The Development of Exoplanet Detection Methods, Techniques and Ideas Before the Kepler Space Telescope
In 1584, Giordano Bruno wrote about “numberless Earths circling around their suns, no worse and no less inhabited than this globe of ours.” Written at a time when Copernicus’ work was under attack by many, he was eventually a victim of the Inquisition but a pioneer in free thought (Finley 90). Now Gaia, MOST, SWEEPS, COROT, EPOXI, and Kepler are just some of the major efforts past and present in the hunt for exoplanets. We almost take those special solar systems and their wonderful complexities for granted, but until 1992 there were no confirmed planets outside our own solar system. But like many topics in science, the ideas that eventually led to the discovery were just as interesting as the finding itself, and perhaps more. That is a matter of personal preference though. Read the facts and decide for yourself.
In 1779 Herschel discovered the binary star system 70 Ophiuchi and began to take frequent measurements in an attempt to extrapolate its orbit, but to no avail. Jump to 1855 and the work of W.S. Jacob. He noted that years of observational data failed to help scientists predict the orbit of the binary star system, with a seemingly periodic nature as to the discrepancy in the distances and angles measured. Sometimes they would be larger than actual and other times they would be less than expected, but it would flip back and forth. Rather than go and blame gravity which worked great, Jacob instead proposes a planet that would be small enough to cause many of the errors to be decreased in nature (Jacob 228-9).
In the late 1890’s, T.J.J. See followed up on this and in 1896 filled a report with The Astronomical Society. He too noticed the periodic nature of the errors and computed a chart as well, having data all the way from when Herschel discovered it. He postulates that if the companion star were about the distance from the central star as the average distance Neptune and Uranus are from our sun, then the hidden planet would be about Mars distance from the central star. He goes on to show how the hidden planet causes the seemingly sinusoidal nature of the outer companion, as seen in the figure. Furthermore, he adds that even though Jacobs and even Herschel found no traces of a planet in 70 Ophiulchi, See was confident that with the new telescopes coming out it was just a matter of time before the matter was settled (See 17-23).
And it was, just less so in favor of a planet. However, it did not out-right eliminate the possibility of one residing there. In 1943, Dirk Reuyl and Erik Holmberg noted after looking at all the data how the fluctuations of the system varied by 6-36 years, a huge spread. A colleague of theirs, Strand, observed from 1915-1922 and from 1931-1935 using high-precision instruments in an effort to resolve this dilemma. Using grating plates as well as parallax readings, the errors from the past were greatly reduced and it was shown that if a planet were to exist, it would be 0.01 solar masses in size, over 10 times the size of Jupiter with an distance of 6-7 AU from the central star (Holmberg 41).
So, is there a planet around 70 Ophiuchi or not? The answer is not, for based on far away the binary system is, no changes of 0.01 seconds of arc were seen later in the 20th century (for perspective, the Moon is about 1800 seconds of arc across). If a planet were in the system, then changes of 0.04 seconds of arc would have been seen at minimum, which never happened. As embarrassing as it may seem, the 19th century astronomers may have had too primitive tools at their hands that caused bad data. But we must remember that any findings of any time are subject to revision. That is science, and it happened here. But as a redeeming quality to those pioneers, W.D. Heintz postulates that an object passed by the system recently and disturbed the normal orbits of the objects, therefore leading to the readings scientists have found over the years (Heintz 140-1).
61 Cygni, Barnard’s Star, and Other False Positives
As the 70 Ophiuchi situation was growing, other scientists saw it as a possible template to explain other anomalies seen in deep-space objects and their orbits. In 1943, the same Strand that helped in observations for 70 Ophiuchi concluded that 61 Cygni has a planet with a mass of 1/60 of the sun or roughly 16 times greater than Jupiter, and it orbits at a distance of 0.7 AU from one of the stars (Strand 29, 31). A paper from 1969 showed that Barnard’s Star had not one but two planets orbiting it, one with a period of 12 years and mass a bit more than Jupiter and the other a period of 26 years with a mass slightly less than Jupiter. Both supposedly orbited in opposite directions of one another (Van De Kamp 758-9). Both were eventually shown to be not only telescopic errors but also because of the wide range of other values different scientists got for the parameters of the planets (Heintz 932-3).
Ironically, one star that was thought to have a companion actually did, just not a planet. Sirius was noted to have some irregularities in its orbit as noted by Bessel in 1844 and by C.A.F. Peters in 1850. But by 1862, the mystery of the orbit was solved. Alvan Clark pointed his new 18 inch objective lens telescope at the star and noted that a faint speck was close to it. Clark had just discovered the 8th magnitude companion, now known as Sirius B, to Sirius A (and at 1/10,000 the brightness, it was no wonder it went hidden for so many years). In 1895 a similar discovery was made of Procyon, another star that was suspected to have a planet. Its star companion was a faint 13th magnitude star found by Schaeberle using the Lick Observatory’s 36-inch telescope (Pannekoek 434).
Other possible planets seemed to pop up in other binary star systems over the ensuing years. However, after 1977 most were put to rest as either a systematic error, faults in reasoning (such as parallax considerations and assumed centers of mass), or simply bad data taken with inadequate instruments. This was especially the case for Sproul Observatory, which claimed to spot wobbles from many stars only to find that constant calibrations of the equipment were giving false readings. A partial list of other systems that were debunked because of new measurements removing the supposed motion of the host star is listed below (Heintz 931-3, Finley 93).
- Iota Cassiopeiae
- Epsilon Eridani
- Zeta Hericulis
- Mu Draconis
- ADS 11006
- ADS 11632
- ADS 16185
The Ideas Become Focused
So why mention so many mistakes about the search for exoplanets? Let me paraphrase something the Mythbusters are fond of saying: failure is not only an option, it can be a learning tool. Yes, those scientists of the past were mistaken in their findings but the ideas behind them were powerful. They looked at orbital shifts trying to see the gravitational pull of the planets, something that many current exoplanet telescopes do. Ironically enough, the masses as well as the distances from the central stars were also accurate to what is considered the main type of exoplanets: hot Jupiters. The signs were pointing in the right direction, but not the techniques.
By 1981, many scientists felt that within 10 years solid evidence of exoplanets would be found, a very prophetic stance as the first confirmed planet was found in 1992. The main type of planet they felt would be found would be gas giants like Saturn and Jupiter, with a few rocky planets like Earth also. Again, very good insight into the situation as it would eventually play out with the aforementioned hot Jupiters. Scientists at the time began to construct instruments that would aid them in their hunt for these systems, which could shed light on how our solar system formed (Finley 90).
The big reason why the 1980’s was more prone to take the search for exoplanets serious was the advancement of electronics. It was made clear that optics needed a boost if any headway was to be made. After all, look at how many mistakes scientists of the past had made as they tried to measure microseconds of change. Humans are fallible, especially their eyesight. So with the improvements in technology it was possible to not rely just on reflected light from a telescope but some more insightful means.
Many of the methods involve making use of the barycenter of a system, which is where the center of mass is for orbiting bodies. Most barycenters are within the central object, like the Sun, so we have a hard time seeing it orbit about it. Pluto’s barycenter happens to be outside of the dwarf planet because it has a companion object, which is comparable in mass to it. As objects orbit about the barycenter, they seem to wobble when one looks at them edge-on due to the radial velocity along the radius from the orbital center. For far away objects, this wobble would be difficult at best to see. How hard? If a star had a Jupiter or Saturn-like planet orbiting it, someone viewing that system from 30 light-years would see a wobble whose net motion would be 0.0005 seconds of arc. For the 80s this was 5-10 times smaller than current instruments could measure, much less photographic plates of antiquity. They required a long exposure, which would remove the precision needed to spot an accurate wobble (Ibid).
Multichannel Astrometric Photometer, or MAP
Enter Dr. George Gatewood of the Allegheny Observatory. During the summer of 1981 he came up with the idea and technology of a Multichannel Astrometric Photometer, or MAP. This instrument, initially attached to the Observatory’s 30-inch refractor, made use of photoelectric detectors in a new way. 12-inch fiber optic cables had one end placed as a bundle at a telescope’s focal point and the other end feeding the light to a photometer. Along with a Ronch grating of about 4 lines per millimeter placed parallel to the focal plane, allows light to be both blocked and enter the detector. But why would we want to limit the light? Isn’t that the valuable intel we desire? (Finley 90, 93)
As it turns out, the Ronch grating doesn’t prevent the entire star from being obscured and it can move back and forth. This allows different portions of the light from the star to enter the detector separately. This is why it is a multichannel detector, because it takes input of an object from several close positions and layers them. In fact, the device can be used to find the distance between two stars because of that grating. Scientists would just need to examine the phase difference of the light due to the movement of the grating (Finley 90).
The MAP technique has several advantages over the traditional photographic plates. First, it receives the light as an electronic signal, allowing for higher precision. And brightness, which could wreck a plate if overexposed, doesn’t affect the signal MAP records. Computers could resolve the data to within 0.001 arc seconds, but if MAP were to get into space then it could achieve a precision of one-millionth of an arc second. Even better, scientists can average the results for an even better sense of an accurate result. At the time of the Finley article, Gatewood felt it would be 12 years before any Jupiter system would be found, basing his claim on the orbital period of the gas giant (Finley 93, 95).
Of course, a few unsaid topics arose during all the development of MAP. One was the use of the radius velocity to measure spectroscopic shifts in the light spectrum. Like the Doppler effect of sound, light too can be compressed and stretched as an object moves towards and away from you. If it is coming towards you then the light spectrum will be shifted blue but if the object is receding then a shift to the red will occur. The first mention of using this technique for planet hunting was in 1952 by Otto Struve. By the 1980s, scientists were able to measure radial velocities to within 1 kilometer per second but some were even measured to within 50 meters per second! (Finley 95, Struve)
That being said, Jupiter and Saturn have radial velocities between 10-13 meters per second. Scientists knew that new tech would need to be developed if such subtle shifts were to be seen. At the time, prisms were the best choice to break up the spectrum, which was then recorded onto film for later study. However, atmospheric smearing and instrument instability would frequently plague results. What could help prevent this? Fiber optics once again to the rescue. Advances in the 80s made them larger as well as more efficient at both collecting light, focusing it, and transmitting it along the entire length of the cable. And the best part is you don’t need to go into space because the cables can refine the signal so that the shift can be discerned, especially when used in combination with a MAP (Finley 95).
Interestingly, the other untouched topic was the use of the electronics to measure the signal of the star. More specifically, how much light we see from the star as a planet transits across the face of it. A noticeable dip would occur in the brightness and if periodic it could indicate a possible planet. Mr. Struve was once again an early proponent of this method in 1952. In 1984 William Borucki, the man behind the Kepler Space Telescope, held a conference in the hopes of getting ideas started as to how best accomplish this. The best method considered at the time was a silicon diode detector, which would take a photon that hit it and convert it into an electrical signal. Now with a digital value for the star, it would be easy to see if less light was coming in. The downside to these detectors was that each could be used for just a single star. You would need many to accomplish even a small survey a sky, so the idea while promising was deemed infeasible at the time. Eventually, CCD’s would save the day (Folger, Struve).
A Promising Start
Scientist sure did try many different techniques to find planets. Yes, many of them were misguided but the effort had to be extended as advances were made. And they did prove to be worthwhile. Scientists used many of these ideas in the eventual methods that are currently used to hunt for planets beyond our solar system. Sometimes it just takes a little bit of a step in any direction.
Finley, David. “The Search for Extrasolar Planets.” Astronomy Dec. 1981: 90, 93, 95. Print.
Folger, Tim. "The Planet Boom." Discover, May 2011: 30-39. Print.
Heintz, W.D. “Reexamination of Suspected Unresolved Binaries.” The Astrophysical Journal 15 Mar. 1978. Print
- - - .“The Binary Star 70 Ophiuchi Revisited.” Royal Astronomical Society Jan. 4, 1988: 140-1. Print.
Holmberg, Erik and Dirk Reuyl. “On The Existence of a Third Component in the System 70 Ophiuchi.” The Astronomical Journal 1943: 41. Print.
Jacob, W.S. “On the Theory of the Binary Star 70 Ophiuchi.” Royal Astronomical Society 1855: 228-9. Print.
Pannekoek, A. A History of Astronomy. Barnes and Noble Inc., New York 1961: 434. Print.
See, T.J.J. “Researches on the Orbit of F.70 Ophiuchi, and on a Periodic Perturbation in the Motion of the System Arising from the Action of an Unseen Body.” The Astronomical Journal 09 Jan. 1896: 17-23. Print.
Strand. “61 Cygni as a Triple System.” The Astronomical Society Feb 1943: 29, 31. Print.
Struve, Otto. “Proposal for a Project of High-Precision Stellar Radial Velocity Work.” The Observatory Oct. 1952: 199-200. Print.
Van De Kamp, Peter. “Alternate Dynamic Analysis of Barnard’s Star.” The Astronomical Journal 12 May 1969: 758-9. Print.
Questions & Answers
© 2015 Leonard Kelley