≡ Menu

Project Blue: Looking for Terrestrial Worlds at Alpha Centauri

Eduardo Bendek’s ACEsat, conceived at NASA Ames by Bendek and Ruslan Belikov, seemed to change the paradigm for planet discovery around the nearest stellar system. The beauty of Alpha Centauri is that the two primary stars present large habitable zones as seen from Earth, simply because the system is so close to us. The downside, in terms of G-class Centauri A and K-class Centauri B, is that their binary nature makes filtering out starlight a major challenge.

Image: The Alpha Centauri system. The combined light of Centauri A (G-class) and Centauri B (K-class) appears here as a single overwhelmingly bright ‘star.’ Proxima Centauri can be seen circled at bottom right. Credit: European Southern Observatory.

If we attack the problem from the ground, ever bigger instruments seem called for, like the European Southern Observatory’s Very Large Telescope in conjunction with the VISIR instrument (VLT Imager and Spectrometer for mid-Infrared) that Breakthrough Initiatives is now working with the ESO to enhance. Or perhaps one of the extremely large telescopes now in the works, like the Thirty Meter Telescope in Hawaii, or the Giant Magellan Telescope in Chile.

And if we did this from space, surely it would be an expensive platform. Except that ACEsat wasn’t expensive, nor was it large. It was designed to do just one thing and do it well.

While NASA turned down Bendek and Belikov’s idea for Small Explorer funding, the striking thing is that it would have fit that category’s definition. ACEsat was designed as a 30 to 45 cm space telescope (you can see a Belikov presentation on the instrument here, or for that matter, read Ashley Baldwin’s ACEsat: Alpha Centauri and Direct Imaging). The small instrument now being proposed by an initiative called Project Blue builds on many of the ACEsat concepts. It would run perhaps $50 million even though the original ACEsat was a $175 million design.

In other words, compared to the $8 billion James Webb Space Telescope, Project Blue’s instrument is almost inexpensive enough to be a rounding error. A privately funded initiative out of the Boldly Go Institute, in partnership with the SETI Institute, Mission Centaur, and UMass Lowell, the telescope shows its pedigree both in its low cost and big scientific return. It seems the ACEsat concept is just too good to go away.

So now we have Project Blue, which is all about seeing the blue of an Earth-like world around one or even both of the Sun-like stars of the Alpha Centauri system. No one discounts the value of the planet already discovered around Proxima Centauri, but the project hopes to find an Earth 2.0, a rocky planet in a habitable zone orbit around a star like our own. That would mean no tidal locking, no small red dwarf primary, and a year measured in months rather than days.

Image: An Earth-like planet around one of the primary Alpha Centauri stars, as simulated by Project Blue.

The project’s new Indiegogo campaign has been set up to raise $175,000 to help establish mission requirements, including the design of an initial system architecture to which computer simulations can be applied by way of testing ideas and simulating outcomes. The launch goal of 2021 is ambitious indeed, as is the low $50 million budget profile, but the project’s backers believe their work can leverage advances in the small satellite industry and imaging systems to pull it off. An explicit goal is to engage the public while tapping the original NASA work.

The project’s connection to NASA is in the form of a cooperative agreement explained on the Indiegogo site:

The BoldlyGo Institute and NASA have signed a Space Act Agreement to cooperate on Project Blue, a mission to search for potentially habitable Earth-size planets in the Alpha Centauri system using a specially designed space telescope. The agreement allows NASA employees – scientists and engineers – to interact with the Project Blue team through its mission development phases to help review mission design plans and to share scientific results on Alpha Centauri and exoplanets along with the latest technology tests being undertaken at NASA facilities. The agreement also calls for the raw and processed data from Project Blue to be made available to NASA within one year of its acquisition on orbit via a publicly accessible online data archive. The Project Blue team has been planning such an archive for broadly sharing the data with the global astronomical community and for enabling citizen scientist participation.

And I notice that Eduardo Bendek is among the ranks of an advisory committee (available here) that includes the likes of exoplanet hunters Olivier Guyon, Debra Fischer, Jim Kasting and Maggie Turnbull. But have a look at the advisor page; every one of these scientists is playing a significant role in our discovery and evaluation of new exoplanetary systems.

Thus we can say that ACEsat lives on in this new incarnation that will benefit from the input of its original designers. The spacecraft would spend two years in low Earth orbit accumulating thousands of images with the help of an onboard coronagraph to remove light from the twin stars, along with a deformable mirror, low-order wavefront sensors, and control algorithms to manage incoming light, enhancing image contrast with software processing methods.

Unlike the major observatories we’re soon to be launching — not just the James Webb Space Telescope but the Transiting Exoplanet Survey Satellite (TESS) — the Project Blue observatory will be dedicated to a single target, with no other observational duties.

A photograph of an Earth-like planet 40 trillion kilometers away gives us a sense of the changes in scale that have occurred since Voyager 1’s ‘pale blue dot’ photograph. But we already knew that Earth was inhabited. Now, gaining spectral information about a blue and green world around a nearby star would allow us to determine whether biosignature gases could be found in its atmosphere, potential signs of life that would mark a breakthrough in our science. The degree of public involvement assumed in the project makes the quest all the more tantalizing.


Comments on this entry are closed.

  • DJ Kaplan September 29, 2017, 14:01

    Of course, Sky & Telescope reports that the Hawai’ian 30-m telescope will continue construction:

    • Ashley Baldwin September 29, 2017, 16:59

      Good. Best temperate site viewing in the Northern Hemisphere (important given the southern location of the other ELTs ). Much better than La Palma though the best viewing of all would have been the mount adjacent ( & above ) the ALMA array at the high Atacama desert. There is a current Japanese proposal to put a 6.5m scope there in the next decade. Atmospheric water vapour content ( a major impediment to ground based optical and IR imaging) drops off precipitously over 5000m ( especially at the already dry Atacama) allowing for much improved imaging , especially IR , limited only by the background (“warm” ) sky ” subtraction” requirement -to about 5 microns.

      Though smaller ,the TMT will have a significantly superior imaging environment at the far higher altitude of Mauna Kea cf the E-ELT at Cerro Amazones. (The location of which ESO IR astronomers were most disappointed with )

      Just the high Antarctic ice “domes” ( C & F ) do better , but largely impracticable given their extreme remoteness.

    • Brett September 30, 2017, 1:16

      I’m hoping they move to La Palma anyways. Even if they win the appeal to the Hawaiian Supreme Court, it’s going to be a PR nightmare building it up there with all kinds of protests and such – and the price/conditions for building it are steep (they’re going to lose three telescopes if they do it).

      I thought they were going to move it anyways, but fought the case just to ensure that it doesn’t become a means to attack the other telescopes atop the mountain.

  • Ashley Baldwin September 29, 2017, 14:03

    Thanks for the kind citation Paul. Project Blue is a bespoke project 30- 45cm coronagraph aiming to image hab zone terrestrial planets around the two Alpha Centauri stars . It was largely for this parochial goal that it wasn’t adopted for one of NASA’s ($170 million) small Explorer programmes . NASA are understandably risk averse with their previous astrophysics budget. Largely but not solely. There was also another contributing factor. Technological maturity fpgiven coronagraph imaging telescopes require a number of critical new developments to operate as planned. This non exhaustive list includes the star occluding coronagraph , low and high order wavefront error management via adaptive optics ( cf on ground based scopes ) and telescope stabilisation essential for the many months of observation required to achieve the projects goals. At the time of ACEsat’s submission all of the above needed a lot of expensive development .

    The names mentioned in the article , such as Guyon, Bendek and Belikov in particular are world leads in these technologies . Thanks in no small part to the pre formulation work carried out for the WFIRST imager and coronagraph the hardware ( and software ) have now been matured to a NASA “Technological Readiness level” , TRL of 5 as per its mission timetable . To be considered flight worthy for NASA a TRL of 6 and beyond is the benchmark . This will be reached and surpasses within the next couple of years with coronagraph development ( the most essential ) in particular being speeded up by the recent introduction of the Decadel imaging Test Bed. At the same time the high precision deformable mirrors required to manage “speckles” ( stray light that escapes the coronagraph to reach and blur the imaging plane ) will also reach maturity along with the adaptive optics to manage low and high order wavefront errors . All this in combination with telescope stabilisation mitigation strategies ( micro thrusters and reaction wheel vibration isolation chambers) some of which will reach a high TRL when employed on JWST. Software algorithms crotical to imaging planets in close binary systems like Alpha Centauri is also being refined with the added advantage of being uploaded into any telescope space bus PC right up and after launch.

    Hopefully WFIRST will launch circa 2025 with an exoplanet coronagraph and imager ,CGI , on board . However there are currently two independent coat reviews ongoing to see whether the CGI can be included within budget. NASA does not need a repeat of JWST were the cost of the technological progression required was woefully underestimated with the inevitable and well publicised huge overspend. There can be no repeat if there are to be future ” great observatories” . These will report before the end of the year. Hopefully in the affirmative , but if not it may be that Project Blue or even a revamped ACEsat ( with suitably matured technology ) may yet fly in the “technological demonstrator ” capacity envisaged for the WFIRST CGI. With the cost of development now covered, it may be that a larger aperture could be built to allow imaging of other nearby Sun like stars making it a more attractive proposition for NASA astrophysics programme assessors .
    An interesting time ahead.

  • DJ Kaplan September 29, 2017, 14:06
  • Gary Wilson September 29, 2017, 14:30

    AceSat is a brilliant concept, and very affordable. I don’t understand why it wasn’t funded by NASA. We have to get to know our immediate interstellar neighborhood far better than we do at this point.

    • Ashley Baldwin September 29, 2017, 16:31

      To narrow a science focus on the Alpha Centauri system with a significant chance of finding nothing. This in combination with the substantial development required by many of its key technologies at the time of submission. See post above.

      It is possible that should the coronagraph not prove cost effective and get dropped from WFIRST, that all the technological development that went into maturing its exoplanet imaging won’t be wasted and a revised ( and much improved ) ACEsat concept submitted perhaps as a future Medium Explorer mission . The simple concept is robust and scaleable up to atleast 1.5m . Such a telescope would not just be limited to just Alpha Centauri . Adopting an ACEsat imaging approach at just three key “habitability ” wavelengths targetibg exoplanet discovery and low level atmospheric characterisation ,without an expensive spectrograph and sensor array, funding could be reserved to push up aperture and sensitivity . Current Medium Exporer short listed exoplanet characteriser FINESSE has a 0.75m aperture in front of a an NIR sensor and spectrograph. A larger aperture might be possible for the same budget by using the cheaper silicon carbide mirror substrate and shorter visible wavelengths both employed by ACEsat . Wavelengths detectable by established and cheap CCDs or even CMOS based sensor arrays rather than the much more pricey envelope pushing NIR HgCdTe equivalents required for the more detailed characterisation of FINESSE. At the cost of aperture size .

      Or maybe even as a significantly more potent Exo-C class funded Probe ( $ 1 billion plus launch and operations ) mission given the suggestion that such an astrophysics programme could come out of the next Decadel.

  • Spaceman September 29, 2017, 14:46


    Having you happened to hear anything new lately about how the Alpha Centauri B RV planet searches are progressing? Are RV teams looking for planets around just Alpha Centauri B or around Alpha Centauri A too?

    • Paul Gilster September 29, 2017, 17:03

      I don’t have anything new, although recall Debra Fischer’s comment in these pages in late June:

      “The stability of the echelle spectrograph at CTIO was not sufficient to obtain radial velocities with needed precision. We then obtained NSF funding to build CHIRON (commissioned in 2012) and started over! This spectrograph was stable over short time baselines, but by 2012, the projected separation of the two stars A and B was so small (about 5 arcseconds) that our spectra for one of the stars was always contaminated by light from the other.”

      The spectral situation will be getting better as the apparent separation grows. Will need to do a follow-up on RV and Alpha Centauri A and B some time soon — thanks for the idea.

      • Ashley Baldwin September 29, 2017, 17:27

        That widening gap will be key to the success of both RV and direct imaging techniques . In terms of Project Blue or better still ACEsat “heavy ” the hiatus isn’t necessarily a bad thing giving time to mature the technology ready for a future bid submission. Bendek, Guyon and Belikov are all driven . Driven even , to developing an exoplanet imaging mission of some type having devoted years of their takaented careers to maturing the technology and various interchangeable concepts necessary . Personally I feel that to come to fruition ( whichever way ) any concept will have to image more than Alpha Centauri alone.

      • Michael Fidler September 30, 2017, 10:44

        What about using nulling interferometry to null out Alpha Centauri A or B?

        • Ashley Baldwin October 1, 2017, 5:54

          Good question. Why not ? Essentially because it’s much simpler and cheaper to use preexisting deformable mirror that is part of the ACEsat optical train in combination with the Multi Star Wavefront Correction software algorithm . Their role is to remove stray star light that seeps past the coronagraph and into the telescope proper to create high order wavefront errors , “speckles” , that interfere with the final image. The two Alpha Centauri stars are sufficiently far enough apart that you don’t see the other one in the same field of view during imaging , just some of its light. The coronagraph functions to stop that in the same way as it blocks out the light of the star being directly observed .

          ACESat ( and Project Blue even more so ) is by necessity a very low budget attempt to do something that should cost billions ( like WFIRST or Exo-C) for just tens of millions . Nulling interferometry can been used in the same way as coronagraphy but is far less well developed and more complex so consequently expensive at this time though ultimately it should offer another ( perhaps better ) route to the same goal. There were high hopes for it a decade a go when TPF was still on the cards , but it seems to have fallen behind a bit during the latest round of technological development coming out of WFIRST preformulation work . This has prioritised Hybrid Lyot style coronagraphs and the PIAA version employed in ACEsat. -the latter because it has easily the highest light throughput to the image plane ,circa 90% -something extremely useful when you do have much to use or lose with a small aperture. By further embedding it in the secondary and tertiary mirrors of the telescope itself you reduce the number of optics in the optical train increasing throughput further and increasing imaging stability – utterly crucial in precision exoplanet imaging .

          Bendek and Belikov cover all of this ( even fielding a question similar to your own) in a one hour SETI Institute lecture on ACEsat that is available on YouTube and a great introduction to the whole concept of exoplanet direct imaging. Well worth a watch .

          • Michael Fidler October 2, 2017, 7:03

            Thanks Ashley, just donated to the project!
            Have been looking at multi exoplanet systems trying to see how K dwarfs stack up against G and M dwarf type stars. Thinking that they would be somewhere between the very compact systems of M dwarfs and the large systems like our Sun. Well that’s not the case, almost all the multi systems in this list – https://en.wikipedia.org/wiki/List_of_multiplanetary_systems – are close in systems with planets orbiting from less then 1 day to 200 days from the M to F class stars! Now I know Kepler was picking up the short period planets more often but the radials where also showing short periods. So is our Sun just a freak system or is this because of the bias toward shorter period systems? The systems with outer long period planets still had many close in short period planets. Does this bodes well for Alpha Centauri A or B since it’s binary and would ACEsat be able to pick them up?

          • ljk October 2, 2017, 10:17

            Here is the video of the ACEsat talk from The SETI Institute online:


            • Ashley Baldwin October 2, 2017, 15:41

              Bias to short period systems. Current RV spectroscopes like HARPS have at best a 50 cm/s sensitivity . The Earth would give a 10cms change , so we about are five times above finding terrestrial planets in the hab zones of Sun like stars . The floor is now determined not by spectroscope sensitivity but stellar photospheric activity. Modern high res soectrgrofh a like EXPRES possess the resolution but it’s the novel “photospheric modelling” algorithms they also employ that it is hoped will push sensitivity down to true Earth analogues. Call it adaptive optics for spectroscopy ! We should start seeing some exciting results over the next decade , especially in combination with Gaia. ( see below )

              Essentially the distance between a star and orbiting planet as seen from Earth is measured in terms of angular separation. Expressed as an angle. By way of illustrious , if observed from about 30 light years , the Sun / Earth angle would be a tiny 100mas. This is obviously determined by the actual star /planet distance but also by how far away the system is from Earth. Essentially the further away the smaller that angular separation and the smaller the requirement for the Inner Working Angle of any coronagraph or Starshade orientated telescope to see the planet. ( it’s brightness matters too obviously and this determines the aperture size of any telescope required ) . So to see Earth from 30 light years the IWA of a coronagraph would have to be less than 100mas.

              In terms of K stars ,in lying in between G and M dwarfs as you point out -their hab zones will too. Anywhere from 0.3 AU out to beyond 1 AU depending on the spectral subclass. In binary systems there will be a stable planetary orbital limit beyond which the gravitational influence of the companion star causes disruption. This is quite close for Alpha Centauri , around 2.5 AU -but still outside both stars hab zones.

              There are many promising binaries /multiples within 30 light years ( the rough imaging limit for the proposed HabEX 4m telescope ) as you say , many with far wider separations than Alpha Centauri . The 40 Eridani System for instance , is 16.5 ltyrs away and consists of a K1 primary orbited a comfortable 400 AU away by an M4 and white dwarf binary . No planets as yet , but a prime target for the latest precision RV studies and direct imaging eventually too . The system is about as old as our own , so the primary is a good potential candidate for habitable planet and SETI . ( known as Keid , it was the home star of “Vulcan” in Star Trek) .

              RV spectroscopy only expresses mass as a minimum, the erstwhile ” msini” . Where ” i ” is orbital inclination and mostly unknown. However, this is where Gaia could come into its own. Everyone knows that it should find tens of thousands of wide orbiting Saturn mass or larger gas giants via astrometry during its observations , especially if extended to near ten years . Great in its own right. But …The crucial part is that it will determine these planets masses and most important of all , orbital characteristics . Best for the extended mission, and best of all for nearer stars say within 100 light years or so. Astrometry unlike RV favours wider planet discovery and constraint . This characterisation includes the orbital inclination. It now seems that Gaia has now been adapted to see even the nearest and brightest stars ( originally it was only thought to work down to a magnitude of 7) . Just the stars that will have been searched by precision RV . Assuming the orbital inclinations of any outer gas giants represent the orbital plane of all planets in any star system ( a reasonable enough assumption most of the time) then we have the ” i “of msini and thus the EXACT ( rather than minimum) mass of any other planets discovered by RV within the same system. Such as hab zone terrestrials .

              The often forgotten WFIRST exoplanet microlensing survey is important as an “all planet” survey rather than favouring close or far out bodies . It will give the first large estimation of true planetary architecture and hopefully put to rest any current systematic bias you allude to .

    • Ashley Baldwin September 29, 2017, 17:13

      They will be part of Debra Fischer’s 100 Earths” project ( there is a nice Yale lecture she gives on this on YouTube ) starting operations imminently with the newly commissioned EXPRES spectrograph at the Discovery Telescope , the also recently commissioned ESPRESSO at the VLT and the NEID spectrograph commissioning 2019 on the WIYN telescope . They are all suitably high res enough to theoretically find Earth mass hab zone planets but the first and the last also have near nightly telescope time to bin large numbers of observations quickly and also employ state of the art “photospheric ” modelling algorithms to minimise the stellar noise that has until recently created a sensitivity floor above the 10-20 cm/s precision necessary.

  • andy September 29, 2017, 18:10

    Hopefully planet searches will turn up something orbiting either A or B, but given the potential for an accretion-hostile environment around the stars, the possibility of a disappointing no-planet scenario is one that we have to consider. Fortunately the mutual inclination between the Alpha Centauri AB orbit and Proxima appears to be below the critical angle for Kozai oscillations, which is a plus for the survival of any planets that managed to form there.

    Incidentally it may well be that Proxima is a captured star and not an original member of the Alpha Centauri system, see this arXiv preprint by Feng & Jones: the dynamics appear to be plausible and there does seem to be a discrepancy in the metallicities.

    • Ashley Baldwin October 1, 2017, 5:57

      Yes. Be interesting to see what effect that had on planetary formation /architecture around the other two stars .

    • DJ Kaplan October 2, 2017, 12:25

      Agreed, we shouldn’t rule out anything at this point.

  • DCM October 1, 2017, 16:26

    With an ion engine maybe an astronomical probe could get close enough in a reasonable time to send close images.

    • ljk October 2, 2017, 10:14

      Your post prompted me to check into ion engines as propulsion systems for interstellar probes:


      The answer from here seems to be we would need some serious technological advancements for ion power to be feasible for any kind of interstellar mission.

      • DCM October 2, 2017, 12:02

        I was afraid of that. I had in mind a telescope & maybe other instruments with a library of programs for possible discoveries. We could when the probe was close enough to see which were likely to be needed send the appropriate orders, to arrive in about 4 years, and which it would carry out autonomously while sending back results. Of course we could use Centauri sunlight for additional power. I was hoping we could ramp up enough speed to get there in maybe 15 or so years. Looks like I’ll never see another solar system up close.

        • ljk October 2, 2017, 14:57

          Before then we may develop telescopes either on Earth or in near space that could do much of what you are suggesting. Just look at how much we have learned about exoplanets already with the instruments we have now.


          One potential answer to why advanced ETI have not visited us may be that they have very sophisticated telescopes in space that can tell them most of what they would want or need to know about Earth and other worlds throughout the galaxy without having to send even a probe here.

          Just over twenty years ago we knew hardly any exoplanets. And until last year I thought anyone getting really serious about interstellar exploration was a pipe dream. Now look where we are and just imagine where we will be in the next 20 to 30 years.

  • DJ Kaplan October 2, 2017, 18:51

    Has ESPRESSO finally been installed at Paranal?
    My Italian is rusty. ;)
    This could be a game-changer.

    • epsilon October 3, 2017, 7:58

      yes, it achieved “first light” on sept 25th, according to the web report.

      • DJ Kaplan October 4, 2017, 13:14

        Thank you, I was thinking that by “first light” they meant through the spectrograph alone, not mounted on one of the telescopes yet.

        I’m standing by for new discoveries.
        (Motivated to take care of my health so I live long enough to see it!)

  • andy October 3, 2017, 4:57

    Talking of small missions dedicated to single targets, has anyone got any news on PicSat (nanosatellite for observing the Beta Pictoris transit)? It looks like their website has gone offline.

  • Michael Fidler October 4, 2017, 22:32

    A little off topic but an update to nearby stars with 14 new ones within 25 pc and new distance to TRAPPIST-1 that increases the planets diameter by 4%. Less dense more water???
    The Solar Neighborhood. XXXX. Parallax Results from the CTIOPI 0.9 m Program: New Young Stars Near the Sun.

    23 06—2MASS J23062928-0502285 (TRAPPIST-1) is an M7.5
    dwarf first identified by Gizis et al. (2000). Later, the CTIOPI
    1.5 m program measured its first ptrig at 82.58 ± 2.58 mas, or
    ∼12.2 pc, (Costa et al. 2006) based on a 3.3 yr baseline. The
    CTIOPI 0.9 m program has also observed it since 2004, and
    we present a new parallax of 78.76 ± 1.04 mas in Table 1.
    Because these two parallaxes are independent measurements, we
    calculate its weighted mean ptrig to be 79.29 ± 0.96 mas, which is
    ∼4% farther than the distance measured at 1.5 m. During this
    12.2 yr period, we did not detect any perturbations in the
    astrometric residuals, which are shown in Figure 8.
    Because of its proximity in the solar neighborhood, 2MASS
    2306-0502 has been the target of extrasolar planet searches.
    Gillon et al. (2016) first detected three Earth-sized transiting
    planets, and recently Gillon et al. (2017) discovered four more
    transiting Earth-sized planets around this cool dwarf. With our
    improved parallax presented herein, we estimate that the radii
    of all seven planets as well as the host star would be 4% larger
    than previously reported.
    Gillon et al. (2017) reported that the Spitzer Space Telescope
    detected two flaring events during 20 days of observation in
    2016 September. We also detected one flaring event in 2009
    July, which is illustrated in Figure 9. As reported in Table 3,
    the overall variability is 11.6 mmag in the I band, which is less
    than the 20 mmag limit for CTIOPI to consider a system
    significantly variable. If the entire flaring event is removed, the
    mean variability drops to 8.2 mmag.

  • Michael Fidler October 6, 2017, 22:27

    Telescope attachment allows ground-based observations of new worlds to rival those from space.

    A new, low-cost attachment to telescopes allows previously unachievable precision in ground-based observations of exoplanets—planets beyond our solar system. With the new attachment, ground-based telescopes can produce measurements of light intensity that rival the highest quality photometric observations from space. Penn State astronomers, in close collaboration with the nanofabrication labs at RPC Photonics in Rochester, New York, created custom “beam-shaping” diffusers—carefully structured micro-optic devices that spread incoming light across an image—that are capable of minimizing distortions from the Earth’s atmosphere that can reduce the precision of ground-based observations. A paper describing the effectiveness of the diffusers appears online on October 5, 2017, in the Astrophysical Journal.




  • Michael Fidler October 6, 2017, 22:32

    Forgot; The paper – https://arxiv.org/pdf/1710.01790.pdf