Dark Matter Interpretation of the Super-Kamiokande Antineutrino Excess and Predictions for JUNO
Listen
Radio episode about this paper
Transcript
Introduction to the show: ident: High Energy Physics: Phenomenology. The day in particle phenomenology: models, signatures and constraints.
Kai: We're looking at the paper: "Dark Matter Interpretation of the Super-Kamiokande Antineutrino Excess and Predictions for JUNO".
Mira: Here's the gist of the paper. Super-Kamiokande has reported a small excess of electron antineutrino events in the 20 MeV energy range, in the search for the diffuse supernova neutrino background This signal is interpreted as a possible indication of dark matter that annihilates dominantly into neutrinos, pointing to a thermal dark matter candidate with s-wave annihilation and with mass in the tens of MeV range Neutrino experiments, including JUNO, will be able to test this hypothesis in the coming years.
Kai: Experimental Context and Signal Interpretation The study investigates the intriguing possibility that the SK antineutrino excess is an indication of sub-GeV dark matter (DM) The existence of DM in the Universe is well established observationally Self-annihilations of local MeV-scale DM particles into neutrinos would be a smoking-gun signature at neutrino detectors in the same energy ballpark of DSNB searches.
Mira: The authors analyze the SK data from all six runs, including those after the Gd-loadings, and evaluate the statistical significance of the excess when interpreted as a DM signal.
Kai: They consider both the case in which DM annihilates directly into neutrinos, leading to a very peaked signature, or into dark particles that subsequently decay into neutrinos with a typically broader spectrum.
Paper summary: Kai: So, to wrap up, the authors of "Dark Matter Interpretation of the Super-Kamiokande Antineutrino Excess and Predictions for JUNO" are essentially saying that this excess in the super-kamiokande data points toward a direct dark matter annihilation into neutrinos >
Mira: Specifically, they've interpreted it as an indication of a direct dark matter to neutrino pair annihilation for a dark matter mass around twenty two point one MeV >
Lev: And the key feature here is that the required annihilation cross-section fits what we expect from thermal freeze-out calculations for allowed local dark matter profiles >
Kai: It's interesting because this points to a thermal dark matter candidate with s-wave annihilation and a mass in the tens of MeV range, which fits into richer dark sector extensions of the standard model >
Mira: And the implication is that neutrino experiments like JUNO are going to be able to test this hypothesis in the coming years by looking for this specific signal >
Conclusion: Kai: So we’ve been looking at this paper from Super-Kamiokande about that small excess of antineutrino events in the twenty MeV energy range, and now we’re getting to the conclusion and what it really means for us on Earth.
Mira: This paper, titled "Dark Matter Interpretation of the Super-Kamiokande Antineutrino Excess and Predictions for JUNO," essentially takes that small bump they saw in their data—that excess of electron antineutrinos—and they are trying to explain it as something from dark matter annihilation.
Lev: It suggests this dark matter is thermal, meaning it was produced in the early universe, and it annihilates dominantly into neutrinos. They pinpoint a mass for this particle at around twenty two point one MeV, which is right in that low-energy range we’re looking at for these searches.
Kai: That twenty two point one MeV mass seems pretty specific, and the authors are saying the annihilation cross-section they found fits what we calculate from thermal freeze-out, which is actually pretty good news because it means their dark matter model is consistent with how much of it should exist in our universe.
Mira: Right, so they’re linking an astrophysical observation—the neutrinos detected by Super-Kamiokande—to a particle physics concept—dark matter annihilation cross-section. The real takeaway is that this isn't just a random fluctuation; it could be a sign of something new existing in the dark sector that we haven't seen before.
Lev: From an engineering standpoint, if this twenty two point one MeV mass is correct, then future experiments like JUNO, with its three percent energy resolution, are perfectly positioned to look for this kind of sharp signal because it would really stand out against the background noise.
Kai: Exactly. The paper proposes that these new results from Super-Kamiokande and the plans for new neutrino detectors like JUNO give us a clear path forward to test if this dark matter hypothesis is actually real.
Mira: And while they have identified this direct annihilation scenario as the simplest explanation, they also leave room for other possibilities, like dark matter annihilating into something else that then decays into neutrinos.
Lev: That’s important because it tells us we can’t just stop at one idea; we need to keep pushing the search for different decay chains or mediators.
Kai: So, in a nutshell, this paper is a strong call to action for future neutrino experiments to get ready for these specific kinds of signals from dark matter.
Mira: And the real question now is whether JUNO will be able to confirm this twenty two point one MeV mass signal when they run their data.
Alessandro Granelli, Silvia Pascoli, Salvador Rosauro-Alcaraz
Instituto de F༿isica Corpuscular (IFIC) · Dipartimento di Fisica e Astronomia, Universitá di Bologna · INFN, Seziona di Bologna · Universitá Clermont-Auvergne, CNRS
hep-ph, astro-ph.CO
Submitted: 2026-05-19
Updated: 2026-10-27
Comments: 11 pages, 2 figures; typos fixed and clarifications added; version accepted for publication in SciPost Physics
Journal ref: SciPost Physics 21 (2026) 106
DOI: 10.21468/SciPostPhys.21.4.106
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Importance score: 84/100
The gist: Super-Kamiokande has reported a small excess of electron antineutrino events in the 20 MeV energy range…
Key concepts
- Diffuse Supernova Neutrino Background (DSNB)
- This is the integrated neutrino flux from all past distant supernovae. It serves as a background signal that neutrino experiments monitor, and the excess observed in SK was initially considered a possible DSNB signal.
- Thermal Dark Matter Freeze-out
- This describes how dark matter particles reached their current abundance in the universe. The required annihilation cross-section to match the observed dark matter density (from PLANCK data) is calculated based on this thermal freeze-out scenario.
- J-factor
- The J-factor is a measure of the astrophysical uncertainty related to dark matter distribution. It is calculated as the line-of-sight integral of the square of the local dark matter density, reflecting how much DM is present along our line of sight.
- Direct vs. Indirect Annihilation
- Direct annihilation means dark matter particles annihilate directly into neutrinos (a sharp energy peak). Indirect annihilation involves DM annihilating into other particles first, which then decay into neutrinos (a broader spectrum). Both channels were analyzed to explain the SK signal.
Terminology
Summary
Super-Kamiokande has reported a small excess of electron antineutrino events in the 20 MeV energy range, in the search for the diffuse supernova neutrino background This signal is interpreted as a possible indication of dark matter that annihilates dominantly into neutrinos, pointing to a thermal dark matter candidate with s-wave annihilation and with mass in the tens of MeV range Neutrino experiments, including JUNO, will be able to test this hypothesis in the coming years. p. 0
Experimental Context and Signal Interpretation
The study investigates the intriguing possibility that the SK antineutrino excess is an indication of sub-GeV dark matter (DM) The existence of DM in the Universe is well established observationally Self-annihilations of local MeV-scale DM particles into neutrinos would be a smoking-gun signature at neutrino detectors in the same energy ballpark of DSNB searches. The authors analyze the SK data from all six runs, including those after the Gd-loadings, and evaluate the statistical significance of the excess when interpreted as a DM signal. p. 1 They consider both the case in which DM annihilates directly into neutrinos, leading to a very peaked signature, or into dark particles that subsequently decay into neutrinos with a typically broader spectrum. p. 1
Theoretical Framework for Dark Matter Annihilation
In the standard scenario of thermal freezeout, the DM relic abundance is controlled by the total annihilation cross-section σtot, in particular by ⟨σtotv⟩FO with v the DM relative velocity and ⟨·⟩FO the thermal average computed at freeze-out. Explaining the value of omegaDMh2 = 0. p. 112 as measured by PLANCK requires ⟨σtotv⟩FO ≃ 3. p. 18 (4.3) × 10−26cm3 s−1 for mDM = 10 (100) MeV for a real scalar/Majorana DM. p. 1 Annihilations into charged particles or photons cannot saturate the value required and annihilations uniquely into light stable or long-lived dark sector particles are also disfavoured since the latter would contribute to the effective number of relativistic degrees of freedom affecting Big Bang Nucleosynthesis (BBN). p. 1 Therefore, annihilation into neutrinos is the Standard Model (SM) channel of choice, the one that can (directly or indirectly) dominate DM annihilations and lead the generation of the present DM relic abundance. p. 1 p. 1
Methodology for Signal Detection
The differential diffuse antineutrino flux from nonrelativistic DM annihilations is given by an equation involving the annihilation cross-section and the J-factor. The J-factor in Eq. p. 2 (1) is defined as a line-of-sight integral of the square of the local DM density. p. 2 p. 2 A widely adopted parameterisation for the local DM distribution is the generalised version of the Navarro-Frenk-White (gNFW) profile. p. 2 p. 2 Using this profile, a best-fit ± 1σ value of Javg = 6+3−1 and 4 (3) ≲ Javg ≤ 17 (30) at 2σ (3σ), reflecting the relatively large astrophysical uncertainties. p. 2 The analysis uses a χ2-analysis to assess the sensitivity to DM annihilations. p. 3
Key Findings and Future Prospects
For direct DM annihilations (green), the data prefer a DM mass of mbest-fit DM = 22.1 MeV at the best fit point, and Javg⟨σv⟩0 ∈ [1. p. 3 p. 38×10−25, 2.3 × 10−24] cm3 s−1 at 2σ. p. 3 p. 3 The annihilation cross section into neutrinos is compatible with that required for thermal dark matter freeze-out. p. 3 If DM annihilates into scalars that subsequently decay into neutrinos, the preferred mass depends on the width of the neutrino spectrum, which is controlled by the mass difference between DM and the scalar. p. 3 p. 3 The signal has a significance of 2. p. 157σ for direct annihilation, 2. p. 459σ for indirect annihilation with nearly degenerate DM and scalar (∆ = 0. p. 403), and 2.45σ for indirect annihilation with larger mass splitting (∆ = 0. p. 45), to be compared to the SK DSNB case. p. 4 New experiments like JUNO are ideally suited to detect IBD, with an exceptional energy resolution of 3%. p. 4 The peaked DM signal clearly stands above both the background and the DSNB contribution, and would thus be readily distinguishable. p. 4
Discussion and Conclusion
The excess can be interpreted as an indication of a direct DM DM → νν¯ annihilation for a DM mass mDM of around 22.1 MeV, with an annihilation cross-section compatible with thermal freeze-out for allowed local DM profiles. p. 4 p. 4 p. 4 Other possibilities could also be considered. p. 4 The paper concludes that new results from SK as well as from new neutrino experiments, in particular JUNO and HK, have the potential to test this hypothesis in the forthcoming future. p. 5 Time will tell. p. 5
Acknowledgements
The authors thank Michele Lucente, Sergio Palomares-Ruiz and Filippo Sala for useful discussions The work of A.G. is supported by the Spanish grant PID2023-148162NB-C21 (MCIN/AEI/10. p. 513039/501100011033) co-funded by the European Union (FEDER), and in part by the European Union’s Horizon Europe programme under the Marie Sk lodowska-Curie Actions – Staff Exchanges (SE) grant agreement No. p. 5 101201278 – DarkSHunt - ERC - 2024 ADG. p. 5 The work of SP has been partly funded by the European Union under the Horizon Europe’s Project: 101201278 – DarkSHunt - ERC - 2024 ADG. p. 5
References
The paper lists numerous references supporting its claims, including works on Super-Kamiokande experiments, constraints on MeV dark matter particles from BBN and the CMB, and analyses of neutrino-dark matter interactions. The extensive list of citations demonstrates the broad theoretical context for interpreting this anomaly in particle physics. The references provide the necessary background for understanding both the experimental setup and the theoretical predictions for dark matter annihilation channels. p. 0 The references also include recent results from collaborations like JUNO and Hyper-Kamiokande, which are crucial for testing the proposed hypothesis in future experiments. p. 5 The bibliography shows a comprehensive approach to linking astrophysical observations with particle physics models of dark matter.
Final Summary Statement
The excess can be interpreted as an indication of a direct DM DM → νν¯ annihilation for a DM mass mDM of around 22.1 MeV, with an annihilation cross-section compatible with thermal freeze-out for allowed local DM profiles.
--- Page 1 ---
Dark Matter Interpretation of the Super-Kamiokande Antineutrino Excess and Predictions for JUNO Alessandro Granelli a,†, Silvia Pascoli b,c,‡ and Salvador Rosauro-Alcaraz d,§ Instituto de F´ısica Corpuscular (IFIC), CSIC-Universitat de Val´encia, Parc Cient´ıfic, C/ Catedr´atico Jos´e Beltr´an 2, E-46980 Paterna, Spain Dipartimento di Fisica e Astronomia, Universitá di Bologna, via Irnerio 46, 40126, Italy INFN, Sezione di Bologna viale Berti Pichat 6/2 viale Berti Pichat 6/2 Dipartimento di Fisica e Astronomia, Universitá di Bologna via Irnerio 46, 40126 Italy INFN Sezione di Bologna viale Berti Pichat 6/2
--- Page 2 ---
Super-Kamiokande has reported a small excess of electron antineutrino events in the 20 MeV energy range, in the search for the diffuse supernova neutrino background We interpret this signal as a possible indication of dark matter that annihilates dominantly into neutrinos, pointing to a thermal dark matter candidate with s-wave annihilation and with mass in the tens of MeV range Neutrino experiments, including JUNO, will be able to test this hypothesis in the coming years
--- Page 3 ---
Gd-loaded SK (SK-Gd) has significantly increased the sensitivity to, e.g., cosmicray spallations, atmospheric and reactor neutrinos, and the Diffuse Supernovae Neutrino Background (DSNB) The latter is the integrated neutrino flux originated from all the past distant supernovae [5–7] Intriguingly, an excess of electron antineutrino (¯νe) events in the energy range relevant to DSNB searches, specifically around an antineutrino energy Eν¯ ≈ 22 MeV, has been reported at SK, already at ∼ 1. p. 05σ when operating with pure water [15] and later after the second Gd-loading [17] (see also [18]) When interpreted as a DSNB signal and considering only post-Gd-loadings data [17], the background-only hypothesis is disfavoured at 0. p. 05σ − 1.7σ, depending on the adopted DSNB model [19–38] and neutron tagging approach
--- Page 4 ---
Here, we analyse the SK data from all six runs, including those after the Gd-loadings, and evaluate the statistical significance of the excess when interpreted as a DM signal We consider both the case in which DM annihilates directly into neutrinos, leading to a very peaked signature [43], or into dark particles that subsequently decay into neutrinos with a typically broader spectrum
--- Page 5 ---
In the standard scenario of thermal freezeout, the DM relic abundance is controlled by the total annihilation cross-section σtot, in particular by ⟨σtotv⟩FO with v the DM relative velocity and ⟨·⟩FO the thermal average computed at freeze-out Explaining the value of omegaDMh2 = 0.12 as measured by PLANCK [50] with s-wave annihilations requires ⟨σtotv⟩FO ≃ 3. p. 18 (4.3) × 10−26cm3 s−1 for mDM = 10 (100) MeV for a real scalar/Majorana DM – somewhat larger than the canonical 3 × 10−26cm3 s−1, see e. p. 1g. [51]. These values would be double in the case of complex scalar/Dirac DM. p. 1 In the DM mass range of interest mDM ∼ (10 − 100) MeV, annihilations into charged particles or photons cannot saturate the value required [52–59] and annihilations uniquely into light stable or long-lived dark sector particles are also disfavoured since the latter would contribute to the effective number of relativistic degrees of freedom affecting Big Bang Nucleosynthesis (BBN). p. 1 An alternative is to have DM annihilations into dark sector particles that promptly decay dominantly into neutrinos, directly or via decay chains, as typical in rich dark sectors [65,66]. p. 1
--- Page 6 ---
As per the case of direct annihilations above, decays into charged particles, photons or stable/longlived light dark sector particles are constrained to be subdominant [67]. Therefore, annihilation into neutrinos is the Standard Model (SM) channel of choice, the one that can (directly or indirectly) dominate DM annihilations and lead the generation of the present DM relic abundance. DM-neutrino interactions have been studied both from a theoretical point of view [68,69], often linking it to neutrino mass generation at the subGeV scale [70–72], and for their cosmological consequences. p. 1 Annihilations DM DM → νν¯ can proceed via the exchange of dark sector mediators. p. 1 Typically, SM gauge invariance can be guaranteed by invoking e. p. 1g. mixing between neutrinos and heavy neutral leptons (HNLs) In the case of annihilation into dark sector particles, all couplings involved refer to the dark sector and can be sizable so that sufficiently large cross-sections can be easily achieved. p. 1
--- Page 7 ---
For concreteness, we consider two options to explain the SK ¯νe excess: DM annihilating directly into a neutrino-antineutrino pair, or DM annihilating into a pair of dark particles, say two real scalars ϕ, each promptly decaying into a neutrino-antineutrino pair. We note that other possibilities are viable, for instance DM annihilating into one neutrino and one HNL or into two HNLs. p. 1 The antineutrino flux is given by an equation involving the annihilation cross-section and the J-factor. p. 2 In the case of DM directly annihilating into a neutrino-antineutrino pair, the antineutrino spectrum is monochromatic, reading dNν¯/dEν¯ = δ(Eν¯ − mDM). p. 2 In the case of DM annihilating into two scalars first, the formal expression for the flux remains unchanged, provided that ⟨σv⟩0 is understood as the annihilation cross-section into scalars, and dNν¯/dEν¯ is doubled and uniform, with an average energy of mDM/2 and a width of δEν¯ = mDM√2∆p1 − ∆2/2/(1 − ∆), where ∆ ≡ (mDM − mϕ)/mDM with 0 < ∆ ≤ 1. p. 2
--- Page 8 ---
The J-factor in Eq. (1) is defined as a line-of-sight integral of the square of the local DM density. A widely adopted parameterisation for the local DM distribution is the generalised version of the Navarro-Frenk-White (gNFW) profile. By using this profile, we obtain a best-fit ± 1σ value of Javg = 6+3−1 and 4 (3) ≲ Javg ≤ 17 (30) at 2σ (3σ), reflecting the relatively large astrophysical uncertainties. p. 2 The analysis yields the following best-fit ± 3σ value for the local DM density at R0: ρ0 = 0. p. 266+0.16−0.2 GeV cm−3. p. 0
--- Page 9 ---
SK Collaboration has so far performed six analysis in search for the DSNB, accumulating more than 6000 days of data taking We perform a χ2-analysis of all SK data in order to assess the sensitivity to DM annihilations. In Fig. 1, we present the results of the analysis For direct DM annihilations (green), the data prefer a DM mass of mbest-fit DM = 22. p. 31 MeV at the best fit point, and Javg⟨σv⟩0 ∈ [1.8×10−25, 2.3 × 10−24] cm3 s−1 at 2σ. If DM annihilates into scalars that subsequently decay into neutrinos, the preferred mass depends on the width of the neutrino spectrum, which is controlled by the mass difference between DM and the scalar. Higher values of Javg⟨σv⟩0 require in this case still allowing consistency with the thermal freeze-out one but at a lower statistical significance. p. 4
--- Page 10 ---
For illustration, we show in Fig. 2 the predicted number of events in JUNO for 147 kton × yr, using the best-fit results we obtained for the direct annihilation case We find that the peaked DM signal clearly stands above both the background and the DSNB contribution, and would thus be readily distinguishable. p. 4 New results from SK as well as from new neutrino experiments, in particular JUNO and HK, will be able to test this hypothesis in the coming future. p. 5 In particular, JUNO is a 20 kton LSc detector ideally suited to detect IBD, with an exceptional energy resolution of 3%. p. 4
--- Page 11 ---
We have proposed an explanation for the SK electronantineutrino excess around 20 MeV as a signal of DM (direct or indirect) annihilations into neutrinos We find that the excess can be interpreted as an indication of a direct DM DM → νν¯ annihilation for a DM mass mDM of around 22.1 MeV, with an annihilation cross-section compatible with thermal freeze-out for allowed local DM profiles. This is the simplest case, pointing to DM annihilating primarily into neutrinos via s-wave Other possibilities could also be considered. p. 4
--- Page 12 ---
We point out a possible intriguing coincidence with the 511 KeV line due to positron annihilations in the bulge Despite being of long standing, this anomaly has not yet received a convincing astrophysical explanation and DM annihilations into electron-positron pairs have been advocated as a possible source. Recent analyses suggest that a spiky profile can induce the observed flux for DM masses up to 20 MeV without contradicting constraints from disk emission and inflight positron annihilation with the interstellar medium, and even for higher injection positron energy [93]. p. 5 Notice that the branching ratio into electron-positrons is required to be much smaller than unity, compatible with our requirement of dominant annihilation into neutrinos In the case of DM annihilations directly into neutrinos and into electron-positrons, the DM mass implied by the SK signal, 22. p. 51 MeV, is in the range required for the 511 KeV line [92]. p. 5 In the case of annihilation first into dark scalars, it should be taken into account that the energy injection of the electron/positrons from the scalar decays are much lower, making the higher DM mass again compatible with the 511 KeV signal. p. 5
Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.
Improvements for AI systems
-
Improved Dark Matter Search Capabilities: AI systems can now search for sub-GeV dark matter annihilation signals by predicting a
peaked signature
from direct annihilation DM DM to neutrinos or indirect annihilation into dark particles that decay into neutrinos, as described in Figure 2. -
Enhanced Experimental Discrimination: AI can distinguish between the predicted peaked signal and the broad DSNB contribution in upcoming experiments like JUNO by using its excellent energy resolution of 3% to separate different energy features.
-
Parameter Estimation for DM Properties: AI systems can estimate best-fit DM mass and local dark matter density values based on Super-Kamiokande data, such as the best fit DM mass of 22.1 MeV from direct annihilation scenarios.
-
Cross Section Compatibility Assessment: The AI can assess whether the required annihilation cross section is compatible with the thermal freeze-out requirement for s-wave annihilation, finding that for direct DM DM to neutrinos annihilation, it is
compatible with that required for thermal dark matter freeze-out.
-
Model Flexibility in Interpretation: AI systems can explore alternative explanations by considering different mass splittings, such as the
indirect annihilation into neutrinos via intermediate real scalars,
and determine the corresponding best-fit parameters.
Sources
- First Gadolinium Loading to Super-Kamiokande
- Second gadolinium loading to Super-Kamiokande
- Diffuse supernova neutrino background with up-to-date star formation rate measurements and long-term multidimensional supernova simulations
- Neutrinos from dense environments : Flavor mechanisms, theoretical approaches, observations, and new directions
- Supernova Relic Neutrino Search at Super-Kamiokande
- Supernova Relic Neutrino Search with Neutron Tagging at Super-Kamiokande-IV
- Diffuse Supernova Neutrino Background Search at Super-Kamiokande
- Search for astrophysical electron antineutrinos in Super-Kamiokande with 0.01wt% gadolinium-loaded water
- The Diffuse Supernova Neutrino Background is detectable in Super-Kamiokande
- Diffuse neutrino flux from failed supernovae
- Shockwaves in Supernovae: New Implications on the Diffuse Supernova Neutrino Background
- Spectrum of the Supernova Relic Neutrino Background and Metallicity Evolution of Galaxies
- Diffuse neutrinos from luminous and dark supernovae: prospects for upcoming detectors at the O(10) kt scale
- Diffuse neutrino supernova background as a cosmological test
- Diffuse neutrino background from past core-collapse supernovae
- Search for Diffuse Supernova Neutrino Background with 956.2 days of Super-Kamiokande Gadolinium Dataset
- Diffuse Supernova Neutrino Background from extensive core-collapse simulations of $8$-$100 {\rm M}_\odot$ progenitors
- Fundamental physics with the diffuse supernova background neutrinos
- Impact of binary interactions on the diffuse supernova neutrino background
- Stellar Collapse Diversity and the Diffuse Supernova Neutrino Background
Related papers
- MadNIS at NLO
- The NANOGrav 15-year data set: A Limit on Continuous Spin Gravitons
- DIS-secting the universe with high-energy neutrinos
- Crowdsourcing Gravitational Waves from Superradiant Axions
- Classification of g-modes for neutron stars with a strong transition: Novel universal relation including slow stable hybrid stars
- Scalar axion field of toroidal electromagnetic pulses