Digital twin unveils hidden exoplanets

An international team led by Northwestern University has developed an innovative computational method to identify exoplanets hidden by the glare of their host stars. The technique, called Physics-based learning for imaging planets (FLIP), functions as a "digital twin" that creates a virtual representation of the stellar system.

Applied to observations already acquired by the James Webb Space Telescope (JWST), this technology has increased planet detection sensitivity by up to 15 times compared to previous methods. The results of the study, published in the journal Nature, allow for the correction of microscopic instrument imperfections that hindered the detection of the faintest objects.

The search for planets beyond our Solar System, known as exoplanets, represents one of the most important frontiers in modern astronomy. Direct imaging, which involves taking actual photographs of these distant worlds, is particularly challenging because the planet's light is extremely faint compared to the intense glare of the star it orbits. For decades, this technical problem has limited the discovery of planets, especially those that are smaller or farther from their star.

The James Webb Space Telescope, launched in 2021, represented a technological leap in our ability to observe the universe, but even its most advanced instruments must contend with the problem of stellar glare. The new Flip technique was born precisely to overcome this historic limitation: instead of relying solely on reference images, it combines physical modeling with machine learning to digitally subtract the star's light. This approach finally allows the revelation of planets that were previously completely hidden, opening new possibilities for characterizing Earth-like extrasolar worlds.