Re: OT: Atomic nucleus excited with laser: a breakthrough after decades

They state a centre frequency of roughly 2 PHz and a decay time of 630s, which would put the Q in the 1e19 ballpark. Prodigious. No wonder it was hard to find.

Jeroen Belleman

"the correct energy of the thorium transition was hit exactly, the thorium nuclei delivered a clear signal for the first time. "

I wonder what that signal was.

Presumably the thorium nucleus absorbs the photon, then remits it when it decays back to the ground state, presumably not in the original direction.

The life-time of the excited state is 630sec when the thorium atoms are presented in a CaF2 crystal. It you hit the crystal briefly with precisely the right frequency, then observed a slowly decaying fluorescent signal at the same wavelength, you'd have a clear enough signal (though not all that much of it).

In fact they gradually stepped up the exciting beam wavelength from

148.2 to 150.3 nm.,and observed a fluorescence peak at around 148.38 nm.

The observed central wavelength of the nuclear transition amounted to

148.3821(5) nm, equivalent to a transition energy of 8.35574(3) eV, which was consistent with the 1 σ-uncertainty of the value reported in radiative-decay experiments but with 800-fold improved precision.

The implication is that their excitation wavelength wasn't all that precise either and will need to be made even more precise for nuclear clock work.

I wonder if they could use it to get Doppler shifts from continental drift?

It says so in the paper: Fluorescent UV light.

Jeroen Belleman

The Time guys have been looking for this forever, so to speak.

It's the only atomic kernel transition with any degree of coupling to electromagnetic radiation. This will be orders of magnitude better than such as lattice clocks.

There will be a flood of papers.

Joe Gwinn

They aren't tuning to a resonance, but to the difference between two close resonances.

Probably not. The technique to used to generate very precise laser wavelengths does seem to be difficult and demanding to work with.

The few people who can do it will have a field day, but they will only generate a few papers - it takes time to do the work and more time to write it up.

Nuclear energy levels aren't "resonances" but quantum states, and the transition between them isn't a "resonance" either, though one can talk about the kind of resonance that would behave in a similar way.

The current definition of the second uses something similar: Some hyperfine resonance of cesium. Normal resonances are in the optical domain, but hyperfine ones are RF.

In nuclei, normal transitions are in the gamma domain, and hyperfine ones are in the domain of optics. It's just a change of scale, if you will.

Jeroen Belleman

Which puts them in the RF frequency domain where counting cycles of the continuous sine reference waveform is relatively easy.

Likewise for H-maser another favourite local time reference signal.

Although there will be some big practical difficulties counting cycles of a waveform at 8eV which is up into the UV. What is the current highest frequency that a semiconductor divider is capable of accepting?

I know that there are some optical logic circuits about but how capable are they at near UV light?

You can't mix this thing down without losing its fidelity. I know how to double optical frequencies but how do you halve or quarter them?

Something involving optical frequency combs might work.

Jeroen Belleman

I don't know if there is a way to divide a lightwave-sorts of frequency down into the electronic domain. Much less gamma ray frequencies.

Even the small differences cited here are still optical.

Yes, it will be combs and etalons.

I'm waiting for a flood on the Time Nuts reflector.

Joe Gwinn

Am 08.05.24 um 17:22 schrieb Joe Gwinn:

I have posted already a pointer to this thread here on time nuts.

:-) Gerhard dk4xp

Is there any way to divide a lightwave down into the electronic frequency domain?

Rubidium clocks use an indirect way that doesn't actually divide.

Not to my knowledge. The usual way is down-mixing. The optical frequency comb provides a way to generate an accurately known optical local oscillator, so to speak.

Jeroen Belleman

[snip]

Hmm. It had to be true, but I never connected the dots there. What is mechanism by which this is achieved? References?

Thanks,

Joe Gwinn

Not really. There are optical parametric oscillators, but their phase noise is horrible by comparison. A 1-cm-long crystal produces a nice tunable output, but its line width will be c/1cm wide.

Degenerate OPOs exist, whose signal and idler are at the same frequency, but I believe their phase noise is not that different—there’s an additional degree of freedom in the signal/idler relationship that would have to be constrained somehow.

Cheers

Phil Hobbs

Don’t have the reference handy, but the basic idea is to use a modelocked system Ti:sapphire laser at 750 nm to generate ~100-fs pulses, then use fiber/grating pulse compression to bring that down to a few femtoseconds, followed by a holey fiber to broaden the spectrum to more than an octave.

Jan Hall is one of the best instruments guys ever.

Cheers

Phil Hobbs