Table of contents

The year 2005 is the centenary of Einstein's four famous papers that were published in 1905. This anniversary is being widely celebrated all over the world and, indeed, 2005 has been dubbed World Year of Physics.

The year 2005, however, also marks the fiftieth anniversary of the first operation of Essen and Parry's caesium beam atomic frequency standard at the NPL in May 1955. While Einstein's papers signalled a revolution in physics and in our understanding of the natural world, the first atomic clock signalled a revolution in time-keeping that has become, among other things, one of the most powerful tools in pushing back the frontiers of Einstein's theories of special and general relativity. The atomic clock has also had consequences for navigation comparable to those brought about by Harrison's mechanical clocks almost exactly two hundred years before. Harrison's H3 was completed in 1757 and H4 in 1759.

The atomic clock, and the creation of an atomic time scale that quickly followed, led ten years later to the adoption of an atomic definition for the SI second in Resolution 1 of the 13th General Conference on Weights and Measures, 1967/68. This marked the end of time-keeping based on the movements of the heavenly bodies that had beaten the rhythm of the days and the seasons since the dawn of human civilization.

Fifty years on is a good occasion to look back, to look forward and at the same time to examine where we are today, in terms of measuring time. While we still arrange for our atomic clocks to show noon when the sun is overhead on the Greenwich meridian, everything else has changed in the fifty years since 1955.

In this special issue of Metrologia the reader will find articles on the development of the atomic clock, its theory and practice, how the first atomic time scale was devised and formally introduced and how we maintain atomic time today, as well as articles looking forward to even more accurate clocks and time scales. Included also are articles on the commercial development of atomic clocks of various types and on some of their applications. At the beginning there is a deliberate emphasis on the history of the introduction of atomic time, including the technical problems to be resolved and the personalities involved. You will see that it includes one article based on notes left by Louis Essen himself, for which we are most grateful to his son-in-law, Mr Ray Essen, for permission to use them and to Dale Henderson of the NPL, who arranged them for publication here.

I hope that this issue will stand as a reference for years to come and I am most grateful to all those who have contributed. I also wish to thank most particularly Norman Ramsey, whose name is indelibly associated with atomic clocks, for having contributed the first article to this special issue.

This review of the history of early atomic clocks includes early atomic beam magnetic resonance, methods of separated and successive oscillatory fields, microwave absorption, optical pumping and atomic masers.

To commemorate the fiftieth anniversary of the development of the first atomic frequency standard, we present some notes about the work of Louis Essen at the National Physical Laboratory. In addition, we publish below some personal recollections of Essen on his work, which have previously been available only on the Internet (http://www.btinternet.com/~time.lord/TheAtomicClock.htm).

When the reference of a unit is changed, the task remains of expressing the new definition in the terms of the previous one. In two centuries, for instance, the definition of the metre was changed six times. Around 1955, the definitions of the second as the SI unit of time were changed from astronomical ones, based on the rotation (or revolution) of the Earth, to a definition stemming from a quantum phenomenon. The expression of the new atomic second in terms of the previous definitions required a number of actions and measurements. The aim of this paper is to review these actions, some of which, fifty years later, are being forgotten for a lack of adequate documentation.

This paper summarizes the creation and technical evolution of atomic time scales, recalling the parallel development of their acceptance and the remaining problems. We consider a consequence of the accuracy of time measurement, i.e. the entry of Einstein's general relativity into metrology and its applications. We give some details about the method of calculation and the characteristics of International Atomic Time, and we show how it is disseminated at the ultimate level of precision.

The caesium beam frequency standard using the separated oscillatory field technique proposed by Ramsey in 1950 has seen intensive development over the last fifty years. Its practical implementation as a primary time standard made possible the realization of the second with a precision better than that obtained by means of astronomical measurements and provided the basis for its redefinition in terms of atomic properties in 1967. This paper describes the basic principles underlying the operation of such atomic standards, reviews the progress made during the last fifty years since its invention, and provides an update on the state-of-the-art accuracy and frequency stability achieved today in this field.

The primary clocks CS1 and CS2 have been developed and operated by the Physikalisch-Technische Bundesanstalt, Braunschweig, Germany. By their contributions to the definition of the scale unit of International Atomic Time they have provided access to the realization of the SI second over decades with exceptional accuracy. They have stood out against other primary clocks by the novelty of their design concept, their robustness of construction and last but not the least by their almost continuous operation for many years. Their properties have changed with time, but during the last 7 years their uncertainty, u_B, has been estimated as 8 × 10⁻¹⁵ (CS1) and 12 × 10⁻¹⁵ (CS2). Comparisons with PTB's cold-atom frequency standard CSF1 (u_B = 1 × 10⁻¹⁵) over 3.5 years revealed that CS2 and CSF1 agreed well within the uncertainty u_B(CS2), whereas CS1 frequency deviates slightly more from CSF1 than u_B(CS1).

Time metrology in Russia in the second half of the twentieth century has been marked, as in other advanced countries, by the rapid development of time and frequency quantum standards and the beginning of atomic time-keeping. This brief review presents the main developments and studies in time and frequency measurement, and the improvement of accuracy and atomic time-keeping at the VNIIFTRI—the National Metrology Institute keeping primary time and frequency standards and ensuring unification of measurement. The milestones along the way have been the ammonia and hydrogen masers, primary caesium beam and fountain standards and laser frequency standards. For many years, VNIIFTRI was the only world laboratory that applied hydrogen-maser clock ensembles for time-keeping. VNIIFTRI's work on international laser standard frequency comparisons and absolute frequency measurements contributed greatly to the adoption by the CIPM of a highly accurate value for the He–Ne/CH₄ laser frequency. VNIIFTRI and the VNIIM were the first to establish a united time, frequency and length standard.

We describe and review the current state of the art in atomic fountain clocks. These clocks provide the best realization of the SI second possible today, with relative uncertainties of a few parts in 10¹⁶.

The basic operational requirements and present technology for high stability hydrogen maser (H-maser) oscillators are described. Design features of a high stability maser oscillator for use in space, maser operation at 0.5 K and present uses of high stability H-masers are discussed.

The first commercial caesium-beam clock was the Atomichron, developed by National Company. The first unit was delivered in 1956. This paper first presents a brief history of commercial caesium-beam clocks. It then covers many of the sources of frequency error and frequency instability. Some details of the technical design of the caesium-tube and electronics are presented. Finally, we present some possible directions for future commercial caesium clocks.

Using microfabrication processes, we have been able to construct physics packages for vapour cell atomic frequency references 100× smaller than previously existing versions, with a corresponding reduction in power consumption. In addition, the devices offer the potential for wafer-level fabrication and assembly, which would substantially reduce manufacturing costs. It is anticipated that a complete frequency reference could be constructed based on these physics packages with a total volume below 1 cm³, a power dissipation near 30 mW and a fractional frequency instability below 10⁻¹¹ over time periods from hours to days. Such a device would enable the use of atomically precise timing in applications that require battery operation and portability, such as hand-held global positioning system receivers and wireless communication systems.

Surprising as it might seem, it is possible to phase-coherently track, synthesize, count and divide optical frequencies of visible laser sources. In essence, the technologies described here now allow direct connection of basically any frequency from DC to 1000 THz. Modern 'self-referenced' femtosecond mode-locked lasers have enormously simplified the required technology. These revolutionary new systems build on a long history of optical frequency metrology that spans from the early days of the laser. The latest systems rely heavily on technologies previously developed for laser frequency stabilization, optical phase-locked-loops, nonlinear mixing, ultra-fast optics and precision opto-electronic metrology. Using examples we summarize some of the heroic efforts that led to the successful development of harmonic optical frequency chains. Those systems played critical roles in defining the speed of light and in redefining the metre. We then describe the present state-of-the-art technology in femtosecond laser frequency combs, their extraordinary performance capabilities and some of the latest results.

The evolution of atomic frequency standards since Essen's atomic clock fifty years ago has been considerable both in respect of microwave and optical standards. In particular, the development of trapping and laser cooling techniques for both atoms and ions has played a major role. This paper reviews the status of the development of single cold trapped ion and cold trapped atom optical frequency standards. Recent results show comb measurements of trapped ion optical frequency standards with accuracies close to Cs fountain limited operation. The factors affecting future stability and reproducibility are discussed. The opportunities for future standards capable of approaching reproducibility at the 10⁻¹⁸ level are outlined, together with the likely limitations arising.

Soon after the advent of atomic clocks 50 years ago, their unprecedented uncertainties and their widespread use required that the definition of time scales and clock comparison procedures be considered rigorously within the theoretical framework of general relativity. We review the present procedures and show that the current relativistic modelling is adequate for the need of clocks and time transfer techniques. Accounting for the expected improvements in clock and time transfer technology, we investigate some possible implications in relation to the relativistic models.

The traditional applications of clocks and oscillators in navigation and scientific research continue to grow as new and more widespread applications emerge. Today, clocks and oscillators regulate the speed and efficiency of virtually every human endeavour ranging from the mundane to the exotic. In this paper we discuss some of the applications that require the service of high performance clocks and oscillators. In particular, we discuss space applications of clocks for navigation and scientific experiments. We also present a discussion of satellite navigation systems, especially the Global Positioning System, and comment on time dissemination over the Internet.