• Abstract and Figures
• Public Full-text

N3AS-21-004 Future Searches for Light Sterile Neutrinos at Nuclear Reactors Jeffrey M. Berryman,1, 2, ∗ Luis A. Delgadillo,3, y and Patrick Huber3, z 1Department of Physics and Astronomy, University of Kentucky, Lexington, KY 40506, USA 2Department of Physics, University of California, Berkeley, CA 94720, USA 3Center for Neutrino Physics, Department of Physics, Virginia Tech, Blacksburg, VA 24061, USA We study the optimization of a green-field, two-baseline reactor experiment with respect to the sensitivity for electron antineutrino disappearance in search of a light sterile neutrino. We consider both commercial and research reactors and identify as key factors the distance of closest approach and detector energy resolution. We find that a total of 5 tons of detectors deployed at a commercial reactor with a closest approach of 25 m can probe the mixing angle sin2 2θ down to ∼ 5 × 10−3 around ∆m2 ∼ 1 eV2. The same detector mass deployed at a research reactor can be sensitive up to ∆m2 ∼ 20 − 30 eV2 assuming a closest approach of 3 m and excellent energy resolution, such as that projected for the Taishan Antineutrino Observatory (TAO). We also find that lithium doping of the reactor could be effective in increasing the sensitivity for higher ∆m2 values. I. INTRODUCTION the 5 MeV bump is entirely unresolved but is likely not due to new physics [17]. The experiments we will dis- Neutrino physics as an experimental science started cuss here will contribute data on neutrino fluxes, but with the discovery of Cowan and Reines using neutrinos this is not a focus of our work. Instead, we ask what from a reactor in 1956 [1]. Since then, reactor measure- the best possible reactor neutrino experiment would look ments have been a mainstay in the quest to understand like to either find a light sterile neutrino or to exclude neutrino properties. KamLAND [2], Double Chooz [3], a sizable mixing. We also focus exclusively on electron Daya Bay [4] and RENO [5] have played central roles in (anti)neutrinos and their potential mixing with a sterile establishing the oscillation of the three active neutrino neutrino. This choice is motivated, to some degree, by flavors. In the near future, the JUNO experiment will the fact that all data on νe ! νe andν ¯e ! ν¯e at this point provide the best measurement of the so-called solar pa- are mutually consistent. The same cannot be said of the 2 νµ ! νe and νµ ! νe data sets; additionally, the global rameters θ12 and ∆m21 as well as determine the mass ordering [6]. In the run-up to the 2011/2012 measure- disappearance data are known to be inconsistent with the global appearance data when interpreted in the con- ment of θ13, the question of how to predict the reactor antineutrino spectrum received renewed attention [7,8] text of a truly sterile fourth neutrino [11, 14, 18, 19]. On and the surprising result was an upward correction to top of all of this, an eV-scale sterile neutrino of the sort the predicted inverse beta decay (IBD) event rate by ap- indicated by terrestrial oscillation anomalies is severely proximately 6%. This, in turn, gave rise to the so-called constrained by cosmological observations [20{23]. Our reactor antineutrino anomaly (RAA) [9], which would be hope is that a careful study of neutrino disappearance at reactors will facilitate future analyses of the global neu- naturally explained if theν ¯e mixed with an additional species of neutrino with a mixing angle sin2 2θ ' 0:1 and trino data set. 2 ∆m2 ≥ 1 eV2. For a review of the status of the field at The reason nuclear reactors are the electron neutrino that time, see Ref. [10]. Enormous experimental progress source of choice is threefold: has been made since then and we refer the reader to the 1. They are free, in the sense that they are constructed global fitting literature for details, see, e.g., Ref. [11{13]; for purposes other than neutrino physics. for a more experiment-centered review, see Ref. [14]. As far as reactor neutrino experiments are concerned, 2. They are very bright, with experimentally accessi- ble fluxes of up to 1013 cm−2 s−1. arXiv:2104.00005v1 [hep-ph] 31 Mar 2021 modern experiments (i.e., those conducted after 2011) all rely on a comparison of measured event rate spec- 3. They produce antineutrinos which can be cleanly tra at different baselines and are thus independent of detected via IBD, which both provides a flavor tag flux predictions. Nonetheless, reactor flux predictions and has a large3 cross section of approximately 6 × have been subject to intense study (for a summary, see 10−43 cm2 per fission. Refs. [12, 15]). Moreover, the rate anomaly has not, as yet, been resolved, though it may be less significant than originally suggested.1 The spectral anomaly known as sistent solution to the RAA by shifting the ratio of 239Pu to 235U in the integrated beta spectrum. In this case, beta and neutrino data would be in good agreement with both summation and con- version calculations of the reactor flux. ∗ jeff[email protected] 2 This paper deals only with electron flavor neutrinos and/or an- y [email protected] tineutrinos and we rely on context to disambiguate these two z [email protected] cases. 1 Recent work [16], if confirmed, would provide a simple and con- 3 relative to other weak-interaction processes. 2 Other sources of electron neutrinos which have been allowed baseline partially offsets the lower absolute rate considered for sterile neutrino searches include radioac- of antineutrino production. We consider for each of these tive sources [24, 25], beta-beams [26], kaon beams [27], scenarios in turn. beam-driven beta-decay sources [28] and stored muon beams [29]. None of these other electron neutrino sources shares all of the advantages of nuclear reactors | in par- II.1. Methodology ticular they all would be quite costly and require de- tectors much larger than what will be considered here. We consider pseudo-experiments in which the spec- On the other hand, some of these approaches like beam- trum of reactor antineutrinos is observed at two base- driven beta-decay sources and stored muon beams would lines. We treat the detector as point-like, but the reactor yield far better sensitivities than possible at a reactor. core has finite physical extent; we treat the latter as a But the question remains: what sensitivities are a pos- perfect sphere. Because of this, we cannot use the os- sible with a reactor source? This is the subject of the cillation probability in Eq. (II.1) as is | we must flux- present study. average Pee over the production region. The effect of this is to average out high-frequency oscillations: if the oscil- II. GREEN-FIELD STUDIES lation wavelength is comparable to or smaller than the size of the production region, then the experiment will have muted sensitivity to the associated value of ∆m2. For the neutrino energies available at nuclear reactors Details on this procedure and the resulting average sur- and for the baselines that we will consider (less than vival probability are given in AppendixA. 100 m), oscillations among the three neutrinos of the The analysis window is taken to be E ⊂ [2:0; 8:0] Standard Model (SM) are negligible. Therefore, if a ster- prompt MeV; we generally assume a bin size of 250 keV, but ile neutrino4 exists and participates in oscillations, then occasionally consider finer spacing. We assume a con- any oscillations observed over these distances would be stant, nonzero background in each energy bin, ignoring attributable to the new state. We may describe these os- any systematic uncertainty on this background and as- cillations in the two-flavor limit, writing the ν survival e suming that the near and far baselines experience the probability as same background rate per unit exposure. We consider 2 two energy response models as limiting cases of the de- 2 2 ∆m L Pee = 1 − sin 2θ sin : (II.1) tector resolution. The first is the response model of 4E ν the PROSPECT experiment, which the collaboration Here, L is the baseline over which the neutrinos propa- has provided in the Supplemental Material to Ref. [30]. 2 The response may be roughly described as a Gaussian gate and Eν is their energy. The parameter ∆m is the p difference in the squares of the masses of two neutrino with width ∼ 7%= E= [MeV], though it includes a non- mass eigenstates; sin2 2θ is the mixing angle describing Gaussian tail down to low energies. On the other end the amplitude of the oscillations. Probing sterile neutri- of the spectrum is the upcoming Taishan Antineutrino 2 2 Observatory (TAO) experiment [31], for which the reso- nos amounts to determining sin 2θ and ∆m . p The overarching purpose of this study is to determine lution will be of the order ∼ 1%= E [MeV]. We assume the optimal configuration(s) for a two-baseline reactor that the TAO detector response is perfectly Gaussian. neutrino experiment with realistic operational assump- For a given experimental configuration, specified by tions. We categorize these by the class of reactor at the near and far baselines (Lnear and Lfar, respectively), which they are conducted | at either a commercial or a we form the following χ2 function: research reactor facility. The salient difference between 2 Nbins N N 2 2 ! these is that the former is much larger (R ∼ 2 m) than X M − φiP sin 2θ; ∆m ; fηjg χ2 = i i the latter (R ∼ 20 cm); commercial cores are also more σN powerful and produce a larger flux of antineutrinos.

Load more

  13

Copy Link