Astronomers Built a Digital Twin of the Webb Telescope. It Found a Planet Hiding in Plain Sight

Astronomers have created a program to find exoplanets that we have just missed before.

Sometimes, photographing a planet around another star is a bit like trying to spot a moth fluttering beside a lighthouse beam. The moth is there; you just can’t see it, because the lighthouse drowns everything around it in light.

Now, astronomers have found a clever way to switch off the lighthouse. Instead of building a bigger, steadier telescope, they built a virtual one. Using a “digital twin” of the James Webb Space Telescope’s Near Infrared Camera (NIRCam), researchers modeled stellar glare so precisely they could essentially wipe it from Webb’s images in post. When they did, a giant planet that had sat unnoticed in two-year-old Webb data popped right out.

The findings were published in the journal Nature.

Lost in the glare

Most known exoplanets were detected indirectly. Imaging one directly is far harder, because of the sheer contrast.

“Planets are a million times fainter than their stars,” said Northwestern’s Rodrigo Ferrer-Chavez, who led the study. “So, to image exoplanets, we need to remove that glare. That is a very tricky problem because even the tiniest changes in the telescope can affect how that glare appears. Slight shifts or tilts — even on the scale of just a few nanometers — can change the pattern of light. Our method models how light travels through the telescope and accounts for those tiny changes, allowing us to remove the glare much more precisely.”

Rather than leaning on reference photos alone, the researchers simulated how starlight travels through Webb, bouncing off its mirrors and landing on its detectors. Then they let the computer hunt for the exact set of flaws and fluctuations that would reproduce what the real telescope recorded. Get the model right, and you know precisely what the star’s glare should look like — so you can remove it cleanly.

The team calls its method FLIP, short for Physics-based Learning for Imaging Planets. FLIP creates a “digital twin” of the telescope. A virtual replica that simulates how starlight travels through the telescope’s mirrors and reaches its detectors.

Co-author Katie Bouman of Caltech looks forward to what the new process can accomplish.

×

Get smarter every day...

Stay ahead with ZME Science and subscribe.

Daily Newsletter
The science you need to know, every weekday.

Weekly Newsletter
A week in science, all in one place. Sends every Sunday.
No spam, ever. Unsubscribe anytime. Review our Privacy Policy.

Thank you! One more thing...

Please check your inbox and confirm your subscription.

“What’s exciting is that we’re showing how computation can help overcome limitations of the telescope itself,” Bouman said. “By combining a physical model of the instrument with optimization techniques from machine learning, we have been able to remove starlight much more precisely, making it easier to detect faint exoplanets and even reveal one that was previously hidden in the JWST data.”

There’s a catch, however: A model this flexible could explain away a real planet as glare. So the team planted fake planet signals in the data and confirmed FLIP recovered them.

A planet hiding in the archive

To put FLIP to the test, the team reanalyzed existing Webb observations of two stars with known planets, HIP 65426 and 51 Eridani. FLIP proved two to 15 times more sensitive to faint planets than standard techniques.

This is how they got to 51 Eridani b, a young Jupiter-like world first spotted in 2015 from the ground. Webb had observed the system at a wavelength of 4.3 microns, a band astronomers use to sniff out carbon dioxide in planetary atmospheres. Yet when the data were processed the conventional way, the planet simply wasn’t there.

Run through FLIP though and the same data revealed 51 Eridani b exactly where, and how, it should appear. It’s the first time the planet has been seen at that wavelength. At the planet’s position, FLIP improved sensitivity roughly fourfold.

It isn’t a new planet, but that makes it a perfect test: astronomers knew the answer, and FLIP found it where older methods failed.

“It was science left on the table because we didn’t have the right tools,” said Jason Wang, an assistant professor of physics and astronomy at Northwestern and study co-author.

RelatedPosts

Machine Learning Helps Astronomers Find 10,000 New Planet Candidates in Existing Data
Swiss-European satellite CHEOPS finished its test phase and will start peering out in space
The Exoplanet that grew a second atmosphere
Earth’s Longest Ice Age May Have Repeatedly Thawed and Refrozen for 56 Million Years

The project grew out of a setback. Ferrer-Chavez first tackled the problem as an undergraduate at Mexico’s Autonomous University of Yucatán, in a thesis co-mentored by Wang and Bouman. None of his approaches worked, but Wang credits those dead ends with paving the way.

The payoff could be considerable. Because FLIP works on data already collected, astronomers can now comb through Webb’s archive for planets that were imaged but never noticed. And since the technique isn’t tied to Webb’s particular design, the team plans to adapt it for NASA’s Nancy Grace Roman Space Telescope and for ground-based observatories.

The biggest prize lies further out. NASA’s proposed Habitable Worlds Observatory aims to directly image Earth-like planets around Sun-like stars, roughly 10 billion times fainter than their hosts. If software can mop up instability after the fact, the engineers building it may get some breathing room.

Right now we don’t actually know of any planets around other stars that resemble the ones in our own solar system. But FLIP could flip the script on that problem.

“It’s not that they don’t exist. It’s that we cannot see them,” Ferrer-Chavez said.