Natural History Magazine
Want conclusive proof that planets exist in the universe? You are standing on one. But before you complain about what constitutes admissible evidence, consider that it’s easier to find something for which you already have an example than to find something that is predicted solely by somebody’s theory.
A pessimistic estimate for the number of planets in the universe is nine. A more realistic estimate is about one sextillion (one followed by twenty-one zeroes). If, for some reason, you have a problem visualizing the magnitude of this number, consider that it’s a million times larger than the total number of sounds and words ever uttered by all humans since the dawn of the species. You can, of course, search for these planets just for the sake of finding planets. You will probably find some. But why search for planets unless you can simultaneously ask the question whether the planets you find resemble the ones you know, and whether any of them can sustain life? Indeed, these questions guide most planetary searches.
Often we make the implicit assumption that our solar system (which we know contains life, whether or not the life can be considered intelligent) is neither special nor unusual nor peculiar. A more troublesome thought is that our solar system is not ordinary precisely because it possessed characteristics that led to the evolution of beings who ask these questions—a point made recently by George Wetherill, a Carnegie Institution Planetary Astrophysicist, at a conference on the search for extra-solar planets. Wetherill’s comments notwithstanding, I remain in the “We’re ordinary” camp and proceed with the next step to create a catalog of nearby stars that make good candidate hosts for planets.
To theorize and discover other worlds was not always greeted with praise headlines. In the year 1600, such thinking cost Brother Giordano Bruno his life. In the Square of Flowers in Rome, the Catholic Church burned Bruno (naked) at the stake—not for being an ordinary heretic but for being an
impertinent and pertinacious heretic. Bruno’s crime? He reasoned that the universe must be infinite because otherwise it would have to exist in one place rather than another, which conflicted with his philosophical sensibilities. From the vastness of this infinite universe Bruno then concluded there must be many other worlds beyond Earth. Fortunately today, scientists who hold such views are allowed to live. And their discoveries are heralded with front-page headlines.
Armed with these expectations and templates, the discovery of planets around other stars ought to be a cinch. We first look in the galaxy for stars that have similar temperature, size and age as the Sun. We call them solar-type stars. We then look for large gaseous planets. These are Jupiter-like planets. We then strain to find smaller rocky planets that resemble Earth. And of course Earth happens to be a very good example of an Earth-like planet.
The problem is that the simple act of looking does not always reveal what you seek. Your ability to resolve information about a distant object relies on three factors: 1) the stability and transparency of the medium you are looking through, 2) the size of your eye’s pupils, and 3) the wavelength of light with which you observe. For these reasons, the widely repeated claim that an astronaut can resolve the Great Wall of China from orbit is highly dubious. (Indeed, Interstate 10—from Florida to California—is several times as wide and fifty percent longer than the Wall yet nobody in orbit waxes poetic about the visibility of the freeway system in the United States.) Assuming visible light and no clouds or other obscuring atmospheric conditions, an astronaut trying to resolve the Wall would need eyeball pupils about the diameter of the average human head. Consequently, using large telescopes in orbit above Earth’s atmosphere will always improve your vision of the rest of the universe.
The most obvious way to discover a planet around another star is by direct detection. But planet detection remains one of the most challenging things you can do with a telescope. Heroic efforts have been launched by persistent astronomers armed with clever techniques and state-of-the-science hardware. When using the visible light part of the spectrum, it’s not uncommon for the host star to be 100 million times brighter than the reflected light from its planets. Therein lies most of the detection problem. When using the infrared part of the spectrum, however, the star might be only ten million times brighter. The energy radiated by planets, however feeble it may be, typically peaks in the infrared, which maximizes a planet’s chances of being detected.
You might also want to detect a planetary system in the act of forming. Current theories of star formation show that as a gas cloud collapses to form the host star, an extended orbiting disk of gaseous, rocky, and icy material can be left behind. Like the planets that are formed from it, the disk can also be detected in the infrared. Indeed, the first discoveries (in the 1980s) of stuff orbiting stars other than the Sun were of these proto-planetary disks. The most famous star among them is Vega, which starred in the recent Hollywood film Contact. The film’s authors (science writer Ann Druyan and noted astronomer Carl Sagan) knew, of course, that Vega has this disk of debris. So when intelligent radio signals are detected from the system, a scientist in the film questions the likelihood of life having evolved on a such a young planetary system, especially since it is still debris-filled and planets are probably still forming. Sure enough, we later learned in the film that the intelligent life evolved somewhere else and used the Vega system as an outpost.
For studies of proto-planetary disks, many questions remain unanswered. Are they more predominant around low mass stars or high mass stars? Suppose the system gives birth to binary or multiple stars (which account for nearly half of all star systems in the galaxy); what then becomes of the disk? What determines how far away from the host star a massive planet like Jupiter can form. Is a nine- or ten-planet system common? Or does the typical planetary system have just one or two (or thirty or forty) planets? Are there some stars that form disks but that never collapse further to form planets? Are we asking the right questions?
Planet orbits can be strongly unstable in binary and multiple star systems. The constantly shifting gravitational allegiance from one star to another inhibits the tidy circular and elliptical orbits that we have come to know in our single-star solar system. For planets to avoid being ejected, their orbits must be large compared with the separation among the host stars. In this picture, the planets are distant enough for the combined gravity of the stars to feel like the gravity of a single star with the combined mass. Another way to avoid getting ejected is to have a tight orbit around just one of the stars, with an orbital distance that is small compared with the separation among the stars.
The best we can do now is collect as much data as is technologically possible and slowly fill in the holes left by our ignorance of planet formation.
A powerful method to detect planets, but does not result in a headline-ready picture, uses the time-honored Doppler shift in the frequency of light from the host star—named for Christian Johann Doppler, the nineteenth century German physicist who first measured a shift in the pitch of sound of a train whistle as it approached and then receded. The shift turned out to be a general feature of all waves emitted (or reflected) by something in motion, including light waves. Yes, you may blame Mr. Doppler for your recent highway tickets. The Doppler shift returned from your speeding car is electronically calculated by a police radar gun.
Contrary to common expectations, planets do not orbit stars that are fixed at the center of the system. Both planets and stars orbit their common center of gravity, which is not always (in fact, is almost never) at the center of the host star. As planets swing in their orbits, the star responds by making tiny loops of its own. The more massive the planet, and the larger its orbit, the bigger the loop that the star makes in response. When seen approximately edge-on, the wavelength (or frequency) of a star’s light will shift back and forth as the star executes its tiny orbits around the system’s center of gravity. The Doppler shift allows us to deduce the corresponding back-and-forth motion of the host star and infer the presence of one or more planets. When there is a single planet, the Doppler shift is simply periodic. For multiple planets, however, the Doppler shift can have a complicated signature of multiple jiggles, which must be decoded to infer the exact number of planets that are responsible. Once during the taping of a television interview about the discovery of planets I attempted to describe this tiny circular motion of a host star by jiggling my hips. This was a mistake because my hips were all that ended up on the network’s evening news.
Unfortunately, Earth-size planets do not induce a large enough jiggle in a host star to be detected with the limited precision of current instruments. Earth is less than one three-hundred-thousandths the mass of the Sun. Understandably, Jupiter-sized planets do leave a detectable Doppler signature in the light of the host star, which accounts for why the first dozen planets ever detected around ordinary stars were Jupiter-size planets: they are simply easier to find.
The growing number of extrasolar planets includes several host stars that are bright enough to be seen in the nighttime sky with binoculars. Those that initially received quite a bit of press included 51 Pegasi, 47 Ursae Majoris, and 70 Virginis, each discovered by the American planet-hunting team of Geoffrey Marcy and Paul Butler. These genitive code-names are simply the number, in order of brightness, of the stars in the constellation Pegasus (the Flying Horse), Ursa Major (the Great Bear—a part of which we all call the Big Dipper), and Virgo (the Virgin). No surprises here. All these stars resemble the Sun in age and temperature, and sure enough, they each have a Jupiter-sized planet.
Things got weird, however, when we learned that the Jupiter-like planets around 51 Pegasi, and 70 Virginus are uncomfortably close to their host stars. We had no theory for a jumbo gaseous planet to form, much less survive, in close orbit to a host star. For a nascent planetary system we might expect the central star to compete for material with the innermost planets. We know that high-velocity gas spews forth from newborn stars and forms strong “winds” that blow away nearby material, thus inhibiting the formation of close-in, Jupiter-style planets. The planet around 51 Pegasi is a mere 5 million miles from its host star. At one eighth Mercury’s average distance from the Sun, it has no business being there. But there may be some salvation for our theories. Recent studies of 51 Pegasi suggest that the Doppler variation in its light, from which the existence of a planet was inferred, may be due to fluctuations in the star’s surface instead. If true, then theories of planet formation are, for the moment, safe.
The royalty of Doppler-discovered planets are those discovered in orbit around pulsars—the neutron-rich remains of a high mass star that has violently exploded. Since pulsars rotate rapidly and pulse (as you might have expected), you can time the arrival of these pulses with extremely high precision and look for Doppler variations that are small enough to discover tiny, low-mass chunks of rock like Earth. When this was first done in 1992, an Earth-sized planet was discovered that had an orbital period of 184 days. Nobody had ever discovered a planet by this method, or by any other method, so skepticism was high. Skepticism was further piqued because one of the last places you would ever expect to find planets in the galaxy is in orbit around the remains of a high mass star that had blown most of its guts all over the galaxy. With those guts went most of the gravity that formerly kept the system bound together. So we expect all planets the star might have had to fly away into interstellar space. We are left with no other choice. The planet—small and rocky—must have formed in the explosion of the star’s remains.
But shortly thereafter, the pulsar planet discovery was retracted as having been a spurious signal in the data—the details of Earth’s elliptical motion around the Sun had not been fully accounted for. Notice that the planet’s advertised period, 184 days, is almost exactly one half an Earth year, which had raised a few suspicious eyebrows when it was first announced. When the detailed motion of Earth’s elliptical orbit around the Sun was subtracted (it had been approximated as a circle), the previously announced planet disappeared from the data. Such are the precarious pathways of cosmic discovery.
In the same year, another group headed by the astrophysicist Aleksander Wolszczan, had the audacity to report a planet around another pulsar. But their data were unimpeachable. Indeed more data revealed a multiple planet system with objects that resembled Earth more than Jupiter. Remember that all of this was deduced from the undulating Doppler signature in the pulsar’s pulses. But do we understand pulsars sufficiently well to be sure that some unknown mechanism is not at work masquerading as the signature of a planet? The true test came from a prediction that if indeed they are planets, and if they have the gravitational attractions we expect for them, then the interaction of their gravitational fields would produce slight changes in their orbits and thereby impart an additional signature in the pulsing patterns of the pulsar. This was, in fact, measured. They are, in fact planets. And we do, in fact, understand pulsar pulses.
In the public’s imagination, pulsar planets do not capture the romance that planets around real stars do because in the end, everybody is interested in the prospect of life. And so was Dan Goldin, the Head of NASA, when he announced that one of NASA’s goals is to obtain an image of a rocky (Earthlike) planet with high enough resolution to identify continents and oceans. How? With kilometer-long arrays of space-based telescopes that are programmed to function as though they were a single, large telescope with super-duper high resolution. Telescope arrays such as these are known as interferometers and have always provided, in their many applications, the highest resolution measurements of any available technology.
You can bet that if space travel ever becomes affordable, and if we somehow manage to travel to the stars, then the first stops on the interplanetary bus route will be those stars that harbor the Earthlike planets that we have already snooped out. Apart from the obvious technological and biological barriers of interstellar travel (at 100 miles per second, which is over ten times faster than humans have ever traveled though space, it would still take nearly 8,000 years to reach the closest star to the Sun), such a move is hardly justifiable scientifically until every nook and every cranny of every planet in our own solar system is searched for life.
In selecting a planet to visit, assuming you want to discover life, much can be learned from the analysis of its atmosphere. With space-based spectrographs, you can also decode the light and deduce the chemical composition of the planet’s atmosphere. The composition of Earth’s air distinguishes us from the other planets in the solar system because oxygen (O2) is common in the presence of photosynthesis, where carbon dioxide (CO2) gets its oxygen stripped while donating its carbon atom to the growth of life. Oxygen also allows for the development of an ozone (O3) layer, which shields the planet’s surface from harmful ultraviolet rays and allows complex molecules to thrive.
And if our sensitive spectrographs also detect smog, ozone-destroying chloro-fluorocarbons, hydrocarbon contaminants, soot from global deforestation, and localized atmospheric radiation belts, we will know for sure that we have found intelligent life.