One good site in four: Vienna's thorium crystal aims a clock out of the vacuum chamber

Put thorium-229 into a calcium fluoride lattice at a position where the surrounding electric field is lopsided, and the nuclear transition stops being one frequency. Three of the four positions the Vienna Institute of Technology team probed sent light back at several wavelengths, which is the signature of that uneven field ^[2], and an uneven field is what pulls the clock's timing off ^[6]. The remaining site answered on a single wavelength ^[3], meaning its local environment is symmetric enough to leave the line unsplit. One usable site in four is a 25 percent hit rate ^[14], and it puts the gate on this device in the furnace rather than the laser bench.

The diagnostic also explains why the calendar reads the way it does. Each frequency point costs six minutes of instrument time, so an hour buys about ten points ^[13]. Schumm's own comparison of the work to cooking, or to alchemy, with repeated small changes until something comes out right ^[12], is an honest description of a search run at that rate over more than a decade.

The precision numbers deserve less deference than the crystal does. The account opens with a clock up to ten times better than existing ones ^[1], then reports that the group expects at least three orders of magnitude of improvement by year end ^[15]. Those two statements are a factor of 100 apart ^[16], and neither is tied to a named baseline instrument. What the team says it is actually chasing is not precision but size: getting the clock small enough to be deployed outside the lab as early as possible ^[15].

That ordering of priorities is the part worth taking seriously. Conventional atomic clocks read electron jumps between energy states and sit underneath satellite navigation and the tracking of financial transactions worldwide ^[7], but they buy their accuracy with vacuum chambers and shields that occupy substantial lab space ^[8]. A nuclear transition is already wrapped in the atom's own electron cloud, which screens stray electric and magnetic fields without external hardware, and a doped crystal runs at room temperature ^[9]. A prototype that could eventually fit on a small chip ^[17] is therefore not a better lab instrument. It is a clock that can be carried.

The dependency underneath all of it is the isotope. Flipping a neutron generally calls for a gamma-ray laser, and none has been built, which is why the work uses ultraviolet light on thorium-229 instead ^[10]. The precision case rests on that higher frequency, since larger energy jumps can be interrogated more often and so divide time more finely ^[11]. The report names no second isotope and no fallback excitation route, so every claim above is a claim about one nuclide in one host crystal.