Seeds of time from the cesium atom - A new standard in atomic clocks

Science Dimension volume 7 issue 1 April 1975

Fleming .
Short wave listeners from all parts of the world who receive the CHU time signal send reaction reports to NRC. These are welcomed and acknowledged. During the 19th century, the development of the railway and the telegraph was followed by the establishment of standard time zones. Sir Sanford Fleming of Canada advocated uniform zones on an international scale. By his plan, the world would be divided into twenty-four zones, each fifteen degrees wide, the first one centred on the Greenwich Meridian. Within each zone, the time would be the same, and the boundary would mark the place where the time would change abruptly by one hour. In practice today, the time zone boundaries tend to conform to national or geographical divisions, so that the divisions differ considerably in shape and size. (National Archives)

Over the years there have been countless descriptions of time. A file that wears and makes no noise. The arbitrary division of eternity. A sandpile we run our fingers in. The life of the soul.

For centuries, men have sought the most accurate means to measure and monitor time. Their search has ranged from tracking the regular motions of heavenly bodies to probing the energies of atoms.

A significant advance in atomic timekeeping has lately been made in the Time and Frequency Section of the National Research Council of Canada's Division of Physics. A new cesium beam standard, Cs V, will soon begin operation as the world's most accurate clock.

Years ago, the measurement of time was based on the period of the earth's daily rotation. When this proved non-uniform, astronomers adopted the period of revolution of the earth around the sun for the year 1900 as the basis for the measurement of time. This "Ephemeris Time" now forms the basis for many astronomical calculations and forecasts but is not readily available for day to day use.

Similarly, man-made devices for recording and keeping track of time have evolved through sun dials, hourglasses, water clocks, mechanical clocks, pendulum clocks and quartz crystal oscillators. Today, time measurements depend on the cesium atomic clock whose frequency has been measured carefully in terms of Ephemeris Time. Neither atomic nor Ephemeris Time vary with the seasons or the years.

The "atomic" clock is so named not because it is powered by atomic energy but because certain fundamental properties of the atom are used to provide a definition of time.

In a magnetic field, atoms of the element cesium 133 can adopt either of two energy levels corresponding to differences in total angular momentum. Additionally, each level contains a hyperfine structure of closely-spaced sub-levels. In practice, a transition or Rabi resonance can be excited from the lower to the upper level, between two particular hyperfine levels, at a frequency of 9,192,631,770 cycles per second (Hz) in the microwave region. The international (SI) second was thus defined in 1967 by the International Bureau of Weights and Measures as the duration of 9,912,631,770 periods of the radiation corresponding to this transition.

Early experiments on cesium beam standards were conducted in the mid-1950's at the National Bureau of Standards in Washington, D.C., and the National Physical Laboratory in Teddington, England. Activity at NRC began in 1957 with the construction of the cesium beam atomic standard Cs I. In 1963, an improved primary frequency standard, Cs III, was developed and served to calibrate first quartz crystal oscillators and later commercial cesium beam secondary clocks when atomic timekeeping activities started in the following year. NRC also monitored the rates of a group of atomic clocks maintained by the Dominion Observatory, which was then responsible for official time in Canada.

In 1970, the NRC group assumed the entire timekeeping function for the country and thus became responsible not only for research and development of atomic time and frequency standards but also for the generation and dissemination of official time. The staff and equipment from both the Observatory and NRC were combined to form a new section within the Physics Division.

A battery of commercial (secondary) atomic clocks, calibrated and regulated by Cs III, became the basis of the Canadian time scale. Frequent calibrations were necessary since these secondary clocks, with stability and accuracy much poorer than for a primary standard, all tended to run either fast or slow.

There is an important distinction between clocks and frequency standards. Clocks run continuously and generate a time scale indicating total elapsed time from an arbitrary zero. On the other hand, frequency standards such as Cs III are set in operation only at intervals in order to check the rate or frequency of continuously running clocks. In an analogous way, a wristwatch can be adjusted periodically to the time kept by a more accurate electronic timepiece.

To combine in one instrument the accuracy of a primary frequency standard with the capability of continuous operation in a clock generating its own time scale, plans were initiated at NRC in 1970 to construct an improved standard, Cs V.

Dr. Allan Mungall of the Time and Frequency Section was responsible for the overall design and was assisted by Mr. Herman Daams and Mr. Ralph Bailey (now retired).

Since its construction in 1973, Cs V has functioned experimentally as a frequency standard, but the changeover to continuous operation will make it the most accurate and stable clock in the world. By contrast, the primary cesium standards of the other countries are not intended for use as clocks but as frequency standards to measure the mean rate of an ensemble of commercial atomic clocks.

"I feel it is far more important to build a good, stable primary clock than to take an average of a large number of less stable secondary clocks," Dr. Mungall says. "Other countries use elaborate averaging procedures to arrive at a mean time scale but I've always felt that, physically, this is the wrong approach."

Although an ensemble of secondary clocks can provide a fairly stable rate for long periods, the component clocks tend to be of identical design and consequently all eventually drift in the same direction. Inevitably, small frequency shifts arise from aging of electronic components, changes in magnetic shields or a build-up of cesium metal in their cavities. As a result, frequent recalibration by a stable primary reference becomes necessary.

Dr. Mungall concludes: "We think our system will ultimately produce a better time scale, will require fewer people to run it and will cost much less to operate."

The 4 m (13 ft) long Cs V looks nothing like a conventional clock. For its accuracy it depends not on a pendulum or mainspring but on a highly uniform magnetic field in the cesium transition region.

A 5 g (0.2 oz) charge of metallic cesium is initially heated to 80C (176F) in a stainless steel oven situated at one end of Cs V. At this temperature, normally solid cesium melts to a dense silvery liquid with sufficient vapor pressure to produce a beam of free cesium atoms which is then shot between the poles of a state selector magnet. Here the atoms are separated into two beams corresponding to different energy levels.

Both beams then pass through a long evacuated space maintained at a uniform low magnetic field (the C field) in which the hyperfine atomic splitting is maintained.

Two different types of atomic transitions can then occur. In one case, a low frequency transition can be excited in the cesium beam by a series of axially-oriented coils. By monitoring this transition, scientists can determine the magnitude and uniformity of the C field throughout the magnetically shielded region.

In the second case, which corresponds to actual clock operation, the cesium beam is subjected to a source of microwave power (called a Gunn diode oscillator) which excites the atomic transition. The closer the oscillator frequency is to the atomic resonance frequency the more atoms in the beam will undergo transition.

The beam then passes from the microwave and C field region through a second state selector magnet onto a platinum-iridium hot-wire detector. Here, the arriving cesium atoms are ionized, then collected as an electric current.

The amplitude of this current serves to indicate what proportion of the cesium atoms have been excited and, in turn, how close the microwave exciting frequency is to the precise cesium atomic transition frequency. (The microwave frequency is produced from a 5 MHz (5Mc/s) quartz crystal oscillator by a series of electronic multipliers and synthesizers).

Changes in the cesium ion current serve as an indication of error and a feedback signal acts to tune the frequency of the crystal oscillator. Thus the microwave and crystal oscillator frequencies are locked to the cesium atomic frequency which is the basis of the international (SI) second.

"In effect, we are interrogating the atoms to find if our frequency is correct or not," explains Dr. Mungall.

Once the correct frequency has been established and remains stable, the output frequency from Cs V can be counted down by a series of electronic dividers and used to generate regular pulses at one second intervals which appear as numbers on a digital clock - precise atomic time.

Although the design of Cs V is similar to its predecessor, Cs III, certain modifications lead to a tenfold improvement in accuracy and stability.

For example, cesium oven and detector units are mounted at each end of the instrument making bi-directional beam operation possible. By averaging the results from operation in both directions, scientists can compensate for any differences in phase between the two microwave fields used to excite the atomic transitions.

"We want to realize the primary standard definition as accurately as possible," notes Dr. Cecil Costain, head of the Time and Frequency Section, "so that measurements have meaning not only today but in 50 or 100 years."

Scientists expect Cs V to maintain an accuracy of one part in 1013. In other words, the atomic clock would be off by no more than three seconds in one million years.

The only other instrument of comparable accuracy is the hydrogen maser, a frequency standard based on the hydrogen atom as a microwave oscillator. Because of a greater short-term (up to 1000 seconds) frequency stability, two NRC-built masers can be used periodically to evaluate Cs V. However, these instruments are currently less suitable as clocks since their performance gradually deteriorates, showing a frequency shift over a longer time.

With Cs V, NRC will bring substantially improved accuracy to its dissemination of time and frequency to outside users in Canada and to its participation in international time comparisons with other countries through the Bureau International de l’Heure (BIH).

. Mungall
Cs V will begin operation shortly as the world's most accurate clock in the Time and Frequency Section of NRC's Division of Physics and will form the basis for Canada's official time scale. Dr. Allan Mungall (right) was responsible for the overall design and was assisted by Mr. Herman Daams (left) and Mr. Ralph Bailey. (NRC)

The NRC time scale is also compared regularly with those of other countries through reception at the main time laboratory of the 100 kHz (100 kc/s) pulsed radio signals of the Loran-C navigation system. The Loran chain, a series of stations which extends across the United States, spans the north Atlantic and reaches into Norway and Europe, is operated by the United States Coast Guard and is closely related to the time scale of the U.S. Naval Observatory (USNO). It provides the most convenient regular communication network between various national time laboratories.

In addition to the use of Loran-C, time comparisons using television signals provide links between NRC, the USNO and the National Bureau of Standards in Boulder, Colorado, and also between the NRC time laboratory and the independently operating clocks which provide the standard frequency and time signals transmitted by station CHU in Ottawa.

Although CHU generates its own signal via a secondary cesium standard (installed in 1963), daily cross-comparisons are made with the main time laboratory to ensure that the station's frequency and time conform to the standards agreed upon internationally. Continuous English and French voice announcements are transmitted over three short wave operating frequencies, 3,330, 7,335 and 14,670 kHz (kc/s).

Mr. Sidney Sheard, who is responsible for CHU's operation, notes: "Our time service via radio is available to anyone possessing even the most inexpensive short wave receiving set. Present-day users range all the way from interested listeners to government, scientific and industrial organizations for whose operation precise time is essential."

Time is also disseminated by telephone line from the main laboratory to other government departments. The telephone signal provided to the Canadian Broadcasting Corporation is retransmitted daily at 13:00 EST over the trans-Canada English network and at 12:00 daily on the French-language system.

In addition, many new methods of distribution are currently being investigated. These include dissemination of time by satellite through television broadcasts, and by high speed digital data transmission lines for aircraft collision-avoidance systems.

Another service planned for the near future is serial digital time code transmission over standard telephone lines. By this system, a user anywhere in Canada who owns a commercial digital clock can dial a given telephone number to place his unit in contact with a code generator clock at NRC's time and frequency laboratory. An electronic time code provided by NRC then acts by a data link through the telephone receiver to correct the commercial clock's time automatically to within one millisecond.

Dr. Costain foresees many applications for this service, particularly in the area of control voice-communication networks such as used by many police forces. In this system, audio tapes can be legally indexed with a time readout which would be displayed digitally on a suitable receiver to accompany a playback of the voice.

Another possible use is a digital display of time on home television receivers.

"I believe this is the best way to distribute time accurately and conveniently to the public," Dr. Costain says. Modern timekeeping seems undoubtedly more efficient than timekeeping in Shakespeare's day: "It shall be what o'clock I say it is." (Petruchio, Act IV, Scene II, The Taming of the Shrew). For Petruchio, determining the hour of day was usually a matter of guesswork. However, with the advent of precise modern atomic clocks such as Cs V, these words now describe the certainty of NRC's role in providing official time for Canada.

Reprinted courtesy of National Research Council of Canada