Interview with Suzanne Aigrain, University of Oxford, by John Strachan
In this interview we ask expert exoplanet researcher Professor Suzanne Aigrain from Oxford University why she is a keen participant in exoplanet research and what her views are on the prospects and difficulties of detecting small exoplanets such as those that may exist around Proxima Centauri.
Can you tell me what first got you interested in astronomy, and exoplanets in particular?
I have a background in physics, and started getting interested in astronomy during internships at the Institute of Astronomy in Cambridge while I was an undergrad. My interest in exoplanet research really came about after I graduated, when I spent 6 months as a trainee at the European Space Agency’s ESTEC research centre in the Netherlands, before starting my PhD. This was an exciting time, as the first transiting exoplanet had just been discovered, and I was given the opportunity to work on this very hot topic. I was hooked, and decided to continue in this field of research for my PhD.
Detecting small exoplanets using the radial velocity method appears to be very difficult. The debate which you have been involved in concerning the existence or not, of a planet around Alpha Centauri is an example of one such case. Can you give your views on the main difficulties involved in detection of small exoplanets using the radial velocity method?
There are many difficulties which have to be overcome in order for the radial velocity (RV) method to succeed in detecting small exoplanets. Here are a few of the most important:
- Instrumental precision: The radial velocity of the star has to be measured accurately enough. Current state-of-the-art RV spectrographs such as HARPS and HARPS-N have accuracy down to 1 m/s, or even slightly below. This is sufficient, with many measurements, to detect Earth-mass planets in the habitable zone of some small stars, such as M dwarfs. However an Earth-mass planet in the habitable zone of a Sun-like star causes a variation of only 10 cm/s, over an entire year… With new calibration techniques such as laser combs, though, the precision of RV spectrographs is still improving, and the next generation of experiments such as ESPRESSO are likely to be able to achieve this precision.
- Stellar activity: The apparent RV of a star can vary even if there are no planets orbiting it. For example, most stars have some starspots (the stellar analogue of sunspots)—regions where the magnetic field of the star is particularly strong. Starspots appear darker than the rest of the star’s surface, and as the stars rotate, they will hide a part of the star that is first moving towards the observer, then away. This will lead to an apparent change in RV that could easily be mistaken for a planet. There are all sorts of other effects, mainly due to magnetism and convection (the hot gas inside the star bubbling up to the surface and back down again) that can cause subtle and complex RV variations. To detect small planets despite these, we must model them in detail, and this is a very active area of research today.
- Patchy observations: Because our instruments are located on the Earth, we can only observe a given star during the night, when the sky is clear and the star is up in the sky. Additionally, in RV we typically observe one star at a time, so we must chose between the different targets we want to observe. As a result, the observations of each star have many gaps, and this can make it even more difficult to distinguish between planetary and stellar signals.
- Comparing models: When we analyse the observations of a given star, we don’t know in advance how many planets it has. To find out, we must try models with different numbers of planets and compare them, and at the same time we must also account for the activity of the star, and the noise of our measurements. This, combined with the patchy nature of the data, makes analysing RV data a very complex, time-consuming and challenging process.
These difficulties came in to play in the case of Alpha Centauri B. The observations were dominated by the signal from the companion star Alpha Centauri A (Alpha Centauri is a binary star), and by the activity of the star. There was also a tiny signal with a period of 3.2 days. This signal became stronger after modelling and subtracting the binary and activity contributions—strong enough for the authors of the original study to report the detection of an exoplanet. However, we ran some simulations using synthetic data with the same time sampling, noise properties, and the same kind of activity signal, but no planet, and the 3.2 day signal was still there. That means it couldn’t have come from an exoplanet. We actually think it was an artefact of the time sampling, that just got boosted by the complex modelling needed to remove the activity. Now that we know this sort of thing can happen, we can look out for it in future, hopefully avoiding a repeat of this sort of problem.
Do you think the daily observing of Proxima Centauri during the three months of the Pale Red Dot Campaign will significantly increase the chances of detection?
A daily observing strategy determined in advance will certainly help. This dense set of observations, captured over the next two months, will help with the modelling of the star’s activity signal. We know that Proxima Centauri rotates slowly, so by gathering many observations in a short space of time, we should be less sensitive to the effects of starspots. If any possible planets are found during the new observations, we can then look back at the considerable data set of previous observations, to confirm that they are real.
If the campaign does detect an exoplanet, what characteristics of the exoplanet do you think we will be able to determine from the observations and what follow up observations would you recommend?
If an exoplanet is detected by the radial velocity method some of its orbital parameters will be determined, in particular the period and eccentricity of the exoplanet. We also get a lower limit on the mass of the exoplanet relative to the star. It is a lower limit only because we do not know the inclination of the orbit. If the exoplanet happens to transit across the disk of the star, then we will know that the orbit is edge on, and thus we will obtain the true mass of the planet, as well as its radius relative to the star. Together these give us its mean density, from which we can tell something about its composition (mainly gaseous, mainly rocky, or something in between?). However, only a small fraction of exoplanets transit, so we’d have to be particularly lucky.
For larger planets, with an extended gas envelope, observing the transits in multiple wavelengths can enable us to probe the composition of its atmosphere. However, for an Earth-like planet, even with the best instruments available, such as the future James Webb Space Telescope (JWST), this kind of measurement may not be possible: its atmosphere may be too thin, and the exoplanet may be too cool to see features in its spectrum during a secondary eclipse (when the planet passes behind the star). We may have to wait until we are able to image the exoplanet directly, by blocking out the light of the star using something like a coronagraph. Some JWST instruments are equipped with a coronagraph, and there is also a project to launch a “starshade” which would act as an external coronagraph for JWST (a project call the New Worlds Explorer). If we were able to isolate the light of the planet, we could then extract its spectrum and learn about the temperature and composition of its atmosphere.
What do you think the impact of finding an Earth-like exoplanet around Proxima Centauri or other stars near to the Earth could be?
This would be a major discovery. It would confirm that such planets must be very common, as we already suspect based on statistical results derived from the Kepler mission: Dressing and Charbonneau (2013) estimated 40-50% of M stars have at least one Earth-sized (up to 1.6 Earth radii) exoplanets in their habitable zone. It would give us an extra impetus to search for more exoplanets in the solar neighbourhood, and to invest in the technology we need to study them in detail. Knowing there are many other worlds potentially like our own in the Galaxy would also change the way we think of our own place in the universe.
What do you think the chances are that, if an exoplanet is detected, it is similar to Earth and that it may harbour life?
If Proxima Centauri hosted any planets substantially more massive than the Earth in its habitable zone, they should have been detected already during previous observations. Therefore, if a new planet is detected as part of the Pale Red Dot campaign, it is likely to be similar in mass to the Earth. Whether such a hypothetical exoplanet might harbour life, though, is very hard to know. Existing models of planet formation, and data from the Kepler satellite and associated RV follow-up, suggest that planets below 1.6 Earth radii are likely to be rocky and to have thin atmospheres, akin to Earth. Larger planets tend to have larger gas atmospheres and may be more like Neptune than the Earth. So if the planet’s mass was small enough, it would probably be rocky. But whether it would also have developed life – that is anyone’s guess!
The Pale Red Dot Campaign is one of several campaigns aimed at detecting exoplanets. Can you comment on any ongoing campaigns which you are particularly interested in and which may help to detect exoplanets in the solar neighbourhood?
There are very many exoplanet campaigns ongoing or just about to start—too many to list—and many of them are exciting. Two that I am particularly interested in at the moment, are the K2 mission and the TERRA Hunting Experiment (THE) at the Isaac Newton telescope (INT) in La Palma. K2 is the Kepler “Second Light” mission and uses the Kepler space observatory. What particularly interests me about K2 is that it is observing some nearby young open clusters, which are groups of young stars that formed out of the same cloud of gas and dust. This represents our first opportunity to search for young transiting planets and directly learn about their early evolution. THE is an experiment, proposed by Didier Queloz from Cambridge University, which would involve installing a high precision radial velocity spectrograph called HARPS3 (a copy of HARPS and HARPS-North) on the INT, and upgrading the telescope to be fully robotic. This instrument would then be used in a long term campaign (5-10 years) to observe a small number of nearby solar type stars every day. The use of a dedicated instrument over such a long period of time will greatly improve the chances of finding Earth-like exoplanets round these stars.
What do you see as the future for exoplanet detection and characterisation over the next ten years?
The first two purpose-built instruments for direct detection of exoplanets on large telescopes, SPHERE on the Very Large Telescope (VLT) and the Gemini Planet Imager (GPI) on the Gemini telescope, have just started large surveys for young, massive planets on wide orbits around nearby stars. In the next two to three years JWST, TESS (the Transiting Exoplanet Survey Satellite) and ESPRESSO will all come online, improving our ability to detect exoplanets and measure their properties from space and from the ground. The ongoing GAIA astrometric mission should also find many high-mass (>Jupiter-mass) long-period exoplanets in the solar neighbourhood, and the small photometric satellite CHEOPS will search for transits of previously known planets.
By 2025 the PLATO space mission will search for exoplanets among relatively bright stars with the aim of being able to detect Earth sized planets in the habitable zone around solar-like stars. The European Extremely Large Telescope (E-ELT) with its large aperture will be able to directly detect exoplanets, down to perhaps rocky sized exoplanets, and using high resolution spectroscopes will be able to record a number of their spectra. In the very long term, the ultimate goal is to be able to directly image and take spectra of Earth-like planets around Sun-like stars, and search for signs of biological activity in their atmospheres. This will require a large, space telescope with a state-of-the-art coronagraph, as recently proposed for example in the High Definition Space Telescope (HDST) report.
About the interviewee. Professor Suzanne Aigrain is head of a research group at the University of Oxford which focuses on the detection and characterisation of exoplanets and their host stars. She was born and educated in France and moved to the UK for her undergraduate studies, where she has remained since, except for a 6-month spell at the European Space Agency’s ESTEC research centre in the Netherlands just after finishing her undergraduate degree at Imperial College London. She completed her PhD on Planetary Transits and Stellar Variability at the University of Cambridge. Since then she has held post doctoral positions at Cambridge and lectureship positions at the Universities of Exeter and Oxford. During this time she has worked as a Co-Investigator or Participating scientist on major collaborations including CoRoT, Kepler, K2, TESS and PLATO. Her research group’s website is www.splox.net and she occasionally tweets as @AirborneGrain.