Download Detecting and Characterizing Extrasolar Planets: A New Frontier in Astronomy and more Study notes Law in PDF only on Docsity!
Other Planetary Systems
The New Science of Distant Worlds
L E A R N I N G G O A L S
13.1 Detecting Extrasolar Planets
◗ Why is it so difficult to detect planets around other stars? ◗ How do we detect planets around other stars?
13.2 The Nature of Extrasolar Planets
◗ What have we learned about extrasolar planets? ◗ How do extrasolar planets compare with planets in our solar system?
13.3 The Formation of Other Solar Systems
◗ Can we explain the surprising orbits of many extrasolar planets? ◗ Do we need to modify our theory of solar system formation?
13.4 Finding More New Worlds
◗ How will we search for Earth-like planets?
c h a p t e r 1 3 • Other Planetary Systems 395
How vast those Orbs must be, and how inconsiderable this Earth, the Theatre upon which all our mighty Designs, all our Navigations, and all our Wars are transacted, is when compared to them. A very fit consideration, and matter of Reflection, for those Kings and Princes who sacrifice the Lives of so many People, only to flatter their Ambition in being Masters of some pitiful corner of this small Spot. —Christiaan Huygens, c. 1690
A
little more than a decade ago, all of planetary sci- ence was based solely on the study of our own solar system. Then, beginning in 1995, a dramatic change occurred as scientists began to detect planets around other stars. More than 250 such planets were already known by 2007, and new discoveries are coming rapidly. We are even beginning to learn about the characteristics of these distant worlds. The discovery of planets around other stars represents a triumph of modern technology. It also has profound philosophical implications. Knowing that planets are com- mon makes it seem more likely that we might someday find life elsewhere, perhaps even intelligent life. Moreover, having many more worlds to compare to our own vastly enhances our ability to learn how planets work and may help us better understand our home planet, Earth. The study of other planetary systems also allows us to test in new settings our nebular theory of solar system formation. If this theory is correct, it should be able to ex- plain the observed properties of other planetary systems as well as it explains our own solar system. In this chapter, we’ll focus our attention on the exciting new science of other planetary systems.
Detecting Extrasolar Planets Tutorial, Lessons 1–
13.1 Detecting
Extrasolar Planets
The very idea of planets around other stars, or extrasolar planets for short, would have shattered the worldviews of many people throughout history. After all, cultures of the western world long regarded Earth as the center of the uni- verse, and nearly all ancient cultures imagined the heavens to be a realm distinct from Earth. The Copernican revolution, which taught us that Earth is a planet orbiting the Sun, opened up the possibility that planets might also orbit other stars. Still, until quite re- cently, no extrasolar planets were known. In this first sec- tion we’ll discuss why the detection of extrasolar planets presents such an extraordinary technological challenge and how astronomers have begun to meet that challenge.
Before we begin, it’s worth noting that these discoveries have further complicated the question of precisely how we define a planet. Recall that the 2005 discovery of the Pluto- like world Eris [Section 12.3] forced astronomers to recon- sider the minimum size of a planet, and the International Astronomical Union (IAU) now defines Pluto and Eris as dwarf planets. In much the same way that Pluto and Eris raise the question of a minimum planetary size, extrasolar planets raise the question of a maximum size. As we will see shortly, many of the known extrasolar planets are con- siderably more massive than Jupiter. But how massive can a planet-like object be before it starts behaving less like a planet and more like a star? In Chapter 16 we will see that objects known as brown dwarves , with masses greater than 13 times Jupiter’s mass but less than 0.08 times the Sun’s mass, are in some ways like large jovian planets and in other ways like tiny stars. As a result, the International Astronom- ical Union defines 13 Jupiter masses as the upper limit for a planet.
◗ Why is it so difficult to detect
planets around other stars?
We’ve known for centuries that other stars are distant suns (see Special Topic, p. 386), making it natural to sus- pect that they would have their own planetary systems. The nebular theory of solar system formation, well established by the middle of the 20th century, made extrasolar planets seem even more likely. As we discussed in Chapter 8, the nebular theory explains our planetary system as a natural consequence of processes that accompanied the birth of our Sun. If the theory is correct, planets should be com- mon throughout the universe. But are they? Prior to 1995, we lacked conclusive evidence. Why is it so difficult to detect extrasolar planets? You already know part of the answer, if you think back to the scale model solar system discussed in Chapter 1. Recall that on a 1-to-10-billion scale, the Sun is the size of a grapefruit, Earth is a pinhead orbiting 15 meters away, and Jupiter is a marble orbiting 80 meters away. On the same scale, the distance to the nearest stars is equivalent to the distance across the United States. In other words, seeing an Earth-like planet orbiting the nearest star besides the Sun would be like looking from San Francisco for a pinhead orbiting just 15 meters from a grapefruit in Wash- ington, D.C. Seeing a Jupiter-like planet would be only a little easier. The scale alone would make the task quite challeng- ing, but it is further complicated by the fact that a Sun- like star would be a billion times as bright as the light reflected from any planets. Because even the best tele- scopes blur the light from stars at least a little, the glare of scattered starlight would overwhelm the small blips of planetary light.
c h a p t e r 1 3 • Other Planetary Systems 397
Jupiter
Sun
center of mass
Not to scale!
Jupiter actually orbits the center of mass every 12 years, but appears to orbit the Sun because the center of mass is so close to the Sun.
The Sun also orbits the center of mass every 12 years.
Jupiter half an orbit later
Sun half an orbit later
Figure 13.1 This diagram shows how both the Sun and Jupiter actually orbit around their mutual center of mass, which lies very close to the Sun. The diagram is not to scale; the sizes of the Sun and its orbit are exaggerated about 100 times compared to the size shown for Jupiter’s orbit.
Interactive Figure
Direct detection is preferable, because it can tell us far more about the planet’s properties. However, current telescopes are not quite up to the challenge of direct detection, at least for planets around ordinary stars. As a result, nearly all extrasolar planets discovered to date have been found by indirect techniques. Let’s now explore detection techniques in a little more detail.
Gravitational Tugs To date, nearly all extrasolar planets have been detected by observing the gravitational tugs they exert on the stars they orbit. This type of detection is indi- rect because we discover the planets by observing their stars without actually seeing the planets themselves. Although we usually think of a star as remaining still while planets orbit around it, that is only approximately cor- rect. In reality, all the objects in a star system, including the star itself, orbit the system’s “balance point,” or center of mass [Section 4.4]. To understand how this fact allows us to dis- cover extrasolar planets, imagine the viewpoint of extra- terrestrial astronomers observing our solar system from afar. Let’s start by considering only the influence of Jupiter, which exerts a much stronger gravitational tug on the Sun than the rest of the planets combined. The center of mass between the Sun and Jupiter lies just outside the Sun’s visi- ble surface (Figure 13.1), so what we usually think of as Jupiter’s 12-year orbit around the Sun is really a 12-year orbit around the center of mass; we generally don’t notice this fact because the center of mass is so close to the Sun itself. In addition, because the Sun and Jupiter are always on opposite sides of the center of mass (otherwise it wouldn’t be a “center”), the Sun must orbit this point with the same 12-year period as Jupiter. The Sun’s orbit traces out only a very small circle (or ellipse) with each 12-year period, be-
cause the Sun’s average orbital distance is barely larger than its own radius. Nevertheless, with sufficiently precise mea- surements, extraterrestrial astronomers could detect this orbital movement of the Sun. They could thereby deduce the existence of Jupiter, even without having observed Ju- piter itself. They could even determine Jupiter’s mass from the Sun’s orbital characteristics: A more massive planet at the same distance would pull the center of mass farther from the Sun’s center, giving the Sun a larger orbit and a faster orbital speed around the center of mass.
S E E I T F O R Y O U R S E L F To see how a small planet can make a big star wobble, find a pencil and tape a heavier object (such as a set of keys) to one end and a lighter object (perhaps a small stack of coins) to the other end. Tie a string (or piece of floss) at the balance point—the center of mass—so the pencil is horizontal, then tap the lighter object into “orbit” around the heavier object. What does the heavier object do, and why? How does how your setup corre- spond to a planet orbiting a star? You can experiment further with objects of different weights or shorter pen- cils; try to explain the differences you see.
Now let’s add in the effects of Saturn, which exerts the second greatest gravitational tug on the Sun. Saturn takes 29.5 years to orbit the Sun, so by itself it would cause the Sun to orbit their mutual center of mass every 29.5 years. However, because Saturn’s influence is secondary to that of Jupiter, this 29.5-year period appears as a small added effect on top of the Sun’s 12-year orbit around its center of mass with Jupiter. In other words, every 12 years the Sun would return to nearly the same orbital position around its center of mass with Jupiter, but the precise point of return would move around with Saturn’s 29.5-year period. By measuring this motion carefully from afar, an extraterres- trial astronomer could deduce the existence and masses of both Jupiter and Saturn after a few decades of observing. The other planets also exert gravitational tugs on the Sun, which further affect the Sun’s orbital motion around the solar system’s center of mass (Figure 13.2). These extra effects become increasingly difficult to measure in practice, but extremely precise observations would allow an extra- terrestrial astronomer to discover all the planets in our solar system. If we turn this idea around, you’ll realize that it means we can search for planets in other star systems by carefully watching for the tiny orbital motion of a star around the center of mass of its star system. Two techniques allow us to observe the small orbital motion of a star caused by the gravitational tugs of planets. (1) The astrometric technique uses very precise measure- ments of stellar positions in the sky ( astrometry means “measurement of the stars”) to look for the stellar motion caused by orbiting planets. (2) The Doppler technique can detect orbital motion through changing Doppler shifts
398 p a r t I I I • Learning from Other Worlds
2015 2005 1980
2010
2020
1975 20001965
2025
1990
1995
1960
1970
1985
0.0005 arcsecond radius of Sun
center of mass
Figure 13.2 This diagram shows the orbital path of the Sun around the center of mass of our solar system as it would appear from a distance of 30 light-years away for the period 1960–2025. Notice that the entire range of motion during this period is only about 0.0015 arcsecond, which is almost 100 times smaller than the angular resolution of the Hubble Space Telescope. Neverthe- less, if alien astronomers could measure this motion, they could learn of the existence of planets in our solar system.
in a star’s spectrum [Section 5.5]; a star’s orbital motion will produce alternating blueshifts as the star moves toward us in its orbit and redshifts as it moves away. Each technique has advantages and limitations.
The Astrometric Technique The astrometric technique has been used for many decades to identify binary star sys- tems, since two orbiting stars will move periodically around their center of mass. The technique works especially well for binary systems in which the two stars are not too close together, because the stellar motions tend to be larger in those cases. In the case of planet searches, however, the expected stellar motion is much more difficult to detect. For example, from a distance of 10 light-years, a Jupiter- size planet orbiting 5 AU from a Sun-like star would cause its star to move slowly over a side-to-side angular distance of only about 0.003 arcsecond—approximately the width of a hair seen from a distance of 5 kilometers. Remarkably, with careful telescope calibration astronomers can now measure movements this small, and instruments currently under development will be 5 to 10 times more precise. How- ever, two other complications add to the difficulty of the astrometric technique. The first complication comes from the fact that the far- ther away a star is, the smaller its side-to-side movement will appear. For example, while Jupiter causes the Sun to move by about 0.003 arcsecond as seen from 10 light-years away, the observed motion is only half as large when seen from 20 light-years away and one-tenth as large when seen from 100 light-years away. The astrometric technique therefore works best for massive planets around relatively nearby stars. The second complication arises from the time required to detect a star’s motion. It is much easier to detect larger
The planets in our solar system have familiar names rooted in mythology. Unfortunately, there’s not yet a well-accepted scheme for naming extrasolar planets. Astronomers still generally refer to extrasolar planets by the star they orbit, such as “the planet orbit- ing the star named... .” Worse still, the stars themselves often have confusing or even multiple names, reflecting naming schemes used in star catalogs made by different people at different times in history. A few hundred of the brightest stars in the sky carry names from ancient times. Many of these names are Arabic—such as Betelgeuse, Algol, and Aldebaran—because of the work of the Arabic scholars of the Middle Ages [Section 3.2]. In the early 1600s, German astronomer Johann Bayer developed a system that gave names to many more stars: Each star gets a name based on its constellation and a Greek letter indicating its ranking in bright- ness within that constellation. For example, the brightest star in the constellation Andromeda is called Alpha Andromedae, the second brightest is Beta Andromedae, and so on. Bayer’s system worked for only the 24 brightest stars in each constellation, be- cause there are only 24 letters in the Greek alphabet. About a cen- tury later, English astronomer John Flamsteed published a more extensive star catalog in which he used numbers once the Greek letters were exhausted. For example, 51 Pegasi gets its name from Flamsteed’s catalog. (Flamsteed’s numbers are based on position within a constellation rather than brightness.)
As more powerful telescopes made it possible to discover more and fainter stars, astronomers developed many new star catalogs. The names we use today usually come from one of these catalogs. For example, the star HD209458 appears as star number 209458 in a catalog compiled by Henry Draper (HD). You may also see stars with numbers preceded by other catalog names, including Gliese, Ross, and Wolf; these catalogs are also named for the as- tronomers who compiled them. Moreover, because the same star is often listed in several catalogs, a single star can have several different names. Some of the newest planets orbit stars so faint they have not been previously cataloged. They then carry the name of the observing program that discovered them, such as TrES- for the first discovery of Trans-Atlantic Exoplanet Survey, or OGLE-TR-132b for the planet orbiting the 132nd object scruti- nized by the Optical Gravitational Lensing Experiment. Objects orbiting other stars usually carry the star name plus a letter denoting their order of discovery around that star. If the second object is another star, a capital B is added to the star name, but a lowercase b is added if it’s a planet. For example, HD209458b is the first planet discovered to be orbiting star number 209458 in the Henry Draper catalog; Upsilon Andromedae d is the third planet discovered to be orbiting the twentieth brightest star (be- cause upsilon is the twentieth letter in the Greek alphabet) in the constellation Andromeda. Many astronomers hope soon to devise a better naming system for these wonderful new worlds.
S P E C I A L T O P I C The Names of Extrasolar Planets
400 p a r t I I I • Learning from Other Worlds
This graph shows the Doppler curve for a Sun-like star with a 0.8 Mj planet in a nearly circular orbit at an orbital distance of 1.67 AU.
For a more massive planet in a similar orbit, we observe a larger Doppler shift with the same planet.
For a planet in a more eccentric orbit, we observe an asymmetric Doppler curve.
For a planet with a similar mass in a closer orbit, we observe a larger Doppler shift with a shorter period.
time (days)
0 250 500 750 1000 1250 1500 time (days)
0 250 500 750 1000 1250 1500
150
100
star’s velocity (m/s)
50
0
50
150
100
150
100
star’s velocity (m/s)
50
0
50
150
100
HD 10697b
HD 187085b
HD 4208b
HD 27442b
Figure 13.5 Sample data showing how measurements of Doppler shifts allow us to learn about extrasolar planets for different types of orbits. The points are data for actual planets whose properties are listed in Appendix E.4. Notice that the data points are repeated for additional cycles to show the patterns.
considerably smaller gravitational tug on their stars than the planet orbiting 51 Pegasi. Moreover, by carefully ana- lyzing Doppler shift data, we can learn about the planet’s orbital characteristics and mass. After all, it is the mass of the planet that causes the star to move around the system’s center of mass, so for a given orbital distance, a more mas- sive planet will cause faster stellar motion. We can derive an extrasolar planet’s orbital distance using Newton’s version of Kepler’s third law [Section 4.4]. Recall that for a small object like a planet orbiting a much more massive object like a star, this law expresses a rela- tionship between the star’s mass, the planet’s orbital period, and the planet’s average distance (semimajor axis). We gen- erally know the masses of the stars with extrasolar planets (through methods we’ll discuss in Chapter 15), and the Doppler data tell us the orbital period, so we can calculate orbital distance. We determine orbital shape from the shape of the Dop- pler data curve. A planet with a perfectly circular orbit travels at a constant speed around its star, so its data curve would be perfectly symmetric. Any asymmetry in the Dop- pler curve tells us that the planet is moving with varying speed and therefore must have a more eccentric (“stretched out”) elliptical orbit. Figure 13.5 shows four examples of Doppler data for extrasolar planets and what we learn in each case.
Study the four velocity data curves in Figure 13.5. How would each be different if the planet were: (a) closer to its star? (b) more massive? Explain.
In some cases, existing Doppler data are good enough to tell us whether the star has more than one planet. Re- member that if two or more planets exert a noticeable gravitational tug on their star, the Doppler data will show the combined effect of these tugs. In 1999, such analysis was used to infer the existence of three planets around the star Upsilon Andromedae, making this the first bona fide, multiple-planet solar system known beyond our own. By 2007, at least 20 other multiple-planet systems had been identified. The Doppler technique also tells us about planetary masses, though with an important caveat. Remember that Doppler shifts reveal only the part of a star’s motion di- rected toward or away from us (see Figure 5.23). As a re- sult, a planet whose orbit we view face-on does not cause a Doppler shift in the spectrum of its star, making it im- possible to detect the planet with the Doppler technique (Figure 13.6a). We can observe Doppler shifts in a star’s spectrum only if it has a planet orbiting at some angle other than face-on (Figure 13.6b), and the Doppler shift
T H I N K A B O U T I T
c h a p t e r 1 3 • Other Planetary Systems 401
Not to scale!
Not to scale!
center of mass
center of mass
b We can detect a Doppler shift only if some part of the orbital velocity is directed toward or away from us. The more an orbit is tilted toward edge-on, the greater the shift we observe.
a If we view a planetary orbit face-on, we will not detect any Doppler shift at all.
We view this planet’s orbit face-on, so it has no velocity toward or away from us...
We view the orbit of this planet and star at an angle, so part of the star’s motion is toward us on one side of the orbit, creating a blueshift...
... and part of the star’s motion is away from ... therefore the star also lacks any motion toward or away us on the other side, creating a redshift. from us, which means we detect no Doppler shift.
Figure 13.6 The amount of Doppler shift we observe in a star’s spectrum depends on the orientation of the planetary orbit that causes the shift.
The Doppler technique directly tells us a planet’s orbital period. We can then use this period to determine the planet’s orbital dis- tance. If the planet were orbiting a star of exactly the same mass as the Sun, we could find the distance by applying Kepler’s third law in its simplest form: In fact, many of the planets dis- covered to date do orbit Sun-like stars, so this law gives a good first estimate of orbital distance. For more precise work, we use New- ton’s version of Kepler’s third law (see Mathematical Insight 4.3), which reads:
In the case of a planet orbiting a star, p is the planet’s orbital pe- riod, a is its average orbital distance (semimajor axis), and and are the masses of the star and planet, respectively. ( G is the gravitational constant; ) Because a star is so much more massive than a planet, the sum is pretty much just that is, we can neglect the mass of the planet compared to the star. With this approxima- tion, we can rearrange the equation to find the orbital distance a :
We will discuss how we determine stellar masses in Chapter 15; for now, we will assume the stellar masses are known so that we can calculate orbital distances of the planets.
Example: Doppler measurements show that the planet orbit- ing 51 Pegasi has an orbital period of 4.23 days; the star’s mass is 1.06 times that of our Sun. What is the planet’s orbital distance?
Solution:
Step 1 Understand: We are given both the planet’s orbital period and the star’s mass, so we can use Newton’s version
a L (^) A^3 GM star 4 p^2
p planet^2
M star + M planet M star ;
G = 6.67 * 10 -^11 m^3 / 1 kg * s 22.
M 2
M 1
p^2 =
4 p^2 G 1 M 1 + M 22 a^3
p^2 = a^3.
of Kepler’s third law to find the planet’s orbital distance. How- ever, to make the units consistent, we need to convert the given stellar mass to kilograms and the given orbital period to seconds; we look up the fact that the Sun’s mass is about
Step 2 Solve: We use these values of the period and mass to find the orbital distance a :
Step 3 Explain: We’ve found that the planet orbits its star at a distance of 7.8 billion meters. It’s much easier to interpret this number if we state it as 7.8 million kilometers or, better yet, convert it to astronomical units, remembering that 1 AU is about 150 million kilometers or
We now see that the planet’s orbital distance is only 0.052 AU— small even compared to that of Mercury, which orbits the Sun at 0.39 AU. In fact, comparing the planet’s 7.8-million-kilometer distance to the size of the star itself (presumably close to the 700,000-kilometer radius of our Sun), we estimate that the planet orbits its star at a distance only a little more than 10 times the star’s radius.
a = 7.81 * 109 m * 1 AU 1.50 * 1011 m
= 0.052 AU
1.50 * 1011 meters:
= 7.81 * 109 m
= 3 F
6.67 * 10 -^11 m
3 kg * s 2 *^ 2.12^ *^10
(^30) kg
4 * p^2
1 3.65 * 105 s 22
a L (^) A^3 GM star 4 p^2
p planet^2
p = 4.23 day *
24 hr 1 day
3600 s 1 hr = 3.65 * 105 s
M star = 1.06 * M Sun = 1.06 * 12 * 1030 kg 2 = 2.12 * 1030 kg
2 * 1030 kilograms.
M AT H E M AT I C A L I N S I G H T 13.1 Finding Orbital Distances for Extrasolar Planets
c h a p t e r 1 3 • Other Planetary Systems 403
of a planet with an orbit like that of Jupiter. The Doppler technique also presents a practical difficulty: The extremely precise radial velocity measurements require a relatively large telescope, so only a relatively small number of stars can be studied with this technique. The limitations of the Doppler technique explain what may at first seem like surprising facts: Most of the extra- solar planets discovered to date orbit closer to their stars than similar planets in our solar system, and we have yet to discover any planets at all with Earth-like masses. Both these facts may simply be selection effects of the Doppler technique; that is, the technique tends to find (or “select”) massive planets in close orbits much more easily than any other type of planet. Planets with masses similar to Earth would have such weak gravitational effects on their stars that we could not use the Doppler technique to find them with current technology, while planets orbiting far from their stars have such long orbital periods that it might take decades of observations to detect them. Thus, the current lack of evidence for Earth-mass planets or jovian-mass planets in distant orbits does not necessarily mean that such planets are rare.
Transits and Eclipses A third indirect way of detecting distant planets does not require the observation of gravita- tional tugs at all. Instead, it relies on searching for slight changes in a star’s brightness caused by a planet passing in front of it or behind it. If we were to examine a large sample of stars with plan- ets, a small number of them—typically one in several hun- dred—will by chance be aligned in such a way that one or more of its planets will pass directly between us and the star once each orbit. The result is a transit , in which the planet appears to move across the face of the star. We occasionally witness this effect in our own solar system when Mercury or Venus crosses in front of the Sun (see Figure S1.5). Other star systems are so far away that we cannot actually see a
planetary dot set against the face of the star as we can for Mercury or Venus set against the face of the Sun. Neverthe- less, a transiting planet will block a little of its star’s light, allowing us not only to detect the planet’s existence but also to calculate the planet’s size in comparison to that of its star (see Mathematical Insight 13.3). Because we usually know the star’s size (through methods we’ll discuss in Chapter 15), transit observations allow us to determine planetary sizes. Detecting planets through transits requires many re- peated observations, because most stars exhibit intrinsic variations in brightness. To be confident that an orbiting planet is responsible for a dip in brightness rather than variability in the star itself, we need to see at least several occurrences of the telltale pattern of dimming that occurs during a transit. If this repeated dimming occurs with a regular period, then it is very likely telling us the orbital period of a transiting planet. We can then calculate the planet’s orbital distance and mass.
Which of the following types of planet is most likely to cause a transit across its star that we could observe from Earth? (a) a large planet close to its star; (b) a large planet far from its star; (c) a small planet close to its star; or (d) a small planet far from its star. Explain.
The first success of the transit method came during fol- low-up studies of a planet that had already been discovered with the Doppler technique. The planet, which orbits a star called HD209458, was already known to complete an orbit every days. Thus, when astronomers observed the star to undergo dips in brightness every days (and at just the times that the Doppler measurements said the planet would be moving across our line of sight), they realized they were observing repeated transits by the planet (Figure 13.7). This discovery greatly advanced our understanding of the
T H I N K A B O U T I T
1
4
star
planet (^2 )
6 5
–2.0 0.0 2. time (hours)
relative brightness
(visible light)
1 2 3
–2.0 0.0 2. time (hours)
relative brightness
(infrared light)
6 (^45)
When the planet passes behind the star, we say it is eclipsed by the star.
We observe a transit when the planet passes in front of the star.
Figure 13.7 The artist’s conception (center) shows the planet orbiting the star HD209458. The graphs show how the star’s brightness changes during transits and eclipses, which each occur once with every -day orbit. During a transit, the star’s brightness drops for about 2 hours by 1.7%, which tells us how the planet’s radius compares to the radius of its star. During an eclipse, the infrared signal drops by 0.25%, which tells us about the planet’s thermal emission.
(^3 )
404 p a r t I I I • Learning from Other Worlds
planet itself. First, because the transits revealed that we view the planet’s orbit edge-on, it told us that the mass derived from the Doppler technique was the planet’s true mass and not just a minimum mass. Second, the amount of the dip in the star’s brightness allowed us to determine the planet’s radius and hence its volume. We were thereby able to learn the planet’s density (by dividing its mass by its volume). Because the radius and density match the values expected for a jovian planet, we had acquired the first strong evidence that a planet in another star system really does resemble the jovian planets of our own solar system. Transits can also tell us about the composition of a planet’s upper atmosphere or exosphere (Figure 13.8). To see how, consider what happens if a planet like Jupiter passes in front of its star. From a distance, Jupiter looks like a solid disk extending out as far as its cloud tops. During the tran- sit, this “solid” disk blocks all the starlight coming from directly behind it. Now, suppose the Jupiter-like planet also has a low-density upper atmosphere extending above its cloud tops. The gas in this upper atmosphere would absorb starlight at specific wavelengths that depend on its compo- sition. For example, if the planet’s upper atmosphere con- tained sodium gas, the star’s spectrum would show stronger sodium absorption lines during the transit than at other times. We’d thereby learn that the planet contains sodium in its upper atmosphere. Planets that pass in front of their stars during a transit can also pass behind their stars, in which case the star blocks the light from the planet. Such an event is called an eclipse (see Figure 13.7), because the star blots out the light from the planet in much the same way that the Moon can blot out the light of our Sun during a total solar eclipse [Section 2.3]. Observing an eclipse is much like observing a transit. In both cases, we measure the total light from the star and planet, searching for a small dip in brightness. Be- cause the star is so much brighter than the planet, the dip
in brightness is smaller during an eclipse than a transit. To maximize the effect of the eclipse, astronomers therefore observe at infrared wavelengths, because planets emit most of their own radiation in the form of infrared thermal emission while stars emit more at visible wavelengths. Even so, the total drop in light was only 0.25% when the Spitzer Space Telescope observed an eclipse of the planet orbiting the star HD209458. This small change allowed astronomers to calculate the planet’s total amount of thermal emission and confirm that the planet’s temperature is over 1,100 K. In fact, much as for planets in our own solar system, the infrared brightness can be used to determine how much starlight the planet reflects or absorbs (see Mathematical Insight 10.1), and the crude infrared spectrum can be used to identify gases in the planet’s atmosphere (see Figure 5.14). The first successful observations of this type were announced in 2005, and it is likely that this technique will be improved and applied to more planets by the time you are reading this book. The method of observing extrasolar planets through transits and eclipses has some unusual strengths and weak- nesses. The most obvious weakness is that it can work only for the small fraction of planetary systems whose orbits are oriented edge-on to Earth. A second weakness is that the method is biased in favor of planets with short orbital peri- ods—and hence with orbits close to their stars—both be- cause these planets transit more frequently and because we must observe repeated transits before we can be confident of a discovery. Counterbalancing these weaknesses is a very important strength: With sufficiently precise measurements of stellar brightness, the transit method ought to be able to reveal planets far smaller than is currently possible with the astrometric or Doppler techniques. Indeed, as we’ll dis- cuss in Section 13.4, NASA is currently preparing a mission (called Kepler ) specifically intended to search for transits by Earth-size planets.
wavelength
100%
wavelength wavelength
relative intensity
100% light blocked by planet
relative intensity
Absorption line depths are exaggerated for clarity.
This graph shows the sodium lines in the star’s spectrum.
This planet orbits in 3.5 days. When it is not in front of its star, we see only the spectrum of the star itself.
The star appears 1.7% dimmer when the planet passes in front of it, so we can figure out the planet’s size compared to the star.
The planet’s upper atmosphere absorbs additional light at wavelengths that depend on its composition.
The sodium lines are deeper when the planet is in front of the star, telling us that the planet’s atmosphere contains sodium.
Figure 13.8 This diagram shows how transit observations can give us information about the composition of an extrasolar planet’s extended upper atmosphere or exosphere.
406 p a r t I I I • Learning from Other Worlds
Jupiter
Sun
center of mass
Not to scale!
Jupiter half an orbit later
Sun half an orbit later
(^1) Gravitational Tugs: We can detect a planet by observing the small orbital motion of its star as both the star and its planet orbit their mutual center of mass. The star’s orbital period is the same as that of its planet, and the star’s orbital speed depends on the planet’s distance and mass. Any additional planets around the star will produce additional features in the star’s orbital motion.
1a (^) The Doppler Technique: As a star moves alternately toward and away from us around the center of mass, we can detect its motion by observing alternating Doppler shifts in the star’s spectrum: a blueshift as the star approaches and a redshift as it recedes. This technique has revealed the vast majority of known extrasolar planets.
1b (^) The Astrometric Technique: A star’s orbit around the center of mass leads to tiny changes in the star’s position in the sky. As we improve our ability to measure these tiny changes, we should discover many more extrasolar planets.
The search for planets around other stars is one of the fastest growing and most exciting areas of astronomy. Although it has been only a little more than a decade since the first discoveries, known extrasolar planets already number well above 250. This figure summarizes major techniques that astronomers use to search for and study extrasolar planets.
Jupiter actually orbits the center of mass every 12 years, but appears to orbit the Sun because the center of mass is so close to the Sun. The Sun also orbits the center of mass every 12 years.
Current Doppler-shift measurements can detect an orbital velocity as small as 1 meter per second—walking speed.
The change in the Sun’s apparent position, if seen from a distance of 10 light years, would be similar to the angular width of a human hair at a distance of 5 kilometers.
(^201520051980)
2010
2020
1975 1965 2000
2025
1990
1995
1960
1970
1985
center of mass
0.0005 arcsecond radius of Sun from 30 light years
or
bi to
fu
ns
ee npl anet
to Earth
starlight redshifted
starlight blueshifted
stellar motion caused by tug of planet
COSMIC CONTEXT FIGURE 13.10 Detecting Extrasolar Planets
c h a p t e r 1 3 • Other Planetary Systems 407
IRIR
(^2) Transits and Eclipses: If a planet’s orbital plane happens to lie along our line of sight, the planet will transit in front of its star once each orbit, while being eclipsed behind its star half an orbit later. The amount of starlight blocked by the transiting planet can tell us the planet’s size, and changes in the spectrum can tell us about the planet’s atmosphere.
(^3) Direct Detection: In principle, the best way to learn about an extrasolar planet is to observe directly either the visible starlight it reflects or the infrared light that it emits. Our technology is only beginning to reach the point where direct detection is possible, but someday we will be able to study both images and spectra of distant planets.
This infrared image shows a brown dwarf called 2M1207 (blue)...
... and what is probably a jovian planet (red) in orbit around it.
Artist’s conception of another planetary system, viewed near a ringed jovian planet.
When the planet passes behind the star, we say it is eclipsed by the star.
We observe a transit when the planet passes in front of the star.
1
4
star
planet (^2 )
6 5
–2.0 0.0 2. time (hours)
relative brightness
(visible light)
1 2 3
–2.0 0.0 2. time (hours)
relative brightness
(infrared light)
(^56)
4
c h a p t e r 1 3 • Other Planetary Systems 409
produce transits and for which we also have mass data from the astrometric or Doppler techniques.
● Composition: We learn composition from spectra. Tran- sits can provide limited information about the compo- sition of a planet’s upper atmosphere if the star shows absorption during the transit that is not present at other times. Eclipses can also provide limited spectral infor- mation. More detailed information about composition requires spectra from direct detections.
We are now ready to see what we’ve learned about the extra- solar planets discovered to date. (See Appendix E.4 for de- tailed data.)
Orbits Much as Johannes Kepler first appreciated the true layout of our own solar system [Section 3.3], we can now step back and see the layout of many other solar sys- tems. Figure 13.11a shows the orbits of known extrasolar planets superimposed on each other, as if they were all or- biting a single star; the dots indicate the minimum masses of the planets found through the Doppler technique. Despite the crowding of the orbits when viewed this way, at least two important facts should jump out at you. First, notice that only a handful of these planets have orbits that take them beyond about 5 AU, which is Jupiter’s dis- tance from our Sun. Most of the planets orbit very close to their host star. Second, notice that many of the orbits are clearly elliptical, rather than nearly circular like the orbits
of planets in our own solar system. These facts are even easier to see if we display the same information on a graph (Figure 13.11b). Look first at the green squares represent- ing the planets in our own solar system; notice that they are located at the distances you should expect and all but Mercury have very small eccentricity, meaning nearly cir- cular orbits. Now look at the red dots representing extra- solar planets. Quite a few of these planets orbit their stars more closely than Mercury orbits the Sun, and none are located as far from their stars as the jovian planets of our solar system. Many also have large orbital eccentricities, telling us that their elliptical orbits have very stretched-out shapes. As we’ll see shortly, both facts provide important clues about the nature of these extrasolar planets.
Should we be surprised that we haven’t found many planets orbiting as far from their stars as Saturn, Uranus, and Neptune orbit the Sun? Why or why not?
At least 20 stars have so far been found to contain two or more planets, and one system has four known planets (Figure 13.12). This is not surprising, since our own solar system and our understanding of planet formation suggest that any star with planets is likely to have multiple planets. We will probably find many more multiple-planet systems as observations improve. However, one fact about these
T H I N K A B O U T I T
8 6 4 2 0 2 4 6 8
distance from central star (AU)
distance from central star (AU)
8 6 4 2 0 2 4 6 8
10 M J
1 M J–10 M J
0.1 M J–1 M J
<0.1 M J
average orbital distance (AU)
orbital eccentricity
0.01 0.
Mercury
Venus Earth
Mars Jupiter
Jupiter’s orbit
Saturn Neptune
Uranus
1.00 10.
Orbits of Extrasolar Planets Orbital Properties of Extrasolar Planets
a This diagram shows all the orbits superimposed on each other, as if all the planets orbited a single star. Dots are located at the aphelion (farthest) point for each orbit, and their sizes indicate minimum masses of the planets.
b The data from (a) are shown here as a graph. Dots closer to the left represent planets that orbit closer to their stars, and dots lower down represent smaller orbital eccentricities. Green dots are planets of our own solar system.
Most known extrasolar planets orbit their stars much closer than Jupiter orbits our Sun.
Many extrasolar planet orbits are eccentric...
...compared to planets in our solar system.
Figure 13.11 Orbital properties of 164 extrasolar planets with known masses, distances, and eccentricities.
410 p a r t I I I • Learning from Other Worlds
2:
2: 2:
2:
3:
Sun
Mercury Venus Earth Mars Jupiter
HD
HD
HD
HD
HD
HD
HD
HD
HD
HIP14810HIP
HD
HD
HD
Ups And 55Cnc Gliese 876
Gliese 581
HD
orbital period (days)
1 10 100 1000
10 M J
1 M J–10 M J
0.1 M J–1 M J
orbital resonance
<0.1 M J
Figure 13.12 This diagram shows the orbital distances and ap- proximate masses of the planets in the first 20 multiple-planet sys- tems discovered. The four highlighted systems are the ones with the best data to show planets in orbital resonances. For example, the “3:1” for 55 Cnc indicates that the inner of the two indicated planets completes exactly three orbits while the outer planet com- pletes two orbits. When reading orbital periods from the graph, be sure to notice that the axis is exponential, so that each tick mark represents a period 10 times longer than the previous one.
40
20
0
60
80
100
mass of planet compared to Jupiter (units of Jupiter masses)
number of known extrasolar planets 0.001 0.01 0.1 1 10 100
Most planets discovered so far are more massive than Jupiter.
No planets with mass as small as Earth’s have been discovered.
Figure 13.13 This bar chart shows the number of planets in different mass categories for 196 extrasolar planets with known minimum masses. Notice that the axis uses an exponential scale so that the wide range of masses can all fit on the graph.
other planetary systems is surprising: Already, we’ve found at least five systems that seem to have planets in orbital res- onances [Section 11.2] with each other. For example, four systems have one planet that orbits in exactly half the time as another planet. In our own solar system we’ve seen the importance of orbital resonances in sculpting planetary rings, stirring up the asteroid belt, and even affecting the orbits of Jupiter’s moons. As we’ll discuss shortly, orbital resonances may also have profound influences on extrasolar planets.
Masses Look again at Figure 13.11. The sizes of the dots indicate the approximate minimum masses of these plan- ets; they are lower limits because nearly all have been found with the Doppler technique. The mass data are easy to see if we display them as a bar chart (Figure 13.13). We see clearly that the planets we’ve detected in other star systems are generally quite massive. Most are more massive than Jupiter, and only a few are less massive than Uranus and Neptune. The smallest detected as of late 2007 is five times as massive as Earth (which has a mass of about 0.003 Ju- piter mass). If we go by mass alone, it seems likely that most of the known extrasolar planets are jovian rather than ter- restrial in nature. Should we be surprised by the scarcity of planets with masses like those expected for terrestrial worlds? Not yet. Remember that nearly all these planets have been found indirectly by looking for the gravitational tugs they exert on their stars. Massive planets exert much greater gravita- tional tugs and are therefore much easier to detect. As we
discussed earlier, the abundance of very massive planets among these early discoveries is probably a “selection effect” arising because current planet-finding techniques detect massive planets far more easily than lower-mass planets.
Sizes and Densities The masses of the known extrasolar planets suggest they are jovian in nature, but mass alone cannot rule out the possibility of “supersize” terrestrial planets—that is, very massive planets made of metal or rock. To check whether the planets are jovian in nature, we also need to know their sizes, from which we can calculate their densities. If their sizes and densities are consistent with those of the jovian planets in our solar system, then we’d have good reason to think that they really are jovian in nature. Unfortunately, we lack size data for the vast majority of known extrasolar planets, because they have only been de- tected by the Doppler technique and their orbits are not oriented to produce transits. However, in the relatively few cases for which we have size data from transits, the sizes and densities do indeed seem consistent with what we expect for jovian planets. The first planet for which we obtained size data (the planet orbiting the star HD209458) has a mass of about 0.63 Jupiter mass and a radius of about 1.43 Jupiter radii. In other words, it is somewhat larger than Jupiter in size but smaller than Jupiter in mass, with an average density about that of Jupiter. This low density is only partially surprising. The planet orbits very close to its star, with an average orbital distance of 0.047 AU (only 12% of Mercury’s distance from the Sun), so we expect the planet to be quite hot. The high temperature should cause its atmosphere to be “puffed up,” giving the planet a lower average density than Jupiter, although we cannot yet explain why its den- sity is quite as low as it is. We currently have size and density information for several other extrasolar planets (see Ap- pendix E.4), and in all but one case they are also larger and
1 4
412 p a r t I I I • Learning from Other Worlds
Composed primarily of hydrogen and helium 5 AU from the Sun Orbit takes 12 Earth years Cloud top temperatures 130 K Clouds of various hydrogen compounds Radius 1 Jupiter radius Mass 1 Jupiter mass Average density 1.33 g/cm^3 Moons, rings, magnetosphere
Composed primarily of hydrogen and helium As close as 0.03 AU to their stars Orbit as short as 1.2 Earth days Cloud top temperatures up to 1,300 K Clouds of “rock dust” Radius up to 1.3 Jupiter radii Mass from 0.2 to 2 Jupiter masses Average density as low as 0.2 g/cm^3 Moons, rings, magnetospheres: unknown
Jupiter “Hot Jupiters” orbiting other stars
Figure 13.14 A summary of the expected similarities and differences between the real Jupiter and extrasolar “hot Jupiters” orbiting Sun-like stars.
summarizes the similarities and differences expected be- tween the real Jupiter and “hot Jupiters.” Although we have not yet (as of late 2007) detected any planets as small in mass as Earth, we have found a few planets with masses only a few times that of Earth. While we have no information about the radii or composition of these planets, their masses are much too low for them to be Jupiter-like. They could potentially be small versions of planets like Uranus or Neptune, or alternatively like large versions of Earth (making them what some people call “super-Earths”). Either way, water is probably a major in- gredient of these planets. The planet called Gliese 581c, detected in mid-2007, is particularly interesting. This planet orbits only about 0. AU from its star. However, because its star is much cooler and dimmer than our Sun, this orbit places the planet within the habitable zone —the zone of distances from a star in which temperatures should allow for the existence of liquid water on a planet’s surface [Section 24.3]. Thus, if Gliese 581c has water on its surface, it could well prove to be the first world besides Earth known to have surface oceans of liquid water.
13.3 The Formation of
Other Solar Systems
The discovery of extrasolar planets presents us with an op- portunity to test our theory of solar system formation. Can
our existing theory explain other planetary systems, or will we have to go back to the drawing board? As we discussed in Chapter 8, the nebular theory holds that our solar system’s planets formed as a natural conse- quence of processes that accompanied the formation of our Sun. If the theory is correct, then the same processes should accompany the births of other stars, so the nebular theory clearly predicts the existence of other planetary systems. In that sense, the recent discoveries of extrasolar planets mean the theory has passed a major test, because its most basic prediction has been verified. Some details of the theory also seem supported. For example, the nebular the- ory says that planet formation begins with condensation of solid particles of rock and ice (see Figure 8.13), which then accrete to larger sizes. We therefore expect that planets should form more easily in a nebula with a higher propor- tion of rock and ice, and in fact more planets have been found around stars richer in the elements that make these ingredients. Nevertheless, extrasolar planets have already presented at least one significant challenge to our theory. According to the nebular theory, jovian planets form as gravity pulls in gas around large, icy planetesimals that accrete in a spin- ning disk of material around a young star. The theory there- fore predicts that jovian planets should form only in the cold outer regions of star systems (because it must be cold for ice to condense), and that these planets should be born with nearly circular orbits (matching the orderly, circular motion of the spinning disk). Massive extrasolar planets
c h a p t e r 1 3 • Other Planetary Systems 413
with close-in or highly elliptical orbits present a direct chal- lenge to these ideas.
◗ Can we explain the
surprising orbits of many
extrasolar planets?
The nature of science demands that we question the valid- ity of a theory whenever it is challenged by any observation or experiment [Section 3.4]. If the theory cannot explain the new observations, then we must revise or discard it. The surprising orbits of many known extrasolar planets have indeed caused scientists to reexamine the nebular theory of solar system formation. Questioning began almost immediately upon the dis- covery of the first extrasolar planets. The close-in orbits of these massive planets made scientists wonder whether some- thing might be fundamentally wrong with the nebular the- ory. For example, is it possible for jovian planets to form very close to a star? Astronomers addressed this question by studying many possible models of planet formation and reexamining the entire basis of the nebular theory. Several years of such reexamination did not turn up any good rea- sons to discard the basic theory. While it’s still possible that a major flaw has gone undetected, it seems more likely that the basic outline of the nebular theory is correct. Scientists therefore suspect that extrasolar jovian planets were indeed born with circular orbits far from their stars, and that those that now have close-in or eccentric orbits underwent some sort of “planetary migration” or suffered gravitational in- teractions with other massive objects.
Planetary Migration If the hot Jupiters formed in the outer regions of their star systems and then migrated in- ward, how did these planetary migrations occur? You might think that drag within the solar nebula could cause planets to migrate, much as atmospheric drag can cause satellites in low-Earth orbit to lose orbital energy and eventually plunge into the atmosphere; however, calculations show this drag effect to be negligible. A more likely scenario is that waves propagating through a gaseous disk lead to migra- tion (Figure 13.15). The gravity of a planet moving through a disk can create waves that propagate through the disk, causing material to bunch up as the waves pass by. This “bunched up” matter (in the wave peaks) then exerts a gravi- tational pull back on the planet that reduces its orbital en- ergy, causing the planet to migrate inward toward its star. Computer models confirm that waves in a nebula can cause young planets to spiral slowly toward their star. In our own solar system, this migration is not thought to have played a significant role because the solar wind cleared out the gas before it could have much effect. But planets may form earlier in some other solar systems, allowing time for jovian planets to migrate substantially inward. In a few cases, the planets may form so early that they end up spiraling
into their stars. Indeed, astronomers have noted that some stars have an unusual assortment of elements in their outer layers, suggesting that they may have swallowed planets (including the migrating jovian planets themselves and possibly terrestrial planets shepherded inward along with the jovian planets).
Encounters and Resonances Migration may explain the close-in orbits, but why do so many extrasolar planets have highly eccentric orbits? One hypothesis links both migration and eccentric orbits to close gravitational en- counters [Section 4.5] between young jovian planets form- ing in the outer regions of a disk. A close gravitational encounter between two massive planets can send one planet out of the star system entirely while the other is flung inward into a highly elliptical orbit. Alternatively, a jovian planet could migrate inward as a result of multiple close encounters with much smaller planetesimals. Astron- omers suspect that this type of migration affected the jo- vian planets in our own solar system. Recall that the Oort cloud is thought to consist of comets that were ejected outward by gravitational encounters with the jovian plan- ets [Section 12.2]. In that case, the law of conservation of energy demands that the jovian planets must have mi- grated inward, losing the same amount of orbital energy that the comets gained. Gravitational interactions can also affect orbits through resonances. Recall that Jupiter’s moons Io, Europa, and Ganymede share orbital resonances that cause their orbits to be more elliptical than they would be otherwise (see Figure 11.20b). Models show that similar resonances be- tween massive jovian planets could make their orbits more eccentric, explaining why some extrasolar planets that orbit at Jupiter-like distances have surprisingly high eccentrici- ties. Other kinds of resonances could lead to planetary mi- gration or ejection of a planet from its system altogether. Alternatively, planetary migration might force two planets
The orbiting planet nudges particles in the disk...
... causing material to bunch up. These dense regions in turn tug on the planet, causing it to migrate inward.
Figure 13.15 This figure shows a simulation of waves created by a planet embedded in a dusty disk of material surrounding its star.
