The search for signs of extraterrestrial civilizations has long centered on the classic Dyson sphere idea: if someone builds a giant structure around a star to harvest its energy, that structure should radiate excess heat in the mid-infrared range. For decades this has been the main tool for technosignature searches. A new preprint proposes a different kind of trace — not thermal, but mechanical.
The concept is called Stellar J-Harvesting, or "harvesting a star's angular momentum." The paper's author, Sahin Torlakcik, suggests that if an engineered system could extract a star's rotational angular momentum and convert it into energy, the star would end up spinning noticeably slower than its peers — neighbouring stars of the same age, mass and chemical composition. No excess infrared light would be required.
Taking rotation away from a star
Angular momentum is a physical quantity tied to how fast a body spins around its axis. Stars, like planets, are born with a certain reserve of it and gradually lose it through stellar winds — a natural process astronomers already model and use to estimate stellar ages.
The author derived a formula linking the energy that could be extracted to the resulting change in a star's rotation period. The paper outlines several hypothetical channels for such a coupling with a star, though the specific engineering mechanism remains an open question. What matters is the testable consequence: anomalously slow rotation compared with stars of the same colour and surface gravity.
Testing 6,725 Kepler stars
The search used data from the Kepler mission — a space telescope that spent years monitoring the brightness of hundreds of thousands of stars in search of exoplanets, gathering along the way a huge archive of stellar rotation periods.
Stars were sorted by colour and surface gravity, a step that allows comparing genuinely similar stars rather than the whole sample indiscriminately. The dataset was then run through eight false-positive filters, weeding out binary systems, unusual evolutionary stages and other sources of error. The clean sample that remained contained 6,725 main-sequence stars of classes F, G and K — stars similar to the Sun or somewhat hotter or cooler.
Two candidates, more ordinary explanations
Within the clean sample, two candidates stood out with deviations above 4 sigma — stars rotating significantly slower than statistically expected for their type.
Further checks tempered the excitement. Data from Gaia DR3, which measures the positions and motions of billions of stars, and imaging from the infrared telescope WISE pointed to entirely mundane causes: likely unresolved binary systems (two stars observed as one) or low metallicity, meaning a lower abundance of heavy elements, which affects how quickly a star loses angular momentum through its wind.
The main result is a search framework: angular-momentum technosignatures are testable with existing stellar-rotation catalogues, and the strongest outliers define concrete targets for spectroscopic, imaging, and radio follow-up.Sahin Torlakcik, study author
For that reason, the author makes no detection claim. Instead, the study sets a cautious upper limit: the occurrence of strong Stellar J-Harvesting signals among Kepler-type stars is fewer than 4.5 per 10,000 stars.
What the new method delivers
The main value of the work lies not in the two candidates but in the approach itself. It demonstrates that technosignatures tied to stellar rotation can be searched for using existing catalogues, without new telescopes or missions.
The strongest deviations from expected rotation do not simply disappear, either: they become concrete targets for follow-up observations — spectroscopy, high-resolution imaging and radio observations capable of ultimately confirming or ruling out a natural origin for the anomaly.
