Ultralow energy calibration of LUX detector using ${}^{127}\mathrm{Xe}$ electron capture

We report an absolute calibration of the ionization yields ($Q_y$) and fluctuations for electronic recoil events in liquid xenon at discrete energies between 186 eV and 33.2 keV. The average electric field applied across the liquid xenon target is $180\mathrm{V}/\mathrm{cm}$. The data are obtained using low energy ${}^{127}\mathrm{Xe}$ electron capture decay events from the 95.0-day first run from LUX (WS2013) in search of weakly interacting massive particles. The sequence of gamma-ray and x-ray cascades associated with ${}^{127}\mathrm{I}$ deexcitations produces clearly identified two-vertex events in the LUX detector. We observe the K-(binding energy, 33.2 keV), L-(5.2 keV), M-(1.1 keV), and N-(186 eV) shell cascade events and verify that the relative ratio of observed events for each shell agrees with calculations. The N-shell cascade analysis includes single extracted electron (SE) events and represents the lowest-energy electronic recoil in situ measurements that have been explored in liquid xenon.

Figure 1

Decay scheme of ${}^{127}\mathrm{Xe}$ [25] with units of keV. The ${}^{127}\mathrm{Xe}$ decays via electron capture to ${}^{127}\mathrm{I}$. The percentage above the transition arrow is the gamma-ray intensity as a fraction of parent (${}^{127}\mathrm{Xe}$) decay.

Figure 2

A schematic illustrating atomic electron capture (a K-shell electron in this case), for a ${}^{127}\mathrm{Xe}$ nucleus, which is converted into ${}^{127}\mathrm{I}$ in an excited state. The excited ${}^{127}\mathrm{I}$ nucleus can subsequently deexcite via emission of one or more gamma rays (or IC electrons). The atomic structure deexcites through x-ray (or Auger electron) cascade emissions.

Figure 3

Schematics (not to scale) of ${}^{127}\mathrm{Xe}$ decay events in the LUX detector where both the x ray and gamma ray have ER interactions in the active volume. Due to the relatively short MFP, the x-ray ER interaction site is considered the same as where the initial nuclear EC happens. Depending on the component of the gamma-ray travel in the vertical direction, the subsequent drift readout of the event in the S2 can appear as two S2s merged with each other (left), as a small (x-ray deposition) S2 followed by a large (gamma-ray deposition) S2 (middle), or a large S2 followed by a small S2 (right).

Figure 4

The time trace, using 10 ns samples, in LUX summed across all 122 PMTs for a ${}^{127}\mathrm{Xe}$ decay event with K-shell electron capture in LUX data (top). The event appears as a clear double scatter. An S1 signal is followed by two well separated S2s. The first (green) S2 is due to K-shell x-ray ER, and the second (blue) S2 is due to the 203 keV gamma-ray ER in the LXe volume. The two S1s from both ERs overlap and appear as one S1 (cyan). The (bottom) plot is a closeup of the two S2s.

Figure 5

Double-scatter events associated with ${}^{127}\mathrm{Xe}$ decays. The “First vertex” is the first S2 ordered by drift time. The S2 sizes have been converted to the number of electrons initially drifted away from the interaction site. The two arced loci of events with higher and lower signal sizes are due to the deexcitation from 375 keV state and 203 keV state, respectively. The first four populations to the left of the line $\mathrm{S}{2}_1=\mathrm{S}{2}_2$ on the 203 and 375 keV bands are EC double scatter events with well isolated K-, L-, M-, and N-shell x-ray S2s from right to left. The populations labeled SE represent a background of single extracted electrons that are emitted from the liquid surface into the gas. This later feature is not centered on ${\log }_{10}$ ($\mathrm{S}{2}_1$ [electron]) $=0$ because efficiency corrections are made to calibrate the corresponding signals at the interaction sites (see Sec. 3b for more details).

Figure 6

Histograms of vertical separation ($1\mu \mathrm{s}$ is equivalent to 1.51 mm) between two vertices of EC events with the 203 keV gamma-ray emission for K-(cyan), L-(red), M-(green), and N-(blue) shell x rays. The distributions are fitted with the model described by Eq. (1). The solid black curves are the fits to data. The dashed black curves are the extrapolations from those fits.

Figure 7

X-rays’ ER charge spectra. The S2 size has been converted to the number of electrons escaping recombination at the interaction site. The peaks from right to left are due to K-, L-, M-, and N-shell x rays respectively. The fits shown use Gaussian functions.

Figure 8

The black (solid and dashed) histogram shows the N-shell x-ray charge signals (with background). The red (solid and dashed) lines are the data-driven background model. Both dashed black and red lines mainly populated by SEs are not used for signal extrapolation. The blue data points are the N-shell x-ray charge spectrum after background subtraction fit with a Gaussian function (blue curve).

Figure 9

The reconstructed energy spectrum of single-scatter events in the energy range of interest. The “Signal Window” indicates the energy range where N-shell x-ray signals (Fig. 8 solid blue) are exploited, while “BG Window” indicates the energy range where the N-shell x-ray calibration background model (Fig. 8 solid red) is defined.

Figure 10

The $Q_y$ mean values (green points) measured from ${}^{127}\mathrm{Xe}$ EC events (this work) at $180\mathrm{V}/\mathrm{cm}$. The red bar indicates the systematic uncertainty on the ${}^{127}\mathrm{Xe}$ measurement due to $g_2$. Also shown are the $Q_y$ measured from the LUX tritium calibration [21] at $180\mathrm{V}/\mathrm{cm}$, of the NEST v0.98 model [11] at $180\mathrm{V}/\mathrm{cm}$, and measured by the neriX experiment [35] at $190\mathrm{V}/\mathrm{cm}$. The light gray band on the tritium measurement indicates the systematic uncertainty due to $g_2$.