In a dark matter detector, the gas acts like photographic film: a passing particle leaves a track of free electrons and ions. But how bright that track is depends on exactly how the particle loses energy in the gas. For light ions like fluorine, the data have been imprecise.

The Kanagawa group built a special chamber with wires, filled with CF4 gas at just 0.06 atmospheres. They sent a beam of fluorine ions with energies from 5 to 50 kiloelectronvolts into it. Each ion knocked out electrons as it traveled through the gas, and the chamber counted how much charge collected on the wires.

They found that at 30 keV, a single fluorine ion creates on average 0.45 ion-electron pairs per kiloelectronvolt of energy deposited. This number barely changes with energy, but drops at the lowest values — the ions start colliding more often with gas nuclei, wasting energy without ionizing.

Now detector developers have a reliable reference point: knowing the ionization yield, they can convert a signal into the particle's energy. It's like calibrating a scale before weighing — without calibration, any reading would be guesswork.