Raymond Davis Jr.: The Chemist Who Counted Atoms One at a Time
The story of Homestake is often told as a triumph, and it was. It is more useful told as what it actually felt like: a chemist doing an experiment nobody else could check, producing a number nobody believed, and continuing for a quarter of a century while the field waited for him to find his mistake. He never found one, because there was not one.
The Absurd Practicality of the Method
Bruno Pontecorvo had proposed in 1946 that a neutrino striking chlorine-37 turns it into argon-37, and that argon, being a noble gas, could be flushed out of the liquid and counted. Davis, a chemist at Brookhaven rather than a physicist, took the proposal literally.
The working fluid was perchloroethylene - dry-cleaning fluid - chosen because it is cheap, chlorine-rich and commercially available by the tanker. He installed 380,000 litres of it, some 615 tonnes, in a tank at the 4,850-foot level of the Homestake gold mine in South Dakota. The depth was necessary: nearly 1,500 metres of rock removes the cosmic-ray background that would otherwise swamp everything.
The expected signal was about one argon-37 atom produced per day in the entire tank. Every few weeks he purged the tank with helium, swept out the accumulated argon, and counted its radioactive decays in a proportional counter small enough to hold in one hand.
It is worth pausing on what that requires. Extracting roughly fifteen specific atoms from 615 tonnes of liquid, and being confident that you got most of them and that they are the right ones, is a chemistry problem of a severity that has few parallels. Davis calibrated it by adding a known number of argon atoms and measuring how many came back.
The Number That Would Not Go Away
From the first results in 1968, the count came out at roughly a third of what the standard solar model predicted. Not zero, which would have suggested the method failed. Not the predicted value. A stable, reproducible third.
There were two obvious explanations and both pointed away from anything interesting. Either Davis's extraction efficiency was worse than he believed, or John Bahcall's solar model overestimated the neutrino production rate. Almost everyone assumed one of those, and for a long time nobody could rule either out.
So Davis kept running, for twenty-five years, improving calibrations and accumulating statistics. Bahcall kept refining the solar model, and it kept predicting the same flux. The discrepancy did not narrow. It hardened.
This became known as the solar neutrino problem, and it is one of the better examples in modern physics of an anomaly that was neither explained away nor prematurely believed. The field held it open for three decades, which is the correct response to a result you cannot account for and cannot dismiss.
The Resolution
Davis's detector was sensitive only to electron neutrinos. That fact was known throughout, and it was the loophole, though nobody could confirm it without a detector that could see the other types.
Super-Kamiokande demonstrated in 1998 that neutrinos change type in flight. The Sudbury Neutrino Observatory then measured both the electron neutrino flux and the total flux of all three types from the Sun, and found the total matched Bahcall's model while the electron component was about a third of it.
Davis had been measuring correctly the entire time. Two thirds of the solar neutrinos were arriving as types his chlorine could not see. The solar model was right, the experiment was right, and what was missing was oscillation - which required neutrinos to have mass, which the Standard Model had said they did not.
He received the Nobel Prize in Physics in 2002, aged 88, sharing it with Masatoshi Koshiba and Riccardo Giacconi. He died in 2006. Bahcall, whose model had been doubted alongside the measurement, did not receive one; he died in 2005.
Why the Story Is Worth Keeping
The obvious reading is persistence rewarded, and that is true but thin. The more useful reading is about what the field did with a result it could not accept.
Nobody suppressed Davis's number, and nobody adopted it either. It was published, repeated, checked, and left standing as an open problem for thirty years while people worked on both the chemistry and the solar model. The resolution came from an independent experiment using a completely different technique, which is the only kind of resolution that settles anything.
It also matters that Davis kept the experiment running rather than defending the result rhetorically. The argument was won by twenty-five years of data, not by insistence, and when the answer arrived it vindicated him from a direction he had not been arguing for.
The same mine now houses DUNE, on the same level, a few hundred metres from where his tank stood.
Frequently asked questions
What did the Homestake experiment actually measure?
Solar neutrinos, by counting argon-37 atoms produced when a neutrino converts a chlorine-37 nucleus. The detector held 615 tonnes of perchloroethylene about 1,478 metres underground.
How few atoms are we talking about?
Roughly one argon atom produced per day in the entire tank. Davis extracted and counted them every few weeks, which means separating about fifteen specific atoms from 615 tonnes of liquid.
Why did nobody believe the result?
Because it was a third of the prediction, and the two straightforward explanations were an extraction problem or an error in the solar model. Neither could be ruled out for a long time, and both were less remarkable than the truth.
What was the actual explanation?
His detector could only see electron neutrinos. Two thirds of the solar neutrinos had changed type in flight before arriving. Confirming that required Super-Kamiokande in 1998 and the Sudbury Neutrino Observatory in 2001 and 2002.
Did he live to be vindicated?
Yes. He received the Nobel Prize in 2002 at the age of 88 and died in 2006. John Bahcall, whose solar model had been doubted alongside the measurement, did not receive one and died in 2005.
