As of now, four thousand planets have been discovered orbiting other suns. Most of these have not been detected by pointing a telescope and seeing or photographing an actual, visible disc, or dot for that matter. A few may have been. For instance, the star Fomalhaut, also known as α Piscis Austrini, has a nearby dot which is not, however, orbiting in the plane of the star’s apparent discs of planetesimals, which has been seen through a telescope, but it appears that this is actually a cloud of débris, which probably explains why it’s visible from here. Therefore the main techniques use indirect evidence. There are two major methods used at the moment, although another was used in the past.
The most fruitful procedure uses the Doppler Effect. This works like this. A star has a spectrum with lines in it representing the elements which absorb particular wavelengths of light, so it looks like this:
This is the solar spectrum, although many of the lines have been omitted. Incidentally, this means that in theory if you look closely enough at a rainbow, you might expect to see gaps in the colour bands like this, but I imagine some kind of blurring effect takes place which makes it impossible. It’s generally known that if an emergency vehicle rushes past, the sound of the siren drops in pitch as it recedes from the listener. The same thing happens with light, which is why very distant objects in the Universe look redder than they would if they were stationary relative to Earth. The position of these bands is very precise because they’re defined by the energy levels of electrons in atoms, which in turn can be calculated using the principles of quantum mechanics. Therefore, quite minor movements in stars towards or away from us produce measurable alterations in the positions of these lines. And they do shift when planets are involved. Taking Jupiter as an example in our own solar system, it orbits around 5.2 AU (an AU, astronomical unit, is the mean distance of Earth from the Sun) from the Sun and has a mass about a thousandth of the Sun’s. Planets orbit in ellipses one of whose foci is ideally occupied by the centre of a star. However, this is not strictly accurate. In fact both the star and the planet orbit a common centre with the planet’s orbit much larger than the star’s. In the case of Jupiter and the Sun, this centre of gravity is a thousandth of the distance between the two, which is 150 000 kilometres from the centre of the Sun. Since the Sun’s radius is 696 000 kilometres, the furthest point of the photosphere from this point is around 850 000 kilometres away, and takes 11.86 years to move round an approximate circle of 5 340 000 kilometres at a velocity of only fourteen metres per second. This doesn’t sound fast enough to me to be detectable from anywhere or to have a notable influence on the absorption lines in the spectrum, and this is because Jupiter is far out and therefore doesn’t move the Sun around much. But suppose a planet had the same mass as Jupiter and only took a week to orbit the Sun. This would imply a semimajor axis (mean diameter of the orbit) of only twenty-two million kilometres and would lead to the Sun describing a 22 000 kilometre wide “orbit”, moving at around 400 kph. This is quite a bit faster, and shows how the easiest planets to detect by this method are those which are large and close to stars – “Hot Jupiters” as they’re known. This is known as the “radial velocity” method.
Another approach is to use transits, which are like solar eclipses in that they are when planets cross in front of stars. Taking the example of Jupiter again, if it were orbiting at a distance of eleven million kilometres from the Sun it would cross the disc once a week, taking somewhat under a day to do so, and would reduce the light seen from interstellar distances by about one percent, which is easily detectable. Many exoplanets have been picked up like this. The drawback is that the orbit of the planet round its star has to be aligned with ours. It’s no good for a planet whose orbit isn’t almost edge-on because that planet will never occlude its sun from our perspective. This is known as “transit photometry”. Like the radial velocity method, this is biassed towards large planets close to their stars, since they transit more often and cut out more light. One interesting side-effect of this method is that it can also be used to detect moons orbiting these planets, since like the stars in the radial velocity method, they will cause their primaries to describe small elliptical paths which will advance or delay their transit slightly, but because the planets detected are next to their stars they will usually lose the gravitational tug of war with them to retain any moons which might form or be captured.
There was an older technique which turned out to be an embarrassing failure. In the mid-twentieth century into the ’80s, attempts were made to detect the “wiggle” used by the radial velocity method when a star and planet were presumed to orbit at close to a right angle to Earth. Instead of moving in an almost straight line through space, a star and planet will always be slightly wavy. This method was thought to have detected planets in the 61 Cygni, Krüger 60 and 70 Ophiuchi systems, and orbiting Barnard’s Star, Lalande 21185 and ε Eridani. Strong doubts were cast upon the idea that variations that tiny could be detected from this distance, and in the end it turned out that they were caused by the lenses being taken out of telescopes and cleaned at regular intervals before being put back in in slightly different positions. Hence all of this lot went out of the window, which is a shame because the planets concerned were eminently “sensible”: they tended to look like the gas giants in our own solar system. It later turned out that Barnard’s Star does in fact seem to have a planet orbiting it at 0.4 AU which is several times the mass of Earth but probably rocky, if it exists, with a surface temperature likely to be around -170°C, assuming no atmosphere and so forth but still likely to be very cold. 70 Ophiuchi was actually the first system (it’s binary) to have the astrometric method used on it, in 1855, but it turned out that any body at the distance predicted would’ve been unstable. However, this has once been used successfully, to detect a planet seven times the mass of Jupiter orbiting Van Briesbock’s Star, a red dwarf nineteen light years away in the constellation Auriga.
Pulsars can also have planets. Pulsars are basically the remnants of collapsed massive stars not quite big enough to become black holes, and are effectively enormous atomic nuclei made entirely of neutrons inside and about ten kilometres across. They spin very fast, varying from about once every few seconds to about a thousand times a second, and if oriented in the right direction emit radiation from their poles which is detected as pulses from here, hence the name. Their spin is very precise, although it gradually slows and can vary due to star quakes. Although their surface gravity is ludicrously strong, this makes no difference to any of their possible planets because their mass is still almost the same as their parent stars, and they undergo the same perturbation as is used in the radial velocity method. This causes slight delays and advances to the pulses on a regular basis, and allows planets down to about the mass of Mars to be discovered. Unfortunately, because they’re outrageously radioactive and magnetic, there doesn’t seem to be any chance that any of them could support life as human beings can understand it – i.e. chemical and organic. The planet illustrated above has been named Poltergeist and orbits a pulsar designated PSR B1257+12, also known as Lich, along with three other planets. It’s about 2 300 light years away.
I mentioned previously that planets and moons can cause clumping and gaps in discs of matter orbiting their primaries. This happens with Jupiter and the asteroid belt. If you plot a bar chart of the periods of the asteroids in the belt, you get something like this:
If you then draw a diagram of these clumps, this emerges:
These are “rings” of asteroids whose time to orbit the Sun is almost in harmony with Jupiter’s in various ways, such as ⅔ or ½, named after prominent members of the clumps. Due to being regularly exposed to Jupiter’s gravity, they tend to be pulled away from their original orbits to ones which are slightly greater than that period. This is thought to be a factor in the formation of planets in our own Solar System. For instance, Earth orbits the Sun slightly less than a dozen times each Jovian year. Asteroids are far too sparse to be visible as rings from elsewhere in our stellar neighbourhood, but there are other star systems which still have the densely-strewn discs of planetesimals we once had while our own solar system was still formng, and in some of these gaps could be used to detect the locations of planets.
There are various other methods for detecting planets, but most of them suffer from the same problem. They tend to be much better at finding planets which are unlikely to be like ours. The transit and radial velocity methods tend to find the large close planets referred to as “Hot Jupiters”, and there do seem to be a lot of these. The astrometry method has so far only succeeded once and again needs a massive planet to work, as well as one which is quite far out. The pulsar timing method can easily detect Earth-sized planets but not ones which are remotely habitable because they’re close to powerful ionising radiation sources, and the detection of gaps in protoplanetary discs implies that planets are still forming in those systems. The difficulty is in identifying any Earth-like planets. However, the fact that there are a very large number of Hot Jupiters in itself makes the presence of such planets in their systems less likely. It’s possible that these planets started out at a considerable distance from their stars and fell inward. If this happened, they’d be likely to disrupt the orbits of any Earth-like planets further in, and their presence in an inner solar system from the start could also have a similar effect, so their very ubiquity could make the Universe a less friendly place for the likes of us.
However, this perception of the Universe as a place where Earth-like planets are rare could simply be an artifact of the methods used to detect planets elsewhere in the Galaxy. There are other ways to find planets but for several decades the only ones available were guaranteed only to locate those which were completely unsuitable for anything like us to live on. Consequently, for all we know there could be loads of planets like ours out there but we wouldn’t know about them.
I have to some extent caricatured this situation. In fact, although it’s true that none of the methods I mentioned are any good at discovering Earth-like planets, there are others which have been used more recently. For instance, there’s gravitational microlensing. The gravity of stars bends light passing from behind them, and this would be bent in a different way were there an accompanying local planet. This works better for planets viewed from further out from the centre of the Galaxy, because there are more stars near the centre than the edge, and are therefore more likely to be in the constellation of Sagittarius as that’s the direction of the centre of the Galaxy. One example is MOA-2007-BLG-192Lb:
This planet is three thousand light years away and is around 3.3 times our mass. This means that if it has the same density as Earth, it would have a surface gravity only 50% higher than ours and would also be a rocky planet like this one. However, since it orbits a brown dwarf, a star which is not large enough to have fusion reactions and is therefore quite cool, this planet is likely to be frozen solid and may even be considerably larger than that because it’s substantially made of ice. Even so, it’s a somewhat Earth-like planet which has been detected by this method.
To conclude then, there could still be plenty of Earth-like planets out there. Although the systems containing hot Jupiters may well be screwed, on the whole it’s currently diffcult to see planets like ours and the principle of mediocrity, which is that we should assume our situation is about average in the absence of further information as opposed to being special, suggests that we should presume that there are plenty of Earth-like planets until we can demonstrate otherwise. Therefore the currently rather hostile view we have of other solar systems is probably the result of how we look for planets, because we are selecting the “easy” cases rather than the ones which are most like our own situation. So far as we know, there are still likely to be many Earths out there and many solar systems like ours.
A Box Of Chocolates 