June 22, 2005
Learning From the Search for Elusive Earth Cousins
What have scientists learned in a decade of searching for extrasolar planets? Are there other solar systems just like our own waiting to be discovered, or are our Sun and its contingent of planets in some way unique? In this interview with Astrobiology Magazine, Professor of Astronomy Geoff Marcy, one of the world's leading planet-hunters, reflects on recent other-worldly discoveries and speculates on what surprises may lay in store.
Astrobiology Magazine -- Geoff Marcy, professor of astronomy at the University of California, Berkeley, and director of the Center for integrative Planetary Science, leads a team of planet-hunters credited with the discovery of more than 100 planets that orbit nearby stars.At a recent symposium on extrasolar planets, Marcy spoke with Astrobiology Magazine Managing Editor Henry Bortman about recent discoveries and the likelihood of finding other solar systems like our own.
Astrobiology Magazine (AM): You've found Jupiter- or Neptune-mass planets orbiting about 6.5 percent of the 1300 stars you've been monitoring in your radial-velocity survey. Almost all of these planets are closer to their host stars than Jupiter is to our own Sun. I know you need to see a planet's complete orbit to figure out its mass and its distance from its star. By now, though, I would imagine you have some strong hints about how many of the remaining stars have Jupiters or Saturns farther out from their stars. What kinds of indications are you getting about the prevalence of giant planets?
Geoff Marcy (GM): Something like 5 percent of the remaining stars show a velocity variation, which has to be due to a companion. But even if we see a linear or near-linear velocity increase, followed by a decrease, you still don't know what the orbital period is, because it could go for decades before it comes back up again. You might have caught it near the top, and now you're catching it going over the top, but it could go down for 3 decades or more. So, sadly, we can't really constrain the orbital period or the mass of a planet, or even know whether it is a planet, based on an incomplete orbit.
Most of them probably are planets. They're probably a Jupiter mass, or a little bit more, out at 5 AUs or 10 AUs or 15 AUs, where the orbital period is 10 or 20 or 30 years. So, as we all would have guessed as children, there probably is a population of Jupiters sitting out there at Jupiter and Saturn-like distances, but we just haven't watched them long enough to confirm that.
That's what's happening with many of our planets. We first caught them on the rise, then they hooked over. We kept taking data, and now they've come back around. Once the orbit closes, that's when we publish a paper. So we've learned that, with patience, they all become full orbits after a while.
AM: So when all of these orbits have closed, you estimate that you'll end up with about double the number of giant planets that you have now, that roughly 12 percent of the stars will have giant planets?
GM: Yeah. Right now 6.5 percent of our stars have Jupiters and Saturns. That's a done deal. But if you just mildly extrapolate to these longer-period ones, it's probably 12 percent, out to say 20 AUs. So something like 12 percent of all stars have a Jupiter or a Saturn like our own, that is to say, roughly the same mass, in a solar-system-like orbit. On the other hand, 85 percent of the stars don't have a Jupiter or a Saturn, which I think is interesting to note. Some of the planetary systems clearly don't have giant planets. Maybe they don't have any planets. But we can rule out the giant planets.
AM: What about Neptune-mass planets?
GM: We have just found the first three Neptune-class planets, with minimum masses of 15, 18, and 21 Earth-masses. They orbit 55 Cancri, HD 190360, and Gliese 436. For comparison, our Neptune has 17 Earth-masses. Most remarkably, we also found a planet with a likely mass of only 7 Earth-masses, orbiting the star Gliese 876. This planet is probably rocky, a "super-Earth" with a radius only twice that of our Earth. So we have found the first planets that resemble the terrestrial planets in our Solar Systems, albeit larger.
AM: So you know for certain that some 85 percent of your stars don't have a Jupiter or a Saturn. But you can't tell yet whether most of them have a Neptune or only a handful do?
GM: Empirically you're right. Of the 85 percent that show definitively they don't have a Jupiter or a Saturn, they could have Neptunes, they could have Earths, and so on. But we have no information. If you wanted my guess, any star that's reasonably isolated - single stars, or stars that have a distant stellar companion - almost certainly had a protoplanetary disk around it when it was young, and those disks almost have to make planets. This is a guess, so put a red flag about what I'm about to say. Based on theory, the guess would be that protoplanetary disks did exist around virtually all of the stars that we currently don't see any planets around at all. Those stars probably have Neptunes and Earths. If you had to bet, you'd bet there are Earth-size, Venus-size, Mars-size, maybe even Neptune-size planets around 80 percent of all the stars, 90 percent, maybe even virtually all of them. It's hard to avoid it.
If I may just elaborate, because it's a very exciting issue for astrobiology and the prevalence of Earths in the galaxy: The only way a star can form is by gas accreting onto the star, and it does so conserving angular momentum, creating a disk. And then the viscosity of that disk drains the material onto the star. So almost every star must have had a protoplanetary disk for a few million years. And therefore, it almost certainly must have made Earths. Why would some disks make Earths and Jupiters and others not? So, the odds are that our non-detections, the 85 percent of the stars that haven't yet shown planets still have lower-mass planets. That would be the bet, without any evidence.
AM: What about orbital eccentricity? The planets in our solar system have nearly circular orbits. But that doesn't appear to be true of the planets around other stars. Most of them have eccentric orbits.
GM: To me, that's the most dramatic discovery of all that we've learned about extrasolar planets. And the anthropocentric surprise is that of the 104 extrasolar planets that my team has discovered, 90 percent of them have eccentricities, elongations of the orbits, greater than those seen among the planets in our own solar system. And most of them are much greater: 0.25 is the average eccentricity. By comparison, Jupiter has an eccentricity of 0.05.
AM: What does a 0.25 eccentricity look like?
GM: If you take the average distance of a planet from its star, the eccentricity determines how close in it comes and how far out it goes. A 0.25 eccentricity brings the planet in 25 percent closer than its average, and it swings the planet out 25 percent farther than its average. Most of the comets have eccentricities of 0.8 or 0.9. Halley's comet has eccentricity of 0.97. They're way eccentric.
AM: You said you were studying about 1300 stars. Does that include many M dwarfs?
GM: We have 150 of these low-mass M-dwarf stars on our survey. We've been watching them for 3 or 4 years. Of the 150, only 2 of them have shown planets: Gilese 436, which has a Neptune that we recently discovered; and Gilese 876, which has 3 planets (including the 7 Earth-mass one), with orbital periods of 30, 60, and 1.9 days. Two out of 150 is a low occurrence rate, relative to the Sun-like stars. So the early suggestion - and it's more than a suggestion - is that the low-mass stars, the M dwarfs, have fewer Jupiters and Saturns in orbits comparable to those that we seen in the solar-type stars. Because we would have seen them if the M dwarfs had them.
Now you could argue that maybe M dwarfs have plenty of Jupiters and Saturns, but they orbit so far away that we haven't detected them yet. I suspect not. Much more likely, and I don't think I'm grasping too far - one or two theorists have looked into this - is that low-mass stars formed out of low-mass molecular cloud cores, in turn making low-mass protoplanetary disks. So it really isn't much of a surprise, in retrospect, that low-mass stars would be associated with lower-mass planets, Neptune-mass or lower. One theoretical paper has suggested that the mass of the planets would be roughly proportional to the mass of the host star. So M dwarfs, being a third of a solar mass, might only make Saturns and below.
AM: You said that your guess is that planets are likely to form around most stars. Let me ask you to speculate further. How likely do you think habitable planets are?
GM: That's harder, and of course, I'm asked that a lot. You know, the problem is that the properties of a planet that render it habitable are still under fairly serious debate. There was the book, Rare Earth, that raised a number issues, and there are other issues about habitability that are still being learned. How stable does the rotation axis have to be? The orbit probably has to be circular enough that liquid water would persist for a long time. So what fraction of earthlike planets would be in circular enough orbits that liquid water could persist for billions of years? There's the question of a carbon cycle that allows greenhouse effects to remain stable, and so on.
These are issues that are just not well understood. So I don't think anybody knows. If I were to articulate our lack of knowledge, if earthlike planets form near 1 AU, 50 percent of them retain their liquid water, making oceans and lakes, and they're habitable by the definitions that we suggest. It could also be that it's one in a million. I think it's fair to argue the other end of the spectrum: that retention of liquid water, stably, not freezing or vaporizing, for a few billion years, to let Darwinian evolution do its thing and create creatures that can write cello concertos - that kind of stability could be a rarity.
I don't think anybody can speak intelligently about the fraction of rocky planets at 1 AU that might retain liquid water. In fact, there's even a first-order question, which is, How many rocky planets even have the amount of liquid water that we have on the Earth? Too much and you have a water world. Too little and it gets absorbed into the silicates, hydrated rocks, and you don't have any liquid water on the surface. We have a rather special amount of liquid water on the Earth. It gives us oceans that have enough thermal inertia to remain liquid and maintain stable temperature. So I think we're not in a good position right now to make an educated guess about whether habitable worlds are around 50 percent of all the stars, or 1 in a million. It could be either way, as far as I can tell.
On the Net: