Low Energy Phonon Bursts Created By Fast Neutron Damage

Abstract

Solid state athermal phonon calorimeters used in the search for low mass dark matter or coherent neutrino-nucleus interactions have long observed a large excess of events below several hundred eV. The relaxation of damage created by the interaction of fast cosmic ray neutrons with the detector has been proposed as a source of these excess events. By comparing neutron exposed detectors to control detectors, we report the first measurement of phonon bursts caused by damage created by fast neutrons. Differences in the spectral shape, the rate dependence on thermal history, and the observed spectral rate scaled to the neutron exposure between irradiated and control detectors suggest that our observed LEE backgrounds are not dominated by neutron damage-induced phonon bursts.

Figures (18)

• Figure 1: Diagrams showing the proposed neutron damage and relaxation process. (Left) Fast neutrons recoil off individual atoms in the detector crystal. (Center Bottom) The energetic "primary knock-on atom" ricochets through the detector crystal, impacting additional atoms, creating extended defects. As atoms are displaced from the regular crystal lattice, this defect constitutes a metastable energetic state. (Right Bottom) After creation, the defect relaxes into a lower energy configuration, emitting athermal phonons. (Right Top) Photograph of the detectors, mounted next to each other in the cryogenic package. In each of the sub-diagrams, the tungsten TES films are marked in pink, while the aluminum films are marked in green.
• Figure 2: (Left) Set A spectra, measured approximately 6 hours after cooldown. Yellow corresponds to a detector irradiated with fast neutrons from AmBe and DD sources (see text), while blue corresponds to the control (unexposed) detector. For comparison, purple shows the spectrum on day 3.4 in the 1 mm detector in Ref. Run57Paper. (Right) Spectra measured for the detectors in set B, measured approximately 1.1 days after cooldown. Red corresponds to the detector exposed to DD neutrons (see text), green is control, and light blue is the "CPDv2" detector spectrum shown in Ref. Run57Paper.
• Figure 3: Rate of events between 15 and 30 eV (where neutron-induced excess events dominate) observed in the detectors in sets A and B. Irradiated detectors are shown in bright colors with crosses, control detectors are shown in dark colors with dots. Dashed lines show fits to Eqn. \ref{['eqn:pwr_law']}, with 1 sigma uncertainties shaded. Dotted lines are fits with the power law exponent fixed (Eqn. \ref{['eqn:pwr_law_fix']}).
• Figure 4: Rate of 15-30 eV events in the set A detectors, showing the change in observed rate after warming up to 50 K for three days. Dashed lines show the same fits as in Fig. \ref{['fig:time']}, while dotted lines show additional power law fits (Eqn. \ref{['eqn:pwr_law_warmup']}). Note that the warm up period significantly reduces the observed background in the irradiated device.
• Figure 5: Diagram showing the channels energy can partition into following a neutron interaction in the a detector crystal. We determine the spectrum of incoming cosmic ray neutrons using EXPACSEXPACS1EXPACS2EXPACS3EXPACS4, and take the AmBe spectrum from Ref. AmBeStudy. Here, we estimate the amount of (non-ionizing) nuclear recoil energy as a proxy for the total amount of damage done to the crystal, which we assume is related to the rate of delayed phonon bursts from damage relaxation, i.e. LEE-type events. We make this estimate using two techniques (see text): a NIEL based model which directly estimates the nuclear recoil energy from the incoming neutron spectrum, and a model that uses Geant4 to simulate the energy deposited in the crystal by incoming neutrons and a Lindhard yieldLindhard1961 model to estimate the nuclear recoil energy (as opposed to ionizing energy) created in these events.