The solar neutrino problem
From 1968 onwards every experiment that counted neutrinos from the Sun found fewer than the solar model predicted. The deficit was resolved in 2002 not by a flaw in the Sun or in the detectors but by new physics: electron neutrinos turn into other flavours.
Calculator: how many solar neutrinos pass through you? →
Beginner
The Sun as a neutrino factory
The Sun shines because hydrogen nuclei in its core fuse into helium. Each time four hydrogen nuclei become one helium nucleus, two neutrinos are released. These solar neutrinos fly straight out of the Sun and reach the Earth eight minutes later. About 60 billion of them cross every square centimetre of your skin each second, day and night, since the Earth does not block them.
Light from the Sun's surface left the core long ago. Neutrinos, by contrast, come straight from the furnace. Measuring them is therefore the only direct way to check what is happening at the centre of the Sun right now.
A prediction and a test
In the early 1960s the astrophysicist John Bahcall calculated how many neutrinos the Sun should produce, using a detailed computer model of its interior, later called the standard solar model. The chemist Raymond Davis Jr. proposed to catch them1. In the Homestake gold mine in South Dakota, almost 1500 metres underground, he filled a tank with 615 tonnes of a chlorine-rich cleaning fluid. A neutrino can turn a chlorine atom into an argon atom, and Davis counted those argon atoms one by one.
Only a third
The first results in 1968 were a surprise: Davis found only about a third of the predicted number. For the next decades the solar neutrino problem remained open. Either the solar model was wrong, or the experiment was wrong, or something happened to the neutrinos on their way. Davis refined his method and kept measuring until 1994; the deficit stayed2.
Other detectors joined. In Japan, the water detector Kamiokande saw neutrinos arriving from the direction of the Sun, proving they really were solar, but found only about half the expected number3. Two experiments using the metal gallium, SAGE in Russia and GALLEX in Italy, were sensitive to the most common, low-energy solar neutrinos. They too saw only a little over half4. Different experiments saw different shortfalls, which was hard to explain by simply turning down the Sun's temperature.
The answer from Canada
Since 1962 physicists had known that neutrinos come in more than one kind, or flavour. The Sun makes only electron neutrinos, and the early detectors could only see electron neutrinos. The Italian-Soviet physicist Bruno Pontecorvo had suggested that neutrinos might change flavour on the way. If so, the missing neutrinos were not missing, just disguised.
The Sudbury Neutrino Observatory, 2 kilometres underground in Canada, settled the question. It used 1000 tonnes of heavy water, whose deuterium nuclei can be split by neutrinos of any flavour. In 2002 SNO reported that the total number of solar neutrinos matched the solar model, while only about one third arrived as electron neutrinos5. The Sun was fine; the neutrinos had changed. Davis and Masatoshi Koshiba shared the Nobel Prize in 2002, and SNO's director Arthur McDonald shared it in 2015.
Help from the Sun's own matter
Part of the flavour change happens inside the Sun itself. Passing through the dense solar interior, neutrinos interact gently with electrons, which can strongly enhance the change. This is called the MSW effect, after the physicists who described it. Later the Italian detector Borexino measured the Sun's neutrinos in detail, energy by energy, including those from a rarer fusion process involving carbon, nitrogen and oxygen7. What began as a problem became a precision tool for studying both neutrinos and the Sun.
Quiz
Test your knowledge
4 questions on this level · Print as worksheet
Intermediate
Neutrinos from hydrogen burning
The Sun converts hydrogen to helium mainly through the pp chain, with about 1 % of its energy coming from the CNO cycle. The net reaction is
From the solar luminosity one obtains a total flux at Earth of about 6.5 × 10¹⁰ neutrinos per cm² per second. Most are low-energy pp neutrinos (below 0.42 MeV). Smaller components include the monoenergetic ⁷Be neutrinos at 0.862 MeV and the rare ⁸B neutrinos, which reach about 15 MeV and dominate water Cherenkov detectors. John Bahcall's standard solar model predicts each component.
Homestake: the first deficit
In 1964 Bahcall and Raymond Davis Jr. published back-to-back papers proposing a chlorine experiment1. The reaction ³⁷Cl + νₑ → ³⁷Ar + e⁻ has a threshold of 0.814 MeV and is mostly sensitive to ⁸B and ⁷Be neutrinos. Davis's 615-tonne tank at Homestake reported its first result in 1968, well below prediction. The final average over 1970–1994 was 2.56 ± 0.23 SNU2, compared with model predictions of about 8 SNU. One SNU is one capture per 10³⁶ target atoms per second.
Kamiokande and the gallium experiments
For two decades many physicists suspected the experiment or the solar model. Kamiokande in Japan provided the first independent test from 1987. Detecting neutrino–electron elastic scattering in water, it measured energy, time and direction event by event. The electrons pointed back to the Sun, and the ⁸B flux was about half the prediction3.
The gallium experiments SAGE (Baksan, Russia, from 1990) and GALLEX/GNO (Gran Sasso, Italy, 1991–2003) used ⁷¹Ga + νₑ → ⁷¹Ge + e⁻ with a threshold of 0.233 MeV, which reaches the pp neutrinos. The pp flux is fixed almost directly by the solar luminosity, so it was the cleanest test. Both found about 65–70 SNU against a prediction of roughly 130 SNU4. Crucially, the deficits differed: about one third for chlorine, one half for water and gallium. No change in the core temperature could produce this energy dependence, which pointed to the neutrinos themselves.
SNO settles the question
Herbert Chen proposed in 1985 to use heavy water, in which neutrinos of all flavours can break up deuterium through the neutral current. The Sudbury Neutrino Observatory began in 1999 with 1000 tonnes of D₂O. It measured three reactions with ⁸B neutrinos:
- charged current, νₑ + d → p + p + e⁻, sensitive only to electron neutrinos;
- neutral current, ν + d → p + n + ν, equally sensitive to all three flavours;
- elastic scattering on electrons, mostly νₑ with a smaller contribution from νμ and ντ.
In 2002 SNO found a total ⁸B flux of 5.09 × 10⁶ cm⁻² s⁻¹, in agreement with the solar model, but an electron-neutrino flux of only 1.76 × 10⁶ cm⁻² s⁻¹5. The remaining two thirds had arrived as νμ or ντ, a 5.3σ result. In the same year the reactor experiment KamLAND confirmed the oscillation parameters with antineutrinos.
The MSW effect
In matter, electron neutrinos feel an extra potential from charged-current scattering on electrons. Lincoln Wolfenstein described this in 1978, and Stanislav Mikheyev and Alexei Smirnov showed in 1985 that it can cause a resonant, almost complete conversion in the Sun6. For high-energy ⁸B neutrinos the survival probability is about 0.3; for low-energy pp neutrinos, where matter effects are small, it is about 0.55.
Borexino's spectroscopy
Borexino at Gran Sasso, running from 2007 to 2021, used ultrapure liquid scintillator to measure individual components: ⁷Be (2007), pep (2012), pp (2014) and, in 2020, neutrinos from the CNO cycle7. Its results trace the predicted transition between vacuum-dominated and matter-dominated oscillation.
Quiz
Test your knowledge
4 questions on this level · Print as worksheet
Expert
Solar fluxes and their predictions
In the standard solar model the pp flux, about 6 × 10¹⁰ cm⁻² s⁻¹, is tied to the luminosity constraint to about 1 %. The ⁸B flux scales roughly as Tc²⁴ with the core temperature and carries a model uncertainty of order 12 %; ⁷Be scales roughly as Tc¹¹. For the radiochemical rates, Bahcall and Pinsonneault (2004) predicted 8.5 ± 1.8 SNU for chlorine and 131 (+12/−10) SNU for gallium8. Since the 2000s the models have been split by the solar metallicity problem: high-metallicity (GS98) and low-metallicity (AGSS09) abundance sets predict ⁸B and CNO fluxes differing by roughly 20 % and 30–40 % respectively, while low-Z models disagree with helioseismology.
Three decades of deficits
- Homestake, ³⁷Cl, Eth = 0.814 MeV: 2.56 ± 0.16 (stat) ± 0.16 (syst) SNU over 1970–19942, about 30 % of prediction.
- Kamiokande, ν e scattering, Eth ≈ 7 MeV: ⁸B flux 2.80 ± 0.19 ± 0.33 × 10⁶ cm⁻² s⁻¹, (48 ± 3 ± 6) % of prediction3; the angular peak towards the Sun established the solar origin.
- SAGE, ⁷¹Ga, Eth = 0.233 MeV: 65.4 (+3.1/−3.0) (+2.6/−2.8) SNU; GALLEX/GNO combined about 69 SNU4; gallium average around 66 SNU, about half of prediction.
Fits with free solar fluxes found the ⁷Be component apparently suppressed more strongly than ⁸B, which is impossible astrophysically because ⁸B is produced from ⁷Be. This was the strongest model-independent hint of new particle physics before SNO. Proposed solutions included vacuum oscillations (the Gribov–Pontecorvo picture), several MSW branches (small mixing angle, large mixing angle LMA, LOW), and spin–flavour precession via a neutrino magnetic moment.
SNO
With the three channels CC, NC and ES on deuterium and electrons, SNO measured in 2002 φCC = 1.76 ± 0.05 ± 0.09, φNC = 5.09 (+0.44/−0.43) (+0.46/−0.43) and φμτ = 3.41 (+0.66/−0.64), all in units of 10⁶ cm⁻² s⁻¹5. The non-zero νμτ flux at 5.3σ and the agreement of φNC with the model removed the astrophysical explanations. The ratio φCC/φNC ≈ 0.34 measures the high-energy survival probability. KamLAND's reactor disappearance (2002) and spectral distortion (2004) selected LMA uniquely.
The MSW mechanism
Coherent forward scattering adds a potential V = √2 GF ne for νₑ relative to νμ,τ6. In a two-flavour approximation the effective mixing angle in matter satisfies
In the solar core ne ≈ 100 NA cm⁻³, so for Δm²₂₁ ≈ 7.5 × 10⁻⁵ eV² matter dominates above a few MeV. Neutrinos produced as nearly pure ν₂ then propagate adiabatically, giving Pee ≈ sin²θ₁₂ ≈ 0.31 (with small θ₁₃ corrections). Below about 1 MeV vacuum averaging applies and Pee ≈ 1 − ½ sin²2θ₁₂ ≈ 0.57. The predicted rise between the two regimes, the upturn of Pee in the 2–5 MeV region, has not been clearly resolved; ⁸B data from Super-Kamiokande and SNO appear somewhat flatter than expected. Super-Kamiokande also reported a day–night asymmetry of about −3 % from regeneration in the Earth, at under 3σ.
Borexino and the CNO cycle
Borexino's pp measurement gave (6.6 ± 0.7) × 10¹⁰ cm⁻² s⁻¹, consistent with the luminosity constraint and the first direct evidence of the primary fusion reaction7. Its 2020 detection of CNO neutrinos, (7.0 +3.0/−2.0) × 10⁸ cm⁻² s⁻¹ with the no-CNO hypothesis rejected at more than 5σ, was the first direct proof that the CNO cycle operates in stars7. Using the ²¹⁰Bi constraint from ²¹⁰Po tracking in a thermally stabilised detector, the 2022 update gave 6.6 (+2.0/−0.9) × 10⁸ cm⁻² s⁻¹; combined with ⁷Be and ⁸B it disfavours the low-metallicity B16-AGSS09met model at 3.1σ. Current global fits obtain sin²θ₁₂ = 0.307 (+0.012/−0.011) and Δm²₂₁ = (7.49 ± 0.19) × 10⁻⁵ eV², the latter dominated by KamLAND.
Status
The solar neutrino problem in its original form is solved: the deficits are explained by three-flavour oscillations with the LMA-MSW solution. The remaining questions concern precision: the shape of Pee(E) in the transition region, the solar composition, the CNO flux at the 10 % level, and searches for non-standard interactions that would distort the MSW prediction. JUNO, Hyper-Kamiokande, DUNE and SNO+ will address these.
Quiz
Test your knowledge
4 questions on this level · Print as worksheet
Open questions
Open points are the energy dependence of the electron-neutrino survival probability between 1 and 5 MeV, where the predicted upturn has not been clearly seen, and the solar metallicity problem, which a more precise CNO flux could settle.
Terms to explore
- Solar neutrinos
- Solar neutrino problem
- Standard solar Model
- Proton-proton chain
- CNO cycle
- Homestake Chlorine Experiment
- Kamiokande
- SAGE (Soviet–American Gallium Experiment)
- GALLEX / GNO
- SNO (Sudbury Neutrino Observatory)
- Borexino
- CNO neutrino discovery
- MSW effect
- Solar metallicity problem
- KamLAND
People
- John Norris Bahcall
- Raymond Davis Jr.
- Masatoshi Koshiba
- Arthur B. McDonald
- Bruno Pontecorvo
- Lincoln Wolfenstein
- Stanislav Pavlovich Mikheyev
- Alexei Yuryevich Smirnov
- Herbert H. Chen
- Till A. Kirsten
- Vladimir N. Gavrin (Vladimir Nikolaevich Gavrin)
- Gianpaolo Bellini
On the timeline
- 1964Bahcall and Davis propose a solar neutrino test
- 1968Homestake finds too few solar neutrinos
- 1969Oscillations proposed to explain the solar deficit
- 1978Wolfenstein: matter changes neutrino oscillations
- 1985Mikheyev and Smirnov discover resonant conversion
- 1989Kamiokande sees solar neutrinos in real time
- 1991SAGE begins the gallium measurements
- 1992GALLEX observes solar pp neutrinos
- 1998Homestake publishes its final solar result
- 2001SNO shows solar neutrinos change flavour
- 2002SNO measures the total solar neutrino flux
- 2002Nobel Prize for neutrino astrophysics
- 2007Borexino measures beryllium-7 solar neutrinos in real time
- 2014Borexino measures pp neutrinos spectroscopically
- 2020Borexino detects neutrinos from the CNO cycle
- 2015Nobel Prize for neutrino oscillations
References
- J. N. Bahcall, Solar neutrinos. I. Theoretical, Phys. Rev. Lett. 12, 300 (1964); R. Davis Jr., Solar neutrinos. II. Experimental, Phys. Rev. Lett. 12, 303 (1964), doi:10.1103/PhysRevLett.12.300, doi:10.1103/PhysRevLett.12.303
- B. T. Cleveland et al., Measurement of the solar electron neutrino flux with the Homestake chlorine detector, Astrophys. J. 496, 505 (1998), doi:10.1086/305343
- Y. Fukuda et al. (Kamiokande Collaboration), Solar neutrino data covering solar cycle 22, Phys. Rev. Lett. 77, 1683 (1996), doi:10.1103/PhysRevLett.77.1683
- J. N. Abdurashitov et al. (SAGE Collaboration), Phys. Rev. C 80, 015807 (2009); M. Altmann et al. (GNO Collaboration), Phys. Lett. B 616, 174 (2005), doi:10.1103/PhysRevC.80.015807, doi:10.1016/j.physletb.2005.04.068
- Q. R. Ahmad et al. (SNO Collaboration), Direct evidence for neutrino flavor transformation from neutral-current interactions in the Sudbury Neutrino Observatory, Phys. Rev. Lett. 89, 011301 (2002), doi:10.1103/PhysRevLett.89.011301
- L. Wolfenstein, Neutrino oscillations in matter, Phys. Rev. D 17, 2369 (1978); S. P. Mikheyev, A. Yu. Smirnov, Sov. J. Nucl. Phys. 42, 913 (1985), doi:10.1103/PhysRevD.17.2369
- Borexino Collaboration, Neutrinos from the primary proton–proton fusion process in the Sun, Nature 512, 383 (2014); Experimental evidence of neutrinos produced in the CNO fusion cycle in the Sun, Nature 587, 577 (2020); Phys. Rev. Lett. 129, 252701 (2022), doi:10.1038/nature13702, doi:10.1038/s41586-020-2934-0, doi:10.1103/PhysRevLett.129.252701
- J. N. Bahcall, M. H. Pinsonneault, What do we (not) know theoretically about solar neutrino fluxes?, Phys. Rev. Lett. 92, 121301 (2004), doi:10.1103/PhysRevLett.92.121301
Last updated: · Editorial standard
Cite this page
Text
Neutrino Atlas (2026). “The solar neutrino problem”. Retrieved 22 November 2026, from https://neutrino-atlas.com/learn/the-solar-neutrino-problem/BibTeX
@misc{neutrinoatlas2026nthesolarneutrinoproblem,
title = {The solar neutrino problem},
author = {{Neutrino Atlas}},
year = {2026},
url = {https://neutrino-atlas.com/learn/the-solar-neutrino-problem/},
note = {Accessed 2026-11-22},
language = {english}
}