How soon could extrasolar planets have been discovered? Options

[Mongo]

Extrasolar pulsar planets were first discovered in 1992, and the first planets around main-sequence stars in 1995. However, it seems clear that main-sequence exoplanets could have been found much earlier, except for the fact that nobody actually looked for them in a way that would have succeeded.

The standard assumption was that all, or almost all, planetary systems followed the architecture of our own system -- a few terrestrial planets at distances ranging from a few tenths of an AU to a few AUs, with larger gas- and ice-giants further out. This would result in systems that are basically impossible to detect with the technology of the time, except perhaps for long-term (years to decades) astrometric studies of nearby dwarf stars. It was known that planetary transits would be easily detectable, but they were thought to be so unlikely due to the geometry, as well as being infrequent in the few existing transiting systems, that they were not worth the resources spent searching for them. Radial velocity searches were considered to be out of the question, with stellar RV shifts ranging up to a few m/s at best, well below the best sensitivity of the spectroscopes of the time.

In fact, "Hot Jupiters" are common, with RV swings measured in the tens or hundreds of m/s, and transiting planets are so common and easy to spot that many current search programs use telescopes no larger than typical amateur telescopes. Today's detectors are better, of course, but that could be compensated for with a slightly larger telescope.

I assume that a planet search program consisting of an all-sky transit search with one or more 30cm-class telescopes would produce several dozen candidates (with at least three observed transits each) orbiting fairly visually bright stars within a few months. Each of those would be examined spectroscopically, and each would be soon found to have RV swings of 50-200 m/s or more. With the existing spectroscopic classification of the primary stars, this would result in fairly accurate masses and radii for each of these planets, as well as expected daytime temperatures. It would be easy to calculate the frequency of "Hot Jupiters" around main-sequence stars (around 1%), and the "inflated Jupiter" effect should be apparent in the data.

So assuming that some astronomer manages to obtain funding and telescope time to conduct such a program, when was the astronomical photometer and spectroscope technology advanced enough for this to have succeeded? The 1960s? 1970s? Surely by the 1980s, at the latest, which seems very late to me.

[JCG]

I'm one of the old-timers who was involved in several surveys looking for extrasolar planets in the 1980s and 90s. Hindsight is always 20/20 in any prospect as it is here, no offense intended, it's a fact I deal with, too. There are couple of factors you need to consider when speculating on why didn't "they" just (fill in the blank) back the day. In the case of extrasolar planets, there are a couple of practical factors to consider: CCDs were not as sensitive and low-noise as they are today and were nowhere near as easy to use. The first forays into extrasolar planets using CCDs used a 512x512 device (the so-called Galileo CCD, the offspring of the CCD developed at JPL for use on the imaging system for the Galileo spacecraft) that was one-of-a-kind and jealously fought over by any and all astronomers for their special subject whether planetary astronomy was used for multiband imaging, spectroscopy, planetary, stellar, galactic, you name it research. No telescope committee would have ever considered using that or any existing CCD for looking for planetary transits, even though the technique was being studied theoretically at the time. Remember also that the frequency of extrasolar planets was "0" so no seriously-believed calculation could be performed to argue for using a precious science-grade CCD for planet searches as opposed to quasar studies, galaxy studies, supernova searches, etc. That calculation got pulled up by its bootstraps, finally (the comment about RV observations is correct).

Another reason is "telescope" time. Back in the day there were not a lot of 1-meter class telescopes to be handed out. And, there was very little money to build and support them (operationally). And each telescope got caught in the same problem of what is the best use of the resource, i.e., what are the most pressing scientific problems that have a high probability of being answered in a minimal amount of telescope time.

The first searches for extrasolar planets using CCDs starting with the imaging of the circumstellar disk of Beta Pictoris by Smith & Terille using a stellar coronograph and the Galileo CCD (there was only one of these state-of-the-art devices at the time). I and a few others confirmed their results by building stellar coronographs from spare parts and other existing intruments. Although simple in concept, the coronograph is a difficult beast to use if one wants quantitative results - there is much, much more to an observation than just theoretical SNR. We were allowed only a few hours on a 2.2-m telescope (UH/Mauna Kea) with the Galileo CCD hoping it would not be cloudy. Fortunately, we were successful. Subsequent dedicated searches for "really hot Jupiters"around red dwarves using the CCD/coronograph technique (imaging near 900 nm where CCD sensitivity starts falling off and "really hot Jupiters" might show some detectable emission in the NIR) were allocated time because we were able to show quantitatively how much time we needed to reach scientifically useful results. Even so, we had to fight ferociously for just a few nights of time. There are just too many other really neat scientific problems to address given the number of telescopes, the number of CCDs available and the chances of cloudy nights. We found a few red dwarf binaries but no Jupiters.

Another issue that limited observations at the time is that of the software and algorithms needed to rapidly process the data (along with all the calibration, quality control, etc., that goes along with processing). No one ever looked forward to reading hundreds of 9-track tapes into the communal computer system for rudimentary image processing algorithms.

That said, one type of search that was very successful was the search for the Kuiper Belt objects and beyond by Dave Jewett. That's because he could calculate the expected discovery rate in quantitative terms and the telescope time needed for even a null result (which was scientifically interesting) was acceptable. So lots of telescope time was dedicated to that search to the consternation of the galactic astronomers but the great joy of the planetary astronomers.

The advent of relatively inexpensive highly, sensitive, large format scientific grade CCDs (and fast computers) has changed the whole search strategy for a number of problems. I remember when Gene and Carolyn Shoemaker (and Glo Helin) performed their NEO asteroid searches with film(!). And when Tom Gehrels started using a CCD with TDI to look for them his discovery numbers were far too low at first because the CCD he was using was very small (miniscule!) in format when compared with what we have today, but the best he could get. A modern cell phone camera (CMOS) array could do better! Consider what the PanSTARRS 1.4 Gpixel (yes, gigapixel) camera can do compared to the original Shoemaker searches. And the new ATLAS system will have 8 cameras of 100s Mpx on 8 small (0.5-m) telescopes to do automatic and autonomous searches for "city-busters" among the NEOs. All of this could be done with Schmidt cameras, film plates, and lots of graduate students, but nowhere(!!!) near as efficiently.

The history of looking for transits of planets is analogous. It wasn't done in the past because the time to do it had not yet come. It had to be pulled up by its bootstraps with every other aspect of observational astronomy (with really good arguments for "me first") trying to cut the straps for the available resources. Fortunately, the pioneers never gave up and their arguments eventually won the day.

P.S. I visit this site every so often and don't comment much, but this particular conversation caught my eye and I thought I could help with some perspective. Keep up the good work. JCG