| Lecture |
|
The Search for and Discovery of Extra-Solar Planets |
Prior to 1995, we had just one planetary system to study: ours. For any scientist or statistician, it is impossible to do any kind of comparison study if there is just one data point! Our prediction for other systems was Sun-like stars with one or two Jupiter-sized planets. These planets would have orbital periods longer than 5 years or so. In order to detect any radial velocity variation in the parent stars, variation which was bound to be small, the astronomers needed to have highly stable equipment, good seeing conditions, and large telescopes. It was also believed that these observations would have to span a period of at least 10 years. The first group to start looking for these small motions of stars was led by Dr. Gordon Walker of the University of British Columbia.
The Canadian team started observing dozens of nearby, solar-type stars in 1981, using the 4-meter Canada-France-Hawaii telescope perched high above the Pacific Ocean on Mauna Kea. Dr. Walker was joined by astronomers Stephenson Yang and Alan Irwin of the University of Victoria, and eventually then-graduate students Andrew Walker and Ana Larson. When enough data were gathered and analyzed, this team, and the rest of the astronomy community, had a big surprise: there was no evidence of even a single extra-solar planet. Were we then alone? Were all of our theories of planets being formed as a natural by-product of their stars wrong? The Universe had yet another surprise in store for us.
The student will be able to:
We cannot see an extrasolar planet. Why Not? Planets reflect the light from their parent star, and are approximately 1 billion times less luminous! Seeing a planet directly would be like reading the inscription on a flash bulb as it went off in your face, or seeing a gnat flying right next to a street light a couple of blocks away, or seeing a dying ember next to a blast furnace.
Planets are very close to their star, and even for the closest stars we do not have the resolution needed to see the planets. Astronomers have a way of describing angles in the sky based upon how far away the objects are. One of the units they use is called the "parsec." (We don't need to be concerned with the actual definition here, just remember that 1 parsec is equal to 3.26 light years.) If we were 1 parsec away from the Sun, looking back, the angle between the Earth and the Sun would be a teeny-tiny 1/3600 of a degree! The nearest star is about 4.2 light years away, and so if we were there, the angle between the Sun and the Earth would be even less. So, how about Jupiter? At 1 parsec away, Jupiter would be 1/720 of a degree away from the Sun, from our point of view. The angular size of 1/3600 of a degree is called an arc second. The largest telescopes on Earth, and even the Hubble Space telescope, currently can resolve objects that are about 1/1800 of a degree or 1/2 arc second apart. Unfortunately, we still have the overpowering light from the parent star that blocks our view of the planet.
To understand how we detect planets based upon the Doppler shift of the light of THE PARENT STARS, we need to return briefly to why the stars move in the first place. Recall that the star and its planets are all orbiting a center of mass. To simplify our lesson here, let's assume that we have a star that has just 1 planet, a planet much like Jupiter. We can vary the closeness of the planet to its star as well as the mass of the planet. You should experiment with changing these variables and see what happens. [If you do not remember exactly what is meant by the Doppler shift of light, then return to the lecture on light and review that section.]
For the orbital animations I showed in class, go to: http://csep10.phys.utk.edu/, click on "sample chapter" link shown in left-hand frame, click on "The Modern Synthesis," "Universal Gravitation," "Kepler's 3rd Law." Scroll down a bit in this sub-window, and see a side-bar that says "Effect of the Center of Mass." You should see a link to a Java applet that demonstrates Newton's modified form of Kepler's third law. Play around with the eccentricity of the orbit and the masses of the bodies. There are other very good illustrations and animations of the fundamentals of gravity, motions, and discoveries.
The amount of wobble detected in a star due to a planet will depend on how the system is oriented towards our line of sight. If we are looking at it nearly edge on, such as is seen in Fig. A above, then we are seeing close to the maximum amount of motion; if we are looking at it pole on, as is seen in Fig. B above, we would not see any motion as there is no component of the movement towards or away from us. When we talk about the mass of a planet that has been discovered orbiting another star, we are really talking about the minimum mass that it is; the planet could possibly be more massive.
The amount of wobble depends on the mass of the planet--the more massive the planet the more the star wobbles. Technically, what is really happening is that the center of mass of the star and planet moves slightly closer to the planet, and so the star must move in a slightly large orbit about the center of mass. The period of the wobble depends on how close the planet is to the star (Kepler's Third Law!). The closer the planet is to the star, the shorter the period. Now, given the information so far, what would the "ideal" conditions be for detecting a planet around another star?
A planetary transit is basically an eclipse. A planet crosses the disk of a star and we observe its shadow.
The geometry of the system has to be such that the planet will actually cross in front of the star. If the plane of the orbit of the planet is tilted too much, the planet will not transit. Even during a transit, the size of the planet is small compared to the star and so the amount of starlight that is blocked by the planet is exceedingly small. Astronomers have instruments that can detect these small decreases in the light from the star.
Source: Capabilities of Various Planet Detection Methods
"Photometry measures the periodic dimming of the star caused by a planet passing in front of the star along the line of sight from the observer. Stellar variability on the time scale of a transit limits the detectable size to about half that of Earth for a 1 AU orbit about a 1 Mo star or Mars size planets in Mercury-like orbits with four years of observing. Mercury-size planets can even be detected in the habitable-zone of K and M stars. Planets with orbital periods greater than two years are not readily detectable, since their chance of being properly aligned along the line of sight to the star becomes very small. Giant planets in inner orbits can also be detectable independent of the orbit alignment, based on the periodic modulation of their reflected light. For the 10% of these that have transits, the transit depth can be combined with the mass found from Doppler data to determine the density of the planet as has been done for the case of HD209458b and see if these inner giants are "inflated". Doppler spectroscopy and astrometry (SIM) measurements can be used to search for any giant planets that might also be in the systems discovered using photometry. Since the orbital inclination must be close to 90° (sin i=1.) to cause transits, there is very little uncertainty in the mass of any giant planet detected.
Source: Capabilities of Various Planet Detection Methods
"Astrometry is used to look for the periodic wobble that a planet induces in the position in the sky of its parent star. The minimum detectable planet mass gets smaller in inverse proportion to the planet's distance from the star. For a space-based astrometric instrument, such as the planned Space Interferometry Mission (SIM), that could measure an angle as small as 2 micro-arcsec, a minimum planet of mass of 6.6Me could be detected in a 1 year orbit around a 1 Mo star that is 10 pc from the Earth and a 0.4 MJ planet in a 4 year orbit. The FAME mission (Full-sky Astrometric Explorer) has an angular resolution of 50 micro-arcsec and the minimum detectable planet mass for it at 10 pc is shown by the descending orange line. From the ground, the Keck telescope is being equipped to measure angles as small as 20 micro-arc seconds, leading to a minimum detectable mass in a 1 AU orbit of 66Me for a solar-mass star at 10 pc. The limitations to this method are the distance to the star and variations in the position of the photometric center due to star spots. There are only 33 non-binary solar-like (F, G and K) main-sequence stars within 10 pc of the Earth. The furthest planet from its star that can be detected is limited by the time needed to observe at least one orbital period. This limit is indicated by the dashed light-blue vertical line chosen to be at 10 years in the figure below. There are no planet detections that have been confirmed using this method."
The astrometric approach (image from
University of New Mexico).
There are additional methods for detecting extrasolar planets: microlensing, gravitational redshift, and the ultimate direct detection. We won't be considering these methods here.
Excluding the multiple, small, rock-like planets discovered orbiting a neutron star (planets which must have formed from material expelled from a massive star that had died and thus different from the planets discussed here), the first discovery of an extra-solar planet (a planet lying outside of our planetary system) came in 1995 by two French astronomers (see text). Amazingly enough, this planet did not have an orbit of 10 or so years, or even a few years, or even a few months. This planet had an orbital period of less than 5 days! The immediate response from many astronomers was that this could not be a planet--the amplitude of the radial velocity variations indicated that the planet was approximately the mass of Jupiter, but much closer to its star than Mercury is to the Sun. No way!
True to scientific form, astronomers across the globe set up programs to observe the parent star, 51 Pegasi (star in the constellation of Pegasus that was given a catalog number), in hopes of disproving that the variations were due to a planet. Among the potential explanations were huge star spots, surface oscillations, and a star seen nearly pole-on. None of the subsequent observations could disprove the possibility that this was a planet orbiting a star. Even though we cannot see the planet, and cannot say for a fact that it is actually there (science does not set out to prove anything anyway), a Jupiter-like planet orbiting a Sun-like star BEST describes the observations. Had the Canadian team had this star on its list and had observations been more frequent than a few days every 3-4 months, they would have had the scoop. Such is the world of discovery.
Your text does an outstanding job of discussing the ongoing discoveries of planets around other stars and the possibilities that maybe some of these planets are at distances from their stars where the conditions would be conducive for life to arise. Be sure to read this section carefully, especially the details of what kinds of observations are used to detect these planets, why we think they really are planets, and the problems inherent in the fact that we do not (at least not yet) have the technology to actually see any of these planets. The links provided below to the extra-solar planet web sites will help you share in the excitement of knowing that we are probably not alone.
Since 1995, close to 100 planets have been detected. The following two graphs (follow the links for a closer look) summarize what a galactic zoo they are:
Also take a look at this table from the "exoplanet.org" web pages: PLANET TABLE.
These observations result from what astronomers call an observational bias. Because of Newton's Law of Gravitation, and Kepler's Law relating period and orbital distance, these planets are the ones were most likely to see. Newton's Law states that the force is proportional to mass, and inversely proportional to distance, therefore big planets close to their stars will move the star the fastest. Kepler's Law states the farther a planet is from a star, the longer it takes to orbit. Therefore in order to observe 1 orbit (and hence exlude other possible reasons for the observed doppler shift), it takes a long time. Jupiter orbits at 5.2AU, and takes about 12 years to orbit the sun. As these biases are removed, new planets will undoubtedly be discovered.
The Future of Extra-Solar Planet Searches
A lot of money is being poured in this research, with NASA having many missions in the planning stages:
New Ideas About Planet Formation
Image from
KH 15D press conference
Wesleyan University, Van Vleck Observatory
Stars strongly suspected of having one or more planets:
Known planetary systems in nice tabular format.
SEDS Other Planetary Systems (A MUST-VISIT SITE)
Find these stars at the PBS worlds
A whole site on exoplanets
A binary star suspected of losing one of its planets
Kepler to Search for Terrestrial Planets
Last updated on:
