[time-nuts] Atomic clocks - Why alkali metals?

Magnus Danielson magnus at rubidium.dyndns.org
Sun Oct 30 15:46:50 UTC 2011

On 10/29/2011 08:19 PM, Attila Kinali wrote:
> Moin,
> Over the past years, i've read quite a bit on the different flavours
> of atomic clocks and how they are build. I found explenations for
> most questions i had, but for one thing: Why do all current designs
> (with the exception of trapped ion optical clocks) use alkali metals
> as reference? Is it because they can be aproximated as single electron
> systems due to the one electron in the valence orbit?
> Related to this is the question why only H, Rb and Cs are used.
> Although, from my point of view there isnt anything that speaks
> against using Li, K or Na, these are not used at all. At least i
> couldnt find any papers or other documents describing frequency
> standards build on these elements.

I did a search a few years ago and collected this list of frequencies:


Actually, Thallium was a contender for the primary reference, as it is 
known already back then to be less susceptible to magnetic pulling 
effects than caesium. Thallium beams would mean about 3 times higher 
frequency, so there was some question about the repeatability at the 
time. The magnetic susceptibility is less of an issue now with servoed C 

You will also find that there is plenty more available. The above is not 
a complete list, but as far as I got with the sources I had at the time. 
I'm sure that I could get better data and more isotopes if I dug around 


More information about the time-nuts mailing list