National Physical Laboratory

Rb Fountain

NPL's Rubidium fountain frequency standard is designed to reach the highest frequency stability and accuracy among these types of standards.

Fig 1: Design of the NPL Rb fountain
Fig 1: Design of the NPL Rb fountain

The main advantage of Rb atoms is their small collision cross section at the low temperatures used in atomic fountains. The collision cross section of Rb atoms is approximately 30 times smaller than Cs atoms. This makes it possible to reduce the systematic frequency shift related to collisions between the Rb atoms. It also allows us to operate the Rb fountain with a larger number of atoms, which reduces the quantum projection noise of the standard.

The NPL Rb fountain consists of the following distinctive features:

  • The main magneto optical trap (MOT) in 111-configuration is continuously loaded by Rb atoms from the LVIS MOT [1] source of cold atoms, which loads about one billion atoms into the main MOT in just 100 ms.
  • The fountain uses an under coupled Ramsey microwave cavity, with a quality factor of 28500, which is very close to the ideal cavity (Q=32100) of the same geometry. This is intended to suppress the corresponding distributed cavity phase frequency shifts.
  • This is the first fountain in which the active control of the temperature of its interrogation region is performed by water pumped through the water jacket of the C-field tube. This makes it possible to tune the Ramsey cavity in resonance with the frequency of the measured clock transition and to control the Blackbody radiation frequency shift.

Accuracy and frequency stability of the standard

The first characterisation of the Rb fountain was performed in 2009 [2, 3]. In 2012, the distributed cavity phase frequency shift of the fountain was measured and calculated [4], which improved the accuracy of the fountain down to 2.4×10-16. The modification of the fountain in 2013 has further improved its accuracy down to 2.1×10-16.

The fractional frequency stability, which is currently limited predominantly by the noise of the local oscillator, is measured to be 7×10-16 after one day of averaging.

On the other hand, the intrinsic frequency stability of the Rb fountain standard is estimated to be about 1×10-16 after one day, which is still to be confirmed with a more stable local oscillator.

Measurements of the Rb hyperfine splitting frequency

There are multiple Rb clocks, including the clocks used in GPS systems, which depend on the measurement of the frequency of the ground state hyperfine splitting of Rb atoms.

Therefore, the precise knowledge of that frequency is essential.

In 2010, and more recently in 2013, the comparison of the frequency of the NPL Rb fountain to the NPL Cs primary atomic fountain standard CsF2 has given us a value for the ground hyperfine splitting frequency of:

NPL (2010): f=6 834 682 610. 904 308(5.2) Hz,

NPL (2013): f=6 834 682 610. 904 314(6) Hz - measurement still in progress [5]

which fall below the old CCTF recommended value of 2004, but agree well with the new CCTF recommended value of 2012, based on the LNE-SYRTE 2012 measurement:

LNE-SYRTE, CCTF Recommended (2012): f=6 834 682 610.904 312(3) Hz


Fig 2: Existent accurate measurements of the Rb ground state hyperfine splitting
Fig 2: Existent accurate measurements of the Rb ground state hyperfine splitting


  1. Compact magneto-optical sources of slow atoms
    Yu. B. Ovchinnikov
    Optics Comm, 249, 473-481 (2005)
  2. Accurate rubidium atomic fountain frequency standard
    Yuri Ovchinnikov and Giuseppe Marra
    Metrologia, 48, 87-100 (2011)
  3. Reply to comment on 'Accurate rubidium atomic fountain frequency standard'
    Yuri Ovchinnikov and Giuseppe Marra
    Metrologia, 48, 448-449 (2011)
  4. Development of NPL Rb fountain frequency standard
    Yuri Ovchinnikov
    Proceedings of European Time and Frequency Forum 2012, 96-100 (2012) [arXiv:1206.1158]
  5. Yu. B. Ovchinnikov, K. Szymaniec, S. Edris, to be published (2013)


For further information, please contact Yuri Ovchinnikov

Last Updated: 23 Sep 2013
Created: 7 Aug 2007


Please note that the information will not be divulged to third parties, or used without your permission