Gas-Cell Laser Clock Technologies for Atomic Frequency Standards

Status date

This project’s objectives concern the introduction of laser optical pumping to the well-established Rubidium (Rb) atomic clock technology, in view of replacing the lamp-pumped devices today available as compact frequency references. With the proven benefits of such laser-pumped Rb clocks, the clock’s frequency stability can be improved to < 6x10-13 at 1 second and < 1x10-14 at one day of integration time, with low frequency drift < 3x10-13/day. The advantages of laser-pumped Rb clocks have already been demonstrated in laboratory studies, but no commercially available laser-pumped clock meeting the full specifications was developed yet, partly also due to lack of suitable laser diodes.

The present project aims to bridge this gap by pushing the required key technologies for the clock, and to demonstrate a laser-pumped Rb clock meeting the above stability requirements from a core physics package of 1 litre volume only. Furthermore, this clock shall be based on technologies suitable for the development of a space atomic clock. Such a high-stability compact atomic clock will find its applications in satellite navigation (GALILEO), telecommunication, and science missions.

Image: CAD design view of the clock demonstrator (Physics package).

click for larger image


Key issues of the project were: identification and evaluation of a suitably robust single-mode laser diode; development of a scheme for reproducible buffer-gas filling to the Rb cells; noise reduction on the atomic clock signal by passive subtraction of the laser noise (using two clock cells) and implementation of a low phase-noise microwave source.

The lifetime of the clock depends on the spectral lifetime of the laser diode, i.e. on the time during which the Rb wavelength can be reached by adjusting the operating parameters. Also, the full clock should be able to operate under vacuum. Two long-term experiments were set up to address these points.


Compared to today’s lamp-pumped Rb clock, this project’s laser-pumped clock will profit from improved short-term stability, thanks to the exploitation of laser optical pumping and a low-noise microwave source. Furthermore, the independent control of light wavelength and intensity – possible only with the laser – allows controlling different systematic effects that limit the long-term stability and thus to improve on this clock specification.

With respect to other atomic clock technologies, the project development has the advantage to provide an atomic clock with excellent frequency stability (see “project objectives” above), which today is reached by the passive Hydrogen maser technology only. This project’s development however paves the way to realise this frequency stability from a drastically more compact Rb clock device, which is of high relevance for any future space application (up to 10 times lower volume and mass).


In an atomic clock, an atomic reference line (the “clock transition”) is detected and used as a stable reference. The frequency of a quartz local oscillator (LO) is stabilized to this reference line, and the quartz then gives the stabilized clock output.

The laser-pumped clock design employs an intrinsically single-mode Distributed Feedback (DFB) laser diode, emitting at the Rb D2 wavelength around 780.24nm. Its precise frequency is actively stabilized to Rb reference lines from a frequency stabilisation reference setup containing a small Rb cell. The laser light prepares a Rb atomic vapour contained in a Rb “clock” cell into a polarized state. Microwave radiation at ≈ 6.8GHz is applied to the atoms via a microwave cavity, and – if in resonance – the clock transition is detected on a photo-detector placed behind the clock cell.

A servo loop using this signal is used to stabilize the quartz LO, from which the microwave radiation is derived. Also the frequency of the atomic clock transition slightly varies in response to changing physical conditions, which finally limits the clock stability. Such systematic effects include variations due to, e.g., small fluctuations in the frequency or intensity of the pump light (“light shift”), applied microwave power, or temperature instabilities of the clock cell.

All these dependencies have to be precisely controlled in a Rb clock. For example, a well-balanced mixture of inert buffer gases has to be added to the clock cell, in order to minimize its sensitivity to variations of the clock cell’s temperature.

Image: Basic functional scheme of a laser-pumped Rb clock.

click for larger image


The project started with studies and development on the required key technologies for a Rb clock. Generation of narrow and low-noise clock transition signals for improved short-term stability was investigated, using narrow-band laser diodes and microwave synthesizers with very low phase-noise. Improved stabilisation of the laser frequency to a separate reference cell and precise control of the buffer-gas content filled to the clock cells were developed to improve on the long-term clock stability. Finally, a clock demonstrator was realised to demonstrate the improvements on the various technologies in a compact clock device.

Current status

The project was concluded in 2010 with the realisation and experimental evaluation of a clock demonstrator. The demonstrator’s core physics package has a volume of 1.1 litres only, and includes the full setup with laser source, optics, and two atomic resonator modules (for passive noise cancellation). Laboratory-type control electronics was used.

The measured short-term stability is < 5x10-13 τ -1/2, only marginally above the shot-noise limit of 4x10-13 τ-1/2. At timescales of τ > 104s the clock stability is < 3x10-14, only slightly higher than the project stability goal of 1x10-14. This is due to imperfections in the thermal isolation of the clock cells that will be easily corrected by introduction of a second stage of thermostats. Frequency drifts are ≤ 2x10-13/day, already at level with the very best lamp-pumped Rb clocks.

First results on long-term studies launched indicate that it will be possible to operate the laser diode precisely at the required Rb wavelength for many years, as required to provide good lifetime of the overall instrument (e.g. 12-15 years for GALILEO clocks). A lamp-removed GALILEO RAFS module pumped by a diode laser was also operated successfully under vacuum and without anomalies for more than 4 months.