Ghost Worlds

Pawel Artymowicz recently had one of those “fiction is reality” moments. As he was crossing the border into the U.S., an immigration officer asked him what he did for a living. Artymowicz, a U of T astronomy professor, responded that he was a theoretical astrophysicist.

“And what is that?” the officer asked, a little suspiciously.

“Oh,” said Artymowicz, eschewing all technical descriptions of his work. “I study how planets outside our solar system form.”

“Ah, you mean like Class M planets,” said the official, proudly recalling how the writers of Star Trek denoted Earth-like planets in the far reaches of the galaxy.

This accidental conjoining of the research interests of a scientist and the enthusiasm of a science-fiction fan would have been unlikely 20 years ago. At that time, no extra-solar planets of any sort had been discovered for Artymowicz or anyone else to study.

But in 1992, Penn State University astronomer Alexander Wolszczan published evidence of the first planet to be found outside our solar system – a distant, rocky orb circling a pulsar in the constellation of Virgo. Since then, “everything has changed,” says Debra Fischer, an astronomer at San Francisco State University and a recent guest of U of T’s department of astronomy and astrophysics. In October, Fischer delivered a public lecture at Convocation Hall on extra-solar planets as part of the department’s 100th anniversary celebrations. The study of these planets, she says, has grown from an intriguing diversion to one of the hottest fields in astronomy physics. “In the beginning, it was like stamp collecting,” she says. There was a planet here, and a planet there. But now astronomers are starting to compare our own solar system to other planetary systems and are being forced to rethink long-held theories of how stars and planets come into being. U of T researchers are among those trying to integrate these discoveries into a broader picture of stellar and planetary evolution. One day, the search for these distant celestial bodies may yield the Holy Grail of planetary astronomy – the discovery of another Earth.

Sci-fi books and movies would have us believe the galaxy is teeming with hundreds of humanoid civilizations inhabiting planets that look a lot like Earth. But the astronomical evidence to support this view is so far lacking. Almost all of the 150-plus extra-solar planets that scientists have detected are gas giants – hundreds of times bigger than Earth. Many of these immense planets hug their parent stars in tight orbits, completing a full circuit in just a few days. (Even Mercury, the speediest planet in our solar system, requires 88 days to orbit our parent star, the sun.) And while Earth and its siblings travel around the sun in near-circles, many of these newly discovered planets move in highly elliptical orbits. In the jargon of astronomy, they have “high orbital eccentricities.” The surface temperature on these planets is furnace-hot much of the time. Life almost certainly could not develop under these conditions.

Is it possible that other Earth-like planets exist, but have so far escaped our detection? The recent wave of discoveries certainly makes the existence of other Earths likely, says Artymowicz, but scientists don’t know how many smaller, rocky planets will be found in the galaxy. “I don’t think we’re at the point where we can reliably predict the number,” he says. “But there is no physical reason why terrestrial planets shouldn’t be there.”

If there is another Earth out there, astronomers are unlikely to see it just yet because of the techniques they use to detect planets. The radial velocity method, which has been used for several years, is biased toward finding large planets with tight orbits, says U of T astronomer Ray Jayawardhana. Through radial velocity, a scientist can infer the existence of a planet by observing its influence on the light of its parent star. Suppose we’re viewing a far-off star system from its edge, says Jayawardhana. An orbiting planet will spin toward us for part of its year and away from us for a similar amount of time. Its parent star will also move very slightly – tugged by its planet toward us and away from us in a regular cycle. This distinctive wobble causes subtle shifts in the light of the star. By observing the system for several orbital periods with a telescope and a spectrograph (which measures the intensity of light at different wavelengths), astronomers can pin down the distance of the planet from its sun, and estimate the planet’s mass. The radial velocity technique tends to locate large planets in close orbits because these planets cause their parent stars to wobble most. Finding smaller planets or planets moving in wider orbits is more challenging. Still, as astronomers refine the radial velocity method, they believe they’ll be able to spot planets only a few times larger than Earth (they’re already detecting objects the size of Uranus and Neptune, which are about 15 times as massive as Earth).

At the same time, astronomers are honing another planet-detection technique, the transit method. Consider once again that we’re observing a distant planetary system edge-on. Light from the star would seem to dim ever so slightly when a planet passed in front of it. If, for example, the planet completes an orbit every 10 days, we would have to watch the star for a month or two – noting a slight dimming of the star’s light on each pass of the planet – to be confident of the planet’s existence. The smaller the planet, the more powerful the telescope we would need to detect it.

Alien astronomers viewing our solar system edge-on could make a similar set of observations. “If you had a sensitive enough telescope, you would actually see the Earth transit the disc of the sun,” says Norman Murray, the associate director of the Canadian Institute for Theoretical Astrophysics at U of T. “And a year later you’d see it transit again – and you’d know it was a planet and not a bird or something flying over your telescope.”

The transit method is a promising detection technique; so far, astronomers using it have found about a half-dozen planets. Plans call for sophisticated orbiting telescopes (successors to NASA’s Hubble Space Telescope) that will look for the periodic transits of many stars – possibly leading to the discovery of thousands of extra-solar planets in the coming decades, including, in all likelihood, some planets that resemble Earth.

In the meantime, Murray and the astrophysicist Matthew Holman of Harvard University have devised a way for astronomers to infer the existence of Earth-sized planets without actually seeing them transit their host stars. Their idea, published in the journal Science last winter, involves carefully timing the transits of huge planets the size of Jupiter. The key is gravity. In a system with a Jupiter-sized planet and a smaller Earth-sized planet, the smaller body will induce slight irregularities in the orbit of the larger body. (In a similar way, astronomers in the 19th and early 20th centuries were able to use irregularities in the orbit of Uranus to infer the existence of Neptune and Pluto.)

Think of our hypothetical alien astronomers dozens of light years away, watching our solar system. With powerful enough telescopes, they could detect transits of Jupiter against the sun. If these alien astronomers monitored Jupiter for several decades, they would notice that the time between successive transits was not exactly the same. They could use this discrepancy to infer the existence of at least one other planetary body. (They would likely presume the existence of Saturn, since its gravitational pull would have the greatest effect on Jupiter’s transit times.) If they had even more powerful telescopes, capable of detecting Earth’s transits, they would discover irregularities in our orbit, too. “Such astronomers would see variations in the times between transits in the order of 10 minutes, due primarily to the influence of Venus,” explains Murray.

Murray and Holman’s technique of scrutinizing transit times would allow astronomers to determine properties of the unseen planet that they can’t with radial velocity. The planet’s mass can be calculated, based on its effect on the orbit of the larger planet. Astronomers could also work out the size of the orbit as well as its eccentricity. And if astronomers are really lucky, and see both planets transit the host star, they can also calculate the sizes of the planets. If you know the size and the mass, you can determine density. “So you can immediately say whether the planets are terrestrial or gas,” says Murray. Terrestrial planets are where life is most likely to be found.

To comprehend these strange new worlds, we need to understand how these planets formed – a line of inquiry that Murray and several other U of T astronomers are actively pursuing. “A theory of planet formation would tell us, in principle, what fraction of stars, like our sun, harbour Earth-like planets,” Murray explains. It would also give astronomers a better idea of where to look for them, he says. But the extra-solar planets found so far are tough to explain using our existing theories.

The prevailing view is that a planetary system begins as a slowly spinning, immense ball of gas. The hot, central part becomes the star, while the material far from the core flattens and evolves into a Frisbee-shaped cloud of debris. This cloud – the proto-planetary accretion disc – is thought to exist for about 10 million years before dissipating, and provides the raw materials from which planets eventually form. The basic scenario is still believed valid; what is hotly debated are the details of the process.

One problem with the traditional model is that it implies that giant gas planets should form far from their parent stars. After all, this is where we find them in our solar system. But it’s not where we see gas giants in extra-solar planetary systems. “We had an understanding of how our system formed, how the Earth fits into the planetary system and how the conditions for life evolved in our solar system,” explains Artymowicz. “There was quite a shock when we discovered that other solar systems are different.”

Now, astronomers are trying to fine-tune the old model. At present, they’re torn between two competing scenarios. In the core-accretion model, planets are born when small chunks of rock, sand-grain-sized debris and dust collide within the disc. As the rocky core grows, its gravity draws in more dust and gas from its surroundings. If it’s large enough, over millions of years it will keep on gathering gas until it becomes a giant planet, like Jupiter. If it is smaller, it will become a rocky planet like Earth. A problem with this scenario is that the accretion process is too slow; giant gas planets may not have enough time to form. In a competing scenario, the disc-instability model, denser patches of gas and dust undergo a sudden collapse, causing one or more planets to form in a mere thousand years.

One U of T theorist, however, believes that gas giants can form according to the core-accretion model at a much faster rate than previously imagined. Roman Rafikov, recently from the Institute for Advanced Study in Princeton, New Jersey, has been examining the competing models. The Astronomical Journal just published his argument that a giant planet orbiting a star at a distance equivalent to Neptune’s orbit in our solar system can form “on a time-scale of about 10 million years.”

While the question of how planets form may seem esoteric, it bears directly on the likelihood that other Earths exist, says Murray. The disc-instability model is neutral on the formation of terrestrial planets – they may or may not form. But the core-accretion model requires terrestrial planets to form. Under that model, gas giants are simply terrestrial planets that, over millions of years, continued to gather gas. In other words, if the core-accretion model is correct, Earth-like planets may be commonplace.

The search for another Earth will intensify over the next several years, with the launch of a new generation of space-based telescopes and the construction of immense new telescopes on the ground. Among the most ambitious ground-based projects is the proposed Thirty Metre Telescope, which, when completed by 2015, will be the world’s largest. U of T is one of 15 Canadian universities co-operating on the project, with backing from the National Research Council and several U.S. institutions. A number of U of T scientists are playing major roles in the project, including astronomers Ray Carlberg and Bob Abraham and physicist Pekka Sinervo, U of T’s dean of Arts and Science.

With these new telescopes, astronomers may make their most tantalizing finding yet: a terrestrial planet orbiting within the “Goldilocks zone” of its parent star (the narrow ring that is neither too hot nor too cold for life to evolve). But the diverse and ever-increasing trove of strange new worlds that scientists have already found has triggered a revolution in astronomy.

“It has been a tremendously exciting 10 years,” says Jayawardhana. “After centuries of people talking about it, we have finally found not one, not two, but more than 150 planets around other stars. It’s truly remarkable.”

Dan Falk is a Toronto science journalist and the author of Universe on a T-Shirt: The Quest for the Theory of Everything (Penguin Canada). Additional reporting by Stephen Strauss.