SEATTLE—The most exhilarating science conference I've ever been to took place in San Antonio 15 years ago this week, when planet hunters Geoff Marcy and Paul Butler announced they had found two planets orbiting sunlike stars beyond our solar system. Coming a couple of months after another team, led by Michel Mayor and Didier Queloz, made the first such discovery, Marcy and Butler's presentation was a veritable Woodstock of astronomy, such was the crush of scientists trying to get into the room. I'll never forget the chill that went up my spine when they said one of the planets orbited within the habitable zone of its star, the range of distances where liquid water could be stable.
The excitement has scarcely dissipated since then. This week Marcy, now at the University of California, Berkeley, spoke again at the American Astronomical Society's winter meeting here to take stock. He proceeded to demolish the prevailing theoretical models for how planets form. Observers in any field of science take a peculiar pleasure in seeing their theorist colleagues collapse into sobbing heaps, but it happens with unnerving regularity with exoplanets. Modelers have consistently failed to predict the diversity of planetary systems out there. And they are the first to admit it. "These models are crap," says Hal Levison of the Southwest Research Institute in Boulder, Colo. "They may be the best we can do, but they are still crap."
Most of the hundreds of exoplanets that astronomers have found are Jupiter-size, but a growing number look tantalizingly like Earth. At this week's meeting came a milestone with the announcement of Kepler-10b [see artist's conception above]. Not only is it the smallest world yet discovered around a sunlike star—1.4 Earth-radii and 4.6 Earth-masses—it has a high density that matches an Earthly composition of rock and iron. NASA's Kepler space observatory discovered it by seeing a periodic dimming of the star's light, which suggested that a planet was crossing in front, and the Keck Observatory on Mauna Kea in Hawaii confirmed that the star is wobbling slightly in response to a planet's gravity. (Even smaller planets orbit a pulsar, a very un-sunlike star, but they get less attention because astronomers are seeking closer cousins to our own solar system.)
Kepler has also found a multiple-planet system around the star Kepler-9. What makes it so cool is that the timing of the dimming varies slightly, confirming that the planets are gravitationally perturbing each other—the first time the dance of the planets has been observed in all its glory. The variation provides a reality check on the existence of planets and an independent way to ascertain their masses. The system has, at least, two Saturn-mass worlds and a 1.6-Earth-radius one.
These discoveries further demonstrate that planetary systems come in all shapes and sizes. Theorists have long since gotten over their surprise that most systems look nothing like our solar system. Still, they argue, systems should have some common features. Marcy went down the list. Planets should go around their stars on nearly circular orbits lying in the equatorial plane of their stars. They should move in the same direction (clockwise or counterclockwise) as the star spins. Over eons, planets can leave their birthplace and migrate to other parts of their system—a process that should clear out any smallish planets from the region immediately around the star. Moreover, to migrate inward, a planet must transfer angular momentum to more distant material; thus, any tightly orbiting planets—including the plentiful breed of massive planets known as hot Jupiters—should be accompanied by more distant ones.
Marcy took down these predictions one by one. Lots of planet orbits are highly elliptical and tilted. As many as a third move in the opposite direction as their star spins, a fact deduced from the way the periodic dimming of starlight brings out the Doppler shift of the spinning star. Perhaps an eighth of sunlike stars have close-in smallish planets. And not a single close-in planet has a more distant companion. Marcy called the success rate of theorists "shocking and a little disturbing."
Theorists dispute some of Marcy's specifics but broadly agree with his critique. "The issue of hot Jupiters without additional companions needs serious attention," says Doug Lin of the University of California, Santa Cruz. "There is no obvious reason for it." The question is whether he and others can fix their basic scenario for the genesis of planets or will have to give it up.
That scenario, known as core accretion, supposes that planets start as small grains of dust that agglomerate to progressively larger sizes. Some sweep up gas and grow into giants. An alternative scenario, gravitational instability, holds that planets start as large clouds of gas that fragment and collapse under their own weight, much as stars are thought to do, but on a smaller scale. It has fallen into disfavor, though, because it would explain gaseous Jupiters but not rocky Earths.
Although Levison bluntly calls core-accretion models "crap", he sees their failure as one of implementation, not of basic principle. Planet formation is complex, and modelers don't have enough computing power to run full-up simulations for all possible permutations of planetary systems. To make statistical predictions, they rely on simplified computer codes, and this, Levison says, leads to the failings that Marcy identifies. "The only lesson to be learned from the fact that the models do not reproduce the observations is that the modelers need to work harder," he says. "I do not think it endangers our basic concepts of planet formation." Lin agrees: "I still believe core accretion scenario is a good paradigm to develop the theory of planet formation. But it needs to become more sophisticated."
Others are less sanguine. Over the past several months, Sergei Nayakshin of the University of Leicester has mashed up the two standard scenarios to create a radical new one. In it, gravitational instability first creates a family of giant planets at large distances from the star, way out past the present-day orbit of Neptune. Within each, solid material settles to the core. They all migrate inward, and the star strips the innermost planets of their outer gaseous layers, reducing them to their rocky cores [see video below]. Accretion rounds them out. In the core-accretion scenario, gas giants are bulked-up rocky words. In Nayakshin's, it’s the other way around: rocky worlds are slimmed-down gas giants.
Nayakshin claims his model can account for the discrepant observations. The early stages of the process are fast and messy and could lead to orbits of all sorts of shapes and tilts. Because planets form far from their stars, any material left over after they migrate need not be enough to give rise to new planets. "Multiple planets aren't needed in hot-Jupiter systems in my model," he says.
That said, Nayaksin's model has yet to receive the same scrutiny as the other scenarios have. When it comes to the universe of planets, nature has a way of confounding scientists' best ideas.