Two Sydney students unlock a sharper view from the James Webb Space Telescope
Two innovative students from Sydney have led a globally significant effort to sharpen the James Webb Space Telescope’s vision. By tackling onboard instrument distortions, they boosted the telescope’s ability to detect Earth-like planets and observe distant cosmic phenomena with unprecedented clarity. The project combined engineering prowess, astrophysics insight, and advanced computation to extend JWST’s scientific reach from a lab to the vastness of space.
From a hardware hitch to a digital breakthrough
The work centers on the Aperture Masking Interferometer (AMI), a precisely engineered metal device that slots into JWST’s cameras to boost resolution. AMI was designed to counter disturbances that blur ultra-delicate measurements of light. Yet at very high resolutions, unforeseen electronic distortions created a “bl brighter-fatter” effect, where light overloaded pixels and smeared the data. This left researchers facing crisper edges in theory but fuzzier images in practice.
Enter Dr. Louis Desdoigts and his colleague Max Charles, both PhD students at the University of Sydney. Desdoigts spent two years developing a massive software solution, while Charles applied it to real observational data. Their mentor, Professor Peter Tuthill of the University of Sydney’s School of Physics, supervised the project and provided the conceptual backbone for AMI’s role in JWST’s suite of instruments.
As the team confronted the problem, Associate Professor Benjamin Pope of Macquarie University suggested a bold approach: build a digital twin of AMI—a detailed, physics-informed computer model that could simulate how the instrument behaves. By understanding how AMI should behave, the team could reverse the distortions that produced blurry results. Desdoigts described the effort as a fusion of physics and machine learning, a method that translates intricate instrument physics into a practical image-restoration tool.
Testing the model on spectacular cosmic targets
Charles put the model through its paces, processing data from three celestial objects. He focused on a striking spiral pattern of gas and dust around a pair of stars in Cygnus, some 6,000 light-years away, and on a jet of matter racing around a black hole at a galaxy’s core. In addition, the team captured the volcanic craters and crisp rim of Jupiter’s moon Io, demonstrating that the corrected images could reveal fine-scale features previously blurred by instrumental effects.
“We know what Io should look like from spacecraft that orbit Jupiter,” Charles noted. “Seeing it so clearly from JWST—despite the telescope’s greater distance—was a powerful proof of concept.” For Desdoigts, the achievement was a validation of patient, methodical work in the lab: “It’s incredibly rewarding to see a software solution extend the telescope’s scientific reach—without ever leaving the lab.”
Impact and future potential
The fix for AMI is more than a single success. It unlocks JWST’s potential to detect extremely faint objects near bright sources, a category central to finding exoplanets in habitable zones and studying faint galaxies at the edge of the observable universe. Tuthill emphasized that the approach could benefit other instruments beyond AMI, suggesting widespread implications for telescope technology worldwide.
Since their groundbreaking work, Desdoigts has continued in academia as a postdoctoral researcher at Leiden University in the Netherlands, while Charles joined him for a recent Europe trip. The two even commemorated their collaboration with matching tattoos—a lighthearted reminder of their shared achievement and the instrument that helped them reach it.
Closing thoughts
“This is way more widely valuable than a single achievement,” said Tuthill. The Sydney team has demonstrated that a thoughtful blend of physics-based modeling and machine learning can unlock latent capabilities in existing technology, potentially accelerating discoveries of Earth-like worlds and deeper insights into the cosmos. As JWST continues its mission, the doors opened by this software could lead to a future of sharper, more informative cosmic snapshots.
