Ancient sundials read the time from the position of the Sun. Modern high-precision optical atomic clocks, by contrast, keep time using atoms in a laboratory, shielded from environmental disturbances. Now a team from the University of Colorado Boulder, the U.S. National Institute of Standards and Technology (NIST), Pennsylvania State University and other institutions reports that it has built a "solar optical clock" that uses sunlight itself as a frequency reference. The preprint was posted in October 2026. The team measured the center of a solar absorption line to one part in a billion of the optical frequency, and also detected solar oscillations and changes during an eclipse. The aim is not to keep accurate time. It is to apply the precision measurement techniques of optical atomic clocks to astronomical observation, with an eye toward finding Earth-like exoplanets.
How a laser is locked to an iron absorption line in the Sun
The research team of Connor Fredrick, Ryan K. Cole, Ryan Terrien, Suvrath Mahadevan and Scott A. Diddams posted the preprint "A Solar Optical Clock" to arXiv on October 20, 2026. The device does not generate clock ticks directly from sunlight. Instead, the team exploited the fact that neutral iron atoms (Fe I) in the solar atmosphere absorb light at a wavelength of 1565.28 nm. By locking a laser to the center frequency of this absorption line, they can measure tiny changes in the frequency of sunlight.
In a conventional optical atomic clock, a laser is locked to the transition frequency of atoms whose environment is tightly controlled, for example by laser cooling in a vacuum. In this device, the absorption line of iron atoms in the solar atmosphere serves as the frequency reference. The frequency of the locked laser is measured with an optical frequency comb synchronized to a reference based on the SI second. An optical frequency comb uses many lines of light spaced at regular frequency intervals to measure the optical frequency of lasers and other sources with great precision.
The research team and New Scientist's report describe it as the first optical clock to use sunlight. However, the purpose of the device is not to keep accurate time like an ordinary clock, but to study how the Sun's absorption line changes. According to an explanation from the University of Colorado Boulder research group, conventional high-precision optical atomic clocks achieve frequency stability and accuracy at the level. The figure reported here is the precision with which the center frequency of the absorption line was measured, and is not a number that can be compared on equal terms with a clock's stability or accuracy.
Thermal heterodyne detection and averaging over thousands of seconds reach 30 cm/s
The setup uses a solar tracker and light-collecting optics to feed sunlight into a single-mode optical fiber. The optics are designed to capture light from the entire solar disk as uniformly as possible. The collected sunlight is mixed with laser light acting as a local oscillator and measured using a technique called thermal heterodyne detection.
Heterodyne detection overlaps two light waves of different frequencies and reads the frequency information of the original light from their difference. The team modulated the laser frequency at 5 kHz and used lock-in detection to generate an error signal showing how far the laser is from the center of the iron absorption line. This signal continuously corrects the laser frequency to hold it at the line center. The frequency of the locked laser is then counted precisely via the optical frequency comb.
Measurement precision depends heavily on the averaging time. With one-second averaging, the center frequency of the absorption line was measured to about 3 MHz. Converted to line-of-sight velocity, this is roughly 4.7 m/s, meaning the iron absorption line, which is about 4.3 GHz wide, was resolved to about 1/1,400 of its width.
When the team extended the averaging time to several thousand seconds and removed a drift proportional to time from the data, the precision approached about 200 kHz. This corresponds to a line-of-sight velocity of about 30 cm/s, or (one part in a billion) of the absolute optical frequency. The researchers attribute the drift to a combination of the orientation of the Sun's rotation axis as seen from the observation site and slight asymmetries in the light-collecting optics.
In the team's numerical simulations and analytical calculations, the precision obtained in the experiment matched the theoretical photon-noise limit for thermal heterodyne spectroscopy. This means measurement precision is limited mainly by statistical fluctuations in the incoming photons. The achievement lies in reaching high frequency-measurement precision on a single solar absorption line; it does not show that sunlight is stable enough to serve as a time standard.
Detecting the Sun's 5.5-minute oscillations and velocity changes during an eclipse
The team verified that the device can capture real solar physics through two observations. First, they binned the Sun's Doppler velocity data in 100-second intervals and analyzed the power spectral density, finding a clear peak at a period of about 5.5 minutes. This corresponds to solar acoustic oscillations (p-mode oscillations), in which sound waves traveling through the Sun's interior make its surface vibrate. The phenomenon is important for studying the Sun's internal structure and motion, and has long been investigated with existing solar instruments.
The second check used data from a partial solar eclipse observed in Boulder, Colorado, in April 2024. At maximum, about 80% of the solar disk was covered by the Moon. Because the Sun rotates, the Doppler shift of light differs between the side approaching Earth and the side receding from it. Normally, collecting light from the entire disk largely cancels the two effects, but when an eclipse blocks light from one side, that balance breaks and the apparent line-of-sight velocity changes. This phenomenon is called the Rossiter-McLaughlin effect, and it is also used in observations of exoplanets transiting their host stars.
The team used two instruments at the same time: the new device that continuously locks the laser to the center of the absorption line, and a conventional heterodyne instrument that sweeps the laser frequency to measure the whole absorption line. Both detected the Rossiter-McLaughlin effect during the eclipse. After removing the linear drift caused by solar tracking, the frequency change reached an amplitude of about 500 m/s in terms of line-of-sight velocity.
Under normal observing conditions, the continuous-lock method showed measurement precision nearly an order of magnitude higher than the sweeping method. The paper compares the results of both methods, but the absolute values corresponding to this difference in precision are not stated explicitly in the published materials we could verify. As the eclipse progressed and the amount of sunlight fell, the signal-to-noise ratio of the heterodyne signal dropped and the Allan deviation, a measure of frequency stability, rose. This change matched the decline in irradiance measured by the direct pyrheliometer mounted on the solar tracker.
The team did not go as far as using the eclipse data to estimate details such as the Moon's size or the Sun's surface structure. The goal was to show that precision measurement of a single absorption line, using optical clock technology, can capture velocity changes caused by real astronomical phenomena.
Prospects and challenges for finding Earth-like exoplanets
One application the research aims at is the search for Earth-like exoplanets. When a planet with a mass similar to Earth's orbits a Sun-like star, the planet's gravity makes the star move slightly. Detecting this motion through the Doppler shift of the star's light requires a line-of-sight velocity precision better than 10 cm/s.
Currently, few astronomical spectrographs can reach this level. In general, they split starlight by wavelength with large diffraction gratings and improve precision by averaging changes across hundreds to thousands of absorption lines. By contrast, this device reached a precision of about 30 cm/s using just one iron absorption line in sunlight. Unlike conventional approaches that average many lines, it can directly track the center frequency of an individual absorption line with high precision.
The team estimates that combining this technology with a 25–40 m class large telescope and adaptive optics could reach a precision comparable to this solar experiment for nearby Sun-like stars in a few hours of observation. Adaptive optics corrects light distorted by atmospheric turbulence and helps deliver starlight efficiently to the detector. However, the only celestial object actually observed in this work was the Sun, and the precision for distant stars is a prediction based on simulations.
Moreover, even if starlight can be measured with high precision, variations caused by the star's own activity pose a problem. In particular, granulation, a convective phenomenon in which hot gas rises and cooler gas sinks on the stellar surface, can produce line-of-sight velocity variations on the scale of tens of cm/s even over observations lasting a few hours. These may exceed the signal from an Earth-like planet, so improving instrument precision alone will not lead to planet detection.
The team therefore proposes simultaneously measuring multiple absorption lines that differ in magnetic sensitivity and in the depth at which they form in the stellar atmosphere. The idea is to exploit the fact that each line is affected differently by stellar activity, separating variations due to granulation and magnetic activity from the Doppler shift caused by a planet's gravity. By extending the single-line measurement demonstrated here to multiple lines, they also aim to improve precision without relying on long averaging times.
The experimental results and the figures for the applications envisioned by the team are summarized below.
| Target / system | Reported figure | Conditions | Basis |
|---|---|---|---|
| Solar optical clock (long averaging) | About 30 cm/s (about 200 kHz, relative precision ) | Solar Fe I absorption line (1565.28 nm), after averaging over thousands of seconds and removing linear drift | Measured result reported in the preprint |
| Solar optical clock (1-second averaging) | About 4.7 m/s (about 3 MHz) | 1-second averaging; about 1/1,400 of the 4.3 GHz linewidth | Measured result reported in the preprint |
| Precision needed to detect Earth-like exoplanets | Below 10 cm/s | Doppler velocity measurement of Sun-like stars with Earth-like planets | Target level; not achieved in this experiment |
| Conventional high-precision optical atomic clocks | Frequency stability and accuracy at the level | Atoms in a laboratory shielded from environmental influences | Track record of existing optical atomic clocks |
Piet Schmidt of Germany's national metrology institute, PTB, also took a cautious view in comments to New Scientist. Schmidt found the configuration of an optical clock that uses sunlight, rather than tightly controlled laboratory atoms, interesting, but pointed out that as an astronomical instrument it does not yet surpass existing stellar spectroscopy techniques. Because light from distant stars is faint, he said, use of this approach would be greatly limited. Indeed, in the eclipse observation the signal-to-noise ratio fell as sunlight dimmed, confirming that the small number of incoming photons constrains measurement precision.
Schmidt instead pointed to observations of the Sun itself as a more realistic application, for example high-precision studies of solar surface eruptions, sunspots and the effects of the solar wind. These phenomena were not directly measured in this work, but with the Sun's abundant light, precision spectroscopy using optical clock technology could become a new observational tool.
For exoplanet searches, the concrete next steps are actual stellar observations with 25–40 m class telescopes and the development of a system that measures multiple absorption lines simultaneously. Whether the team can separate stellar-surface variations on the scale of tens of cm/s and detect a planetary gravitational signal below 10 cm/s will largely determine the direction of the research.
