[Bogie_smiles (OP)]

Question: I think that an atomic clock uses a count of the frequency of the cesium atom to establish the rate that it measures the passing of time. Is that an electromagnetic pulse that is emitted by the cesium atom at a highly exact frequency?

[evan_au]

Atoms absorb radiation at exactly the same frequencies that they emit radiation.

But in a cesium clock, you need to select the cesium atoms so they are all in exactly the same quantum state at the start of the experimental procedure, or you will get conflicting results.

The clock radiates these selected atoms with that specific frequency, and checks that they are transformed into the other quantum state. So they actually detects absorption by the atoms, rather than the radiation from the atoms.

If the frequency is a bit off, there won't be as many atoms transformed into the other state. This dip in output is used to tune the operating frequency of the clock.
 
See more about the latest cesium clocks at: https://www.wired.com/2014/04/nist-atomic-clock/

[Bogie_smiles (OP)]

Let me test my understanding: The Cesium atom has many frequencies, but to work best for the atomic clock, the atoms have to be coaxed to a particular precise frequency. I’m trying to understand the term “oscillation frequency” as it applies to the frequency setting. Is that a single frequency at which the clock is set to function, or is it two frequencies between which the atoms oscillate as the atom changes state?

[evan_au]

The Cesium atom has many frequencies

Cesium vapor can absorb (and quickly re-radiates) photons at many optical frequencies, corresponding to electrons jumping between different shells.
- These photons have moderately high energy - enough to trigger your eyes to see light. But the frequency was too high for electronics to count (until fairly recently).
- These photons are typically generated by subjecting the vapor to a hot electrical discharge
- the high temperature causes the frequency to vary a lot between different atoms.

Cesium atoms have another way to absorb and reradiate photons, with the electron staying in the same (lowest-energy) shell
- The outer electron of a cesium atom can be considered like a little magnet
- The Cesium nucleus can also be considered like a little magnet
- There is a very small difference in energy between the state with these little magnets lined up in one direction or the other. These are called hyperfine energy levels.
- This is a very low energy signal, which could (in theory) be used close absolute zero.
- The low temperature means that all atoms give very much the same frequency: 9,192,631,770 Hz, in the microwave band.
- This frequency was fairly easy for a national laboratory to count with electronic circuits in the 1960s - and now the circuits in your cellphone would be able to count this fast.
- Because the energy is very low, this state can last for a long time - the longer time period you use to measure it, the more accurately you can count it's frequency

the atoms have to be coaxed to a particular precise frequency.

The capability to absorb this frequency is inherent in the structure of the cesium electron structure.

To get the strongest possible signal, you want all of the atoms ready to absorb this frequency, so they sort the atoms using magnets or lasers, to discard any that don't have their electrons in the lowest-energy "ground" state.

I’m trying to understand the term “oscillation frequency” as it applies to the frequency setting.

Rather than try to detect the very weak radiation of a smallish number cesium atoms emitting a frequency of 9,192,631,770 Hz (a very difficult measurement), they take the opposite approach and bombard the cesium atoms with an intense signal at 9,192,631,770 Hz.

To produce this intense signal, the cesium clock contains a fairly accurate electronic circuit which oscillates at approximately 9,192,631,770 Hz.
- The clock then measures what percentage of the cesium atoms' miniature magnets have been "flipped" into the opposite state
- If it is a high percentage, the local oscillator frequency must be exactly 9,192,631,770 Hz

Is that a single frequency at which the clock is set to function

- To know that this is the highest possible percentage, the local oscillator frequency is varied a tiny bit either side of this center frequency, to ensure that it has reached the highest possible percentage
- So it is a small range around a center frequency
- By measuring the average frequency of this high power oscillator, it is easy to count off the number of oscillations in a second - and any other multiple.

So a fairly accurate oscillator circuit is made into an extremely accurate oscillator by using the precise hyperfine absorption frequency of the cesium atom.

or is it two frequencies between which the atoms oscillate as the atom changes state?

The cesium atom will absorb a photon at 9,192,631,770 Hz when it moves from the lower to the upper hyperfine state.

The cesium atom will emit a photon at 9,192,631,770 Hz when it moves from the upper to the lower hyperfine state. But as noted before, measuring this tiny signal is too hard, so the atoms initially in the upper state are discarded before the measurement even starts.

See: https://en.wikipedia.org/wiki/Atomic_clock

Footnote: These days, it is getting easier for a national laboratory to count optical frequencies, and the latest and greatest atomic clocks use optical frequencies, rather than microwave frequencies.

[Bogie_smiles (OP)]

Thank you for the good explanation. I'm getting it I think. Quite amazing where technology has gotten, and so fast.

[evan_au]

Talking to a friend who works with National time standards...

The two hyperfine levels (at zero magnetic field) actually split into 16 hyperfine levels in a magnetic field, dictated by two quantum numbers F and m_F).
There are always some magnetic fields around, so you want to select the ones which are least affected by a magnetic field.
So actual devices apply a small magnetic field (called a "C-Field") which splits all of these levels off to different frequencies.
Then the local oscillator just operates on the two that are least affected by the magnetic field:
- F= 4 and m_F= 0
- F= 3 and m_F= 0

Thus the local oscillator is actually set ≈1.5Hz higher in frequency than the SI defined 9,192,631,770 Hz.

The C-Field can be used for fine frequency adjustments, as fine as 6 parts in 10^-15 for recent models (eg HP 5071a).

For a lot more details, see pages 17 & 18 of: http://www.hpl.hp.com/hpjournal/pdfs/IssuePDFs/1973-09.pdf

@alancalverd, can you advise the significance of quantum numbers F and m_F?