Scientists in the United States and New Zealand have shown how a new type of frequency comb based on dual pump lasers has the potential to finally transform comb technology from a niche scientific instrument into mass-market chip-based products (Nature, doi: 10.1038/s41586-026-11086-2). They say that their octave-spanning, self-aligning microcomb opens up the possibility of compact devices for navigation, timekeeping and sensing.

Stretching past an octave

Frequency combs act like electromagnetic clockwork mechanisms, enabling otherwise hard-to-measure optical frequencies to be pinned down using electronic counters. A comb generates rapid-fire laser pulses that in the frequency domain consist of a broad spectrum of equally spaced very narrow teeth. The value of any tooth can be calculated by multiplying its position in the series with the constant frequency spacing—equal to the pulse repetition rate—and then adding a frequency offset. It is thanks to the fact that the repetition rate usually lies in the microwave portion of the spectrum that optical frequencies can be recorded electronically.

These devices were first realized in the 1990s by tapping the output from bulky mode-locked lasers. More recently, researchers have shown how to generate the toothy output by directing light from continuous-wave pump lasers at tiny ring-shaped resonators fabricated on chips and exploiting nonlinear interactions to generate a multitude of sidebands around the pump frequency.

A major challenge in making frequency combs is the need to measure the frequency offset—in effect the gap between the lowest end of a device’s spectrum and its first tooth. This can be done by doubling the frequency of the lowest tooth and generating beats between that and a tooth an octave higher (since this frequency difference equals the offset frequency). However, even the output from most laser materials or resonator platforms falls well short of an octave.

Physicists John Hall and Theodor Hänsch won the 2005 Nobel Prize in Physics in part for showing how to use a photonic crystal fiber to stretch the spectrum from a titanium-sapphire frequency comb over more than an octave. Since then, other researchers have found they can span an octave with the output from chip-integrated microcombs that produce isolated light pulses known as dissipative Kerr solitons. However, because these devices produce a cascading series of sidebands from a single low-power seed pump, the teeth at the comb’s extremities are usually weak and noisy, requiring additional apparatus to reliably measure the frequency offset and limiting their use in real-world applications.

More well-defined teeth

In the latest work, Grégory Moille and Kartik Srinivasan,  of NIST/University of Maryland, USA, Miro Erkintalo, University of Auckland, New Zealand, and colleagues from other US institutions show how this shortcoming could be overcome by instead exploiting what are known as parametrically driven cavity solitons from resonators driven by a pair of continuous-wave pump lasers. The scheme results in a comb that spans the gap between the frequencies of the two pumps, yielding sharp, well-defined end teeth.

Building on work carried out in Belgium and the United Kingdom in 2021, as well as their own recent proof-of-principle experiment, Srinivasan and colleagues have now shown the commercial potential of such an approach by demonstrating a couple of key practical ingredients. One of these was to produce the new variety of soliton using on-chip pumps with outputs separated by an octave. The other was to show that such a device can generate a single frequency comb across the full spectrum defined by the pumps, with Kerr nonlinearities aligning what would otherwise be three distinct offset frequencies.

Demonstrations and next steps

The researchers carried out their demonstration using microring resonators with diameters of a few tens of micrometers fabricated in silicon nitride on commercial four-inch wafers. They showed that microrings fabricated at different positions across a wafer, or even on different wafers, all produce self-aligned frequency combs, with the only difference being slight shifts in the absolute pump frequencies due to variations in the thickness of silicon nitride. They say that these shifts can be compensated for by integrating tiny heaters into the devices, requiring of the order of 100 milliwatts of electrical power.

They argue that such devices could have multiple applications thanks to the flexibility provided by the nearly harmonically related pump lasers (one having almost double the frequency of the other). With the frequency offset established off-chip by measuring the beat frequency between the two pumps (the lower of the two being frequency doubled), they say they now have an additional degree of freedom to stabilize repetition rates or lock the device’s output to a timekeeping atom.

In their lab, Moille, Srinivasan and coworkers demonstrated what they describe as the three core tasks of frequency-comb metrology. First of all, they demonstrated how to generate specific optical frequencies by working up from given microwave standards. Then, they operated their clockwork in the opposite direction, producing low-noise radiation at micrometer or millimeter wavelengths from their low-noise lasers. Finally, they stabilized the ticking of a microwave clock by locking the frequency comb’s output to the radiation from rubidium-87 atoms in a vapor cell.

Toward real-world use

The researchers acknowledge, however, that their technology as it currently stands is not yet ready for the real world. Their device has a high repetition rate, which they say is a virtue insofar as it limits the density of a comb’s teeth and thereby prevents excessive dilution of the pump power. But such repetition rates are too high to be measured electronically, meaning they currently have to add a bulky phase modulator to their setup in order to step the frequency down.

They argue that instead pairing their comb with a low-repetition-rate device provides an “efficient path to field deployment” by keeping the package’s overall size, weight and power down to reasonable levels. What’s more, they add, their self-aligned comb scheme could be applied to a low-repetition-rate comb spanning a single octave, should it be possible in future to build and operate such a device economically.