University of Colorado at Boulder; winner, Nobel Prize for Physics, 2005
Measuring frequencies is regarded as being the way to achieve high resolution and accuracy in physical measurements. As these applications become more demanding, eventually the actual methods of measuring frequency and even the basic standard of frequency become limiting factors. What has happened over the last quarter century is the following:
With the introduction of laser stabilisation ideas in the late 1960s, the various national standards labs across the world recognised the value to optical frequency measurement, as leading for example to an improved measurement of the speed of light, surely one of our most important physical constants. These efforts were team approaches - our National Institute of Standards and Technology team used about a dozen person-years to follow the tradition universally employed in the lower frequency domains of radio and microwaves. Namely, in the absence of any really crisp non-linear effect, we could only multiply frequency by a factor of two per step, based on using second-harmonic generation in a non-linear detector or a non-linear crystal. The present standard of frequency is based on Cesium atoms (which resonate at about 10 billion oscillations per second), that we need to compare to the lightwave frequency domain of oscillations at a frequency about 100 000-fold faster. To achieve such a factor of 100 000 in steps of a factor of two will need about 16 steps, implemented as perhaps three stages of microwaves and 13 stages of lasers. This leads to laboratory-filling experimental set-ups, with a dozen or more individual lasers set up to form a measurement ‘chain’ from the microwave standard up to the visible.
Veniamin Chebotayev (Novosibirsk) and Theodor Haensch (Stanford) in the late 1970s pointed out that if one had a series of pulses regularly spaced in time, then a spectral analysis of this broadband light would reveal a regular series of frequencies, much like the teeth of a comb. Professor Haensch’s group made some pioneering measurements at Stanford, which clearly suggested the possibility and utility of such an approach. But the calibration of the comb needed another 20 years to become easily available, based on the very short-time pulses emitted by Titanium-doped Sapphire crystal lasers operating in a ‘self-mode-locked’ regime. These had stable repetitive pulses, very short in time (only a few optical cycles!). The peak power, corresponding to the very short times, was exceedingly high and offered a possibility for high-order non-linear possible responses, without destruction of the non-linear material.
The final critical technology was the development of special optical fibres, where detailed internal structures surrounding the tiny fibre core could be designed to affect and finally control the variation of travelling speeds for the range of colours within pulses of light. In 1999 a Bell Labs group showed white light being produced from coherent laser pulse input to such a micro-structured fibre, and the scientific community then had the final technique/component needed to make optical frequency measurements possible and even convenient. Professor Haensch’s group (by then he was at the Max Planck Institute for Quantum Optics in Munich) and my JILA group in Boulder were abruptly put into a friendly - but intense - competition to bring all these tools together.
The resulting Optical Frequency Comb technology was celebrated by our joint publication less than a year afterward. Now, rather than a multi-laboratory set-up with at least a half-dozen PhD’s attending fussy lasers of many types, there is even a commercial realisation of the Optical Frequency Comb measurement system in an ‘airlines luggage’ scale of package. A new laser can be brought into the lab, unpacked, and measured accurately (how about 15 digits?) before lunch, and perhaps even before morning coffee time. It is truly a powerful and valuable innovation! By now there are a half-dozen realistic candidate atoms being studied as possible future replacement for the Cesium frequency standard, which already is giving 15 digits accuracy and is by far the most reproducible of the basic standards represented in the seven basic standards of metrology (mass, length, time, temperature, quantity of material, electrical current, and optical radiative intensity.)
It is breathtaking to think that after 40 years of hard-won incremental advances in Cesium-based clocks, at least four of the new optically-based clocks show capability a decade and more improved over these ultimate Cesium results - and this is only the opening volley from the Optical Frequency Standards! I definitely want to watch the impact of this new Optical Frequency Comb innovation for the next few years - it’s going to be something else. I wonder if possibly this will represent the largest single-step advance in measurement science. Perhaps these factors were considered in the selection process leading to the 2005 Nobel Prize choices.