Scintillation yield and time dependence from electronic and nuclear recoils in liquid neon

We have performed measurements of scintillation light in liquid neon, observing a signal yield in our detector as high as (3.5 $\pm$ 0.4) photoelectrons/keV. We measure pulse shape discrimination efficiency between electronic and nuclear recoils in liquid neon from 50 to 300 keV nuclear recoil energy. We also measure the ${\mathcal{L}}_{\mathrm{eff}}$  parameter in liquid neon between 30 and 370 keV nuclear recoil energy, observing an average ${\mathcal{L}}_{\mathrm{eff}}$$=0.24$ above 50 keV. We observe a dependence of the scintillation time distribution and signal yield on the pressure and temperature of the liquid neon.

Figure 1

Schematic representation of the MicroCLEAN scintillation cell.

Figure 2

The gas handling system for the MicroCLEAN runs described in this paper. The circulation pump consists of a heater that evaporates the liquid; the gas then flows through the charcoal trap before being reliquefied.

Figure 3

Sample event from a single PMT digitized by the CAEN V1720 waveform digitizer. More than 8 $\mu$s of presamples are recorded to measure the baseline.

Figure 4

Schematic of the neutron scattering setup.

Figure 5

The energy calibration obtained by fitting the Compton edge and photoabsorption hump produced by 511-keV $\gamma$ rays to a simulation of the detector. The green lines denote the fit range. We currently have no satisfactory explanation for the excess of events between 40 and 200 keV in comparing the data to the simulation.

Figure 6

The energy calibration obtained from the ${}^{22}$Na source checked by comparing background data to simulations containing ${}^{40}$K and ${}^{208}$Tl components. We clearly see the two Compton edges of the 1.4- and 2.6-MeV $\gamma$ rays produced by these sources. We do not attempt to simulate any other backgrounds, so the divergence of data and Monte Carlo (MC) below 1000 keV is not surprising. There is evidence for a nonlinearity in the energy scale at these high energies that is likely related to the response of the top PMT, and we correct for this using Eq. (4).

Figure 7

Energy spectrum of a background-subtracted ${}^{83}$Kr${}^{\mathrm{m}}$ run in neon. As discussed in the text, the observed light yield for these data is $(3.0\pm 0.3)$ photoelectrons/keVee. The resolution is $19\%$ (with $\sigma /E$ of the Gaussian fit shown in red) at 41.5 keV.

Figure 8

The light yield in the detector as a function of time with the circulation pump to the charcoal trap engaged for a liquid temperature of 28.7 K as determined by the ${}^{22}$Na source. The light yield reached a maximum at $(3.5\pm 0.4)$ photoelectrons/keVee. The error bars are statistical only and do not include the uncertainty introduced by the single-photoelectron response of the bottom PMT.

Figure 9

Mean prompt fraction as a function of energy for electronic and nuclear recoils at three different temperatures. The error bars include both statistical and systematic uncertainties and are generally dominated by a $2.5\%$ uncertainty stemming from differences in the two PMTs.

Figure 10

ERC vs energy for the three data sets. The $x$ axis is given in relevant units for dark matter searches, keVr, by assuming ${\mathcal{L}}_{\mathrm{eff}}$ $=$ 0.24 (see Sec. 4).

Figure 11

PSD for the 26.7 K data set individually, along with a fit to the statistical model. The solid line shows the model prediction with additional noise terms set to zero. We note that the effect of the additional noise is most relevant for energies above 100 keVr.

Figure 12

Fits for the nuclear recoil scintillation efficiency in liquid neon for all energies. All data sets include results from each temperature setting, except for the data at 178 keVr that were only taken at 26.7 K.

Figure 13

The observed nuclear recoil scintillation efficiency vs nuclear recoil energy in neon, along with the Lindhard $+$ Birks model described in the text.

Figure 14

The PSD cut in the organic scintillator to select neutron events. Only events within the pentagon are included in the data set.

Figure 15

A sample TOF cut. The TOF cut window was 10 ns, and only events at the beginning of the TOF peak were selected, as these events are more likely to be single-scatter events.

Figure 16

The observed energy spectrum for 368.7-keVr events as the location of the TOF cut (as defined by ${\mathrm{TOF}}_{\mathrm{L}}$ and ${\mathrm{TOF}}_{\mathrm{H}}$) is moved by $\pm 10$ and $+20$ ns. The nominal cut around the front side of the TOF peak is the case where ${\mathrm{TOF}}_{\mathrm{L}}<\mathrm{TOF}<{\mathrm{TOF}}_{\mathrm{H}}$. The degradation in the observed peak as the TOF cut moves away from standard location is clear.

Figure 17

The observed energy spectrum for 28.9-keVr events as the location of the TOF cut (as defined by ${\mathrm{TOF}}_{\mathrm{L}}$ and ${\mathrm{TOF}}_{\mathrm{H}}$) is moved by $\pm 10$ and $+20$ ns. The nominal cut around the front side of the TOF peak is the case where ${\mathrm{TOF}}_{\mathrm{L}}<\mathrm{TOF}<{\mathrm{TOF}}_{\mathrm{H}}$. The degradation in the observed peak as the TOF cut moves away from standard location is clear.

Figure 18

The average traces for 27.4 K data, along with fits to two-component and three-component exponential models. Neither model accurately describes the intermediate time regime.

Figure 19

Examples of the four-component model fit to 26.7 and 28.8 K data.

Figure 20

The fitted longest and shortest time constants, ${\tau }_1$ and ${\tau }_2$, for neon scintillation as a function of neon temperature. The error bars represent the combined estimated statistical and systematic uncertainties, derived as described in the text.

Figure 21

The fitted intermediate time constants, ${\tau }_3$ and ${\tau }_4$, for neon scintillation as a function of neon temperature. The error bars represent the combined estimated statistical and systematic uncertainties, derived as described in the text.

Figure 22

The fitted weights of the four-component model for neon scintillation as a function of neon temperature. The error bars represent the combined estimated statistical and systematic uncertainties, derived as described in the text.

Figure 23

The total signal yield plotted with the intensities of the long, short, intermediate, and residual components as a function of temperature multiplied by the total signal yield of four runs taken over the course of three days.