The tank sat more than a mile down in a working nickel mine, filled with the purest water a laboratory could make.
It had been poured in to calibrate instruments before the switch to a more expensive liquid.
Then something arrived that the tank was never built to catch.
Not many, and not often.
Over six months of quiet running, a handful of faint double flashes turned up in the record, and every one of them had started 150 miles away.
What an antineutrino leaves behind when it finally hits something
A reactor makes these particles by the enormous quantity while it runs.
They carry no charge, weigh almost nothing, and pass through rock as though the rock were not there.
Catching one is a matter of waiting for the rare occasion when a particle strikes a proton head on, which turns that proton into a neutron and throws out a positron.
The positron is the first flash.
It moves faster through water than light does through water, which is legal and produces a cone of blue light behind it, the optical equivalent of a sonic boom.
The neutron is the second flash, and it comes late.
It wanders for a moment, then gets captured by a hydrogen nucleus, which releases a gamma ray at 2.2 million electron volts.
Two flashes, microseconds apart, in that order.
That pairing is the signature, and it is the only thing separating a real event from background noise.
Why the room was full of water at all
The detector sits under about 1.2 miles of rock.
That depth is not scenery. The overburden screens out the cosmic ray shower that would otherwise bury a signal this faint under millions of false coincidences every day.
Inside the cavern is an acrylic sphere about 40 feet across, watched by roughly 9,500 photomultiplier tubes.
It held around 860 tons of ultrapure water.
And it held that water for an unglamorous reason, because the instrument was being commissioned and the water was there to be characterized rather than to catch anything.
Which is the part worth sitting with.
The medium in the tank was a placeholder, and the team went looking for reactor antineutrinos in it because it was there.
What the record says
The dataset came from 190 days of running in 2018.
Two independent analysis methods were used to pull the coincidences out, and the combined result reached 3.5 standard deviations, which is evidence rather than discovery.
The flux came from Ontario stations at Bruce, Darlington and Pickering.
The nearest of them sits 150 miles off, running the way it runs every day, doing nothing unusual at all.
The result went into a physics journal in early 2023.
One caveat on the counting. The per method candidate numbers circulating in secondary coverage do not reconcile with each other, and the published breakdown is not something a reader can easily check.
So treat the number of events as a handful rather than a figure, because that is how the collaboration itself describes it.
Why so few is still the point
Water is a poor detector medium and always has been.
Scintillator, the oily liquid these experiments normally use, throws off far more light per interaction, which is why every previous reactor antineutrino result came from a tank full of it.
The usual fix for water is doping.
Add gadolinium and neutron capture becomes far easier to see, and most proposals for water based antineutrino detection assume you will have to.
This one did not.
A collaboration member put the reason plainly, saying it worked in a water detector because their neutron detection efficiency was high, which is an engineering statement rather than a physics one.
That distinction matters commercially.
An array of plain water tanks costs a fraction of the same volume in scintillator, and water is available everywhere.
What a commissioning fill turned into
The obvious application is watching reactors you are not allowed inside.
An operating core emits antineutrinos continuously and cannot be run without emitting them, so a detector at distance is in principle a verification tool that no operator can spoof.
Distance remains the hard part.
Flux falls off with the square of range, so monitoring at any useful standoff means a far larger tank or far clearer water, and those are engineering problems with no cheap answer.
The tank itself has moved on.
It has since been filled with the scintillator it was always meant to hold, and the water phase is over, which makes this result a byproduct of a plumbing schedule.
Buried instruments keep doing this.
A telescope under a Japanese mountain was built to look for ghost particles from dead stars, and the useful trick is nearly always the unplanned one.
The university summary has the plain language version. The preprint has the statistics.
Read the whole thing?
Get the week's signal, not the noise
Our sharpest reporting on energy, climate and nature — free, once a week.
