FTPI-MINN-20-28

Constraints on Decaying Sterile from Solar Antineutrinos

1, 2, 3, 1, 2 Matheus Hostert ∗ and Maxim Pospelov 1School of Physics and Astronomy, University of Minnesota, Minneapolis, MN 55455, USA 2William I. Fine Theoretical Physics Institute, School of Physics and Astronomy, University of Minnesota, Minneapolis, MN 55455, USA 3Perimeter Institute for Theoretical Physics, Waterloo, ON N2J 2W9, Canada (Dated: September 9, 2021) Solar experiments are highly sensitive to sources of ν ν conversions in the 8B neutrino flux. In this work we adapt these searches to non-minimal sterile→ neutrino models recently proposed to explain the LSND, MiniBooNE, and reactor anomalies. The production of such sterile neutrinos in the Sun, followed the decay chain ν4 νϕ ννν with a new scalar ϕ results in upper limits 2 → → for the neutrino mixing Ue4 at the per mille level. We conclude that a simultaneous explanations of all anomalies is in tension| | with KamLAND, Super-Kamiokande, and Borexino constraints on the flux of solar antineutrinos. We then present other minimal models that violate parity or lepton number, and discuss the applicability of our constraints in each case. Future improvements can be expected from existing Borexino data as well as from future searches at Super-Kamiokande with added Gd.

I. INTRODUCTION The flux of antineutrinos from the Sun at the MeV energies is negligible [18], which remains an excellent ap- Beyond neutrino mixing, the study of solar neutrinos proximation down to tens of keV in energy [19]. Com- provides important input to the standard solar Model bined with the fact that the detection cross section for (SSM) [1,2] and has been used to search for several phe- νe is much larger and easier to measure compared to nomena beyond the SM of particle physics. One example that of νe, this makes experiments sen- is neutrino decay, originally proposed as an alternative sitive to very small fluxes of antineutrinos [20–22]. The solution to the solar neutrino problem [3,4]. Indeed, current sensitivity reaches fluxes as small as a few times 5 8 after precision measurements of the solar neutrino oscil- 10− of the B neutrino flux [23–28]. These searches lation parameters by KamLAND [5], strong constraints have been discussed in the context of new physics, such as large ν ν oscillations. This Lepton number (LN) on the lifetimes of ν2 and ν3 have been obtained [6,7] as → data is consistent with no additional neutrino disappear- violating process is rather small in most theories, be- 2 ance. ing suppressed by (mν /E) , but can be enhanced due Recently, non-minimal neutrino decay models have re- to spin-flavor precession [29, 30]. The latter arises from ceived interest in the literature, where exotic decays of the coupling of a large neutrino magnetic moment to the solar magnetic field, which induces ν ν conversions, a relatively heavy (mexotic mactive) and mostly-sterile e → x neutrino are invoked to explain longstanding experimen- followed by flavor transitions into νe due to matter ef- tal anomalies at short baselines (SBL). One category of fects. Another possibility to generate such LN violating models concerns “visible” decay, where signatures is neutrino decay. For instance, neutrino mass new sterile states are produced and decay back to visible models where LN is a spontaneously broken global sym- active neutrinos. Originally proposed in Ref. [8] as an metry predict the existence of a pseudo-goldstone boson explanation to the Liquid Scintillator J, the majoron [31, 32]. In these models, solar antineu- trinos may be produced from the decay ν ν J, which (LSND) anomaly [9, 10], this scenario has now been re- 2 → 1 visited [11, 12] in light of recent data of short-baseline is enhanced in dense matter [33]. This possibility of pro- duction from neutrino decay is, in fact, quite general and arXiv:2008.11851v2 [hep-ph] 7 Sep 2021 ν ν and ν ν appearance at MiniBooNE [13– µ → e µ → e 15], as well as νe disappearance at reactors [16, 17]. can be realized in any LN violating model with neutri- Due to the small mixing angles required in this expla- nos that decay sufficiently fast, be they ν2, ν3, or the nation, the effects of attenuation in solar neutrino fluxes new mostly-sterile state ν4 discussed here (see Ref. [34] is small. Yet, the total number of heavy neutrinos pro- for an early discussion in the context of a 17 keV sterile duced is large, and if these states undergo sufficiently neutrino). distinctive decays or scattering inside a detector, they In this work, we explore a new possibility where lep- can be searched for. In this article, we point out that if ton number can, in fact, be conserved but the decay of a antineutrinos are produced in the decay of these heavy new light boson leads to a large flux of antineutrinos. We neutrinos, then they are strongly constrained by exist- derive limits on the electron flavor mixing with ν4, work- ing searches for neutrino-antineutrino transitions in solar ing only with the gauge-invariant and parity-conserving neutrino experiments. model of Ref. [12]. Focusing solely on Dirac neutrinos, we show that our bounds exclude virtually all of the parameter space preferred that can simultaneously ex- ∗ [email protected] plain LSND and MiniBooNE, as well as the region of 2

2 2 such new sterile states with eV masses are in strong ten- m4 = 300 eV, mφ/m4 = 0.9, Ue4 = 1 10− | | × 40 10− sion with cosmological observations, which has prompted

νe 6 dΦ (8B) several studies to resolve this by invoking secret interac- 10 dEν σIBD tions [41–45]. ] 41

1 10−

− Visible sterile neutrino decays are, therefore, a natural 5 10 “next-to-minimal” explanation to SBL anomalies to con-

MeV 42 10− ] 2

1 sider. The advantages of this scenario are that it does 4 −

s 10 not necessarily lead to strong correlations between ap- [cm

2 σν e e− − 43 σ pearance and disappearance channels, the mass scale of 10− 103 [cm the new sterile state is not fixed by the oscillation length σνe e ν − of the experiments, and that it already contains a secret E

dΦ 44 d 102 10− interaction mechanism, possibly alleviating tension with

νe dΦ (decay) cosmology. In this work, we focus on the decay of steriles dEν 1 45 10 10− with eV to hundreds of keV masses to a new scalar ϕ, as 0 5 10 15 discussed in Refs. [8, 11, 12, 46, 47]. In all such visible Eν [MeV] decay scenarios, heavy neutrinos decay to mostly-active neutrinos via ν νϕ , where more neutrinos can be pro- 4 → FIG. 1. The solar neutrino energy spectrum from 8B (shaded duced from ϕ νν decay if ϕ is massive as in Ref. [12]. orange) and resulting antineutrino spectrum from the decays Here visible refers→ to the detectability of the decay prod- of ν4 (shaded grey). We also show the inverse beta decay ucts, in contrast to models where neutrinos decay to the (IBD) and neutrino-electron scattering cross sections on an wrong-helicity states that do not feel the weak interac- overlaid axis. tions (up to tiny helicity-flipping terms proportional to 2 2 mν /Eν ). Such visible decays can explain the anomalous νe-like interest for reactor anomalies. They also disfavor most events at SBL experiments by means of a sub-dominant but not all parameter space suggested as a solution to population of ν4 states in neutrino beams, which often the MiniBooNE anomaly. For models with Majorana 1 decays to νe-like daughters . One typical prediction is neutrinos, the constraints become even stronger due to that the spectrum of daughter νe and νe neutrinos is ν4 νϕ ννν decays. → → softer than the initial flux of ν4 parents and associated The paper is organized as follows. In SectionII we neutrinos, skewing the effective flavor conversion towards review the benchmark model for decaying sterile neutri- lower energies. While this brings a mild improvement nos, and in SectionIII we discuss generic aspects of solar over the oscillation fit to the low energy excess observed antineutrino searches. The resulting constraints, future at MiniBooNE, it leads to less satisfactory energy spectra prospects, and alternative search methods are then dis- at LSND, which is compatible with a signal that grows cussed in SectionIV. We dedicate SectionV to a survey in energy. In addition, the neutrino flux at LSND comes of minimal alternative models for decaying steriles, and from both π+ and µ+ decay at rest, yielding a large and conclude in SectionVI. monochromatic νµ flux, and a spectrum of νµ and νe. Since only the νe component is detected via the IBD pro- cess, the presence of a neutrino-to-antineutrino transition II. DECAYING STERILE NEUTRINO in the decay chain can convert the large νµ flux to signal, since ν4 states can be produced in pion as well as muon decays. This is a crucial point in the study of Ref. [12], The most significant deviations from the three- which found improved compatibility between LSND and neutrino paradigm at SBLs are the LSND excess of ν e MiniBooNE regions of preference when this conversion is events, with a statistical significance of 3.8 σ when inter- significant. preted under a ν oscillation hypothesis, and the Mini- µ For concreteness, we focus on the gauge-invariant and BooNE excess of ν -like events, with a significance of e parity-conserving model of Ref. [12], wherein a SM singlet 4.8 σ when interpreted under a ν ν and ν ν os- µ e µ e ν is introduced and equipped with sizable couplings to a cillation hypothesis. Reactors at very→ short-baselines→ also s new scalar singlet ϕ. The sterile neutrino can then couple have some evidence of ν disappearance [16, 17], but in e to light and mostly-active neutrinos in a gauge-invariant that case the neutrino flux predictions are highly uncer- tain and harder to control [35, 36]. Despite the large sig- nificance of these anomalies, they remain unsolved. Their standard interpretation under oscillations of a eV-scale 1 sterile neutrino leads to strong tensions between differ- Constraints on this scenario have been obtained in Ref. [48] using the near detector of NOνA and T2K, as well as MINERνA and ent data. This is driven mainly by the absence of anoma- PS-191. We note that the constraints have been obtained under lous results in νµ disappearance experiments [37, 38], as simplified assumptions, and that a detailed study with total sig- appearance and disappearance channels are strongly cor- nal efficiency, as well as appropriate uncertainties is needed in related in the oscillation scenario [39, 40]. In addition, order to derive reliable constraints. 3

2 3 pair, giving rise to our signal. Overall, the process of m4 = 300 eV, Uµ4 = 1 10− | | × interest is 105 ν (E ) ν(E ) + ϕ(E ) (2) ] U 4 ν4 1 ϕ e4 2 1 4 | 10 | = 10 → − −2 &ν(E2) + ν(E3).

MeV 3 U 10 | e4 2 1 = 10 | 3 For most cases of interest, E m , so if ν (ν ) is − − ν4 4 4 4 s  2 2 Borexino produced via weak interactions, it will be left-handed − 10 (right-handed) polarized to a very good approximation.

[cm KamLAND 11 We then assume all heavy neutrinos to be polarized with ν 1

E 10 dΦ

d mφ/m4 = 0.9 KamLAND 21 a definite helicity h = 1 for neutrinos and h = +1 4 − 4 mφ/m4 = 0.5 SK-IV for antineutrinos. Nevertheless, due to the assumption of 0 10 mφ/m4 = 0.1 parity conservation for the ϕ interactions with neutrinos, both helicity flipping (HF) and helicity conserving (HC) 0 5 10 15 20 decay channels are allowed. Assuming all neutrinos to E [MeV] ν be ultra-relativistic, we find the squared-amplitudes for h4= 1 h polarized ν4 − ν ϕ decay, FIG. 2. The experimental limits on solar νe at 90% C.L. → as a function of the neutrino energy. The shaded regions 2 3 are excluded by Borexino (blue) [25], KamLAND 2011 [24] 2 2 2 E1 Mν− ν ϕ = gϕ Usi m4 , (3) (pink), KamLAND 2021 [28] (grey), and SuperK-IV 2020 [27] 4 → − | | E i=1 ν (yellow). The different new physics predictions are also shown X 2 3 3 as solid curves assuming Uτ4 = 0 and Uµ4 = 10− . 2 E | | | | 2 2 2 2 1 Mν− ν+ϕ = gϕ Usi m4 (1 rϕ) (4) 4 → | | − − Eν i=1   X fashion by means of mixing between the heaviest neutrino where rϕ = mϕ/m4. Integrated over phase space, both state, ν4, and the active flavors. The relevant Lagrangian is given by channels contribute identically to a total decay rate of

2 2 LAB gϕ m4 2 2 2 L = gϕνsνsϕ + mαβνανβ, (1) Γ (ν νϕ) = U U (1 r ) . (5) − 4 → 16π E | s4 si| − ϕ α,β i ν X X where the neutrino mass mechanism is left unspecified Our decay rate is in agreement with Refs [12, 50]. Note and assumed to not play a role in the low-energy phe- that helicity conserving decays prefer larger E1 values, nomenology. In the mass basis, the neutrino mass eigen- while helicity flipping decays prefer smaller values of E1. states are given by νi = Uαi∗ να, with α e, µ, τ, s , Therefore, for our present application, helicity-flipping α ∈ { } and U a unitary mixing matrix. Under the assumption decays are important since the antineutrinos from the P of parity conservation in the sterile sector, U is identi- subsequent scalar decay tend to be more energetic. Also cal to the extended Pontecorvo–Maki–Nakagawa–Sakata important is the limit r 1, where the scalar particle has (PMNS) matrix, now 4 4. We return to this issue in Sec- most of the ν energy regardless→ of the helicity structure × 4 tionV. In the decays of ν4 and ϕ, only the three lightest of the decay. This is the scenario with the most energetic mass states are produced, and so the it is useful to define antineutrinos in the final state, for which a simultaneous 3 the low-energy flavor stateν ˆs = i=1 Uαiνi. For most explanation of MiniBooNE and LSND is most successful. processes of interest, however, the non-unitarity correc- The scalar decay length in the lab frame to leading P tions introduced by working withν ˆs instead of the full order in the small mixing elements is flavor states ν is small and appears only at order U 4. s | α4| Unless stated otherwise, we refer toν ˆs as simply ν from 3 g2 m2 now on, as the mass eigenstates have decohered on their ΓLAB(ϕ νν) = ϕ ϕ U U 2. (6) → 8π E | si sj| way from Sun. The new scalar does not couple directly i,j=1 ϕ to the SM, and loop-induced couplings will ultimately X depend on the UV completion of the model and its neu- As expected, the scalar decays are doubly suppressed by trino mass mechanism (see, for instance, Refs. [31, 49]). small mixing elements, and so it tends to decay more slowly than ν4. Nevertheless, the decay of both parti- Due to mixing, heavy neutrinos with masses below the cles can be considered prompt within astrophysical ob- MeV scale would be produced in the Sun via the same jects. Finally, note that only due to parity conserving processes responsible for ν production at a rate U 2 nature of the scalar interaction, both left- and right- e | e4| times smaller. Once produced, the ν4 mass eigenstates handed antineutrinos are produced. In this case, only immediately decay to a light neutrinos and the scalar the right-handed antineutrinos (ν+) are relevant for de- boson. The scalar then decays to a neutrino-antineutrino tection through weak interactions. 4

III. SOLAR ANTINEUTRINOS In addition to these, SNO has also set limits at the SNO 3 level of Pν ν (Eν [4, 14.8] MeV) < 8.3 10− [23] e→ e ∈ × The flux of MeV antineutrinos from the Sun in the SSM at 90% C.L. All limits quoted above assume a total 8B 6 2 1 is negligibly small. The largest antineutrino flux at MeV flux of 5.88 10 cm− s− , except Borexino which as- × 6 2 1 energies comes from small fractions of long-lived radioac- sumes 5.46 10 cm− s− , and KamLAND which as- × 6 2 1 tive isotopes in the Sun, namely 232Th, 238U, and mainly sumes 5.94 10 cm− s− . At the lowest energies, a × 40K. This give rise to an antineutrino flux on Earth of bound can also be obtained by noting that the number about 200 cm 2 s 1 with E 3 MeV [18]. This com- of elastic ν e scattering events in solar neutrino experi- − − ν . − ponent, however, is still 6 orders of magnitude smaller ments decreases if too many ν4 states are produced, both than the geoneutrino flux at the surface of the Earth at due to lower νe e cross sections and suppressed νe flux. − these energies, and can be safely neglected. At larger These effects, however, are insensitive to variations of the energies, photo-fission reactions produce an even smaller total νe flux below the tens percent level. The predictions 3 2 1 flux of antineutrinos of about 10− cm− s− [18]. It from the sterile neutrino decay model are compared with is only down at the much lower energies of tens of keV the 8B flux in Fig.1. The independent bounds quoted that antineutrinos start being produced in thermal reac- by KamLAND, Borexino, and SK-IV are shown in Fig.2 tions at a similar rate to neutrinos with fluxes as large as a function of Eν . 9 2 1 as 10 cm− s− [19]. The strength of the limits above is mostly due to the Existing limits on the flux of solar antineutrinos are large cross section for Inverse beta decay (IBD) on free usually quoted in terms of an energy-independent proba- protons at MeV electron-antineutrino energies. Beyond 8 bility Pνe νe of conversion of B neutrinos into antineu- dominating over the neutrino-electron elastic scattering trinos. The→ most stringent limits were obtained by Kam- cross section by about two orders of magnitude (see LAND in 2011 [24] Fig.1), this channel has a distinct signature that drasti- cally reduces backgrounds. After produced, the positron KamLAND 2011 5 Pν ν − (Eν 8.3 MeV) < 5.3 10− , (7) annihilates and the final state neutron is quickly captured e→ e ≥ × by the free protons. This results in a double-bang signal which was recently improved in 2021 [28], with a positron kinetic energy Te Eν 1.8 MeV, and a delayed emission of a 2.2 MeV' gamma.− The cross sec- KamLAND 2021 5 ≈ Pν ν − (Eν 8.3 MeV) < 3.5 10− , (8) tion for this process is well understood at high [54] and e→ e ≥ × low [55] energies, and relatively simple formulae that are and by Borexino in 2019 [25] valid for all energy regimes have been derived by Ref. [56]. Borexino 5 In this work we implement the latter calculation, which Pν ν (Eν 1.8 MeV) < 7.2 10− , (9) e→ e ≥ × is provided as machine-friendly data files by Ref. [57]. all at 90% C.L. In addition, SuperKamiokande (SK) a. Backgrounds For Borexino, reactor neutrinos represent the largest source of backgrounds, but are ef- has derived limits on extraterrestrial νe sources during phases I, II and III [51], but the high energy thresh- fectively constrained by DayaBay measurements. At- mospheric neutrino events with genuine IBD scattering olds of Eν > 17.3 MeV make them irrelevant for the study of 8B neutrinos. For SK phase IV (SK-IV), im- or otherwise inherit large uncertainties from the atmo- provements to the trigger system were implemented and spheric neutrino flux and cross sections, but represent the detection of neutron capture on Hydrogen was made only a small contribution (6.5 3.2 events). Finally, the 238U and 232Th geoneutrino fluxes± are energetic enough possible, lowering thresholds to Eν > 13.3 MeV [52]. The constraint on solar antineutrino flux was found to to contribute to the IBD sample, but are only signifi- SK IV 2013 4 be P − − < 4.6 10 . Recently, further im- cant up to 3.2 MeV. Borexino omits the contribution of νe νe − provements→ to the neutron× tagging algorithm lowered this geoneutrinos from the Earth’s mantle in their estimation, which is conservative. This component is the most likely value to Eν > 8.3 MeV [27], and using the data 2008 to 2018 the limit was improved to explanation for the 2σ excess seen in the lowest energy bin [25] (see also their≈ latest geoneutrino analysis [58]). SK IV 2020 4 P − − (E 8.3 MeV) < 3.6 10− . (10) The reactor neutrino flux at KamLAND is dominant νe νe ν → ≥ × below 8.3 MeV, but contributes only about 2.2 events A previous preliminary result was shown in Ref. [26] and above that value. Due to the smaller overburden at Kam- an even more recent update was presented in Ref. [53]. LAND and SK, they suffer from larger spallation back- Loading of Gd in the SK water tank is expected to greatly grounds, coming mainly from radioactive decays of 9Li. improve the neutron tagging efficiency, and would al- After muon tagging and fiducial volume cuts, these are low for much more stringent limits. With projections reduced to less than 5 events at both locations. The large on the signal selection efficiency and background reduc- number of neutrino-electron scattering events presents a SK IV Gd tion, Ref. [27] finds that a limit of P − − (E νe νe ν background for SK. For this reason, a cut is applied re- 5 → ≥ 8.3 MeV) . 2.2 10− could be achieved with 0.2% Gd quiring small shower angles with respect to the direction loading [27]. Lowering× the energy threshold of the trigger of the Sun, cos θ < 0.9. This does not impact IBD could further improve these projections. events as the positron angle with respect to the incom- 5

2 2 2 2 m4 = 300 eV, mφ/m4 = 0.10, Ue4 = 1 10− m4 = 300 eV, mφ/m4 = 0.90, Ue4 = 1 10− | | × | | ×

25 ν4 νeνeνe (2.0 events) 30 ν4 νeνeνe (9.6 events) → → reactors reactors atm atm 20 25 atm atm data 20 data 15 15 10 Events/MeV Events/MeV SK-IV 10 SK-IV 5 5

0 0 9 10 11 12 13 14 15 16 9 10 11 12 13 14 15 16 Eν /MeV Eν /MeV

FIG. 3. The inverse beta decay spectrum at Borexino (top row), KamLAND 2011 (middle row), and SK-IV 2020 (bottom row), all as stacked histograms. The hatched histograms show the background estimations by the collaboration, and the filled histogram (grey) our prediction of visible neutrino decays. All plots assume Uµ = Uτ4 = 0. | 4 | | | ing neutrino is significantly larger ( cos θ 0) than in where stands for total exposure of the experiment, the predominantly forward process ofh elastici ≈ scattering. P N is the flavour transition probability for Solar νˆs νe The observed event spectra and background predictions antineutrinos→ averaged over the radius of the Sun (see by the respective collaborations are shown in Fig.3. AppendixA), and

dΦν4 dΦνe = U 2 , (12) A. IBD Rates from Decays e4 dEν4 | | dEν dec dΓ dP 1 dΓν4 νˆsϕ ϕ νˆsνˆs The largest observable flux from sterile neutrinos = → → . dE1dE3 Γν4 Γϕ dE1 dE3 would come from ν4 states produced via weak interac- tions in the decay of 8B. The number of IBD events in a Note that Eq. (11) is the analogue of Eq. (9) from given experiment can be computed as Ref. [12], and is simpler since we work with very long

ν4 dec baselines and under the assumption that the number of dNevents dΦ dP 8 = σ(E ) P (E ), initial νµ states is negligible. For the flux of B neutri- 3 νˆs νe 3 dE dE dE N dE dE dE → νe ν4 1 3 ν4 1 3 nos, dΦ /dEν , we implement the high-metallicity fluxes (11) in the SSM AGSS09-B16 [59], where the total 8B neu- 6

2 2 m4Γ4 = 1 eV , m4 = 300 eV, mφ/m4 = 0.9 m4Γ4 = 10 eV , m4 = 100 eV, mφ/m4 = 0.9

0.012 β 0.020

decay KamLAND 2021 KamLAND 2021 KamLAND 2011 KamLAND 2011 0.010 SuperK-IV SuperK-IV 99% C.L. Borexino 0.015 Borexino excluded 0.008 99% C.L. OPERA

2 exclusion 2 | | LSND 4 4

µ 0.006 µ 0.010

U LSND U | | OPERA KARMEN β KARMEN

0.004 decay 0.005

0.002 All w/o cosmo MiniBooNE Reactors preferred All w/o cosmo MiniBooNE 0.000 0.000 Reactors preferred 0.00 0.01 0.02 0.03 0.04 0.00 0.01 0.02 0.03 0.04 0.05 U 2 U 2 | e4| | e4|

FIG. 4. Limits from solar antineutrino searches on the active-heavy neutrino mixing at 99% C.L. The regions required to explain the short-baseline anomalies in a Dirac sterile neutrino decay model are also shown (99% C.L.). For reactors, a preferred interval 2 2 2 2 in Ue4 is shown and is independent of Uµ4 . On the left we show the m4Γ4 = 1 eV case and on the right m4Γ4 = 10 eV , | | | | where the ν4 is shorter-lived. Our bounds are the same for the two cases.

2 With the predicted number of IBD events at each solar m4Γ4 = 1 eV , m4 = 100 eV, mφ/m4 = 0.5 0.020 neutrino experiment, we implement our statistical test KamLAND 2021 (described in detail in AppendixB) to place limits on KamLAND 2011 the active-heavy mixing angles. Our χ2 test statistic 99% C.L. SuperK-IV models solar neutrino flux and experimental backgrounds 0.015 excluded Borexino uncertainties through bin-uncorrelated nuisance param- eters with Gaussian errors. Both flux and background uncertainties are fixed at 10%, except for the SK-IV, for 2 | LSND

4 which we inflate those to 20%. For the KamLAND 2021

µ 0.010

U dataset, we combine the data into 2-MeV-wide bins when | OPERA performing our fit. We have also performed a total rate fit β -decay KARMEN for each of the energy bins to obtain model-independent 0.005 limits on the solar antineutrino flux. We find similar to those provided by the collaborations, within 50%. Our All w/o limits are always weaker, and therefore more conserva- LSND MiniBooNE tive, when compared to the official ones. For SK-IV, 0.000 Reactors preferred no model-independent limit was shown in the final arti- 0.00 0.02 0.04 0.06 0.08 0.10 cle [27], so we show our own result in Fig.2. U 2 | e4|

FIG. 5. Same as Fig.4, but for mϕ/m4 = 0.5. In this regime IV. RESULTS LSND and MiniBooNE are harder to combine due to softer ν spectrum obtained via electron mixing. We show only the e We plot our 90% C.L. limits in Fig.4 as a function of combination of MiniBooNE with other datasets, excluding U 2 and U 2 for m /m = 0.9. On the same axes, LSND, as reported in Ref. [12]. The solar antineutrino spec- | e4| | µ4| ϕ 4 trum is also softer, but can effectively constrain the preferred we show the preferred (grey-shaded) regions obtained in combined region at 99% C.L. Ref. [12] to explain LSND and MiniBooNE individually, as well as the combined fit to MiniBooNE, LSND and global data (except reactors and cosmology) as “All w/o cosmo”. Weaker constraints from the OPERA [60] and 6 2 1 trino flux is 5.46 10 cm− s− . For low-metallicity KARMEN [61] neutrino experiments, as well as beta de- models, our constraints× on the new physics coupling are cay kink searches are shown as dashed grey lines. We weakened by about 20%. pick two particular cases with the shortest ν4 and ϕ life- 7

U 2 = U 2 = 0.00 U 2 = 0.01, U 2 = 0 | µ4| | τ4| | µ4| | τ4| 99% C.L. excluded 99% C.L. excluded

10 2 10 2

2 − 2 − | | 4 4 e e U U | |

KamLAND 2021 KamLAND 2021 KamLAND 2011 KamLAND 2011 SuperK-IV SuperK-IV Borexino Borexino 3 3 10− 10− 0.2 0.4 0.6 0.8 1.0 0.2 0.4 0.6 0.8 1.0 R = mφ/m4 R = mφ/m4

FIG. 6. Limits obtained in this work in the plane of active-heavy mixing versus the ratio R = mϕ/m4, at 99% C.L. On the left we show the case of m4 = 300 eV, but we note that our curves are mostly insensitive to the overall scale of m4 and mϕ in the region of interest for short-baseline anomalies.

times to compare against our limits, corresponding to parent particle ν4. These limits are independent of the 2 2 m4Γ4 = 1 eV and m4Γ4 = 10 eV . These lifetimes absolute scale of mϕ and m4, provided these masses are are achieved for couplings close to the perturbativity below the Q-values of the 8B decays and above the light 2 2 2 2 2 limit, namely gϕ = (1.5) 10− /( Ue4 + Uµ4 ) and neutrino masses. 2 2 2 2 × 2 | | | | gϕ = (12) 10− /( Ue4 + Uµ4 ), respectively. It is Accounting for perturbativity bounds on gϕ and the clear that an× explanation| | of| LSND| is in large tension baselines of LSND and MiniBooNE, a lower bound on with solar antineutrino searches for all three experiments m4 of 100 eV can be obtained for neutrino mixing ≈ 2 3 we consider. It should also be noted that the region with of the order of Ue4 10− . Below this value, ν4 is 2 | | ≈ large Uµ4 which is not excluded by our curves is ex- too long-lived and does not lead to interesting signatures cluded| by| MiniBooNE itself. As expected, KamLAND at short-baseline experiments. The preferred regions for leads to the strongest bounds despite its large neutrino MiniBooNE and LSND shift to larger mixing angles when energy thresholds 8.3 MeV. Borexino and SK-IV bounds either ν4 or ϕ are longer-lived, while our constraints re- are competitive, with the latter performing better for main unaffected. The parameters used in Figs.4 and5 harder antineutrino spectra. are chosen so as to minimize the decay length of new par- In Fig.5, we show our constraints for the case ticles. Bounds from kink searches in beta decay become the strongest above m 5 keV, and peak searches in mϕ/m4 = 0.5. A simultaneous explanation of Mini- 4 & BooNE and LSND is more challenging now, and only meson decay preclude an explanation of LSND with the a global fit including MiniBooNE but not LSND is avail- current model at the MeV scale. able (“All w/o LSND”). In this case, we constrain the We note that a massless scalar has also been discussed region preferred by MiniBooNE significantly. Lower val- as an explanation of the MiniBooNE and LSND anoma- ues of mϕ/m4 are even more challenging from the point of lies [11], although it is disfavored with respect to the view of explaining the SBL results, as ϕ becomes longer best-fit point of Ref. [12] at 99% C.L. Interpreting that lived and helicity-flipping decays of ν4, which lead to soft study in the present lepton-number and parity conserv- daughter spectra, become more important. ing model, one would obtain no constraint from solar an- Changing the heavy neutrino mass but keeping the ra- tineutrino searches. In that case, however, light neutrinos will also decay. We comment more on this possibility and tio R = mϕ/m4 fixed leaves our bounds unaltered for all the mass range of interest. Lowering R, on the other others inV. hand, weakens our bounds slightly due to the softer solar antineutrino spectrum, although the weakening saturates A. Future opportunities once below mϕ/m4 0.5. We show constraints on the electron mixing angle≈ in Fig.6. The strongest constraints are obtained for vanishing muon and tau mixing, and for Borexino has collected an additional (500) days of R 1, as in that case ϕ carries most of the energy of its data on top of the 2771 days already analyzedO in Ref. [25]. → 8

In addition, the improvements made by the collaboration however, that light-sterile mixing parameters governing in the latest geoneutrino analysis [58], such as enlarged light neutrino decay are related to those of SBL anoma- fiducial volume and improved background rejection, can lies in a model-dependent fashion. In principle, but not be implemented in the solar antineutrino search. Just without fine-tuning, the correlation between U , U , | e4| | µ4| in terms of exposure, this represents an improvement of Us∗4Usj , and Usi∗ Usj for i, j < 4 may be relaxed. In 40% with respect to the values we use. models| | with LN| violation| or LN charged scalars (see be- a. Synergy with DSNB Solar antineutrino searches low), provided several constraints are satisfied, solar an- will become even more stringent with upcoming efforts tineutrinos may become relevant again for massless ϕ as to detect νe events from the DSNB [62, 63]. The SK de- the light-neutrino decays ν2 ν1ϕ are open. tector is expected to detect this neutrino flux with the → addition of Gd to its detector volume [64]. The large neutron capture cross sections on Gd and the emission of V. ALTERNATIVE MODELS: VIOLATING 8 MeV gammas will help reduce backgrounds and lower PARITY AND LEPTON NUMBER the νe detection threshold to neutrino energies as low as the IBD threshold, provided Ee > 0.8 MeV [65]. The Various other possibilities for visible sterile neutrino increased detection efficiencies at lower energies, and re- decay exist, depending on the Dirac or Majorana nature duced accidental and mis-reconstructed backgrounds will of neutrinos, as well as on the parity structure of the improve on the limits we set, being limited by intrinsic sterile neutrino sector. While we only focused on the reactor and atmospheric νe backgrounds, but also by an parity conserving model discussed above, we would like exponentially rising spallation background. Future large to dedicate this section to understanding if other minimal liquid-scintillator detectors, such as the Jiangmen Under- extensions of the SM by a singlet sterile neutrino and a ground Neutrino Observatory (JUNO) [66] and the Jin- scalar are subject to our constraints. For clarity, we focus ping neutrino experiment [67], can also improve on cur- on SM extensions with a single new sterile neutrino: νs = rent constraints with an expected threshold of E 8.5 L R R ν & νs + νs in the Dirac case and ν in the Majorana case. MeV [68]. In the far future, observatories capable of ac- Our findings for the minimal models are summarized in cumulating larger numbers of DSNB events, such as the TableI. proposed detector THEIA [69], would play an important a. Dirac L(ϕ) = 0 We start with a generalization role in searching for solar antineutrinos. We note that of Eq. (1) by writing in the event of a detection of the DSNB, one could also constrain the models considered here by requiring small D 0 L R L − = gϕνs νs ϕ + h.c. (15) DSNB absorption by relic neutrinos [70, 71]. int = g (V ) (V ) ν P ν ϕ + h.c., b. Light neutrino decay Even in the parity conserv- ϕ L si∗ R sj i R j ing model discussed so far, one can avoid solar antineutri- where index summation is understood. For complex g , nos by resorting to a massless ϕ. In that case, however, ϕ this is the most generic parametrization of the vertex. the light mostly-active neutrinos will decay. For typical We implicitly diagonalized the Dirac mass matrix by parameters relevant for the SBL anomalies, this decay means of two unitary matrices V and V , defined by will happen within 1 AU, both visibly and invisibly. For L R νL = (V ) νL, νR = (V ) νR, where νL,R are the (chi- instance, consider ν ν ϕ and ν ν ϕ decays with s L si i s R si i i 2 → 1 3 → 1 ral) mass eigenstates. Note that the enlarged PMNS ma- normal ordering and m1 0. For a coupling of gϕ = 1, ≈ trix is defined as UPMNS = VL when the charged lepton we find 0 mass matrix is diagonal. Abandoning the assumption of 5 LAB 10− E2 parity conservation in the sterile sector that was made cτ2 0.9 AU 2 , (13) ≈ Us1Us2 10 MeV previously, VL = VR, one can have allow for VL = VR by | |    L 6 R 5 choosing different Yukawa couplings for ν and ν . By 10− E3 s s cτ LAB 0.03 AU . (14) breaking parity at the level of the Dirac mass matrix, it 3 ≈ U U 2 10 MeV | s1 s3|    is possible to independently tune the couplings appear- In the convention adopted by the literature, these corre- ing in the operators νiPRν4 and νiPLν4. In practice, this 0 5 0 spond to τ2 /m2 = 4.4 10− s/eV and τ3 /m3 = 1.3 allows to tune the rate for visible and invisible decays 6 × × 10− s/eV. For inverted ordering, both ν1 ν3ϕ and of ν4− neutrinos. The same is true for the decay of the ν ν ϕ decays are of the faster kind in Eq. (→14). On top scalars, which can be either visible or invisible, depending 2 → 3 of the cosmological issues with such short lifetimes (for on VL and VR. In these models, a connection to the SBL recent discussions, see Refs. [72, 73]), the largest gϕ val- anomalies through visible decays always predicts visible ues relevant for the allowed regions are already excluded solar antineutrinos provided ϕ is heavy enough to decay. by laboratory experiments, such as SNO [74], which is b. Dirac L(ϕ) = 2 One can also introduce scalars consistent with no disappearance of solar neutrinos (see carrying LN.| These| type of scalars have been usually dis- Refs. [7, 75]). Other datasets have also been discussed to cussed in the context of Majoron models, but for our constrain the lifetime of light neutrinos, including mea- current purposes, we assume no particular connection surements of the flavor ratios of cosmic neutrinos [76] to neutrino masses. We consider a model with a Dirac and of the Glashow resonance [77]. It should be noted, field νs, and a complex scalar ϕ carrying lepton number 9

+ Minimal Models Parametric Limit Polarized ν4− decay Scalar decays Spectrum of ν Expected Signals

+ + + (VL)s4(VR)si (VR)s4(VL)si ν4− ν−ϕ / ν ϕ ϕ ν−ν− / ν ν Hard Solar/SBL Dirac L(ϕ) = 0 | | ∼ | | → → + + + D 0 (VR)s4(VL)si 0 ν4− ν ϕ ϕ ν ν Soft Solar/SBL Lint− | | → → →

(VL)s4(VR)si 0 ν− ν−ϕ ϕ ν−ν− None SBL | | → 4 → → + + + gL gR ν4− ν−ϕ∗ / ν ϕ∗ ϕ∗ ν−ν− / ν ν Hard Solar/SBL Dirac L(ϕ) = 2 ∼ → → − + D 2 gR 0 ν4− ν ϕ∗ ϕ∗ ν−ν− Hard Solar/SBL Lint− → → → + + gL 0 ν− ν−ϕ∗ ϕ∗ ν ν None None → 4 → →

Majorana + + + M ν4− ν−ϕ / ν ϕ ϕ ν−ν− / ν ν Hard Solar/SBL Lint − → →

TABLE I. Minimal models for sterile neutrino decay to light neutrinos. Here ν and i stand for all light neutrinos mass eigenstates

(i < 4), and ν4− for the heavier left-handed polarized neutrino. The first two columns show the minimal model considered and the judicious choices of its parameters to achieve a certain parity structure. The third and fourth columns show the decay channels allowed in that model, separated by all possible helicity final states. The penultimate column shows the kind of visible + solar antineutrinos (ν ) energy spectrum is predicted in the model. This depends on the HF or HC nature of the ν4− decay. Note, however, that the spectrum may change in the limit mϕ/m4 1. For the minimal Majorana neutrino model, one always obtains a prediction for visible solar antineutrinos. See the main text→ for definitions.

L(ϕ) = 2. In all generality, we can write d. Simplified models Finally, we note that Refs. [8, − 11] work with simplified models, and do not specify the D 2 c c Lint− = gL(νs) PLνsϕ + gR(νs) PRνsϕ + h.c. (16) origin of the g ν ν ϕ vertex. Although this operator L ⊃ e e 4 = gL(VL)si(VL)sjνiPLνjϕ may arise from a Lagrangian as simple as Eq. (1), it can be considered more generically as a by-product of non- + gR(VR)si(VR)sjνiPRνjϕ + h.c., 2 renormalizable operators such as (LH) ϕ and (LH)νsϕ. where again we implicitly diagonalized the Dirac mass In these effective models, the active-heavy mixing neces- matrix with VL and VR. In this case, even for parity sary for ν4 production in most accelerator experiments, 2 conserving mass matrices, one can violate parity in the Uµ4 , is independent from ge, which controls the decay | | neutrino-ϕ interactions by tuning the arbitrary gL and rate of ν4 νeϕ. In this case, the only way to generate νe → + + gR couplings. In this case, ν4− states produced in the appearance at LSND is via muon decays, µ e νeν4. 2 → Sun will always decay to visible antineutrinos provided It also follows that the mixing Ue4 may be parametri- | | gR = 0, independently of the decay products of ϕ∗. For cally small, turning off ν4 production in the Sun via mix- 6 8 8 + this model, one may attempt to explain SBL anomalies ing. Four-body decays of the type B Be e ν4ϕ are 2 → 2 7 with only the decay products of scalar produced in ν− negligible as kaon decays constrain g Uµ4 < (10− ). 4 e | | O decays by setting gR 0. In that case, no visible solar If a vector particle is introduced instead, the cosmo- antineutrinos appear.→ logical history is yet even more involved. We do not c. Majorana neutrinos A final possibility is to aban- study this case here, although our solar antineutrino don LN and work with Majorana neutrinos. In this case, bounds would also apply to parity-conserving scenarios a minimal model can be built with only νR and a scalar with small modifications. Note that our constraints are ϕ. LN is violated by the νR Majorana mass term, and not relevant for fully invisible sterile neutrino decays, as the most general interaction Lagrangian in this case is invoked to relax the tension between SBL appearance and disappearance tension in Refs [38, 78]. M R c R Lint = gR(ν ) ν ϕ + h.c. (17) = gR(V )si(V )sjνiPRνjϕ + h.c., where now we implicitly diagonalized the Majorana mass VI. CONCLUSIONS matrix by means of a single unitary matrix V = UPMNS0 . In this case, all light neutrinos as well as antineutrinos Puzzling results from some of the short-baseline neu- are visible due to the reduced number of degrees of free- trino experiments will eventually find an explanation dom. Both HC and HF decays of ν4 are controlled by the with more data coming from the SBN program at Fer- same parameters, and cannot be disentangled as easily. milab [79–81] and the π+ decay-at-rest experiment at Solar antineutrinos could appear in this case if all other J-PARC, JSNS2 [82]. At the moment it is possible to constraints are satisfied. speculate that some form of new physics in the neutrino 10 sector is responsible for the deviation of LSND and Mini- effect, the degrading of average energy for the SN neu- BooNE results from theoretical expectations within the trinos, can be constrained with the observed signal of minimal three-generation neutrino model. Among such SN1987A. This has been explored to constrain neutrino speculations is the class of models where the excess of self-interactions inside supernovae by the requirement antineutrinos at LSND and excess of low-energy electron- that neutrinos carry sufficient energy to the outer layers like events at MiniBooNE is due to a prompt production of the collapsing star [83]. We point out, however, that a and decay of dark sector particles. This new sector is more general statement can be made regarding neutrino likely to comprise a heavier, mostly sterile neutrino, that energy loss outside the dense environment, which is in- can be produced via neutrino mixing in meson decays dependent of the explosion mechanism. Details of this and nuclear reactions. Such heavier neutrino can gener- effect will be addressed in a future publication. ate a cascade decay to an unstable bosonic mediator and light neutrino, giving rise to the admixture of electron antineutrinos in the flux at the end of the decay chain. ACKNOWLEDGMENTS We have shown that up to some model dependence one should expect that regular nuclear processes in the The authors would like to thank Linyan Wan and San- Sun create an antineutrino flux. Such flux is stringently dra Zavatarelli for correspondence on the SK-IV and constrained by most of the solar neutrino experiments, at Borexino experimental capabilities. We also thank Ivan 5 a O(few 10− ) level owing to a larger cross sections for Esteban and Joachim Kopp for discussions. MP is sup- × the IBD processes, and additional structure to the signal ported in part by U.S. Department of Energy (Grant No. that has been exploited to cut on backgrounds. After desc0011842). This research was supported in part by application of these constraints, our results disfavor large Perimeter Institute for Theoretical Physics. Research at part of the parameter space of the model in Ref. [12], and Perimeter Institute is supported by the Government of disfavor this mechanism as an explanation of the LSND Canada through the Department of Innovation, Science excess, while significantly narrowing possible parameter and Economic Development and by the Province of On- space for the MiniBooNE excess. The simulations used tario through the Ministry of Research, Innovation and to produce the results in this paper are publicly available Science. on github2. In general, our limits disfavor large ν ν transi- tions that could improve the combined fit of→ LSND and Appendix A: Solar Flavor Transitions MiniBooNE data. Such transitions could in principle be avoided if ϕ is lighter than the lightest neutrino state, When ϕ decays to light neutrinos, it decays into the 3 in which case, mixing angles and CP phases have to be state νˆ = U ∗ ν . The average transition prob- | si i=1 si | ii fine-tuned to avoid light neutrino decays. The alternative ability forν ˆs to exit the Sun and be detected as a νe on models with parity violation or apparent LN violation the surface ofP the Earth under the (good) approximation presented in SectionV may avoid ν ν transitions even of adiabatic flavor conversion is simply for massive ϕ, but would require a case-by-case→ study of the SBL physics and additional constraints. 3 U m 2 U 2 i=1 | si | | ei| Our constraints add to the existing list of problems of Pνˆ ν N = , (A1) h S → e i UD 2 + U 2 + U E 2 the decaying sterile neutrino solutions to the SBL puzzle. | eP4| | µ4| | τ4| Chiefly among them is cosmology and astrophysics. As is which depends on the mixing matrix elements in mat- well known, new and relatively strongly interacting states m 2 ter Usi at the production point. Here, ... denotes can be populated by the thermal processes leading to a weighted| | average over the production regionh i and the the modifications of observed quantities, such as primor- subscript N refers to taking the non-canonical normal- dial nucleosynthesis yields and/or total amount of energy ization ofν ˆs into account. We have neglected Earth density carried by neutrinos at late times. In addition, matter effects, which for antineutrinos leads to a reduc- these models are likely to cause strong modifications to tion (increase) of P2 e (P1 e) below 10%, and as- the supernovae neutrino spectrum. One reason for such sumed the unitarity of→ the 4→ 4 mixing∼ matrix U. Ne- modification is the possibility of the neutrino number- glecting all CP phases, we follow× Ref. [84] and write changing processes, such as νν ννϕ νννν. Given → → U = R23R34R24R14R13R12, where Rij = R(θij) is the relatively strong couplings in the models of Refs. [11, 12], usual rotation matrix in the (i, j) plane. Note that if we the underlying cross sections are far greater than weak assume Uτ4 = c34s23s24+c23s34 = 0, then to leading or- interaction cross section, meaning that the neutrinos can − 2 2 2 2 2 der in the small angles, Ue4 θ14 and Uµ4 c23θ24, share energy and maintain their chemical equilibrium im- with the rest of the mixing| matrix| ∼ elements| of| the∼ active mediately after they leave the star. The main physical 3 3 sub-matrix retaining their usual definition. × 2 2 2 In the limit where ∆m21 ∆m31 ∆m41, the flavor evolution of solar neutrinos can be described by an ef- fective two-neutrino model, with in-matter modifications 2 ‡ github.com/mhostert/SolarDecayingSteriles. only on θ12. Up to corrections proportional to the new 11

U 2 = 0.01, U 2 = 0.0, U 2 = 0.0 U 2 = 0.01, U 2 = 0.001, U 2 = 0.0 U 2 = 0.01, U 2 = 0.01, U 2 = 0.0 | e4| | µ4| | τ4| | e4| | µ4| | τ4| | e4| | µ4| | τ4| m 2 m 2 m 2 Pνˆs νe Us2 ν Pνˆs νe Us2 ν Pνˆs νe Us2 ν 1.00 h → iN h| | iN 1.00 h → iN h| | iN 1.00 h → iN h| | iN U m 2 U m 2 U m 2 U m 2 U m 2 U m 2 h| s1 | iN h| s3 | iN h| s1 | iN h| s3 | iN h| s1 | iN h| s3 | iN 0.75 0.75 0.75

0.50 0.50 0.50 Probability Probability Probability 0.25 0.25 0.25

0.00 0.00 0.00 1 0 1 1 0 1 1 0 1 10− 10 10 10− 10 10 10− 10 10 Eν/MeV Eν/MeV Eν/MeV

m 2 ν m 2 ν m 2 ν Pνˆs νe Us2 N Pνˆs νe Us2 N Pνˆs νe Us2 N 1.00 → N h| | i 1.00 → N h| | i 1.00 → N h| | i U m 2 U m 2 U m 2 U m 2 U m 2 U m 2 h| s1 | iN h| s3 | iN h| s1 | iN h| s3 | iN h| s1 | iN h| s3 | iN 0.75 0.75 0.75

0.50 0.50 0.50 Probability Probability Probability 0.25 0.25 0.25

0.00 0.00 0.00 1 0 1 1 0 1 1 0 1 10− 10 10 10− 10 10 10− 10 10 Eν/MeV Eν/MeV Eν/MeV

FIG. 7. Full flavor transition probability as well as the relevant mixing matrix element in matter, all averaged over the 8B neutrino production region in the Sun and properly normalized. From left to right, we increase the muon-heavy mixing, showing neutrino transitions on the top row and antineutrino transitions on the bottom.

2 2 mixing angles (θ14, θ24, and θ34) as well as to θ13, the as it should be since ∆m21 ∆m31. new mixing angle in matter is  sin 2θ ∆m2 tan 2θM 12 12 , (A2) 12 ' ∆m2 cos 2θ A 12 12 − CC Appendix B: Statistical Method with ACC = 2√2Eν GF Ne(r) for neutrinos (antineutri- nos) proportional± to the electron density at the produc- tion region. Since the sterile component does not feel When deriving upper limits on the mixing angles, we neutral-current interactions, the neutral-current poten- minimize the following log-likelihood function tial also modifies the relation above. Terms proportional to A = √2E G N (r), where N (r) is the neutron NC ∓ ν F n n number density in the Sun, are small, however. This is D = 2 µ (θ,~ β~) D + D ln i (B1) because they are proportional to the new small mixing i i i ~ ~ L " − µi(θ, β)# angles, as well as due to the smaller number of neutrons i X 2 in the Sun, Nn/Ne . 1/2. In our simulation, we fol- βi,j + 2 , low Ref. [84] and keep all corrections in θ14, θ24, θ34, σ i,j i,j and θ13 in the transition probabilities. As an approxima- X tion, we assume the neutral-current potential to follow the same radial dependence as the charged-current one, where θ~ stands for the vector of physics parameters (e.g., with A /A = 1/4. NC CC − U 2), β~ the vector of nuisance parameters with individ- The full transition probabilites using Eq. (A1) in | α| the energy region of the 8B flux are shown in Fig.7 ual entries βj and associated Gaussian errors σj. As an approximation, we assume to follow a χ2 distribution for both neutrinos and antineutrinos. In the limit L U , U 0, the scalar decays produce mostly-ν when estimating our confidence intervals. | µ4| | τ4| → e states, and Pνˆs νe Pνe νe . In this case, the fla- The most important systematics for our study are the vor evolution is→ similar' to→ the standard MSW effect, uncertainties on the total 8B solar neutrino flux and to- where neutrino (antineutrinos) undergo resonant (non- tal backgrounds numbers. To be conservative, we as- resonant) adiabatic flavor conversion, in which produc- sign each energy bin two normalisation systematics, one m tion of ν2 at the center of the Sun is enhanced (sup- exclusive to the new physics prediction, modelling un- pressed). For Uµ4 , Uτ4 = 0, the situation is more com- certainties in the solar flux, and one exclusive to back- plex, but the high| | energy| | 6 behavior can be understood by grounds. All normalization systematics are assumed to taking the limit θ12 π/2 (0) in the mixing elements be uncorrelated, which is conservative, and are assigned U 2 for neutrinos (antineutrinos).→ Note that νm ν , 10% Gaussian errors. | si| 3 ∼ 3 12

hi hj Appendix C: Polarized decay rates which are also valid for νi νj ϕ decays. We have defined → 2x (1 β)(1 r2 ) To produce TableI, we computed the decay rates in C(x) = − − − ϕ , (C7) each channel explicitly. We collect all results for ν4 and 2β ϕ decays assuming massless neutrino final states in each (1 + β)(1 r2 ) 2x F (x) = − ϕ − , (C8) one of the models discussed. The total decay rate for 2β (—)hj hi νi νj ϕ can be obtained by summing each polarized matrix→ element as which apply for helicity conserving and helicity flipping channels, respectively. Note that C(x) grows while F (x) decreases monotonically with x. For scalar decays ϕ hi hj 2 → max ν ν , we compute M to find x M 2 i j h1h2 hi,hj hihj | | Γ4 = | | dx, (C1) min m4β16π 2 2 2 x P M = gϕ (VL)si(VR)sj mϕ, (C9) Z | −−| | | M 2 = g (V ) (V ) 2m2 , (C10) | ++| | ϕ R si L sj| ϕ where β = p4 /E4 is the velocity of ν4 in the laboratory with all other combinations vanishing in the limit of | max| 2 frame, and xmin = (1 β)(1 rϕ)/2. Similarly, for ϕ massless final states. h h ± − → (—) i (—) j 2. Dirac L(ϕ) = 2 case νi νj decays, | |

h1 Now, switching to Eq. (16), the amplitudes for νi h2 → max 2 νj ϕ decays are xϕ M hi,hj hihj Γϕ = | | dxϕ, (C2) min mϕ16π 2 2 2 xϕ P M = 4 gR (VR)si(VR)sj mi C(x), (C11) Z | −−| | | M 2 = 4 g (V ) (V ) 2m2C(x), (C12) | ++| | L L si L sj| i 2 2 2 M + = 4 gL (VL)si(VL)sj mi F (x), (C13) where β = p /E is the ϕ velocity in the laboratory | − | | | ϕ | ϕ| ϕ 2 2 2 frame and (x )max = (1 β )/2. M+ = 4 gR (VR)si(VR)sj mi F (x). (C14) ϕ min ± ϕ | −| | | h1 h2 The amplitudes for νi νj ϕ decays can be obtained with the substitution L→ R. For scalar decays ϕ νh1 νh2 , we compute M ↔ 2 to find → i j | h1h2 | 2 2 2 M = 4 gL (VL)si(VL)sj mϕ, (C15) | −−| | | 1. Dirac L(ϕ) = 0 case M 2 = 4 g (V ) (V ) 2m2 , (C16) | ++| | R R si R sj| ϕ where again the amplitudes for decays into antineutrinos Making use of Eq. (15) and neglecting light neutrino can be obtained by L R. hi hj ↔ masses, the amplitude squared for νi νj ϕ decays is given by → 3. Majorana case

2 2 2 In the Majorana case, one can make use of the ex- M = gϕ (VL)si(VR)sj mi C(x) (C3) | −−| | | pressions for the Dirac L(ϕ) = 0 case, keeping in mind 2 2 2 M++ = gϕ (VR)si(VL)sj m C(x), (C4) that V = V , and that an additional overall factor of | | | | i R L 2 2 2 2 should be included for the νi total decay rates and an M + = gϕ (VR)si(VL)sj mi F (x), (C5) | − | | | 2 2 2 overall factor of 2/(1 + δij) for the total ϕ νiνj decay M+ = gϕ (VL)si(VR)sj mi F (x), (C6) → | −| | | rate.

[1] J. N. Bahcall, A. M. Serenelli, and S. Basu, Astrophys. [3] J. N. Bahcall, N. Cabibbo, and A. Yahil, Phys. Rev. J. Suppl. 165, 400 (2006), arXiv:astro-ph/0511337 [astro- Lett. 28, 316 (1972), [,285(1972)]. ph]. [4] S. Pakvasa and K. Tennakone, Phys. Rev. Lett. 28, 1415 [2] M. Agostini et al. (BOREXINO), (2020), (1972). arXiv:2006.15115 [hep-ex]. 13

[5] S. Abe et al. (KamLAND), Phys. Rev. Lett. 100, 221803 [35] J. M. Berryman and P. Huber, Phys. Rev. D 101, 015008 (2008), arXiv:0801.4589 [hep-ex]. (2020), arXiv:1909.09267 [hep-ph]. [6] A. S. Joshipura, E. Masso, and S. Mohanty, Phys. Rev. [36] J. M. Berryman and P. Huber, (2020), arXiv:2005.01756 D66, 113008 (2002), arXiv:hep-ph/0203181 [hep-ph]. [hep-ph]. [7] J. F. Beacom and N. F. Bell, Phys. Rev. D65, 113009 [37] M. Dentler, A. Hern´andez-Cabezudo, J. Kopp, P. A. (2002), arXiv:hep-ph/0204111 [hep-ph]. Machado, M. Maltoni, I. Martinez-Soler, and [8] S. Palomares-Ruiz, S. Pascoli, and T. Schwetz, JHEP T. Schwetz, JHEP 08, 010 (2018), arXiv:1803.10661 [hep- 09, 048 (2005), arXiv:hep-ph/0505216 [hep-ph]. ph]. [9] C. Athanassopoulos et al. (LSND), Phys. Rev. Lett. 77, [38] A. Diaz, C. A. Arg¨uelles,G. H. Collin, J. M. Conrad, 3082 (1996), arXiv:nucl-ex/9605003 [nucl-ex]. and M. H. Shaevitz, (2019), arXiv:1906.00045 [hep-ex]. [10] A. Aguilar-Arevalo et al. (LSND), Phys. Rev. D64, [39] N. Okada and O. Yasuda, Int. J. Mod. Phys. A12, 3669 112007 (2001), arXiv:hep-ex/0104049 [hep-ex]. (1997), arXiv:hep-ph/9606411 [hep-ph]. [11] A. de Gouvˆea,O. Peres, S. Prakash, and G. Stenico, [40] S. M. Bilenky, C. Giunti, and W. Grimus, Eur. Phys. J. JHEP 07, 141 (2020), arXiv:1911.01447 [hep-ph]. C1, 247 (1998), arXiv:hep-ph/9607372 [hep-ph]. [12] M. Dentler, I. Esteban, J. Kopp, and P. Machado, Phys. [41] B. Dasgupta and J. Kopp, Phys. Rev. Lett. 112, 031803 Rev. D 101, 115013 (2020), arXiv:1911.01427 [hep-ph]. (2014), arXiv:1310.6337 [hep-ph]. [13] A. A. Aguilar-Arevalo et al. (MiniBooNE), Phys. Rev. [42] S. Hannestad, R. S. Hansen, and T. Tram, Phys. Rev. Lett. 98, 231801 (2007), arXiv:0704.1500 [hep-ex]. Lett. 112, 031802 (2014), arXiv:1310.5926 [astro-ph.CO]. [14] A. A. Aguilar-Arevalo et al. (MiniBooNE), Phys. Rev. [43] L. Vecchi, Phys. Rev. D 94, 113015 (2016), Lett. 121, 221801 (2018), arXiv:1805.12028 [hep-ex]. arXiv:1607.04161 [hep-ph]. [15] A. Aguilar-Arevalo et al. (MiniBooNE), (2020), [44] Y. Farzan, Phys. Lett. B 797, 134911 (2019), arXiv:2006.16883 [hep-ex]. arXiv:1907.04271 [hep-ph]. [16] G. Mention, M. Fechner, T. Lasserre, T. A. Mueller, [45] J. M. Cline, Phys. Lett. B 802, 135182 (2020), D. Lhuillier, M. Cribier, and A. Letourneau, Phys. Rev. arXiv:1908.02278 [hep-ph]. D83, 073006 (2011), arXiv:1101.2755 [hep-ex]. [46] Y. Bai, R. Lu, S. Lu, J. Salvado, and B. A. Stefanek, [17] M. Dentler, A. Hernandez-Cabezudo, J. Kopp, M. Mal- Phys. Rev. D93, 073004 (2016), arXiv:1512.05357 [hep- toni, and T. Schwetz, JHEP 11, 099 (2017), ph]. arXiv:1709.04294 [hep-ph]. [47] Z. Moss, M. H. Moulai, C. A. Arg¨uelles, and J. M. Con- [18] R. A. Malaney, B. S. Meyer, and M. N. Butler, Astro- rad, Phys. Rev. D97, 055017 (2018), arXiv:1711.05921 phys. J. 352, 767 (1990). [hep-ph]. [19] E. Vitagliano, J. Redondo, and G. Raffelt, JCAP 1712, [48] V. Brdar, O. Fischer, and A. Y. Smirnov, (2020), 010 (2017), arXiv:1708.02248 [hep-ph]. arXiv:2007.14411 [hep-ph]. [20] E. K. Akhmedov, Phys. Lett. B255, 84 (1991). [49] X.-J. Xu, (2020), arXiv:2007.01893 [hep-ph]. [21] R. Barbieri, G. Fiorentini, G. Mezzorani, and [50] C. W. Kim and W. P. Lam, Mod. Phys. Lett. A5, 297 M. Moretti, Phys. Lett. B259, 119 (1991). (1990). [22] A. Acker, A. Joshipura, and S. Pakvasa, Phys. Lett. [51] K. Bays et al. (Super-Kamiokande), Phys. Rev. D85, B285, 371 (1992). 052007 (2012), arXiv:1111.5031 [hep-ex]. [23] B. Aharmim et al. (SNO), Phys. Rev. D70, 093014 [52] H. Zhang et al. (Super-Kamiokande), Astropart. Phys. (2004), arXiv:hep-ex/0407029 [hep-ex]. 60, 41 (2015), arXiv:1311.3738 [hep-ex]. [24] A. Gando et al. (KamLAND), Astrophys. J. 745, 193 [53] A. Giampaolo, “The Diffuse Supernova Neutrino Back- (2012), arXiv:1105.3516 [astro-ph.HE]. ground at Super-Kamiokande,” (2021). [25] M. Agostini et al. (Borexino), (2019), arXiv:1909.02422 [54] C. H. Llewellyn Smith, Gauge Theories and Neutrino [hep-ex]. Physics, Jacob, 1978:0175, Phys. Rept. 3, 261 (1972). [26] W. Linyan, Experimental Studies on Low En- [55] P. Vogel and J. F. Beacom, Phys. Rev. D60, 053003 ergy Electron Antineutrinos and Related Physics, (1999), arXiv:hep-ph/9903554 [hep-ph]. Ph.D. thesis, Tsinghua University (2018), avail- [56] A. Strumia and F. Vissani, Phys. Lett. B564, 42 (2003), able at http://www-sk.icrr.u-tokyo.ac.jp/sk/_pdf/ arXiv:astro-ph/0302055 [astro-ph]. articles/2019/SKver-Linyan.pdf. [57] A. M. Ankowski, (2016), arXiv:1601.06169 [hep-ph]. [27] K. Abe et al. (Super-Kamiokande), (2020), [58] M. Agostini et al. (Borexino), Phys. Rev. D 101, 012009 arXiv:2012.03807 [hep-ex]. (2020), arXiv:1909.02257 [hep-ex]. [28] S. Abe et al., (2021), arXiv:2108.08527 [astro-ph.HE]. [59] N. Vinyoles, A. M. Serenelli, F. L. Villante, S. Basu, [29] C.-S. Lim and W. J. Marciano, Phys. Rev. D37, 1368 J. Bergstr¨om,M. Gonzalez-Garcia, M. Maltoni, C. Pe˜na- (1988), [,351(1987)]. Garay, and N. Song, Astrophys. J. 835, 202 (2017), [30] E. K. Akhmedov, Sov. Phys. JETP 68, 690 (1989), [Zh. arXiv:1611.09867 [astro-ph.SR]. Eksp. Teor. Fiz.95,1195(1989)]. [60] N. Agafonova et al. (OPERA), JHEP 07, 004 (2013), [31] Y. Chikashige, R. N. Mohapatra, and R. D. Peccei, Phys. [Addendum: JHEP07,085(2013)], arXiv:1303.3953 [hep- Lett. 98B, 265 (1981). ex]. [32] G. B. Gelmini and M. Roncadelli, Phys. Lett. 99B, 411 [61] B. Armbruster et al. (KARMEN), Phys. Rev. D65, (1981). 112001 (2002), arXiv:hep-ex/0203021 [hep-ex]. [33] Z. G. Berezhiani and M. I. Vysotsky, Phys. Lett. B199, [62] S. Horiuchi, J. F. Beacom, and E. Dwek, Phys. Rev. 281 (1987), [,288(1987)]. D79, 083013 (2009), arXiv:0812.3157 [astro-ph]. [34] Z. G. Berezhiani, G. Fiorentini, M. Moretti, and [63] C. Lunardini, Phys. Rev. Lett. 102, 231101 (2009), A. Rossi, Z. Phys. C54, 581 (1992). arXiv:0901.0568 [astro-ph.SR]. 14

[64] J. F. Beacom and M. R. Vagins, Phys. Rev. Lett. 93, [75] J. M. Berryman, A. de Gouvea, and D. Hernandez, Phys. 171101 (2004), arXiv:hep-ph/0309300 [hep-ph]. Rev. D 92, 073003 (2015), arXiv:1411.0308 [hep-ph]. [65] C. Simpson et al. (Super-Kamiokande), Astrophys. J. [76] M. Bustamante, J. F. Beacom, and K. Murase, Phys. 885, 133 (2019), arXiv:1908.07551 [astro-ph.HE]. Rev. D 95, 063013 (2017), arXiv:1610.02096 [astro- [66] F. An et al. (JUNO), J. Phys. G43, 030401 (2016), ph.HE]. arXiv:1507.05613 [physics.ins-det]. [77] M. Bustamante, (2020), arXiv:2004.06844 [astro-ph.HE]. [67] J. F. Beacom et al. (Jinping), Chin. Phys. C41, 023002 [78] M. H. Moulai, C. A. Arg¨uelles, G. H. Collin, J. M. (2017), arXiv:1602.01733 [physics.ins-det]. Conrad, A. Diaz, and M. H. Shaevitz, (2019), [68] S. J. Li, J. J. Ling, N. Raper, and M. V. Smirnov, Nucl. arXiv:1910.13456 [hep-ph]. Phys. B944, 114661 (2019), arXiv:1905.05464 [hep-ph]. [79] D. Cianci, A. Furmanski, G. Karagiorgi, and [69] M. Askins et al. (Theia), Eur. Phys. J. C 80, 416 (2020), M. Ross-Lonergan, Phys. Rev. D 96, 055001 (2017), arXiv:1911.03501 [physics.ins-det]. arXiv:1702.01758 [hep-ph]. [70] Y. S. Jeong, S. Palomares-Ruiz, M. H. Reno, and [80] M. Antonello et al. (MicroBooNE, LAr1-ND, ICARUS- I. Sarcevic, JCAP 1806, 019 (2018), arXiv:1803.04541 WA104), (2015), arXiv:1503.01520 [physics.ins-det]. [hep-ph]. [81] P. A. Machado, O. Palamara, and D. W. Schmitz, Ann. [71] M. Bustamante, C. Rosenstrøm, S. Shalgar, and Rev. Nucl. Part. Sci. 69, 363 (2019), arXiv:1903.04608 I. Tamborra, Phys. Rev. D 101, 123024 (2020), [hep-ex]. arXiv:2001.04994 [astro-ph.HE]. [82] S. Ajimura et al., (2017), arXiv:1705.08629 [physics.ins- [72] M. Escudero and M. Fairbairn, Phys. Rev. D 100, 103531 det]. (2019), arXiv:1907.05425 [hep-ph]. [83] S. Shalgar, I. Tamborra, and M. Bustamante, (2019), [73] M. Escudero, J. Lopez-Pavon, N. Rius, and S. Sandner, arXiv:1912.09115 [astro-ph.HE]. (2020), arXiv:2007.04994 [hep-ph]. [84] A. Palazzo, Phys. Rev. D 83, 113013 (2011), [74] B. Aharmim et al. (SNO), Phys. Rev. C 88, 025501 arXiv:1105.1705 [hep-ph]. (2013), arXiv:1109.0763 [nucl-ex].