≡ Menu

Exoplanet Imaging via a Fast New Camera

The world’s largest superconducting camera by pixel count has been deployed at the Subaru Telescope at Mauna Kea in Hawaii. This is a technology we’ll want to watch, for it assists the effort to image exoplanets directly from the surface of the Earth, a goal that not so long ago would have seemed impossible. But it can be done, and we have a new generation of extremely large telescopes (ELTs) on the way, so the progress in support technology for such installations is heartening.

The new device is called the MKID Exoplanet Camera (MEC), with the four-letter acronym standing for Microwave Kinetic Inductance Detector. A superconducting photon detector was first developed as far back as 2003 at Caltech and the Jet Propulsion Laboratory, paving the way for devices that can operate at wavelengths ranging from the far-infrared to X-rays. The MEC comes out of the laboratory of Ben Mazin at the University of California at Santa Barbara as part of an effort that includes contributions from both US and Japanese scientists.

The MKID Exoplanet Camera operates in the optical and near infrared, running at 90 millikelvin, or 1/1000th of a Kelvin, which is close to absolute zero. The technology involved can read out data thousands of times per second, according to the MEC’s developers, which plays directly into the success of adaptive optics systems that are designed to correct for atmospheric distortions. Current adaptive optics methods bend a telescope’s mirror at a rate of thousands of times per second, using complex algorithms to produce an image as it if were taken in space.

The problem: Planets most likely to be found with today’s adaptive optics are young worlds still glowing with the heat of their formation. Mazin points to HR 8799, a system with four gas giants, each of which is more massive than Jupiter, as the kind of catch currently available, and indeed, HR 8799 has been confirmed by direct imaging with the Keck and Gemini telescopes in Hawaii. The planets are still hot and glowing as the system matures. Moving into the range of smaller, cooler worlds will take exquisite collaboration between adaptive optics and the camera.

The MKID technology allows Mazin’s MEC to determine the energy of each photon as it hits the detector. Sarah Steiger is a UC-Santa Barbara doctoral student who worked on the project:

“This allows us not only to determine a planet’s brightness, but also to get a spectrum (the brightness as a function of energy), which can reveal additional information about an exoplanet’s properties, such as its age, mass and potentially atmospheric composition.”

Image: The 20440 pixel MKID device designed for MKID Exoplanet Camera is the highest pixel-count superconducting detector array at any wavelength. Credit: UC-Santa Barbara.

The fast data rates available with an MKID mean that the technology can work interactively with an observatory’s adaptive optics system to remove scattered and diffracted starlight, which allows the detection of exoplanets much fainter than can currently be imaged. In terms of astrobiology, says Olivier Guyon, that means we can one day turn the MEC to nearby exoplanets that can be characterized in greater detail than before. Guyon is the project scientist in charge of the Subaru Coronagraphic Extreme Adaptive Optics (SCExAO) instrument:

“We’re not going to be able to do that with Subaru, or with any of the current telescopes, because they’re just a bit too small. But we’re preparing for the next big step, which is to deploy exoplanet imaging cameras on larger telescopes such as the Thirty Meter Telescope. When those telescopes come online, the same technologies, the same camera, the same tricks will allow us to actually look for life.”

Ahead for Mazin’s team is the refinement of the software and algorithms that make MEC effective, with fast optical correction being the focus for the next several years. For more on the camera and its ongoing development, see Walter et al., “The MKID Exoplanet Camera for Subaru SCExAO,” Publications of the Astronomical Society of the Pacific Vol. 132, No. 1018 (17 November 2020). Abstract.


Comments on this entry are closed.

  • Enzo November 25, 2020, 15:01

    I hope that they are ready to work with E-ELT because TMT hasn’t even started construction and, in the current environment, I don’t think it will for a while.

  • Alex Tolley November 25, 2020, 16:22

    If the camera can detect the energy of the photons directly, and handle this at kHz speeds, does this imply that the data processing can exclude photons from the star and just leave the ones from the planet[s] to be imaged, effectively making a starshade or coronagraph?

    • Robert November 26, 2020, 14:08

      I should think you would still prefer the starshade or coronagraph and allow the photon discrimination algorithms to clean up the stray photons directly from the star. Otherwise the algorithms might be overwhelmed. Also, the photons from the planets also come from the star originally anyway so the changes on some may be subtle.

      • Alex Tolley November 26, 2020, 16:14

        The pap-er certainly has an image that used a coronagraph to hide the star so you are probably right. Much of the value seems to be removing the speckle noise to clean up the image and increase the contrast. It just seemed to me that with the data acquired, the computation could (in theory) replace the coronagraph. It could do a lot more in both astronomical and non-astronomical applications. I don’t know how the 2 processing algorithms work but it seems to me that this might also be a domain worth applying machine learning to improve the image quality.

  • Alex Tolley November 25, 2020, 16:36

    The arxiv version of the full paper detailing the technology can be found here.

  • Spaceman November 25, 2020, 17:35

    An exciting technological development, indeed. There has been a lot of talk as of late about how broadband internet satellite constellations, like SpaceX’s Starlink, will have a profound effect on ground-based observational astronomy in the coming years. To what extent, will these vast satellite constellations impact ground-based observations of exoplanets?

    • Wojciech J November 27, 2020, 18:51

      I believe there was a brief proposal that if those satellites do significantly alter observations, we should tax the companies operating them 1% of income to fund space observatories.

      • Robin Datta November 27, 2020, 21:41

        It might perhaps be prudent to make it standard practice to coat every satellite with carbon black or some equivalent paint job. And to make their outgoing communications either highly directional and well collimated towards terrestrial receivers or towards relay satellites at higher altitudes. EM “noise” pollution and space debris should be policed up with no hesitancy.

      • Michael Fidler November 27, 2020, 23:59

        Well, there is a lot of confusion as to what is taking place with these satellites. For one, the closer to the equator the less time in twilight, at 45 degrees north or south on June 21st the satellites are illuminated for a much longer periods. Two, the high slant angle of the sun on the bottom surface satellite at twilight would cause a perfect reflection off into deep space. Three, a number of baffles on the bottom surface like in telescope tubes should reject a large percentage of the light. Just coating the surface with a flat black may not be the best solution. Black silicon consists of clusters of microscopic vertical pillars, or nanowires, so incoming light bouncing between individual silicon nanowires cannot escape the structure and makes the material darker. I would think there may be many other possibilities.

        • Michael Fidler November 29, 2020, 10:19

          June 21/December 21

  • Robin Datta November 25, 2020, 18:32

    Sort of like a “hand count/audit” of all the incoming photons! By corraling the photons of interest, thus maximizing signal-to-noise ratio, the maximum fidelity for the available photons may be achieved.

  • Neil Stahl November 27, 2020, 11:03

    The comment about getting a spectrum of the planet seems a little misleading. I believe we can already do that. This seems to allow “us” to do that without a coronagraph, hence far more easily and more often. And when it’s more developed maybe from multiple planets at once?

  • Michael Fidler November 29, 2020, 10:16

    ABB and Nüvü to Deliver Exo-planet Cameras for NASA’s Roman Space Telescope.
    November 29, 2020
    A two-year contract awarded to ABB from NASA’s Jet Propulsion Laboratory will see key ABB/Nüvü Camēras technology fly onboard the space telescope in 2025, on course to capture the first spaceborne images of planets outside our solar system.

    The Nancy Grace Roman Space Telescope, NASA’s future space observatory, is due to launch in 2025 in search of other earth-like worlds. It carries two instruments: one to study the mystery of dark energy distribution in the cosmos; and the first dedicated exoplanet imaging camera in space, the CGI (CoronaGraph Imager). Within the CGI will be two high sensitivity cameras with electronic cores developed by ABB together with Nüvü.