Are we alone? There is perhaps no more important single scientific question. People have pondered this issue from the very dawn of sentience, wondering if other, similar, beings inhabited a distant mountain range or the other side of an ocean. The history of humanity is largely one of exploration and expansion, and while at first this was limited to the Earth’s surface, in the last few decades only the power of our interplanetary rockets has kept us from exploring our wider environment. As the science-fiction author Ray Bradbury proclaimed when the Viking landers arrived on Mars in July 1976, “There is life on Mars, and it is us.”

Now we are poised to move far beyond our own planetary system — if not yet physically, then at least intellectually — as we learn how many other habitable worlds exist in our galaxy. The expectation of most astronomers is that Earth-like planets will be commonplace; in other words, Earth-size planets will be found orbiting at distances from their parent suns that would allow liquid water to exist on or near the planet’s surface. This prediction is not simply a conceit concocted by wild-eyed planet hunters, but is a logical outcome of what we have learned in the last decade about extrasolar planetary systems. All of the evidence points to the inescapable conclusion that rocky, terrestrial planets similar to the Earth should be orbiting unseen around many, if not most, of the myriad of stars that we see when we look up at the night sky.

This prediction of the likelihood, or frequency, of Earth-like planets orbiting Sun-like stars is about to be tested rigorously by the Kepler Space Telescope, which is designed specifically to determine the percentage of solar-type stars that harbour Earth-like worlds. Due for launch early this month, NASA’s latest telescope will search for extrasolar Earths using a method known as transit photometry. When an Earth-sized planet with a suitably inclined orbit passes between its star and the line of sight to the Earth, it blocks a tiny fraction — about one 10,000th — of the star’s light. Kepler is capable of measuring these tiny, periodic dimmings, and will monitor more than 100,000 stars for at least three and a half years. In the process it should discover dozens of Earth-like planets. In short, we are about to determine the value of a key factor in any estimate of the prevalence of life in the universe: the frequency of habitable worlds where life could originate and evolve.

Early exoplanetology

Until 1995 the only known planets outside our solar system were two scorched rocks in orbit around a very un-Sun-like pulsar, PSR B1257+12. This shortage of extrasolar planets was not due to a lack of effort in trying to find them. Since 1980 a Canadian research team led by Gordon Walker of the University of British Columbia had been searching a sample of nearly two dozen Sun-like stars for the telltale “Doppler wobble”, or periodic redshift and blueshift in the star’s spectrum, caused by the orbital motion of the host star around the system’s centre of mass. By 1992 the researchers had found no definitive evidence for even a single extrasolar planet comparable in size to Jupiter. They therefore stopped their search and wrote a paper describing the upper limits (of the order of one Jupiter mass) on the masses of any possible planets lurking undetected around the target stars. When this paper was published in the autumn of 1995, the field of extrasolar planetology seemed to have come to a premature dead end.

Two months later, Michel Mayor, an astronomer at the Geneva Observatory in Switzerland, made a startling announcement: he and his colleague Didier Queloz had found the first reproducible evidence of a Jupiter-mass planet orbiting a solar-mass star. Mayor had only begun his planet search in 1994, but his target list contained more than 100 stars — including, crucially, 51 Pegasi, located about 50 light-years away from Earth. The newly discovered planet, named 51 Pegasi b, has a mass at least half that of Jupiter and causes a periodic Doppler shift that the Canadian team could have detected — if 51 Pegasi had been one of their target stars.

With an orbital period of just four days, 51 Pegasi b occupies an orbit 100 times closer to 51 Pegasi than Jupiter is to the Sun, making it the first example of the class of planets known as “hot Jupiters”: gas giants with surface temperatures of about 1500 K, which is closer to those of low-mass stars than of Jupiter (150 K). The existence of 51 Pegasi b was confirmed a week later by the astronomers Geoff Marcy and Paul Butler, then at San Francisco State University and the University of California at Berkeley, respectively, who had been searching for gas giants with Doppler spectroscopy since 1988. After Mayor’s announcement, they combed through their seven years of accumulated data on over 100 stars. They quickly found two more “Doppler planets”, 70 Virginis b and 47 Ursa Majoris b. This time, the gas-giant planets had more sensible orbital periods of about 0.33 and three years, making them warm and cool Jupiters, respectively.

The field of extrasolar planets grew enormously after the Swiss and US teams announced their discoveries in early 1996. To date, these two teams have found most of the more than 300 Doppler planets now known to astronomers. Such planets are surprisingly common, with over 10% of solar-type stars appearing to have gas-giant planets with orbital periods less than about five years, with even more planets orbiting at larger distances. Hot Jupiters like 51 Pegasi b occur around about 1% of solar-type stars: if the Canadian team had searched a sample of 100 stars, as the Swiss did, they might have found the first hot Jupiter.

On the theoretical side, the wave of new exoplanet data has left models of gas-giant planet formation in ferment, with two viable mechanisms that may explain the observed range in mass (roughly Saturn mass to 13 Jupiter masses) and relatively high frequency of gas-giant exoplanets. The first, “disk-instability”, mechanism would allow gas giants to form in even the shortest lived protoplanetary disks. Such disks occur in regions of massive star formation. There is good evidence that the solar system formed in such a commonplace area, implying the ubiquity of similar systems. The other possible mechanism relies on “core accretion”, where a roughly 10-Earth-mass core forms and then slowly accretes sufficient gas from the disk to become a gas giant. If the disk gas disappears faster than the cores form, the result is “failed cores’’ similar to the ice-giant planets Uranus and Neptune.

From hot Jupiters to super-Earths

Short-period planets like the hot Jupiters have a 10% chance of having their orbits inclined such that the planet passes in front of its host star as seen from Earth. By measuring the fractional change in the star’s brightness during these transits, we can calculate a planet’s size relative to the size of the star, which can in turn be estimated based on its stellar classification. Hence, transit measurements offer final proof that the Doppler-planet candidates actually are gas-giant planets similar to Jupiter and Saturn. The 10th hot Jupiter found using Doppler spectroscopy by Butler (by then at the Carnegie Institution) and his colleagues turned out to be the first transiting planet, HD 209458 b.

Transits also provide the opportunity to study a planet’s atmosphere as starlight passes through it during a primary eclipse (which occurs when the planet passes in front of its star), or as the planet’s thermal emission is blocked during a secondary eclipse (planet behind the star). Clever observations with the Hubble and Spitzer space telescopes have allowed astronomers to detect the presence of water, carbon dioxide, methane, sodium and hydrogen in the atmospheres of several known hot Jupiters. Transit surveys are also being used to discover new hot Jupiters, with over 50 such planets having been found so far using this method followed up with Doppler confirmations.

Continued improvements to Doppler spectrometers have allowed the Mayor, Marcy and Butler groups to discover not only Jupiter-type planets but also short-period planets with masses as little as four times that of the Earth. Most of these hot and warm “super- Earths” are known to be accompanied by gas-giant siblings at greater orbital distances. This configuration — reassuringly similar to that of the terrestrial and gas-giant planets in our solar system — suggests that these super-Earths formed closer to their parent stars than the gas giants did. Theoretical models of planet formation firmly predict that these planets must be at the high-mass end of the range of rocky extrasolar planets; hence, rocky planets with even smaller masses must exist as well.

Estimates of the frequency of hot and warm super- Earths based on Doppler surveys imply that perhaps a third of all solar- and lower-mass stars have such planets. Mayor’s team has even found one star, HD 40307, that has three orbiting super-Earths with masses of about four, seven and nine times that of the Earth (see “Super-Earth trio”). Several other multiplanet systems have also been found using Doppler spectroscopy. The 55 Cancri system, for example, contains at least five known planets, with the innermost planet being a hot super-Earth with an orbital period of 2.8 days.

Total eclipse of a star

In addition to the Doppler-wobble and transit methods, a third technique, known as microlensing, has also recently been used to discover planets. When a foreground star happens to pass directly in front of a distant background star, the background star’s light can be bent toward the Earth by the gravity of the foreground star. The result is a gradual brightening and dimming of the background star as the unseen foreground star passes in front of it. If the foreground star has a planet in orbit at a distance similar to the asteroid belt in our solar system, then observations will reveal an additional, sharper brightening and dimming of the background star that can only be caused by a third body. The critical distance is known as the Einstein radius, since Einstein predicted this effect in 1936. However, he assumed that it could never be observed in practice because the chance of having the two stars line up in exactly the right configuration when viewed from the Earth would be small, and the chance of observing it at the right time even smaller.

Acting independently, two teams of astronomers (led by Ian Bond from the Institute of Astronomy in Edinburgh, UK, and Andrzej Udalski of the Warsaw University Observatory in Poland) discovered the first microlensing planet in 2004. The new planet had a mass 2.6 times that of Jupiter, and was designated OGLE2003-BLG-235/MOA2003-BLG-53 b as its discovery involved observations by both Bond’s Microlensing Observations in Astrophysics (MOA) team and Udalski’s Optical Gravitational Lensing Experiment (OGLE) team.

This discovery was made possible by Udalski’s participation in the OGLE search for microlensing events, which involves monitoring hundreds of thousands of stars in our galactic bulge with a modest 1.3 m telescope at the Carnegie Institution’s Las Campanas Observatory in Chile. Einstein could not have foreseen such a dedicated effort to test his 1936 prediction.

Microlensing has detected eight planets to date: five gas giants and three cold super-Earths with masses in the range from about 3 to 13 Earth-masses, including one multiple system containing planets analogous to Jupiter and Saturn. The cold super-Earths discovered using microlensing appear to be similar to the solar system’s ice-giant planets, Uranus and Neptune, because of their similar masses and orbits at large distances where ices will be common. Combined with the Doppler discoveries of abundant gas giants, and hot and warm super-Earths, astronomers have identified extrasolar analogues for each of the three basic classes of solar-system planets.

Seeking (higher) resolution

One of the major goals of planet hunters is to take a spatially resolved image of a planet, so that the planet’s atmosphere can be more easily distinguished from that of the host star. Six claims for the first image of an extrasolar planet were advanced in 2008 alone.

Perhaps the most convincing of the six is the image of three possible planets in orbit around the star HR 8799 by Christian Marois and his colleagues at Canada’s Herzberg Institute of Astrophysics (see “Seeing new worlds” and Physics World January 2009 p5, print edition only). The masses of the planets are estimated to be in the range of 7 to 10 Jupiter masses, and they orbit HR 8799 at distances of about 24, 38, and 68 astronomical units (the mean distance of the Earth from the Sun), distances comparable to those of the ice giants and Kuiper belt in our solar system.

One of the main uncertainties in estimating their masses is in estimating their ages: given a theoretical model of how a newly formed planet cools with time, the age of the star is used as the age of the planet, allowing the planet’s mass to be guessed. Marois has estimated that HR 8799 is about 60 million years old, but if the star were considerably older than that, then the masses of the three orbiting bodies would be greater than about 13 Jupiter masses. This limit is important: once over it, such “planets” can burn deuterium and so are classed as brown dwarf stars rather than planets (see “An exoplanet encyclopaedia”).

Only further observational and theoretical work will decide which, if any, of these six candidate images is indeed that of a planetary system. These ambitious efforts presage the ultimate goal of obtaining direct images of nearby Earth-like planets and studying their atmospheres for evidence of molecules associated with habitable (carbon dioxide, water) and inhabited (oxygen, methane) worlds.

Given the discoveries of the last decade, we cannot argue that the solar system is a fluke of the universe. While the chaotic nature of the planet-formation process rules out finding an exact analogue, it is clear that planetary systems similar to our own must be a common feature of our galactic neighbourhood. Indeed, only last month astronomers using the CoRoT space telescope discovered the smallest sized transiting exoplanet to date: a fiery new world less than twice the radius of the the Earth that orbits its star every 20 hours.

The Kepler mission will determine how frequently Earths occur in our galaxy: do 1% of Sun-like stars have Earth-like companions? 10%? 100%? Given that there are billions of Sun-like stars in our galaxy alone, the number of Earth-like worlds must be similarly immense. Whatever the answer, by the time Kepler finishes its primary mission in 2013, we will know just how crowded the universe really is.