Planets of Other Suns

Exoplanets, planets of other stars. How did we discover them?

It took a long time to do so, with plenty of false alarms along the way. I am old enough to remember when Peter van de Kamp’s 1960’s and 1970’s claimed discovery of planets around Barnard’s Star was taken very seriously. But it was later discovered that maintaining his telescope caused the largest observed effects, and that discovery is now discredited. In 1991, Andrew Lyne, M. Bailes, and S.L. Shemar claimed that they had discovered a planet orbiting pulsar PSR 1829-10, a planet with orbit period half a year. But they then discovered that they had made a small error in their accounting for the Earth’s position, and they retracted their discovery.

The first confirmed discovery of an exoplanet was in 1992, when Alexander Wolszczan and Dale Frail discovered two planets orbiting pulsar PSR 1257+12. This was followed by a third one in 1994. The first one for a “normal” sort of star was in 1995, when Michel Mayor and Didier Queloz discovered a planet around the Sunlike star 51 Pegasi.

These discoveries were followed by numerous other ones, and some 3500 planets are now considered confirmed to exist. In particular, the Kepler space telescope’s observations have yielded a large number of discoveries, about 2/3 of all known exoplanets.

More below the fold.

We have been able to learn very little about most exoplanets, like their orbits and/or their masses and/or their radii. But what we have learned has often been very startling. The Solar System looks pretty much what one would expect a planetary system to look like, but many exoplanet systems look very different in some ways. In particular, “warm Jupiters” and “hot Jupiters” are common, and some planets are unusually close to their stars. Some such planets also have very eccentric orbits, elongated with one end much closer than the other to the planet’s star. There also appears to be a gap between presumably rocky “super Earths”, up to 1.75 Earth radii, and presumably icy and gassy “mini Neptunes”, starting at 2 Earth radii. A large number of multiplanet systems have been detected, but there seems to be an excess of single-planet systems over what one would expect from the multiplanet ones, the “Kepler dichotomy”.

A further problem is that we have been able to learn the radius of an exoplanet, or its mass, but seldom the two together. That makes it difficult to get an idea of exoplanets’ bulk compositions. So if an Earth-mass planet was mostly a huge ocean, it would be significantly larger than the Earth. Likewise, if an Earth-sized planet was mostly a huge ocean, it would be significantly less massive than the Earth. So we don’t know for sure how many seemingly Earthlike planets have super oceans.

Even for Earth mass and Earth size, our problems don’t end there. Our homeworld has plenty of water on its surface, but that water is only about 0.00023 of its total mass. How much water is in its interior has been a subject of a lot of hand-wavy estimates, but a plausible number that I’ve seen is 10 oceans. So we don’t know for sure why the Earth has enough water to cover much of its surface but not enough to cover its highest possible mountains. So if the Earth had 10 oceans of water added to its surface, all its land would be covered up, but its overall density would not be affected much.

If the Earth had much less water, it would likely have much less of volatiles like nitrogen and carbon dioxide. It would also be lifeless in the sense of no organisms, meaning that it would have much less atmospheric oxygen, if any. This means that it would have a very thin atmosphere. Something that makes desert planets with breathable air like Dune’s Arrakis and Star Wars’s Tatooine very implausible.

Super oceans also present habitability problems, because for a thick enough ocean (> 200 km for 0 C), the lower parts of the ocean will freeze into high-pressure phases of ice. That will make it hard for hydrothermal-vent organisms to reach the surface.

Also, if an Earthlike planet is out of the “habitable zone” or liquid-water zone, its water will either boil off, as in Venus, or freeze, as in Mars.

It must be noted that the Solar System would be very difficult to detect over interstellar distances. That is evident for plotting it against known exoplanets. Most of the exoplanets are on the massive and near side of a line stretching from small and very near to large and very far — and the Solar System’s easier-to-detect planets fall on that line: Venus, Earth, Jupiter. The others are even more difficult.

Now for observing them.


One can observe some exoplanets directly, like four planets of star HR 8799. Those planets were observed in near-infrared because their star is a very young star, and the planets still have plenty of their original heat. Even so, it was necessary to block out the light of their star.

But the large majority of exoplanets have been observed indirectly, and I will start out with some history of indirect observations of stars.


The story begins around 2000 years ago, when Greek astronomer Ptolemy named one of the stars in the constellation Perseus after a Gorgon named Medusa, a horrible monster that legendary hero Perseus had killed. Medusa looked like a woman with snakes for hair, and looking at her could turn one into stone. In the Middle Ages, Arab astronomers named it Al Ghûl (“The Ghoul”), after a kind of demon that likes to eat human flesh. But it was in 1667 that Italian astronomer Geminiano Montanari noted that it varied in brightness, and in 1783 that British astronomer John Goodricke noted that it varied periodically. It has a period of 2.86 days, and it is normally at magnitude 2.1, but for 10 hours in each cycle, it is at magnitude 3.4. He suggested that there was some dark object orbiting the star, or else that the star had a giant starspot. In 1881, American astronomer Edward Charles Pickering presented evidence that Algol was two stars orbiting each other and eclipsing each other, an eclipsing binary, and in 1889, German astronomer Hermann Carl Vogel showed that it was a “spectroscopic binary”, a pair of stars whose orbital motions were evident in its spectrum.

Astronomers have found large numbers of eclipsing binaries, including some in other galaxies. They are very useful for finding properties of stars, and also useful for finding distances.

Turning to exoplanets, there have been several searches for exoplanet transits, exoplanets eclipsing their stars. These include with ground-based telescopes, like WASP and MEarth, and with space-based ones, like CoRoT and Kepler. The latter one looked continuously at some 145,000 stars in Cygnus and Lyra from 2009 to 2013. It stopped because of the failure of two of its reaction wheels, necessary to keep it pointed at those stars. Since then, it has been aimed along the plane of the Earth’s orbit, with a broader mission: K2.

It is necessary to look at all those stars because transits are the result of line-of-sight coincidences. It is not very probable that a planet’s orbit plane is nearly edge on. For the Earth around the Sun, it is 1/215. But nevertheless, a large number of planets has been detected by this technique. Also a problem is different orbit orientations. Though the Solar System’s planets have nearly coplanar orbits, they are inclined enough to each other to make it very improbable to see two or more planets transiting. That is less of a problem for very close planets, however, and many such planets have been found.

It has also required very precise photometry, measurement of amount of light. Jupiter transiting the Sun produces a relative brightness dip of 0.01, and the Earth 0.0001, close to Kepler’s limit. But low-mass stars are much smaller, and planet transits much easier to see for them.


We go back in time again, this time to around 1679 – 1687, when British astronomer-physicist Sir Isaac Newton was working out how the planets move as a consequence of his three laws of motion and his law of gravity. He noticed that not only does the Sun pull on the planets, the planets also pull on the Sun, and they pull it away from the Solar System’s barycenter or center of mass.

Advancing to 1779, we find British astronomer William Herschel starting to search for stars very close together, for measuring the parallax across the Earth’s orbit. He found many more than he expected, some 800 in all. Observing them over some 20 years, he concluded that they were orbiting each other because of their gravitational pulls on each other.

Starting in 1819, German astronomer Friedrich Wilhelm Bessel measured the positions of over 50,000 stars. He joined the race to find a star’s parallax, and he was the first to succeed, in 1838, with 61 Cygni. Then in 1844, he concluded that stars Sirius and Procyon had “dark companions”, from their moving as if they were in binary-star systems but with the other stars seemingly absent. In 1862, American astronomer Alvan Graham Clark discovered Sirius’s companion, and in 1896, American astronomer John Martin Schaeberle observed Procyon’s companion.

Around the time of that latter discovery, astronomers were starting to take spectra of stars and to observe stars’ spectral lines. They noticed that some stars’ spectral lines move, and that some stars have paired spectral lines that move in opposite phase relative to each other. This is from blueshifting and redshifting from the stars alternately moving toward us and away from us relative to their barycenters. These stars are called spectroscopic binaries, and one of the first to be discovered was Algol.

Now to planets.

— Positions

The first indirect method used to search for exoplanets was astrometry, measurement of stars’ position. Back in the nineteenth century, some astronomers claimed that there was astrometric evidence for a planet of star Omicron Ceti, but it was eventually discredited. Then in 1948, American astronomer Peter van de Kamp started observing Barnard’s Star with his observatory’s telescope. By painstakingly measuring the star’s position relative to nearby stars on photographic plates, he hoped to discover evidence of planets pulling on it. In 1963, he announced that the star has a planet with 1.6 Jupiter masses at 4.4 AU (“astronomical units”, the average distance between the Earth and the Sun). Then in 1969, he claimed that the star had two, with masses 1.1 and 0.8 Jupiter masses.

But in 1973, George Gatewood and Heinrich Eichhorn failed to observe what PvdK claimed to have observe, and in that year, John J. Hershey discovered that maintenance on PvdK’s telescope had observable effects comparable to what PvdK claimed for planets. So PvdK’s planets are now discredited.

The largest Solar-System deviation is due to Jupiter, about 0.5 milliarcseconds at 10 parsecs / 32.6 light years / 7 times the distance to Alpha Centauri. Saturn has an effect about half that, and the Earth a much smaller effect.

— Distances

Turning to distance along the line of sight, one can precisely measure changes in it with a pulsar’s pulses, since pulsars’ rotation is very regular. A pulsar’s planets make its pulses arrive alternately arrive early and late, and pulsar planets have indeed been found in that way. Only four planets have successfully been found by that means, with three orbiting one pulsar, PSR 1257+12, and one orbiting another pulsar, PSR 1620-26.

— Radial Velocities

But the most successful means of detecting exoplanets with gravity has been with radial velocity, like with spectroscopic binary stars. The planet of 51 Pegasi was detected by that means, and large numbers of other planets have also been detected by that method.

Radial-velocity mass measurements have an ambiguity: the tilt of the planet’s orbit plane. What we observe is a sort of projected mass, mass multiplied by a factor that measures how close to edge-on the orbit is. So it is difficult to get actual masses, except for cases where we know the orbit inclination, as for transits.

For the Solar System, Jupiter is fairly easy to discover with its induced radial velocity of 12 m/s, but one has to observe over its period of 12 years to be sure about it. The Earth has a much shorter period, but makes only 0.1 m/s, over a hundred times less.

— Perturbations

Planets not only pull on their stars, they also pull on each other.

This was first used in the Solar System to measure the masses of planets and moons, and to discover new ones. In 1846, French astronomer Urbain Leverrier and British astronomer John Couch Adams both calculated where an additional planet would be by studying discrepancies in planet Uranus’s motions. It was discovered the next year by German astronomer Johann Gottfried Galle, only about 1d away from Leverrier’s predicted position. There were some remaining discrepancies in Uranus’s orbit, and they were proposed to be from an additional planet. At first, Pluto seemed to be this planet, but it was found to be too small for that. This discrepancy was eventually resolved when Voyager 2 flew by that planet, providing an improved mass value.

Leverrier also measured Venus’s mass by its perturbations on nearby planets, doing that because that planet has no moons. He found a value (0.83) close to what was measured with spacecraft (0.815). He was less successful for Mercury and Mars, however. He proposed that there was some planet inside Mercury’s orbit that was giving it some excess perihelion precession. However, no intra-Mercurian planets were found, and recent upper limits on sizes of intra-Mercurian asteroids are around 20 km. Instead, that extra precession was due to a modification of Newtonian gravity: Einstein’s general relativity.

What works for planets can also work for moons, and the masses of some of the moons of Jupiter and Saturn were found from their mutual perturbations. This was helped by some of the moons having orbital resonances with each other. This makes the moons in such sets alternately move ahead and behind over several orbits.

Turning to exoplanets, transit observations are precise enough to find small variations in them from the planets’ perturbations on each other: transit timing variations or TTV’s. These have been used to estimate the masses of several exoplanets, and some exoplanets have been discovered from the TTV’s that they make in known ones. In 2011, American astronomer Sarah Ballard became the first one to do so, inferring the existence of planet Kepler-19c from its perturbations on planet Kepler-19b.

Gravitational Microlensing

We go back in time yet again, to 1801. German astronomer Johann Georg von Soldner calculated how much light would be bent by gravity if it obeys Newtonian mechanics. When German physicist Albert Einstein worked out general relativity, he discovered that it predicted a value twice as large as Soldner’s. British astronomer Sir Arthur Eddington and others then measured the deflection of starlight near the Sun during a total solar eclipse in 1919, and they found some rather rough agreement. But it nevertheless became big news. Later work, like with radio telescopes in separate continents observing quasars and combining their observations to act like an Earth-sized telescope (VLBI), showed much better agreement.

Then in 1979, what looked like two identical quasars was detected, SBS 0957+561. They turned out to be two gravitational-lens images of the same quasar. Several other gravitational-lens multiple images have since been detected.

In 1992, the Optical Gravitational Lensing Experiment (OGLE) was started, for looking for evidence of brown dwarfs and other such hard-to-see objects in our Galaxy. When one of them passes in front of a star, it will do gravitational lensing on that star’s light, making its apparent brightness change. It was later joined by Microlensing Observations in Astrophysics (MOA), and both observer groups have found some 49 exoplanets using this technique.

So there you have it. We have observed numerous exoplanets, but they are sometimes very unlike Solar-System planets, and much about them continues to remain mysterious. But there are several efforts in the works to look for more exoplanets and to observe them in more detail.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: