[time-nuts] IEEE Spectrum - Dec 2017 - article on chip-scale atomic frequency reference

[Tom Van Baak]

Mark,

> In the standards definitions that include "at sea level", the question these days is "which sea level?".

Chris,

> So does that mean e.g. NIST and BIPM need to measure the acceleration at
> their respective locations to within parts in 10^17 or 10^18 in order to
> compare their frequency standards?

Yes, all national timing labs do this to one degree or another. To operate and compare clocks at that level of precision you need to accurately know your geopotential, which is sort of like knowing the acceleration of gravity, or elevation.

But it's not one-to-one as you suggest. A 1 meter change in elevation corresponds to a frequency offset of about 1e-16. So for 1e-18 levels of performance you "only" need to know g, or your elevation to 1 cm accuracy.

> That seems not practical.

It is practical, and necessary, and really cool!

Here are some papers that will give you an idea how much work it takes to make clocks at the 1e-16 and 1e-17 level. I mean, it's not like you just throw some cesium atoms in a bottle, rub the lamp, and out comes a genie singing 9192.631770 MHz.

These two examples describe the complexity of a primary cesium standard:

"Accuracy evaluation of the primary frequency standard NIST-7", 2001
http://tf.nist.gov/timefreq/general/pdf/1497.pdf

"Accuracy evaluation of NIST-F1", 2002
http://tf.nist.gov/timefreq/general/pdf/1823.pdf

In the first paper, see especially tables 1, 3 and 4 for an idea of the corrections they must apply. You'll notice that the largest correction is gravitational. Therefore part of their job in making a primary standard is to measure gravity at the exact point where the cesium atoms operate. And yes, that gets you in the dirty world of what's underground, what mountains are nearby, where's the water table this week, what shape the earth really is, and the phase of the moon, etc.

These two examples describe the complexity of precisely measuring gravity in order to calibrate an atomic clock:

"The relativistic redshift with 3 × 10−17 uncertainty at NIST, Boulder, Colorado, USA", 2003
http://tf.boulder.nist.gov/general/pdf/1846.pdf

"A re-evaluation of the relativistic redshift on frequency standards at NIST, Boulder, Colorado, USA", 2017
http://tf.boulder.nist.gov/general/pdf/2883.pdf

Really, all four papers are worth a quick read, even if you just look at the tables and photos.

/tvb