≡ Menu

Photonic Chip Boosts Exoplanet Detection

The Australian Institute of Physics Congress ends today in Brisbane, concluding a schedule of talks that can be viewed here. Among the numerous research presentations was the description of a new optical chip for telescopes that should help astronomers tease out the image of a planet through thermal imaging, nulling out the light of the host star. The new photonic chip could be a replacement for bulk optics at the needed mid-infrared wavelengths.

Harry-Dean Kenchington Goldsmith, a PhD candidate who built the chip at the Australian National University Physics Center, says that the same technology that allows astronomers to penetrate dust clouds to see planets in formation will also be used to study the atmospheres of potentially life-bearing planets. ANU’s Steve Madden describes the chip as an interferometer that “adds equal but opposite light waves from a host sun which cancels out the light from the sun,” making it possible to detect the much fainter light of a planet. He likened the chip’s operation to the methods used in noise cancelling headphones.

This is important work because to understand planet formation, we have to learn more about how planets emerge from the gaseous disk surrounding the parent star. Doing this requires direct imaging at mid-infrared wavelengths, a method that has yielded young planets in formation as well as large exoplanets in distant orbits from their hosts. The next step will be to improve our methods of blanking out the star’s light for better imaging and, as this ANU news release points out, spectroscopic analysis of planetary atmospheres.


Image: The telescope chip developed at Australian National University. Credit: Stuart Hay/ANU.

It turns out that Ronald Bracewell (yes, that Bracewell) discussed these issues back in 1978, proposing what has come to be called the Bracewell nuller, which causes light from a star to destructively interfere, cancelling the star’s light so the planet can be observed. Thinking ahead to our own era of exoplanet discovery, Bracewell suggested using the method as a way to find planets around other stars. Along with a coronagraph (a different technique altogether), a nulling interferometer was considered for use in two high visibility but ultimately cancelled missions, NASA’s Terrestrial Planet Finder and the European Space Agency’s Darwin.

The ANU team brings nulling techniques to a photonic chip. This morning while digging around in the paper on this work, I received a comment from Centauri Dreams reader Mike Fidler, who pointed readers to a shorter (though highly technical) description of the chip in one of ANU’s publications. The key point here is that using photonic chips we can replace some operations normally handled through conventional optics. From the paper on this work:

Photonic integrated circuits are established as the technique of choice for a number of astronomical processing functions due to their compactness, high level of integration, low losses, and stability. Temperature control, mechanical vibration and acoustic noise become controllable for such a device enabling much more complex processing than can realistically be considered with bulk optics.

These are major benefits, making this a technology we’ll continue to follow with interest.

Ronald Bracewell’s original paper on nulling concepts was “Detecting nonsolar planets by spinning infrared interferometer,” Nature 274 (24 August 1978), 780–781 (abstract). For Goldsmith and Madden’s work, see Goldsmith et al., “Chalcogenide glass planar MIR couplers for future chip based Bracewell interferometers,” in Proc. SPIE 9907, Optical and Infrared Interferometry and Imaging V, 990730 (August 4, 2016) (preprint).


Comments on this entry are closed.

  • Alex Tolley December 8, 2016, 11:31

    A perusal of the papers Mike provided suggests a useful range of 3.8-4.2 microns in the IR. This range doesn’t seem extensive enough for the absorption spectrum of potential biosignature gases, e.g. water, oxygen, methane, etc.

    However, the technique looks very promising and I hope the useful range can be extended with different materials so that these chips can facilitate atmosphere characterization.

    As an aside, I wonder if, in principle, these interference chips could be used for microscopy.

  • Horatio Trobinson December 8, 2016, 11:38

    Massively, incredibly prescient, Bracewell.

  • Michael C. Fidler December 8, 2016, 19:42

    This should work well for the detection of exoplanets around M dwarfs and brown dwarfs, Proxima Centauri temperature is around 3000K and radiates well in the L band at 4 microns. The nearby brown dwarfs would also be good targets. So if we are lucky maybe they can get an image of Proxima Centauri b! It would be nice if NASA could do a upgrade to the Hubble telescope with this new photonic device, when it has matured. That’s another question, just how hard would it be to send a mission to Hubble with the new manned spacecraft coming online? Would a cargo vessel need to be sent up with the new instruments? It is relatively easy to reach the Hubble scope from low earth orbit so has anyone talked about the potential for upgrades?
    Found a good pdf file on habitability around Red Dwarfs.
    Living with a Red Dwarf:

  • Geoffrey Hillend December 8, 2016, 22:29

    Alley Tolley I don’t think that the chip is NOT limited to the 3.8 to 4.2 microns. It cancels all the light from the star and what is left is the exoplanet light. Are you sure that the 3.8 to 4.2 refers to the near infra-red which is dust penetrating since dust is invisible at the near infra-red in proto-planetary nebula?

    • Alex Tolley December 9, 2016, 12:37

      Did I misinterpret their paper?

      This from the abstract:

      Chalcogenide glasses are well known for their transparency to beyond 10000 nm, and the first results from coupler devices intended for use in an interferometric nuller for exoplanetary observation in the Mid-Infrared L’ band (3800-4200 nm) are presented here showing that suitable performance can be obtained both theoretically and experimentally for the first fabricated devices operating at 4000 nm.

      I thought figure 6 was confirmation of that range.

      If I misunderstood, then please correct me.

  • john walker December 9, 2016, 5:18

    Goldsmith’s paper talks of a “usable” contrast ratio without quantifying that. Odd. But, I’ve read of 10 -4 at 1500nm. If anyone knows of PIC nulling at visible wavelengths and associated contrast ratios I’d very much like to know more.

  • Jeff Wright December 9, 2016, 13:46
  • Geoffrey Hillend December 9, 2016, 17:30

    Alex Tolley. You didn’t misinterpret it. I didn’t read it so I didn’t understand the technology, so I assumed the chip could remove most of the starlight like a multiple space telescope would. https://en.wikipedia.org/wiki/Terrestrial_Planet_Finder
    I not an expert on telescope technology. Where is this abstract and what is the page number?

  • Alex Tolley December 9, 2016, 18:53

    Mike Fidler posted the link in the “Asteroid thought experiment” thread:


    “Chalcogenide Glass Planar MIR couplers for future chip based Bracewell Interferometers”

  • Geoffrey Hillend December 10, 2016, 20:09

    Alex Tolley Thanking for posting it. I don’t completely understand the technology. I don’t understand why it says that 3800 to 4200 nanometers or 3.8 to 4.2 microns is the Mid infrared. I thought the mid infrared was beyond 5 microns or 5000 nanometers which is the electromagnetic absorption region of some bio signature gases. http://www.ipac.caltech.edu/outreach/Edu/Regions/irregions.html

    At any rate it says that it is transparent to beyond 10000 nanometers which is what one wants since you want to cancel out the light closer to the visible range like the near infra red, so the mid infrared can pass through. Mid infrared gas absorption: Co2 absorbs at 15 microns, and H20 20 microns, Methane 7.6m . Oxygen is at 1.7 m.

    I don’t understand why would we want to null out the near or mid infrared? The visible light yes but the near infrared?

  • Geoffrey Hillend December 10, 2016, 20:14

    I think I got it. Maybe it is just the light of the star and not the exoplanets.

  • Benjamin Pope December 11, 2016, 13:38

    Lovely piece of research. I might mention this is not the only project doing on-chip nulling:



  • Ashley Baldwin December 11, 2016, 13:53

    Anything new can only help . Surely this involves the same principles as the more established and researched Visible Nulling Coronagraph which has can operate between shorter 550 -1500 nm and was central to the sadly defunct TPF-C concept ? Though much less developed it hopefully also offers the unique ability of the VNC ( amongst corongraphs ) to operate with segmented mirrors ( more likely to be the basis of any future large space telescope ) . The shorter wavelength of the VNC offers better resolution though . Though MIR , thermal wavelength, potentially offers more potent spectral characterisation and the ability to see through dust clouds . Most “biosignature ” gases have absorption lines across a wide range of wavelengths though and not just in the MIR . It would be ironically be more suited to the JWST or similar bandwidth telescope than Hubble which only operates into the NIR and isn’t likely to be economically upgraded even if it lasts to 2021 when its funding ends . This regardless of the hopeful return of manned missions to LEO .

    This technology sounds much less developed than the VNC and although a new technique it is still dogged by the same issues that have held back the VNC , namely the reduction in SNR as wavelength increased to even the NIR and the all limited size of its inner working angle ( how close to the star a planet could be imaged ) . More importantly and as with all coronagraphs there remains a need for extremely high quality mirrors and related precision wavefront control. This is traditionally compensated by deformable mirrors in the light chain , DMs . The latter technology is finally maturing to the point where such coronagraph designs might finally deliver efficiently with the introduction of 1024 actuator DMs, which will be as important to this new corongraph as any that went before . All helping to dispose of leaked star light “speckles” and also supplemented with the “final common pathway ” requirement for potent post imaging processing techniques .

    Much of the improvement has been slowly incremental and driven by the WFIRST concept to the point that I’m sure we will have the technology to directly image Earth sized habitable zone planets next decade with the limiting factor becoming finance . That said I’m sure the more potential light suppressing techniques available for exoplanet imaging the better as each one has its strengths and weaknesses and the common technology they all share/require is clearly being driven forward by the research into each design to the point where “critical mass” is reached and we finally have the common precision and stability required . The general principles behind most coronagraphs have been around for over a decade but it’s only been in the last year or so that the required hardware has begun to approach the required standard required. Exciting times approach .

    It may ultimately be that the perfection of all these coronagraph designs along with ACEsat and Project Blue like innovations show the way to low cost , high performance exoplanet imaging telescopes that are in themselves cheaper than risky $100 million plus servicing/ upgrade missions ( even to LEO let alone SEL2) . $120 million ACEsat already scales up to a 1.5m aperture & it was only a prototype .

  • Geoffrey Hillend December 11, 2016, 17:51

    In none of the above papers does it say that the photonic chip does not cancel light at the visible wavelength. In the press release above, they say the chip itself is a photonic interferometer which cancels the starlight so we can see the faint Infra-red or near infra red and so I have to assume that includes the visible band as well. It is not a Bracewell nuller which uses two or more telescope mirrors?

    Also to get technical, some biosignature gases like Co2 can absorb at different frequences depending on atmospheric pressure which is related to the altitude of absorption: rina.eas.gatech.edu/EAS8803_Fall2009/Lec6.pdf

  • Geoffrey Hillend December 11, 2016, 17:52
  • Geoffrey Hillend December 11, 2016, 18:04

    I really don’t expect that the photonic chip cancells all the visible light but if one says that it cancels out the light of the star to see the exoplanets in the press release then it is misleading to imply that it does if it really doesn’t which is why I assume that it does. Maybe it still needs and interferometer with two or more mirrors to do that. It would be nice if it didn’t. I also like to see if it does.

  • john walker December 12, 2016, 17:02

    @Benjamin Pope: The Dragonfly instrument was developed partly from fellow Australian researchers. A completely different approach to chip based “nulling” comes from FIT in Melbourne(Florida ;)) using a charge injection device (CID) as described here: https://arxiv.org/pdf/1511.03715v1.pdf