The Sun’s missing neutrinos: a tank, a calculation and a long disagreement
Deep in a South Dakota mine, a few atoms of argon carried news from the Sun. Making sense of their unexpectedly small number took decades of work on both the experiment and the theory.
How do you find a handful of atoms in a huge tank?
The Homestake experiment did not take a picture of the Sun. In a mine in South Dakota, Raymond Davis Jr. and his collaborators filled a large tank with roughly 615 tonnes of a chlorine-bearing cleaning fluid. Very occasionally, a solar electron neutrino could turn a chlorine-37 nucleus into argon-37.
That change offered a chemical way to count rare interactions. After an exposure, the team extracted argon from the liquid and placed it in small counters. Radioactive decays of argon-37 then provided evidence of the atoms created during the run.
The scale of the task was extraordinary: recover a tiny number of atoms from an enormous amount of liquid, then distinguish their decays from background. Efficiency of extraction, efficiency of counting and unwanted production of argon all mattered. The tank’s size alone could not make the answer reliable.
A solar calculation had to become a detector prediction
John Bahcall and his collaborators calculated the neutrinos expected from nuclear reactions inside the Sun. But predicting what Homestake would count required another step: fold those neutrino energies together with the probability that chlorine would respond to each one.
The chlorine reaction has an energy threshold of about 0.814 MeV. The abundant, low-energy neutrinos from the main proton–proton reaction lie below it. Homestake therefore did not count all solar neutrinos with equal sensitivity; the smaller high-energy component was particularly important to its prediction.
That distinction is essential when reading a claim that the experiment saw only a fraction of the expected signal. It was a fraction of the predicted chlorine capture rate. It was not a census showing that the Sun produced the same fraction of every neutrino in its nuclear reaction chains.
- 01Expose chlorine
Rare electron-neutrino captures create argon
- 02Extract the atoms
Recover argon after an exposure
- 03Count its decays
Infer a capture rate, allowing for losses
The low count kept coming back
Results emerging from the late 1960s showed fewer captures than expected. The disagreement persisted as the experiment continued and the solar calculations improved. It became known as the solar neutrino problem.
There were several serious possibilities. The solar model might have assigned the wrong rates to important reactions. An extraction or counting problem might have reduced the experimental yield. Or the neutrinos might not have reached Earth in the form assumed by the calculation.
Both sides had work to do. The experiment checked whether argon could be recovered and recognized; solar researchers examined nuclear inputs, composition and the structure of the Sun. A disagreement between prediction and measurement becomes more interesting when both have survived specific attempts to find an ordinary mistake.
The tank could answer one question very well
Homestake’s strength was also a limit on what it could tell us. Extracting accumulated argon produced a capture rate averaged over an exposure. It did not provide an event-by-event direction, arrival time or neutrino energy.
At solar energies, the chlorine reaction was sensitive to electron neutrinos. If some of those neutrinos arrived as muon or tau flavor, they would not contribute to that same capture signal. A low rate could therefore reflect a change during the journey, not just a weak source.
Homestake alone could not separate all those possibilities. That was a reason to build complementary instruments. A detector that measured a direction, or reached lower energies, or responded differently to flavors could ask a question the chlorine tank could not.
A new comparison finally changed the argument
Later solar experiments used other materials and methods. SNO’s heavy water made a particularly revealing comparison possible: measure the arriving electron-neutrino component and also the total flux of the three active flavors.
SNO found that the electron component was low while the total active flux agreed with solar expectations. KamLAND then tested the relevant oscillation picture with antineutrinos from terrestrial reactors. A question rooted in solar modelling could now be checked using a very different source.
Homestake’s importance was not that its first low count already proved the final explanation. It kept a difficult discrepancy alive long enough for more discriminating measurements to become possible. The resolution grew out of that sustained exchange between calculation, chemical technique and new detector designs.
Primary sources & revision
- Brookhaven National Laboratory · Raymond Davis Jr. and solar neutrinos
- Particle Data Group · Neutrino Masses, Mixing, and Oscillations (2024)
- John Bahcall · Solving the mystery of the missing neutrinos
- John Bahcall · How the Sun shines
- SNO Collaboration · Direct Evidence for Neutrino Flavor Transformation (2002)
- KamLAND Collaboration · First Results: Evidence for Reactor Anti-Neutrino Disappearance (2002/2003)
First published and source-checked on 9 October 2026. Original explanatory text and diagrams by Sandbox Physics. Illustrations are schematic; no experimental event records are reproduced here. This is an independent educational publication, not an official Nobel or experiment collaboration publication.