Planet Hunters Talk

EB with a Companion in K2 C1

  • JKD by JKD

    KID 201158453 = EB (p=7.079d) with planetary transit at 2000.43BKJD

    Posted

  • robert_gagliano by robert_gagliano in response to JKD's comment.

    Very nice pickup JKD. I checked the Lc and unfortunately couldn't find any repeat transits. Might be a new circumbinary. I would suggest you alert the science team...send Meg an email or message.

    Posted

  • ajamyajax by ajamyajax

    201158453:

    Well plenty speculative, but maybe a repeat of around P=40.86 for a circumbinary PC... I did get an ok fit with minimal data points. So maybe a warm Jupiter here if we are lucky. The short duration might suggest a luminous object such as a small star or a brown dwarf, again if it is real. p.s. a P=33.x period also seemed possible, but those two possible transit shapes looked less similar I thought.

    s1=1980.48 p1=7.0775 d1=0.19 (4.56 hours)

    s2=1983.657 p2=7.0775 d2=0.13 (3.12 hours)

    s3=2000.453 p3=40.86 d3=0.12 (2.88 hours)

    T1

    T2

    F3

    Posted

  • Shellface by Shellface

    I've tried to write out all of my thoughts here a couple of times, but it's all so snakey and long-winded that I cannot finish it. Instead, have some diagrams.

    On the left are the transit timing variations for the binary transits, and on the right is the same for the eclipses. Both are in units of days, and relative to the first epoch. Note the differing scales. The differing shapes of the variations - seemingly linear with abrupt changes for the primary and closer to sinusoidal for the secondary - are quite obvious.

    Timing variations of the binary transits and eclipses have been used to measure the mass of Kepler circumbinary systems, or if not, provide upper limits. They are more complex than planetary TTVs because they are not dominated by resonant interaction, but instead show the results of quasi-stochastic gravitational interaction and relativistic effects.

    Does this support the notion of the presence of a circumbinary companion? Well, maybe. The transits and eclipses are narrow - both look grazing, to me - so there is not much data covers each event. It's a possibility that a lack of good time resolution could result in biases in the derived timings. It is thus important to check whether these timings are consistent with a physically plausible model, which is a bit out-of-reach considering the lack of system parameters at the current juncture.

    Could this really be a circumbinary planetary system? Eh. Maybe. With only one event that can really be called a transit, if it is one, to work with, it's difficult to say much. It's an open possibility that an object could only transit once, even if it has a orbital period shorter than the dataset. It would be unwise to draw any strong conclusions with such limited understanding of the system.

    I will be keeping this in mind, but there is a lot that could go wrong here.

    Posted

  • ajamyajax by ajamyajax in response to Shellface's comment.

    SF, what are your thoughts on the stellar situation? Is there a chance of a blend here where the actual duration of at least one good transit is longer than the observed?

    EPIC, 2MASS, J mag, H mag, K mag, J - H, H - K, (J-H spectral type, stellar mass est) (H-K spectral type, stellar mass est)

    201158453 , 2MASS J11355537-0505380 , 13.291 , 12.832 , 12.762 , 0.459 , 0.07 , ('K2V', 0.85) , ('G1V', 1.07)

    Posted

  • Shellface by Shellface

    The lightcurve seems compatible with the primary being an early K-dwarf. Though more uncertain than typical due to the grazing eclipse, I would wager that the secondary is early M. There are nearby stars visible in the 2MASS image, so blending isn't likely.

    I'm not sure what you mean in the second bit. The stellar events seem fine, quite compatible with being on target. In turn, none of the possible planetary transits are complete, so it is not possible to draw any real conclusions regarding their shape. The one event that has enough to really look at seems fine. Perhaps its impact parameter is small, because the ingress is short (under the assumption that this is a transit).

    Also, let me just dump the part of my written junk that is presentable.

    Let us turn to the system in question. This lightcurve contains one fairly clear transit-like event that is not due to the binary, but does not have an unambiguous second event, the one aja pointing out being tentative. The profile of the clear event is decent, but the data for its ingress is missing. Conversely, the data for the weak event has little more than an ingress, if interpreted as a transit. Limits on the durations of the events can be inferred from the available data as ~ 2 - 3.5 hours for the clear event and ~ 1 - 3 hours for the weak one.

    It is obviously important to find parameters for the binary for a circumbinary system, and not much can be said about the possible transits without it. The colours of the source give a spectral type of K2 or so; given the fairly large ratio of transit/eclipse depths, this is probably a good representation of the primary spectral type, because the secondary probably does not contribute too much to the total flux. It can be seen that both the transits and eclipses of the binary are approximately perfectly V-shaped. Though transits become more V-shaped as radius ratio increases, eclipses do not follow the same trend, but do become V-shaped for grazing events. So, it can be inferred that the eclipses, and perhaps the transits, are grazing. Because of this, the radius and temperature ratios cannot be clearly determined, but the companion is probably end-K - early M. The transit morphology suggest that the components are dwarfs.

    The non-zero phase offset between transits and eclipses means that the stellar orbit is eccentric. The eclipses occur at phase 0.45 (where they would be at 0.50 for a circular orbit), so the eccentricity is probably not large; the phase offset gives e cos ω ≈ -0.079, where omega is the argument of periapsis; the negative value of e cos ω implies 90 < ω < 270 degrees, because eccentricity cannot be negative. Because cos ω can only lie in the interval [-1, 0], the eccentricity must be >~0.079.

    Large out-of-transit variability can be seen in the lightcurve, and its varying function indicates that it is due to spots. The principle period is 5.46 days, which is probably pseudo-synchronous with the binary orbit (it avoids a 1:1 resonance due to the eccentricity of the system driving tidal equilibrium to a shorter period). If my calculations are correct, this gives e ≈ 0.125, so ω ≈ 130 or 230 degrees. The eclipses being substantially shorter than the transits favours the former, but this is not concrete.

    Posted

  • ajamyajax by ajamyajax

    201158453: running first tests with new corrected data and looks like my previous ideas here were processed more like glitches in their Best LC data correction. (See orange bands below)

    T1a
    T1b

    Data Credit: HAT Surveys — HATNet — HATSouth

    Astrophysical Sciences at Princeton University

    http://k2.hatsurveys.org/archive/

    Posted

  • Shellface by Shellface in response to ajamyajax's comment.

    Hm. That looks a lot worse.

    Maybe the transit-like event is/was non-astrophysical? Because that really doesn't look good with that improved reduction.

    Posted

  • Martti_Holst_Kristiansen by Martti_Holst_Kristiansen

    The possible planetary transit candidate (2000.44 BKJD) is initialized with missing data points, this could be explained by the fact that the corrected Vanderburg lighcurves have all bad data points (SAP_QUALITY > 0) removed before producing a lightcurve:

    enter image description here

    This being said, perhaps there is another opportunity, see 2033.44 BKJD:

    enter image description here

    Posted

  • ajamyajax by ajamyajax in response to Martti Holst Kristiansen's comment.

    Yes, saw that also as another possible spec transit with the repeat being ~2033.42 BJD, but to me that looks like a glitch in the Hat Survey corrected data (Best LC) as well. Feel free to check that out for yourself in the new data though.

    s3=2000.43 p3=32.99 d3=0.17 (4.08 hours)

    T3X

    Posted