by SiPM dark noise. The lower part (ADC<200) of red is due to accidentals. The per-panel muon detection efficiency and the total event rate with the ADC<200 cut were measured with three stacked stave panels, where the coincidence of the top and bottom panels triggers through-going cosmic muons and the central panel is for evaluation. The result is summarized in Table 1.7-1. Note that the coincidence here is among SiPMs within one panel (central panel), where the panel-contained events such as environmental gammas are not suppressed. Given that the rate in the table is comparable to the actual cosmic muon rate (i.e., the contribution from the other sources does not inflate the data rate), we will record all the intra-panel coincidence events. Inter-panel coincidences will be applied in offline analysis; in this way, optimization of the cut can be performed based on the detection efficiency of the as-installed multi-panel geometry for the actual muon angular distribution, in order to achieve the optimal balance between live-time loss and other background sources. Table 1.7-1. Per-panel muon detection efficiency and event rates, for various coincidence settings. Coincidence Threshold Detection Efficiency (%) Hit Rate (cps/panel) 1 99.83 ± 0.01 178 2 99.75 ± 0.02 160 3 97.72 ± 0.05 136 4 94.30 ± 0.09 112 During the installation of the detector system, we found that the end-cap panels needed some slight modification to fit into the space. The modified end-cap panels were shipped to KIT in February 2017, and are waiting for the next detector service period scheduled in summer 2017.

  • Page 29

    14 1.8 Data quality for the KATRIN experiment S. Enomoto, L. Kippenbrock, and D. S. Parno∗ We are continuing to develop a system to ensure the quality of the data used during the final KATRIN analysis1 . Our primary goals are to exclude data taken during times when the apparatus is operating out of specification and could introduce unacceptable systematic effects, and to identify data for which special care is necessary in the analysis stage. In the last year, we have coordinated with the various subsystem task groups to generate a nearly complete draft list of several hundred sensors (out of more than 10,000 in the experi- ment) whose readings could indicate a critical problem with data quality. In certain cases, the data-quality indicators are not simple sensor readings, but the products of specialized offline analyses; laser Raman spectroscopy results, which give the isotopic composition of the source gas, are a prime example. This list of indicators is the first step toward building a database of data-quality criteria, which cannot be finalized until the completion of commissioning and sensitivity studies for all subsystems. With other collaboration members, we are leading a task group to draft an overall data- quality policy for the experiment, including sensor categorization and procedures for setting, documenting, and changing validity ranges. Several aspects of the data-quality infrastructure have now been implemented in KATRIN analysis code, including fine-grained, numerical alert levels that could provide early warning of dangerous drifts in operational parameters. We have also completed a stability study of data-quality indicators in the FPD system, based on a week of data-taking during a summer 2015 commissioning period. Fig. 1.8-1 shows the results for a typical indicator, the current drawn by the front-end electronics on the 8V supply line. Variation over time was quite small, with almost all readback values contained within a range of 1 mA. No correlation of the sensor value with detector rate was observed. We have proposed a fairly wide validity range that would have invalidated only 40 s of data over the weeklong test period. Figure 1.8-1. Stability data for a sensor in the focal-plane detector system, the current drawn by the front-end electronics on the 8V supply line. Left: Weeklong time series of readings in Summer 2015; blue dashed lines indicate the proposed validity window. The isolated spike to 0 A is due to a readback error. Right: Histogram of current readings. ∗ Presently at Carnegie Mellon University, Pittsburgh, PA. 1 CENPA Annual Report, University of Washington (2016) p. 3.

  • Page 30

    UW CENPA Annual Report 2016-2017 April 2017 15 TRIMS 1.9 Status of the Tritium Recoil-Ion Mass Spectrometer (TRIMS) L. I. Bodine∗ , T. H. Burritt, C. Claessens† , S. Enomoto, M. Kallander, Y. -T. Lin, E. Machado, R. Ostertag‡ , D. S. Parno§ , D. A. Peterson, R. F. H. Robertson E. B. Smith, T. D. Van Wechel, and D. I. Will The Tritium Recoil-Ion Mass Spectrometer (TRIMS) experiment is designed to measure the branching ratio to the molecular bound state 3 HeT+ following the β-decay of the molecule T2 1 . This branching ratio has been predicted by modern molecular final state theories, but the prediction disagrees with experimental results from the 1950s. Many other predictions from the same theoretical framework, however, do agree with the experiments. The TRIMS experiment will use modern instrumentation and analysis methods to re-examine this observ- able, so that the discrepancy can be addressed. The knowledge of this branching ratio will provided a verification of the theoretical framework which has been applied to describe the final state distribution of the β-decay spectrum measured by KATRIN. In March 2016, the TRIMS apparatus was moved from the Physics/Astronomy Building to CENPA Hot Lab2 . We completed the vacuum system reconstruction and reconnected the external electronics. We carried out performance tests for our equipments, such as leak-checking our vacuum system and testing the stability of our magnet power supplies and high-voltage power supply; the system has been stable. We have resumed data-taking with a CAEN DT5720 digitizer and run slow controls with a LabJack12 through the ORCA data-acquisition system3 . The BEANS and KAFFEE analysis codes from KATRIN4 were integrated into the TRIMS data-processing chain. We use BEANS to convert data, both the PIPS detector waveform data and slow-control data, from ORCA to ROOT and digi- tally process the waveforms through a pair of offline trapezoidal filters. We use KAFFEE to upload the processed data to a dedicated server. A new preamp was also designed to improve the resolution of the detectors. Currently, our best PIPS detector has an energy resolution of FWHM = 2.5 ± 0.4 keV, which essentially meets our stringent requirement of FWHM = 2.3 keV. In attempt to mitigate the high-voltage micro-discharges reported last year2 , we made several design changes during the reconstruction. Most importantly, we replaced the stainless- steel internal electrodes with a gold-plated electrode and a negative 100V-biased mesh elec- trode. The gold-plated electrode on the 60 kV end was designed to reduce the number of ions knocked out by β particles, whereas the mesh on the ground end was designed to prevent ∗ Departed March, 2016. † University of Mainz, Mainz, Germany. ‡ Karlsruhe Institute of Technology, Karlsruhe, Germany. § Presently at Carnegie Mellon University, Pittsburgh, PA. 1 CENPA Annual Report, University of Washington (2013) p. 15. 2 CENPA Annual Report, University of Washington (2016) p. 12. 3 CENPA Annual Report, University of Washington (2012) p. 25. 4 CENPA Annual Report, University of Washington (2014) p. 5.

  • Page 31

    16 the secondary electrons produced at ground from being accelerated by the high voltage. The mesh design also removes a Penning trap that arose in the previous design. The assembly work was done in a laminar flow cabinet to assure the requirements for ultra-high vacuum compatibility. While the performance was significantly better than that achieved in 2015, our high- voltage commissioning revealed that micro-discharges persisted. Analyzing the energy spec- trum, we observed that these micro-discharge events resemble proton (or hydrogen ion) events. We therefore performed a glow-discharge cleaning by supplying 3 kV to a decay chamber filled with 2 × 10−2 Torr of argon in order to drive out residual hydrogen on the β electrode. The cleaning removed almost all micro-discharge events, but it also sputtered gold atoms from the electrode to the inside surface of the decay chamber. We replaced the chamber and substituted a bakeout procedure for the glow-discharge cleaning. The bakeout, which should serve to remove residual hydrogen from the entire vacuum system, has been successfully completed. We are preparing for further high-voltage commissioning tests. 83m 1.10 The tritium and Kr gas-handling system for the TRIMS experiment M. Kallander, Y. -T. Lin, E. M. Machado, D. S. Parno∗ , and R. G. H. Robertson The TRIMS T2 gas-handling manifold was originally designed to diffuse three types of gases through a leak valve into the main vacuum system of the apparatus. The original gas system was a three-legged manifold constructed entirely of VCR4 and VCR8 hardware. T2 and H2 were to come from 50 cc cells on two of the legs, and a dry, neutral gas such as Ar or N2 would come through the third leg from a large compressed gas cylinder. The only means of pumping the manifold was through a leak valve, which was the single point of connection between the vacuum system and the manifold. In the course of using the manifold, this revealed itself to be an unanticipated shortcoming of the gas-handling system. Additionally, early on in the experiment, it was known that 83m Kr was going to be used as a calibration source of mono-energetic conversion electrons to test the silicon detectors used by TRIMS. It was envisioned that this testing could occur only once, at the very beginning of the experiment, after which the generator would be removed from a leg of the gas manifold. Later, it was decided that TRIMS should have a permanent, in situ 83m Kr generator to allow calibration of the detectors, as well as to allow for mapping of the electric field inside the decay chamber. The generator, previously used by the Project 8 experiment, takes the form of zeolite beads embedded with 83 Rb. The design of the gas-handling manifold has been modified to include this 83m Kr genera- tor, as well as a separate pumping line for the gas manifold. Fig. 1.10-1 shows a diagram of the TRIMS apparatus. The section labeled “Hydrogen Bottle” will eventually be replaced with the T2 cell. The addition of the section labeled “B” allows the gas-handling manifold to be pumped independently of the main vacuum system. The system is now assembled, pumped, ∗ Presently at Carnegie Mellon University, Pittsburgh, PA.

  • Page 32

    UW CENPA Annual Report 2016-2017 April 2017 17 baked, and ready to receive T2 . The TRIMS experiment underwent a UW Radiation Safety Office panel review, and was approved for the usage of T2 and 83 Rb. Decay Relief Z Chamber Turbo RV1 AV3 Purifier Regulator Cold Nitrogen UV Window Mixing AV1 X Trap Getter A SV1 Krypton AV2 AV4 SV3 Pirani Getter Glass Glass Insulator Insulator AV7 E CV1 CV2 Hydrogen Pirani SV2 SV4 Pirani Leak 50cc RGA Y AV5 LV1 B C SV5 D Ion SRG X: 2400cc A: 3100cc D: 110cc Hydrogen Ion Y: 510cc B: 260cc E: 62cc Bottle AV6 Z: 3700cc C: 72cc 50cc Figure 1.10-1. A schematic diagram of the TRIMS apparatus with calibrated volumes MAJORANA 1.11 Overview of the MAJORANA DEMONSTRATOR T. H. Burritt, M. Buuck, C. Cuesta∗ , J. A. Detwiler, Z. Fu, J. Gruszko, I. Guinn, D. A. Peterson, R. G. H. Robertson, K. Ton† , and T. D. Van Wechel The MAJORANA DEMONSTRATOR is a neutrinoless double-beta (0νββ) decay search employ- ing a 44.1-kg array of high-purity germanium (HPGe) detectors, 29.7 kg of which are enriched in 76 Ge. The experiment is currently running at the Sanford Underground Research Facility (SURF) in Lead, SD. The detectors are arranged in strings of 4 to 5 on low-background elec- troformed copper supports which are then placed in two cryogenic lead- and copper-shielded modules (see Fig. 1.11-2). The primary technical goal of the DEMONSTRATOR is the achieve- ment of a radioactive background of 3 counts/ton/year within a 4-keV region of interest surrounding the 2039-keV Q-value for 76 Ge 0νββ decay. Such a low background level would justify deeper investment in a much larger ton-scale experiment with sufficient sensitivity to definitively search for 0νββ decay for inverted hierarchical neutrino masses. In the past year, the collaboration completed construction of the second module of HPGe detectors, and the experiment is now running in low-background mode with both modules in a completed shield. The radioassay program instituted to minimize natural radioactivity in MAJORANA detector materials was published in 20161 . A first measurement of the total muon flux in the underground laboratory using the MAJORANA DEMONSTRATOR veto system was ∗ Presently at Centro de Investigaciones Energéticas, Medio Ambientales Y Tecnológicas, Madrid, Spain. † Departed in June, 2016. 1 N. Abgrall et al., Nucl. Instr. Meth. Phys. Res. A, 828, 22-36 (2016).

  • Page 33

    18 Figure 1.11-2. Schematic of the MAJORANA DEMONSTRATOR apparatus. recently accepted for publication1 . More recently we published a search for mono-energetic peaks in our low-energy data (5-100 keV), placing new limits on models of pseudoscalar and vector Dark Matter, solar axions, electron decay, and Pauli Exclusion Principle violation 2 . We have recently submitted a manuscript on our calibration system3 , and are preparing a submission on our enriched germanium processing R&D, as well as a number of other technical topics. We hope to release our first 0νββ decay limit in the coming months, and a full analysis of MAJORANA DEMONSTRATOR backgrounds later this fall. We are also preparing for joint analysis activities with the GERDA collaboration; this combination should achieve unprecedented sensitivity for 0νββ decay searches. The two experiments are already proceeding with plans to combine to build a larger, ton-scale apparatus, the Large Enriched Germanium Experiment for Neutrinoless ββ Decay (“LEGEND”). 1.12 Summary of recent results from the MAJORANA DEMONSTRATOR M. Buuck, C. Cuesta∗ , J. A. Detwiler, Z. Fu, J. Gruszko, and I. Guinn The MAJORANA DEMONSTRATOR was finished with assembly in 2016. The first module came online after the 2015 rework on December 31st, 2015 and the second module came online in August 2016. Both modules took data from August 25th to September 27th 2016 in order to establish the region-of-interest (1900 keV to 3000 keV) background for Critical Decision 4 (CD4). The background in this region is predicted to be independent of the energy, so it is used to determine the projected background for the neutrinoless double-beta (0νββ) decay region-of-interest (ROI), which is a small window around the Qββ -value of 2039 keV. Because the modules were constructed and commissioned independently, each module used separate data acquisition systems and the data were kept separate and analyzed independently. Data ∗ Presently at Centro de Investigaciones Energéticas, Medio Ambientales Y Tecnológicas, Madrid, Spain. 1 N. Abgrall et al., Astropart. Phys. 0927-6505 (2017). 2 N. Abgrall et al., arXiv:1612.00886 [nucl-ex], to appear in Phys. Rev. Lett. (2017). 3 N. Abgrall et al., arXiv:1702.02466 [nucl-ex] (2017).

  • Page 34

    UW CENPA Annual Report 2016-2017 April 2017 19 from Module 1 is labeled as Dataset 3 and data from Module 2 is labeled as Dataset 4. After analysis of these datasets showed no major problems with the running of both modules, we unified their DAQ systems and are now acquiring only one dataset at a time. Summary statistics of Datasets 3 and 4 Dataset 3 was acquired over a period of 32.37 days, of which 31.67 days was live (97.8%). Of that livetime, energy calibration of Module 1 or 2 took up 1.18 days and disruptive commissioning tests took up 0.57 days, with the remaining 29.91 days devoted to the 0νββ- decay search and background level estimation. Figure 1.12-1. Duty cycles of all complete datasets taken so far with the MAJORANA DEMONSTRATOR. Datasets 0, 1, 2, and 3 are with Module 1 only and Dataset 4 is with Module 2 only. Datasets 3 and 4 were acquired simultaneously and are used in this analysis. Of the livetime devoted to the 0νββ-decay search, some fraction is lost to a few sources of deadtime. In Dataset 3, 0.0025 days were lost due to dead time during the periodic (∼ 0.1 Hz/detector) firing of the pulser system. This pulser is implemented primarily to ensure that the detectors remain live during data collection. A further 0.0127 days were cut by the muon veto, and 0.0003 days were lost due to rare events such as digitizer channel dropout. Finally, removal of junk waveforms from the acquired data (pulser retriggers, transient elec- tronic noise, etc.) has an estimated efficiency to retain events in the ROI of 0.9995. This leaves effectively 29.88 days live for Dataset 3. The duty cycles of all complete datasets acquired so far are shown in Fig. 1.12-1. Module 1 contains 20 detectors enriched to approximately 88% 76 Ge and 9 detectors made from natural Ge. Of the enriched detectors, 16 are used in this analysis and have a combined active mass of 12.63 ± 0.19 kg, while 4 of the natural detectors are used in this analysis and have a combined active mass of 2.79 ± 0.06 kg. Multiplying the enriched active mass by the livetime for Dataset 3 gives the exposure used to determine the background level for CD4: 377.41 kg-days.

  • Page 35

    20 The Dataset 4 exposure is computed in the same way as the Dataset 3 exposure. Dataset 4 was live for 25.65 days (79.3%), with 1.17 days calibrating, 0.78 days commissioning, and 23.69 days taking 0νββ data. After subtractions for pulsers, the muon veto, channel dropout, and the data cleaning framework efficiency, the total livetime for DS4 is 23.57 days. Module 2 contains 15 enriched (with 7 used in DS4) and 14 natural (with 7 used in DS4) detectors for a total active mass of 5.47 ± 0.08 kg enriched and 3.95 ± 0.09 kg natural. These values produce an exposure of 128.85 kg-days for Dataset 4. Background level of Datasets 3 and 4 Several standard cuts are applied to the data before generating a final spectrum. First, only events occurring in a single detector are retained, because 0νββ-decay only occurs in one detector at a time. Some detectors are not fully depleted and so are only used to veto events with high multiplicity; these are removed from the analysis. Any waveform that is tagged as ”junk” by the data cleaning process is removed, and every event tagged as simultaneous with a cosmogenic muon interaction by the muon veto is also removed. A tag called A vs. E that identifies events interacting multiple times within a single detector is also applied and tagged events are removed. Finally, the delayed-charge-recovery (DCR) tag (Sec. 1.13) is applied. This tag identifies degraded alpha particle interactions along the passivated surfaces of the detectors by detecting the delayed collection of the electrons liberated in the interaction. After all of these tags are applied and the tagged events are removed, 2 events remain in Dataset 3 between 1900 keV and 3000 keV. This corresponds to an expected background of +0.18 0.16−0.10 counts/kg/month using Feldman-Cousins1 errors. In Dataset 4, zero events remain +0.31 after applying these cuts, which gives an expected background of 0−0 counts/kg/month. A histogram showing the energy spectrum of the sum of Dataset 3 and Dataset 4 is shown in Fig. 1.12-2. DS3 & DS4 - High gain (Enriched) Counts/40keV/kg/day 1 No PSD cuts AvsE cut AvsE & DCR cuts 10−1 10−2 10−3 500 1000 1500 2000 2500 3000 Energy (keV) Figure 1.12-2. Full energy spectrum for sum of DS3 and DS4. The black line is the spectrum with all cuts except for A vs. E and DCR applied. The red line is the spectrum with all cuts except for DCR applied, and the blue line is with all cuts applied. 1 Unified approach to the classical statistical analysis of small signals, Gary J. Feldman and Robert D. Cousins, Phys. Rev. D 57, 3873 (1998).

  • Page 36

    UW CENPA Annual Report 2016-2017 April 2017 21 This can be compared to the background measured in Dataset 1, which was taken with the same module before the copper external radiation shield was fully installed. In Dataset 1, +0.18 we obtained a projected background of 0.24−0.10 counts/kg/month. The energy spectra from Dataset 1, Dataset 3, and Dataset 4 obtained after applying all cuts are shown overlaid in Fig. 1.12-3. The result from Datasets 3 and 4 can also be compared to the Department of En- ergy’s Optimal Performance Parameter goal for background which was 0.6 counts/kg/month after all cuts. We also had an internal design goal of 3 counts/ton/year in the small ROI +5.7 around Qββ (2039 keV). In those units, the result from DS3 is 5.1−3.2 counts/ton/year and +8.6 the result from DS4 is 0−0 counts/ton/year in the 2039 keV ROI. The combined dataset +4.1 gives a rate of 3.7−2.3 counts/ROI/ton/year. As more data is collected, we will obtain a more precise measurement of our background levels, and will eventually be able to determine whether or not we have hit our goal of 3 counts/ROI/ton/year. As most of the events near the ROI appear to be from degraded alpha-particle interactions, our background could po- tentially be further reduced by improving pulse-shape or other kinds of tags of those events. It is also important to understand the ultimate source of these particles so that they can be better mitigated in a future experiment. Enriched- All cuts (High energy) 10−1 Counts/keV/kg/day M1 (DS3) M2 (DS4) M1 (DS1) 10−2 10−3 10−4 500 1000 1500 2000 2500 3000 Energy (keV) Figure 1.12-3. Comparison of energy spectrum for DS1 (Module 1 only), DS3 (Module 1 only), and DS4 (Module 2 only). Of particular interest is the significant reduction in activity near the 0νββ region-of-interest from DS1 to DS3 and DS4. 1.13 Alpha particle discrimination and TUBE T. Bode∗† , M. Buuck, C. Cuesta‡ , J. A. Detwiler, J. Gruszko, I. Guinn, S. Mertens∗† , and M. Willers∗ Alpha particles pose a problematic background in large granular detector arrays, particularly those alphas that originate from Rn progeny that plate out on the detector surfaces during manufacturing and assembly. The geometry of the p-type point contact (PPC) detectors ∗ Technical University of Munich, Munich, Germany. † Max Planck Institute, Munich, Germany. ‡ Presently at Centro de Investigaciones Energéticas, Medio Ambientales Y Tecnológicas, Madrid, Spain.

  • Page 37

    22 implemented in the MAJORANA DEMONSTRATOR make the DEMONSTRATOR relatively in- sensitive to alphas, with the exception of a contribution coming from the passivated surface between the point contact and outer dead layer. In this region, the response to alphas is difficult to characterize. The charge collection properties near this surface can differ for dif- ferent detector models. In the MAJORANA DEMONSTRATOR, events have been observed in which alphas incident on or originating from this surface are significantly degraded in energy, leading to a potential background contribution in the ROI for 0νββ. However, it is also observed that charge mobility is drastically reduced on or near the passivated surface, and is slowly released on the timescale of waveform digitization, leading to a measurable change in slope of the tail of a recorded pulse. It is expected that this effect is due to surface propagation of the electron contribution to the signal, while the holes are collected normally. This matches the model developed by Mullowney et al 1 . Using a filter that can identify the occurrence of this delayed charge recovery (DCR), these events can be identified, allowing for the efficient rejection of passivated surface alpha events in analysis. Validating this model and fully characterizing the alpha interaction rejection efficiency of such a filter requires alpha source scans of the specific detector geometries used. To that end, the CENPA group has collaborated with the GERDA and MAJORANA groups at Technical University of Munich (TUM) to rebuild and adapt an existing alpha-source scanner. This internal scanning cryostat, the TUM Upside down BEGe (TUBE) experiment, is currently being used to take an alpha source scan of a MAJORANA detector. The early results from these measurements have confirmed our model of alpha interactions and validated the use of the DCR pulse shape discrimination technique. The Delayed Charge Recovery (DCR) Cut for Alpha Particle Interactions An alpha interaction cut, which identifies waveforms with a low-mobility charge component like those of Fig. 1.13-1, was developed in 2016 using commissioning data from Module 1 of the DEMONSTRATOR. Based on the results of signal-to-background optimization studies of these commissioning runs, it has been implemented with a single-site bulk (gamma and double-beta decay) event acceptance efficiency of 90% ± 4.1%. The bulk acceptance is determined using calibration spectra. The uncertainty in the acceptance is driven by pulse shape deviations, and is quantified by comparing the acceptance in the 228 Th double-escape peak (a known sample of single-site events without significant charge loss) to the acceptance over a range of energies. A new version of the DCR discriminator, which corrects for the effect of charge trapping in the detector bulk and thus reduces this systematic uncertainty, has been developed and is currently being tested. 1 P. Mullowney, M. C. Lin, K. Paul, et al, NIM A, 662, 1 (2012) 33-44.

  • Page 38

    UW CENPA Annual Report 2016-2017 April 2017 23 Figure 1.13-1. Left: Sample average normalized TUBE waveforms with pole-zero (pulse decay) correction. Waveforms are from alpha source scans with the source incident on the detector surface at different distances (r) from the point-contact, in red and blue, and from a run without the alpha source incident on the detector, in black. The alpha event waveforms show a slow component contribution from the drift of charge along the passivated surface, which is used to identify these events. The shape of the waveforms closely matches that predicted in simulations. Right: Sample simulated waveforms, incorporating passivated surface drift of electrons. The interactions originate at different radii (r) and depths (z) in the detector, with (r,z) = (0,0) mm being the center of the point-contact. The use of the DCR discriminator in the MAJORANA DEMONSTRATOR data sets reduces the rate of background events in the 0νββ region-of-interest by a factor of 25 (Sec. 1.12). Further improvement is expected from the switch to data-taking incorporating multisampling, which will allow longer waveforms to be recorded, and from future improvements to the DCR analysis. Characterizing Alpha Interactions with TUBE Given the importance of the DCR discriminator to the MAJORANA DEMONSTRATOR’s re- sults, the decision was made to scan a MAJORANA detector using the existing TUBE scan- ner1 while continuing preparations to build a similar scanner at CENPA. The TUBE scanner was adapted to accept the dimensions and IR-shine sensitivity of PONaMA-1, a natural- abundance detector made by ORTEC for the MAJORANA Collaboration that has the same geometry as the 76 Ge-enriched detectors currently taking data in the MAJORANA DEMON- STRATOR. In February 2017, a spectral-grade 241 Am source was installed in the system and high-quality data-taking began. Analysis of the TUBE scanning data has begun, and early results (see Fig. 1.13-2) confirm that the DCR parameter identifies alpha event waveforms effectively at a range of radial scanning positions, even when the events cannot be clearly identified by their energy. The 1 M. Agostini, Dissertation, Technical University Munich (2013).

  • Page 39

    24 waveforms of TUBE alpha scan events confirm the interaction model used in our simulations (see Fig. 1.13-1). Figure 1.13-2. Data from TUBE taken without the alpha source incident on the detector, in black, with the source incident on the passivated surface 8.25 mm from the point-contact, in red, and 12.75 mm from the point-contact, in blue. Left: Energy spectra taken with TUBE. The energy of the alpha peak increases with increased scanning radius, as is expected if the electrons are the primary source of delayed charge. Right: The DCR distribution in TUBE. The 90% DCR cut falls at the green line; events falling to the right of this line are rejected. Alpha scan distributions show a clear excess of events at high values of DCR. 1.14 Searching for 2νββ to excited states T. H. Burritt, M. Buuck, C. Cuesta∗ , J. A. Detwiler, Z. Fu J. Gruszko, I. Guinn, D. A. Peterson, R. G. H. Robertson, and T. D. Van Wechel While 2νββ from 76 Ge to the ground state of 76 Se has been observed, the same decay into an excited state of 76 Se has not yet been seen. The MAJORANA DEMONSTRATOR is searching for 2νββ to excited states in 76 Ge. There are three possible excited state decay modes allowed, as shown in Fig. 1.14-1. Due to the reduced Q-values of these modes, these modes have much + longer half-lives than the ground state mode. The 0+ g.s. − 01 decay has the shortest half-life of all the excited state decay modes, expected to be between 1.2 × 1023 yrs − 6.7 × 1024 yrs. The excited state decay modes are accompanied by at least one γ-ray each (at 559.1 keV and + 563.2 keV for 0+ g.s. − 01 ). These photons give the excited state decay modes a unique signa- ture that reduces their backgrounds significantly, making detection by the DEMONSTRATOR feasible. A candidate event would be a multi-detector event, with at least one event falling in a region of interest around the 559.1-keV and 563.2-keV γ energies. Recognizing these events requires identification of multi-detector coincidence events and proper calibration of the detectors. The MAJORANA event builder is the software that detects coincidence events. The event ∗ Presently at Centro de Investigaciones Energéticas, Medio Ambientales Y Tecnológicas, Madrid, Spain.

  • Page 40

    UW CENPA Annual Report 2016-2017 April 2017 25 76 + Figure 1.14-1. Level diagram for 2νββ in Ge. The 0+ g.s. − 01 mode is the most common excited state decay. builder is responsible for converting raw binary files produced by the ORCA DAQ software into ROOT files that are usable by the MAJORANA DEMONSTRATOR anaylsis software. The builder also detects and removes waveforms that have been corrupted. Finally, the builder combines waveforms from multiple detectors that are nearby in time into events. In the last year, the event builder has been upgraded in order to support multi-sampled waveforms from the Gretina digitizers. These digitizers use different sampling rates for the baselines and the falling edge of the waveform than the rising edge, enabling a longer timing window for each waveform without sacrificing information useful for discriminating multi-site events. The event builder has also been upgraded to support ORCA’s quick start mode, which enables dead-time free data collection. The event builder is currently being used to process all of the data for the MAJORANA DEMONSTRATOR. The MAJORANA DEMONSTRATOR’s detectors are calibrated using γ-ray peaks from a 228 Th line source. The spectral lines from these peaks are fit to a peakshape function that is the sum of a gaussian component, low- and high-energy exponentially modified gaussian tail compo- nents, and a step component. Multiple peaks from a calibration spectrum are simultaneously fit, with the peakshape parameters describing the calibration constants and resolution as functions of peak energy. The results of such a fit are shown in Fig. 1.14-2. This technique enables the fitting of small spectral peaks that would not converge if fit on their own. More peaks allows easier study of calibration systematics, such as nonlinearities in the digitizer response. The simultaneous fit is performed using an adaptive Hybrid Monte Carlo (HMC) algorithm. HMC is a gradient-based Markov Chain Monte Carlo (MCMC) technique, which has a convergence time that scales well with number of correlated parameters compared to other MCMC methods. This is necessary due to the large number of parameters, many of which experience strong correlations, involved in a simultaneous fit.

  • Page 41

    26 Figure 1.14-2. The results of the multipeak fitter run on uncalibrated data. Peaks were fit using the Gaussian (teal) and low-energy tail (gold) components. Peaks were fit at 238.6 keV, 241.0 keV, 277.4 keV, 300.1 keV, 583.2 keV, 727.3 keV, 860.6 keV, and 2614.5 keV. Project 8 1.15 Status of the Project 8 neutrino-mass experiment A. Ashtari Esfahani, S. Böser∗ , R. Cervantes, C. Claessens∗ , L. de Viveiros†‡ P. J. Doe, S. Doeleman§ , M. Fertl, E. C. Finnk , J. A. Formaggio¶ , M. Guiguek , K. M. Heeger∗∗ , A. M. Jonesk , K. Kazkaz†† , B. H. LaRoque† , A. Lindman∗ , E. Machado, B. Monreal‡‡† , J. A. Nikkel∗∗ , N. S. Oblath¶k , W. Pettus, R. G. H. Robertson, L. J. Rosenberg, G. Rybka, L. Saldaña∗∗ , P. Slocum∗∗ , J. R. Tedeschik , T. Thümmler§§ , B. A. VanDevenderk , M. Wachtendonk, M. Walter§§ , J. Weintrub§ , A. Young§ , and E. Zayas¶ Project 8 is an effort toward an improved neutrino mass determination by measuring the frequency of the cyclotron radiation emitted by tritium beta decay electrons spiraling in a magnetic field. The technique, called Cyclotron Radiation Emission Spectroscopy (CRES), ∗ Institut für Physik, Johannes-Gutenberg Universität Mainz, Germany. † University of California, Santa Barbara, CA. ‡ Pennsylvania State University State College, PA. § Harvard-Smithsonian Center for Astrophysics, Cambridge, MA. k Pacific Northwest National Laboratory, Richland, WA. ¶ Massachusetts Institute of Technology, Cambridge, MA. ∗∗ Yale University, New Haven, CT. †† Lawrence Livermore National Laboratory, Livermore, CA. ‡‡ Case Western Reserve University, Cleveland, OH. §§ Institut für Kernphysik, Karlsruher Institut für Technologie, Karlsruhe, Germany.

  • Page 42

    UW CENPA Annual Report 2016-2017 April 2017 27 combines the rich history of single-electron trapping with high precision frequency analysis techniques to provide a unique avenue to high resolution electron spectroscopy. The project is organized into four phases to reach the anticipated sensitivity to the neutrino mass of 40 meV. Figure 1.15-1. Spectrogram of the 83m Kr lines at 30.4 keV demonstrated resolution of 3.3 eV approaching the natural line widths1 . The publication of a high resolution 83m Kr conversion electron spectrum (Fig. 1.15-1) spanning over the K, L, M, and N shells of krypton marked the success of phase I of the experimental effort1 . This was the first demonstration of the CRES technique. More sophisti- cated analysis schemes are being developed to facilitate more accurate energy reconstruction2 . Building on the success of phase I, an upgraded tritium compatible cell was developed and built. The new cell, shown in Fig. 1.15-2, is based on a waveguide with a circular cross section increasing both the signal-to-noise ratio and effective volume of the system. The magnetic trap consists of five “z 2 ” coils, which can act independently as harmonic traps, or in concert to form a broad flat region with pinches at the end (termed a “bathtub” trap). To probe the magnetic field generated by the trapping coil, a new Electron Spin Resonance (ESR) magnetometer has been installed on the insert. The new insert was installed last summer and is currently being commissioned with conversion electrons from 83m Kr. Later, the cell will be filled with molecular tritium gas. The microwave background noise in the phase II experiment has also vastly improved compared to phase I, as can be seen in Fig. 1.15-3. Installation of a cryogenic circulator prevents the amplifier noise from being reinjected into the cell. As a result, the overall noise level for the apparatus has decreased. The new circulator has also smoothed noise oscillations by terminating the standing wave in between the first amplifier and waveguide short, which was placed at the bottom end of the waveguide to increase the SNR level by reflecting electron signal back to amplifiers. Another phase II milestone was the development of a tritium handling system. The gas- handling manifold, initially constructed at Yale University, allows the safe handling of the 1 D.M. Asner et al., Phys. Rev. Lett. 114, 1162501 (2015). 2 A. Ashtari Esfahani et al. J Phys. G. 44 054004 (2017).

  • Page 43

    28 98 phase 1, T = 073K Phase 1 phase 1, T = 096K phase 1, T = 117K phase 1, T = 134K phase 1, T = 156K 100 phase 1, T = 163K phase 1, T = 175K Noise Spectrum [dBm/Hz] phase 2, T = 48 K phase 2, T = 58 K phase 2, T = 68 K 102 phase 2, T = 71 K phase 2, T = 90 K phase 2, T = 100 K phase 2, T = 110 K phase 2, T = 120 K 104 Phase 2 106 500 750 1000 1250 1500 1750 2000 2250 2500 Frequency [MHz] Figure 1.15-3. Noise level comparison between Figure 1.15-2. Phase II cell with five trapping phase I and II over a frequency band of 2 GHz. coils and ESR probes. The gas will be loaded Each curve in the plot represents the noise into the cell through the gas inlet. The cali- power at a specific temperature. The direct bration port is also being used to inject signals relation between the noise level and temper- into the waveguide for calibration purposes. ature is apparent in the plot. The phase II This cell is compatible with both Krypton and apparatus has also demonstrated a clear re- Tritium data taking. duction in both noise level and its oscillation. tritium gas. In addition, it provides an 83m Kr calibration source of monoenergetic conversion electrons with energies near the tritium β-decay end point. The design of the system was modified such that the gas-handling system would attach to the existing Project 8 phase II apparatus. Project 8 underwent a UW Radiation Safety Office review panel of the gas- handling manifold and received permission to use T2 . The gas manifold will contain a 2 Ci inventory of T2 and 5 mCi inventory of 83 Rb which serves as a generator for the 83m Kr. The new system also has the capability to be used in phase III (Fig. 1.15-4). Figure 1.15-4. Evan Zayas works on the combined tritium and krypton gas system. This system will be used to source and probe both krypton and tritium in the system. It has been designed for use in both Phase II and III of the experiment.

  • Page 44

    UW CENPA Annual Report 2016-2017 April 2017 29 One of the current efforts of the Project 8 collaboration is the integration of a more sophis- ticated data acquisition system with a significantly higher trigger rate and a multi channel approach that can be extended to the phase III and IV data acquisition requirements. For this purpose the collaboration is currently commissioning a ROACH2 board (Reconfigurable Open Architecture Computing Hardware) which was developed for antenna arrays in radio astronomy. On the ROACH2 board, the incoming signal is digitized at 3.2 GSPS (after the signal has been downsampled by 24.2 GHz). The signal is then fed to an FPGA with three band widths that can be individually tuned to different central frequencies. Following down-conversion, the signal is split and its Fourier transform is computed. Both the time- and frequency domain representation of the signal are then streamed to a server via 10 GbE. With this combination of real-time Fourier transform computation and fast data packet re- ception and processing on the server it is possible to operate either in streaming or triggered mode with no dead-time. Since January 2017 the ROACH2 is connected to the analog signal of the Project 8 experiment and first sets of data have been acquired using the streaming mode of the DAQ software. Fig. 1.15-5 shows 1 second of data with many electron tracks visible. This data has an event rate of a few tracks per second, which could not be recorded with the former Real-time Spectrum Analyzer due to its long dead time and high trigger threshold. Figure 1.15-5. One second of data recorded by the new DAQ system, ROACH2. The spectrogram shows the power in each bin of time and frequency. The lines of higher power are the signal from individual electrons. Phase III planning efforts are underway and will require a larger volume of magnetic field, which requires a new method for harvesting the microwave signal using a phased array of antennas. With the larger instrumented volume we can achieve a neutrino mass sensitivity mν < 2 eV, competitive with current limits from Mainz1 and Troitsk2 . To acquire the necessary statistics over a period of one year, we estimated a required volume of 200 cm3 . It is therefore not practical to conduct this experiment in an enclosed waveguide detector as in phases I and II. For phase III, Project 8 must enlarge its tritium volume so that trapped electrons will emit cyclotron radiation into free space. A used MRI magnet has been acquired and is on site to accommodate the larger experiment, with an open bore of 90 cm. 1 C. Kraus et al. 2005 Eur. Phys. J.C 40:447 (2005). 2 V.N. Aseev et al. Phys. Rev. D 84(11) 112003 (2011).

  • Page 45

    30 A phased array of antenna elements is an effective means to collect free-space radiation. We are investigating one or more ring-shaped arrays of antenna elements with each element amplified and digitized independently. The total available (coherent) signal power increases linearly with the number√of instrumented channels N , while the incoherent noise of each channel contributes only N to the total noise. Performance has been modeled for an 8-cm- radius ring with 48 channels, the maximum that will fit around the circumference (Fig. 1.15- 6). A SNR of 9 dB can be attained in this design. Figure 1.15-6. An 8-cm-radius ring array composed of 48 open-ended waveguide antenna elements in free space. Phases are tuned such that the focus is on the cylindrical axis (left) or 4 cm from the center (right). The goal of phase IV is sensitivity to the full range of neutrino masses allowed by the assumption of an inverted mass hierarchy. An idealized estimate of the Project 8 sensitivity is mν < 40 meV (90% C.L.). In order to circumvent the fundamental limit set by final-state broadening with a molecular T2 source, the Project 8 collaboration is planning to make use of atomic tritium for the first time. 1.16 Development of an atomic source for Project 8 S. Böser∗ , J.Ȧ. Formaggio† , M. Guigue‡ , A. Lindman∗ , E. Machado, R. G. H. Robertson A very attractive feature of the Project 8 method, and the goal of its fourth phase, is the possibility of performing a tritium beta decay experiment with atomic tritium. Doing so removes the line-broadening contribution caused by zero-point motion in the T2 molecule, about a 1-eV FWHM line-width contribution with both statistical and systematic penal- ties for molecular-based experiments. Fig. 1.16-1 compares the atomic and molecular line- widths 1 . The endpoint energy for the atomic decay is about 10 eV smaller than the molecular decay, which places stringent requirements on the purity of an atomic source. A molecular ∗ Institut für Physik, Johannes-Gutenberg Universität Mainz, Germany. † Massachusetts Institute of Technology, Cambridge, MA. ‡ Pacific Northwest National Laboratory, Richland, WA. 1 L. I. Bodine, D. S. Parno, and R. G. H. Robertson; Phys. Rev. C 91 035505 (2015).

  • Page 46

    UW CENPA Annual Report 2016-2017 April 2017 31 10 Atomic T (at -8.1 eV) 8 T2 Relative probability 6 4 2 0 -2 -1 0 1 2 Energy (eV) Figure 1.16-1. Line-width contributions for atomic and molecular tritium. The atomic line is Doppler-broadened, as shown for a temperature of 1K. The molecular broadening is mainly due to zero-point motion in the T2 molecule. contamination larger than 10−5 would represent a significant background at the atomic end- point. Fortunately, the operating conditions for an atomic trap are compatible with very low temperatures (T < 2 K) and molecules are untrapped and freeze out. The basic requirements for an atomic trap as applied to Project 8 are as follows: • a means for producing atomic tritium, • cooling to temperatures at which atoms can be trapped magnetically, • injection into the trap, • and trapping with a useful lifetime of roughly 105 s or greater. We initially considered a cooling scheme involving a chamber with walls at ∼ 150 mK, and the use of 4 He or 3 He as an exchange gas. Atoms of T injected into a region with a static vapor pressure of He would be cooled collisionally, with the exchange gas carrying heat to the walls. Since neither He isotope has a significant magnetic moment, the He is not trapped, but the T can be. More detailed consideration revealed that the scattering cross section of He on H is anomalously low, which calls for a higher density. Given the dependence of vapor pressure on temperature, even for 3 He the resulting temperature is too high for practical magnetic trapping. We have instead turned to a novel scheme that does not involve an exchange gas. Atomic T produced in a source that could be a hot W cracker or an RF discharge is cooled by accommodation on a cold surface to about 30 K. The atoms then pass through a velocity selector that picks out a slice of the thermal distribution having a width corresponding to approximately 30 mK and a central energy equivalent to a magnetic field of 3 T for low-field- seeking states.

  • Page 47

    32 The velocity and state selected beam is prepared in a relatively low field of 0.05 T and enters the magnetic trap by surmounting a wall of 2 T. At the peak field, an RF transition drives the atoms to the high-field seeking state, and they slow further as they enter the central trap volume where the field is 1 T. Finally, circularly polarized radiation drives the atoms irreversibly back to the low-field seeking state, leaving them trapped in the active volume. Fig. 1.16-2 shows the cooling cycle of RF transitions where the central field is 1 T and the axial pinch field is 2 T. Preliminary calculations suggest that the desired density 1 of 1012 40 pinch   20 selector   Frequency, GHz a  à  d     d  à  a     28.804  GHz   56.819  GHz   0 -20 trap   -40 0.0 0.5 1.0 1.5 2.0 2.5 3.0 Magnetic Field, T 15   Figure 1.16-2. Breit-Rabi diagram for atomic tritium showing the RF transitions proposed for trapping atoms injected over a magnetic barrier of 2 T into a trap with a central field of 1 T. cm−3 could be reached and maintained. With substantial levels of RF needed to induce the spin flips, a cyclical fill-measure-fill regimen might be called for, but it may be possible to configure the magnetic trap field arrangement to provide enough isolation between the RF drive and the low-noise amplifiers. SNO 1.17 SNO and the solar-neutrino reaction hep R. G. H. Robertson, N. Tolich, and T. Winchester∗ Following more than a half-century of experimental study of the solar neutrino spectrum, all but 2 of the sources of solar neutrinos have been observed. One is the CNO cycle, limited ∗ Departed August 2016. Presently at Bellevue College, Bellevue, WA. 1 A. Ashtari Esfahani et al. (the Project 8 Collaboration); J. Phys. G: Nucl. Part. Phys. 44 054004 (2017).

  • Page 48

    UW CENPA Annual Report 2016-2017 April 2017 33 experimentally now to about 1% of the total solar neutrino flux, and the other is the hep reaction, p +3 He → 4 He + e+ + ν, (1) which produces the most energetic neutrinos in the solar spectrum. Although the predicted rate for this reaction is very low, the neutrino energy places the upper end of the detected spectrum in a region above 15 MeV that is largely free of interfering backgrounds. The Sudbury Neutrino Observatory (SNO) was an underground, 1 kton heavy water Cherenkov detector in Sudbury, Ontario, Canada, detecting both charged- and neutral- current neutrino-deuteron interactions. The experiment was designed to detect solar neu- trinos, and provided the first clear evidence of neutrino oscillations with a measurement of the neutrino flux from the decay of 8 B. An initial “box” analysis of SNO data to identify hep neutrinos 1 produced two candidate events and an upper limit on the integral hep flux of 2.3 × 104 cm−2 s−1 at 90% CL. A new analysis 2 makes a number of improvements to increase SNO’s sensitivity to hep neutrinos. A new event fitter was developed for existing data from SNO that improves both energy and vertex reconstruction. The fitter is used to remove backgrounds that previously limited the fiducial volume, which is increased by 30%. Since the hep events have relatively high energies, data not previously used because, for example, a calibration source was present, could be included in the analysis. A modified Wald-Wolfowitz test was employed to increase the amount of live time by 200 days (18%) and show that the additional data is consistent with the previously used data. A Bayesian analysis technique was developed to make full use of the posterior distributions of energy returned by the event fitter. In the first significant detection of hep neutrinos, we find (Fig. 1.17-1) the most probable rate of hep events is 3.5 × 104 cm−2 s−1 , which is significantly higher than the theoretical prediction. We also find that the 95% credible region extends from 1.0 to 7.2 × 104 cm−2 s−1 , therefore excluding a rate of 0 hep events at greater than 95% probability. 0.030 0.030 Theoretical Theoretical Posterior Posterior 0.025 0.025 95 % credible region 95 % credible region 0.020 0.020 P(Φh ) P(Φh ) 0.015 0.015 0.010 0.010 0.005 0.005 0.000 0.000 0 20 40 60 80 100 0 20 40 60 80 100 Φh (103/cm2/s) Φh (103/cm2/s) Figure 1.17-1. Posterior distribution for the hep flux from the new analysis. The theoretical prediction is given by the red line. The blue region represents the 95% confidence interval which excludes the theoretical prediction. 1 B. Aharmim et al. (the SNO Collaboration). Astrophys. J. 653, 1545-1551 (2006); hep-ex/0607010. 0.025 2 T. Winchester, PhD thesis, University of Washington, 2016. Theoretical Posterior 95 % credible region 0.020 0.015 P(Φh ) 0.010

  • Page 49

    34 COHERENT 1.18 The COHERENT experiment A. Cox, C. Cuesta∗ , J. Detwiler, A. Eberhardt, E. Erkela, Z. Fu, G. Garvey, D. S. Parno† , D. Peterson, T. D. Van Wechel, and A. Zderic In 2016 we received a UW Royalty Research Fund Award to perform R&D toward fielding an array of NaI(Tl) scintillating detectors at the Spallation Neutron Source (SNS) in Oak Ridge. The SNS produces a pulsed source of stopped pions, which decay to neutrinos with energies reaching tens of MeV. One goal of the experiment would be to measure the coherent elastic neutrino-nucleus scattering (CEνNS) cross section of these neutrinos with the 23 Na nucleus, a Standard Model process that has yet to be observed. Another goal would be to perform an improved measurement of the charged-current interaction cross section of neutrinos with 127 I. This measurement would test several aspects of nuclear transitions relevant to neutrinoless double-beta decay, such as the potential quenching of gA . These goals are being pursued in conjunction with the COHERENT Collaboration, who is deploying additional detector tar- gets for the CEνNS search, including CsI(Tl) scintillating crystals, an array of Ge detectors, and a liquid xenon time projection chamber. CENPA came into possession of a large number of NaI(Tl) crystals that were surplussed from the failed Department of Homeland Security Advanced Spectroscopic Portal (ASP) program. We received 130 crystals of dimension 2” × 4” × 16” with photomultiplier tubes (PMTs) attached, giving a total mass of 1.0 tons. Preliminary testing of the crystals at UW and elsewhere shows that they are high quality. The PMTs require high-voltage (HV) power supplies and a voltage divider (PMT base) to operate. We have designed a PMT base that incorporates the socket and HV power supply recovered from the custom bases produced for ASP. A prototype implemented on a perfboard was tested and implemented on printed circuit boards. The design is now being merged with a dual-output amplifier designed at ORNL to achieve sufficient dynamic range to perform both the low-energy CEνNS search and the high-energy CC scattering measurement with the same data run. Meanwhile, we assisted Duke collaborators with the installation of a prototype 24-crystal array at the SNS (see Fig. 1.18-1). The detectors are held in a compact array surrounded by water shielding to block neutrons from the SNS beam as well as natural gamma radiation in the environment surrounding the detector. The detector has been running since Summer 2016 and data is being shipped via internet to the CENPA mamba cluster for redundant storage and parallel analysis. ∗ Presently at Centro de Investigaciones Energéticas, Medio Ambientales Y Tecnológicas, Madrid, Spain. † Presently at Carnegie Mellon University, Pittsburgh, PA.

  • Page 50

    UW CENPA Annual Report 2016-2017 April 2017 35 Figure 1.18-1. UW post-bac researcher Alex Zderic (right) deploying the 24-crystal proto- type with Duke colleagues at the SNS in Oak Ridge, TN. We have also assisted with validation of and upgrades to the neutrino source-simulation code, originally written in GEANT4 by H. Ray and collaborators at the University of Florida. The original estimates for the detector locations in the simulation were improved, and we verified the experimental hall layout using building surveys provided by ORNL. We also began to modify the output data format to improve flexibility, reliability, and readout speed. Separately, we are working on simulations of both large and small (24 crystal) NaI(Tl) arrays to support the analysis of the test deployment.

  • View More

Get the full picture and Receive alerts on lawsuits, news articles, publications and more!