2 Basics of Time and Frequency Standard 2-1 Definitions of Time and Frequency Stan- dard MORIKAWA Takao The definition of most basic concepts in the time and frequency standards field, time and timescale, depends on the science and technology level of the times and is never immutable. In this paper, the historical change in the definitions of time and timescales such as International Atomic Time and Coordinated Universal Time are reviewed. Keywords Time and frequency standard, Standard time, International atomic Time, Coordinated universal time, International system of units 1 History of time and frequency role in many high-technologies, such as satel- standards lite positioning technology (exemplified by the GPS), synchronization in high-data-rate Time―together with its inverse quantity, digital communication networks, and preci- frequency―stands as one of the most basic sion measurement applications. physical quantities of daily life, included among quantities such as length and mass. 2 Establishment of standard time Since ancient times, humans have attempted to systems measure time more and more precisely, rely- ing on a variety of physical phenomena Humans have developed precise time- observed in nature. It is well known that the measurement technologies by applying the calendar was developed in ancient Egypt best scientific knowledge available in each through celestial observations, a result of the era. Meanwhile, the definition of time itself need for precise time in agriculture, the basic has grown in importance since the 19th centu- industry of the time. As overseas trade devel- ry, with improvements in measurement accu- oped in Europe in the 17th and 18th centuries, racy and the evolution of economic activities. the establishment of a technology to determine In Europe in the 18th century, particularly in longitude precisely, needed for safe and effi- Britain, mechanical clocks became popular cient navigation, became one of the greatest among the middle class. At that time, there challenges for sea-fairing nations. Major was no such thing as standard time; as a refer- European countries set up astronomical obser- ence, noon was determined as the moment the vatories, including the Greenwich Observatory sun crosses the meridian in a given local area. and the Paris Observatory, and encouraged the This method posed no real problems, as development of accurate clocks, in part by human activities were generally limited to offering generous rewards for success. The individual local areas. However, in the 19th marine chronometer was one of the products century, as railways developed and people and of such policies［1］. In today's advanced scien- goods moved more quickly over long dis- tific world, the atomic clock plays a crucial tances, time differences among local areas MORIKAWA Takao 3 became a problem. In response, the London the Convention of the Metre (Convention du and Northwestern Railway determined a stan- Métre) of 1875. The Comité International des dard time in 1848, based on the time in Green- Poids et Mesures (International Committee of wich, England. As this standard railway time Weights and Measures, or CIPM) was formed gained acceptance, Greenwich Time became below the governing organization Conférence adopted by law in 1880 as British standard Générale des Poids et Mesures (the General time［2］. At the same time, discussions were Conference on Weights and Measures, or taking place regarding an international stan- CGPM, which met every four years). The dard time, and in October 1884, the Interna- Bureau International des Poids et Mesures tional Meridian Conference was held in Wash- (International Bureau of Weights and Mea- ington D.C. in the United States to discuss sures, or BIPM) was then established to con- world standard time. The following points duct various metrological activities under the were agreed upon［3］［4］. control of the CIPM. After 1875, a number of ① A single prime meridian for all nations metrological units, including time, were would be adopted in place of multiplicity of defined internationally within the framework initial meridians. of the Convention of the Metre. In Resolution ② The adoption of the meridian passing 12 of the 11th CGPM, 1960, the International through the center of the transit instrument at System of Units (SI unit) was established, and the Observatory of Greenwich would be pro- remains in use to this day［5］. SI units can be posed to member countries as the initial classified into basic units and derivative units, meridian for longitude. combinations of basic units. This method of ③ From this meridian longitude shall be count- dividing units into two classes is not uniquely ed in two directions up to 180 degrees, east determined by physics but is in fact somewhat longitude being plus and west longitude minus. subjective. Nevertheless, from viewpoints of ④ A universal day would be adopted. This international exchange, education, and would not interfere with the use of local times. research, there were many advantages to the ⑤ The universal day would be a mean solar establishment of a single, practical, interna- day, begin for all the world at the moment of tional unit system. Considering these advan- mean midnight of the initial meridian, and be tages, the CGPM has adopted seven clearly counted from zero up to twenty-four hours. defined units (the meter, the kilogram, the sec- ⑥ The astronomical and nautical days should ond, the ampere, the Kelvin degree, the can- start at midnight. dela, and the mol, the last of which was added ⑦ Angular space and time should be expressed to the original basic units in 1971) and 27 by the decimal system. derivative units (including frequency). How- Based on these resolutions, Japan issued ever, although the official definitions of SI Imperial Edict No.51, entitled "Initial Meridi- units are authorized by the CGPM, these defi- an Longitude Calculation Method and Stan- nitions are not immutable. They reflect the dard Time" on July 13, 1886［4］, the first estab- progress of science and technology in each era lishment of a standard time in Japan. At this and are subject to corresponding revision. In time, the second was defined as 1/86,400 of a this context, in 1927 the CIPM established a mean solar day. committee to discuss a number of issues, including revisions to the definition of the var- 3 The Convention of the Metre ious units. In particular, the group thus and definition of the second formed―the Consultative Committee for Time and Frequency (CCTF)―convened to discuss In the late 19th century, an important event the definition of the second. As mentioned took place with respect to the measurement of earlier, the second was initially defined as length, a physical quantity as basic as time: 1/86,400 of a mean solar day. However, astro- 4 Journal of the National Institute of Information and Communications Technology Vol.50 Nos.1/2 2003 nomical measurements had indicated that the "International atomic time (TAI) is the irregularity of the Earth's rotation affected the time reference coordinate established by the accuracy of time measured in this manner. Bureau International de l'Heure on the basis of Thus the 11th CGPM, 1960 adopted the follow- the readings of atomic clocks operating in var- ing definition of the second［5］, based on the ious establishments in accordance with the solar year as defined by the International definition of the second, the time unit of time Astronomical Union. of the International System of Units." "The second is the fraction 1/31,556,925. The definition of TAI was revised as fol- 9747 of the tropical year for 1900 January 0 at lows in 1980 based on the theory of relativity 12 hours ephemeris time." ［5］. This definition, however, posed a problem "TAI is a coordinate time scale defined in in that long-term astronomical observations a geocentric reference frame with the SI sec- would be required to provide accurate meas- ond as realized on the rotating geoid as the urements of the second. Meanwhile, the Cs scale unit." atomic frequency standard, developed in the We cannot ignore the effects of relativity rapidly evolving field of microwave spec- when considering precise time in the context troscopy, was found to be able to determine of modern science and technology. The the length of a second with a much higher effects of relativity on space-time have been degree of accuracy. From 1955 to 1958, the discussed in detail elsewhere［7］. Note that British NPL and the American USNO jointly TAI is designated as starting at 0 sec 0 min 0 measured the frequency of a Cs frequency hr on January 1, 1958 of UT1. Until 1969, standard based on ephemeris time［6］. Using TAI was provided by averaging the times of the results of this measurement, the definition the atomic clocks in what was then the Bureau of the second was revised as follows by the International de l'Heure. Since 1969, TAI has th 13 CGPM in 1967-1968［5］. been provided by averaging the atomic clocks "The second is the duration of 9,192,631,770 of research institutes in many countries, in periods of the radiation corresponding to the order to improve the reliability and short-term transition between the two hyperfine levels of stability of TAI［8］. Most of the clocks used to the ground state of the cesium 133 atom" determine TAI are commercial Cs atomic The above definition assumes that all the clocks with accuracies of approximately 10-12. Cs atoms are in the ideal state, free of any Thus to increase the accuracy of TAI, frequen- physical disturbance. This definition was cy calibration using a primary frequency stan- epoch-making in the sense that for the first dard becomes very important. As a case in time in the history humans adopted the atomic point, in 1976, all the primary Cs frequency time in stead of astronomical time, the latter standards of the U.S.