Researchers generated experimental images of an artificial exoplanet using a novel coronagraph, finding that for star-planet separations above a tenth of the diffraction limit, the images closely matched theoretical predictions. [Image: Boris SV/ Getty Images]
Scientists in the United States have shown how a coronagraph using what is known as spatial mode sorting could be used to image extrasolar planets at quantum limits (Optica, doi: 10.1364/OPTICA.545414). They say that their scheme should make it easier to study very faint exoplanets, including those similar to Earth.
Toward direct imaging
Researchers to date have discovered nearly 6,000 planets orbiting stars other than our sun. Most of the time, however, they have done so using indirect methods, such as measuring a slight dip in a star's output as a planet passes in front of it. These techniques limit what can be learned about an exoplanet's atmosphere, temperature or gravity, for example.
Direct imaging, which captures the starlight reflected off a planet, provides more detailed information on such properties but faces two major challenges. One is the fact that light from an exoplanet is drowned out by that from the star it orbits—in the case of Earth-like planets with orbits potentially conducive to life, these photons account for only about one in every 100 billion detected. The other problem is resolution, given that exoplanets are often separated from their host stars by less than the minimum resolvable distance of space-based telescopes.
Blocking star light
In the latest research, Nico Deshler, University of Arizona; Itay Ozer, University of Maryland; and their colleagues have shown how these problems can be mitigated by using a coronagraph to block direct light from a star while allowing the faint signal from an accompanying exoplanet to reach the detector. In contrast to previous work, the coronagraph blocks only a telescope's fundamental spatial mode, but photons in higher-order modes (crucial for subdiffraction imaging) get through.
The researchers’ “quantum-optimal coronagraph” consists conceptually of two sets of optical mode sorters. Light from the distant scene is focused onto the first sorter, where it is split up into a number of point spread functions—of the fundamental mode upward. All modes apart from the fundamental one then pass through to the second sorter, where they are recombined to form an image at the detector. This image should show the exoplanet but not the host star.
Deshler and colleagues implemented the scheme experimentally using just a single mode sorter made from a multi-plane light converter. Incoming light passes through the sorter, where it is spatially demultiplexed, and then each mode is focused to a blurred spot at a slightly different position on a “sorting plane.” A pinhole mirror is positioned in this plane so as to transmit only the spot corresponding to the fundamental mode, with that light then sent to a beam dump and the higher modes instead bouncing off the mirror before passing back through the mode sorter. The recombined modes create an image on a detector, which is analyzed using a computer.
Experiments and future improvements
At larger separations the signal began to win out, and the recorded intensity profiles came to closely resemble those calculated from theory.
To put their system to the test, the researchers exposed it to two points of light—a brighter one representing the host star and a second, much dimmer one representing the exoplanet that they positioned either slightly above or below the first point. Stepping the second point up and down over very small increments, they recorded and analyzed each resulting image.
They found that when the artificial exoplanet was very close to its star—less than a tenth of the Rayleigh diffraction limit—most photons from the exoplanet were discarded along with those from the star, and experimental noise obscured the exoplanet signal. However, at larger separations the signal began to win out, and the recorded intensity profiles came to closely resemble those calculated from theory.
Deshler and colleagues also established how accurately they could locate the planet compared with a theoretically perfect coronagraph. Setting the star to be 1,000 times brighter than the planet and using a maximum likelihood estimator to analyze their images, the researchers found they could get to within a few percent of the theoretical best over a wide range of sub-diffraction planetary positions.
Being suited to detecting such subdiffraction exoplanets, the researchers reckon their approach could be used to probe Earth-like planets most likely to harbor life. They point out that their experimental system is currently limited by a number of factors, including cross-talk between different optical modes and dark noise in their detector. But they believe these shortcomings could be overcome by employing higher-fidelity optical devices.
They add that their coronagraph could potentially be adapted to deal with the finite size of real stars, as well as to analyze broadband light from exoplanets, picking up specific chemical signatures in their atmospheres. The team notes that broadband sources complicate detection through wavelength-dependent cross-talk, but the researchers say this problem could potentially be overcome by sorting spatial and spectral modes simultaneously.