The same ion, two separate rigs
At the Centre for Quantum Technologies, K. J. Arnold and colleagues compared two room-temperature 176Lu+ single-ion references by correlation spectroscopy. Each has an evaluated fractional frequency uncertainty near 1 × 10−19.[1]
The measured relative frequency difference is [−0.1 ± 5.7 statistical ± 1.0 systematic] × 10−19. Two independent systems of the same species agree at the 5.7 × 10−19 level. The test fills a same-species comparison that had remained outstanding below 10−18.[1]
What the agreement still leaves open
The uncertainty interval on the difference includes zero. That shows the separate systematics of the two rigs do not diverge at the 5.7 × 10−19 scale. A shared drift could still sit in an effect that correlation spectroscopy applies to both, in which case the agreement speaks to a joint shift rather than to the absolute frequency.[1]
A millimetre on the scale of the second
The paper presents this accuracy as ground for international timekeeping, millimetre-scale chronometric levelling, and tests of Lorentz invariance, general relativity, dark matter and variation of fundamental constants. The Nature article says the clock would take more than 260 billion years to lose a second. A new SI second is not adopted by this comparison. The next observable sign is whether a third 176Lu+ reference, built outside this pair, returns the same frequency.[1]