The fantastic success of the Hubble Space Telescope (HST) at a wide range of astronomical observations has spoiled us all—we have come to expect to see images on a frequent basis of gorgeous cosmic locales, from nearby comets to distant galaxies. Taking an image of a nearby extrasolar planet is no more difficult than taking an image of a distant galaxy (V ~ 30 mag), except for the fact that the faint planet is located right

Figure 1. Apparent astrometric detection of a 1.6 Jupiter-mass planet orbiting Barnard's Star on a 24-year period, eccentric (e = 0.6) orbit (solid line), based on yearly means of Sproul Observatory data (van de Kamp 1963).

Figure 1. Apparent astrometric detection of a 1.6 Jupiter-mass planet orbiting Barnard's Star on a 24-year period, eccentric (e = 0.6) orbit (solid line), based on yearly means of Sproul Observatory data (van de Kamp 1963).

next to its much brighter host star. At optical wavelengths, the host star may outshine its planets by a factor of 109 or more. HST was not designed to be able to separate out the light of a planet from its host star.

As a result, essentially all of our information about extrasolar planets has come from indirect detection methods, where the existence of the planet is inferred based on the gravitational reflex motion of its host star as it orbits the center of mass of the entire system. This motion can be detected in several different ways, with the first studies having sought the back-and-forth motion of the star on the plane of the sky with respect to nearly stationary, background stars: the astrometric technique.

In 1937 Peter van de Kamp became the Director of Swarthmore College's Sproul Observatory. In the next year, he began a long term astrometric program to search for low-mass companions to nearby stars with the Observatory's 24-inch refractor telescope. One of the first stars he added to the target list was Barnard's Star, discovered in 1916 by E. E. Barnard. Barnard's Star is a red dwarf with a mass of ~0.15 Mq, close enough at 1.8 pc that only the Alpha Centauri triple system is closer to the Sun, but so faint that it cannot be seen with the naked eye. As a low-mass star very close to the Sun, it is an excellent candidate for an astrometric planet search: a Jupiter-like planet would force Barnard's Star to wobble over a total angle of ^0.04 arcsec, a wobble of several microns on the photographic emulsions used to record the trajectory of Barnard's Star, as it sped across the sky at 10 arcsec per year.

Astrometric claims for very low-mass companions to the stars 70 Ophiuchi and 61 Cygni had been made in 1943 by two different groups, but these claimed detections of objects with masses in the range of 10 to 16 Jupiter masses could not be verified and were soon discarded, if not forgotten. Only after taking 2,400 observations of Barnard's Star for over two decades was van de Kamp ready to announce his discovery of a planet, given this unfortunate prior history. In 1963 van de Kamp announced that he had found the first extrasolar planet: a planet with a mass only 60% greater than that of Jupiter, orbiting Barnard's Star with a period of 24 years (van de Kamp 1963). In order to be certain of its reality, he had waited for an entire orbital period to elapse before making the an-

2.1. Barnard's Star

Figure 2. Refutation of a planetary companion to Barnard's Star, based on an independent data set and astrometric analysis (Gatewood & Eichhorn 1973). The size of the data points indicates their weight in the astrometric solution—smaller data points have larger errors. Only one astrometric axis (x) is shown.

Figure 2. Refutation of a planetary companion to Barnard's Star, based on an independent data set and astrometric analysis (Gatewood & Eichhorn 1973). The size of the data points indicates their weight in the astrometric solution—smaller data points have larger errors. Only one astrometric axis (x) is shown.

nouncement (Figure 1). The semimajor axis of the planet was 4.4 AU, similar to Jupiter's 5.2 AU, and the fact that the orbit was considerably more eccentric than Jupiter's orbit did not raise too many questions about what had been found. Astronomers expected gas giant planets to exist elsewhere, and within a few years, Barnard's Star literally became the textbook example of another star with a planetary system.

On the advice of a senior professor at the University of Pittsburgh who must have doubted the existence of van de Kamp's planet, George Gatewood undertook a second study of Barnard's Star. Investigating an independent collection of photographic images of Barnard's Star, Gatewood used a new plate-measuring engine at the U.S. Naval Observatory to remove the human element from the plate-measuring process. Applying mathematical techniques derived by his thesis advisor, Heinrich Eichhorn, Gatewood analyzed 241 plates taken at the University of Pittsburgh's Allegheny Observatory and at the van Vleck Observatory in Connecticut. Surprisingly, their analysis was not able to confirm the existence of a planet orbiting Barnard's Star (Gatewood & Eichhorn 1973). Figure 2 shows that their data precluded the large amplitude wobble in van de Kamp's data.

The situation got much worse for van de Kamp's planet in the same year, when a colleague of his at the Sproul Observatory published an analysis of the astrometry of GL 793, another nearby red dwarf star. John Hershey had found that both GL 793 and Barnard's Star were wobbling about their proper motion across the sky in much the same way. When one star zigged, so did the other. When one star zagged, so did the other. This could only mean one thing: systematic errors in the Sproul refractor. In retrospect, the spurious zig-zags could be traced to several changes that had been made to the optical system, but evidently had not been completely corrected for in the error terms for the astrometric solutions: a new cast iron cell for the 24-inch lens and new photographic emulsions in 1949, and a lens adjustment in 1957.

After 1973, the strong evidence for Barnard's Star having a gas giant planet began to fade from view and from the textbooks. Astronomers have continued to monitor Barnard's Star for planetary-induced wobbles, as it remains an attractive target. Van de Kamp also continued to pursue Barnard's Star, convinced that some day a planet

Figure 3. Discovery data for the first extrasolar planet around a main sequence star, 51 Pegasi (Mayor & Queloz 1995), detected by measuring the periodic Doppler velocity shift of 51 Pegasi induced by the planetary companion. The solid line shows the Doppler shift expected for a solar-mass star orbited by a 0.6 Jupiter-mass planet on a circular orbit with a semimajor axis of 0.05 AU.

would be found in its grasp. Van de Kamp died in 1995, just before the first reproducible evidence for an extrasolar giant planet was announced to the world.

2.2. 51 Pegasi

The existence of an extrasolar planet can be inferred indirectly, not only by measuring the two-dimensional, astrometric wobble in the plane of the sky, but also by searching for the one-dimensional oscillation of the star to-and-fro along the line-of-sight to the star by measuring the Doppler shift of the star's spectral lines. For a Jupiter-like planet orbiting a solar-type star, the Doppler shift has a semiamplitude of 13 m s_1. Beginning in the late 1970s, a group at the University of British Columbia had pioneered a technique of using a cell of hydrogen fluoride gas in the optical path of the telescope as a source of stable reference lines, which could be used to measure tiny Doppler shifts in stellar spectra. By 1995, however, they had searched for over a decade and had found no unambiguous evidence for planets in the two dozen stars they had been following—their results appeared to place upper limits only on the masses of planets that might exist in orbit around their target stars (Walker et al. 1995). It looked like extrasolar Jupiters might not be very common, contrary to long-standing theoretical and philosophical expectations.

Shortly after the Walker et al. (1995) results appeared, a new claim for the detection of a gas giant planet was announced by a team composed of Michel Mayor and Didier Queloz of the Geneva Observatory. Duquennoy & Mayor (1991) had published the definitive catalog of binary stars in the solar neighborhood, including binaries found by several different methods, but primarily by searching for the Doppler spectroscopic wobble. As a result of this survey, Mayor had a list of roughly 200 nearby solar-type stars that appeared to be single, though a number of them showed evidence for having a very low-mass companion. In 1994, a new spectrometer on the 2 m telescope at the Haute Provence Observatory with a spectral accuracy of ~13 m permitted Mayor to begin a serious spectroscopic search for extrasolar giant planets. By the summer of 1995, Mayor & Queloz had struck oil—the solar-type star 51 Pegasi appeared to be wobbling with an amplitude well above their measurement errors. Following a final, confirmatory observing run in September 1995, Mayor & Queloz (1995) announced their discovery of a ^0.6 Jupitermass planet on a circular orbit (Figure 3), as expected based on Jupiter's low eccentricity (e ~ 0.05) orbit. While the mass and circular orbit of the planet seemed normal, its orbital period was far from the expectations for an extrasolar Jupiter: 4.23 days rather than 12 or so years. This meant that the planet was orbiting its star 100 times closer than Jupiter orbits the sun, at a distance of ^0.05 AU rather than out at 5.2 AU. At that distance from its star, 51 Pegasi's planet would have an atmosphere that was thermally evaporating: it would be a "hot Jupiter."

Was 51 Pegasi's planet real or not? Following the announcement in October 1995, several other groups of astronomers ran to their telescopes to see if they could confirm this audacious claim—and they did. The first team to confirm the reality of 51 Pegasi's planet was that of Geoff Marcy and Paul Butler, using the Shane reflector on Mount Hamilton (Marcy et al. 1997). Several other teams followed in their footsteps and also confirmed the detection of 51 Pegasi's planet. Suddenly the idea of extrasolar planets around solar-type stars did not seem so preposterous anymore.

Marcy and Butler had begun their spectroscopic planet search in the late 1980s, using an iodine cell as a reference source for their spectrometer (since hydrogen fluoride is a deadly gas), and they were soon achieving accuracies of 10 m or better. Given their significant head start on the Swiss group, Marcy and Butler began a frantic effort to reduce their many years of data, allowing them to announce the discovery of two more planets in January, 1996—planets orbiting the solar-type stars 47 Ursae Majoris (Butler & Marcy 1996) and 70 Virginis (Marcy & Butler 1996). The field of extrasolar planets had truly been born. Time Magazine celebrated the event with a cover story breathlessly entitled, "Is Anybody Out There? How the discovery of two planets brings us closer to solving the most profound mystery in the cosmos."

Was this article helpful?

0 0

Post a comment