Difference between eclipsing binaries and exoplanets

[A Alex P]

Kepler space mission has discovered thousands of exo-planet candidates. Why are they just candidates ? Why followup ground based study is required ? Why is Kepler unable to confirm them ?
Second thing, how are scientists going to differentiate between eclipsing binaries mimicking to be exoplanets.

 

[lomidrevo]

Second thing, how are scientists going to differentiate between eclipsing binaries mimicking to be exoplanets.

By a careful analysis of an eclipsing spectroscopic binary, mass, radius and effective temperature (or at least the ratio of temperatures) of both members can be estimated. I guess this is sufficient to be able to determine whether the observed system is a star binary or an exoplanet orbiting a star.

 

[Bandersnatch]

An eclipsing binary produces two equally spaced dips in the light curve, instead of one, as both objects are luminous.

 

[A Alex P]

An eclipsing binary produces two equally spaced dips in the light curve, instead of one, as both objects are luminous.

But we can not differentiate using light curve only. Shape of light curve and space between dips will depend on size and ordit of two bodies. Light curve can be similar in both cases.

 

[Bandersnatch]

But we can not differentiate using light curve only. Shape of light curve and space between dips will depend on size and ordit of two bodies. Light curve can be similar in both cases.

How? You'd need two planets with the same period.

 

[A Alex P]

How? You'd need two planets with the same period.

Ok, in case of similar size eclipsing binaries we can differentiate them from planets. But if binary candidates have a combination of stars such that one star is bright and big, while other is faint and small in size, then it will need further confirmation techniques.

 

[Bandersnatch]

Ok, in case of similar size eclipsing binaries we can differentiate them from planets. But if binary candidates have a combination of stars such that one star is bright and big, while other is faint and small in size, then it will need further confirmation techniques.

Why? With planets you only get dips in the curve when the planet transits the star, not the other way around. With binaries you get dips in both cases.
Like so:

(image source: planethunters.org)

 

[A Alex P]

Why? With planets you only get dips in the curve when the planet transits the star, not the other way around. With binaries you get dips in both cases.
Like so:
View attachment 224670
(image source: planethunters.org)

Ok, now it is clear to me. Thanks.

 

[phyzguy]

I think there is some confusion here. It's not that Kepler can't tell eclipsing binaries from exoplanets. They are easy to tell apart, as Bandernsatch illustrates. The problem comes when you are monitoring the brightness of a star, and there is another (fainter) eclipsing binary star coincident with the star you are measuring. Kepler intentionally defocuses images in order to reduce the error of a brightness measurement by spreading the light of the star across several pixels. This means that what looks like one star can actually contain the star you are interested in and also contain a dimmer eclipsing binary that you don't know is there. When the light from the eclipsing binary dips, you see a very small dip in the star you are monitoring that can look just like the signal of an exoplanet. The secondary dip can be so small that you don't see it. This is one of the main reasons why Kepler candidates are considered "candidates" until the same star has been followed up with a larger telescope at higher resolution to make sure there is no contamination.

 

[snorkack]

Are brown dwarfs classified as stars or exoplanets?

How common are eclipsing brown dwarfs? If this is even known (can they be distinguished from planets)?

 

[stefan r]

Are brown dwarfs classified as stars or exoplanets?

How common are eclipsing brown dwarfs? If this is even known (can they be distinguished from planets)?

A companion brown dwarf would have formed from the same collapsing gas cloud as other stars. A planet formed in a star's accretion disk. The sizes overlap, you could have a brown dwarf that was less massive than some of the large planets.

An eclipsing binary produces two equally spaced dips in the light curve, instead of one, as both objects are luminous.

Stars in elliptical orbits are not likely to have equally spaced eclipses.

A white dwarf in front of a red giant would not make much of a dip in the ultra-violet. A red giant in front of a white dwarf would not create much dip in the red. Planets are also luminous so there is a small dip when they go behind their star. The planet eclipse is much smaller but is in the reflected frequencies. The planet also has phases like the Moon or Venus. A planet is brightest right before and after it disappears.

Kepler space mission has discovered thousands of exo-planet candidates. Why are they just candidates ? Why followup ground based study is required ? Why is Kepler unable to confirm them ?
Second thing, how are scientists going to differentiate between eclipsing binaries mimicking to be exoplanets.

Suppose 2 asteroids pass in front of Kepler. I am not sure, but possibly 2 pieces of nearby dust could also make the right type of dip. If the time between dips is 1.234 years it is not likely that a third asteroid will pass in exactly 2.468 years. The first 2 dips make it a candidate. More telescopes can focus on the star when the transits are supposed to occur (around 2.5, 3.7, 4.9 etc years). Kepler collected a lot of data about a large region. A large telescope can make a more detailed transit curve.

 

[JMz]

A companion brown dwarf would have formed from the same collapsing gas cloud as other stars. A planet formed in a star's accretion disk.

Is this true? It seems typical but not necessary. (1) Some of the hot Jupiters seem to have masses that are close to the big-enough-to-fuse (deuterium anyway), even though they are treated as planets -- presumably arising from the accretion disk -- not stars. Also, two stars that were formed as a sufficiently close binary must have finished their formation by creating a mutual accretion disk, which can be said to be what they formed "from", just as much as any planets did. (2) Conversely, is there any reason to think that the range of objects that form from the collapsing gas cloud is limited to masses high enough to fuse? If not, then some planets must form directly in that cloud.

 

[Bandersnatch]

Stars in elliptical orbits are not likely to have equally spaced eclipses.

How so? They have the same period.
(I don't mean spacing between every dip - I meant spacing between dips of each star)

 

[stefan r]

How so? They have the same period.
(I don't mean spacing between every dip - I meant spacing between dips of each star)

Suppose we have this:

If we are looking from the bottom center straight up we could see an eclipse followed by a second eclipse with 1 second spacing. The next eclipse is 5 seconds later. If we observe from side-center we would have 3 seconds between each eclipse.

If we are talking about the length of the eclipse itself we get the same length only when looking vertically (vertical on upright computer screen). In the horizontal few there is a quick eclipse at periapsis followed by a long duration eclipse at apoapsis.

Looking at more angles you get various lengths but only 2 are equal.

 

[sophiecentaur]

Why are they just candidates ? Why followup ground based study is required ? Why is Kepler unable to confirm them ?

It could be that the mission is to identify a large number of candidates. There is not time to also examine a few individual candidates in depth. The two functions are in fact very different.

 

[A Alex P]

I think there is some confusion here. It's not that Kepler can't tell eclipsing binaries from exoplanets. They are easy to tell apart, as Bandernsatch illustrates. The problem comes when you are monitoring the brightness of a star, and there is another (fainter) eclipsing binary star coincident with the star you are measuring. Kepler intentionally defocuses images in order to reduce the error of a brightness measurement by spreading the light of the star across several pixels. This means that what looks like one star can actually contain the star you are interested in and also contain a dimmer eclipsing binary that you don't know is there. When the light from the eclipsing binary dips, you see a very small dip in the star you are monitoring that can look just like the signal of an exoplanet. The secondary dip can be so small that you don't see it. This is one of the main reasons why Kepler candidates are considered "candidates" until the same star has been followed up with a larger telescope at higher resolution to make sure there is no contamination.

you have written a line " The problem comes when ...". In this you want to say that target star and plus a binary star system revolving around it ( three bodies in total) or something else.

 

[Bandersnatch]

If we are looking from the bottom center straight up we could see an eclipse followed by a second eclipse with 1 second spacing. The next eclipse is 5 seconds later. If we observe from side-center we would have 3 seconds between each eclipse.

If we are talking about the length of the eclipse itself we get the same length only when looking vertically (vertical on upright computer screen). In the horizontal few there is a quick eclipse at periapsis followed by a long duration eclipse at apoapsis.

Looking at more angles you get various lengths but only 2 are equal.

Again, as I wrote earlier - I don't mean spacing between dips made by one star transiting and then the other. I mean that spacing between two transits of one star is the same as the spacing between two transits of the other star (6 seconds in your example). That's what shows that they're both sharing the orbital period, and what differentiates the dips from two planets on different orbits transiting the star.

It could be that the mission is to identify a large number of candidates. There is not time to also examine a few individual candidates in depth. The two functions are in fact very different.

I also think that's it.
If one looks at the light curves taken by Kepler, available for viewing e.g. at planethunters.org, one can see that there is a lot of noise, and observation period is often only long enough to catch a single transit.

 

[A Alex P]

A companion brown dwarf would have formed from the same collapsing gas cloud as other stars. A planet formed in a star's accretion disk. The sizes overlap, you could have a brown dwarf that was less massive than some of the large planets.

Stars in elliptical orbits are not likely to have equally spaced eclipses.

A white dwarf in front of a red giant would not make much of a dip in the ultra-violet. A red giant in front of a white dwarf would not create much dip in the red. Planets are also luminous so there is a small dip when they go behind their star. The planet eclipse is much smaller but is in the reflected frequencies. The planet also has phases like the Moon or Venus. A planet is brightest right before and after it disappears.
Suppose 2 asteroids pass in front of Kepler. I am not sure, but possibly 2 pieces of nearby dust could also make the right type of dip. If the time between dips is 1.234 years it is not likely that a third asteroid will pass in exactly 2.468 years. The first 2 dips make it a candidate. More telescopes can focus on the star when the transits are supposed to occur (around 2.5, 3.7, 4.9 etc years). Kepler collected a lot of data about a large region. A large telescope can make a more detailed transit curve.

1 ) "The sizes overlap, you could have a brown dwarf that was less massive ...". So, we will need mass information using spectroscopy to completely confirm that its exoplanet system ? Or something else.

2 ) "Stars in elliptical orbits are not likely to have equally spaced eclipses." This will be case for same mass binary stars only otherwise they will show dips at unequal space. And size of stars will fix the amount of dip in light curve.

3 ) "A white dwarf in front of a red giant would not make..." Most of the observations are taken in R band using ground based telescopes while doing follow up studies, whats's the reason behind it ?
The second dip is very small in case of planets as you already said it's due to reflected light from planets surface just before it disappears behind star. Is there any possibility for any binary system to give same light curve ?

4) "Suppose 2 asteroids pass in front of Kepler. I am not sure..." This is true for long period candidates. I mean if dips appeared after one year or so then we will definitely need follow up observations on expected dates of transit. But for short period candidates Kepler itself can cover a lot of transits. I mean there will some time limit that Kepler is going to observe one field say for 6 months. In this time of observation it observes same field of stars. In this time if he detects candidates with period less than 1 month it can easily cover their 6 transits but if some of the candidates have period of 3 or more months then it will need follow up observations. But I have seen in list of exoplanet candidates there are unconfirmed candidates with period in days also.

 

[A Alex P]

It could be that the mission is to identify a large number of candidates. There is not time to also examine a few individual candidates in depth. The two functions are in fact very different.

You mean ground based follow up observations are part of mission itself.

 

[phyzguy]

you have written a line " The problem comes when ...". In this you want to say that target star and plus a binary star system revolving around it ( three bodies in total) or something else.

No, the target star plus the eclipsing binary are not in the same stellar system. They are simply in almost the same direction. For example, we might have a target star that is 1000 light years away, and an eclipsing binary that is 3000 light years away, but so close to it on the sky that the two objects look like one object. Then, when the eclipsing binary (that we don't know is there) dims, we see a slight dimming in the combined light from the target star and the eclipsing binary that we interpret as an exoplanet transit.

 

[A Alex P]

No, the target star plus the eclipsing binary are not in the same stellar system. They are simply in almost the same direction. For example, we might have a target star that is 1000 light years away, and an eclipsing binary that is 3000 light years away, but so close to it on the sky that the two objects look like one object. Then, when the eclipsing binary (that we don't know is there) dims, we see a slight dimming in the combined light from the target star and the eclipsing binary that we interpret as an exoplanet transit.

Ok, eclipsing binary and target star are close to resolve as we observe them from earth. I was thinking you are saying a elipsing binary system revolving around target star.

 

[stefan r]

you have written a line " The problem comes when ...". In this you want to say that target star and plus a binary star system revolving around it ( three bodies in total) or something else.

I believe he means an optical double. The target star could be 150 parsecs away and the eclipsing binary 1000 parsecs. An optical double could also be a galaxy or quasar that is mega parsecs or giga parsecs away.

It should be possible to have a planet orbit a star and then transit an optical double. I have not seen anything suggesting that we have found such a thing. Proper motion would usually prevent that from repeating.

1 ) "The sizes overlap, you could have a brown dwarf that was less massive ...". So, we will need mass information using spectroscopy to completely confirm that its exoplanet system ? Or something else.

If it is orbiting a star and it is not itself a star than it is an exoplanet. We do not have instruments strong enough to identify dwarf planets and moons. When those instruments are built someone will have to tighten up the definitions of words we use. A planet has cleared its orbit around the sun.

2 ) "Stars in elliptical orbits are not likely to have equally spaced eclipses." This will be case for same mass binary stars only otherwise they will show dips at unequal space. And size of stars will fix the amount of dip in light curve.

The eccentricity of orbits is not related to the mass. Earth-sun is nearly circular so aliens will observe Earth is eclipsed near 6 months before and after the transit. Elliptical orbits will have uneven eclipse timing regardless of the mass ratio. The exception is an alien observer looking at the eclipses right at apoapsis and periapsis. That exception will be there regardless of the mass ratio.

Not sure what you mean by "fix". You can have partial eclipses where part of the planet/star/moon is overlapping the edge of the circle. When the eclipse is total (or transit) the light curve has a flat bottom. The total depth of the dip is effected by both the size of the objects and the brightness.

Even that is not completely true. The sun is a little bit redder/dimmer at the edge than in the center because of the sun's atmosphere. A mercury transit across the center would get slightly deeper as it passed the middle.

3 ) ... Most of the observations are taken in R band using ground based telescopes while doing follow up studies, whats's the reason behind it ?...

Kepler used that frequency:

...The instrument has a spectral bandpass from 400 nm to 850 nm...

3 )...
The second dip is very small in case of planets as you already said it's due to reflected light from planets surface just before it disappears behind star. Is there any possibility for any binary system to give same light curve ?

I do not know the albedo of a star. The amount of emitted light will be much higher than reflected light. For a star that is very close to it's companion there should be tidal effects. Tides should change the brightness more than reflection but I am not sure of details. For an average case the star will have a nearly constant brightness and a planet has a near 100% cycle in high frequency. The dip is the same/similar the gradual increase over a half orbit is different.

Would be interesting if someone knows how to calculate the light curve for an M-dwarf orbiting a B-dwarf slightly outside the Roche limit. My internet searching failed.

4) ...In this time if he detects candidates with period less than 1 month it can easily cover their 6 transits but if some of the candidates have period of 3 or more months then it will need follow up observations. But I have seen in list of exoplanet candidates there are unconfirmed candidates with period in days also.

Suppose someone runs through a stop sign and hits your car. You know it is his fault. He knows it is his fault. The circumstances are clear. His insurance company will still insist on a formal police report. Obviously the police officer was not there and you were there so how could (s)he know better what happened? I do not know of any reason to doubt the Kepler team if they believe they are certain that a candidate will be confirmed. The planet still needs to be a "candidate" until someone else confirms it.
We also do not want to have a moving target for how many eclipses need to be observed. With several thousand planets you can make estimates of the density of planets and compare their characteristics. Large planets near small stars are easiest to find. That does not mean that a typical planet is large, or that a typical planet orbits a small star, or that a typical planet has a short orbital period. The candidate list can be used to improve the statistics. If Cygnus is behind the sun during a transit we might not have any telescopes able to look at it. Weather on Earth does not make a planet not exist. It just hides the evidence.

 

[snorkack]

If it is orbiting a star and it is not itself a star than it is an exoplanet. We do not have instruments strong enough to identify dwarf planets and moons. When those instruments are built someone will have to tighten up the definitions of words we use. A planet has cleared its orbit around the sun.

WD-1145+017 b has about 2000 km diametre. Smaller than Moon. And smaller than Pluto. There is no reason why satellites cannot be observed, nor dwarf planets.

In Solar System, Ceres orbits at 10,6 degrees from zodiac. Vesta at 7,1 degrees.
An observer who sees Ceres transit but who is a few degrees off Vesta´s orbital plane and therefore cannot see Vesta would have no clue that Ceres has not cleared its orbit and therefore is not a planet. Confirming a planet requires proving nonexistence of bodies that are too massive and too close (though on non-transiting orbits).

 

[stefan r]

WD-1145+017 b has about 2000 km diametre. Smaller than Moon. And smaller than Pluto. There is no reason why satellites cannot be observed, nor dwarf planets.
...

WD-1145+017 b is a great example for detection limits. When the star is more compact the transit is amplified. WD-1145+017 has 2% of the Sun's radius but WD-1145+017 b has 15% of Earth's radius. That makes it a much easier target. You could build a detector capable of finding WD-1145+017 b that could not find a Jupiter sized object around a star 50% wider than the Sun.

The transit time for the large object around WD-1145+017 is around 4 minutes. In the discovery paper they say WD-1145+017 b has a dust cloud larger than Earth. The core object is small but the transit is much longer.

They also found moons around WASP 1407b . That is a gap 4 million km wide in the rings. The ring system is over an au wide. We do not have transit data or images of the moon that created the gaps. It could be orbiting anywhere in the ring or in a wider orbit that creates a resonance.

In Solar System, Ceres orbits at 10,6 degrees from zodiac. Vesta at 7,1 degrees.
An observer who sees Ceres transit but who is a few degrees off Vesta´s orbital plane and therefore cannot see Vesta would have no clue that Ceres has not cleared its orbit and therefore is not a planet. Confirming a planet requires proving nonexistence of bodies that are too massive and too close (though on non-transiting orbits).

If we are not sure whether it should be called a brown dwarf or a planet (at least Jupiter mass) and it also transited twice in Kepler telescope's observation time then can we be confident it cleared an orbit sometime in the past or will within a billion years future?

The astronomical union's definition of a planet says the words "around the sun".

 

[snorkack]

They also found moons around WASP 1407b . That is a gap 4 million km wide in the rings. The ring system is over an au wide. We do not have transit data or images of the moon that created the gaps. It could be orbiting anywhere in the ring or in a wider orbit that creates a resonance.

If we are not sure whether it should be called a brown dwarf or a planet (at least Jupiter mass)

Is the second component of WASP 1407 a brown dwarf planet (and then its satellite making a gap in the rings is a moon) or does it as a brown dwarf qualify as a member of a binary star (and then it is the satellite making the gap in the rings that is a planet)?