1. Beyond Our Solar System: The Search for Extrasolar Planetary Systems
by Dr. William D. Cochran
Outreach Lecture Series Volume 21
Produced by and for the Outreach Lecture Series of the Environmental Science Institute. We request that the use of any of these materials include an acknowledgement of Dr. William D. Cochran and the Outreach Lecture Series of the Environmental Science Institute of the University of Texas at Austin. We hope you find these materials educational and enjoyable.

2. Beyond Our Solar System: The Search for Extrasolar Planetary Systems
The past several years have witnessed a major revolution in our thinking about how planetary systems form and evolve. For decades astronomers have searched for planets around nearby stars. While there had been some claims of possible detections in the 1960s through the 1980s, we now know that those were incorrect. The first truly believable hint that there were planets around other stars came from a very unexpected source in Alex Wolszczan (now at the Pennsylvania State University) detected some objects around the millisecond pulsar PSR B
These are three earth-mass (one in a lunar mass!) planets that formed around a very rapidly rotating neutron star. While these were indeed the very first extrasolar planets ever discovered, many astronomers tend to ignore them, because they do not orbit a “normal” star like our Sun. Nevertheless, they do demonstrate that planets probably do form in just about any circumstances, including many cases that we had not previously been able to imagine.
Why is it important to search for extrasolar planets? We gain a broader understanding of the formation of planetary systems and learn more about our own planet and solar system in the process. Of course, there’s always the hope of finding other “earth-like” planets that might support life.
William D. Cochran
The University of Texas
McDonald Observatory

3. Beyond Our Solar System
What is a planet?
How do we search for planets?
How many have we found?
Our recent discovery
Expectations vs. Reality

4. Introduction
In trying to understand the formation of our Solar System, we have made elaborate observations of the planets and other aspects of our Solar System.
We have used the observations as constraints for elaborate and coherent models to explain its formation.
These models were developed using a single example - our own Solar System.
Our sun is really a very common, ordinary star. There is nothing about our sun that would lead us to believe that it is special in any way. There are literally billions of other stars in our galaxy that are very similar to our sun.
Therefore, we believe that any theory to explain the formation and evolution of the sun should also be generally applicable to most of the other stars near us in the galaxy.
However, we do need to be cautious. One thing that we have learned in science is that when a theory is based on only a single example, there are certain aspects of it that will probably be wrong. Hopefully, most of the theory will be OK, but we may well have to revise parts of it.

5. Is our Solar System the proper paradigm for models of planet formation and planet evolution in general?
Is planetary system formation:
a common natural result of star formation, or
are planets rare around stars like our Sun?
How do the characteristics of a planetary system depend on properties of the central star, such as its mass, rotation, magnetic field, abundance of heavy elements, etc?
In our commonly accepted theory for the formation of the Sun and the solar system, the formation of the planets is a very natural by-product of the formation of the Sun. In fact, in this model it is very difficult to form a star without also forming a planetary system. Thus, our standard theory makes a clear and testable prediction --- planetary systems like our own should be very common around solar-type stars. This is the prediction that planet search programs have been testing.
Of course, our theory of star- and planet-formation was finely tuned to produce our solar system. One interesting question we are trying to answer is what variations of this model will be necessary to produce the other types of planetary systems we might find.

6. What is the diversity of habitats for life in the Universe?
How do planetary properties change and evolve during the life of the star?
What is the diversity of habitats for life in the Universe?
These kinds of questions were unanswerable as long as the only planetary system we knew was our own. However, beginning in 1995, with the discovery of the first planet orbiting a star other than our Sun, we have received many new insights into these questions and many new surprises!
We always think of our planetary system as constant and unchanging. We often want to reject any notions that the planets may not always have been in the same orbits in which we find them now. However, we can see the orbits of small bodies in the solar system, such as comets and asteroids, changing on relatively short timescales. These orbit changes are due to their gravitational interactions with the much larger planets such as Jupiter. However, there is significant evidence that the orbits of all of the planets may have evolved significantly in the early history of the solar system. Such dynamic evolution may turn out to be extremely important in understanding how other planetary systems arrived in the conditions in which we find them today.
Ultimately, one of the major motivations in searching for planets around other stars is the quest for life elsewhere in the Universe. We usually believe that life must develop in close association with a planet. Thus, finding extrasolar planets is the first crucial step in finding extraterrestrial life.

7. Our Current Theory of Star and Planet Formation (From Shu et al. 1987)
B) A protostar with surrounding nebula and disk forms at the center of a core.
A) Cores form within a molecular cloud
C ) A stellar wind starts along the polar axis, and starts to clear out the disk.
D) Infall ends, leaving a newly formed star and disk. Planet formation occurs within the disk.
This figure traces the formation of a star from the initial interstellar cloud of gas and dust which begins to collapse under the force of gravity, through the newly formed protostar, to the final stage of star formation in which the newly formed star is surrounded by a rotating disk of the debris left over from the formation of the star. This disk is made up of gas (primarily hydrogen and helium) and dust particles. It is from the material in this disk that planets are formed. Our standard model of planet formation says that in this last stage of star formation, the dust particles in the disk will collide, stick together, and grow into progressively larger and larger objects. When these objects reach a few kilometers (a mile or so) in size, we call them “planetesimals”. At this size, the gravitational attraction between planetesimals becomes important, and they will continue to collide with each other and grow in size. We believe that this is the way Earth, and the other inner planets (Mercury, Venus, Mars) were formed. This same process was also at work in the outer regions of the disk as well. There, temperatures were much colder, and the solid material included water ice. In the outer regions, there was more material available for accumulation, and the planetesimals grew more rapidly, forming giant planet cores. When one of these cores reached about times the present mass of the Earth, its gravity was sufficient to attract a large amount of the gas in the disk to collapse onto it. This formed the very deep gas envelope of Jupiter and Saturn. Uranus and Neptune are thought to be mostly this core with a much thinner gas envelope.
This cartoon of the basic steps in the formation of a star is from an article by F. H. Shu, F. C. Adams, and S. Lizano in Annual Reviews of Astronomy and Astrophysics, volume 25, pages 23-81, 1987.

8. Young Stellar Disks Source: STScI, NASA
Young stellar disks, planet formation
Source: STScI, NASA
Source: STScI, NASA

9. What do we mean by a planet?
To answer the previous questions, we want to search for planets orbiting other stars and compare their properties to our own Solar System.
However, we cannot just define a planet as any object orbiting around a star. The following types of objects orbit a star:
another star (this is called a binary star system)
a brown dwarf
a planet
A planet is something that is actually hard to define in a rigorous sense, but is something that we know when we see it! However, just because we find some sort of object in orbit around a star, we can not automatically assume that it must be a planet. Most of the stars in the sky are in binary or multiple star systems, in which two or more stars orbit around each other. There are also objects called “brown dwarfs”, which can be thought of as “failed stars”.

10. Classical Definitions of Stars, Brown Dwarfs, and Planets
Star: an object with sufficient mass that it starts and sustains hydrogen fusion.
Brown Dwarf : an object that forms in the same manner as a star but lacks sufficient mass for sustained hydrogen fusion.
Brown dwarfs are generally believed to be <0.08 times the mass of the Sun (< 80 times the mass of Jupiter).
Planet: an object formed in a circumstellar disk by accretion and/or gravitational instability
We do not know if there is a maximum size for a planet but some theorists think they will be < 15 Jupiter masses.
Stars like our sun shine because they are releasing the energy generated by nuclear reactions deep inside their cores. At the center of the sun, the temperatures are several million degrees, and the pressures are much larger than anything which can be generated in a laboratory here on Earth. In this core, hydrogen nuclei (protons) undergo nuclear fusion reactions to form helium. In more massive stars where the temperatures and pressures are even greater, the helium nuclei can fuse together to form carbon nuclei.
In less massive objects, the temperatures and pressures in the stellar core are much lower, and the fusion reactions are much less vigorous. In a star with less than 8% of the mass of our sun, the temperatures and pressures are too small for hydrogen fusion to occur. These objects, which form in the same way as our Sun, simply lack the mass to ignite the nuclear reactions. They will shine for a while by releasing gravitational energy from their formation, and then slowly fade in brightness. We call these “Brown Dwarfs”.
A planet is also a low mass object, but we believe it was formed from the material in the remnant debris disk around a newly formed star, rather than by the same physical process that formed the star. We know that planets such as Mercury or Mars can have very low masses, but we are not really sure of the largest mass that a planet can have. Some astronomers think the upper limit might be about times the mass of Jupiter. While we think planets are formed in a very different process than brown dwarfs, it is quite possible that the mass range of these two types of objects might overlap.

11. How to Search for Planets?
Detection of faint planets next to a bright star is extremely difficult.
There are two approaches to detecting extrasolar planetary systems:
Direct detection attempts to discriminate the light originating from the planet (or reflected from it) from light coming directly from the star.
Indirect detection attempts to detect effects of the planetary system on the light from the star. Usually, this means the detection of orbital motion of the star around the center of mass, or “barycenter” of the star-planet system.
If it were easy to find planets around other stars, it would have been done a long time ago! Planets shine by reflected light. The light that we see from the Moon, or from Jupiter or Venus is just reflected sunlight. Planets also give off heat (infrared light), but our eyes are not sensitive to that. We can see Jupiter at night, because the Sun is not in the sky, masking the much smaller amount of light from Jupiter. In the day, the bright Sun prevents us from seeing Jupiter. (It is possible to see Venus during the day, if you know exactly where to look and have sharp eyes. Try it some time!)
Direct detection attempts to detect the starlight reflected from the planet itself. This is very difficult, because the astronomer must separate the star light from the planet light in some manner.
The alternative is to use the light we can detect, the light from the star, to reveal the presence of the planet. But we must first find something that the planet is doing to that star light that would reveal its presence. The answer is that the star and the planet are in orbit around each other. They both move in a symmetric orbit around an imaginary balance point between them that we call the center-of-mass, or the “barycenter”.

12. Planetary Orbits
The orbit of a planet around a star is an ellipse. The star is at one focus of the ellipse. (This is Kepler’s first law.)
The size of the ellipse is described by the semi-major axis, “a”. a
In 1609, the German astronomer Johannes Kepler published the first two of his three laws of planetary motion. (When first published, these two laws were applied only to the orbit of Mars. But Kepler later extended them to the other planets, and also added a third law of planetary motion.) Kepler’s First Law states that “Each planet moves about the Sun in an orbit that is an ellipse, with the sun at one focus of the ellipse.”
The “long” axis of the ellipse is called the “major axis”, and the short axis is called the “minor axis”. We denote the size of an ellipse by a measurement called the “semi-major axis”, which is often labeled by the letter “a”. The semi-major axis is simply half the length of the major axis.
An ellipse is the set of all points (x,y) in the plane such that the sum of the distances from (x,y) to two fixed points is some constant. The two fixed points are called the foci, which is the plural of focus.

13. Planetary Orbit Shapes
The shape of the ellipse is described by the eccentricity, “e”.
A circle has e = 0.0 :
As the eccentricity increases from 0.0 to 1.0, the ellipse becomes more and more elongated:
We describe the overall shape, or degree of roundness, of the ellipse by the parameter “e”, the eccentricity. Values of “e” can range between 0.0 and 1.0.
A circle is a special case of an ellipse with eccentricity of In this case, the focus is at the center of the circle.
More and more elongated elipses are described by progressively larger values of the eccentricity. For these more eccentric ellipses, the focus moves much closer to one end of the ellipse.
Most of the planets in our solar system have orbits which are nearly circular. Thus, their eccentricities are near zero. Many comets have very eccentric orbits. Their eccentricities are near 1.0.
e =0.6
e = 0.99

14. Direct Detection
In theory, planets around other stars can be detected by imaging in either visible or infrared light.
With a series of images and spectra, you can measure:
semimajor axis: a
eccentricity: e
inclination (the tilt of the planet’s orbit with respect to the plane of the sky)
planet size
temperature
composition
Direct detection of extrasolar planets is, of course, provides the most definitive detection of the planet, and offers the most opportunities to study the properties and characteristics of the planet.
If we can see the planet directly, we can follow its orbit around the star. This tells us the size, shape, and orientation of the orbit. If we can measure the infrared light (the heat) given off by the planet, we can determine the temperature of the planet and its size. If we take a “spectrum” of either the visible or infrared light, we can determine the composition of the planet. a

15. We take a spectrum by spreading out the light into its component colors. We are all familiar with the way a prism will spread sunlight into the different colors, with red at one end and blue and violet at the other end. This is called a spectrum. Our telescopes have very sophisticated instruments which do the same thing. By examining the relative intensity of light of different specific colors, we can determine the composition of the material either giving off or reflecting the light.

16. An isolated planet might be bright enough to observe on its own
An isolated planet might be bright enough to observe on its own. However, we are trying to image a relatively faint object very close to a bright star (like observing a candle near a search light)!
Visible
Infrared
Since the planets “shine” in visible light by reflecting sunlight, they can only reflect a small fraction of the total light given off by the star. Most of the starlight does not strike the planet, and not all of the light that does hit the planet is reflected. (Some is absorbed by the planet, and ultimately heats the planet.) If we were to observe our solar system from a nearby star, most of the planets would appear about a billion times fainter than the Sun in visible light. In the infrared region of the spectrum, they would be between 10,000 times and a million times fainter than the Sun. This, by itself, is not really a problem. Astronomers build big telescopes in order to detect faint objects!
But the problem is to detect the faint object right next to the very bright one. That is very difficult to do with a ground-based telescope. Instead, such direct detection attempts are best done from space-based observatories.

17. This star has a planet around it!
This is an image of the star  Eridani, which was found by UT’s McDonald Observatory Planetary Search Program to have a planet in orbit around it.
This planet is in an orbit which is 3.3 AU from the star and takes 6.8 years to complete one orbit. (“AU” is an abbreviation for “Astronomical Unit”, the average distance from Earth to the Sun. This is about 186 million kilometers, or 93 million miles.) This planet is about 0.86 times the mass of Jupiter.
The planet only shines by reflected light. It is about a billion times fainter than the star, and the starlight reflected off of it is lost in the glare.
Can you find it?

18. Indirect Detection
We are left with using the light from the star to infer the presence of the planet. To do so, we must find something the planet is doing to the star to make its presence known.
The most obvious thing any planet does is to orbit the star.
More correctly, both the planet and the star orbit the center-of-mass (barycenter) of the system.
Since the star is much more massive than the planet, the star moves in a smaller and slower orbit.
In our solar system even the most massive planet, Jupiter, is still only about 1/1047 times the mass of the Sun. The mass of all of the other planets combined does not even add up to the mass of Jupiter. Since the Sun is 1047 times as massive as Jupiter, the center of mass of the solar system is 1/1047 of the way from the Sun to Jupiter. This is just outside the apparent visible “surface” of the Sun. All of the bodies in the solar system (all of the planets, the asteroids, the comets, and the Sun) orbit around this imaginary point, the center of mass. Astronomers call the center of mass the “barycenter”. This term comes from the Greek work “barys”, which means “heavy”.

19. How can we detect the orbital motion of the star and use it to determine the presence of planets?
Radial velocities
Astrometry
Transit photometry
Additional indirect methods (not to be discussed today) include micro-lensing and very accurate pulse timings of pulsars and white dwarf stars.
There are three major techniques of indirect detection that we will discuss in great detail in the next several slides. There are also several other indirect detection techniques which may potentially give interesting results.

20. Star and Planet Orbit Source: STScI, NASA Star and Planet Orbit
Source: STScI, NASA

21. Radial Velocities
We want to measure the changes in the line-of-sight (radial) component of the star as it orbits the system barycenter (center of mass). V*
The technique of “Radial Velocities” has given all of the detections of planets around stars like our sun over the past 8 years. This technique tries to measure the orbital velocity of the star around the system barycenter. The word “Radial” means “along the line of sight”. All we are able to do is measure the velocity along our line of sight. If we look edge-on the plane of the orbit, then the star will be coming directly toward us for part of the orbit, then across our line of sight, then directly away from us, and then across our line of sight again. The velocity will repeat this pattern over and over each time the star orbits the barycenter. If we can measure these changes in the velocity of the star with sufficient precision, we will see a periodic variation in the radial velocity of the star. In this case of looking directly edge-on to the orbit, the maximum and minimum velocities that we see will represent the true orbital velocity of the star. However, if we are looking directly face-on to the orbit, all of the motion of the star will be perpendicular to our line of sight. In that case, we will not see any variation in the radial velocity at all! In the more general case if looking at the orbit from some angle, part of the motion of the star will be along our line of sight, and part will be perpendicular to the line of sight. In this case, we will measure a velocity variation which is somewhat less than the actual orbital velocity of the star. Thus, when we measure a velocity variation of a star, we know that the actual velocity variation of the star is greater than or equal to what we measured. We measure a lower limit to the actual stellar velocity.
Star, M* r*
Planet, mp rp Vp
barycenter

22. Spectroscopic Technique
Copyright 2001 HowStuffWorks

23. What do radial velocity measurements tell us?
We measure the how long it takes the planet to orbit the star.
This tells us the orbital semi-major axis (how far the planet is from the star).
We measure the speed of the star in its orbit.
This tells us the minimum mass of the planet.
The mass might be higher than this minimum value, if we are looking nearly pole-on to the system
We measure the shape of the radial velocity curve.
This tells us the shape of the orbit -- its eccentricity.
Astronomers go to the telescope and measure how the radial velocity of a star varies with time. For some stars, they see the velocity repeat a definite pattern over the years. How do they get information on the planetary orbit from this? We can measure how long it takes the pattern to repeat itself. This is called the “orbital period”. Since the velocity variations are caused by the motion of the star (and the planet) around the barycenter, this orbital period is how long they take to make an orbit -- the size of the planetary “year”. If we can make a good guess of the mass of the star, we can then compute the distance between the star and the planet
The speed of the star in its orbit depends on how large that orbit is. We know the orbital period, which is the time it takes the star to make an orbit. If the orbit is big, the star has to move faster to make it all the way around in one orbital period. We saw earlier that the distance of the star from the system barycenter simply depends on the ratio of the mass of the planet to the mass of the star. More massive planets have the barycenter closer to the planet and thus farther from the star. Thus, in a system with a massive planet, the star must move a long distance during the orbital period, and must move faster. So, the velocity of the star is proportional to the mass of the planet. But remember that the velocity we measure is a lower limit to the actual orbital velocity. Therefore, the planet mass that we compute will be a lower limit to the actual planetary mass.
Orbits with different eccentricities will give a different shape to the change in velocity with time. This allows us to measure the eccentricity of the orbit.
The radial velocity technique is thus most sensitive to massive planets, and to planets close to the star.

24. Examples from our Solar System
For a 1 solar mass star:
Planet Period Speed
Jupiter 11.9 years 13 m/sec (29 mph)
Uranus 84 years 0.3 m/sec (0.7 mph)
Earth 1 year m/sec (0.2 mph)
If we were to look at our solar system from a distance, we would see the Sun moving on its orbit around the solar system barycenter. It will be showing the superimposed velocity variations due to each of the planets in our solar system. The orbit of the Sun is dominated by Jupiter. The reflex motion of the Sun due to Earth is miniscule.

25. Astrometry
Astrometry is the very accurate measurement of the positions of objects (stars, etc.) with respect to a known, fixed reference frame such as distant quasars.
The orbit of a star around a star-planet barycenter will appear as an ellipse when projected on the plane of the sky.
The astrometric signal is largest for:
nearby stars
low mass stars
massive planets
planets far from their parent stars
Astrometry means measuring the positions of the stars in the sky. This is a very traditional field of work for astronomers. This technique is used to measure the orbits of binary star systems, where two stars orbit each other.
The size of the orbital motion of a star due to a planet, however, is very small and is extremely difficult to measure accurately.

26. For our solar system viewed from 10 pc away:
Examples
For our solar system viewed from 10 pc away:
1.0 milliarcsec is 1/3,600,000 degrees!
The current ground-based limits are a few milliarcsec.
To do better, we must use a spacecraft.
So far, there are no good, reliable detections of extrasolar planets via astrometry.
Planet Orbit Size Angular Size Period
Jupiter 5.2 AU 0.5 milliarcsec 11.9 yr
Uranus 19.2 AU 84 microarcsec 84 yr
Earth AU 0.3 microarcsec 1 yr
This shows the angular size of the observed orbit of the Sun around the solar system barycenter when viewed from a distance of 10 parsecs (about 32.6 light years). A “milliarcsec” is seconds of arc. Recall that one degree is broken into 60 minutes of arc, and one minute is broken into 60 seconds of arc.
Thus, 1.0 milliarcsec is 1/3,600,000 degrees!

27. The Motion of Our Sun Around the Barycenter as Seen from 10pc
This shows the actual motion of our Sun around the solar system barycenter from 1960 until The yellow circle is the size of the Sun. Thus, we see that the motion of the Sun describes a spiral pattern of about its own diameter. The motion is dominated by the 11.9 year orbital period of Jupiter, but also shows the effects of the other planets.
0.2 milliarcsec

28. Transit Photometry
A transit occurs when a small body passes in front of a larger body as viewed from the Earth.
In this case, the planet passes in front of the star.
If you are viewing the star when this happens, you will see the brightness of the star decrease because the planet is blocking it.
This requires that the line-of-sight be very close to the orbital plane of the planet around the star.
If we are lucky enough to be looking almost directly edge-on to a planetary system, then once per orbit the planet will pass directly between us and the star, blocking out some of the light of the star. If we make very precise measurements of the brightness of the star, we will notice that the star will dim slightly during this transit event.

29. Transit Technique Copyright 2001 HowStuffWorks

30. How often does this happen?R*
The probability of a transit = R*/a
For a Jupiter-mass body at 0.05AU, this is 10%
For an Earth-mass body at 1AU 0.5%
For a Jupiter-mass body at 5AU 0.1%
The maximum transit duration = (P R*)/(a) (P = orbital period)
For a Jupiter-mass body at 0.05AU, this is ~3 hours
For an Earth-mass body at 1AU ~13 hours
For a Jupiter-mass body at 5AU ~29 hours
The transit depth =(Rp/R*)2
For a Jupiter-mass body (any distance) 1-2%
For an Earth-mass body ~0.008%
The probability of seeing such a transit event is greatest for planets closest to the star, and decreases significantly for planets farther away.
The amount of star light blocked during the transit depends simply on the ratio of the apparent surface areas of the planet and the star. Since radius of Jupiter is about 10% of the radius of the Sun, the area blocked, and thus the amount that the Sun would be dimmed, is 10%x10%, or about 1%.
While the chances of seeing a transit for the planets around any particular star may be low, this method of searching for planets offers some real advantages. Radial velocity measurements of stars must be taken one-by-one. But photometric measurements can be made on thousands of stars at once by taking CCD images of rich star fields. In fact, the background star image on all of these slides is from a region in the constellation Camelopardus that we are surveying for planetary transits using the 0.76-meter telescope at McDonald Observatory.

31. Extrasolar Planet Detection Results (as of November 2002)
There have been over 100 planets found orbiting other stars.
For 11 of the stars, more than 1 planet has been detected.
For all other stars with planets, we so far only know of 1 planet orbiting the star.
This may be a selection effect since it is harder to find smaller planets or planets far from the star
The first extrasolar planet around a solar type star was found in Since then over 100 planets have been found, all via radial velocity techniques. The rate of discovery of extrasolar planets is continually increasing, as our techniques improve. Several planetary systems of more than one planet have been found.
We now have many examples to compare to our Solar System and we find that none have the same properties as our system!

32. These are called “selection effects”.
Caveat!
All of our conclusions about the types of planets that exist are biased by the methods we use to find planets and the properties of those planets to which these methods are most sensitive.
These are called “selection effects”.
When we interpret the results from searches for extrasolar planets, we have to remember that the radial velocity technique, which is responsible for essentially all of the results to date, is most sensitive to massive planets and to planets close to the parent star. Just because we have not yet discovered a planet around any particular star does not mean that the star does not have planets. It just means that the planets that might be there would give a radial velocity signal too small for us to detect unambiguously.

33. Latest Results from McDonald Observatory: The planet around  Cephei A A planet in a close Binary Star System
 Cephei A is the third brightest star in the constellation Cepheus
It is a close binary star (the 2 stars can not be separated by eye
We have found a planet in a 2.5 year period orbit around  Cephei A.
Other planets have been found in wide binary star systems, but never in one with the stars so close together.
This is the latest result from the McDonald Observatory Planet Search program. This star was on the observing list of the pioneering radial velocity planet search program of Bruce Campbell and Gordon Walker, which operated on the Canada-France-Hawaii Telescope from 1980 until They saw the 2.5 year velocity variation, but thought that it was due to internal variability within the star γ Cephei A itself. Our data show that the velocity variations continue to repeat exactly for over 20 years! We have now eliminated any internal causes of the variations, and have concluded that a planetary companion to γ Cephei A is the best explanation of the observed radial velocity variations.
This is not the first planet found around a star in a binary system, but all of the other binary stars with planets have the two stars very far apart from each other. In this system, the stars are much closer together, in this case only about 22 astronomical units on the average.
Most of the stars in the sky are part of a binary- or multiple-star system. This is one aspect in which our Sun is unusual -- there is no other star in orbit around the Sun. If planets can form in a very close binary star system such as this one, then it opens up a large number of additional stars that might have planetary companions to them.

34. This is an artist’s conception of what the planet in the γ Cephei system might look like. The bright star on the right is γ Cephei A and the smaller, fainter, redder star on the left is γ Cephei B. We don’t know if the planet actually has a nice system of rings around it, but all of the gas-giant planets in our solar system have some sort of a ring system. Notice the shadows of the planet cast on the rings by the two stars.
This figure was made by Tim Jones, of McDonald Observatory.

35. “Hot” Jupiters or 51 Peg-Type Stars
The first extrasolar planet detected was orbiting 51 Peg. The first few planets found had properties similar to the planet around 51 Peg (called 51 Peg b).
These objects are in very short-period orbits
They are planets of about the same mass as Jupiter.
51 Peg
P = days
(Mayor and Queloz 1995)
100 50
-50
-100
V (m/sec) 1
Phase
The first extrasolar planet found around a solar type star is the companion to the star 51 Pegasi. This system was found by Michel Mayor and Didier Queloz of the Geneva Observatory in This planet was a complete shock in many ways. The major surprising aspect was that the orbital period is only 4.2 days. Recall that Jupiter’s orbital period is years. This very short period orbit meant that the planet orbited the star at a distance of only 0.05AU, or 1/8 of the distance from the Sun to Mercury. Nothing in our experience led us to believe that a Jupiter-type planet could possibly exist in an orbit so close to the parent star. Our theory of planet formation says that planets can not form so close to the star because the temperatures are too hot, and there is not sufficient material in the circumstellar disk at that location to form such a massive planet. Since then about a dozen other similar planets have been found around other solar type stars. This class of planets are called “51-Peg Type” planets, “hot Jupiters”, or “roasters”.
Of course, radial velocity measurements only gave us a minimum mass for the planet. They don’t tell us what the planet is made of. So, one question that came up immediately was whether these close-in planets are gas-giants like Jupiter, or extremely massive rocky terrestrial planets.
We see the orbital speed increasing and decreasing
every days.
Are these “hot” Jupiters really gas giant planets,
or can we have super-massive terrestrial planets?

36. The Planet Around Star HD 209458
Charbonneau et al. 2000
The planet around HD was discovered by the radial velocity signal.
After discovery, a transit was observed  this gives a good measure of the inclination. From M sin i, we can then derive M.
MP= 0.69 ± 0.05 MJUP
RP = 1.40 ± 0.17 RJUP
density =  = 0.31 ± 0.07 g/cm3
intensity
The star HD was found by radial velocity techniques to have a close-in “51-Peg” type planets. A planet this close to the parent star will have about a 10% chance of being oriented close enough to edge-on to show transits of the planet across the disk of the star. Therefore, as each new “51-Peg” type system was discovered, astronomers closely monitored them to see if their brightness did indeed drop slightly once per orbital period. This star was the first (and so far the only) star to show such periodic transit events. This discovery was made by David Charbonneau and Tim Brown. They found that once every 3.5 days the brightness of the star dropped by 1.7%. This is exactly what would be expected for a planet slightly larger than Jupiter. This constituted the first dramatic confirmation that the radial velocity variations we are seeing are indeed caused by the orbital motion due to a planetary companion. We had always assumed this, but this was an extremely important result.
Definitely a gas giant planet like Jupiter!

37. The 51 Peg-Type Planets Sun to scale Sun NOT to scale
To show the variety of orbits of the “51 Peg” type planets, here we superimpose the orbits of several of these planets on a drawing of the orbits of the inner planets in our solar system. In the drawing on the left, the size of the dot representing the central star is to the same scale as the sizes of the orbits. Some of these planets are indeed orbiting at only 6-10 stellar radii! A very large number of extrasolar planets have been found in orbits with semi-major axes smaller than that of Mercury in our solar system.
Sun to scale Sun NOT to scale

38. Massive Eccentric Planets
In our Solar System, the planetary orbits have low eccentricities (with the exception of Pluto).
This is a natural consequence of planet formation.
This is what we expected for other planetary systems.
100 50
Relative Radial Velocity (m s-1)
Another class of planets that was completely unexpected was the “massive eccentric planets”. We show here the radial velocity data for the companion to the star 16 Cygni B, which was discovered to have a 1.6 Jupiter mass companion by the McDonald Observatory Planet Search program. This planet takes about 800 days to orbit its parent star, but the eccentricity of the orbit is about This is far larger than the eccentricities of any of the planets in our solar system. Our current model for the formation of our solar system will give orbits that are close to circular. How to get these massive eccentric planets is an active subject of research.
-50
-100
1988
1990
1992
1994
1996
1998
2000
2002
Date

39. Increasing Eccentricity
When we look at the list of planets which have been discovered so far, we see that low eccentricities are not common.
If we look at the orbital characteristics of the planets that have been found so far around other stars, we see that they span a very wide range of eccentricities and orbital periods. It seems that highly eccentric planets are quite common. We also see that planetary orbital periods cover virtually all possible orbital periods. Massive Jovian-type planets do not occur only at the large heliocentric distances of the Jovian planets in our solar system. In many other planetary systems, gas-giant planets occur at much smaller distances. This is a difficulty for our current model of planetary system formation.
Increasing Eccentricity
Increasing eccentricity

40. When we look at the list of planets which have been discovered so far, we see that low eccentricities are not common.
Except for the 51 Peg-type systems, the eccentricities are relatively large, independent of mass.
If we look at the orbital characteristics of the planets that have been found so far around other stars, we see that they span a very wide range of eccentricities and orbital periods. It seems that highly eccentric planets are quite common. We also see that planetary orbital periods cover virtually all possible orbital periods. Massive Jovian-type planets do not occur only at the large heliocentric distances of the Jovian planets in our solar system. In many other planetary systems, gas-giant planets occur at much smaller distances. This is a difficulty for our current model of planetary system formation.

41. How massive are the planets we have found?
If we look at a histogram of masses, we see that we have preferentially found objects with small masses.
The radial velocity technique is sensitive to larger masses.
Therefore, there appears to be very few brown dwarfs orbiting these stars.
When we make a histogram of the minimum masses of the known extrasolar planets, we see a very surprising result. Recall earlier that we said that we have to be wary of selection effects biasing our interpretations. Radial velocity techniques, which were responsible for all of these detections, are most sensitive to massive planets. But the observed mass distribution peaks in the lowest mass bin! We have probably found all of the massive planets in orbits of a few years or less around all of the stars surveyed. However, there are undoubtedly a very large number of low-mass planets that we have not yet discovered. Thus, when we correct this diagram for the known selection effects, this peak at low masses will only get much larger. This large peak indicates that there is probably a very large number of lower mass objects remaining yet to be discovered. These would be objects similar to Saturn, Uranus and Neptune.

42. How do we get planets into very close orbits around the stars?
Maybe they form at large distances from the star, and then are moved to their present locations by some process. This is called orbital migration.
There is an elaborate theory that uses tidal forces raised by the planet in the disk of material from which it was formed to cause the inward migration.
Of course, the migration must stop before the planet plunges into the star!
Perhaps the planets formed where we find them today, but by some different process than we think formed our solar system.
Explaining these results is now proving to be a major problem. Astronomers are loathe to abandon our theory for the formation of our solar system. However, this theory has great difficulty in forming gas-giant planets very much closer to the parent star than we find Jupiter in our solar system. It also wants to form planets in nearly circular orbits. If we assume that our model is basically correct, then the question arises as to why we find planets very close to the parent star, and why planets are found in such a wide distribution of eccentricities.
The first approach that was taken was to assume that planets form according to our theory for the solar system, and then some other physical processes put the planets into the orbits in which we find them. One way to move planetary orbits in semi-major axis is through orbital migration. The newly formed planet can interact with the residual proto-planetary disk via gravitational tidal forces. The planet accretes all of the material in the disk at the same radius as it is orbiting. The planet then raises tidal bulges in the disk material both interior and exterior to its orbit. These tidal bulges will then tug on the planet, and either slow it down or speed it up in its orbit, depending on which tidal bulge is more massive. This will then cause the orbit to migrate either inward or outward. For most circumstances, the planets can move inward. However, this tidal migration also causes the orbital eccentricity to decrease nearly to zero. Thus, this mechanism can move planets into the inner regions of these systems, but can not account for the large eccentricities.

43. How do we account for the very large orbital eccentricities?
Interactions with any remaining planetesimals:
A slow inward migration can still occur once the gas disk is gone if there are still a lot of planetesimals remaining.
Gravitational interactions among planets:
A system of multiple planets can become dynamically unstable once the gas disk dissipates. Planets can collide, be ejected, or be put into eccentric orbits.
Dynamical interactions with other stars:
If most stars form in dense star clusters, then the gravity of nearby stars could disrupt newly formed planetary systems.
So, how can we get the large orbital eccentricities? A number of theories have been offered, but none of them really seem to work well in all cases. This is certainly an area in which considerably more theoretical work needs to be done!

44. An amusing cartoon! You can find more similar cartoons on the web at
Cartoon used with permission. Copyright 2001 David Farley.

45. Multi-planet systems around other stars
Our Solar System has many planets, all orbiting the Sun.
Multiple star systems form in hierarchical systems (e.g. a pair of stars orbit another star as a pair).
Thus, a true test for if objects are in a planetary system is to find multiple objects orbiting the same star. St ar M
sin i ( J UP ) P er
iod d a ys
(AU) e HD
83443 . 35 2
986
038 08 17 29 83
174 42 U ps A nd 71 4
617
059 02 05
241 3
828 24
1308 5 2 . 56 31 55 Cn c 93 14 66
118 03
> ?
2920 HD
82943 88
221 6
728 54 1 63
444 16 41
74156 51 60
276 65 7 46
2300 3 47 40 G li e s
876 30 12
130
We would expect several planets to be found around most stars. After all, our solar system has nine planets (or eight, if you chose not to count Pluto). Here I list the characteristics of several of the multi-planet systems which have been found so far. . 27 1 89 61 02
207 10 47
U M a 2 56
1090 5 09 06 76
2640 3 78 HD
168443 7 64 58
295 53 16 96
1770 87

46. Selection Effects in the discovery of extrasolar planets
Selection effects are important! We are most sensitive to the short period systems.
I really do have to emphasize the importance of remembering the selection effects of radial velocity methods when we try to interpret the results. This is called the “discovery space” plot for planet detection, in which we plot the detected planet minimum mass as a function of the system semi-major axis. The large white dots represent the known extrasolar planets. For reference, the major planets in our solar system are indicated by dots labeled by “M’ for Mars, “E” for Earth, “J” for Jupiter, etc. The green diagonal lines show the masses of planets detected by radial velocity techniques, for a variety of system radial velocity amplitudes. At present, radial velocity techniques can achieve about 3 meters/second precision in a single measurement. We hope to improve this to about 1 meter/second soon. This plot shows that such an improvement will allow us to detect lower mass planets. We see that the discovery of planets in longer period orbits is limited by the length of time we have been observing the star. We can not determine an orbital solution until we have seen the system complete at least one full orbit. We are now rapidly closing in on being able to discover true analogs to our solar system.

47. Expectations Reality Planetary Systems are common
So far, about 4% of main-sequence stars have detectable planets.
There are many planets per system.
Only a few systems of multiple planets have been found so far.
Jupiter is the prototype giant planet.
Many objects are more massive than Jupiter. Radii, composition, and structure are unknown.
When groups such as ours started searching for planets around other stars, our expectations were colored by our model for the formation of our solar system. This led us to believe that other planetary systems would have certain characteristics. Here we list several of the expectations that we had before any extrasolar planets were discovered, and compare them to what has been found so far.
Prograde or direct motion is
(1) Rotation or orbital motion in a anticlockwise direction when viewed from the north pole of the Sun (i.e. in the same sense as the Earth); the opposite of retrograde.
(2) The East-West motion of the planets, relative to the background of stars, as seen from the Earth.
All planets are in prograde orbits near the stellar equatorial plane.
Orbital inclinations are unknown.

48. Expectations Reality a > 5AU for gas giants.
There is a very wide range of semimajor axes. How did “hot” Jupiters get so close to their parents?
Nearly circular orbits.
Many highly eccentric orbits.
Terrestrial planets in the inner regions.
Terrestrial planets not yet detectable.
However, once again, we have to realize that the “Reality” that we list is based solely on the 100 or so planets that have been discovered through the fall of As more and more planets are found by a variety of techniques, we will begin to eliminate the influence of the observational selection effects. Then, our perception of “Reality” may well change some!
Configurations are dynamically stable.
Dynamical evolution is probably very important.

49. Summary
As we get longer time baselines and better precision, we will be able to detect systems which look more like our own.
We need larger surveys of more stars to really understand the diversity of planetary systems.
We need to keep modifying our formation models to account for what we have found.
It is incredibly fun to be able to experience a revolution in scientific thought first-hand. That is what is happening right now in this field of extra-solar planets. Our understanding of the true diversity of planetary systems in the galaxy will continue to evolve over the next several years. Eventually, the theory of how these systems form and evolve will begin to catch up with the observations. Hopefully, in the next 5 to 10 years we will have a much clearer understanding of the formation and evolution of planetary systems.

50. Credit: Tim Jones, of McDonald Observatory.

51. Dr. William Cochran Senior Research Scientist
Dr. William Cochran received his Ph.D. from Princeton University in His research interests include extrasolar planetary systems, high-precision measurements of stellar radial velocity variations, variable stars, planetary atmospheres, comets, and asteroids.