D-23: The Hunt for Exoplanets

The following is the first guest post on the series, written by a person called Murmeldjuret on the Stellaris forums. It has been reproduced here with their permission.


Today, by request it will be on how to detect exoplanets. I have made it rather exhaustive, so get your coffee ready.
For those of you wanting to learn of the things that can and will want to kill you, there will be moredeadly friendly lightshows tomorrow.

The main problem with detecting planets is that they are very small compared to a star. They don’t give off light, and not being on fire tends to leave them cold. In fact, we have only directly seen a very small set of exoplanets. Wikipedia lists 20, all of them larger than jupiter. Nasa lists 33. All the other thousands of planets we have “detected” we have actually never seen. They have all been inferred from observing the light of the star they orbit. Inferred is mathspeak for qualified guessing, meaning the changes in starlight is most likely from a planet, but we are not 100% sure. Some of them we are really certain about, some like the previously described Gliese581d are less certain and are considered disproven one year to be proven the next.

There are two primary methods to finding planets in other solar systems. Planetary transit and radial velocity also known as doppler shift. Both require the planet to rotate over or nearly over the star, meaning we can only detect planets around stars whose planetary disk aligns with the star and Earth. We will not be able to detect planets from the majority of stars out there. If a star has not yet been proven to have a planet it might still have one, only we are unable to detect it. I should point out that there are methods to detect planets by two other means but we have seen only a dozen this way.

Planetary Transit:

If the planetary disk intersects a star, for parts of that planet’s year it will lie between us and the star. Much like a solar eclipse from our moon blocks sunlight, only much fainter. By measuring the light from the star over a long time we can see how it varies, and if it has any periodic and regular variation. An ideal planet transition is shown in the diagram below, with one period being a year for the detected planet, the y-axis is total luminosity from the star:

  • A: Planet is obscured by star, only the star is visible. Useful as reference point.
  • B: Planet is reflecting sunlight, we could judge albedo (reflectiveness) of planet if we trust the size and orbital distance of the planet. This is useful since the albedo of different wavelengths is dependent on absorption and absorption is dependent on what elements it is composed of.
  • C: Planet is starting to block sunlight, and depending on how sudden the drop in sunlight is we can see if there is an atmosphere. Given enough spectral and spatial resolution we could determine what elements the atmosphere has, especially watervapor. With a sufficiently precise spectral resolution we can also see if it has molecular oxygen (O2) and carbon dioxide. Right now this is only viable for gas giants, not so for rocky planets.
  • D: Planet is obscuring sunlight by a very fine amount. We can also estimate how much heat the planet captures which is a strong indicator of how much atmosphere it has.

All of these combined with total rotation time and solar mass/energy gives us a equation system that can reasonably well give us size, mass, and orbit of planet. This is how we today say how large it is, how far from the star it is, and if it is in the habitable zone. We can also determine if it has an atmosphere and how much atmosphere it has when we get more detailed equipment. With even better equipment we can also detect oxygen and carbon dioxide, and therefore if it has breathable atmosphere and if carbon material for life exists. All without ever seeing the planet.

Radial Doppler:

Most of you have probably heard of the doppler effect. It is when an object moving towards you increases in frequency (gets blueshifted), and what makes things moving away from you get redshifted. It is why police sirens zooming past sound high pitched while approaching and low pitched after they passed. The same is true of light, and since almost everything distant in the universe is redshifted the universe must be expanding.

Another physical phenomena is that two objects in orbit always rotate around a common centre of gravity. It is very noticeable for binary starsystems, less so for Sun and Earth. But this means that every star with a planet orbits a commoncentre of gravity that is not quite the centre of the star. Therefore every star spins in a very small orbit, barely noticeable from just looking at it.

In the event that the planet is on a disk close to being aligned with us and its star, then we can use radial doppler to see the planet(s). This means that the star is at two points going perpendicular to us (not towards us or away from us) and at one point it is moving towards us, and at another moving away from us. This will in turn by a very small degree dopplershift the wavelength of the star. The star and its spectra are wobbling from the planets gravity on the star. This is periodic with the year of the planet and scales with the relative mass of the planet and its star. Since we can know the mass of the star we can also estimate the orbit and mass of the planet. This can unlike the transition method detect planets that are not transitioning directly over the star. Though we will not find the exact mass unless we can compensate by using other means. The European HARPS installation can determine the speed of a stars rotation within an error of 1m/s.

The biggest drawback of both systems is that both are dependent on the relative mass and size of the planet vs the star.

  • Detecting planets around red dwarves is easy, detecting planets around blue supergiants is nearly impossible.
  • Detecting small planets the size of Earth is very hard, detecting gas giants is easier.
  • Detecting planets far away from the star is very hard, detecting a gas giant in close orbit is easy.

See the chart below of EU’s database to see the mass of detected planets, in jupiter masses. And another below with both mass and distance to their star.

So what does this mean?
We have found over two thousand planets so far. Almost all of them in the past two years. Averagingat almost 3 planets per day. We will detect about 70 new exoplanets from today to release.

This has challenged previous assumptions of planet frequency. Previously, planets were thought to only orbit main sequence stars, were uncommon, and multiplanet starsystem are exotic. We have had to revise our understanding of the universe. The estimation is now that there is on average more than 1 planet per star. Meaning that theMilky Way contains something on the order of 100-400 billion planets, probably more.
The local Virgo Supercluster might contain tens of thousands of trillions of planets. (I just took a reasonable number, do not quote me.) If only one planet in a trillion 1/1000000000000 contains life, there are still over ten thousand life bearing planets in our local supercluster. And our local supercluster is one of many millions of detected superclusters.

If the chance that a planet is inhabitable is 1 in a billion, the Milky Way contains 100-400+ habitable planets, reminescent in order of magnitude to a certain spacegame…

In short, there be planets out there.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s