This Locked Cabinet Holds the Answer to One of the Biggest Questions in Particle Physics

The cabinet door hiding the readout from a secret clock system.
The cabinet door hiding the readout from a secret clock system.
Photo: Ryan F. Mandelbaum

A 50-foot ring topped with white insulation sits attached to wires, pipes, and other electrical components in a warehouse on Fermilab’s northern Illinois campus. Scientists taking data with this device have the potential to rock the field of particle physics to its core, but they’re missing a crucial number to make their final calculation: the ticking speed of a clock that’s kept in a back room hidden in a locked compartment. Today, only two people know this value, and they keep it in hidden envelopes. They’re not telling anyone what it is.

Well, not yet, at least. Currently, a theory called the Standard Model is used to explain the particles that make up our universe and how those particles interact. Physicists have found all of the particles and forces that this theory describes, but there are still countless mysteries in the universe, like the true nature of dark matter or why there’s so much more matter than antimatter, that the Standard Model fails to explain. Various experiments are now probing the Standard Model for cracks, and this year, scientists hope to unveil a measurement from one of them, the Muon g-2 experiment, a measurement that might break from the theory.

“If the number is any different than what the Standard Model predicts, then the only explanation would be that some new particle or or new force was outside of the Standard Model,” postdoctoral researcher Saskia Charity told me as we stood on a platform overlooking the ring of the Muon g-2 experiment.


Nearly a century ago, physicist Paul Dirac made a prediction for the value of the electron’s magnetic moment, called “g,” a number that describes how its spin precesses like a wobbly top in a magnetic field. The value he predicted was 2. Very soon after, measurements revealed that g varied slightly from 2, and physicists began to use the difference between g’s actual value and 2 to probe the internal structure of subatomic particles and the laws of physics more generally. In 1959, CERN in Geneva, Switzerland built the first experiment to measure the g-2 value of a subatomic particle called the muon, basically a heavier, shorter-lived electron. Brookhaven National Lab on Long Island, New York began its own experiment, which concluded in 2001.

Brookhaven published the results of its g-2 experiment in 2004: It’s measurement deviated from what the Standard Model predicted. But the experiment didn’t take enough data for the statistical analysis to definitively prove that the value they’d measured was truly different and not just a statistical fluctuation, at least not to the five-standard-deviation level (known as five sigma) that particle physicists require to call something a discovery. And so, after 10 years, Fermilab decided to pick up the search. It had the magnet from Brookhaven’s experiment shipped along the Atlantic coast, up the Mississippi River, and then to Fermilab on a flatbed truck. Physicists would rerun the experiment with a stronger muon beam.

Scientists at Fermilab have been running the experiment since 2017 and are now working on their analysis of the first wave of results. But human beings are prone to bias, so the scientists have replaced a crucial experimental variable, a clock measurement, with a random number generated by atmospheric noise. Once the researcher’s data analysis code is ironed out and found to work, the researchers will reveal the clock’s ticking speed and input that value into their code.

The true clock measurement comes from a pair of atomic clocks whose readouts sit behind a locked black cabinet in a room full of electronics on racks. Fermilab’s deputy director of research Joe Lykken told me that only two people have access to the time: Lykken and his deputy, Greg Bock. The first run has completed, and the clock measurements sit inside a pair of sealed envelopes, each labeled with Lykken or Bock’s name.


“It’s in my office, which is pretty secure,” Lykken said. He wouldn’t tell me where in his office.

The Muon g-2 experiment.
The Muon g-2 experiment.

But what’s so important about a clock? Fermilab’s experiment works by first directing high-energy protons into a target, producing a shower of more protons, muon’s antiparticle partner called an antimuon, and a particle called a pion—some of which decay into more antimuons. These antimuons travel in a beam 4,000 times around the magnet and then decay into anti-electrons, called positrons, which carry a record of the muons with them. Scientists can calculate the g-2 value based on a ratio of the frequency of precession of the antimuon and the strength of the magnetic field. But calculating a frequency requires knowing the time. The scientists simply switch the random number frequency with the real frequency on the clock to get the final answer, explained Charity. Antimuons are easier for Fermilab to produce, Charity explained, but the g-2 result would be the same as for regular muons.

A ceremony would typically accompany the unblinding of the clock value, but Lykken explained that a select few physicists will likely perform the calculation in private before recreating the result for a larger audience. The first run’s statistics won’t be any stronger than the Brookhaven run, but they will at least confirm whether the presence of the deviation continues or if it was merely a fluctuation. Subsequent runs will tighten the error bars to see whether the deviation hits the five-sigma level. Another blinding (and unblinding) of clock values will accompany those runs.


Measuring a deviation to five standard deviations is just the first part of the story. “Say we release this number and it’s the exciting number the community is expecting—no one will believe us,” Charity said. “We have to be ready to defend everything we did. I’m not afraid of that, but it’s daunting.”

And of course, the result might reveal that the Brookhaven measurement was a statistical fluke all along, as occurred with the infamous “750 GeV bump” at CERN, when hints of a new particle at the Large Hadron Collider disappeared once more data came in. That would still be interesting, as it would rule out a bunch of potential ideas that theorists devised to explain the deviation in the first place.


“Either way, it’s really exciting,” Charity told me. But she hopes the final measured number is the one that contradicts the Standard Model, because it would challenge physicists to rethink what they have long assumed to be true about the universe.

Former Gizmodo physics writer and founder of Birdmodo, now a science communicator specializing in quantum computing and birds

Share This Story

Get our newsletter


Well, the scientific method has always been a way to eliminate human bias from our work. Its why an engineer can build something that *works* and not have it be “science” while a scientist can build something that doesn’t work and have that be “science”. Colleagues will confirm or refute a scientist’s findings by replicating the apparatus and methods. And the loop of testing and experimentation leads to a refined result that then is further refined as others continue to test it. The engineer might build something once, patent the design (or not and just build it) and never iterate or refine or get someone else to confirm the work. It doesn’t mean the engineer did something wrong. It just means its not *science*.

And sometimes in the most complex of things where being able to create a study where things are *right* is of the utmost importance, you have to work blind and see if you can derive the result because you understand everything and not because you feel something *should be* the way it works but because you have modeled the actual way it works. You’d be surprised at how few people can put aside their personal beliefs to even start working on this kind of thing. And even then, bias creeps in and well.... the best way to eliminate bias is to have many people converging on the answer and testing the method. Because while we all have biases we don’t all have the *same* biases and where one person might have a bias others will likely not. And that’s why its important that scientists share knowledge and expose their work to the community - even if it means they won’t personally profit from what they have created.