New Method Pulls a Giant Planet From Webb Telescope Glare

Astronomers have found a way to remove much of the stray light and instrument wobble from James Webb Space Telescope images, and the technique pulled a giant planet out of data where it had been lost in glare. The results were published Nov. 4 in the journal Nature, according to Live Science.

The team hopes the method will help reveal faint planets that look more like those in our own solar system, which current telescopes struggle to see.

What did the astronomers find in the Webb data?

The researchers, led by Rodrigo Ferrer-Chavez of Northwestern University, reanalyzed Webb data from three stars. One of them was 51 Eridani, which sits about 97 light-years from Earth and was already suspected to host a giant planet.

That planet, 51 Eridani b, had been detected earlier by ground-based telescopes. In previous Webb observations, however, it was drowned out by starlight. After the team processed the data with its new approach, the planet stood out clearly, Live Science reported.

Why is it so hard to see a planet next to its star?

Ferrer-Chavez said planets are a million times fainter than the stars they orbit, so imaging them requires removing the glare. Webb uses coronagraphs, instruments that block part of a star's light, a bit like the way the moon covers the sun in a total solar eclipse.

Even the best coronagraphs cannot block all of the light. The telescope itself also moves: even very stable instruments vibrate almost imperceptibly while collecting light. Those tiny shifts leave small mismatches in the images, and they can be enough to hide even very large planets.

How does the new technique work?

The team built a digital replica of Webb that virtually recreates how light travels through the telescope's detectors, without the imperfections caused by movement. To fine-tune the replica, the researchers combined computational models grounded in real-world physics with machine learning techniques similar to those used to train AI.

They named the approach Physics-based Learning for Imaging Planets, or FLIP.

Is this an artificial intelligence model?

The authors say no. Jason Wang, a Northwestern astrophysicist and study co-author, said the physics is already understood, so AI is not needed to work out what is happening inside the instrument. The machine learning serves to tune a model that is built on known physics.

Why does it matter?

The result matters for what astronomers can currently detect. In the nearly 35-year history of exoplanet hunting, most planets found have been very different from those in our solar system. Part of the reason, according to Live Science, is that larger and more reflective worlds are the easiest to spot with today's telescopes.

Ferrer-Chavez said solar-system-like planets are not missing from the universe. In his words, the problem is that "we can't see them." The team hopes FLIP will help astronomers pick out more of these hard-to-spot worlds in existing and future Webb data.

For general readers, the work does not change what is known about planets close to home. It shows that software and physics-based modeling can extend what a telescope already in space is able to reveal, without any new hardware.

What are the caveats?

This is a single study, though it has been peer-reviewed and published in Nature. The planet in the demonstration, 51 Eridani b, was already suspected and had been detected from the ground, so the test showed the method could recover a known world from Webb data. The hope that FLIP will turn up more elusive planets is a stated goal of the researchers, not a result yet reported.

Read next
Webb Telescope Finds Dense Star Clusters That May Have Birthed Black Hole Seeds
Webb Photo Shows Golden Galaxies Bending Light From the Early Universe
Webb Peers Through Dust to Reveal Young Stars Hidden in a Stellar Nursery

Around Prague, based on reporting by Live Science