A scale of time based a frequency standard of 9,192,631,770 hertz obtained by comparing the outputs of a large number of “atomic clocks,” originally cesium-beam clocks but latterly including hydrogen masers and various types of fountain clocks (see below). This time is called International Atomic Time (abbreviated TAI, from the French). The frequency was chosen through astronomical observations¹, to make the second as nearly as possible equal to the ephemeris second, and since 1967 the second of atomic time has been the SI second, because the SI second was redefined to make it the atomic second. By definition, TAI equaled UT1 at 0h 0m 0s on 1 January 1958.
In 1964, the CGPM recognized the atomic second as a way to get the duration of the ephemeris second. In 1967, the 13th General Conference of Weights and Measures in Paris completely abandoned the second based on astronomical observations (at that point the ephemeris second), and resolved:
“That the unit of time in the International System of Units shall be the second, defined as follows:
“The second is the duration of 9 192 631 770 periods of the radiation corresponding to the transition between the two hyperfine levels of the ground state of the cesium-133 atom.”
With the atomic clocks of that time, the new definition increased the precision of the second’s definition by about four orders of magnitude.
The world’s standards laboratories constantly submit their atomic clock measurements to the BIPM for comparison. (Until 1985, this task was performed by the International Time Bureau (abbreviated BIH, from Bureau International de l'Heure).) From this data the BIPM computes a time scale called EAL (Echelle Atomique Libre). Frequency corrections from primary frequency standards are then applied to EAL, to keep the value of the second constant, and the result is TAI.
The stability of atomic time is based on a behavior of atoms: they must absorb radiation of a precise frequency in order to pass from one energy state to another. In the type of clock first adopted as a standard, the isotope cesium-133 is heated in a vacuum to produce a gas. Magnetic fields separate the atoms in the gas that are in a high energy state from those in a lower energy state, and those in a lower energy state are directed toward detectors some way off. On the way, they pass through a beam of radio frequency radiation from an oscillator controlled by a quartz crystal. If the oscillator is running at the frequency cesium-133 atoms can absorb, the atoms pass to the higher energy state. If the frequency of the oscillator drifts, the number of higher energy atoms reaching the detectors drops sharply. The oscillator frequency is constantly and automatically adjusted to maximize the number of high energy atoms reaching the detectors, and that stabilizes the frequency of the oscillator. The best clocks of this type lose or gain less than one second in several million years.
In 1955, the National Physical Laboratory in Britain constructed a cesium-beam clock and over the next three years compared its time with astronomical observations made by U.S. Naval Observatory in Washington, D. C.. From the comparison, they concluded that the clock's mechanism was keeping its oscillator tuned to a frequency of 9,192,631,770 cycles per ephemeris second. Turning this relationship around, one could say that 9, 192, 631, 770 cycles of the oscillator defined a unit equivalent in magnitude to an ephemeris second. This unit was named the atomic second.
Meanwhile the builders of atomic clocks had been using them to keep time at national services. The U.S. Naval Observatory, for example, started an atomic time scale it named A.1 in 1959. In 1961 the United States and United Kingdom synchronized their clocks, making the same step corrections. Two years later, the BIH started A3 on 1 Jan 1958.
On the basis of this work, in 1971 the 14th CGPM approved a time scale to be based on the atomic second and called International Atomic Time (TAI), leaving the definition of the scale to the CIPM, who had already prepared it.
W. Markowitz, R. G. Hall, L. Essen and J. V. L. Perry.
Frequency of Cesium in terms of Ephemeris Time.
Physical Review Letters, volume 1, page 105. (1958)
Greenwich Time and the Discovery of the Longitude.
Oxford University Press, 1980.
Explanatory Supplement to the Astronomical Ephemeris and the American
Ephemeris and Nautical Almanac.
H. M. Stationery Office, 1961.
D. D. McCarthy and J. D. H. Pilkington (editors).
Time and the Earth's Rotation.
Copyright © 2000 Sizes, Inc. All rights reserved.
Last revised: 29 March 2004.