PHY111 Our Evolving Universe:
Lecture Summaries

This HST image of the first-magnitude star Fomalhaut (in the constellation Pisces) has had the star itself blocked out by a device called a coronagraph. With the glare from the star much reduced, it is possible to spot a planet (Fomalhaut b; see inset) orbiting the star at a distance of about 115 AU (115 times the Earth's distance from the Sun). Fomalhaut b is definitely a planet, not a brown dwarf – it is estimated to weigh no more than 3 times as much as Jupiter. It is unusually bright for its mass, and some astronomers speculate that it may have icy rings like Saturn's. Image by the HST Advanced Camera for Surveys; credit NASA/ESA

Summary of Lecture 13 – Detection of Extrasolar Planets

1. There are various ways of attempting to detect planets around other stars:
   • direct observation is very difficult with current technology, because the star's light swamps the much fainter planet:
     • 42 "planets" in 39 systems as of December 2 2013;
     • all very massive (smallest is 2 Jupiter masses, see picture, and several are >10 Jupiter masses and so probably brown dwarfs) and far from their stars (a couple at around 3 AU actually orbit brown dwarfs; the closest to a "proper" star is at 8.5 AU, comparable to Saturn, and the most distant at 2500 AU, 80 times further than Neptune!);
   • astrometry relies on detecting the orbital motion of the star across the sky caused by the gravitational force between star and planet:
     • the oldest method, but very difficult, with several claimed discoveries subsequently retracted;
     • most sensitive to massive planets orbiting relatively far from low-mass stars;
     • only works for nearby stars (otherwise motion not detectable, similar to parallax);
     • unsuccessful so far – only one or two unconfirmed detections;
   • spectroscopic methods rely on detecting (by Doppler shift) the orbital motion of the star in the line of sight:
     • most sensitive to massive planets orbiting close to stars;
     • not dependent on distance, except that fainter stars are harder to get the necessary high-precision spectra for;
     • because we detect only line-of-sight motion, which depends on tilt of orbit to line of sight, only minimum mass of planet can be determined (mass that planet would have in ideal case of edge-on orbit);
     • the most successful method to date: 538 planets in 404 systems discovered as of December 2 2013;
   • transit observations rely on detecting the drop in brightness of the star caused by a planet's passing in front of it:
     • requires extremely precise measurements of stellar brightness, best done from space (to remove effect of atmosphere);
     • most sensitive to large planet (bigger dip) close to star (more frequent dips);
       • but can see much smaller objects – best hope at present of detecting Earth-size objects around normal stars;
     • measures orbital period and radius of planet, but provides no information on mass unless combined with other information, e.g. spectroscopy;
     • requires very nearly edge-on orbit;
     • best combined with spectroscopy: transit implies edge-on orbit, so spectroscopy gives exact mass – combine with size from transit to get density, and hence information on chemical composition;
     • 424 planets in 321 systems discovered as of December 2 2013 – number increasing rapidly;
   • timing of regular variables can be used in a similar way to spectroscopy:
     • extremely sensitive if the parent star's variation is fast and regular;
       • smallest planet discovered this way (orbiting a pulsar) weighs only about the same as the Moon!
     • but can only be used with rapidly and regularly varying parent star;
     • 15 planets in 12 systems (three pulsars, one sdB pulsating variable, seven short-period eclipsing binaries, and one from timing variations in the transit of another planet);
   • gravitational microlensing detects planets when their gravity amplifies the light from a background star:
     • one-off event – only get one chance to see it;
     • limited information – mass of planet, and projected distance from star, but no details about orbit, unless lensing event lasts long enough to see orbital motion;
     • where information available, planets fairly far from their stars (few AU, periods 5-10 years);
     • masses similar to other exoplanets: 0.01 to 3.5 Jupiter masses;
     • 25 planets in 23 systems as of December 2 2013.
2. Properties of the planets:
   • The 1047 extrasolar planets in 794 systems (as of 02/12/13) so far observed around other stars:
     • are fairly massive (49% more massive than Jupiter; only 11% 10 Earth masses or less; 13% of transiting planets 2 Earth radii or less);
     • are mainly gas giants like Jupiter and Saturn, in those cases where their density can be measured (transiting planets), although some of the super-Earths with masses <10 Earth masses are icy or rocky;
     • are fairly close to their stars (63% closer to their star than the Earth, over ⅓ closer than 0.1 AU; only 7% further out than Jupiter);
     • often have quite eccentric orbits (least circular orbit of a major solar system planet is Mercury's, with eccentricity 0.2: about 40% have orbits less circular than this);
     • tend to orbit Sun-like stars high in heavy elements (just over half the stars are within 20% of the Sun's mass, whereas in a fair sample of nearby stars about 70% are class M, i.e. much less massive than the Sun; also, a similar fraction are higher in heavy elements than the Sun, whereas a typical local star is about the same as the Sun).
   • These properties are biased by the detection methods:
   • massive planets close to stars are the easiest to detect, so they will be over-represented;
   • low-mass planets like Earth are almost impossible to detect, so they will be missing;
   • planets with very long periods will be missed by most techniques (except direct imaging), because we haven't been watching for long enough yet;
   • spectroscopic techniques are easier with Sun-like stars than with class M stars, which have very complicated spectra (also, astronomers looking for extrasolar planets really want to find another Earth, so they preferentially look at stars like the Sun!).
   The preference for high-metallicity stars and the non-circular orbits aren't obviously biased, but may be common specifically for massive planets in close orbits, not for planetary systems in general.
   • The hot Jupiters (massive planets in very close orbits) were unexpected:
   • theory says gas giant planets should form far from stars, like Jupiter;
   • likeliest idea is that the planets do form out there and then migrate inwards
   • (modern models of solar system formation suggest this also happened to Jupiter, but orbital resonance with Saturn moved it out again – see PHY106)
   • The preference for stars high in heavy elements was expected:
     • planets believed to form when dust grains in the dusty disc around a young star coalesce into larger bodies;
     • dust is made from heavy elements, not hydrogen and helium;
     • therefore stars high in heavy elements likely to have more circumstellar dust, therefore more likely to have planets;
     • however, planets have been found around stars quite low in heavy elements, e.g. HD155358, an old Sun-like star (class G0) with only one-fifth of the Sun's heavy element content;
   • Some of the stars with detected planetary systems are members of binary pairs:
     • slightly unexpected – many people thought that collapsing gas cloud would form either binary system (if it rotated so fast it tore itself apart) or planetary system (if rotating less fast), but not combination;
     • however simulations do show circumstellar discs formed around members of (not too close) binary systems, suggesting that planets could form;
     • several planets found in wide orbit around both stars of a close binary (very weird).

Web links

• The PowerPoint file for this lecture.
• Nick Strobel has a brief discussion of extrasolar planetary systems; Gene Smith doesn't cover this at all. Probably the most successful planet-hunting team is led by Geoff Marcy and Paul Butler, who have a nice public website.
• The binary star animations came from Richard Pogge of Ohio State University. The transit animation is by Hans-Jorg Deeg and can be found on his homepage or through the Extrasolar Planets Encyclopaedia. The planet formation movie is from Willy Kley's web page; the still pictures of simulations are by Geoffrey Bryden, but he's moved institutions and the originals do not seem to be online anymore. Planet images were from various NASA websites, including Voyager, Galileo and Magellan.
• Discoveries of extrasolar planets are advancing apace. For updated information consult the Extrasolar Planets Encyclopaedia. Another good exoplanet catalogue is the Exoplanet Data Explorer maintained by the highly successful California Planet Survey team; however, this catalogue covers only planets detected spectroscopically.
• A short self test for this lecture.

Go back to main page
Go on to Summary for Lecture 14