Kepler Mission Options

[Decepticon]

If TrES 2 Had any moons would that be detectable?

[Holder of the Tw...]

Moons would appear in the transit data if they were as large as Titan or Ganymede, but given how close they would be to the planet, it is unlikely they would show up as separate signals. They would be averaged into their planet during the time intervals that Kepler could distinguish.

An excerpt from the mission manager updates on the main Kepler site ...

"2009 April 20. Mission Manager Update - The Kepler science team has decided that further refinement of the telescope's focus would significantly improve the mission's science return. The project is therefore proceeding with these adjustments. "

[Syrinx]

http://www.nasa.gov/mission_pages/kepler/n...m-20090424.html

Kepler Mission Manager Update
04.23.09

by Jim Fanson, Kepler, JPL Project Manager

The Kepler telescope's focus has been successfully optimized. This involved moving the primary mirror of the telescope toward the focal plane array, the area where light is focused, by 40 microns (1.6 thousandths of an inch) and tilting it by 0.0072 degrees. Various other calibrations are underway, including: detailed measurement of star images formed by the telescope at various locations on the focal plane; determination of the exact sky coordinates of every one of the camera's 95 million pixels, and mapping of "ghost" images, which result when the light from bright stars reflects off the front of the camera's charge-coupled devices (CCDs), bounces off lenses inside the telescope, and winds up back on the CCDs in another location.

[Holder of the Tw...]

In an earlier reply I made to Deception's question about detecting moons around Tres-2, I gave myself some wiggle room by limiting the answer to "separate signals", without really answering the question (even though the answer "probably not" was implied). There would be some 30 minute intervals where a moon would show up alone, along with an equal number of cases where the planet would too, when the elongation was sufficient. My guess was there would not be enough of these intervals by themselves during the prime mission to add up to a detection. We're talking about a moon orbit here that is, at most, half the size of our own moon's orbit.

Provided, of course, that separate signals are really required. Whether there is a statistical methode for analysing the data in it's entirety, I didn't know, and still don't.

But now I found this little tidbit from Caltech:

"TrES-2 is the first transiting planet - or planet that passes directly between its star and Earth - to be found in an area of the sky known as the "Kepler field", ... Discovering TrES-2 beforehand allows Kepler's astronomers to plan additional observations of it, such as searching for moons."

Okay. My bad. My bad.

[Mongo]

I would have thought that the main technique for discovering a large moon around Tres-2 would be timing variations. Since the planet's primary will be monitored continuously, we should end up with three years worth of transits and their timing. If the transit timing varies periodically with a period significantly smaller than Tres-2's orbital period (it should be possible to disentangle the putative moon's period from Tres-2's orbital period), I would think that the cause would have to be a moon.

As far as I know, no such variations have been found to date (on a limited number of transits), but with the huge number of transit observations expected during Kepler's primary mission, the margin of error for any timing variations should be considerably reduced from the current value.

[Holder of the Tw...]

Again, I really don't know. But my understanding is that Kepler sums up a total of 30 minutes worth of observations ( the individual readouts being much, much shorter) and transmits this average value of half an hour as its reading for each pixel. Maybe I'm wrong about that. But if it is the case, it's a fairly coarse interval we're talking about.

The timing offset for the earth, caused by the moon, would be at most three minutes on either side of the average. Still, if you added up hundreds of transits... who knows?

Edit: from Wikipedia - "The CCDs are read out every three seconds and co-added on board for 15 minutes".

[Hungry4info]

It's very unlikely TrES-2 b has any detectable moons, being so close to its parent star and thus having such a small hill radius.

[Mongo]

I am sure that you are right, for that reason. Another Earth-Luna system might be detectable, provided that there were sufficient transits recorded. How close to its primary could an Earth-Luna twin be and remain stable, and would that result in sufficient monitored transits during the Kepler primary mission to tease out the satellite?

[searches ARXIV]

Timing Detection of Eclipsing Binary Planets and Transiting Extrasolar Moons

We investigate the improved detection of extrasolar planets around eclipsing binaries using eclipse minima timing, and extrasolar moons around transiting planets using transit timing, offered by the upcoming COROT (ESA, 2005), Kepler (NASA, 2007), and Eddington (ESA 2008) spacecraft missions. Hundreds of circum-binary planets should be discovered, and a thorough survey of moons around transiting planets will be accomplished by these missions.

Determination of the size, mass, and density of "exomoons" from photometric transit timing variations

Precise photometric measurements of the upcoming space missions allow the size, mass, and density of satellites of exoplanets to be determined. Here we present such an analysis using the photometric transit timing variation (TTV_p). We examined the light curve effects of both the transiting planet and its satellite. We define the photometric central time of the transit that is equivalent to the transit of a fixed photocenter. This point orbits the barycenter, and leads to the photometric transit timing variations. The exact value of TTV_p depends on the ratio of the density, the mass, and the size of the satellite and the planet. Since two of those parameters are independent, a reliable estimation of the density ratio leads to an estimation of the size and the mass of the exomoon. Upper estimations of the parameters are possible in the case when an upper limit of TTV_p is known. In case the density ratio cannot be estimated reliably, we propose an approximation with assuming equal densities. The presented photocenter TTV_p analysis predicts the size of the satellite better than the mass. We simulated transits of the Earth-Moon system in front of the Sun. The estimated size and mass of the Moon are 0.020 Earth-mass and 0.274 Earth-size if equal densities are assumed. This result is comparable to the real values within a factor of 2. If we include the real density ratio (about 0.6), the results are 0.010 Earth-Mass and 0.253 Earth-size, which agree with the real values within 20%.


So it looks like detection of sufficiently large exomoons via transit timing with both CoRoT and Kepler should be possible, if they exist.

[Holder of the Tw...]

As I mentioned earlier, any moon of Tres-2 would have to be within about half a lunar distance of the planet to be stable. Say about 130,000 mile radius (forgive my english units here) for the orbit. My back of the envelop calculations, using Wiki values, showed a hill sphere of about a lunar distance. Over the long haul, only orbits about half the size of a hill sphere are truly stable, even if the orbit is retrograde.

Earth has a hill sphere of about four lunar distances, so the moon could effectively orbit to twice its distance (or four times for a geologically brief period). You could also keep the moon at its current distance, move the earth to half its current distance to the sun (not recommended), and still have a stable earth-moon system.

The most massive moon in our solar system is Ganymede. The mass ratio of Tres-2 : Ganymede is about 16,000 : 1.

If we put Ganymede in orbit around Tres-2 at 130,000 miles, the center of mass of the system is offset from the center of mass of the planet by 8 miles. At 50 miles/sec orbital speed of the planet about the star, the difference in transit timings at greatest elongation amount to about plus or minus 0.15 seconds from the expected. Over a fifteen minute integration of photometry, you get about a 0.02 per cent lightening or darkening over what you would expect during that interval.

The moon, by the way, would preceed or follow the planet by about 40 minutes at greatest elongation.

I welcome any efforts to check (and correct) my math here.

[Mongo]

The math looks good to me.

Perhaps a more productive technique would be to look for any moons directly. The radius of TrES-2 is about 1.272 times that of Jupiter, or about 89,000 km. A large moon might be around 2,000 or 2,500 km in radius, for an areal ratio of 1270 to 1 (2,500 km) or 1980 to 1 (2,000 km). This would represent an additional dimming of up to about 0.08 percent of the drop due to the exoplanet occurring up to 40 minutes before or after second contact and a compensating rise occurring the same amount of time before or after third contact. Is 0.08% of the full lightcurve drop within the sensitivity of the Kepler detectors?

Actually, if both techniques were possible, both mass and radius (and hence density) of this hypothetical moon would be known, which would imply bulk composition -- very useful to know.

This would certainly be an easier task when dealing with hot Neptunes or superEarths, instead of TrES-2 in particular.

[Syrinx]

http://www.nasa.gov/mission_pages/kepler/n...m-20090501.html

by Jim Fanson, Kepler, JPL Project Manager

Kepler's calibration data collection is drawing to a close. Several hundred data sets have been acquired to characterize and map the optical and noise performance of the telescope and the electronics for the focal plane array (the area where light is focused).

[Syrinx]

http://www.nasa.gov/mission_pages/kepler/n...m-20090507.html

The project will convene a science operations readiness review on Monday, May 11, to determine if the team is ready to commence science data collection.

[Holder of the Tw...]

According to the last mission manager report, they were suppose to decide today whether Kepler is ready to proceed with science observations. Haven't heard anything yet.

Is 0.08% of the full lightcurve drop within the sensitivity of the Kepler detectors?

Depends on how close it is to the star. If Kepler could detect a moon this size with an eighty minute transit time of its own, and you are lucky enough to always catch it far away from the planet, then clearly yes. The 0.08% variance simply matches the size of the object that is detectable, so it would be within the sensitivity. But it won't always be 40 minutes of separation.

The two papers you listed seem to hint at a very optimistic outlook for Kepler finding moons. So I'm at a loss to explain it. These graphs and charts from the Kepler website would indicate to me that our own moon around the earth would not be detectable by Kepler, although earth itself would be.

It would help a lot if the Kepler website addressed the issue directly. Given the fact that a Jupiter size planet, in an earth like orbit, might possibly have earth sized moons, the omission of any talk of Kepler's ability to detect such moons seems a bit of an oversight.

Or maybe I missed something. Like on one of the education/outreach pages?

[Holder of the Tw...]

The hunt is now underway! Best wishes for success.

Status Report

[tacitus]

Indeed. I guess we're entering the "hurry up and wait" stage as we wait for the first announcements of hot Jupiters and the like.

I assume they won't be announcing anything for several months, even if they find some short period planets within the next few weeks?