Logo of the Physikalisch-Technische Bundesanstalt
Panoramic view of the clock hall at PTB with the four caesium clocks CS1, CS2, CSF1 and CSF2.

Atomic time scales

Atomic time scales

Until 1967, the second was defined as a fraction of a rather long time interval between two constellations of an astronomic object (sun in its zenith, vernal equinox) which were supposed to recur periodically. The present definition of the SI second via an atomic characteristic of the element caesium-133 allows the second to be reproduced as frequently as desired, and with much higher accuracy. Since the early 1960s caesium atomic clocks (Cs clocks) which reproduce the SI second within tolerances specified by the manufacturer have been commercially available. The relative tolerances first were of the order of 5·10<small>-11</small> and today are in the range of 5·10<small>-13</small>.

Time institutes usually operate several such clocks to avoid any interruption of the continuous realization of their time scales. By appropriate pooling of the data for such a group of clocks, a group time scale can be realized which is more stable and reliable than the time scales realized with a single clock. The algorithms used for this purpose depend on the number and the quality of the clocks involved. If one clock is added to or removed from the group, it will be ensured that the original rate of the time scale is maintained as accurately as possilble. The reason for this is that for many applications a small instability of the rate is more important than a certain value of the rate.

The free atomic time scales formed in this way are referred to as TA(k) where (k) identifies the operating time institute. Free atomic time scales are not significant for the dating of events but serve as a reference for the measurement of time intervals and rates.

In some national metrology institutes, so-called primary clocks were built which serve to realize the second with a specified small uncertainty. These clocks enabled the institutes to periodically calibrate their free atomic time scales. The NRC (National Research Council, Canada) and PTB were successful in operating their primary clocks continously over many years. Thus TA(NRC) and TA(PTB) could be directly derived from the primary clocks, and their scale unit was (and still is) the SI second as realized with the available primary standards. TA(PTB) has always been derived from one single clock, first CS1, then CS2. Since 2010 the scale unit TA/PTB) is determined by Caeium fountain clocks.