Download Understanding the SI Second: From Astronomical Time to Atomic Clocks and more Summaries Physics in PDF only on Docsity!
Special Issue / PTB-Mitteilungen 122 (2012), No. 1 Time – the SI Base Unit “Second” n
Structure
1 Introduction 2 The Definitions of the Unit of Time 3 Realization of the SI Second 4 Atomic Time Scales: TAI and UTC 5 Time and Frequency Comparisons 6 Dissemination of Time for Society 7 Time and Fundamental Questions of Physics 8 Closing Remarks
1 Introduction
“Time is a strange thing ...”^1. In fact, the SI base unit “the second” has a special position among the units: It is the SI unit which has been real- ized by far with the highest precision – and this is why other base units are defined or realized with reference to it. The metre, for example, has been a derived unit since 1983, defined as the length of the path travelled by light in vacuum during a time interval of 1/ 299 792 458 of a second. The realization of the volt – the unit of the electrical voltage – makes use of the Josephson effect, which links the volt with a frequency via the ratio of two fundamental constants, h/(2e) (e: elementary charge, h: Planck’s constant). All this will be dealt with in the various articles of this publication. But there are other remarkable facts: The only meas- uring instrument most people have on them in daily life is a watch. Time touches every person every day (at least in our civilization). For decades, indicating the time – an essential task for everyday life – has been the privilege of the authorities in town and in the country, and doing so is associ- ated with prestige and has a direct influence on the life of the people. The transition from astronomic time determination to time determination on the basis of atomic clocks – which will be the subject of Section 2 of this article – has, therefore, been associated with all kinds of disputes in the scien- tific circles involved. In countries where the legal responsibilities are less clearly regulated than in Germany (in Germany, the regulation is based on the Units and Time Act), the rivalry between the interest groups continues to this day. The work which was started at the Physikalisch-Technische Reichsanstalt (Imperial Physical Technical Institute
- PTR) in the early 1930s and continued at PTB in Braunschweig after 1950 has given an essential impetus to this transition. In the following, a short survey of earlier defini- tions of the unit of time will be given, followed by sections dealing with the current definition and, thus, with the caesium atomic clock and international cooperation in the field of time measurement. Section 6 will then deal with the popular services which PTB makes available for the dissemination of legal time in Germany. To this day, research in the field of clock develop- ment is undertaken intensively. A research group announcing that it has achieved progress towards an even more exact clock can be sure to find the interest of the media. In Section 7, an answer will be given as to whom this will benefit.
2 The Definitions of the Unit of Time
2.1 The rotation of the Earth as the measure of time People’s natural measure of time is the day. It is defined by the rotation of the Earth around its axis. Following an old cultural tradition, it is subdi- vided into 24 hours, each comprising 60 minutes, with each minute comprising again 60 seconds [1, 2]. If one assigns the moment 12 o’clock to the zenithal culmination of the Sun, the true solar day is obtained as the period of time in-between. Due to the inclination of the Earth’s axis relative to the plane of the Earth’s orbit around the Sun, and due to the elliptical Earth’s orbit, its duration changes during one year by up to ± 30 s. Averaging over the length of days of 1 year leads to the mean solar time, whose measure is the mean solar day dm. Until the year 1956, its 86 400th^ part served as the unit of time, the “second”. It was realized with the aid of high-precision mechanical pendulum clocks and, later on, with quartz clocks, to make – on the one hand – time measures and – on the other hand
Time – the SI Base Unit “Second”
Andreas Bauch* (^1) Hugo von Hoff- mannsthal, Libretto to the opera “The Rose Cavalier”
- Dr. Andreas Bauch, Working Group “Dis- semination of Time” e-mail: andreas.bauch@ ptb.de
n (^) The System of Units Special Issue / PTB-Mitteilungen 122 (2012), No. 1
- the unit hertz (Hz = 1/s) available for frequency measurements. The mean solar time related to the zero meridian passing through Greenwich is called Universal Time (UT). If time measures and time scales are derived directly from the recordings of the moments of star passages through the local meridian, these are falsified by the changes in the position of the axis of rotation in the Earth’s body, the so-called “polar motion”. This polar motion occurs, in particular, at a period of approx. 14 months and, in addition, accidentally. In the time scale called UT1, the predictable periodic portions have been eliminated so that UT1 is (almost) proportional to the angle of rotation of the Earth. The recordings of the planets’ positions and of the Moon’s orbit relative to the starry sky – dated in UT1 – collected over many years – revealed, however, that the duration of the mean solar day, too, is subjected to con- tinuous changes [3]. There is, on the one hand, a gradual deceleration of the Earth’s rotation by tide friction. 400 million years ago, for example, the year had 400 days. Mass shifts inside the Earth’s body and large-scale changes in the sea currents and in the atmosphere lead, on the other hand, to a change in the moment of inertia of the rotating Earth and, thus, in its period of rotation. These effects which have been – partly – understood and are – partly – unpredictable, conceal the gradual deceleration. And as regards proof of the result- ing seasonal variations – this is a topic we will still have to talk about! 2.2 Time and frequency at the PTR and in the early days of PTB If we compare the quartz clock shown in Figure 1
- which was developed at the PTR in the 30s of the last century – with what we are wearing today as wrist watches, it becomes clear why we speak of a “quartz revolution” [4]. Although the contribu- tion of the PTR – and later that of PTB – to this modern development has been comparably small, the success of the PTR was, however, of essential importance for the abandoning of the astronomic time determination. The motivation for the activi- ties at PTR was the emerging of wireless com- munication traffic and the necessity of measuring the transmitter frequencies in the kHz and MHz range. For this purpose, E. Giebe and A. Scheibe developed frequency standards in the shape of luminous resonators: small quartz rods equipped with electrodes in glass flasks filled with a neon- helium gas mixture which were excited to a glow discharge as soon as the frequency of an electrical a. c. field at the electrodes agreed with one of the mechanical resonance frequencies of the tuning fork [5]. The relative measurement uncertainty was between 10–6^ and 10–7. Maintaining the form of the tuning forks, these were later integrated into an electric oscillator circuit, as had been shown before in a similar way by W. G. Cady, G. W. Pierce and W. A. Marrison [6]. Until 1945, a total of 13 quartz clocks had been developed at the PTR. Their rate stability, which was superior to that of pendulum clocks, motivated the Geodätisches Institut Potsdam and the Seewarte Hamburg to fabricate four of these clocks – or two, respectively
- in accordance with the design principle of the PTR. To link the frequency which had been realized with the quartz clocks up with the astronomically defined unit, the PTR received the signals of the time signal transmitter Nauen, checked by the Seewarte Hamburg, and of the transmitters Rugby (UK) and Bordeaux (F) which – in the course of 1 day – all emitted time signals in the frequency band around 16 kHz (wavelength: 18.5 km). In the rate of the quartz clocks, long-term – and obvi- ously deterministic – fluctuations occurred relative to the time signals received and were corrected by the “improvements” of the broadcasting times subsequently published. When these fluctuations occurred on several clocks of sufficiently differ- ent construction, A. Scheibe and U. Adelsberger judged this – already in 1936 – as an indication of periodic fluctuations of the period of the Earth’s rotation [7]. Meticulously, they tried to improve and to document the quality of their clocks. In doing so, they – first of all – had to convince the protectionists of the superiority of mechanical pendulum clocks, rather than the astronomers, who were very well informed about the irregular- ity of the Earth’s rotation [8]. Figure 2 shows the
- probably most frequently published – measure- ment result obtained with quartz clock III. After Figure 1: Quartz clock of the PTR, constructed after 1935; the arrows – from left to right – point to 3 frequency dividers, 3:1, 10:1, 6:1, to amplifiers, control trans- mitter and, at the far right, to the internal thermostat with the tuning fork.
n (^) The System of Units Special Issue / PTB-Mitteilungen 122 (2012), No. 1 time – this had already been proposed by the Eng- lishman Louis Essen in 1955 – using the transition between the hyperfine structure level of the basic state of the caesium-133 atom. But the time for such a radical change in the concept for the defini- tion of the second was not yet mature then. Shortly before, Essen and his colleagues had put the first caesium atomic clock (abbreviation: Cs clock) into operation at the English National Physical Laboratory, Teddington [14]. From 1955 to 1958, the duration of the unit of time valid at that time – that of the ephemeris second – was determined in cooperation with the United States Naval Observa- tory, Washington, to be 9 192 631 770 periods of the Cs transition frequency [15]. A discussion of the measurement uncertainty for this numerical value furnished the value 20 – although nobody seriously believes that it was possible to indicate the duration of the ephemeris second to relatively 2 · 10 –9. Nevertheless, this measurement result pro- vided the basis for the definition of the unit of time in the International System of Units (SI) which was decided in 1967 by the 13th General Conference of Weights and Measures (CGPM) and which is still valid today: “The second is the duration of 9 192 631 770 periods of the radiation corresponding to the transi- tion between the two hyperfine levels of the ground state of the caesium 133 atom.” In the middle of the 1950s, PTB started to work on atomic frequency standards. In addition to initial tentative work on caesium, the 23.8 GHz inversion oscillation in the NH3 molecule was excited in the so-called ammonium maser and later, the 1.4 GHz transition in atomic hydrogen was used for the construction of a hydrogen maser. However, in the end, only the development of the caesium atomic clock in the “Unit of Time” labora- tory, directed by Gerhard Becker, was successful and will be described in further detail below.
3 Realization of the SI Second
3.1 The caesium atomic clock Any clock can be imagined to consist of 3 parts: an internal clock generator (pendulum, balance, oscillation of a quartz); a counting mechanism, which counts the clock rate events and indicates their number; and an energy source (weights, spring, electric energy from battery or electrical connection). The special feature of an atomic clock, however, is that the clock rate is derived from single, free atoms, which is basically possible with different elements. The isotope caesium-133 has proved to be particularly suited for this because it allows a high-precision, infallible and compact clock to be constructed with comparably simple means. Figure 3 illustrates the functional principle of a Cs clock as it has been built since the middle of the 1950s and is still used today: Based on a quartz oscillator, a microwave field of the frequency fp is generated by means of a frequency generator and coupled into the resonance apparatus (see the large rectangle shaded in yellow). fp already suits the transition frequency of the caesium atoms fCs quite well. In the apparatus, a Cs atomic beam is produced in vacuum by heating caesium to approx. 100 °C in the oven. The beam passes a first magnet
- called the “polarizer” – which diverts only atoms in the energy state E 2 into the desired direction. These atoms fly through the Ramsey cavity (named after Norman F. Ramsey, who won the Nobel Prize in 1989). In the two end sections of the length l of the resonator, the atoms are irradiated with the microwave field. In this way, the transition between the two energy levels is excited with a certain prob- ability. This maximum probability and, thus, the maximum number of atoms in the state E 1 behind the resonator are achieved when the resonance condition is complied with, i. e. when fp and fCs exactly agree. The analyzer magnet now deflects especially E 1 -atoms onto the detector where Cs atoms become Cs+^ ions, and from the ionic current an electrical signal ID is obtained. If fp is tuned over the resonance point, a so-called “resonance line” (see Figure 3) appears when the signal ID is recorded. Its linewidth W is given by W ≈ 1/T, with T being the time-of-flight of the atoms through the resonator having the length L. Now: How can the resonance point be found? For this purpose, fp is Figure 3: Cs-clock, schematic representation; fN: standard frequency, fp: frequency for irradiation of the atoms, ID: detector signal, UR: signal for regulation of the quartz oscillator; Insert on the left: detector signal ID as a function of the frequency of the microwave field fp if it is tuned across the resonance point at fCs; the linewidth W is indicated.
Special Issue / PTB-Mitteilungen 122 (2012), No. 1 Time – the SI Base Unit “Second” n modulated around fCs, and the resulting modula- tion is detected in the detector signal ID. From this, the voltage bias UR is obtained by means of which the quartz oscillator is tuned in such a way that fp gets, on average, into agreement with fCs. In this way, the natural oscillations of the quartz frequency are suppressed in accordance with the adjusted control time constant. The atomic resonance then determines the quality of the emitted standard frequency fN (usually 5 MHz). However, if a short electrical impulse is generated after each 5 million periods of fN, the successive impulses have the tem- poral distance of 1 second – with atomic accuracy. 3.2 Systematic frequency uncertainty and frequency instability Much ado is made of the accuracy of the atomic clocks, and the question is posed what this accu- racy is good for (see Section 7). In the following, we will, therefore, briefly describe what is to be understood – in the narrow sense – by “atomic accuracy”. When the transition between atomic energy levels is excited, the maximum transition probability – i. e. the middle of the resonance line
- is recorded at a value of the excitation frequency which never agrees exactly with the resonance frequency f 0 (here: 9 192 631 770 Hz) for undis- turbed atoms at rest. As the term “accuracy” is used in English, we use the composite term fre- quency accuracy in German to describe the agree- ment between the actual value and the desired value of the output frequency (1 Hz, 5 MHz, etc.) of a clock. Metrologically correct is the following procedure: A detailed list of all possible “disturb- ing” effects is laid down in the so-called uncer- tainty budget of a (primary) clock. This uncer- tainty budget contains the quantitative assessment of all effects which lead to a deviation of the realized transition frequency from that which is to be expected in undisturbed atoms at rest. As the relevant physical parameters as well as the theo- retical relations are known to a limited extent only, such an assessment is affected by an uncertainty. As the final result, a combined uncertainty value is determined, and this quantity will be mentioned several times in the following. The practical con- sequence of this – not really exact – knowledge is: Not even an atomic clock is perfect: two clocks will always exhibit slightly different rates. Rate refers to the change in the reading of a clock relative to a reference clock: If yesterday, the difference of the reading was ΔT 1 and it is ΔT 2 today, the rate of the clock is calculated to be (ΔT 2 – ΔT 1 )/1 day. Typical of atomic clocks is a rate of a few nanoseconds per day. If this is written as a relative frequency differ- ence, one obtains multiples of 10–14. The uncertainty budget of Cs clocks contains contributions which depend on the velocity of the atoms (Doppler effect), but also on electric and magnetic fields along the trajectory of the atoms. This is why the magnetic shielding shown in Figure 3 is used. The output frequency of atomic clocks – and generally of electric oscillators – is, as has already been said, subject to systematic – but also statisti- cal – influences. The oscillations of the output frequency of a Cs clock are, for example, associ- ated with the statistically varying number of atoms arriving at the detector, and these varia- tions – quite clearly – become smaller, the larger the number of atoms and the larger the processed signal. In literature, a great number of represen- tations of the respective characteristics, of their calculation and of their interpretation can be found [16] which are indispensable for both the understanding of the properties of the clocks and the selection in special applications. In the follow- ing, some terms will be briefly presented. The quasi-sinusoidal output voltage of a fre- quency standard is described by V(t) = V 0 (1 + ε(t)) · sin{2πν 0 t + Φ(t)}, (1) V 0 , ν 0 being the nominal amplitude and frequency, and ε(t) and Φ(t) the momentary amplitude or phase variations. As additional quantities, we introduce the relative frequency deviations y(t) = dΦ/dt / (2πν 0 ). The statistical parameter most widely used for the relative frequency insta- bility is the Allan variance σ (^) y^2 τ y (^) k 1 τ ykτ 2 ( ) = ( (^) + ( ) − ( )) , (2) which links successive frequency differences with each other which, then in turn, present mean values over time τ. The usual case is a double- logarithmic plotting of σy(τ) as a function of τ. In this representation, different instability contribu- tions can, in many cases, be identified by means of the slope of the σy(τ) graph. Figure 4 (on the right) and Figure 5 represent such “sigma tau diagrams”. They show the duration of the observation times needed to achieve a desired measurement uncer- tainty – and they also show whether the frequency standard is possibly not at all suited for the meas- urement task. 3.3 Commercial atomic clocks Since the end of the 1950s, Cs-clocks have been offered commercially, and today approximately 200 of them are produced annually worldwide. Already in the early 1960s, PTB made use of a commercial atomic clock – called the “atomichron” [17] – to monitor the rates of the quartz clocks in Braunschweig and Mainflingen. Over the course of years, we have come to understand the interfer-
Special Issue / PTB-Mitteilungen 122 (2012), No. 1 Time – the SI Base Unit “Second” n tainty has been assessed to be 8 ∙ 10 –^15 (CS1) and 12 · 10 –^15 (CS2) [22]. Their role in the realization of the International Atomic Time is acknowledged in Section 4, whereas a detailed description of their development and properties can be found in [23]. Figure 6 shows a current photo of the two clocks. Less successful was the development of other clocks on the basis of the same construction principle, CS3 and CS4, which began as early as in
- The idea to select slower atoms by construct- ing the magnetic lenses correspondingly and, thus, achieve a smaller linewidth was – basically
- correct, and the objective to reach also a smaller frequency uncertainty was also almost reached. But – in spite of the great commitment of the staff – the desired operational reliability was never achieved. Finally after completion of the – much more exact – caesium fountain CSF1, this effort did not make sense any more, and the clocks were dismounted. 3.5 The caesium fountains CSF1 and CSF One of the main objectives in the further develop- ment of atomic clocks has always been to extend the time in which the atoms interact with the microwave field and – thus – reduce the linewidth. This promised, at the same time, a smaller fre- quency instability and a smaller frequency uncertainty. When laser cooling was discovered and understood in the middle of the 1980s [24], its importance for clock development was recognized immediately. At PTB, the new atomic clocks CSF [25] and, later on, CSF2 were developed, which are shown in Figure 7 [26]. Although the principle shown in Figure 3 is still applied, the new – and very important – feature is that it is laser radiation which is used to influence the state of energy and the movement of the atoms [27]. Laser cooling in a magneto-optical trap or in optical molasses furnishes “cold” atoms with a thermal velocity of approx. 1 cm/s. Expressed in the temperature unit, this corresponds to a few millionths of kelvin – compared to the 300 kelvin of the environment. These cold atoms are placed on trajectories, as outlined in Figure 8. The time-of-flight T of the atoms above the microwave resonator is 50 times longer than the respective time-of-flight T in CS2. The resonance line, to the centre of which the microwave oscillator is stabilized, is correspond- ingly narrower. With a relative uncertainty of only approx. 1 ∙ 10 –^15 , CSF1 and CSF2 already come quite close to an ideal clock. Comparisons with CS1 and CS2 allow their frequency uncertainty
- which previously could only be estimated – to be verified now. According to these revisions, CS2 agrees well with CSF1 within the uncertainty of 1 σ. In the case of CS1, a deviation of approx. 9 ∙ 10 –^15 is found. Since the beginning of 2010, CSF1 has been used as a frequency reference for the realization of the time scale at PTB. Unlike conventional clocks, CSF1 does not furnish any second pulses which directly represent a time scale. In fact, the output frequency of CSF1 is com- pared with that of a hydrogen maser. From that comparison, a correction of the maser frequency is calculated, and the time scale is derived from the maser corrected with it.
4 Atomic Time Scales: TAI and UTC
A time scale is defined by a sequence of second marks which starts from a defined beginning. International Atomic Time TAI (Temps Atom- ique International) – and especially Coordinated Universal Time (UTC) – allow events in science and technology to be dated. At the same time, Figure 6: PTB's primary clocks CS1 and CS2. Figure 7: The two caesium fountains CSF1 and CSF2, with Dr. Ste- fan Weyers, Head of the “Time Standards” Working Group.
n (^) The System of Units Special Issue / PTB-Mitteilungen 122 (2012), No. 1 UTC provides the basis of the “time” that is in use in everyday life. Since 1988, the Time Department of the BIPM (Bureau International des Poids et Mesures) has been charged with its calculation and propagation. For its calculation, approx. 350 clocks from approx. 70 time-keeping institutes, distrib- uted all over the world, are used. From PTB, the measurement values of the primary clocks CS1 and CS2, of the three commercial caesium clocks and, since around 1990, of the hydrogen maser have been transmitted. The algorithm used to combine all data is designed in such a way that it reliably produces an optimized, stable time scale over an averaging period of 30 days. For this purpose, statistic weights – which follow from the rate behaviour of the clocks during the past 12 months
- are assigned to the contributing clocks when their rates are averaged. The more stable the rate of a clock, the higher its statistic weight. The mean obtained in this way is called EAL (Echelle Atomique Libre, free atomic time scale). In a second step, TAI is obtained, whose scale unit is to agree with the SI second as it would be realized at sea level, i. e. it shall be as long as indicated in the definition of the second (see Section 2.4). For this purpose, the TAI second is adapted to seconds realized with primary standards. In 2009 and 2010, 14 standards were involved. With its clocks CS and CS2 and the two caesium fountains CSF1 and CSF2, PTB has, for many years, taken a prominent position worldwide. No other standard worldwide has transmitted so many measurement values as CS1 and CS2: From CS2, for example, values for 288 months in succession (including August 2010) were transmitted.^2 For a number of years, only these two clocks have been available worldwide for this purpose. Figure 9 shows current meas- urement values of all fountain clocks relative to the – already controlled – time scale TAI. Accord- ingly, the TAI second agrees very well with the SI second, the deviation is exactly known. A discus- sion of the mean offset and of the scattering of the single values would go beyond the scope of this essay. The beginning of TAI was defined in such a way that the 1st^ of January 1958, 0 o’clock TAI, agreed with the respective moment in UT1 (see Section 2.1). From TAI one obtains UTC, which is the basis of today’s universal time system with 24 time zones. UTC results from proposals of the CCIR (Comité Consultatif International des Radiocom- munications) of the ITU (International Telecom- munication Union). According to these proposals, the emission of time signals should be “coordi- nated” worldwide, i. e. it should be related to a common time scale. If TAI had been used for it, the time signals would have been shifted gradually with respect to the indication of a universal time clock, i. e. the moment 12:00 o’clock would have been shifted with respect to the highest position of the Sun at the zero meridian. The latter follows from the adaptation of the duration of the atomic second to the ephemeris second described above. For navigation according to the position of celes- tial bodies, knowledge of UT1 is necessary, and time signals have been emitted since the early 20th century, in particular due to the requirements of navigation. Since in the 1960s, astronomic naviga- (^2) Then the whole thing stopped – the caesium filled in in 1989 had been used up. Since December 2010, CS2 has been ticking again! Figure 8: Function of a foun- tain frequency stand- ard, represented in four time-dependent steps from the left to the right; laser beams are illustrated by arrows (white, if blocked).
- Loading of an atomic cloud;
- Throwing the cloud up by detuning the frequency of the vertical lasers;
- Dilatation of the initially compact cloud due to the ther- mal energy which remains in spite of the cooling. The at- oms fly upwards and downwards through the microwave reso- nator, and a specific population of the two hyperfine structures is reached;
- After the second passage of the mi- crowave resonator, the population of the state is measured by laser irradiation and fluorescence detection.
n (^) The System of Units Special Issue / PTB-Mitteilungen 122 (2012), No. 1 comparisons between the clocks and the respec- tive time scales UTC(k) of the institutes k operat- ing the clocks, and then comparisons of UTC(k) among each other. The former is quasi trivial. For the latter, a standard procedure exists which uses the signals emitted by the satellites of the Ameri- can Satellite Navigation Systems GPS and of the Russian GLONASS. Special time receivers deter- mine the arrival times of the signals of all satellites which are simultaneously visible above the horizon with respect to the local reference time scale and furnish original data of the kind {local time scale minus GPS time T(GPS)} (or GLONASS time). To compare the time scales of two institutes (i) and (k) with each other, the time differences are determined, the measurement data are exchanged and (e. g.) the differences [UTC(i)-T(GPS)]– [UTC(k)-T(GPS)] formed. By averaging typically 500 daily observations with a duration of approx. 15 minutes, the time scales of two time-keeping institutes can be compared worldwide with a sta- tistic uncertainty of approx. 2 – 4 ns (1 s). Not only the clocks themselves, but also the methods of the intercontinental comparisons were – and still are – continuously further devel- oped. By a combination of all methods available, the comparison of caesium fountains could – in an international collaboration – be realized with a statistic uncertainty of 1 ∙ 10–15. The values were averaged over 1 day each [31]. PTB intensively pursues the use of geostationary telecommuni- cation satellites for time comparisons – called “Two-Way Satellite Time and Frequency Transfer (TWSTFT)” – and operates ground stations for the traffic with Europe/USA and Asia. Figure 12 shows a mobile station for the calibration of signal transit times [32] located next to the permanent installa- tions on the Laue Building. The time comparison data collected worldwide can largely be retrieved on servers which are pub- licly available. A time and frequency comparison of the highest accuracy can, thus, be performed with respect to UTC(PTB), with respect to many other realizations UTC(k), or to UTC and TAI und – thus – to the SI second – and this with the accuracy suited for almost every application.
6 Dissemination of Time for Society
The dating of events and the coordination of the various activities in a modern society have been recognized as being so important that in many countries how legal time is to be indicated is regulated by law. This is also the case in Germany. International traffic and communications make it necessary for the times of the countries which are fixed in this way to be coordinated with each other. The basics for this were laid down in October 1884 by the Washington Standard Time Conference [1, 2]. Thereby, the position of the zero merid- ian and the system of the 24 time zones – each one having a geographic longitude of 15° – were determined. After the second had been redefined on the basis of quantities of atomic physics in 1967, the regulation for the legal time valid in Germany also had to be adapted. This was realized by the Time Act of 1978 in which PTB was entrusted with the realization and dissemination of the time which is decisive for public life in Germany. Central European Time (CET) or – if introduced – Central European Summer Time (CEST) were determined as the legal time. CET and CEST are derived from UTC, adding one or two hours: CET(D) = UTC(PTB) + 1h, CEST(D) = UTC(PTB) + 2h. Figure 11: Comparison of Universal Coordi- nated Time UTC with atomic time scales UTC(k) realized in four European time- keeping institutes (k) during one year; MJD 55834 refers to September 30, 2011; red: Istituto Nazion- ale di Richerca di Metrologia, INRiM, Turin; cyan: NPL, Tedding- ton, UK; green: METAS, Swizerland blue: PTB. Figure 12: Establishment of a transportable satellite terminal for calibration of the time comparisons via TWSTFT with the stationary facility (background). In addition, the Time Act also authorizes the German Federal Government to introduce, by way of a statutory ordinance, Summer Time between 1 March and 31 October of each year. The dates for the beginning and the end of CEST are determined by the Federal Government in accordance with the currently valid directive of the European Parlia- ment and of the Council of the European Union and are announced in the Federal Law Gazette. The Time Act of 1978 and the Units in Metrology Act of 1985 were combined to a new, joint act, i. e. the Units and Time Act, which was adopted in 2008
Special Issue / PTB-Mitteilungen 122 (2012), No. 1 Time – the SI Base Unit “Second” n and in which all regulations concerning the deter- mination of time have been taken over without changes. During the past few decades, PTB has used different procedures to disseminate time and frequency information to the general public and to use it for scientific and technical purposes. The long-wave transmitter DCF77 of Media Broadcast GmbH is the most important medium for this because the number of receivers in operation is estimated to be more than 100 million. It is often the case that PTB is known to many Germans and to many people in Europe only due to one service: the control of radio-controlled clocks. The carrier oscillation 77.5 kHz of this emission is used for the calibration of standard frequency generators. Before the war, PTR had already offered a com- parable service, using the “Deutschlandsender”. With DCF77, the time and date of legal time are transmitted in an encoded form via the second marks. In 2009, in memory of 50 years of DCF broadcasting, topics such as the current state of the broadcasting programme, the receiver character- istics, radio-controlled clocks and the history of time dissemination in Germany were intensively addressed in publications [33]. Corresponding services on long-wave also exist in England, Japan and the USA [34]. Since the mid 1990s, PTB has been offering time information via the public telephone network. Computers and data acquisition facilities can retrieve the exact time from PTB with the aid of telephone modems, calling the number 0531
- The major part of the calls (presently approx. 1700 calls per day) comes from the meas- uring stations of different energy suppliers who need this information for the fiscal measurement of natural gas. With the advent of the Internet, a new medium for time dissemination came into being which has meanwhile become extraordinarily popular. Publicly available servers with the addresses ptbtimeX.ptb.de (X = 1, 2, 3) serve to synchronize computer clocks in the Internet with UTC(PTB). Figure 13 shows the number of current accesses to the three servers in the course of one week. During the past few years, the number of accesses has increased to approx. 3000 per second (as shown in the diagram).
7 Further developments and applications
For a great number of technical, military and – last but not least – scientific applications, exact and stable clocks and frequency references are indis- pensable. In the previous sections, the – seemingly
- constant improvement of the accuracy of time measurement has been outlined. This – and also PTB’s role – is impressive, but inevitably, the ques- tion arises: What is the use of all this? Where will the development take us? Let’s deal with the second question first: All atoms and ions have states of energy. Between these, transitions occur – at frequencies that are up to 105 times higher than those we have discussed up to now (i. e.: in the range of the visible or ultraviolet radiation). In that case, schoolbook physics teaches us that the transitions from the higher to the lower state of energy run rapidly and spontaneously. In special cases, the transition probabilities are, however, strongly suppressed and, at the same time, the transition frequencies depend only to a small extent on external parameters. If the frequency of a laser is stabilized to such a transition line, it represents a very good standard for the frequency and the wavelength [35, 36]. In the case of the selected “clock transitions”, linewidths W in the Hz area are obtained which furnish – with a transition frequency from 10^14 to 10^15 Hz – a correspondingly high line quality and, thus, a small frequency insta- bility. The quality gained with the transition from low to higher clock frequencies (mechanical clock: 1 Hz, quartz clock: 10 kHz, atomic clock, so far: 10 GHz) is continued and, consequently, also leads to a reduction in the frequency uncertainty. In the past few years, the development of optical frequency standards has been rapid. At the end of 2010, the smallest value of the frequency uncer- tainty published lay at 9 ∙ 10–18^ [37]. From this, the perspective for a possible redefinition of the second results, and the question arises as to how the atomic transition best suited for it can be found. The category “secondary representation of the second” was created [38]. The categories which have been recognized so far are based on a microwave transition in rubidium and on optical transitions in the ions 88 Sr+ (strontium), 199 Hg+ (mercury) and (^171) Yb+ (ytterbium) and in the neutral atom 87 Sr [36, 38]. Due to the development of frequency comb technology, the comparison of optical frequency Figure 13: Number of the NTP packages received during one week from the three externally accessible NTP servers ptbtimeX: 0 corresponds to 15.12.2010, 0:00 UTC; green: X = 1, red: X = 2, black: X = 3.
Special Issue / PTB-Mitteilungen 122 (2012), No. 1 Time – the SI Base Unit “Second” n was provided by the analysis of the arrival time of the pulse signals with respect to atomic time scales [47]. Many of the approaches searching for “new physics” concentrate on possible deviations from Einstein’s Principle of Equivalence [30]. An experi- ment in this connection which can be easily carried out is the comparison of two atomic clocks with dif- ferent atomic references (e. g. caesium clock versus hydrogen maser) in the time-dependent gravi- tational potential of the Sun – during the annual rotation of the Earth on its elliptical orbit [48]. In the past 20 years, hypothetical infractions of the Principle of Equivalence have – to an ever increas- ing extent – been gradually ruled out. The availabil- ity of more exact clocks and improved possibilities for the comparison of clocks were the prerequisite [49]. In future, this type of time measurement will continue to offer a wide field of activities for PTB which will excellently complement PTB’s tasks of everyday routine.
8 Closing Remarks
“Sometimes I get up in the middle of the night and stop all the clocks. However, we should not fear time – it, too, is a creature of the Father who created all of us”.(1)^ n Literatur [1] G. Dohrn-van Rossum: Die Geschichte der Stunde, München, Wien, Carl Hanser Verlag, 1992 [2] R. Wendorff: Zeit und Kultur, 3. Auflage, Opladen, Westdeutscher Verlag 1985 [3] F. R. Stephenson, L. V. Morrison: Long-term changes in the rotation of the Earth: 700 B:C: to A. D. 1980, Phil. Trans. R. Soc. Lond., A 313 , 47−70 (1984) [4] J. Graf, Hrsg: Die Quarzrevolution, 75 Jahre Quarzuhr in Deutschland, Furtwanger Beiträge zur Uhrengeschichte, Neue Folge, Band 2, Deutsches Uhrenmuseum 2008 [5] W. Giebe, A. Scheibe: Über Leuchtresonatoren als Hochfrequenznormale, Zeitschrift für Hochfrequenz technik und Elektronik, 41 , 83−96 (1933), see also U. Kern: Forschung und Präzisionsmessung, VCH Weinheim, 2008 (sections 3 and 5) [6] A. Scheibe: Genaue Zeitmessung, Ergeb. Ex. Naturw, 15 , 262−309 (1936) (with numerous original quota- tions) [7] A. Scheibe, U. Adelsberger: Schwankungen der astronomischen Tageslänge und der astronomischen Zeitbestimmung mit den Quarzuhren der Phys.- Techn. Reichsanstalt, Phys. Z., 37 , 185−203 (1936) [8] N. Stoyko: Précision d’un Garde-Temps Radio-Elect- rique, Ann. Fr. de Chronom., 221 (1934) [9] A. Scheibe, U. Adelsberger: Die Gangleistungen und technischen Einrichtungen der Quarzuhren der PTR in den Jahren 1932−1944, Manuskript, Berlin, Heidelberg und Braunschweig, 1950, 2 volumes. [10] A. Scheibe, U. Adelsberger, G. Becker, G. Ohl, R. Süss: Über die Quarzuhrengruppe der PTB und Vergleichsmessungen gegenüber Atomnormalen, Z. Angew. Phys., 11 , 352−357 (1959) [11] R. Süß: 10 Jahre Normalfrequenzaussendungen der PTB über den Sender DCF77, PTB-Mitt., 7 8, 357−362 (1968) [12] G. Becker: Von der astronomischen zur astrophysi- kalischen Sekunde, PTB-Mitteilungen, 76 , 315− und 76 , 415−419 (1966) [13] R. A. Nelson, D. D. McCarthy, S. Malys, J. Levine, B. Guinot, H. F. Fliegel, R. L. Beard, T. R. Bartho- lomew: The leap second: its history and possible future, Metrologia, 38 , 509−529 (2001) [14] L. Essen, J. V. L. Parry: An atomic standard of frequency and time interval, Nature, 176 , 280− (1955) [15] W. Markowitz, R. G. Hall, L. Essen, J. V. L. Parry: Frequency of cesium in terms of ephemeris time, Phys. Rev. Lett., 1 , 105−107 (1958) [16] W. J. Riley: Handbook of Frequency Stability Analysis, NIST Special Publication 1065, NIST 2008, basierend auf: Standard definitions of Physical Quantities for Fundamental Frequency and Time Metrology, IEEE-Std 1189−1988 (1988) [17] P. Forman: Atomichron®: the atomic clock from concept to commercial product, Proc. IEEE, 73 , 1181−1204 (1985) [18] J. H. Holloway, R. F. Lacey: Factors which limit the accuracy of cesium atomic beam frequency stand- ards, Proc. Intern. Conf. Chronometry (CIC 64), 317−331 (1964) [19] G. Becker: Stand der Atomuhrentechnik, Jahrbuch der Deutschen Gesellschaft für Chronometrie, 18 , 35−40 (1967) [20] G. Becker, B. Fischer, G. Kramer, E. K. Müller: Neuentwicklung einer Caesiumstrahlapparatur als primäres Zeit- und Frequenznormal an der PTB, PTB-Mitteilungen, 79 , 77−80 (1969) [21] A. Bauch, K. Dorenwendt, B. Fischer, T. Heindorff, E. K. Müller, R. Schröder: CS2: The PTB’s new primary clock, IEEE Trans. Instrum. Meas., IM-36 , 613−616 (1987) [22] A. Bauch, B. Fischer, T. Heindorff, P. Hetzel, G. Petit, R. Schröder, P. Wolf: Comparisons of the PTB primary clocks with TAI in 1999, Metrologia, 37 , 683−692 (2000) [23] A. Bauch: The PTB primary clocks CS1 and CS2, Metrologia, 42 , S43−S54 (2005) [24] S. Chu: The manipulation of neutral particles, Rev. Mod. Phys., 50 , 685−706 (1998), C. Cohen-Tannoud- ji: Manipulating atoms with photons, ibid. 707−720, W. D. Philips: Laser cooling and trapping of neutral atoms, ibid. 721− [25] S. Weyers, D. Griebsch, U. Hübner, R. Schröder, Chr. Tamm, A. Bauch: Die neue Caesiumfontäne der PTB, PTB-Mitteilungen, 109 , 483−491 (1999)
n (^) The System of Units Special Issue / PTB-Mitteilungen 122 (2012), No. 1 [26] V. Gerginov, N. Nemitz, S. Weyers, R. Schröder, D. Griebsch, R. Wynands: Uncertainty evaluation of the caesium fountain clock PTB-CSF2, Metrologia, 47 , 65−79 (2010) [27] R. Wynands, S. Weyers: Atomic fountain clocks, Metrologia, 42 , S64−S79 (2005) [28] A. Einstein: Zur Elektrodynamik bewegter Körper, Annalen der Physik, IV. Folge, Band 17 , 891− (1905) [29] G. Becker, B. Fischer, G. Kramer, E. K. Müller: Die Definition der Sekunde und die Allgemeine Relativ- itätstheorie, PTB-Mitt., 77 , 111−116 (1967) [30] C. M. Will: The confrontation between General Relativity and experimentation, Living Rev., 9 , (2006), http://livingreviews.org/lrr-2006- [31] A. Bauch et al.: Comparison between frequency standards in Europe and the USA at the 10−15^ uncer- tainty level, Metrologia, 43 , 109−120 (2006) [32] D. Piester, A. Bauch, L. Breakiron, D. Matsakis, B. Blanzano, O. Koudelka: Time transfer with nano- second accuracy for the realization of International Atomic Time, Metrologia, 45 , 185−198 (2008) [33] J. Graf: Wilhelm Foerster, Vater der Zeitverteilung im Deutschen Kaiserreich, PTB-Mitt., 119 , 209− (2009); A. Bauch, P. Hetzel, D. Piester: Zeit- und Frequenzverbreitung mit DCF77: 1959−2009 und darüber hinaus, ibid. 217−240; K. Katzmann: Die Technik der Funkuhren, ibid. 241− [34] ITU-Empfehlungen ITU-R TF768-5 “Stand- ard Frequencies and Time Signals” und ITU-R TF583-6 “Time Codes”, zu beziehen unter http:// www.itu.int/ITU-R/index.asp?category=study- groups&rlink=rsg7&lang=en, Related Information “TF583” bzw. “TF768” [35] E. Peik, U. Sterr: Optische Uhren, PTB-Mitteilun- gen, 119 , 123−130 (2009) [36] S. N. Lea: Limits to time variation of fundamental constants from comparisons of atomic frequency standards, Rep. Prog. Phys., 70 , 1473–1523, (2007) [37] C. W. Chou, D. B. Hume, J. C. J. Koelemij, D. J. Wineland, T. Rosenband: Frequency Com- parison of Two High-Accuracy Al+ Optical Clocks, Phys. Rev. Lett., 104 , 070802 (2010) [38] P. Gill, F. Riehle: On Secondary Representations of the Second. Proc. 20th^ European Frequency and Time Forum (Braunschweig 2006), 282– [39] J. L. Hall: Defining and measuring optical frequen- cies, Rev. Mod. Phys., 78 , 1279–1295 (2006), T. W. Hänsch: Passion for precision, ibid. 1297–1309. [40] H. Schnatz , pp. 7– [41] Chr. Tamm, S. Weyers, B. Lipphardt, and E. Peik: Stray-field-induced quadrupole shift and absolute frequency of the 688-THz 171Yb+ single-ion optical frequency standard, Phys. Rev., A 80 , 043403 (2009) [42] E. Peik, Chr. Tamm: Nuclear laser spectroscopy of the 3.5 eV transition in Th-229, Europhysics Letters, 61 , 181–186 (2003) [43] H. Drewes: Science rationale of the Global Geo- detic Observing System (GGOS), Dynamic Planet, Internat. Ass. of Geodesy Symposia vol. 130, eds: P. Tregoning, C. Rizos, Springer (Berlin 2007), 703 [44] D. Piester, H. Schnatz: Hochgenaue Zeit- und Frequenzvergleiche über weite Strecken, PTB-Mitt., 119 , 131–143 (2009) [45] R. F. C. Vessot, M. W. Levine: A test of the equiva- lence principle using a space-borne clock, J. Gen. Rel. and Grav., 10 , 181–204 (1979) [46] I. H. Stairs: Testing General Relativity with Pulsar Timing, Living Rev. Relativity 6 , (2003); http://www.livingreviews.org/lrr-2003- [47] J. Taylor: Binary pulsars and relativistic gravity, Rev. Mod. Phys., 66 , 711−719 (1994) [48] A. Bauch, S. Weyers: New experimental limit on the validity of local position invariance, Phys. Rev., D 65 , 081101 (2002) [49] T. M. Fortier et al: Precision atomic spectroscopy for improved limits on variation of the fine structure constant and local position invariance, Phys. Rev. Lett., 98 , 070801 (2007)
