Microwave frequency standards are key to dissemination of high accuracy time services to UK science, technology and industry, and to compatibility with international timescales through world-wide time transfer and clock comparison.

Caesium fountain primary frequency standards are used to realise the SI definition of the second. Currently. NPL has one operational caesium fountain (NPL-CsF1). This is used to contribute to the international timescale TAI and has participated in an international comparison of primary frequency standards. It has also been used as the frequency reference in measurements of the absolute frequency of an optical clock transition in a trapped ⁸⁸Sr⁺ ion.

Two other atomic fountains, one based on caesium and the other on rubidium, are being developed. This research effort is aimed at realisation of the SI second with reduced uncertainty and improved stability. Low phase noise microwave oscillators are also being developed to support this work.

Caesium fountain contributions to TAI

Contributions from a small number of primary frequency standards, including the NPL caesium fountain primary frequency standard, are used to generate steering corrections to the international time scale (TAI), which is maintained by the Bureau International des Poids et Mesures (BIPM) in Paris.

The measured frequency of the caesium ground state hyperfine transition is offset from that of the unperturbed transition frequency by several systematic effects. These include the second order Zeeman shift, the frequency shift due to collisions between the cold atoms, black body radiation, microwave leakage, etc.

The overall accuracy of our caesium fountain primary frequency standard NPL CsF1 (2 parts in 10¹⁵) is dominated by the uncertainty in the evaluation of the frequency shifts due to microwave leakage and collisions between cold atoms. The evaluation of the uncertainty budget for NPL CsF1 was reported in the 18^th EFTF proceedings and later published in Metrologia. This peer-reviewed publication is a requirement for valid contributions to the international atomic timescale TAI.

In the period February-May 2004 three campaigns were completed to evaluate the step interval of UTC and TAI. A campaign duration of 30 days was required to reduce the uncertainty of the two-way satellite time and frequency link to 1 part in 10¹⁵, similar to the level of accuracy of NPL-CsF1.

Comparison between caesium fountain primary frequency standards

NPL and several other National Measurement Institutes (including IEN, NIST, PTB and SYRTE) performed a comparison campaign of caesium fountain primary frequency standards for about 25 days in October to November 2004. In all laboratories, the fountains were compared locally to hydrogen masers. The masers were compared with each other using three different methods: two-way satellite time and frequency transfer and two different Global Positioning System based methods.

Following this campaign, the frequency differences between the three fountains of IEN, NPL and SYRTE were evaluated. The observed difference between the NPL and SYRTE fountains was below their estimated measurement uncertainty of 1 part in 10¹⁵. The IEN fountain deviated by about two standard deviations from the other two fountains.