Making it easier to find ‘new Earths’

Australian researchers have combined artificial intelligence and photonics to develop a new type of adaptive optics sensor, with the potential to provide ground-based telescopes with a clearer view of planets in distant solar systems.

The results are described in a paper in the journal Nature Communications.

“The main way we identify planets orbiting distant stars is by measuring regular dips in starlight caused by planets blocking out bits of their sun,” explains Barnaby Norris, lead author of the study from the University of Sydney (USyd). “This is really difficult from the ground, so we needed to develop a new way of looking up at the stars.”

Exoplanets are notoriously difficult to study, because they can be up to a billion times fainter than their host star. Separating a star from a planet is about equivalent to standing in Melbourne and attempting to see a coin in Sydney.

Ground-based telescopes face an extra challenge because they must peer through the Earth’s atmosphere, where different temperature layers and interacting wind speeds can distort incoming light.

While this twinkling may delight poets, it’s a constant frustration for astronomers who would much prefer sharp, detailed images. One solution is to cut out the middleman and observe from above the atmosphere – most exoplanet observations, for example, have so far come from orbiting telescopes like NASA’s Kepler mission, with only a small handful made by ground-based telescopes.

But space telescopes are costly to run. Instead, astronomers and engineers have spent the last few decades developing adaptive optics technologies, equipping Earth-bound telescopes with deformable mirrors that can rapidly change shape in order to cancel out the effects of turbulence.

Such mirrors work in conjunction with wavefront sensors that can precisely measure the amount of blurring, and high-speed computers that can compute the corrections in real-time.

This new research uses photonic lantern technology to simultaneously conduct the optical light off the moving telescope and determine the necessary corrections.

“This new sensor merges advanced photonic devices with deep learning and neural networks techniques to achieve an unprecedented type of wavefront sensor for large telescopes,” Norris says.

Stuart Ryder, astronomer from Macquarie University who was not involved in the study, explains that a photonic lantern is like a “broad fibre optic ‘pipe’ for light, which gradually tapers into a bundle of much smaller fibres.”

“A turbulent image entering the photonic lantern causes changes in the relative brightness of the light emerging from the end of each of these fibres, and although this pattern tells us how this light was distorted going in, there’s no simple, fast way to calculate what this is. “

Here, artificial intelligence and machine learning step in – the system can be taught to recognise the links between distortion and a given light pattern, so corrections can be rapidly inferred.

“Unlike conventional wavefront sensors, it can be placed at the same location in the optical instrument where the image is formed,” adds Norris. “This means it is sensitive to types of distortions invisible to other wavefront sensors currently used today in large observatories.”

The new sensor is currently only a concept demonstrated in a laboratory. But there are plans to deploy it in the 8.2-metre Subaru Telescope in Hawaii – one of the largest optical telescopes in the world.

“This is no doubt a very innovative approach and very different to all existing methods,” says Olivier Guyon, astronomer from the Subaru Telescope and the University of Arizona. “It could potentially resolve several major limitations of the current technology.”

If the tests at Subaru are successful, then this new method may revolutionise exoplanet research, opening up new vistas from the ground.

However, Ryder notes: “This technique works best on and around bright, compact objects like stars. For fainter, extended objects like galaxies, we will still require an artificial, laser guide star to determine the right correction.”

Laser guide stars are current cutting-edge technologies in adaptive optics, used on telescopes such as ESO’s Very Large Telescope in Chile.

But photonic lanterns have great promise. Since they’re relatively small and cheap, multiple could be placed across a telescope’s focal plane to provide corrections.

“The corrected optical light can then be fed to a camera or spectrograph, providing Hubble Space Telescope-like data from the ground, something which is currently extremely difficult to achieve,” Ryder adds.

Intriguingly, the applications of this new sensor are not limited to stargazing.

According to USyd co-author Sergio Leon-Saval: “It could be applied in optical communications, remote sensing, in-vivo imaging and any other field that involves the reception or transmission of accurate wavefronts through a turbulent or turbid medium, such as water, blood or air.”