A Discovery of a Candidate Companion to a Transiting System KOI-94: A Direct Imaging Study for a Possibility of a False Positive
Yasuhiro H. Takahashi, Norio Narita, Teruyuki Hirano, Masayuki Kuzuhara, Motohide Tamura, Tomoyuki Kudo, Nobuhiko Kusakabe, Jun Hashimoto, Bun'ei Sato, Lyu Abe, Wolfgang Brandner, Timothy D. Brandt, Joseph C. Carson, Thayne Currie, Sebastian Egner, Markus Feldt, Miwa Goto, Carol A. Grady, Olivier Guyon, Yutaka Hayano, Masahiko Hayashi, Saeko S. Hayashi, Thomas Henning, Klaus W. Hodapp, Miki Ishii, Masanori Iye, Markus Janson, Ryo Kandori, Gillian R. Knapp, Jungmi Kwon, Taro Matsuo, Michael W. McElwain, Shoken Miyama, Jun-Ichi Morino, Amaya Moro-Martin, Tetsuo Nishimura, Tae-Soo Pyo, Eugene Serabyn, Takuya Suenaga, Hiroshi Suto, Ryuji Suzuki, Michihiro Takami, Naruhisa Takato, Hiroshi Terada, Christian Thalmann, Daigo Tomono, Edwin L. Turner, Makoto Watanabe, John Wisniewski, Toru Yamada, Hideki Takami, Tomonori Usuda
aa r X i v : . [ a s t r o - ph . E P ] S e p A Discovery of a Candidate Companion to a Transiting SystemKOI-94: A Direct Imaging Study for a Possibility of a FalsePositive
Yasuhiro H. Takahashi , , ∗ , Norio Narita , Teruyuki Hirano , Masayuki Kuzuhara ,Motohide Tamura , , Tomoyuki Kudo , Nobuhiko Kusakabe , Jun Hashimoto , Bun’eiSato , Lyu Abe , Wolfgang Brandner , Timothy D. Brandt , Joseph C. Carson , ThayneCurrie , Sebastian Egner , Markus Feldt , Miwa Goto , Carol A. Grady , , , OlivierGuyon , Yutaka Hayano , Masahiko Hayashi , Saeko S. Hayashi , Thomas Henning , KlausW. Hodapp , Miki Ishii , Masanori Iye , Markus Janson , Ryo Kandori , Gillian R.Knapp , Jungmi Kwon , Taro Matsuo , Michael W. McElwain , Shoken Miyama ,Jun-Ichi Morino , Amaya Moro-Martin , , Tetsuo Nishimura , Tae-Soo Pyo , EugeneSerabyn , Takuya Suenaga , , Hiroshi Suto , Ryuji Suzuki , Michihiro Takami , NaruhisaTakato , Hiroshi Terada , Christian Thalmann , Daigo Tomono , Edwin L. Turner , ,Makoto Watanabe , John Wisniewski , Toru Yamada , Hideki Takami , Tomonori Usuda *; [email protected] Department of Astronomy, The University of Tokyo, 7-3-1, Hongo, Bunkyo-ku, Tokyo,113-0033, Japan National Astronomical Observatory of Japan, 2-21-1, Osawa, Mitaka, Tokyo, 181-8588,Japan Department of Earth and Planetary Sciences, Tokyo Institute of Technology, Ookayama,Meguro-ku, Tokyo 152-8551, Japan Subaru Telescope, National Astronomical Observatory of Japan, 650 North A‘ohokuPlace, Hilo, HI96720, USA H.L. Dodge Department of Physics & Astronomy, University of Oklahoma, 440 W BrooksSt Norman, OK 73019, USA Laboratoire Lagrange (UMR 7293), Universit´e de Nice-Sophia Antipolis, CNRS, Obser-vatoire de la Cˆote d’azur, 28 avenue Valrose, 06108 Nice Cedex 2, France Max Planck Institute for Astronomy, K¨onigstuhl 17, 69117 Heidelberg, Germany Department of Astrophysical Science, Princeton University, Peyton Hall, Ivy Lane,Princeton, NJ08544, USA Department of Physics and Astronomy, College of Charleston, 58 Coming St.,Charleston, SC 29424, USA Department of Astronomy & Astrophysics, University of Toronto, 50 George St.,Toronto, Ontario, M5S 3H4, Canada Universit¨ats-Sternwarte Mu¨nchen, Ludwig-Maximilians-Universit¨at, Scheinerstr. 1,81679 Mu¨nchen, Germany Exoplanets and Stellar Astrophysics Laboratory, Code 667, Goddard Space Flight Cen-ter, Greenbelt, MD 20771, USA Eureka Scientific, 2452 Delmer, Suite 100, Oakland CA96002, USA Goddard Center for Astrobiology Institute for Astronomy, University of Hawaii, 640 N. A‘ohoku Place, Hilo, HI 96720, 3 –Received ; acceptedSubmitted on 2013/Sep/10USA Department of Astronomy, Kyoto University, Kitashirakawa-Oiwake-cho, Sakyo-ku, Ky-oto, Kyoto 606-8502, Japan Hiroshima University, 1-3-2, Kagamiyama, Higashihiroshima, Hiroshima 739-8511,Japan Department of Astrophysics, CAB-CSIC/INTA, 28850 Torrej´on de Ardoz, Madrid, Spain Jet Propulsion Laboratory, California Institute of Technology, Pasadena, CA, 171-113,USA Department of Astronomical Science, The Graduate University for Advanced Studies,2-21-1, Osawa, Mitaka, Tokyo, 181-8588, Japan Institute of Astronomy and Astrophysics, Academia Sinica, P.O. Box 23-141, Taipei10617, Taiwan Astronomical Institute “Anton Pannekoek”, University of Amsterdam, Postbus 94249,1090 GE, Amsterdam, The Netherlands Kavli Institute for Physics and Mathematics of the Universe, The University of Tokyo,5-1-5, Kashiwanoha, Kashiwa, Chiba 277-8568, Japan Department of Cosmosciences, Hokkaido University, Kita-ku, Sapporo, Hokkaido 060-0810, Japan Astronomical Institute, Tohoku University, Aoba-ku, Sendai, Miyagi 980-8578, Japan 4 –
ABSTRACT
We report a discovery of a companion candidate around one of
Kepler
Objectsof Interest (KOIs), KOI-94, and results of our quantitative investigation of thepossibility that planetary candidates around KOI-94 are false positives. KOI-94has a planetary system in which four planetary detections have been reportedby
Kepler , suggesting that this system is intriguing to study the dynamical evo-lutions of planets. However, while two of those detections (KOI-94.01 and 03)have been made robust by previous observations, the others (KOI-94.02 and 04)are marginal detections, for which future confirmations with various techniquesare required. We have conducted high-contrast direct imaging observations withSubaru/HiCIAO in H band and detected a faint object located at a separationof ∼ . ′′ from KOI-94. The object has a contrast of ∼ × − in H band,and corresponds to an M type star on the assumption that the object is at thesame distance of KOI-94. Based on our analysis, KOI-94.02 is likely to be areal planet because of its transit depth, while KOI-94.04 can be a false posi-tive due to the companion candidate. The success in detecting the companioncandidate suggests that high-contrast direct imaging observations are importantkeys to examine false positives of KOIs. On the other hand, our transit lightcurve reanalyses lead to a better period estimate of KOI-94.04 than that on theKOI catalogue and show that the planetary candidate has the same limb dark-ening parameter value as the other planetary candidates in the KOI-94 system,suggesting that KOI-94.04 is also a real planet in the system. Subject headings: stars: imaging — planets and satellites: individual (KOI-94) —instrumentation: adaptive optics — techniques: high angular resolution
1. Introduction
One of the best ways to determine orbital parameters of extrasolar planets(exoplanets) is the transit method. Particularly the
Kepler satellite, launched in 2009, hasexecuted successful transit observations, resulting in discovery of more than 3,500 planetcandidates . It drastically increases the number of exoplanets we know and has found manymulti-transiting planetary systems.The principal problem related to the transit method is a possibility of a false positive.A false positive in transit surveys means misidentifying a signal caused by an object otherthan a planet orbiting the target star as its true planetary companion. In transit surveys,false positives are induced by some objects, for example eclipsing binaries, backgroundtransiting planetary systems or companions with transiting planets (Fressin et al. 2013)within photometric apertures of target stars. Such objects reduce their brightnessperiodically, and total flux containing both from the target stars and from the mimickingobjects consequently decreases periodically. Since the presence of the false positive sourcescannot be verified only with the transit method, the depressions in light curves cannot betranslated directly into planets orbiting the targets, and follow-up observations are requiredto confirm that the depressions are really caused by planets.Although most of Kepler planetary candidates are waiting to be confirmed, falsepositive rates reported by spectroscopic follow-up observations (Santerne et al. 2012; ∼ Kepler candidates theoretically expected by Morton & Johnson(2011) (less than 5% for over half targets) and Santerne et al. (2013) (11 . ± .
1% forwhole targets) based on Galactic models. The inconsistency has not been elucidated yet, http://kepler.nasa.gov/ 6 –and therefore, suggests that Kepler candidates are required to be confirmed with othermanners, including the radial velocity method, TTVs (Agol et al. 2005), the centroidanalysis (Batalha et al. 2010),
BLENDER analysis (Torres et al. 2004), and the direct imagingobservations.Many follow-up observations with the direct imaging method have been executed so far,but most observations were too shallow to confirm candidates efficiently. For example, letus assume that there is a false positive source within the target star’s photometric aperturewith the magnitude of ∆ m = 10 mag compared to the target. A full occultation of thesource induces a depression with the depth of 100 ppm, which is a typical value caused by anEarth-like planet transiting a Sun-like star. A false positive source with a smaller contrast(∆ m <
10) can cause the 100 ppm depression by its partial occultation, while a source witha larger contrast (∆ m >
10) can induce a shallower depression. Thus, if a depression withthe depth of 100 ppm is detected, direct imaging observations with a contrast of ∆ m = 10mag can fully and efficiently put a constraint on the possibility of a false positive. However,some studies employ direct imaging with a detection limit of ∆ m <
10 mag for confirminga transit detection with the depth of <
100 ppm (e.g. Barclay et al. 2013). Such shallowobservations cannot fully reject false positive sources. Moreover, if a companion candidateis found around the target by the direct imaging, we can evaluate gravitational influencesupon the orbital migrations of the planetary system by the companion.We focus on KOI-94, which was listed on the earliest
Kepler
Object of Interest (KOI)list (Borucki et al. 2011; Batalha et al. 2012). KOI-94 is a relatively faint — Kp = 12 . H = 11 . . +0 . − . Gyr (Hirano et al. 2012) or 3 . ± .
39 Gyr (Weiss et al. 2013). Thissystem has four planet candidates named as KOI-94.01, 02, 03 and 04, which are alsoknown as KOI-94 d, c, e and b, respectively (properties listed on Table 1). Hirano et al. 7 –(2012) discovered the “planet-planet eclipse” phenomenon in the light curves, where theterm means that a planet occults another planet transiting their host star at that time.Combining the event with measurements of the Rossiter-McLaughlin effect (Ohta et al.2005) of KOI-94.01, they confirmed that KOI-94.01 and 03, at least, are real planets andshowed that orbital axes of KOI-94.01 and 03 and the spin axis of the main star KOI-94are well aligned. The fact suggests that the planets have not experienced the planet-planetscattering (e.g. Chatterjee et al. 2008; Nagasawa & Ida 2011) or the Kozai migration(e.g. Kozai 1962; Wu et al. 2007; Fabrycky & Tremaine 2007). In contrast, a boundnessof KOI-94.02 and 04 remains unclear; radial velocities observed by Weiss et al. (2013)represented so low amplitudes that they could detect 04 only at a 2 σ significance and 02 atthe same level of non-detection, and their direct imaging observations had shallow depths(∆ K S = 3 . . ′′
5, 5.9 at 1 . ′′
0) in spite of the depth of KOI-94.04 (131 ppm). Hence, deeperdirect imaging observations for KOI-94 are required to confirm detections of the candidates.In this paper, we present results of high-contrast direct imaging observations forKOI-94 and elucidate the presence of a faint object around the star. Details of the directimaging observations and discussion based on the observations are described in Section 2and 3, respectively. We then show our results and discussion of reduced light curves ofplanetary candidates KOI-94.04 in Section 4. Section 5 summarizes the paper.
2. Deep direct imaging observations for KOI-94 and analyses
In employing the direct imaging observations, we first estimate the distance of KOI-94from the Earth. Given KOI-94’s magnitudes of g = 12 . , i = 12 .
057 (
Kepler
InputCatalog) and metallicity of [Fe / H] = +0 . ± . r -band absolute magnitude M r ∼ . M r and an observed 8 –magnitude of r = 12 .
186 (
Kepler
Input Catalog) for KOI-94 enables us to estimate itsdistance to be ∼
440 pc. On the other hand, by applying an KOI-94’s estimated mass of1 . +0 . − . M ⊙ (Hirano et al. 2012) to Yonsei-Yale isochrone model (Demarque et al. 2004), wecalculate M V ∼ .
9, which can be compared with V -band magnitude of KOI-94 V = 12 . ∼
550 pc. The accuratedistance does not matter for our following discussions, and we therefore adopt the averageof both estimations, ∼
500 pc, in this study.In order to check the presence of possible false positive sources, we observed the staras a part of the SEEDS project (Tamura 2009). The SEEDS has directly imaged stellarcompanions that are the important clue to the origin of close-in exoplanets (Narita et al.2010, 2012), as well as substellar or planetary companions (e.g. Thalmann et al. 2009;Carson et al. 2013; Kuzuhara et al. 2013). We obtained H band images using a high-contrast near-infrared camera HiCIAO (Hodapp et al. 2008; Suzuki et al. 2010) with a 188actuators adaptive optics system (AO188; Hayano et al. 2010). The detector of HiCIAOconsists of 2048 × ′′ × ′′ . The observations were performed on 2012 July 10 , where weemployed angular differential imaging (ADI; Marois et al. 2006) to remove the starlight andthe stellar speckles. The sky was photometric, and the typical full width at half maximum(FWHM) of our observed PSFs was ∼ . ′′ On the observation night, the eclipse of KOI-94.02 or 04 did not occur based on theirtransit information on the KOI catalogue. 9 –degrees in total. We note that the observations were performed without an occulting maskand array saturation, which allow us to carefully calibrate the acquired data sets.For the first attempt in data reductions, we plainly combined all frames after derotatingthem and subtracting halos from them, but spider noises and speckle noises prevented usfrom discussing the presence of faint objects around the target star. We then employedthe locally optimized combination of images (LOCI) algorithm (Lafreni`ere et al. 2007) inorder to remove the noises. As a result, we detect a faint object (hereafter KOI-94 B, orB for short) at a separation of ∼ . ′′ , though some new techniques have been developedfor avoiding the self-subtraction (e.g. Currie et al. 2013). Hence, we simply attempted toapply the classical ADI (Marois et al. 2006) data processing to the images, after subtractingthe halos of KOI-94 PSFs on each frame. Indeed, this additional attempt improved thesignal-to-noise ratio (SNR) of KOI-94 B. A final image made by combining all framesreduced with the classical ADI analysis and its 5 σ contrast curve are shown in Figure 1 The self-subtraction is an inherent problem in the ADI analysis and means a diminutionof a signal’s flux together with various noises, induced by subtraction of the signal itself. Incase that the signal is sufficiently bright and scarcely rotated, like KOI-94 B, LOCI tendsto regard the signal as a noise and, consequently, LOCI attempts to decide coefficients toohard for the sake of removing the signal. As a result, the self-subtraction of LOCI can belarger than that of the classical ADI analysis. 10 –and 2, respectively. We can easily find the faint star in Figure 1 at the position with aseparation of 0 . ′′ ◦ ) from the central star. Figure 2 shows that ourdeep high contrast imaging (the red line) reaches 1 . × − at 0 . ′′ . × − at 1 . ′′ H = 11 . K S band, the difference of wavelengths does not affectthe discussion as long as B has a moderate color of H − K S .
3. The depths of KOI-94.02and 04 in Kp band are also represented in Figure 2 as the dotted lines, because we candirectly compare them with B’s contrast in H band under condition that the transit depthsare constant in various wavelengths. Properties of KOI-94 B are summarized in Table 2 indetail.Under an assumption that KOI-94 B is a single star with the same distance as thatof KOI-94 ( ∼
500 pc), our measured magnitude, H = 18 .
2, for KOI-94 can be convertedinto an absolute magnitude of M H = 9 .
7, which can be compared with 3-4 Gyr isochronesin NextGen model (Baraffe et al. 1998) to infer a mass of ∼ . M ⊙ for B. The measuredprojected separation of B ( ∼ . ′′
6) corresponds to ∼
300 AU at the assumed distance. 11 –
3. Interpretation of the results of the direct imaging observations3.1. Influences of the new companion upon the KOI-94 system
In this section, we examine influences from the companion candidate B found with thedirect imaging. First of all, we note that B does not correspond to any of the reported fourplanets orbiting KOI-94. This is because it is unrealistic that a planet with the separationof ∼
300 AU (see Section 2) orbits the Sun-like star with the period of less than 54 days,the longest value in the KOI-94 system. Since the most important fact is that the candidateB is within the
Kepler photometric aperture (typically a few arcsecs), we discuss influencesfor all hypotheses upon the light curves of KOI-94 in this section. Among the depressionsof KOI-94, because we know that KOI-94.01 and 03 are real planets transiting KOI-94(Hirano et al. 2012; Weiss et al. 2013), only KOI-94.02 and 04 are to be discussed.The following cases can cause depressions in the light curves of KOI-94.02 or 04; (1) areal planet transiting the central star KOI-94, (2) a real planet transiting the companion B,and (3) B being an eclipsing binary. In case of (1), B cannot affect estimates of stellar andplanetary parameters from the light curves seriously, because B is much fainter than thecentral star. In cases of (2) and (3), whether B is bound or not to the central star KOI-94does not become a serious issue. If (2) or (3) is true, planet candidates of KOI-94.02 or 04may not be a real planet orbiting KOI-94. These hypotheses about KOI-94.02 are examinedand concluded in the following paragraphs based on the direct imaging observations. Thehypotheses (2) and (3) about KOI-94.04 are precisely discussed in § § § Kepler magnitude from our H band magnitude and directly compare it with the depths of the depressions of the planetarycandidates in Kp band on the KOI catalogue, we adopt the following analyses.First, if B is physically bound to KOI-94 and a single star, B’s mass should be ∼ . M ⊙ (see Section 2) and we can then calculate its apparent brightness in Kp band to be Kp = 24 . M g = 16 . M r = 15 . H band and the equation Kp = . g + 0 . r for g − r ≤ . . g + 0 . r for g − r > . , (1)where g and r are apparent magnitudes for an apparent Kp magnitude. Consequently weobtain ∆ Kp = 11 .
8, corresponding to a contrast of 1 . × − in Kp band. The contrast,which means an upper limit, is too small to explain the depths of KOI-94.02 (7 . × − ) or04 (1 . × − ). Therefore, in this case, hypothesis (2) is ruled out.Simultaneously, the hypothesis (3) that B is an eclipsing binary bound to KOI-94 canbe discussed. Because each component of the binary would be fainter than the single Mstar, the binary would consist of later type stars than the above estimate. The later starscan produce shallower depressions in the light curves, and thus the hypothesis (3) is alsoexcluded.Next we investigate the situation that B is not bound to KOI-94. In this case, theB’s color cannot be constrained from our measured H -band magnitude and the theoreticalevolutionary trucks, since it is sufficiently possible that B has the distance and age different http://keplergo.arc.nasa.gov/CalibrationZeropoint.shtml 13 –from those of KOI-94. Thus we are not able to directly estimate the Kp magnitude. Weemploy statistical discussions to estimate it based on Howell et al. (2012), who compared Kepler
Input Catalogue with 2MASS catalogue and derived empirical relationships betweeninfrared magnitudes and Kp magnitudes of stars in the Kepler field. Since they did notindicate Kp - H relationship, we substitute our H magnitude into J and K S magnitudesin their equations in order to obtain a rough estimate of B’s Kp magnitude. Then weacquire Kp = 19 . Kp = 21 . Kp - J and Kp - K S relationships, respectively.Since the magnitude of KOI-94 central star is Kp = 12 .
2, the contrasts are ∆ Kp = 7 . . × − and 1 . × − , respectively.Comparing them with the depth of KOI-94.02 (7 . × − ), we suggest that it is difficult toregard KOI-94.02 as a planet transiting B or a false positive induced by B; in other words,an extreme case that B is completely occulted can explain the depth of KOI-94.02. Onthe other hand, the possibility that KOI-94.04 (1 . × − ) is a false positive cannot beexcluded by comparing B’s magnitude and its depth. Hence we conclude that KOI-94.02 isa real planet orbiting KOI-94, assuming the typical color of stars in the field, and discussKOI-94.04’s nature in the following sections. In this subsection, we assume KOI-94.04 as a false positive.First we examine the hypothesis (2) that the planetary candidate orbits B. Becausethe assumption that the companion B is physically bound to the central star is rejected in § . × − with thecontrast of B of (1 . − . × − in Kp band (see § R KOI − . /R B = 0 . − .
8, if any light from KOI-94.04 is neglected. The estimation of R KOI − . /R B can constrain the radius of KOI-94.04. If B has a radius larger than 0 . R ⊙ ,the ratio leads to R KOI − . > . R J ; an exoplanet with a radius larger than 2.1 R J has not been discovered so far (Weiss et al. 2013, see also exoplanets.org). Consideringthe estimate of R KOI − . /R B and the observational knowledge cumulative for exoplanetproperties, we can suggest that a star with a radius of > . R ⊙ can hardly explain thedepressions for KOI-94.04. Note, however, an object with R B < . R ⊙ can account for ourestimated R KOI − . /R B . Obtaining a spectral type of B and determining its radius byfuture observations are important for investigating the possibility of the hypothesis (2).Secondly we investigate the hypothesis (3) that the companion candidate B is a binary.B’s contrast in Kp band and the depth of KOI-94.04 are the same as those in the abovesituation. If the amount of B’s fluxes decreases by 20-70% via the eclipse in the system,the depth can be explained, and thus the possibility of B being a binary is not rejected.Moreover, Fressin et al. (2013) and Santerne et al. (2013) calculated the probability of falsepositive with a given Kepler transit depth, leading to the estimated false positive rate of8 . ± .
0% for KOI-94.04 in combination of their calculations. Consequently, the possibilityof a false positive remains for KOI-94.04. If the secondary eclipses of KOI-94.04 is detected,it becomes strong evidence supporting KOI-94.04 as an eclipsing binary. However, we didnot find a significant secondary eclipse in the following light curve analyses. Thus thepossibility that KOI-94.04 is a planet (i.e. the hypothesis (1)) remains, which is discussedin the next section.
4. Constraints by KOI-94 light curves
Because we cannot fully exclude the possibility that KOI-94.04 is a false positive evenwith the new direct imaging observations, we investigate the possibility of KOI-94.04 being 15 –a real planet by revisiting the KOI-94.04 transit light curves. § § § We here employ
Kepler ’s public data sets (Borucki et al. 2011; Batalha et al. 2012)of Quarter 1 through Quarter 13. In order to detect KOI-94.04 in the light curves, wefirst remove trends on the curves by fitting polynomial functions with masking transitdepressions. Second we fold the curves by KOI-94.04’s period on the KOI catalogue(3.743245 days in Borucki et al. 2011). Third we bin the folded light curves into thousanddata points. Errors of the binned points are based on the scatter of the data points in eachbin. As a result, we see a depression of KOI-94.04.The shape of phase-folded transit light curve looks asymmetric, suggesting that thefiducial period may be incorrect. Therefore, we systematically changed its period fromthe catalogued value with shifts in unit of its error (0.000031 days) to find a symmetrictransit light curve, because a folded transit light curve with an exact period would give asymmetric curve. Figure 3 shows three light curves with various periods and their best fitmodel curves. The − . σ light curve (i.e. P = 3 . − . × . . I ( µ ) = 1 − u (1 − µ ) − u (1 − µ ) , (2)where I is the intensity and µ is the cosine of the angle between the line of sight and theline from the stellar center to the position of the stellar surface. In order to quantify theirsymmetricities, we measured a χ parameter obtained by fitting a transit model function to 16 –the curves. Here, the χ parameter is expressed as χ = X i ( F model ,i − F obs ,i ) σ i , (3)where F model ,i and F obs ,i are the i -th modeled and observed relative flux data, and σ i is itserror. The curves other than − . σ shifted data sets resulted in the higher χ parameters.In addition to χ , limb-darkening parameter u and their radial ratio between the planetcandidate KOI-94.04 and its host star KOI-94, R p /R s , are depicted in Figure 4 as a functionof time shift. In the reduction, we fixed the other limb-darkening parameter u at 0.40(Masuda et al. in prep.) and see the variation of the best-fit u value. The light curveshifted by − . σ (the red dots and line in Figure 3) indicates the smallest χ value inour grid survey (cf. Figure 4), which means 3.743186 days gives the best for its period.The corresponding u and R p /R s are 0 . +0 . − . and 0 . ± . u is especially used for the confirmation of planetary candidate KOI94.04 inthe following section. In contrast to § a/R ∗ = 7 . ± .
59, eccentricity of 0 . ± . . ± . R ⊕ , and mass of 10 . ± . M ⊕ . Surprisingly, the radius and themass lead to an extraordinarily high density of 10 . ± . − . Also, they determinedthat KOI-94.02 has a period of 10.42 days and eccentricity of 0 . ± .
23. Meanwhile,according to their orbital stability analysis, 80% of the random simulations with eccentricorbits reached close encounters between KOI-94.04 and KOI-94.02. Namely eccentricorbits typically become dynamically unstable in the KOI-94 system in their analysis, and 17 –they therefore suggested that the planets should have circular orbits to keep the systemdynamically stable. Nevertheless their radial velocity data support the eccentric orbitsfor KOI-94.02 and 04. This inconsistency allows us to consider the mass or eccentricityestimate for KOI94.02 or 04 possibly inaccurate. Furthermore, though Weiss et al. (2013)also showed TTVs in the KOI-94 system, the TTV of KOI-94.04 was not discussed in theirTTV analysis. Hence, the possibility that KOI94.04 is a false positive cannot be ruled outby the discussions of Weiss et al. (2013).Our limb-darkening analyses also enable us to discuss whether the planetary candidatesorbiting the same host star or not by comparing them. Our analyses show that the limbdarkening parameter u for the − . σ shifted data (i.e. the lowest χ ) is u = 0 . +0 . − . ,under the condition that the u parameter is fixed at 0.40. Our values of u and u areconsistent with those of KOI-94.01, 02 and 03 reported by Masuda et al. (in prep; u ∼ . u ∼ . u and u simultaneously as free parameters. The consistencysupports that KOI-94.04 is orbiting the same host star. In contrast, the u parameter ofthe data sets phase-folded with the catalogued period (i.e. 0 σ ) is u = 0 . +0 . − . , whichis inconsistent with those of the other planets. The discrepancy implies KOI-94.04 doesnot transit KOI-94, but its χ parameter is worse than that of − . σ data sets. The χ parameter suggests that true KOI-94.04’s period may not be the value on the cataloguebut slightly shifted (our result P = 3 . P = 3 . ± .
5. Summary
We have focused on the multiple planetary system KOI-94 and conducted high-contrastdirect imaging observations with Subaru/HiCIAO in order to examine a possibility of falsepositives. As a result of our classical ADI analysis, we have discovered a faint object at aseparation of 0.6 arcsecs with a contrast of ∆ H ∼ × − to the north. Our estimates ofits magnitude in Kp revealed that the insubstantial binary in the background can explainthe depths of KOI-94.04 as a false positive. We have also excluded the possibility of a falsepositive of KOI-94.02 because the depths of KOI-94.02 is almost equal to or larger thanthe contrast of the faint object, assuming the companion candidate being a real companionor a background star with the typical color. On the other hand, our transit analyses showthe limb darkening parameter of KOI-94.04 is consistent with those of the other planetarycandidates in the KOI-94 system, suggesting that KOI-94.04 might orbit the same hoststar. Although we cannot conclude that KOI-94.04 is a planet from our results, we havedemonstrated that the combination of the direct imaging and analysis of transit light curvecan constrain the possibility of a false positive.Our results also suggest that it is not enough to conclude the possibility of falsepositives by shallow or low-contrast direct imaging observations, which have been oftenconducted so far. Considering the fact that Weiss et al. (2013) failed to find the companioncandidate B due to their detection limits, our results suggest to require deep observationswith ∆ m &
10 mag at a few arcsecs for the confirmation of a planet with its transit depth of .
100 ppm rather than conventional and shallow observations for excluding false positives.Furthermore, the discovery of the companion candidate B, if confirmed to be bound, gives 19 –an important clue to dynamical evolution of the planets in the KOI-94 system.This paper is based on data collected at the Subaru telescope, operated by NationalAstronomical Observatory of Japan. We thank the special support for HiCIAO and AO188observations by the staffs. We are also grateful to NASA’s
Kepler
Mission, which obtainedphotometric data of KOI-94. The data analysis was operated on common use data analysisat the Astronomy Data Center, ADC, of the National Astronomical Observatory of Japan.The works by Y.H.T. and T.H. are supported by Japan Society for Promotion of Science(JSPS) Fellowship for Research (DC1: 23-271, No. 25-3183). N.N. acknowledges supportsby NAOJ Fellowship, NINS Program for Cross-Disciplinary Study, and Grant-in-Aid forScientific Research (A) (No. 25247026) from the Ministry of Education, Culture, Sports,Science and Technology (MEXT) of Japan. This work is partly supported by the JSPS fund(No. 22000005). The work by J.C.C. is supported by the U.S. National Science Foundationunder Award No. 1009203. 20 –
REFERENCES
Agol, E., Steffen, J., Sari., R., & Clarkson, W. 2005, MNRAS, 359, 567Baraffe, I., Chabrier, G., Allard, F., & Hauschildt, P. H. 1998, A&A, 337, 403Barclay, T., Rowe, J. F., Lissauer, J. J., et al. 2013, Nature, 494, 452Batalha, N. M., Borucki, W. J., Koch, D. G., et al. 2010, ApJ, 713, L109Batalha, N. M., Rowe, J. F., Bryson, S. T., et al. 2012, ApJS, 204, 24Borucki, W. J., Koch, D. G., Basri, G., et al. 2011, ApJ, 736, 19Carson, J., Thalmann, C., Janson, M., et al. 2013, ApJ, 763, L32Chatterjee, S., Ford, E. B., Matsumura, S., & Rasio, F. A. 2008, ApJ, 686, 580Currie, T., Debes, J., Rodigas, T. J. et al. 2013, ApJ, 760, L32Demarque, P., Woo, J.-H., Kim, Y.-C., & Yi, S. K. 2004, ApJS, 155, 667Fabrycky, D., & Tremaine, S. 2007, ApJ, 669, 1298Fressin, F., Torres, G., Charbonneau, D., et al. 2013, arXiv:1301.0842Hayano, Y., Takami, H., & Oya, S., et al. 2010, Proc. SPIE, 7736, 21Hirano, T., Narita, N., Sato, B., et al. 2012, ApJ, 759, L36Hodapp, K. W., Suzuki, R., & Tamura, M. 2008, Proc. SPIE, 7014, 42Howell, S. B., Rowe, J. F., Bryson, S. T., et al. 2012, ApJ, 746, 123Ivezi´c, ˇZ., Sesar, B., Juri´c, M., et al. 2008, ApJ, 684, 287Kozai, Y. 1962, AJ, 67, 591 21 –Kuzuhara, M., Tamura, M., Kudo, T., et al. 2013, ApJ, 774, 11Lafreni`ere, D., Marois, C., Doyon, R., Nadeau, D., & Artigau, `E. 2007, ApJ, 660, 770Latham, D. W., Bakos, G. ´A., Torres, G., et al. 2009, ApJ, 704, 1107Marois, C., Lafreni`ere, D., Doyon, R., Macintosh, B., & Nadeau, D. 2006, ApJ, 641, 556Morton, T. D., & Johnson, J. A. 2011, ApJ, 738, 170Nagasawa, M., & Ida, S. 2011, ApJ, 742, 72Narita, N., Kudo, T., Bergfors, C., et al. 2010, PASJ, 62, 779Narita, N., Takahashi, Y. H., Kuzuhara, M., et al. 2012, PASJ, 64, L7Ohta, Y., Taruya, A., & Suto, Y. 2005, ApJ, 622, 1118Ohta, Y., Taruya, A., & Suto, Y. 2009, ApJ, 690, 1Santerne, A., D´ıaz, R. F., Moutou, C., et al. 2012, A&A, 545, 76Santerne, A., Fressin, F., D´ıaz, R. F., et al. 2013, arXiv:1307.2003Skrutskie, M. F., Cutri, R. M., Stiening, R., et al. 2006, AJ, 131, 1163Suzuki, R., Kudo, T., Hashimoto, J., et al. 2010, Proc. SPIE, 7735, 101Tamura, M. 2009, American Institute of Physics Conference Series, 1158, 11Thalmann, C., Carson, J., Janson, M., et al. 2009, ApJ, 707, L123Torres, G., Konacki, M., Sasselov, D. D., & Jha, S. 2004, ApJ, 614, 979Weiss, L. M., Marcy, G. W., Rowe, J. F., et al. 2013, ApJ, 768, 14Wu, Y., Murray, N. W., & Ramsahai, J. M. 2007, ApJ, 670, 820 22 –Zacharias, N., Monet, D. G., Levine, S. E., et al. 2004, Bulletin of the AmericanAstronomical Society, 36, 1418This manuscript was prepared with the AAS L A TEX macros v5.2. 23 –Table 1: Properties of planetary candidates in KOI-94Candidate Period (days) Depth (ppm) M pl ( M ⊕ ) R pl ( R ⊕ )KOI-94.01 (d) 22 . ± . ±
11 11 . ± . . ± . . +5 . − . . ± . . ± . +18 − . ± . . ± . . ± . . ± . ′′ ) Position Angle (deg) Contrast ( × − ) ∆ H (mag)0 . ± . . ± .
34 13 . ± . . ± . H is the H band flux ratio to thecentral star. 25 – + Fig. 1.— An H band image of KOI-94 reduced by the classical ADI analysis. The positionof the central star is shown as a cross. A faint companion candidate appears at a separationof ∼ . ′′ -5 -4 -3 -2 -1
0 1 2 3 4 5 6 7 2 4 6 8 10 12 C on t r a s t t o K O I - m [ m ag ] Separation from KOI-94 [arcsec]KOI-94 B KOI-94.02KOI-94.04Limits in H band (This work)KOI-94 B in H band (This work)Limits in Ks band by Weiss et al. (2013) Δ Fig. 2.— A comparison between detection limits for our observations and the contrast ofthe companion candidate B, with the depths of KOI-94.02 and 04. The solid line showsa 5 σ contrast curve in H band with Subaru/HiCIAO for KOI-94 overlaid with a positionof KOI-94 B. The size of the circle is larger than its errors. The two dotted lines depicteach depth of planetary candidates in Kp band. The arrows represent limits by Weiss et al.(2013) in K S band with MMT/ARIES for reference. 27 – R e l a t i v e F l u x typical error 0 σ data0 σ model-3.0 σ data-3.0 σ model-1.9 σ data-1.9 σ model-8.0 × -5 × -5 -0.2 -0.15 -0.1 -0.05 0 0.05 0.1 0.15 0.2 O - C Time from Transit Center [days]
Fig. 3.— Three KOI-94.04 light curves phase-folded by three periods as a function of timefrom the transit centers. The dots and lines in the upper panel represent observational andmodeled relative flux data. Typical error size for each dot is shown at the lower left. Thecolors reflect differences of KOI-94.04’s period; 0 σ (i.e. not shifted), − . σ and − . σ (thelowest χ ) shifted from the KOI-catalogued period are green, blue and red, respectively.Lower panel shows residuals. 28 –
900 920 940 960 980 1000 1020 1040 1060 -3 -2.5 -2 -1.5 -1 -0.5 0 0.5 1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 χ u , R p / R s × Time Shifts in σ χ u R p /R s × Fig. 4.— Variations of parameters for KOI-94.04 as a function of its period. The red lineindicates variation of the χ values. The green and blue dots with error bars represent thelimb-darkening u2