INTRODUCTION

We present a general discussion of this topic, with special emphasis on three specific aspects: detection techniques employed in extrasolar planet searches; the properties of the planetary systems found to date; and the future prospects in the search for other worlds, particularly those which may actually resemble our own Earth.

Meanwhile, van de Kamp, himself director of Sproul Observatory, had been making high resolution photographic plates of Barnard's Star, the second nearest star system to our own, located only 6.0 light years away. He and his students eventually accumulated over 2,000 plates, taken with a large 0.61-meter refracting telescope from 1938 through 1962. His results appeared to show a periodic wobble in the movement of Barnard's Star, which he attributed to the presence of two unseen planetary companions revolving around the star. These bodies, he thought, could be about 0.8 and 1.1 times as massive as Jupiter.

During the 1960's and 1970's, van de Kamp re-examined his data on a number of occasions. In 1975 he announced a new set of results for his two planets—he now proposed the existence of a planet about 0.4 times as massive as Jupiter (hereafter referred to as Jupiter-masses) with an orbital period of 22 years, and a smaller 1.0 Jupiter-mass planet with a period of 11.5 years [Bell, 1997]. Over the course of time, however, his claims were severely questioned, starting with a 1973 paper by astronomers George Gatewood and Heinrich Eichhorn. It is now commonly held that the wobbling effect observed by van de Kamp was probably due to imperfections of the Sproul telescope, on account that new research with better instrumentation has not been able to duplicate his results.

Serious speculation as to the possibility of planets orbiting other stars is also to be found in the early years of exobiology (or astrobiology), a field studying the possible nature and origin of life throughout the universe.

In 1959 physicists Giuseppe Cocconi and Philip Morrison published a paper on the prestigious British journal Nature, suggesting the possibility of using existing radio telescopes to detect signals from intelligent life on nearby star systems. Radio telescopes—that is, instruments consisting of antenna dishes with the ability to receive radio waves from space—had been introduced in the 1930's and were beginning to take hold within mainstream astronomy.

Soon after, in 1960, the Green Bank facility in West Virginia, USA (a part of the National Radio Astronomy Observatory) announced Project Ozma, the first systematic attempt to detect for radio signals from intelligent life. Cornell University astronomer Frank Drake, who independently from Cocconi and Morrison had been doing pioneer work on exobiology, was put in charge of the plan. Drake chose two solar-type stars—stars which are similar to the Sun in mass and evolutionary state—namely, Tau Ceti and Epsilon Eridani, to which he "listened" with a 26-meter radio dish for about 150 hours. Even while he detected only static radio signals, the whole project rested on the premise that these two stars had to have their own planets in orbit around them.

The assumption that extrasolar planets were, perhaps, the rule rather than the exception, was not then as odd as we may think today. Even in the absence of hard evidence, this idea was firmly held by a number of astronomers, such as Green Bank director Otto Struve, and Berkeley professor Carl Sagan. Struve is even recorded as declaring that he "believed that all solar-type stars either had planets or were parts of multiple star systems" [Poundstone, 1999].

In 1961, during an informal exobiology conference held at Green Bank, a formula devised by Drake—and now known as the Drake Equation—was presented to the scientific world. The equation attempts to estimate the number of extraterrestrial civilizations in the Galaxy—at any given moment in time—with the ability to engage in interstellar radio communication.

The equation is written as:   N = R^* f_p n_e f_l f_i f_c L

There were some interesting debates at Green Bank about the possible value of f_p, which, in fact, stands for the fraction of stars that have planets. Struve proposed that f_p must be equal to 0.5, meaning that one out of two stars had to have planets. Four decades later, and in the light of current evidence, it now appears that the actual value may indeed be around 0.5.

By the 1980's, with the advent of orbiting infrared observatories, the first convincing evidence for extrasolar planetary systems began to surface. The detection by the IRAS satellite (1983) of a disk of dust grains around the star Vega, and soon after, around the stars Fomalhaut, Epsilon Eridani and Beta Pictoris, was the first step in the right direction. These protoplanetary disks, as they are now called, are thought to be the formation cocoons of planetary systems. Through time, as the systems evolve, the disc material may gradually accrete into planetesimals, and later into full-fledged planets.

But it was not until the 1990's that the existence of extrasolar planets was to be proven, once and for all.

The discovery itself was very intriguing, since the accepted idea at the time held that planets should be plentiful, but only, perhaps, within main sequence stars. Energy generated by these stars—so called because of their position in the Hertzprung-Russell diagram—is due to the conversion of hydrogen to helium within their core. They are, indeed, what we may call "normal stars", representing about 90% of all visible stars. What makes the discovery a significant one is that these planets actually seem to have formed after the original host star collapsed and became a pulsar [Marcy et al., 2003].

The situation was not only intriguing, but unexpected. Wolszczan was not exactly looking for planets, but rather investigating a pulsar known as PSR B1257+12, which is located in the constellation Virgo at a distance of about 1,500 light years from Earth. It was when he detected an unexplained irregularity in the pulsar's radio emissions that he suspected the presence of unseen companions. Wolszczan was then at the Arecibo Observatory in Puerto Rico (which operates the largest single-dish radio telescope in the world), while Frail, who confirmed much of the original data, was in New Mexico.

"The pulsar planets are like a carbon copy of the inner Solar System" [Croswell, 1997], said Wolszczan after his announcement. Indeed, his planetary system contains some of the smallest planets known to date, with masses of only 0.020, 4.3 and 3.9 times that of Earth. The orbits of these three bodies appear to almost circular, with radii of 0.19, 0.36 and 0.46 astronomical units, comparable in size to the orbits of our planets Mercury and Venus (the smaller, innermost planet was not actually found until 1993).

In 1995, Swiss astronomers Michel Mayor and Didier Queloz announced the discovery of a jovian planet around the star 51 Pegasi. This was the first confirmed detection around a main sequence star. By using the technique of radial velocity variation, Mayor and Queloz were able to measure the planet's periodic perturbations over its host star.

Only recently, with the advent of space-born telescopes and adaptive optics have astronomers been able to achieve the extreme accuracy needed to overcome the aforementioned difficulties. Adaptive optics are computer-controlled systems capable of making minute and instantaneous adjustments to optical surfaces—over a scale of microseconds—thus compensating for the blurring effect of the Earth's atmosphere.

Four main detection techniques have been employed since Wolzsczan's 1991 discovery to find planets outside of our solar system. A description of these techniques follows:

1. Astrometric Displacement—

Through this technique, an astronomer uses photography to measure small, regular variations in the position of a star as directly projected onto the sky [see figure 1]. This was the method used by van de Kamp in his decade-long search attempts. To date, however, only a single, unconfirmed detection has been achieved using this method (two suspected planets orbiting Lalande 21185).

The main limitation of this technique is that astrometric displacements tend to be quite small, even for massive planets orbiting nearby stars. As seen in the figure below, the displacement of our own Sun, due to Jupiter, as seen from a distance of 32.6 light years, would not exceed about 0.8 mili-arc seconds. The precision of Earth-based astrometric measurements is at present about 10 mili-arc seconds, and for space-based measurements, about 1 mili-arc seconds; but this could soon improve to 20 micro-arc seconds and 2 micro-arc seconds, respectively [How to Find Extrasolar Planets].

Related to this issue, is the limitation imposed by the actual duration of the search, since long-period planets would only become evident after perhaps decades of study. Detections by this method would necessarily involve planets with orbital periods shorter than about 10 years.

Figure 1. Astrometric Displacement Source: Shao et al. Comments: This diagram shows the astrometric displacement of the Sun, due to Jupiter, as would be seen from a distance of 32.6 light years. This is precisely what we would expect to see while searching for jovian planets around nearby stars.

2. Radial Velocity Variation—

Through this technique (otherwise known as Doppler spectroscopy), an astronomer uses ultra-sensitive spectroscopy devices to measure small, regular variations in the radial—line-of-sight—velocity of a star [see figure 2]. Most extrasolar planets that have been confirmed to date are due to this method.

This technique seems very effective for main-sequence stars of spectral types about F5 through M. Stars of earlier type (O, B, A and F down to about F4) tend to rotate faster and to pulsate, which makes it difficult to obtain precise radial velocity measurements.

The limitations of this technique are similar to that of astrometric displacements: the extreme sensitivity required to achieve any meaningful result. For example, if we again were to represent the radial velocity variation of our own Sun due to Jupiter, it would only amount to 13 meters per second—in this case, regardless of distance. The precision of radial velocity measurements is at present about 3 meters per second [Marcy et al., 2003].

Another limitation of this technique is that the amplitude of the measured variation in radial velocity will only give us a minimum possible value for the planet's mass. This happens because we are detecting a velocity variation through only the line-of-sight axis, disregarding the fact that the planet's orbit may actually be inclined with respect of this axis. The actual relation between the two quantities is expressed as m = M sin (i), where M stands for the true mass, m for the minimum possible mass, and i for the orbital inclination angle [formula derived from How to Find Extrasolar Planets]. This last quantity, however, is very difficult to ascertain, unless the planet were to be seen in transit across its host, in which case the angle must actually be very close to 90°.

Figure 2. Variation of the radial velocity of 51 Pegasi Source: Mayor and Queloz Comments: The graph shows a cyclical variation of radial velocity with an amplitude of about ±57 meters and a period of 4.23 days. These precise measurements are achieved by means of high resolution spectrographs using iodine cells, which provide reference absorption lines which are then overlaid on to a star's spectrum.

3. Pulsar timing—

Through this technique, an astronomer uses a radio telescope to measure small, regular variations in the timing of the radio pulses emitted by a pulsar. To date, two detections have been achieved using this method: a three-planet system around PSR 1257+12, and a single planet around PSR B1620–26.

4. Transit Photometry—

Through this technique, an astronomer uses a photometric device to measure small, regular variations in the apparent brightness of a star, induced by the passage of the suspected planet through the stellar disk. [see figure 3] To date, only one detection has been achieved using this method, namely, a single planet orbiting the star OGLE–TR–56. One other planet—that of HD 209458—which was originally detected through Doppler spectroscopy, was later confirmed by transit observations.

A big limitation of this technique, however, is that transits may only occur if a planet's orbital inclination with respect to our line-of-sight axis were almost 90°. This is rarely the case, through the probability increases proportionally with the planet's size and inversely with its orbital size. It is estimated that a Jupiter-sized planet orbiting within 0.05 astronomical units of its host has about a 10% chance of transit, while a terrestrial planet may only have about a 0.5% chance [Gaudi, 2003].

The great advantage of the transit photometry method, however, is its accessibility. Even a commercial-grade telescope, with a CCD detector and a laptop computer (all of which may be bought for less than US$5,000) is capable of detecting the light variations produced in a star by the transit of a large jovian planet, with a level of precision which is close to current limits. The precision of Earth-based photometric measurements is at present about 0.1% (equivalent to about 0.001 magnitudes) and for space-based measurements, about 0.001% (0.00001 magnitudes) [Schneider, 2001].

Another advantage of transits is that the timing and intensity of the dimming allow the size of an extrasolar planet to be ascertained, and if this is combined with a mass measurement (which may be obtained from astrometric data, from example) the planet's density may then be found. A density estimate may give clues about the chemical composition and internal structure of a planet.

Figure 3. Diagram and light curve of a simulated transit event Source: Esquerdo et al. Comments: The rapid decrease in a star's luminosity, as well as the minimal amplitude of the variation—as shown by this light curve—is an unmistakeable signature of small body transiting around a star.

In summary, techniques 1 through 3 all result from the same basic idea, that is, measurement of the gravitational perturbation exerted by the suspected planet over its host. In each case, the observed period of variation is, by definition, equal to the planet's orbital period, while the amplitude of variation is proportional to the planet's mass [Schneider, 2001].

Both of the above referred quantities can be combined through Kepler's Third Law to obtain a value for the planet's orbital radius. Light curves from transit observations give additional data about planetary diameters and orbital inclinations, which can be further combined with data obtained from other techniques—as explained above—to infer about the planet's composition and structure. (A light curve is a graph showing how on object's brightness changes through time.)

A number of new, sophisticated search techniques—some of which look very promising for the near future—have been suggested, or even experimented with. Direct imaging is thought to be possible through a method known as nulling interferometry. The goal here is to combine light signals from a number of different telescopes in such way as to suppress the host star, leaving its suspected planet's faint light intact. Photometric microlensing (otherwise known as gravitational lensing) has also been proposed. Through this method, an astronomer uses an exotic effect predicted by Einstein's General Theory of Relativity, namely, the ability of a massive object—by virtue of its strong gravitational field—to bend or distort the incoming light of stars. This is expected to produce a significant brightening of a host star bearing a lensing-capable planet.

An important concept—namely, the difference between Jupiter-like and Earth-like planets—needs to be explained before we embark in any meaningful discussion of this topic. A jovian planet is a large, low-density body made of gases (mainly hydrogen and helium) and liquid ices (such as water ice, methane, ammonia and carbon dioxide), while a terrestrial planet is small, dense body made primarily of iron and silicate rocks, thus actually possessing a solid surface, unlike its jovian counterpart.

The great majority of extrasolar planets so far detected belong to the jovian type, with masses ranging from approximately 0.2 to 10 Jupiter-masses. They tend to move in small orbits with periods sometimes as short as 3 days, and rarely longer than one year. With the possible exception of the planet moving around pulsar PSR B1620–26—which may orbit at a mean distance as large as 64 astronomical units—and an unconfirmed secondary planet of Epsilon Eridani, none has been found to possess an orbital radius larger than 6 astronomical units.

This particular set of properties results from a selection effect inherent to the radial velocity technique, whereas massive planets in smaller orbits will have a stronger gravitational influence on their hosts, thus making themselves apparent more readily to telescopes currently available.

There is, however, a big problem here. According to our current planet formation models, these epistellar jovians (otherwise known as "hot Jupiters") could not have possibly formed as close to their hosts as we see now them. It has been traditionally thought that jovian planets, containing a significant amount of ices and volatiles, can only accrete far away from stellar heat, for otherwise these compounds would evaporate. We do not have at present a conclusive answer to this issue, but it has been suggested that there must be some previously unaccounted mechanism forcing these planets to migrate inwards at some early point during their evolution [Darling].

It has also been found—in contrast to what we see in our Solar System—that a large fraction of these planets move in orbits with a significant degree of eccentricity, usually 0.3 or above. Few of them, however, seem to belong to double or multiple star systems. This may give credit to the suspicions of many astronomers, who doubt that a stable planetary system could actually develop under the complex gravitational interactions that are found within a multiple star system.

To give an idea of the nature of these new planets, the individual properties of three specific planetary systems are summed up below. [All data is taken from the Paris' Observatory's Extrasolar Planet Encyclopaedia website, unless otherwise stated.]

51 Pegasi—
This system consists of a single large planet orbiting around a main sequence, G4-type star located about 50 light years from Earth. The planet has an calculated minimum mass of 0.46 Jupiter-masses, describing an orbit around its host with a period of 4.23 days which corresponds to a mean distance of 0.0512 astronomical units. The orbital eccentricity is about 0.013—close to that of the Earth around the Sun, which would mean a nearly circular orbit—and its inclination has been estimated at nearly 90°.

Upsilon Andromedae—
A system composed of three jovian planets orbiting a wide binary star (perhaps a triple system) located about 44 light years from Earth. The planets have calculated minimum masses of 0.69, 1.89 and 3.75 Jupiter-masses and orbital periods of 4.6170, 241.5, and 1,284 days. The inner planet orbits in a nearly circular orbit, while the outer bodies move in elliptical orbits of about 0.3 eccentricity [see figure 4].

Figure 4. Orbital diagram of the Upsilon Andromedae planetary system Source: University of San Francisco Comments: This is the first multiple-planet system found around a main-sequence star other than the Sun. The innermost planet was detected in 1996 by Geoffrey Marcy and R. Paul Butler, while the two outer bodies would only be confirmed three years later.

HD 209458 (8^th magnitude star in the constellation Pegasus)—
This system consists of a single jovian planet with a substantial atmosphere orbiting around a solar-type star located about 150 light years from Earth. The planet's mass has been determined to be 0.69 Jupiter-masses, and it orbits its host with the accurately determined period of only 3.524738 days, corresponding to a distance of 0.045 astronomical units. The orbit seems to be an almost perfect circle, with an inclination that has been accurately measured as 86.1°.

In view of these discoveries, the old question of whether there may be life in other worlds immediately comes to mind. The answer is inextricably tied to the concept of a stellar habitable zone.

A stellar habitable zone is defined as "an imaginary spherical shell surrounding a star throughout which the surface temperatures of any planets present might be conducive to the origin and development of life as we know it" [Darling]. This is sometimes referred as the range where water—in its liquid form—may exist on the surface of a planet, an important factor which may lead to the eventual evolution of life. Regrettably, in spite of the good assortment of planets found to date, none has yet been found dwelling inside of an habitable zone. The planetary system of 47 Ursae Majoris, however, because of the moderately wide orbital radii of its two jovian planets, is currently considered the closest match.

It is amazing to consider how much we have learned about planets—and planetary systems in general—during the last decade. But then, will it ever be possible to find worlds similar to own, in particular, those actually dwelling within habitable zones?