Experimental parameters for a Cerium 144 based intense electron antineutrino generator experiment at very short baselines

The standard three-neutrino oscillation paradigm, associated with small squared mass splittings $\mathrm{\Delta }{m}^2\ll 0.1{\mathrm{eV}}^2$, has been successfully built up over the last 15 years using solar, atmospheric, long baseline accelerator and reactor neutrino experiments. However, this well-established picture might suffer from anomalous results reported at very short baselines in some of these experiments. If not experimental artifacts, such results could possibly be interpreted as the existence of at least an additional fourth sterile neutrino species, mixing with the known active flavors with an associated mass splitting $\mathrm{\Delta }{m}_{\text{new}}^2\gg 0.01{\mathrm{eV}}^2$ and being insensitive to standard weak interactions. Precision measurements at very short baselines (5–15 m) with intense MeV ${\overline{\nu }}_e$ emitters can be used to probe these anomalies. In this article, the expected ${\overline{\nu }}_e$ signal and backgrounds of a generic experiment which consists of deploying an intense ${\beta }^{-}$ radioactive source inside or in the vicinity of a large liquid scintillator detector are studied. The technical challenges to perform such an experiment are identified, along with quantifying the possible source- and detector-induced systematics and their impact on the sensitivity to the observation of neutrino oscillations at short baselines.

Figure 1

Simplified decay scheme of the ${}^{144}\mathrm{Ce}\text{−}{}^{144}\mathrm{Pr}$ couple. $\beta$ branches with branching ratios greater than 0.001% are displayed, along with the corresponding Log(ft) values, daughter nucleus level energies and spin parities. The main gamma transitions (intensity greater than 0.1%) among the excited states of the ${}^{144}\mathrm{Pr}$ and ${}^{144}\mathrm{Nd}$ nuclei are also displayed, together with their corresponding energies. The full ${}^{144}\mathrm{Ce}$ and ${}^{144}\mathrm{Pr}$ decay scheme data can be found in the National Nuclear Data Center, the Chart of Nuclides database[18].

Figure 2

Estimation of the ${\overline{\nu }}_e$ spectrum without neutrino oscillations (left panel) and of the $\beta$ spectrum (right panel) of the ${}^{144}\mathrm{Ce}\text{−}{}^{144}\mathrm{Pr}$ couple. Both spectra are normalized to two decays.

Figure 3

Differential observable neutrino spectrum from ${}^{144}\mathrm{Pr}$ folded with the IBD cross section.

Figure 4

Expected signal as a function of both visible energy ${\mathrm{E}}_{\text{vis}}$ and baseline L, in the no-oscillation hypothesis, for an ANG positioned either outside (left) or at the center (right) of a liquid scintillator spherical detector. See text for further details.

Figure 5

Sketch of the generic experimental setup.

Figure 6

The upper panel shows the expected signal as a function of L/E. Error bars include statistical uncertainties corresponding to 1.5 years of data taking and a 1.5% uncertainty on the source activity. The no-oscillation scenario is compared to two oscillation scenarios, with squared mass splittings $\mathrm{\Delta }{m}_{\text{new}}^2=0.5$ and $2.0{\mathrm{eV}}^2$ and a ${\sin }^2(2{\theta }_{\text{new}})=0.1$ mixing parameter. The lower panel shows the ratio of the oscillated spectra to the non-oscillated spectrum.

Figure 7

Sensitivity of our generic reference experiment at 90%, 95%, and 99.73% C.L. The inclusion contours of the reactor antineutrino anomaly are shown in the grey shaded areas. Left panel: $\text{Rate}+\text{shape}$ sensitivity. Right panel: Free rate sensitivity. Unless explicitly mentioned, default parameters for sensitivity contours include a ${}^{144}\mathrm{Ce}\text{−}{}^{144}\mathrm{Pr}$ source of 3.7 PBq deployed at 10 m from the detector center for 1.5 years, with 1.5% uncertainty on normalization, vertex and energy resolution at 1 MeV of 15% and 5%, respectively, and contours given at the 95% C.L.

Figure 8

Sensitivity contours of a source experiment at the KamLAND, Borexino, and $\mathrm{SNO}+$ detectors. See Fig. 7 for parameters. Left panel: $\mathrm{Rate}+\mathrm{shape}$ sensitivity. Right panel: Free rate sensitivity.

Figure 9

Illustrative contours for input oscillation signals at $\mathrm{\Delta }{m}_{\text{new}}^2=0.5$, 1, $2{\mathrm{eV}}^2$ and ${\sin }^2(2{\theta }_{\text{new}})=0.1$. Acceptance contours are given at the 99% confidence level. See Fig. 7 for other parameters.

Figure 10

Influence of source activity on sensitivity, with a $\text{rate}+\text{shape}$ analysis (left panel) and a free rate analysis (right panel). See Fig. 7 for other parameters.

Figure 11

Left panel: Influence of normalization uncertainty on sensitivity. Right panel: Influence of source extension on sensitivity. See Fig. 7 for other parameters.

Figure 12

Influence of distance between source and detector center on sensitivity, with a $\text{rate}+\text{shape}$ analysis (left panel) and a free rate analysis (right panel). In the extreme case the overall statistics vary from $6.2\times {10}^2$ events to $2.7\times {10}^4$ events. See Fig. 7 for other parameters.

Figure 13

Influence of data-taking time on sensitivity, with a $\text{rate}+\text{shape}$ analysis (left panel) and a free rate analysis (right panel). See Fig. 7 for other parameters.

Figure 14

Influence of energy resolution on sensitivity, with a $\text{rate}+\text{shape}$ analysis (left panel) and a free rate analysis (right panel). See Fig. 7 for other parameters.

Figure 15

Influence of vertex resolution on sensitivity, with a $\text{rate}+\text{shape}$ analysis (left panel) and a free rate analysis (right panel). See Fig. 7 for other parameters.

Figure 16

Simulation of the number of gamma energy deposits from the steel plate to the target, per one initial gamma ray at 2.185 MeV (left panel) and per one initial neutron shot in a Watt spectrum (right panel). Error bars are statistical only. Since the gamma simulation is much faster and easier to bias, the statistics is higher and error bars are contained in the points. The neutron simulation shows several points with very high error bars, a symptom of poor convergence in the considered bins. It is worth noting that direct neutron energy deposits are not considered here, only the energy deposits of gammas from neutron reactions. High-energy gammas mainly come from neutron capture on iron, as indicated by the peaks in the spectrum in the steel plate. At lower energy, the 2.2-MeV peak signs the neutron capture on hydrogen and the 0.5-MeV peak comes from pair creation.