Why did it take so long to discover exoplanets? Mostly because the signals by which we can detect them are subtle. Even Jupiter is 1000 times less massive than the Sun and it will induce a very small reflex motion on the Sun as it orbits it or eclipse one percent of its light. These are all subtle effects. Currently, three major methods are used to detect exoplanets, and each of them have their issues and selection effects and limitations. The Doppler method was the workhorse for the first decade or so of exoplanet detection, and it's still generating new discoveries every week. In the Doppler method, a giant planet or even a smaller planet pulls the star that it orbits about a common center of gravity, and induces a reflex motion. That reflex motion is not directly visible, but the wobble exerted on the star can manifest as a doppler shift as the star wobbles towards us and away from us as the exoplanet orbits. That effect is only visible if the orientation of the orbit is near the plane of the sky, and essentially, the Doppler method is giving us the mass of the planet multiplied by the sine of the inclination angle. If the orbital system is in the plane of the sky, no doppler shift is present and we cannot detect it. On average, this means that we detect only half the signal for a system at random orientation, but we can still understand the statistical properties. The eclipse or transit method has been enormously successful recently in particular from the Doppler satellite launched by NASA a few years ago. In this method, the system has to be essentially orbiting in the equatorial plane coming towards us or away from us, and that means that it can pass in front of the star and eclipse slightly the star's light just for a short while. But it's a repeatable signal and so multiple transits will confirm that planet exists. From space, the stability photometry means that very subtle eclipses and transits can be detected far less than a percent. This is what has allowed earth mass planets to be detected for the first time. The eclipse method is essentially a statistical technique. Since many systems will not be oriented suitably to have an eclipse, many stars must be observed either sequentially or the same time to find the tiny fraction that at any given time well be able to show an eclipse. The third method is actually the most direct and most obvious, imaging the exoplanet. It also turns out to be the hardest. For something like a Jupiter, one part in a 100 million of the star's light is reflected. For an Earth, it's one part in a billion. This reflected light is seen in the presence of the glare of the much brighter star at a very small angle from that star at a large distance. This makes it extremely difficult to tease out the weak reflected light from the exoplanet. It's fairly easy to calculate the activity of different planets and with just a few numbers. Basically, we need to know the mass and size of the planet compared to the star that it orbits and its distance from the star. All the various techniques scale in different ways with distance, size, and mass. For direct detection, the inverse square law is the enemy. If the distance of the planet from the star doubles, the amount of reflected light goes down by the square of that number or a factor of four. This means more distant planets are much harder to detect even though they separate at larger angles from the parent star. The analogy here is trying to see a firefly sitting right next to the glare of a stadium floodlight. It's an extremely difficult thing to do. In practice, it requires telescopes with adaptive optics where the response of the telescope to a point source can be beautifully modeled so that the starlight can be exactly subtracted out, leaving the faint residual reflected light from the exoplanet. This is hard enough to do that it was first done recently in 2008. The method also benefits from moving to longer wavelengths. Stars like the Sun have their radiation declining as you move into the near infrared because their peak radiation is in the visible regime whereas planets have their peak radiation in the infrared because they're much cooler objects. So the contrast between an exoplanet and a star is much better in the infrared than the optical regime, sometimes by an order of magnitude or a factor of 20. Exoplanets have even been discovered serendipitously, in particular, in the Hubble Space Telescope archive. One beautiful example in the formal hub system was an exoplanet seen twice moving through the fringes leftover after subtracting the bright starlight, and extremely subtle signal only noticeable because two different times were showing parts of the orbit of the exoplanet. It's estimated that 50 to 100 exoplanets may be sitting in the HST Archive in data that was taken for completely different purposes. Researchers are currently going through it painstakingly seeing if they can tease out exoplanets in images that have been made of other nearby bright stars. After the first successes, people have gone on to measure actual solar systems with the imaging technique. In one case, three planets has been observed and another four, all in one image. The advantage of the imaging technique is that, of course, data can be taken sequentially to show the orbits of those planets, and infer their masses. In each of these systems, imaged in their infrared, infrared radiation is more than would be expected from the sunlight creating the temperature for that exoplanet. It appears that the exoplanet has its own internal heat source, and they have formed very recently and therefore is cooling. Seeing is believing, but only a couple of dozen exoplanets have ever been imaged. By far, the majority of the thousands of exoplanets found have been detected by more indirect methods, in particular the Doppler method. Any orbiting system, the planet and the star orbit a common center of gravity. So there's a small reflex motion caused by the planet on the star. That angular wobble is not detectable with ground-based telescopes. It might however be in the future detectable by space-based observatories. For the moment, ground-based telescopes can fairly easily detect that wobble by the Doppler shift of that induced motion on the star. This indeed was the way that most exoplanets were discovered before 2010. With the launch of the Doppler satellite, transits or eclipses took over as the primary way to find exoplanets, and most of the large numbers being found in the last few years are coming from the Doppler mission. There are three primary ways to detect exoplanets. The most obvious is to make an image. But there's a turns out to be the hardest, and has only succeeded a few dozen times. Even a giant exoplanet reflects 100 million times less light than the star that it orbits emits. So that star has to be beautifully modeled and subtracted out before the residual reflected light is visible. The imaging method works best for cold planets fairly far out in their orbits, and its works best done at infrared wavelengths where the contrast is improved. The more course for detecting exoplanets until 2010 was the Doppler method, and indirect method whereby the wobble, on the parent star, induced by the giant planet can be detected as a periodic Doppler shift. Since 2010, the Doppler satellite has produced hundreds and, in fact, thousands of new exoplanet detections by the transit or eclipse method. This method, however, only works when the orientation of the orbit is almost in the equatorial plane, such that the planet approaches us, and goes directly away from us. The Doppler effect can detect almost every inclination except when the orbit is in the plane of the sky.
