avatar University of Washington Inc Services


  • Page 1

    ANNUAL REPORT Center for Experimental Nuclear Physics and Astrophysics University of Washington May, 2006 Sponsored in part by the United States Department of Energy under Grant #DE-FG02-97ER41020.

  • Page 2

    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 photos, from top to bottom: Stephan Schlamminger and Todd Wagner, adjusting the apparatus for the Eöt-Wash equivalence principle test. Jessica Mitchell, assembling the e- gun for testing the KATRIN detector. The prototype PIN diode pixilated array, built by the Washington Technology Center, for the KATRIN detector.

  • Page 3

    UW CENPA Annual Report 2005-2006 May 2006 i INTRODUCTION CENPA pursues a broad program of research in nuclear physics, astrophysics and related fields. Research activities are conducted locally and at remote sites. CENPA is a major participant in the Sudbury Neutrino Observatory (SNO), the KATRIN tritium experiment and the Majorana double-beta decay experiment. The current program includes “in-house” research on nuclear collisions and fundamental interactions using the local tandem Van de Graaff, as well as local and remote non-accelerator research on fundamental interactions and user-mode research on relativistic heavy ions at large accelerator facilities in the U.S. and Europe. We thank our external advisory committee, Baha Balantekin, Russell Betts, and Stu- art Freedman, for their continuing valuable recommendations and advice. The committee reviewed our program in May, 2005. A comprehensive analysis of the complete data taken in the Sudbury Neutrino Observa- tory when salt was present in the heavy water was completed and published. The SNO data provides a precise measure, 33.9+2.4 −2.2 degrees, of the “solar” mixing angle. In combination with other experiments (KamLAND in particular), the mass splitting ∆m212 is found to be (8.0+0.6 −0.4) × 10 −5 eV2 . The Neutral-Current Detection array has been operating in production mode in SNO since November 2004. The solar-neutrino-live fraction has been good, > 60 %, and most of the remaining time is calibration data. The array produces a clear neutron signal at the level anticipated. The UW-designed data acquisition systems, ORCA for NCDs and SHaRC for the photomultiplier system, have been integrated and are operating successfully. We finished a determination of the mass of the lowest T=2 state in 32 S with a pre- cision/accuracy of approx 0.3 keV. This allowed for a stringent test of the Isobaric Mass Multiplet Equation, where we found a significant discrepancy. Our result provides the best test of the limits of the approximations inherent in the IMME and of its utility for predicting masses away from the valley of stability. Construction of the KATRIN experiment is proceeding apace with contracts placed for the major components and construction started on the experimental hall. Commissioning has started on the prespectrometer using the ORCA based data acquisition system. The prespectrometer’s internal electrode, that was supplied by CENPA, has successfully passed the extreme high vacuum test. In the US, the detector design document and a project execution plan have been completed, allowing detailed design and cost estimating to begin. Using the CENPA electron gun, we have been working closely with manufacturers to optimize the properties of PIN diode arrays we are considering for the focal plane detector. The experiment remains on track to begin data taking in Fall ’09. The Majorana Scientific Collaboration proposes to search for neutrinoless double-beta decay by building an array of 86% enriched 76 Ge segmented radiation detectors that serve as both source and detector. In September, the collaboration received a favorable review by the Joint NSAC-HEPAP Neutrino Scientific Assessment Group (NuSAG)1 sub-committee. 1 Neutrino Scientific Assessment Group, Recommendations to the Department of Energy and the National

  • Page 4

    ii Efforts at CENPA this past year have concentrated on developing a conceptual design report, carrying out a variety of simulation studies within the Majorana- GERDA (MaGE) simulation framework, and R&D studies of the potential for future use of an active liquid Ar shield. The Object Oriented Realtime Control and Acquisition (ORCA) system, which was devel- oped at CENPA, has been expanded to support many additional CAMAC modules, several stepper motor controllers, and general process control capabilities. ORCA has been success- fully taking production data for the SNO NCD experiment and, in addition, is being used in development systems at UW, LANL, PNNL, FZK, and MIT associated with the SNO, KATRIN, and Majorana projects. In nuclear astrophysics, we have begun our precision 4 He + 3 He fusion cross section measurements, and we have produced several test 23 Na and 22 Na targets for our planned 22 Na(p,γ)23Mg study. We have observed that much of the correlation structure in RHIC heavy ion collisions is dominated by low-Q2 parton scattering and fragmentation, observed for the first time with our analysis techniques. Because low-Q2 parton scattering is possibly the dominant formation mechanism and the best probe of the colored medium produced at RHIC we have pursued this phenomenon in elementary collisions, first in p-p collisions and more recently with an extensive analysis of e+ -e− fragmentation functions. A coherent picture of nonperturbative QCD processes at small energy scales is emerging. The HBT interferometry analysis activity in the STAR experiment at RHIC continues to provide provocative results that challenge theoretical ideas. Our recent work in developing the distorted wave emission function (DWEF) model has proved very successful in simultaneously reproducing HBT radii and the magnitude and shape of the pion momentum spectrum. New results show that to explain STAR data the DWEF model prefers an emission temperature of 193 MeV, the same temperature that lattice gauge calculations predict for the transition from a quark-gluon plasma to a hadronic phase in the medium. We have also shown that the space-time part of the emission function can be scaled with participant number to the one-third power to fairly accurately predict the observables of non-central Au+Au collisions and central and non-central Cu+Cu collisions. We have developed a “spin pendulum” for a novel torsion-balance test of CP and Lorentz symmetries. Our upper limit on the energy required to reverse the direction of an electron spin about an arbitrary direction fixed in inertial space is roughly 10−21 eV, which is comparable to the electrostatic energy of two electrons separated by 10 astronomical units. Our value is well below the benchmark expectation, based on the electron and Planck masses, m2e /MP = 2 × 10−17 eV. Our tests of the gravitational inverse-square law have shown that the law holds down to the “dark energy length scale” of 85 micrometers. We are developing a next-generation instrument that should increase our sensitivity at the 50 micron length scale by about a factor of 50. Science Foundation on a United States Program in Neutrinoless Double Beta Decay: Report to the Nuclear Science Advisory Committee and the High Energy Physics Advisory Panel, 2005.

  • Page 5

    UW CENPA Annual Report 2005-2006 May 2006 iii Five CENPA graduate students obtained their PhD degree during the period of this report. 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 Prof. Derek W. Storm, Executive Director, CENPA, Box 354290, University of Washington, Seattle, WA 98195; (206) 543- 4080, or storm@npl.washington.edu. 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 or to be quoted without permission of the authors. In each article the names of the investigators are listed alphabetically, with the primary author underlined, to whom inquiries should be addressed. Derek Storm, Editor Barbara Fulton, Assistant Editor

  • Page 6

    TANDEM VAN DE GRAAFF ACCELERATOR A High Voltage Engineering Corporation Model FN purchased in 1966 with NSF funds, operation funded primarily by the U.S. Department of Energy. See W. G. Weitkamp and F. H. Schmidt, “The University of Washington Three Stage Van de Graaff Accelerator,” Nucl. Instrum. Meth. 122, 65 (1974). Recently adapted to an (optional) terminal ion source and a non-inclined tube #3, which enables the accelerator to produce high intensity beams of helium and hydrogen 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. Additional ion species 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. In addition, we are now producing a separated beam of 15-MeV 8 B at 6 particles/second. BOOSTER ACCELERATOR See “Status of and Operating Experience with the University of Washington Superconducting Booster Linac,” D. W. Storm et al., Nucl. Instrum. Meth. A 287, 247 (1990). The Booster is presently in a “mothballed” state.

  • Page 7

    UW CENPA Annual Report 2005-2006 May 2006 v Contents INTRODUCTION i 1 Neutrino Research 1 SNO 1 1.1 Status of the SNO Project . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 SNO NCDs 2 1.2 NCD-array pulse-shape fitting . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 1.3 NCD electronics calibrations . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 1.4 Livetime studies in the SNO NCD phase . . . . . . . . . . . . . . . . . . . . 4 1.5 Determination of the efficiency function of the NCD MUX system through raised threshold source runs . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 1.6 SNO muon chamber data acquisition verification . . . . . . . . . . . . . . . . 7 KATRIN 8 1.7 The CENPA contribution to the KATRIN neutrino mass experiment . . . . . 8 1.8 The design of the KATRIN detector system . . . . . . . . . . . . . . . . . . . 10 1.9 Electron gun for profiling silicon detectors for KATRIN . . . . . . . . . . . . 12 1.10 Developing a complete Geant4 simulation of the KATRIN detector region . . 13 1.11 Estimating background rates for KATRIN from cosmic rays ionizing the gaseous tritium source . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 1.12 KATRIN preamplifier design and development . . . . . . . . . . . . . . . . . 15 Majorana 17 1.13 The Majorana neutrinoless double-beta decay experiment . . . . . . . . . . . 17 1.14 DAQ and SOH for Majorana . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 76 1.15 The Majorana sensitivity to excited-state double-beta decays of Ge . . . . . 19 1.16 Development of simulation and analysis frameworks for Majorana . . . . . . . 21

  • Page 8

    vi 1.17 MaGe simulation of LANL clover detector with AmBe neutron source . . . . 22 1.18 Simulation of the SLAC electron beam dump experiment using MaGe . . . . 23 1.19 Studies of α-contamination on high purity Ge crystals . . . . . . . . . . . . . 24 1.20 LArGe, liquid argon compton supressed germanium crystal . . . . . . . . . . 25 1.21 Thermal modeling of the MEGA cryostat cooling . . . . . . . . . . . . . . . 26 1.22 Monte Carlo simulation studies of the LoMo counting facility . . . . . . . . . 27 2 Fundamental Symmetries and Weak Interactions 28 Torsion Balance Experiments 28 2.1 Torsion balance search for spin coupled forces . . . . . . . . . . . . . . . . . . 28 2.2 Testing the weak equivalence principle using a rotating torsion balance . . . . 30 2.3 Progress toward improved torsion fibers . . . . . . . . . . . . . . . . . . . . . 31 2.4 Tests of the gravitational inverse-square law below the dark-energy length scale 32 2.5 Small force investigations for LISA . . . . . . . . . . . . . . . . . . . . . . . . 33 2.6 Progress on the torsion pendulum based axion search . . . . . . . . . . . . . . 34 2.7 Eöt-Wash data acquisition development . . . . . . . . . . . . . . . . . . . . . 35 2.8 Test of Newton’s second law for small accelerations . . . . . . . . . . . . . . . 36 2.9 Status of the APOLLO lunar laser ranging project . . . . . . . . . . . . . . . 37 2.10 APOLLO laser safety system . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 Weak Interactions 39 2.11 The electron-capture branch of 100 Tc . . . . . . . . . . . . . . . . . . . . . . . 39 2.12 Parity non-conserving neutron spin rotation experiment . . . . . . . . . . . . 41 2.13 Status of the ultra-cold neutron Aβ experiment at Los Alamos . . . . . . . . 42 2.14 Monte Carlo calculations for experiments at LANL (UCNA) and the ILL . . 43

  • Page 9

    UW CENPA Annual Report 2005-2006 May 2006 vii 3 Nuclear Astrophysics 45 3.1 Astrophysical S-factor for the 3 He + 4 He reaction: energy loss and beam heating in the gas cell . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 3.2 Ge efficiency measurements and calculations for the 4 He + 3 He experiment . 47 3.3 Completion of the measurement of 8 B beta decay . . . . . . . . . . . . . . . . 49 3.4 Completion of the measurement of 8 B beta decay: Si detector response . . . 50 22 3.5 Measurement of the Na(p,γ)23Mg reaction rate . . . . . . . . . . . . . . . . 51 4 Nuclear Structure 53 32 4.1 The F t value of the β-delayed proton decay of Ar . . . . . . . . . . . . . . 53 4.2 Measurement of the absolute γ branches in the decay of 32 Cl . . . . . . . . . 54 32 33 4.3 S(p, γ) and its importance in calibrating a Ar beta-delayed proton spectrum 55 4.4 Mass of the lowest T = 2 state in 32 S: A test of the Isobaric Multiplet Mass Equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56 5 Relativistic Heavy Ions 57 5.1 Opacity and chiral symmetry restoration in heavy ion collisions at RHIC: the DWEF Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57 5.2 Using lattice-gauge predictions in DWEF calculations . . . . . . . . . . . . . 59 5.3 Non-Gaussian HBT Correlations . . . . . . . . . . . . . . . . . . . . . . . . . 61 5.4 Studies of energy losses of fast charged particles . . . . . . . . . . . . . . . . 63 5.5 Summary of event structure research . . . . . . . . . . . . . . . . . . . . . . . 64 5.6 Describing fragmentation functions with beta distributions . . . . . . . . . . . 65 5.7 Identified hadron fragment distributions . . . . . . . . . . . . . . . . . . . . . 67 5.8 Fragment distributions from identified partons . . . . . . . . . . . . . . . . . 68 5.9 Parton energy dependence of fragmentation functions . . . . . . . . . . . . . 69 5.10 Scaling violations from beta-distribution fragmentation functions . . . . . . . 70 5.11 Correlations on transverse rapidity and fragmentation functions . . . . . . . . 71

  • Page 10

    viii 5.12 Angular asymmetry of low-Q2 parton fragments and non-perturbative parton scattering and fragmentation . . . . . . . . . . . . . . . . . . . . . . . . . . . 72 5.13 Precision centrality determination in A-A collisions . . . . . . . . . . . . . . 73 5.14 Two-particle correlations with identified particles . . . . . . . . . . . . . . . . 74 5.15 CI and CD angular correlations in 200 GeV Au-Au collisions . . . . . . . . . 75 5.16 Novel fluctuation and correlation analysis methods . . . . . . . . . . . . . . . 76 5.17 GUI for submission and monitoring of STAR data analysis jobs . . . . . . . . 77 5.18 Web interface for viewing histograms of data analysis . . . . . . . . . . . . . . 78 6 Electronics, Computing, and Detector Infrastructure 79 6.1 Electronic Equipment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79 6.2 Improvements to the ORCA DAQ system . . . . . . . . . . . . . . . . . . . . 80 6.3 Single-molecule electrophoresis of RNA through a biological nanopore . . . . 81 6.4 Laboratory computer systems . . . . . . . . . . . . . . . . . . . . . . . . . . . 82 7 Accelerator and Ion Sources 84 7.1 Ion sources, injector deck and accelerator crew training . . . . . . . . . . . . . 84 7.2 Van de Graaff accelerator operations and development . . . . . . . . . . . . . 85 7.3 Physical plant maintenance, repairs and possible upgrade . . . . . . . . . . . 86 8 Outside Users 87 8.1 Studies of the low energy fission of the actinides using surrogate reactions . . 87 8.2 Department of Astronomy Beowulf cluster . . . . . . . . . . . . . . . . . . . . 89 8.3 Molecular dynamics of proteins and peptides . . . . . . . . . . . . . . . . . . 90 9 The Career Development Organization: year six. 91 10 CENPA Personnel 92 10.1 Faculty . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92 10.2 CENPA External Advisory Committee . . . . . . . . . . . . . . . . . . . . . . 92

  • Page 11

    UW CENPA Annual Report 2005-2006 May 2006 ix 10.3 Postdoctoral Research Associates . . . . . . . . . . . . . . . . . . . . . . . . 92 10.4 Predoctoral Research Associates . . . . . . . . . . . . . . . . . . . . . . . . . 93 10.5 Research Experience for Undergraduates participants . . . . . . . . . . . . . . 93 10.6 Professional staff . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94 10.7 Technical staff . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94 10.8 Administrative staff . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94 10.9 Part Time Staff . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95 11 Publications 96 11.1 Published papers: . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96 11.2 Invited talks, abstracts and other conference presentations: . . . . . . . . . . 99 11.3 Degrees Granted, Academic Year, 2005-2006 . . . . . . . . . . . . . . . . . . . 105

  • Page 12

  • Page 13

    UW CENPA Annual Report 2005-2006 May 2006 1 1 Neutrino Research SNO 1.1 Status of the SNO Project J. F. Amsbaugh, G. A. Cox, J. A. Detwiler, P. J. Doe, C. A. Duba, G. C. Harper, M. A. Howe, S. R. McGee, A. Myers, N. S. Oblath, K. Rielage,∗ R. G. H. Robertson, L. C. Stonehill,∗ T. D. Van Wechel, B. VanDevender and J. F. Wilkerson The Sudbury Neutrino Observatory detector is now in its third configuration for solar neutrino detection, in which 3 He-filled proportional counters have been deployed in the heavy water. There are 36 ‘strings’ of individual counters filled with 3 He and another 4 filled with 4 He for investigation of backgrounds. The total deployed length of counter is 398 m. Production running in this mode began in November, 2004, and will continue until December 31, 2006, at which time the collaboration will begin decommissioning SNO in order to return the heavy water to its owners. The array has been running successfully and produces a clear neutron signal that is approximately of the magnitude expected from neutral-current disintegration of deuterium by solar neutrinos. Data are being recorded under blindness protocols designed by the UW group. Two strings (one 3 He and one 4 He) are presently out of commission with breakdown or high-rate noise problems. Another eight have various minor problems that can be dealt with in data processing. Data are recorded in two parallel streams, one a high-speed integrating shaper-ADC path to handle bursts such as a supernova, and the other a low-rate path in which the ionization current signals are fully digitized at 1 Gs/s for 15 µs. The information contained in the two data streams provides powerful means for sorting valid neutron events from other types. Analysis of the radioactivity in the “NCDs” by studying the time-correlated alpha decays characteristic of the U and Th chains indicates that U is about a factor of 10 below the specified 25 pg/g level, and Th is about a factor of 3 above the specified 2 pg/g level. The additional photodisintegration background implied is about 5% of the solar neutrino signal, which compromises slightly but not seriously the obtainable accuracy. During this phase of SNO operation UW is responsible for the maintenance of the NCD array, and we have on occasion sent technical personnel to site on short notice to deal with problems. We maintain complete functional systems at UW and in the underground control room at SNO to aid in tests, code development, and diagnostics. Many subsidiary mea- surements in support of analysis have been carried out with the UW system. We continue to provide our yearly average of 150 SNO operator shifts on site, the largest number (to- gether with Oxford) in the collaboration. Both data-acquisition (DAQ) and NCD experts are available on call round the clock. Presently at Los Alamos National Laboratory, Los Alamos, NM 87545.

  • Page 14

    2 SNO Neutral Current Detectors (NCDs) 1.2 NCD-array pulse-shape fitting N. S. Oblath and R. G. H. Robertson We are developing a method for fitting the pulses from the Neutral-Current-Detection (NCD) Array at SNO. The NCD Array is a set of 40 strings of 3 He and 4 He proportional counters installed in the heavy-water volume of SNO. The purpose of the array is to capture neutrons from the Neutral-Current (NC) neutrino interaction that occurs in the heavy water (νx +d → n+p+νx ). The pulses primarily come from neutron-capture events, 3 He(n,p)3 H, or the alpha- decays of radioactive contaminants in the NCDs themselves. The back-to-back proton and triton in the neutron-capture events and the alpha particles ionize the gas in the counters, and the ionization electrons are collected on the central-wire anode creating a current pulse. To accurately count the number of neutrons detected we need to be able to separate neutron pulses from backgrounds such as alphas. The pulse-shape analysis group within SNO is working on several methods of identifying pulses. Our method uses deconvolution to simplify the pulses prior to fitting them with simulated pulses. The entire algorithm is a three-stage process. First, after the pulse is extracted from the data, two pre-fitting routines are used. One determines the z-position of the pulse, or where along the length of the NCD the event occurred. This routine takes advantage of the fact that certain frequencies in the frequency-space representation of a pulse will cancel upon reflection from the bottom of the NCD. By looking for those minima in each pulse the z position can be determined. Determining the z position is necessary so the reflection can be deconvolved in the second stage. The z-position-determination algorithm has had the best results so far of any previous z-position efforts. The other prefitting routine makes initial guesses for three remaining event parameters that describe the track orientation and radial location. This will be done with a multilayer- perceptron neural-network by the ROOT class TMultiLayerPerceptron. Two networks will be trained, using sets of simulated neutrons and alphas respectively. The two networks will be used to guess the parameters under the assumptions that the event was due to an alpha particle and that it was due to a neutron capture. The second stage, deconvolution, removes various effects that modify the pulse shapes, such as the reflection of the pulse from the bottom of the NCD string, and the various electronics components. The deconvolution is done in Fourier space. A point-spread function is simulated for each effect, which describes how a delta function is transformed. Once all of the deconvolutions are complete the pulse is transformed back into the time representation. The final stage is the actual pulse fitting using simulated neutron and alpha pulses. Minuit (as implemented in the ROOT class TMinuit) will be used to minimize the χ2 between the data and simulation, once assuming the event is from an alpha particle, and once assuming the event is from neutron capture. Fits of simulated pulses have shown that the fitting routine works well, and work is under way to identify the best way to separate neutron-capture and alpha events. Care is needed because the parameter space is highly correlated. All stages of the process have been shown to work and efforts to optimize each of them are underway.

  • Page 15

    UW CENPA Annual Report 2005-2006 May 2006 3 1.3 NCD electronics calibrations G. A. Cox, M. A. Howe, S. R. McGee, A. W. Myers, K. Rielage, R. G. H. Robertson, T. D. Van Wechel and J. F. Wilkerson The NCD Electronics Calibration Program consists of a weekly probing of the system’s shaper/ADC cards, multiplexers, and digital oscilloscopes. The tests include threshold mea- surements, linearity checks, and characterization of the multiplexers’ logarithmic amplifiers. A known test signal is injected at the front end of the electronics and its output is measured, thereby showing the effects of the various electronic settings. These automated tests, which have been performed since the beginning of data acquisition, have shown that our electronics system is stable.1 Work on these electronics calibrations over the last year has been more of the same. Due to a more advanced analysis approach of the NCD data, our software model of the NCD electronics system has become increasingly more sophisticated. The transformation of the physics and calibration pulses through each component in the system is now included. Methods are also being developed to create a single “transfer function” that can be used to convolve the effects of the electronics with simulated NCD events. In April 2005, a special data set was produced to facilitate characterization of our elec- tronics model. The data acquisition system is robust enough to allow us to acquire data with the hardware in various configurations. Using various setups, the pulse transformation effects caused by individual electronics components were isolated and measured. One problem that had been plaguing the electronics calibration analysis was the deter- mination of the uncertainties of the parameters which characterize the logarithmic amplifiers in the system. The uncertainties estimated by the MINUIT package from a χ2 minimization routine for an single event were smaller than expected, based upon a very large sample of measurements. The uncertainties returned by the fitting routine were approximately a factor of 10 too small. The problem was due to small error bars on each bin of the digitized wave- form. The reason that our bin errors were too small was because they were estimated from the RMS of the baseline. However, the bandwidth in our system filters out some of the high frequency noise and effectively smoothes our baseline (the digitizers have a 1GHz sampling rate). The bandwidth filtering produces a correlation between each of the 1ns wide bins of the waveform to its nearby bins, a smaller measured RMS along the baseline, and subsequently results in a smaller bin error estimate. Monte Carlo and simulation software was written to produce simulated waveforms. By producing a very large number of simulated waveforms and then performing a χ2 minimization analysis and histogramming the resulting best-fit logamp parameters, the uncertainty of the logamp parameters can be accurately estimated from their distributions. Automation of this work is currently ongoing. 1 CENPA Annual Report, University of Washington (2005) p. 35.

  • Page 16

    4 1.4 Livetime studies in the SNO NCD phase J. A. Detwiler Two important changes to the SNO detector in the NCD phase impact the estimate of the experiment’s livetime, essential for converting event counts into measured fluxes. The first is the addition of the NCDs themselves, which are operated as a distinct detector system from the PMTs used in the experiment’s previous data taking phases and hence require their own livetime estimate. The second change is the incorporation of the “NHIT monitor” diag- nostic, whose operation imposes a negligible deadtime but significantly impacts the livetime calculation. The NCD and PMT signals are digitized and read out with independent electronics and data acquisition systems. The livetime for the PMT data stream is obtained from a 10 MHz clock on the master trigger card (MTC), which is referenced to the GPS time standard at periodic intervals. The NCD system has its own 10 MHz clock to determine its livetime. However, since this clock is not directly referenced to the GPS time standard, we have measured the relative frequency shift and jitter between it and the 10 MHz clock on the MTC. The measured frequency shift and jitter were small enough that estimating the NCD livetime from the 10 MHz clock on the MTC would contribute negligible uncertainty. The same comparison was performed with a second 50 MHz clock on the MTC. The two comparisons were consistent with each other. The NHIT monitor is a diagnostic that regularly measures the thresholds of the NHIT triggers in order to identify drifts in those thresholds quickly so that action can be taken when necessary to maintain trigger efficiency while keeping data taking rates at acceptable levels. The diagnostic works by pulsing a particular set of PMT channels while measuring the response of the trigger. The process imposes a negligible deadtime on the detector; however, a fraction of the NHIT monitor events “steal” PULSE GT triggers, a periodic trigger signal whose count forms the basis of a run’s livetime estimate. As a result, the NHIT monitor triggers must be identified and incorporated into the the PULSE GT count in order to correctly calculate livetime. A worry was expressed that NHIT monitor events were not being tagged correctly by the trigger, and therefore would be miscounted. In response, an independent method to efficiently tag the NHIT monitor events was developed based on their hit pattern in the PMT channels. From this, it was demonstrated that the NHIT monitor events were indeed being properly tagged by the trigger, and hence the livetime calculation could be performed without additional complications.

  • Page 17

    UW CENPA Annual Report 2005-2006 May 2006 5 1.5 Determination of the efficiency function of the NCD MUX system through raised threshold source runs S. R. McGee and L. P. Mannisto The Neutral Current Detector (NCD) data acquisition system at the Sudbury Neutrino Ob- servatory (SNO) has two independent triggering paths; fast and slow. The fast (or shaper) path triggers on the integrated energy of an event. Intended to collect as much data as possible during a supernova-neutrino shower, this path also allows for a determination of the neutron acceptance efficiency of the NCDs during high-rate neutron source runs. The slow (or MUX) path triggers on the instantaneous current of the NCD event and allows for particle identification through various pulse-shape analysis algorithms. Any event acceptable for analysis in the NCD phase must have triggered both the MUX and the shaper systems to allow for identification of the pulse and an accurate measure of its energy, respectively. Calibrations are done periodically at SNO with neutron-emitting sources to calibrate the nominal MUX threshold efficiencies. However, due to changes in the NCD hardware or increased noise pick-up in the NCD system from ambient sources, it occasionally is necessary to temporarily raise the MUX thresholds on some or all of the NCDs to keep the MUX trigger rate low. The level to which the threshold is raised is chosen to minimize the effect of dead- time in that particular instance, so it is necessary to know accurately the MUX threshold efficiency over a wide range of threshold values. To determine these efficiencies when the MUX threshold is raised, a collection of source runs were taken in various locations (to optimize the neutron flux through a particular set of NCDs) and with various values of raised MUX thresholds. Since this is a test of the correlation efficiency between two triggering systems, it is necessary to optimize the neutron efficiency while minimizing the amount of noise that may effect the efficiency calculation (i.e., events that may naturally trigger only the shaper or the MUX independently). For this reason, runs where the 67-Hz AmBe source is less than 1 m from the NCD are being considered for the determination of the raised MUX threshold efficiency. Runs where the source is further than 1 m may be used as a check of consistency. To determine the efficiency function, a fit is made to the efficiency (i.e., the fraction of events in which both a shaper and a MUX trigger are present) versus raised MUX-threshold value as shown in Fig. 1.5-1. This fit is made at least three times for different increasing shaper lower limits. As can be seen in the plot, as the lower limit of the shaper increases the efficiency remains higher for increased range of MUX-threshold values. Currently, a two-dimensional polynomial function has been shown to fit well to most strings for most shaper ranges. Refinement of the fitting technique may include adding a second function in the range of lower MUX threshold values for better fits to the higher shaper ranges. Also, a method is being investigated to better determine the error bars on the MUX threshold values.

  • Page 18

    6 Figure 1.5-1. Plot of the efficiency of neutron acceptance on string 0 versus the MUX threshold value setting above the nominal value (in bits). Each curve is a fit to data taken with a fixed value of the shaper lower limit. The left most curve corresponds to the lowest setting of the lower limit. The curves to the right correspond to increasing shaper lower limits, illustrating a trend of increasing efficiency as the lower limit of the shaper is increased. Data shown here were taken during a central source run (source is at the midplane of SNO and is placed less than 1 m from string 0).

  • Page 19

    UW CENPA Annual Report 2005-2006 May 2006 7 1.6 SNO muon chamber data acquisition verification M. A. Howe and B. L. Wall An external measurement of through-going muons at the Sudbury Neutrino Observatory is underway. The purpose is to verify the accuracy of muon-track reconstruction algorithms used in SNO. The University of Washington contribution to the project is the development and verification of the data acquisition system for the project. MIT is handling the detector assembly. The full detector system includes eight gas drift chambers, measuring 101”x47”x3”, in a 4x2 stack running in a Geiger-counter mode. Each chamber has 16 channels connected to a LeCroy 3377 TDC. Two scintillator paddles are used to trigger the TDC for read out. Data readout and control is done via a G4 Macintosh by the ORCA program developed by Mark Howe. (a) (b) Figure 1.6-2. (a) Scintillator paddle and muon chamber setup. (b) Muon spectrum from the scintillator paddles. A scaled down version was assembled here at the University of Washington (see Fig. 1.6- 2). A single channel of a muon chamber ran in coincidence with two scintillator paddles in order to verify the operation of the data acquisition system. Four half-hour runs were done in common-start single-word mode, at 4 nanosecond resolution, with a 4 microsecond window. Each channel would count up to 16 hits in a triggered event, which means multiple peaks generated in an event would all get counted. Data in Fig. 1.6-2 from an ADC connected to the PMTs verifies that triggers being generated are from muons. A clear peak centered at bin 479 is seen from minimum ionizing particles. A total of 6165 triggers were generated in a region above bin 300 indicating a muon rate of 51.3 muons per minute, approximately one third of the expected rate calculated from the 154 cm2 of scintillator overlap region. Of the 6165 triggers generated there were 2992 TDC events with a measured accidental rate of 1 count per minute.

  • Page 20

    8 KATRIN 1.7 The CENPA contribution to the KATRIN neutrino mass experiment T. H. Burritt, P. J. Doe, G. C. Harper, M. A. Howe, M. J. Leber, A. W. Myers, R. G. H. Robertson, B. VanDevender, T. D. Van Wechel, B. L. Wall and J. F. Wilkerson This has been a very exciting year as construction activity on the KATRIN experiment rapidly moves towards a peak. Of the additional 12 M Euro required to increase the sensitivity of the experiment to 200 meV, 11 M have been secured and contracts have been released for the major components of the experiment. These include the window-less gaseous tritium source, the main spectrometer vessel and the building, which will house the experiment. Technical challenges in designing the gaseous source have resulted in delays but additional engineering resources have been identified and the first tritium runs are expected to begin in Fall 2008. The US participants in the KATRIN collaboration propose to provide two crucial com- ponents of the experiment, the data acquisition system (DAQ) and the focal plane detector. To date the US contingent has supplied the data acquisition system that is being used to commission the pre-spectrometer and an electrode system in the pre-spectrometer that is used to reduce backgrounds. Both will provide important benchmarks for the final DAQ and the electrode system being designed for the main spectrometer vessel. Activity on the US side has also been intense. Joseph Formaggio accepted a tenure track position at MIT and has taken with him the responsibility to provide the detector calibration system and the active veto and shield. This expansion of the US KATRIN contingent is a very positive move but now places an emphasis on management. Keith Rielage accepted a position at the Los Alamos National Laboratory. We are fortunate to welcome a new post- doctoral member to the collaboration, Brent VanDevender. The UW/CENPA will remain the lead institute in the US collaboration, housing both the spokesperson and the project manager. In July 2005 we submitted to the KATRIN collaboration the Detector Design Document which laid out the design solution to meeting the needs of the focal plane detector and DAQ. This design is shown in Fig. 1.7-3. Since then we have been engaged in R&D to further refine the design. This R&D is scheduled to be completed in July 2006 and production of the fabrication drawings will begin. In October 2005 we submitted to the DOE a proposal for funding the US contribution to the KATRIN experiment. This has been circulated for scientific review. At the time the DOE strongly encouraged us to formulate a Project Execution Plan (PEP). This PEP underwent review in January and the very helpful suggestions for improvement are being incorporated in the plan. Central to the success of the US effort is the focal plane detector itself. This consists of a 10-cm diameter PIN diode array of several hundred pixels. A detector of this size, coupled with our requirement of a thin (50 nm) dead layer on the entrance window is very unusual.

  • Page 21

    UW CENPA Annual Report 2005-2006 May 2006 9 As a result, many of the manufacturers who have quoted on providing the array have included relatively costly R&D. Magnet Detector array Pump and calibration ports Electronics Magnetic flux tube Gate valve Veto and shield Support structure Figure 1.7-3. A cross-section showing the Figure 1.7-4. The prototype PIN diode ar- main components of the focal plane detector. ray from WTC showing the hexagonal pixel shapes. We are working with the manufacturers to measure the thickness of the dead layer using our electron gun at UW. One exciting development is the discovery that the Washington Technology Center (WTC) on the UW campus has the capability of manufacturing such detector arrays. We have contracted with WTC to provide three prototype arrays for study, the first of which is shown in Fig. 1.7-4. The identical hexagonal shape of each pixel allows pixel-to-pixel comparison to ensure there is no variation in performance across the face of the detector. The final array will employ a dart board pixel arrangement which better addresses the physics needs. If successful the WTC effort promises considerable cost savings. The focal plane detector system is a complex apparatus that is a critical component in the success of the KATRIN experiment. Providing a commissioned detector and DAQ in time for the beginning of tritium data taking in fall 2008 will be a challenge that will draw heavily on the talents of the CENPA support staff.

  • Page 22

    10 1.8 The design of the KATRIN detector system T. H. Burritt, G. C. Harper, M. J. Leber, R. G. H. Robertson, B. VanDevender, B. L. Wall and J. F. Wilkerson In July 2005, with the release of the Detector Design document, the fundamentals of the design of the focal plan detector system were established. This enabled detailed design of the detector system to proceed. The main components of this design are shown in Fig. 1.8-1. MAGNET COOLING ELECTRODE DETECTOR LN2 PORT VETO, FLUX TUBE SHIELD ELECTRONICS GATE VALVE VACUUM PUMP PORT Figure 1.8-1. Cross-sectional view of the focal plane detector system showing the main components. The detector housing is located inside the 40 cm diameter bore of a 3 to 5 tesla supercon- ducting magnet which focuses the electrons from the beta decay onto the face of the detector. Also housed inside the bore of this magnet is an inert shield of lead and OFHC copper used to reduce background radiation and an active veto of plastic scintillator to register cosmic ray muons. The inert shield and active veto will be designed and manufactured by MIT. In order to gain easy access to the detector housing the magnet is mounted on a rail support structure that allows both the magnet and the veto system to be simply slid aside. In addi- tion, this support structure allows the position of the magnet to be adjusted with respect to the detector in order to center the magnetic flux tube on the face of the detector. Details of the detector vacuum housing are shown in Fig. 1.8-2. The detector housing contains two levels of vacuum. The extreme high vacuum of 10−11 mbar contains the detector and may be separated from the main spectrometer by a 250 mm gate valve. This gate valve allows the detector system to be operated independently of the rest of the experiment if so desired. To reduce electronic noise, the detector, its support structure and preamplifiers operate at ≈ 120K and therefore are surrounded by an insulating, medium vacuum, of ≈ 10−6 mbar. The detector and front-end electronics are mounted at the end of a conical copper electrode attached to a cylindrical ceramic insulator. This ceramic

  • Page 23

    UW CENPA Annual Report 2005-2006 May 2006 11 CERAMIC ISOLATION RING LIQUID NITROGEN INPUT COPPER ELECTRODE DETECTOR ELECTRONICS FLUX TUBE ULTRA HIGH VACUUM MEDIUM VACUUM Figure 1.8-2. Details of the detector vacuum housing and cooling. insulator is cooled by liquid nitrogen. A total of 10 W of heat must be removed from the detector and electronics. The ceramic insulator is attached to the ultra high vacuum cham- ber which contains the pumping ports and port to insert calibration sources and an electron gun which can be swept across the face of the detector to calibrate individual pixels with a variable, mono-energetic electron beam. If low energy radioactive backgrounds prove to be higher than acceptable it is possible to apply post acceleration to the electrons, raising them above the radioactive background. This is achieved by applying up to 30 kV to the copper electrode, detector and electronics. These backgrounds come principally from materials in the immediate surrounds of the detector such as the detector mount, the electrical connections to the individual pixels and the vacuum signal feed-through. To control these backgrounds we are working closely with manufacturers and assaying all materials to be used in construc- tion. Another design challenge that arises due to the use of high voltages in intense magnetic fields is the possibility of enhanced ExB breakdown, even at the extreme levels of vacuum surrounding the detector and electrode. R&D into the questions of radioactive backgrounds, custom components such as ceramic insulators and feedthroughs will continue into July after which it is expected that the design will be frozen and fabrication drawings will be produced.

  • Page 24

    12 1.9 Electron gun for profiling silicon detectors for KATRIN P.J. Doe, G. C. Harper, B. Kuffner∗ and J. A. Mitchell The monoenergetic electron source developed in 20031 to profile electron backscattering with respect to incident angle of the large area silicon detectors to be used in the KATRIN ex- periment has been extensively modified to accommodate large (10 cm diameter) multi-pixel detectors. The apparatus uses an electron gun and an einzel lens that were previously trans- lated via an external UHV X-Y translator. There is also an oil-free pumping system,2 and the associated stand, diagnostics, power supplies, and sample chamber. The sample chamber has been completely redesigned. The gun produces low energy (of order 1 eV) electrons by UV photoemission from a stainless steel surface. A UV grade silica fiber and hypodermic needle collimator direct photons from a mercury arc UV lamp onto a 1 mm diameter spot on the emission surface. The electrons are accelerated through a potential that can be varied up to −30 kV. The electron beam is focused to a 1 mm diameter spot on the device under test (DUT) 0.6 m from the emission surface by an einzel lens operating at about half of the accelerating potential. Electrostatic and beam transport studies of the gun and beam optics were been done using the ion optics program SIMION.3 It is now possible to scan the DUT under the electron beam in both transverse axes with an internal, vacuum-compatible X-Y translator table4 having a range of ±5 cm in each axis. The table may be rotated to any angle up to 60◦ with respect to the longitudinal axis. Cooling and noise-free pumping are provided by a small dewar of liquid nitrogen (LN2 ). The boiled off vapor from the LN2 is used to cool the detector. Two E-type chromel-constantan thermocouples are used to measure the temperatures of the DUT and the detector cooling ring. All of the required parts have been procured and assembled. All of the utilities are accessible from the top 0.5 m diameter flange. The flange can be removed with an existing crane and placed on a work stand. The work stand allows the operator to rotate the flange 180◦ and clamp in place with no assistance to fully expose all of the serviceable parts of the assembly. ∗ Berufsakadamie Karlsrhue University of Cooperative Education, Karlsruhe, Germany. 1 CENPA Annual Report, University of Washington, (2003) p. 65. 2 Varian Vacuum Products, Lexington, MA 02421. 3 SIMION 3D, version 6.0, David A. Dahl, Idaho National Engineering Lab, Idaho Falls, ID 83415. 4 Newmark Systems Inc., Mission Viejo, CA 92691.

  • Page 25

    UW CENPA Annual Report 2005-2006 May 2006 13 1.10 Developing a complete Geant4 simulation of the KATRIN detector region J. A. Formaggio,∗ M. L. Leber, R. G. H. Robertson and J. F. Wilkerson KATRIN is a next generation tritium beta-decay experiment with an expected sensitivity an order-of-magnitude better than previous experiments. To achieve this level of sensitivity, detector-related backgrounds must be limited to 1 mHz. To guide the detector design and understand the detector backgrounds, a complete Geant4 simulation is being developed. Muons, neutrons, and radioactive isotopes have been simulated, including the 238 U chain, the 232 Th chain, the 210 Pb chain, and 40 K. Since the last annual report,1 much progress has been made. More materials and their expected radioactive impurities have been added, including the connections between detector and mount, conflatT M flanges, and insulators for the high-voltage electrode. A plastic scin- tillator records energy depositions and a time stamp, so simultaneous events from cosmic-ray secondaries in the beta detector can be vetoed. Although the simulation geometry is not complete, it can be used to guide the detector design. Copper shielding has shown as much as 80% higher rates from cosmic-ray secondaries than lead, so a pure lead or lead-copper combination will be used for shielding. A thicker detector has lower intrinsic noise, but may have a higher detection efficiency for background gammas. Fig. 1.10-1 shows the rate increase from thorium decays in the connections behind the detector, which can be as high as 15%. A greater effect will be from muon secondaries because only neutral particles, like photons, enter the high magnetic field region. Depending on the post acceleration voltage, a thicker detector may increase rates from muon secondaries as much as 80%. ×10 -6 Rate [mHz/500eV] 20 0.3 mm Det 18 16 0.5 mm Det 14 12 10 8 6 4 0 10 20 30 40 50 60 70 80 90 100 Energy Deposited [keV] Figure 1.10-1. comparison of the background rates from the electrical connections behind the silicon detector for different thickness detectors. Progress is being made toward a complete model for detector-related backgrounds in the KATRIN experiment. ∗ Presently at Massachusetts Institute of Technology, Building 26-568, 77 Massachusetts Ave, Cambridge, MA 02139. 1 CENPA Annual Report, University of Washington (2005) p. 39.

  • Page 26

    14 1.11 Estimating background rates for KATRIN from cosmic rays ionizing the gaseous tritium source J. A. Formaggio,∗ M. L. Leber, R. G. H. Robertson and J. F. Wilkerson KATRIN’s source is a 10 meter long tube filled with high purity tritium gas.1 Electrons ejected from the gas molecules by cosmic rays could create a background to the beta de- cay measurement if they are in the energy region near the endpoint, 18.6 keV. Since these background electrons are created in the source, they have the same detection efficiency as beta decay electrons. Just like the beta electrons, only those created in the guided magnetic flux tube with an opening angle of 51 degrees will reach the detector. Using GEANT4, the background rate from these delta electrons has been calculated. For this simulation, the source geometry is drastically simplified. A stainless steel tube, 1 cm thick walls, is filled with hydrogen gas. Muons,2 high energy protons and neutrons,3 and ambient neutrons4 impinge on the tube and ionize the gas. The delta electrons produced in the simulation with the relevant energy and opening angle are counted and converted to a rate of ionization in the KATRIN source. Rate Hz/500 eV 10-4 10-5 10-6 10-7 10-8 10-9 0 10 20 30 40 50 60 70 80 90 100 Delta Electron Energy keV Figure 1.11-1. Rate of delta ray production in the source will be the background rate in the detector conservatively assuming perfect transmission. The lower limit on the energy window is from the main spectrometer voltage setting, the upper limit from the detector energy resolution. Fig. 1.11-1 shows the spectra of ejected electrons, which falls like E12 , matching theoretical predictions. In the simulation the energy lost by the cosmic rays is consistent with the stopping power of hydrogen gas. The rate of delta ray production in the region of interest, 18.5-25 keV, is 1.33 ± 0.02 µHz. This is well below the design goal of 10 mHz background, and no shielding is necessary for KATRIN’s source. Presently at Massachusetts Institute of Technology, Building 26-568, 77 Massachusetts Ave, Cambridge, ∗ MA 02139. 1 KATRIN Design Report, (2004) http://www-ik.fzk.de/ katrin/publications/index.html. 2 A. J. Da Silva, Development of a low Background Environment for the Cryogenic Dark Matter Search, Ph.D. dissertation, University of British Columbia, Vancouver, Canada. 3 S. Eidelmann et al., Phys. Lett. B 592 1, 24 (2004). 4 M. Yamashita, et al., J. Geophys. Res. 71, 16, 3817 (1966).

  • Page 27

    UW CENPA Annual Report 2005-2006 May 2006 15 1.12 KATRIN preamplifier design and development H. Gemmeke,∗ R. G. H. Robertson and T. D. Van Wechel Two different charge sensitive preamplifier designs are currently under development by the CENPA electronics shop. One of the designs will be selected for use on the multi pixel Si PIN diode detector array at the focal plane of the KATRIN experiment. Since the preamplifiers will be cryogenically cooled, the goal is to minimize power dissipation. Another criterion is that the preamplifier design have low noise, with a goal of 600 eV FWHM. The first preamplifier is of a conventional design with a folded JFET cascode amplifier as the first stage. The second preamplifier is of a more unconventional design, a parametric amplifier. The preamplifier design consisting of a folded cascode JFET input stage followed by a low power op amp second stage is nearly complete. For the cascode input stage a compromise has to be made between the output noise level and power dissipation. In general increasing the JFET current (Id) decreases the noise level. The noise level also decreases as the JFET’s drain to source voltage (Vds) is increased. This effect is much weaker than that of the noise to current dependence. In the case of the folded cascode, another noise contributor is the current noise due to resistor Rd between the junction of the drain of the JFET and the emitter of the folded cascode PNP transistor, and the positive power supply. This resistor which sets the JFET current produces a noise current that is inversely proportional to the square root of resistor value. This implies that for a given power supply voltage and Vds, that if the value of this resistor is reduced to increase Id, to lower the FET’s noise contribution, part of this reduced noise contribution is lost due to the increased noise contribution of Rd. Since the noise level is only loosely dependent upon Vds, it has been shown by circuit simulations that the best compromise if one desires to keep the supply voltage low for minimum power dissipation, is to set Vds to a low value such as between one and two volts so that for a given supply voltage, the value of Rd may be maximized. The lower Vds reduces the amplifier bandwidth somewhat, but since it is still over an order of magnitude greater than the Gaussian shaper’s bandwidth, it is acceptable. The proposed design has an overall preamplifier power dissipation of 84 to 90 mW, of which 72 mW is in the cascode stage and 12 to 18 mW is in the second stage op amp. The supply voltages are plus and minus 6 volts, with Vds set to 1.7 volts and Id at 10 mA. Pspice simulations with a detector capacitance of 15 pF predict that the noise level at the output of a 4-µS Gaussian shaper following the preamplifier is 564.8 eV FWHM at 300K and 453.1 eV FWHM, if the preamplifier is cooled to 200K. If Id is reduced to 5 mA, the total power dissipation is reduced by 30 mW, but the noise level increases to 605.6 eV FWHM at 300K and 490.5 eV FWHM at 200K. This folded cascode design also has potential use for the Majorana experiment and with modifications for Nanopore DNA sequencing. (See section 6.3.) The main motivation for the parametric amplifier is that the first stage has much lower power dissipation than a conventional amplifier design. The disadvantage is in circuit com- ∗ Forschungszentrum Karlsruhe, Institut fur Prozessdatenverabeitung und Elektronik, POB 3640, 76021 Karlsruhe, Germany.

  • Page 28

    16 plexity. Parametric amplifiers are devices that provide amplification through the variation of a circuit parameter, in our case capacitance. The input stage of the parametric preamplifier has a varactor bridge consisting of a pair of reverse biased varicap diodes pumped in anti phase by a local oscillator(LO) signal. The junction of the varicap diodes is both the input and output of the first stage. The DC and low frequency input voltage appearing at the junction determines the degree of unbalance of the bridge with the pump voltage at the LO frequency developed at the diode junction being proportional to the degree of bridge unbal- ance. The detector is AC coupled through a coupling capacitor in series with an inductor acting as a low pass filter to the junction of the varicap diodes. Charge from the detector due to ionization events redistributes to the varicap diode junction producing a change in voltage at the diode junction modulating the pump voltage. The detector signal which is at a very high impedance, modulates the pump signal which is at a low impedance resulting in a large power gain. The pump voltage output is capacitively coupled to a second stage RF amplifier. The output of the second stage is transported outside of the cryostat to the room temperature environment where it is synchronously demodulated by a phase sensitive detector. A prototype with a LO frequency of 200 MHz was constructed late last year as a proof of concept. It demonstrated that the concept works but had poor noise performance due to the poor impedance matching between the first and second stage. Also the power dissipation was quite high due to the use of a commercial broadband RF amplifier as a second stage. The new design currently under development will have a tuned MOSFET amplifier dissipating 6 mW for the second stage. Also using varicaps with a higher capacitance versus voltage slope and increasing the LO amplitude will help to improve the signal to noise ratio. The new design will use an active mixer for the phase sensitive detector, rather than the diode based Double Balance Mixer used by the prototype. A means of adjusting the relative phase shift between the LO and RF inputs of the mixer needs to be developed to compensate for circuit phase shifts and propagation delay.

  • Page 29

    UW CENPA Annual Report 2005-2006 May 2006 17 Majorana 1.13 The Majorana neutrinoless double-beta decay experiment J. F. Amsbaugh, J. A. Detwiler, P. J. Doe, A. Garcı́a, M. A Howe, R. A. Johnson, K. Kazkaz, M. G. Marino, S. R. McGee, R. G. H. Robertson, A. G. Schubert, B. A. VanDevender and J. F. Wilkerson Neutrinos continue to provide some of the most exciting opportunities in understanding our universe. Neutrinoless double-beta decay (0νββ) provides the physics community with the opportunity to build on our successes in understanding the neutrino and crafting a new standard model. Determining if neutrinos are Dirac or Majorana particles is one of the most important questions facing the physics community today. The Majorana experiment aims to answer this question. For the first time we can mount experiments that probe the neutrino mass region below the upper limits set by direct kinematical searches (tritium) and suggested by observational cosmology, while planning scaled approaches that can address the lower bounds of mass defined by the atmospheric and solar + reactor neutrino oscillation experiments. Our proposed method uses the well-established technique of searching for 0νββ in high- purity Ge diode radiation detectors that play dual roles of source and detector. The technique is augmented with recent improvements in signal processing, detector design, and advances in controlling intrinsic and external backgrounds. Progress in signal processing from segmented Ge-diode detectors offers significant benefits in rejecting backgrounds, reducing sensitivity of the experiment to backgrounds, and providing additional handles on both signals and backgrounds through multi-dimensional event reconstruction. Development of sophisticated Cu electroforming methods allow the fabrication of ultra-low-background materials required for the construction of next generation detectors. The envisioned Majorana experiment will consist of one or more modules containing 57 high-resolution intrinsic germanium detectors, each with mass ∼1.1 kg of Ge enriched to 86% in 76 Ge. The crystals will be deployed in an ultra-low-background electroformed Cu cryostat, located deep underground within a low-background shielding environment. Observation of a sharp peak at the ββ endpoint would quantify the 0νββ decay rate, demonstrate that neutrinos are Majorana particles, indicate that Lepton number is not conserved, and provide a measure of the effective Majorana mass of the electron neutrino. The physics goals for the first phase of Majorana are to: • Probe the quasi-degenerate neutrino mass region above 100 meV. • Demonstrate that backgrounds, at or below 1 count/ton/year in the 0νββ - decay peak 4-keV region of interest, can be achieved that would justify scaling up to a 1 ton or larger mass detector. • Definitively test the Klapdor-Kleingrothaus claim to have observed 0νββ - decay in 76 Ge in the mass region around 400 meV.

  • Page 30

    18 These goals are consistent with recent recommendations from the DNP/DPF/DAP/DPB Joint Study on the Future of Neutrino Physics and the conclusions on 0νββ reported by the Neutrino Scientific Assessment Group (NuSAG). The Majorana Scientific Collaboration consists of about 100 scientists and 16 collabo- rating institutions from four countries with extensive experience in double-beta decay and ultra-low-background experiments. CENPA collaborators are currently involved in a number of research and development related projects as described in the following sections. 1.14 DAQ and SOH for Majorana J. A. Detwiler, M. Howe, A. Myers, T. Van Wechel and J. F. Wilkerson We have developed an initial conceptual plan for the Majorana data acquisition (DAQ) and State of Health monitoring (SOH). The DAQ systems include the digitization electron- ics, software management of the readout, and data storage. The SOH system encompasses hardware and software safety monitors and interlocks for detector and support systems, as well as electronics health and some data quality monitoring. We drafted a list of tasks relevant to the implementation of the DAQ and SOH, and incorporated this list into the Majorana work-breakdown structure. We estimated the funds and labor necessary to complete each task based on price estimates for materials and services, and based on our experiences in previous experiments such as SNO. We identified several places where different options may allow for considerable savings or risk management. For example, the digitization electronics may be built in-house, or commercial systems may be purchased. Various points of interface with other Majorana systems were also identified and discussed among collaborators. The DAQ and SOH plans and costing were incorporated into the Majorana reference design and the draft proposal was submitted to the DOE. We are in the process of refining that reference design to optimize cost and schedule versus physics impact of the experiment, and updating the proposal to reflect those refinements.

  • Page 31

    UW CENPA Annual Report 2005-2006 May 2006 19 76 1.15 The Majorana sensitivity to excited-state double-beta decays of Ge S. R. Elliott,∗ V. M. Gehman,∗ K. Kazkaz, D.-M. Mei∗ and J. F. Wilkerson The Majorana experiment1 will be searching for neutrinoless double-beta (0νββ) decays of 76 Ge. The anticipated background levels in the Majorana detector are sufficiently low to allow for a search for two neutrino double-beta decays to an excited state of the daughter nucleus. Measuring the half life of this excited state two-neutrino double-beta (ES2νββ) decay will provide additional direction in the development of the theory behind calculating nuclear matrix elements.2 Ge is expected to decay to the first excited 0+ state (also referred to as the “0+ 76 1 ” state) of 76 Se, with two accompanying cascade gammas with energies 563 and 559 keV. 76 Ge may also decay to the 2+1 state, with only one accompanying 559 keV gamma. While the phase- space factor for this latter decay is roughly an order of magnitude larger than that of the former, the matrix elements are relatively suppressed by roughly two orders of magnitude.2 Thus decays to the 0+ 1 state are predicted to be more prevalent. Furthermore, from an experimental standpoint, it is reasonable to expect a search for a coincidence involving two cascade gammas rather than just one will offer greater background reduction capabilities. For these reasons, the Majorana experiment is focusing on decays to the 0+ 1 state. The Majorana detector will consist of modules each containing an array of germanium crystals enriched to 86% 76 Ge. The array contains 3 layers of 19 crystals arranged in a hexagon, with four unique crystal locations per layer (see Fig. 1.15-1). We have performed preliminary calculations of the efficiency of this module to the ES2νββ signal using MaGe, a GEANT4- and ROOT-based simulation framework collaboratively developed by the Ma- jorana and GERDA3 collaborations. Requiring a very simple 3-crystal cut, with two of the crystal energies being within 2 keV of 559 and 563 keV, the efficiency is calculated to be 0.983(1)%. The crystals may be segmented, though, to reduce the backgrounds to the 0νββ signal. One possible segmentation scheme involves 2 radial (i.e., “pie wedge”) by 3 axial (i.e., “hockey puck”) segments. Using a very simple 3-segment analysis that parallels the 3- crystal cut above, the efficiency is increased an additional 0.635(3)% to 1.618(3)% total. The efficiency of the module for this decay may be further increased with alternative analytical methods involving only energy cuts and relaxing the 3-crystal or 3-segment requirements. Such additional methods are currently under study. The segmentation of the crystals may be higher, for instance up to 6 × 6 segments. In this case, the additional, simple 3-segment efficiency is 0.466(2)%, or 1.449(2)% total–a bit lower than in the 2 × 3 case. The reason for this is that with 6 × 6 segmentation, the individual segments have smaller volume, and while the gamma rays are therefore more likely to escape the segment in which the decay occurred, they are disproportionately less likely to deposit their full energy in a single segment. Relaxing the 3-segment requirement, however, may increase the efficiency of the 6 × 6 segmentation over that of the 2 × 3 scheme, as the driving ∗ P-23, MS H-803, Los Alamos National Laboratory, Los Alamos, NM 87545. 1 http://majorana.pnl.gov. 2 J. Suhonen and O. Civitarese, Phys. Rep. 300, 123 (1998). 3 http://www.mpi-hd.mpg.de/ge76.

  • Page 32

    20 Figure 1.15-1. Simulation rendering of the Majorana detector module. There are four unique crystal locations per layer (typical crystal used in simulations are 0, 1, 7, and 8 in each of the three layers). The shielding, cryostat, and most of the support structure are omitted for clarity. The vertical bars are copper tubes comprising part of the support structure. consideration then is only whether or not the gamma rays escape the segment in which the decay occurred, and the dependence on the photopeak efficiency of the gammas is reduced. ES2νββ decays are expected to have a half life 10-100 times as long as the half life of ground state 2νββ decays, 1.9 × 1021 years.4 Assuming a half life of 1023 years, then, for the excited state decay, a single Majorana module will observe 25 such events in 8 months of live time. This live time requirement may be reduced when analysis methods to increase the signal efficiency are developed. The backgrounds to the ES2νββ decay come primarily from uranium and thorium in the germanium crystals, support structure, cryostat, and shielding, as well as 60 Co and 68 Ge in the crystals. At the anticipated levels of radioisotope contamination, the number of background events expected to survive the simple 3-crystal and 3-segment analysis cuts is only 0.002 counts over 8 months of live time. Thus this signal is essentially “backgroundless”. If more analysis cuts are developed to increase the sensitivity to the excited state signal, the backgrounds will have to be re-analyzed as well to calculate their level of contribution. 4 H. V. Klapdor-Kleingrothaus et al., Eur. Phys. J. A 12, 147 (2001).

  • Page 33

    UW CENPA Annual Report 2005-2006 May 2006 21 1.16 Development of simulation and analysis frameworks for Majorana J. A. Detwiler, R. A. Johnson, K. Kazkaz, M. G. Marino and A. G. Schubert Monte Carlo (MC) radiation transport simulation models have been developed for Majorana using MaGe, an object-oriented MC simulation package based on the Geant4 toolkit and op- timized for simulations of low-background germanium detector arrays. MaGe is being jointly developed by the Majorana and GERDA1 collaborations, using professional programming techniques in consultation with collaborators from the National Energy Research Scientific Computing Center (NERSC) at LBNL. MaGe defines a set of physics processes, materials, constants, event generators, etc. that are common to these experiments, and provides a uni- fied framework for geometrical definitions, database access, user interfaces, and simulation output schemes in an effort to reduce repetition of labor and increase code scrutiny. Our group has been instrumental in coding much of the backbone of MaGe, including the framework for handling general detector geometries, the list of simulated physics processes, and various event generators. We have also been involved in coding quality control, and have implemented several specific detector geometries within MaGe. We are in the process of validating the various physics processes simulated by MaGe in order to extend its applicability an increase confidence in MaGe results. These studies are discussed in more detail elsewhere in this Annual Report. We are also in the process of adding the capability to simulate pulse shapes generated by energy deposits in solid-state germanium detectors. In order to organize the analysis of simulated and test data, and to prepare for analysis of physics data, we have proposed and begun to implement a software analysis framework for Majorana. The C++ framework is based on the ROOT data analysis toolkit, and is archi- tectured with a “modular processing design”, which standardizes computations performed at the event-loop level. In modular processing, the analysis is divided into several modules, each of which performs specific tasks at three different stages of data processing: at the beginning and end of the full analysis, at the beginning and end of each run processed, and at the event-by-event level. The initial implementation of this framework has been used to analyze background simulations for the Majorana reference design, and to study the background re- jection efficacy of various detector options. We are in the process of optimizing the design of the framework and adding pulse shape analysis capabilities. 1 The GERDA Collaboration, Nucl. Phys. B 143, p. 567 (2005).

  • Page 34

    22 1.17 MaGe simulation of LANL clover detector with AmBe neutron source A. G. Schubert The Majorana and Gerda collaborations have jointly developed MaGe, a Geant4 and ROOT based simulation framework, to aid in the design and analysis phases of both experiments. Comparisons between experimental data and MaGe simulation results can be used to verify MaGe’s performance. This report describes a MaGe simulation of an experiment performed at LANL. The experiment used a detector consisting of four Ge crystals arranged in a clover con- figuration. The detector was surrounded by lead shielding, and an Americium Beryllium (AmBe) neutron and gamma source was located outside of the shield. A slab of polyethy- lene neutron moderator was placed within the lead shield, between the source and detector. Energy spectra were collected from the Ge crystals for various moderator thicknesses. The experimental setup was modeled within MaGe. A comparison of experimental data and simulation results appears in Fig. 1.17-1. Overall agreement between the data sets helps to validate the MaGe package and verify calculations of experimental run time. Discrepancies between the locations of some peaks are currently being investigated. AmBe Clover cts/s/keV 1 10-1 10-2 10-3 0 500 1000 1500 2000 2500 3000 Energy [keV] Figure 1.17-1. Comparison of experimental (black) and simulated (grey) energy spectra from a single crystal, for six-inch thickness of polyethylene moderator.

  • Page 35

    UW CENPA Annual Report 2005-2006 May 2006 23 1.18 Simulation of the SLAC electron beam dump experiment using MaGe M. G. Marino Understanding the background created by neutrons involves the simulation of their creation, propagation, and interaction with the detector as well as the verification of all associated code. This paper presents work completed concerning the verification of neutron production and transport within the MaGe simulation package. An experiment has been done at the Stanford Linear Accelerator Center (SLAC) which involved a high energy (28.7 GeV) electron beam incident upon an aluminum beam dump.1 Neutrons were created mainly by photonuclear interactions initiated by bremsstrahlung pho- tons from the decelerating electrons, and the neutron time-of-flight and energy spectra were measured outside a steel shield and a concrete shield of variable width. A detailed description of the setup of the SLAC beam dump experiment can be found in the reference cited. A 28.7 GeV electron beam was incident upon a cylindrical aluminum beam dump inside a shield housing with an opening to allow the beam to enter. A detector was placed laterally through the shields from the aluminum beam dump and the neutron energy and time of flight (TOF) spectra were measured for shield concrete shield widths of 9 ft, 11 ft, and 13 ft. Some results are shown in Fig. 1.18-1 (FLUKA and experimental data obtained from S. Roeslera et al.2 ). 10-12 dΦ/dE (neutron cm-2 MeV-1 electron-1) 10-13 MaGe FLUKA 10-14 Exp 10-15 10-16 10-17 10-18 1 1.5 2 2.5 3 1 1.5 2 2.5 3 1 1.5 2 2.5 3 Log(E) (MeV) Log(E) (MeV) Log(E) (MeV) (a) Energy spectra. 10-2 dC/dtTOF (Counts s-1 electron-1) MaGe -3 10 FLUKA 10-4 10-5 1 1.2 1.4 1.6 1.8 2 2.2 1.2 1.4 1.6 1.8 2 2.2 1.2 1.4 1.6 1.8 2 2.2 Log(t ) (ns) Log(t ) (ns) Log(t ) (ns) TOF TOF TOF (b) TOF spectra. Figure 1.18-1. A comparison of calculated and experimental energy and TOF spectra for three shield widths (L to R: 9 ft, 11 ft, 13 ft). 1 S. Taniguchi et al., Nucl. Instrum. Methods A, 503, 595 (2003). 2 S. Roeslera et al., Nucl. Instrum. Methods A, 503, 606 (2003).

  • Page 36

    24 1.19 Studies of α-contamination on high purity Ge crystals J. A. Detwiler, R. A. Johnson, M. G. Marino and J. F. Wilkerson The Majorana experiment will search for the neutrinoless double beta decay (0νββ) of 76 Ge → 76 Se using 60 kg modules of high purity germanium detectors enriched to 86% 76 Ge. The current lower limit on the half-life of this reaction in 76 Ge is 1.9 × 1025 years. A full under- standing of our backgrounds is a necessary condition if we are to be successful in discerning such a small signal. We report on progress made on understanding surface contamination issues using MaGe, a Geant4 based Monte-Carlo simulation program for the Majorana and Gerda collaborations. Each 60 kg module in Majorana consists of 57 close-packed 1.1 kg germanium crystals surrounded by a copper cryostat and various “small parts” such as copper support rods, teflon trays, and wiring. We came up with an algorithm to randomly sample any general surface in our simulation, as Geant4 currently has no satisfactory method for sampling the surfaces of complicated volumes. This algorithm was implemented into the MaGe package, and the results have been quite satisfactory. Fig. 1.19-1a shows randomly generated points that are uniformly distributed on the surfaces of a single crystal, support tray, and contact ring in a 60 kg module for Majorana. Surface contaminants that may deposit energy into our region of interest(2039 ± 2 keV) include the 238 U and 232 Th radioactive decay chains. The decay daughters of Rn, 210 Pb especially, can be particularly problematic. 222 Rn is an airborne gas which quickly decays to 210 Pb. The chain 210 Pb→210 Bi→210 Po→206Pb deposits a 5.3 MeV α particle from the 210 Po decay. If 210 Pb plates out on a surface, the long half-life (26 years) will produce a steady stream of 5.3 MeV α particles for a long time. The simulated spectrum in Fig. 1.19-1b is from 210 Pb plated out on all the inner surfaces of a 60 kg module. -9 ×10 210 Pb Surface Contamination Counts/keV/cm2/Initial Decay 0.22 0.2 0.18 0.16 0.14 0.12 0.1 0.08 0.06 0.04 0.02 0 2000 2020 2040 2060 2080 2100 Energy [keV] (a) (b) Figure 1.19-1. (a) Surface points randomly generated on the surface of a crystal, contact rings, and support tray. (b) 210 Pb contamination spectrum around the region of interest.

  • Page 37

    UW CENPA Annual Report 2005-2006 May 2006 25 1.20 LArGe, liquid argon compton supressed germanium crystal C. E. Aalseth,∗ J. F. Amsbaugh, P. J. Doe, M. H. Toups and J. F. Wilkerson We have continued using the LArGe apparatus1 to further understand background suppres- sion in a high purity germanium crystal diode by identifying the Compton scattered gamma ray. This test apparatus consists of a 55 mm diameter by 53 mm P-type crystal suspended in a bath of liquid argon (LAr) with a minimum of 7 cm of LAr surrounding it. The LAr acts as an active shield, emitting scintillation light peaked at 128 nm. The scintillation light is wavelength shifted by a non-metallic visible light reflector, ESR2 , in order that it can be detected by a hemispheric 20 cm photomultiplier tube(PMT) in the LAr. The PMT signal is then used as an anti-coincident veto on the HPGe signal to eliminate events that deposit energy in both the crystal and LAr. The initial test run results were reported last year with background suppression consistent with 0.5 radiation lengths of LAr surrounding the HPGe. During the subsequent run, the HPGe crystal exhibited high leakage current while biased. This is probably due to contamination of the crystal resulting from a leak to atmosphere in the crystal storage container. After several vacuum bake-outs the crystal performance was still unacceptable and it is now being refurbished by the manufacturer. A crystal replacement was not available for use. While studying the PMT response without a crystal present, it was noticed that the re- flector material acts both as a wavelength shifter and as a visible light guide. A quantitative estimate of the efficiency of these two processes could not be made using the existing appa- ratus. Modifications were therefore made to allow a summer student (M.T.) to investigate samples of light guide material coated with a wavelength shifter and examine the light guiding effect. A coating technique with polystyrene doped to about 10% TPB, tetraphenyl butadi- ene, and dissolved in toluene was used on samples of ESR and acrylic. PMT only spectra were collected with background only and an external 207 Bi source. Unfortunately, the results are inconclusive due to difficulties with the readout electronics and the investigation will be repeated. Much of our effort this past year has been directed towards designing an apparatus, which will improve the efficiency of the Compton shield. A larger dewar, 87 cm ID by 137 cm tall has been acquired and will provide three radiation lengths of LAr shield around the crystal. A pressurized liquid nitrogen reservoir maintains the LAr in a liquid state while providing long periods of stable operation between filling the reservoir. To reduce sources of background, the HPGe crystal will be suspended in the argon using low background materials. The front- end components of the preamplifiers will be mounted close to the HPGe to optimize the noise performance. Four PMTs in the LAr will collect the light that has been wavelength shifted and reflected by ESR material, similar to our current apparatus. The crystal handling components are finished and the cryogenic components construction will begin soon. ∗ Pacific Northwest National Laboratory, Richland, WA 98352. 1 CENPA Annual Report, University of Washington (2005) p. 46. 2 also known as VM2000, from 3M Corp.

  • Page 38

    26 1.21 Thermal modeling of the MEGA cryostat cooling J. F. Amsbaugh and D. Nikic∗ The Majorana double beta decay1 experiment has a multiple HPGe crystal prototyping and demonstration project known as Multiple-Element Gamma Assay or MEGA.2 MEGA makes use of a very high purity electroformed copper cryostat to reduce radioactive background. The design can accommodate up to 8 copper inner cans(IC), each containing a pair of HPGe crystals, on a cooling plate (CP) ring inside an annular vacuum insulated cryostat. The CP is cooled by a long coaxial cold finger exiting the side of the annulus. A LN2 tank removes heat from the far end of the cold finger. To minimize background contributions from materials near the crystals, no black body (BB) radiation shields between the cold parts and the warm cryostat walls are used. The copper IC surrounds and cools the crystals. It also absorbs BB radiation from the cryostat wall. Sufficient heat conductance to the LN2 is required to keep the temperature below the HPGe’s maximum operating temperature. We studied the cryostat thermal performance using ANSYS finite element analysis soft- ware3 by first calculating the maximum steady state temperature rise of the ICs as they conduct the absorbed BB radiation to the CP held at a 100K boundary condition (BC). This is shown in column ∆TIC in Table 1.21-1. Next we calculated the maximum temperature rise in the CP and cold finger assembly as it conducts the IC heat loads into the LN2 (80K) with BB radiation and thermal contact conductivities accounted for. This is shown in column ∆TCP in Table 1.21-1. The maximum temperature of the IC’s is obtained by adding the two delta T’s to the liquid nitrogen boiling point (80 K) to get the temperature given in column Temp. in Table 1.21-1. Solutions were obtained for copper emissivities, ǫ = 0.01, 0.05, 0.1, 0.2. Finally, transient cool-down calculations of the previous model starting at 300K and modeled 96 or 192 hours were done. These results (not reported) are for an empty model since HPGe crystal heat capacity was not included but are consistent with the steady state calculations, verifying good convergence. ǫ ∆TIC ∆TCP Temp. The material properties and contact conductivities 0.01 0.48 K 5K 86 K (IC = 7000W/m2K and CF = 1500W/m2 K) used were 0.05 2.4 K 26 K 109 K typical for smooth OFHC copper. The ǫ of polished 0.1 4.8 K 52 K 137 K copper4 ranges from 0.04 to 0.01and lightly oxidized is 0.2 9.6 K 105 K 195 K about 0.2. Table 1.21-1. Steady State ANSYS results. Further study and comparison with cool down data is needed, but the temperature results shows that copper surface ǫ and the contact conductivities need to be carefully considered and optimized. We have also accomplished our goal of gaining experience using ANSYS. Presently at Boeing Company, Seattle, WA. ∗ 1 Information available online at http://majorana.pnl.gov 2 K. Kazkaz et al., IEEE Trans. Nucl. Sci. 51, 1029 (2004). 3 ANSYS Inc., Canonsburg, PA 15317. 4 Y.S. Touloukian and D.P. DeWitt, Thermophysical Properties of Matter, Vol.7 Thermal Radiative Prop- erties, p.136, IFI/Plenum, New York, 1970-

  • Page 39

    UW CENPA Annual Report 2005-2006 May 2006 27 1.22 Monte Carlo simulation studies of the LoMo counting facility J. A. Detwiler, J. A. Formaggio,∗ R. A. Johnson, M. G. Marino, S. R. McGee, A. G. Schubert and J. F. Wilkerson Researchers at Pacific Northwest National Laboratory (PNNL) and the University of Wash- ington (UW) have been collaborating on the design of an ultra-low-background counting facility to be constructed at Lower Monumental Dam (LoMo) on the Snake River in Eastern Washington State. Monte Carlo (MC) radiation transport simulation models were developed for a LoMo prototype screening detector and for the existing 329/17A Ultra-Low-Background screening detector (17A) using MaGe,1 an object-oriented MC simulation package based on the Geant4 toolkit and optimized for simulations of low-background germanium detector arrays. MaGe is being jointly developed by the Majorana and GERDA2 collaborations. System performance studies were performed with the PNNL 17A detector, since the LoMo cryostat and crystal design is still at the conceptual stage. These initial studies were based on several short calibration runs taken with the 17A detector using 60 Co and 22 Na sources. The agreement between MC and data was very nice over almost the entire energy scale. Efficiencies and sensitivities for standard sample geometries were calculated for the 17A detector and compared with data where available. MaGe’s arbitrary sample geometry framework allows for such calculations to be performed for any given sample. Copper purity requirements for the LoMo depth were assessed with an analytical calcula- tion based on the method outlined in Heisinger et al.3 for estimating the spallation production of radioactive isotopes in underground experiments. The total activity of all radioactive spal- lation products at LoMo depth was found to be 290.4 µBq/kg, where most of the activity is from 57 Co, 58 Co, and 60 Co. The equivalent activity for intrinsic 238 U and 232 Th in the copper was found to be 23.4 and 71.5 µBq/kg, respectively. Efforts to verify cosmogenic activation physics in MaGe are ongoing so that these calculations can be backed-up with full end-to-end MC simulations. For the LoMo geometry, 238 U and 232 Th decay chains were simulated in the copper can and coldplate surrounding the crystal, and in the passive lead shielding. These simulations indicated that backgrounds from lead impurities are about an order-of-magnitude lower than those from the copper at LoMo depths, and hence are not large enough to warrant an inner layer of ancient lead or high-purity copper. The ambient γ radiation at the LoMo site was measured by PNNL researchers and simulated with MaGe. This work indicated that no more than 10 inches of lead is needed to reduce the external γ background below the levels expected from cosmogenic activation of copper. These studies implied that significant cost savings could be obtained with a simplified passive shielding design without impacting the physics. However, further studies on the shielding of direct cosmogenic backgrounds, particularly backgrounds from spallation neutrons, will need to be performed when the µ- nuclear and hadronic physics in MaGe have been verified. ∗ Presently at Massachusetts Institute of Technology, 77 Massachusetts Ave, Cambridge, MA 02139. 1 The Majorana Collaboration, Nucl. Phys. B 143, p. 544 (2005). 2 The GERDA Collaboration, Nucl. Phys. B 143, p. 567 (2005). 3 B. Heisinger et al., Earth Planet. Sci. Lett. 200, p. 345 (2002).

  • Page 40

    28 2 Fundamental Symmetries and Weak Interactions Torsion Balance Experiments 2.1 Torsion balance search for spin coupled forces E. G. Adelberger, C. E. Cramer and B. R. Heckel We have completed a year-long search for new spin-coupled forces using our second-generation spin pendulum.1 The spin pendulum, a net spin dipole containing approximately 6 × 1022 polarized electron spins, rotates about its fiber axis at a constant angular velocity. Conse- quently, its affinity for a preferred direction in space (or lack thereof) appears as a modulation of its angular position at the rotation frequency. The amplitude of this modulation is itself modulated both daily and annually as the orientation of the fiber axis changes with the earth’s rotation and orbit around the sun. Analyzing data taken over thirteen months for signals modulated at solar or sidereal periods, we constrain the energy required to flip a spin in a hypothetical spin-coupled field to be less than 10−21 eV. Throughout the year, we performed systematic tests to measure the sensitivity of our apparatus to disturbances caused by tilt, temperature, magnetic fields, and local gravitational gradients. No significant corrections to the data had to be applied. The constraint on spin-coupled forces can be interpreted as a limit on possible simultane- ous Lorentz and CPT violation as described by Colladay and Kostelecký2 as well as coupling to an astronomical source via pseudoscalar bosons3 or massless bosons constrained only by rotational and translational invariance.4 In the Kostelecký framework, we provide an upper limit on the CPT-violating parameter |b̃e | ≤ 6.7 × 10−21 eV that should be compared to the benchmark value m2e /MP = 2 × 10−17 eV, where me and MP are the electron and Planck masses, respectively. Fig. 2.1-1 shows our limits on pseudoscalar boson coupling compared to the previous results reported by Youdin et al.,5 Ni et al.6 and Wineland et al.7 1 CENPA Annual Report, University of Washington (2004) p. 7. 2 D. Colladay and V. A. Kostelecký, Phys. Rev. D 55, 6760 (1997). 3 J. E. Moody and F. Wilczek, Phys. Rev. D 30, 130 (1984). 4 B. A. Dobrescu and I. Mocioiu, private communication. 5 A. N. Youdin et al., Phys. Rev. Lett. 77, 2170 (1996). 6 W. T. Ni et al., Phys. Rev. Lett. 82, 2439 (1999). 7 D. J. Wineland et al., Phys. Rev. Lett. 67, 1735 (1991).

  • Page 41

    UW CENPA Annual Report 2005-2006 May 2006 29 Figure 2.1-1. 2σ upper limits on |gP e gS |/(h̄c) as a function of interaction range λ for the e gP gS −r/λ potential V (r) = − 8πme c σe · ∇ r . Our results, previous work by Youdin et al., Ni et al. e and Wineland et al. are indicated by solid, dashed, dash-dotted and long sloping dashed lines, respectively. We do not give constraints for 10 km < λ < 103 km because integration over the terrestrial surrounding is not reliable in this regime.

  • Page 42

    30 2.2 Testing the weak equivalence principle using a rotating torsion balance E. G. Adelberger, K-Y Choi, J. H. Gundlach, B. R. Heckel, S. Schlamminger, H. E. Swanson and T. A. Wagner We are using a torsion balance rotating with a constant angular velocity about its fiber axis to test the weak equivalence principle. Four titanium and four beryllium test bodies are mounted on the pendulum body in a dipole configuration. A violation of the equivalence principle would yield a differential acceleration of the two materials towards a source mass. Differential accelerations are detected as modulations of the angular position of the pendulum at the rotation frequency. This acceleration can be analyzed for a variety of source masses covering a range from one meter to infinity. An equivalence-principle violating acceleration can be expressed as a Yukawa type interac- tion of a new composition dependent interaction, with a strength relative to the gravitational interaction (α) and an interaction range (λ). Of special interest for long range interaction are accelerations towards the sun and towards the center of the galaxy. The former would modulate our lab fixed acceleration with a daily period, the latter with a sidereal period. In the past year we took two data sets, lasting 78 and 52 days. Before and after each data set we measured the gravity gradient fields from a nearby hill. In addition, we determined the gravity multipole moment of the pendulum and we performed several systematic tests to measure the sensitivity of our apparatus to disturbances caused by tilt, temperature and magnetic fields. It turned out that our first data set had a large susceptibility to a temperature gradient across the torsion balance. This problem was reduced significantly for the second data set, which was taken with a different fiber. Using the measurements of the second data set we improved our previously published limit1 on the equivalence-principle violation by a factor of five for local source masses. Fig. 2.2- 1 shows the 2σ upper limit on new Yukawa type interactions coupling to the baryon num- ber. The limiting systematic effect was a tilt induced signal. Furthermore we tested the equivalence principle for astronomical sources, such as the Sun and the center of our galaxy, with a 12 times greater sensitivity than previous torsion balance experiments. We have measured a rel- Figure 2.2-1. New 2σ limits on a Yukawa type force coupling to baryon number. The ative differential acceleration towards the sun lines labeled “New” give the result of last of ∆a/a = (1.1±1.6)×10−13 (1σ uncertainty). year’s data. The white region is now excluded from the parameter space. 1 Y. Su et al., Phys. Rev. D 50, 3614 (1994).

  • Page 43

    UW CENPA Annual Report 2005-2006 May 2006 31 2.3 Progress toward improved torsion fibers J. H. Gundlach, C. A. Hagedorn and S. Schlamminger We report here on continued efforts to fabricate improved torsion fibers for our LISA and small-force torsion balance experiments. As our experiments improve, we have found that one major obstacle to achieving greater sensitivity is the thermally generated torque noise from the torsion fibers themselves. The magnitude of the torque noise is proportional to the square root of the ratio of a fiber’s torsion constant to its mechanical quality factor Q. We have made many efforts toward decreasing this ratio, primarily by increasing Q. Furthermore, for our experiments, it is essential that our pendulums remain electrically neutral. Traditionally this has been achieved by using grounded conducting metallic wires as torsion fibers. In a quest to find higher quality factors, we have made fibers from quartz glass. The mechanical quality factor of quartz/fused silica is among the largest known. Quartz is a very good electrical insulator, making the problem of electrical grounding no longer trivial. We have refurbished a thermal evaporator that has a very large bell jar to permit the metalization of entire quartz fibers as they hang vertically. In addition to allowing us to metalize fibers, the evaporator is an excellent multi-purpose vacuum chamber. We have installed, and continue to refine, a simple opto-electronic system that allows us to assess the Q of our fibers before and after coating. The evaporator’s powerful turbo pump allows us to work at pressures markedly below 10−5 Pa, enabling us to work with residual gas pressures lower than those found in our small force experiments. These features, along with the user-friendliness of the bell jar, make it a useful test bed for new ideas. With these systems in place, we are in a position to fabricate new fibers and quantify their properties. We have been able to make conducting fibers whose properties are slightly superior to those of our current tungsten fibers. Improvement is likely, and its extent is a topic of current research. If we are able to find a way to use unmetalized fibers, we should see at least a factor of ten improvement in our torque noise floor. We are investigating alternate methods of charge management to see if such an advance is possible for our torsion balances. As our fibers have improved it has 13.6 also become necessary to deepen our Amplitude (radians) understanding of the effects of resid- 13.5 ual gas on our pendulum’s Q. At present, the Qs of our small force ex- periments are dominated by losses in 13.4 the fiber. Using our new Q measure- ment system, we have shown that, 13.3 for some pendulum geometries, the 0 250 500 750 Time (Oscillator Periods: 90 s) system’s Q will be dominated by gas Figure 2.3-1. Excerpt from our data for an uncoated damping if improved fibers are used. quartz fiber. The fitted curve is an exponential decay This result has important ramifica- with Q = 130000. The tungsten fibers presently used tions for future experimental design. in our operating experiments have Q ∼ 4000.

  • Page 44

    32 2.4 Tests of the gravitational inverse-square law below the dark-energy length scale E. G. Adelberger, T. S. Cook, B. R. Heckel, C. D. Hoyle, D. J. Kapner∗ and H. E. Swanson For the past year we have been working to confirm or disconfirm the apparent short-ranged gravitational anomaly seen in the initial data of our 21-fold symmetric torsion pendulum.1 With a thinner bottom attractor and modified apparatus we have completed two additional data sets, improved our modeling and analysis software, and explored numerous systematic effects. This search has not yielded any reasonable culprit to explain the discrepancy of the initial data set, however the bulk of the data now seem to confirm Newtonian gravity (see Fig. 2.4-1). Progress is also being made on the “Wedge” pendulum,2 which will be the next phase of our test of Newton’s inverse-square law. The pendulum and attractor are made of 50µm thick rhenium foil with a 120-fold symmetric wedge pattern made by electric discharge machining. They are currently being cut by the instrument shop. The wedge pendulum will also get a number of apparatus improvements including a new drumhead type electrostatic screen that will improve the ease of clearing dust and an angle encoder for precise measurement of attractor rotation. Figure 2.4-1. PRELIMINARY 95% exclusion limits on |α| and λ, the strength and range of a Yukawa potential. The line labeled “Eöt-Wash 2006” indicates the sensitivity of our most recent data set as compared to our previously published limits labeled “Eöt-Wash 2004”. ∗ Presently at Kavli Institute for Cosmological Physics, University of Chicago, Chicago IL, 60637. 1 CENPA Annual Report, University of Washington (2005) p. 13. 2 CENPA Annual Report, University of Washington (2005) p. 14.

  • Page 45

    UW CENPA Annual Report 2005-2006 May 2006 33 2.5 Small force investigations for LISA∗ J. H. Gundlach, C. A. Hagedorn, B. R. Heckel, M. J. Nickerson, C. T. O’Rourke, S. E Pollack and S. Schlamminger We are conducting low-noise force investigations for the Laser Interferometer Space Antenna (LISA)1 being developed jointly by NASA and ESA for the study of gravitational waves from astronomical sources. Our work has concentrated on quantifying small forces which act on the proof masses housed within the LISA sciencecraft. Our measurements are carried out under similar operating conditions of LISA and within the relevant frequency band of 0.1 to 100 mHz. In this manner reliable models may be formulated which predict the performance and thereby direct the design of LISA. In the past year we have improved our existing low- noise torsion balance to better estimate these forces. Our previous pendulum and field mass plate were both gold coated glass. To better simulate the LISA case of a Au/Pt proof mass and housing we have made a new pendulum and field mass out of conducting materials. Our new gold coated silicon pendulum is slightly larger than the old pendulum yielding a 1.7 increase in force sensitivity. The new opposing surface consists of two, electrically disconnected, gold coated pieces of copper placed a few hundred microns side-by-side. The two plates, together larger than the pendulum, simulate better the enclosed proof mass of LISA. A translation stage allows us to change the plate- pendulum separation by up to 10 mm. The current planned separation between the proof masses and housings in LISA is 4 mm. Using the control electrodes installed last year we run the pendulum in electrostatic feedback so that the surface of the pendulum remains parallel to the surface of the plates. By changing the electrostatic potential on each plate and measuring the required torque needed to maintain the pendulum position we have determined the electrostatic contact potentials between each plate and the pendulum. Fig. 2.5-1 contains our results for the first few months of doing these measurements. We noticed a change in one plate of about 15 mV over the first month, but otherwise the two plates appear to have relatively constant contact potentials within 10 mV over three months of observation. Figure 2.5-1. Plate-pendulum contact potentials during the first few months with our new pendulum and field mass bi-plate. Aside from some variation at early times, the contact potentials appear to be constant within 10 mV. ∗ Not supported by DOE CENPA grant 1 http://lisa.jpl.nasa.gov

  • Page 46

    34 2.6 Progress on the torsion pendulum based axion search E. G. Adelberger, B. R. Heckel, S. A. Hoedl, D. Schultheis and H. E. Swanson Calculations indicate that a torsion pendulum search for the axion will offer an improvement of 1018 over the most recent measurement1 for an axion mass of ∼ 200 µeV. The axion is the result of the hypothesized Peccei-Quinn symmetry and is a favored cold dark matter candidate.2 The mass is constrained by the known flat geometry of the universe to be heavier than 1 µeV (λa < 20 cm), and is constrained by the neutrino flux from SN1987A to be lighter than 1000 µeV (λa > 0.02 cm). Note that microwave cavity searches probe for light axions (1.2 µeV < ma < 12.4 µeV). A torsion-pendulum based search is possible because the axion mediates a macroscopic pseudo-scalar potential between polarized and unpolarized fermions. By observing the motion of a germanium planar torsion pendulum (source of unpolarized fermions) position near the pole faces of an energized ferromagnet, we can observe such a force. In the past year, we have constructed and evaluated three different electromagnet ge- ometries, in addition to modeling and designing an appropriate cooling system to maintain a constant magnet temperature. In addition, we have developed a method for potting the magnet in a highly filled epoxy, enabling the magnet to operate at vacuum pressures near 10−7 Torr. The characterization of our first magnet assembly is complete, and a correspond- ing silicon pendulum, designed with the measured magnetic field gradients in mind, is in the process of fabrication. We intend to have a first publication of this experiment complete by the end of the calender year. 1e-15 1e-20 Proposal Ni Youdin (gs gp)/(−h c) 1e-25 1e-30 1e-35 Θ = 10-9 1e-40 Θ = 10-14 1e-45 1 10 100 1000 10000 Mass (µeV) Figure 2.6-1. Our expected sensitivity to the axion electron-nucleon coupling as a function of the axion Compton wavelength compared with recent experimental searches and the expected coupling for different values of ΘQCD ; a picture of our axion-pendulum apparatus. 1 W. T. Ni et al. Phys. Rev. Lett. 82, 2439 (1999). 2 L. J. Rosenberg and K. A. van Bibber, Phys. Rep. 325, 1 (2000).

  • Page 47

    UW CENPA Annual Report 2005-2006 May 2006 35 2.7 Eöt-Wash data acquisition development H. E. Swanson and the Eöt-Wash Group Data acquisition for the two rotating balances and the short-range balance has been ongoing for the last year, and no significant changes have been made to their respective computer codes or hardware. A new data acquisition system has been assembled for the axion search apparatus. The hardware consists of a PC with a 2.4 GHz Athalon, 1 GB of RAM and National Instruments multifunction DAQ and GPIB interfaces (PCI-6229 and GPIB-USB-HS). We were able to reuse much of the software developed for the other instruments with the exception of the rou- tines used to communicate with the DAQ hardware. When the short-range data acquisition system was upgraded we chose a National Instruments E-series DAQ board. The hardware interface used here is newer, provides a more useful mix of ADC, DAC and Digital I/O chan- nels and is also less costly. It is unfortunately not compatible with the E-Series routines used in the short range system. We have developed a new set of software procedures to interface with the board’s NI-DAQmx drivers. The system is currently operational and has so far been used to investigate temperature stability in the apparatus environment.

  • Page 48

    36 2.8 Test of Newton’s second law for small accelerations K-Y. Choi, J. H. Gundlach, S. Schlamminger and C. D. Spitzer One of the outstanding problems in astrophysics is the unexpected flatness of galactic rotation curves. Traditionally, dark matter has been introduced to solve this problem. A phenomeno- logical solution explaining the rotation curves can be formulated by modifying Newton’s sec- ond law. As early as 1983, M. Milgrom suggested such a modification, called MOdified New- tonian Dynamics (MOND).1 Newton’s second law was to be replaced by F = mg µ(a/a0 )a. The function µ(a/a0 ) approaches one, and thereby restores Newton’s second law, for accel- erations much larger than the characteristic acceleration a0 . For accelerations smaller than a0 the function reduces to µ(a/a0 ) = a/a0 . A torsion pendulum can be used to test such a modification of Newton’s second law.2 The proposed modification would change the linear differential equation of motion to a nonlinear equation, without an analytic solution. We have programmed a numerical simulation using a realistic model of the pendulum to predict the torsional excursion as a function of time. We then fit the lowest frequency in this trace as a function of amplitude. The lowest frequency in such a model increases for smaller amplitudes as a function of a0 . We used the Newwash torsion balance (see Sec. 2.2) to measure the frequency of the pen- 2.5 dulum for small amplitudes. The pendulum, 2.0 starting at rest, assumes on average an am- Frequency (mHz) plitude of approximately 100 nrad within one 1.5 period. In order to measure at small ampli- tudes, we stopped the data collection after the 1.0 pendulum has acquired a significant torsional 0.5 amplitude. We then damped the pendulum and started a new run. By doing this many 0.0 -8 -7 -6 1.0x10 1.0x10 1.0x10 times we managed to get a reasonable distri- Amplitude (rad) bution of amplitudes down to 20 nrad. We di- vided the recorded torsion angle as a function Figure 2.8-1. The measured frequency as a function of amplitude. The solid line is the of time into 1600 s long sections. To each of best fit, yielding a0 = (7 ± 14) × 10−13 cm/s2 , the section, we fit the average amplitude and the dashed line shows the predicted frequency frequency of the oscillation. The results of the for a0 = 10−11 cm/s2 . fits are shown in Fig. 2.8-1. From a very preliminary analysis of our data we get a0 = (7 ± 14) × 10−13 cm/s2 . This value is three orders of magnitude better than the previously published test3 of F=ma and four orders of magnitude smaller than the required a0 for a MOND fit of the galactic rotation curves. Current MOND scenarios require that no absolute accelerations are present. This condition cannot be obtained on earth for extended periods of time. 1 M. Milgrom, Astrophys. J. 270, 365 (1983). 2 E. Fischbach encouraged us to perform the measurements described here. 3 A. Abramovici and Z. Vager, Phys. Rev. D 34, 3240 (1986).

  • Page 49

    UW CENPA Annual Report 2005-2006 May 2006 37 2.9 Status of the APOLLO lunar laser ranging project∗ E. G. Adelberger, C. D. Hoyle, H. E. Swanson, T. D. Van Wechel and the APOLLO Collaboration Fundamental properties of gravity, including the Strong Equivalence Principle and Ġ/G, are best probed by the shape of the lunar orbit. APOLLO (Apache Point Lunar Laser-ranging Operation)1 will measure this shape with unprecedented precision by detecting pulsed laser light (532 nm) returned from retroreflectors on the lunar surface. We aim for an order of magnitude improvement upon current measurements that characterize the lunar orbit at the 4 mm level.2 We have made much progress toward achieving “science-quality” data from APOLLO in the last year. All major mechanical work was completed and initial operation/acquisition software allowed for the first ranging operations in July, 2005 and the first detection of lunar return photons in October (see Fig. 2.9-1). To date, we have received several thousand lunar returns. This quantity of photons, collected in only a few hours of observation time, is equivalent to the number acquired over several years at other ranging stations. Additionally, the round-trip travel time of the photons agrees with the JPL lunar orbit prediction within several hundred picoseconds (corresponding to a distance uncertainty of only a few cm). The remaining hardware and software tasks required to achieve millimeter precision have recently been completed, and we are presently analyzing our first potential science-quality data. Figure 2.9-1. Return time histogram of the first lunar photons detected by APOLLO. The peak position and width are consistent with the JPL lunar orbit prediction. In addition to these accomplishments, other key items completed in the previous year include development of an automated laser safety interlock system, as well as installing an infrared camera that will ensure that the beam does not intercept overflying aircraft. Finally, the Python/Tkinter data acquisition user interface software, developed here at CENPA, has achieved a mature state with many useful features that will allow remote operation of the entire system in the coming months. ∗ Not supported by DOE CENPA grant 1 T. W. Murphy, et al., Nucl. Phys. B, 134, 155 (2004). 2 J. G. Williams, S. G. Turyshev and D. H. Boggs, Phys. Rev. Lett. 93, 261101 (2004).

  • Page 50

    38 2.10 APOLLO laser safety system∗ E. G. Adelberger, C. D. Hoyle, H. E. Swanson and The APOLLO Collaboration APOLLO, the lunar laser ranging program at the Apache Point Observatory (APO) operates a green (532 nm) pulsed laser with an average power of about 2 watts. This laser is classified as a class 4 vision hazard. The purpose of the safety system is two fold; to help prevent eye damage resulting from inadvertently looking directly into the laser beam or its reflections and to comply with the FAA requirement that the laser not be pointed at an aircraft in flight. The FAA has the further requirement that whenever the laser is active there will be two observers with the ability to immediately shut down the laser should they see any aircraft approach the general location of the beam. There are spotter stations located on the catwalk on either side of the telescope. The spotters operate hand held kill switches that plug into weather proof outlets at each station. As an additional safety measure we purchased a long wavelength infrared (LWIR) camera system that is able to detect aircraft in its field of view and provide a fast signal to shut down the laser. The camera has been mounted on the truss that supports the telescope’s secondary mirror and bore sighted along the observing axis. The safety system makes use of a spring loaded shutter that interrupts the laser beam before it leaves the enclosure. This shutter requires an active signal to open and is independent of the laser’s electronics. Safe laser operation is imposed by requiring that the following conditions (interlocks) are satisfied before the shutter can be opened or remain open: The door to telescope dome must be shut. The telescope motor controller is not set to “emergency stop”. Neither Left nor right Spotter has activated a kill switch. The LWIR camera’s field of view remains clear. APOLLO’s control computer signals it’s OK to range. The operator’s console is located in the telescope’s control room. A key operated switch enables the shutter. The console has a diagnostic display showing the status of the interlocks and push buttons for opening and closing the shutter. There is a second operator’s console located near the APOLLO apparatus in telescope dome. The interlock logic is implemented in an Allen Bradley SLC500 Programmable Logic Controller. ∗ Not supported by DOE CENPA grant

  • View More

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