Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                

Precision laser-based measurements of the single electron response of spherical proportional counters for the NEWS-G light dark matter search experiment

Q. Arnaud et al. (NEWS-G Collaboration)
Phys. Rev. D 99, 102003 – Published 20 May 2019
PDFHTMLExport Citation
Abstract
Authors
Article Text
  • INTRODUCTION
  • EXPERIMENTAL SETUP
  • ANALYSIS METHODOLOGY
  • APPLICATIONS
  • CONCLUSION AND DISCUSSION
  • ACKNOWLEDGMENTS
  • APPENDICES
    • References

    Abstract

    Spherical proportional counters (SPCs) are a novel gaseous detector technology employed by the NEWS-G low-mass dark matter search experiment for their high sensitivity to single electrons from ionization. In this paper, we report on the first characterization of the single electron response of SPCs with unprecedented precision, using a UV-laser calibration system. The experimental approach and analysis methodology are presented along with various direct applications for the upcoming next phase of the experiment at SNOLAB. These include the continuous monitoring of the detector response and electron drift properties during dark matter search runs, as well as the experimental measurement of the trigger threshold efficiency. We measure a mean ionization energy of W=27.6±0.2  eV in Ne+CH4 (2%) for 2.8 keV x-rays, and demonstrate the feasibility of performing similar precision measurements at sub-keV energies for future gas mixtures to be used for dark matter searches at SNOLAB.

    • Figure
    • Figure
    • Figure
    • Figure
    • Figure
    • Figure
    • Figure
    1 More
    • Received 26 February 2019

    DOI:https://doi.org/10.1103/PhysRevD.99.102003

    © 2019 American Physical Society

    Physics Subject Headings (PhySH)

    1. Research Areas
    Particle dark matter
    1. Techniques
    Dark matter detectorsData analysisGaseous detectors
    Gravitation, Cosmology & AstrophysicsStatistical Physics & ThermodynamicsInterdisciplinary Physics

    Authors & Affiliations

    Click to Expand

    Article Text (Subscription Required)

    Click to Expand

    References (Subscription Required)

    Click to Expand
    Issue

    Vol. 99, Iss. 10 — 15 May 2019

    Reuse & Permissions
    Access Options
    Author publication services for translation and copyediting assistance advertisement

    Authorization Required

    Log In
    Other Options
    ×

    Images

    • Figure 1
      Figure 1

      The experimental setup as described in Sec. 2. The 213 nm laser light is sent through an optical fiber splitter to both the PD (which triggers the acquisition) and the SPC to extract photoelectrons from the inner surface of the vessel. The two panels on the right show a typical PD signal (raw pulse on top panel) together with the resulting SPC signal (treated pulse on bottom panel) from a single electron reaching the sensor and undergoing an avalanche of average gain. The time delay between the SPC and the PD pulse corresponds to the drift time of the electron from the surface to the sensor.

      Reuse & Permissions
    • Figure 2
      Figure 2

      Energy spectra (SPC channel) of laser-induced events associated with low (left panel) and high (right panel) PD pulse amplitudes. In both panels, the fit of our model (described in Sec. 3a) to the data is shown as a solid red line. The relative contribution of the null events and single and multiple electron events are shown as solid orange, green and grey lines, respectively. The top axis gives the energy scale in the average number of primary electrons ⟨PE⟩ based on the best fit value of the mean gain. The reduced χ2 of the fit (normalized to the 121 degrees of freedom) is indicated as well.

      Reuse & Permissions
    • Figure 3
      Figure 3

      Results from the fit of the energy spectra for laser-induced events recorded during the same run as in Fig. 2. On all panels results are reported as a function of the mean amplitude of the PD signals in each of the eight data sets. The color code indicates if the results are obtained from performing an individual fit (black) or a joint fit of all subsets (blue). Top left panel: Markers correspond to the best fit values of the mean number of electrons μ while the two lines show the result from the fit of a first-order polynomial to the markers of the matching color. Both methods confirm the linearity between μ and the laser power. Top right panel: Values of the χ2 normalized to the number of degrees of freedom for each subset are shown both for the individual fits (black dots) and joint fit (blue dots). The green (orange) dashed lines correspond to the 1σ (2σ) confidence intervals derived from the chi-squared distribution with 121 degrees of freedom to assess the goodness of individual fits. Bottom panels: Black markers correspond to the best fit values of θ (left) and of the mean gain ⟨G⟩ (right) for each data set. The best fit values from the joint fit are reported as a solid blue line. The light blue band corresponds to the 1σ confidence region calculated with a profile likelihood approach. These correspond to θ=0.09±0.02 and ⟨G⟩=30.26±0.21  ADU. Note that 1  ADU=186±19  SE (secondary electrons).

      Reuse & Permissions
    • Figure 4
      Figure 4

      Energy spectrum of non-laser-induced events recorded during laser calibration measurements in Ne+CH4 (2%) at 1.5 bar with HV1=1150  V. The spectrum clearly shows the 270 and 2822 eV lines of x rays from electron capture in the L- and K-shell of Ar37, respectively. The energy scale is determined based on the position of the 2822 eV peak. The dashed line indicates the analysis threshold that was set at 100 eV. The solid red line indicates the fit of our model to the data. Our modeling of the detector response accounts both for primary ionization statistics with the Conway-Maxwell-Poisson (COM-Poisson) distribution and for statistical fluctuations of the avalanche gain with the Polya distribution. See core text for more details.

      Reuse & Permissions
    • Figure 5
      Figure 5

      Top panel: Energy spectrum of laser-induced events with (blue markers) or without (black markers) triggering on the SPC channel. The energy scale is indicated—on the top axis—in the average number of primary electrons ⟨PE⟩ based on the determination of the mean gain from the fit of our model (solid red line) to the total energy spectrum (black markers). Bottom panel: Relative fraction of events triggering on the SPC channel as a function of the energy. The error bars indicate the statistical (binomial) uncertainty at the corresponding energy. The red and blue curves show the trigger efficiency curves as derived from the two methods discussed in the core text. These yield a trigger threshold value of Eth1=0.562  ⟨PE⟩ (Eth2=0.565  ⟨PE⟩) and a standard deviation of σth1=0.114  ⟨PE⟩ (σth2=0.115  ⟨PE⟩).

      Reuse & Permissions
    • Figure 6
      Figure 6

      Monitoring of the stability of the gain over time using a UV laser. The top panel shows the distribution in SPC pulse amplitude vs time of laser-induced events corrected for the laser instability using the PD pulse amplitude. The middle and bottom panels show the distribution of Ar37 2822 eV events before (middle panel) and after (bottom panel) correcting for gain variations using the position of laser-induced events. The red markers indicate the center of a Gaussian fitted to amplitude spectra for slices in time of ∼15  min width.

      Reuse & Permissions
    • Figure 7
      Figure 7

      Monitoring of the stability of the mean drift time of laser-induced events. The red markers indicate the center of a Gaussian fitted to drift time spectra for slices in time of ∼20  min width.

      Reuse & Permissions
    • Figure 8
      Figure 8

      Effect of the bias induced on the estimators of the mean gain ⟨G⟩ (top panel) and of the Polya parameter θ by ±5% random fluctuations of the laser power. As explained in more detail in the core text, simulated data sets which are generated by Monte Carlo account for ±5% random fluctuations of μ on an event-per-event basis. Input values used for the simulations are indicated between the two panels. On each panel, the black markers and the associated error bar are derived from the mean and standard deviation of a Gaussian fit to the distribution of the best fit values of the corresponding parameter of interest. As attested to by their small deviation with respect to the dashed blue line that indicates the input value of θ and ⟨G⟩ in the simulation, the bias is negligible.

      Reuse & Permissions
    ×

    Sign up to receive regular email alerts from Physical Review D

    Log In

    Cancel
    ×

    Search

    Article Lookup

    Paste a citation or DOI
    Enter a citation
    ×