Hunting exoplanets using the transit method

6 min readOct 10, 2019

--

With this post, I would like to offer some insights into how I found exoplanets using the transit method. If you are interested in other methods, I recommend reading the following two posts:

Planet Hunters

At Planet Hunter, you can help scientists find exoplanets by analyzing data and marking things of interest, which are ultimately looked at by scientists to verify the exoplanet, or whatever else it might be.

I used to analyze stellar light curves on Planet Hunters, back when users still had a lot of freedom when it comes to inspecting the data, which made it a lot more fun to participate as a citizen scientist. Along the way, I have analyzed over 4,000 light curves from the Kepler Space Mission, and thus I have seen various interesting things.

The transit method

When a planet passes in front of its host star, a tiny percentage of the light from the star is blocked by the planet, which measures as a dip in the light curve. Such an event where the star’s light is blocked by something orbiting it is called a transit.

This dimming, depending on the distance of the planet to its host star, is around 0.008% for an Earth-sized planet around a Sun-like star, and a significant 1.7% for HD 209458 b, which is 1.35 times the size of Jupiter (and 15.13 times the size of Earth).

This dimming is picked-up by photometric sensors of ground-based telescopes or space-based telescopes like Kepler. So let’s have a look at a few light curves I came across.

Light curves

Below is a fairly typical light curve of a giant star (SPH22295814) with a radius 8.4 times greater than the Sun. Large stars are volatile, and so they tend to show an irregular and noisy light curve. In my experience light curves of such stars are not very conducive to finding exoplanets, as the noise tends to camouflage indications of any transit events.

Type of star: Irregular variable giant
Radius: 8.4 R☉
Apparent visual magnitude: 10.9
Spectral class: K
Temperature: 4,780 K

Transit light curves

What we are looking for is a significant dip in the light curve. The length of this dip indicates the amount of time it took for the planet to transit; the depth of the dip indicates the amount of stellar light blocked by the planet. Below you can see the transit light curves of various exoplanets.

These light curves of Kepler’s first five planet discoveries show not only drop in star brightness as the planet transits the star, but an indication of the planet’s inclination — how far from the center the planet is passing across the star. (Image credit: *NASA/Kepler Mission*)

Major transit events

The image below might excite you, because there are some very obvious transit events. Additionally, we can see a high degree of periodicity, which is good because planets obviously continue orbiting their star, so you would expect to see a repeating pattern of dips. As such, a planet is verified by multiple periodic transit events.

Type of star: Quiet dwarf
Radius: 1.1 R☉
Apparent visual magnitude: 13.3
Spectral class: G
Temperature: 5,603 K

But I’m afraid this is not a planet transiting. Rather, this light curve is very characteristic of an eclipsing binary, which is when two stars eclipse each other as they orbit each other. Hence also you can see two distinct depths to the dips in the light curve. In the image below you can see more clearly what is going on with the light curve above, where the deeper dip shows the blocking of light when the smaller star is behind the larger star, and the more shallow dip shows when the smaller star is in front of the larger star.

Eclipsing binaries

Now you can probably guess what the light curves below show as well.

Indeed, eclipsing binaries again. When you are hunting for planets, it can be disappointing when you analyze a light of light curves and find nothing of interest. But finding eclipsing binaries I always find exciting. Obviously not nearly as exciting as finding exoplanets, but all the same, you don’t encounter them often, and they do look beautiful. I think it’s amazing how the orbits of two stars, their light output, and how much light they block from each other are encoded into such a simple visualization.

Type of star: Quiet dwarf
Radius: 1.4 R☉
Apparent visual magnitude: 13.2
Spectral class: G
Temperature: 5,683 K

The image below shows another eclipsing binary, but this time the light shows greater variability, which contributes to the depth of the dips, thus giving the impression that these transits may be from different objects. The fact that the light output is so variable brings into question whether the dips are indicative of transit events, or starspots.

To verify that indeed they are transits from planets, one would have to analyze the other quarters of the data (currently no longer possible on Planet Hunters) to see if the suspected transit pattern repeats.

Exoplanets

Now we get to the exciting part. In the image below, you can see three distinct dips of different depths.

Type of star: Quiet dwarf
Radius: 1.0 R☉
Apparent visual magnitude: 14.3
Spectral class: F
Temperature: 6,221 K

Upon seeing this, I looked at the other quarters of data and wrote down all the days at which transit events occurred. I then looked for the periodicity of all three transits, and identified the following planets:

Planet 1 transits: 3.9, 54.7, 105.5, 156.2, 207.0, 257.8, 308.7, 358.9
Planet 2 transits: 33.0, 70.6, 108.2, 145.7, 183.3, 220.9, 258.5, 296.1, 333.2, 370.8
Planet 3 transits: 14.1, 89.4, 164.6, 239.7, 314.4, 389.6

I found the periodicity of these planets to be:

Planet 1: 50.8 day orbit
Planet 2: 37.6 day orbit
Planet 3: 75.3 day orbit

There were a few transits I could not make sense of given the periodicity observed: 202.1, 277.3, 352.0

Still, it’s wonderful to be able to deduce the existence of exoplanets as well as the time it takes the planets to orbit their host star, just from a light curve.

Here are the light curves of a few other exoplanets I found:

Variable stars

Let me end this post by showing some other peculiarities found. In the images below you can see the light curves of variable stars with very short-period variability in light output, of the order of 0.1 days or even shorter. This speed in variability conflicts with the shutter speed of the sensors, as Kepler is unable to resolve such rapid variations.[1] As a result, you get these beautiful patterns: