Search for a circum-planetary material and orbital period variations of short-period Kepler exoplanet candidates
AAstron. Nachr. / AN , No. 88, 789 – ?? (2006) / DOI please set DOI!
Search for a circum-planetary material and orbital period variations ofshort-period Kepler exoplanet candidates
Z. Garai ,(cid:63) , G. Zhou , J. Budaj , , and R.F. Stellingwerf Astronomical Institute, Slovak Academy of Sciences, 059 60 Tatranska Lomnica, Slovak Republic Stellingwerf Consulting, 11033 Mathis Mtn Rd SE, Huntsville, AL 35803, United States of America Research School of Astronomy and Astrophysics, Australian National University, Canberra, ACT 2611, AustraliaReceived dd mmm yyyy, accepted dd mmm yyyyPublished online dd mmm yyyy
Key words
Stars: planetary systems – techniques: photometricA unique short-period ( P = 0 . days) Mercury-size Kepler exoplanet candidate KIC012557548b has been dis-covered recently by Rappaport et al. (2012) . This object is a transiting disintegrating exoplanet with a circum-planetarymaterial – comet-like tail. Close-in exoplanets, like KIC012557548b, are subjected to the greatest planet-star interactions.This interaction may have various forms. In certain cases it may cause formation of the comet-like tail. Strong interactionwith the host star, and/or presence of an additional planet may lead to variations in the orbital period of the planet.Our main aim is to search for comet-like tails similar to KIC012557548b and for long-term orbital period variations. Weare curious about frequency of comet-like tail formation among short-period Kepler exoplanet candidates. We concentrateon a sample of 20 close-in candidates with a period similar to KIC012557548b from the Kepler mission.We first improved the preliminary orbital periods and obtained the transit light curves. Subsequently we searched for thesignatures of a circum-planetary material in these light curves. For this purpose the final transit light curve of each planetwas fitted with a theoretical light curve, and the residuals were examined for abnormalities. We then searched for possiblelong-term changes of the orbital periods using the method of phase dispersion minimization.In 8 cases out of 20 we found some interesting peculiarities, but none of the exoplanet candidates showed signs of acomet-like tail. It seems that the frequency of comet-like tail formation among short-period Kepler exoplanet candidatesis very low. We searched for comet-like tails based on the period criterion. Based on our results we can conclude thatthe short-period criterion is not enough to cause comet-like tail formation. This result is in agreement with the theoryof the thermal wind and planet evaporation (Perez-Becker & Chiang 2013). We also found 3 cases of candidates whichshowed some changes of the orbital period. Based on our results we can see that orbital period changes are not caused bycomet-like tail disintegration processes, but rather by possible massive outer companions. c (cid:13) Close-in exoplanets are subjected to the greatest planet-starinteractions. It may have various forms. (1) Heavy irradia-tion changes the atmospheric structure and creates a deeptemperature plateau or a stratosphere (Hubeny, Burrows &Sudarsky 2003; Burrows, Budaj & Hubeny 2008; Knutsonet al. 2008; Fortney et al. 2008). (2) Strong irradiation drivesthe mass loss from the planet (Burrows & Lunine 1995;Guillot et al. 1996). It was detected in HD 209458b (Vidal-Madjar et al. 2003, 2004), and in HD 189733b (Lecave-lier des Etangs et al. 2010; Bourrier et al. 2013). Severaltheoretical studies were devoted to this subject (e.g. Yelle2004; Tian et al. 2005; Hubbard et al. 2007; Lopez, Fort-ney & Miller 2012).
Kurokawa & Kaltenegger (2013) de-veloped a combined model of atmospheric mass loss calcu-lation and thermal evolution calculation of a planet.
Owen &Wu (2013) concluded that evaporation is the driving force ofevolution for close-in Kepler exoplanets. It is therefore plau- (cid:63)
Corresponding author: e-mail: [email protected] sible to assume that some close-in exoplanets may be sur-rounded by a circum-planetary material. (3) Strong interac-tion with the host star forces the planet into synchronizationand circularization of its rotation and orbit (Zahn 1977; Tas-soul & Tassoul 1992; Bodenheimer, Lin & Mardling 2001).(4) Strong interaction with the host star, and/or presence ofan additional planet may lead to variations in the orbitalperiod of the planet (Agol et al. 2005; Holman & Murray2005; Ford et al. 2011, 2012; Steffen et al. 2012; Maciejew-ski et al. 2013; Mazeh et al. 2013).Kepler mission discovered thousands of new transitingextrasolar planet candidates (Borucki et al. 2011). The un-precedented photometric precission of Kepler has made itpossible to detect transits by Earth size planets (Fressin etal. 2012; Borucki et al. 2012), planetary optical secondaryeclipses and phase variations (see e.g. Jackson et al. 2014;Sanchis-Ojeda et al. 2014; Esteves, De Mooij & Jayaward-hana 2013 and references therein), amongst other subtlephenomena.A unique close-in Mercury-size Kepler exoplanet can-didate KIC012557548b has been discovered recently by c (cid:13) a r X i v : . [ a s t r o - ph . E P ] O c t
90 Z. Garai et al.: Search for a circum-planetary material and orbital period variations. . . R e l a t i v e F l u x Orbital Phase
Fig. 1
The transit light curve of KIC012557548b with thetransit (Borucki et al. 2011; analyzed by Budaj 2013). It isbest represented by an exoplanet with a comet-like tail.
Rappaport et al. (2012) . Unlike all other exoplanets it ex-hibits significant variability in the transit depth (Fig. 1). Theshape of the transit is highly asymmetric, with a significantbrightening just before the eclipse, sharp ingress followedby a smooth egress (Fig. 2). The light curve of this planetwas studied in more detail by
Brogi et al. (2012) , Budaj(2013) , Kawahara et al. (2013) , Croll et al. (2014) , and van Werkhoven et al. (2014) . The planet also has an ex-tremely short orbital period of 0.65356(1) days (15.6854hours). The host star that is apparently being occulted isKIC012557548, a V = 16 magnitude K-dwarf with T eff (cid:39) K. Rappaport et al. (2012) suggested that the planethas size not larger than Mercury, and is slowly disintegrat-ing/evaporating, creating a comet-like tail.
Perez-Becker &Chiang (2013) proposed a model of the atmosheric escapevia the thermal wind. It is effective only for planets whichare less massive than Mercury. Gravity of the more massiveplanets would provide too deep potential barier for the wind.Another close-in Kepler exoplanet candidate KIC8639908b( R p ≤ . R ⊕ ), detected recently also by Rappaport etal. (2014) with an orbital period of 0.910022(5) days, ex-hibits a distinctly asymmetric transit profile, likely indica-tive of the emission of dusty effluents, and reminiscent ofKIC012557548b. The host star has T eff (cid:39) K, it is a V= 15.9 magnitude K-dwarf. Mass loss and possible comet-like tail was detected also in GJ 436b in Ly α by Kulow etal. (2014) .In this paper, we search for evidence of disintegration inthe shortest period Kepler exoplanet candidates, by exam-ining their transit light curves for asymmetric features. Thelayout of the paper is as follows. In
Section 2 , we provide anoverview of our exoplanet sample. In
Section 3 , we describethe data analysis and in
Section 4 our motivation.
Section 5 is the main part of our work. It describes obtained results.Our findings are summarized in
Section 6 . Our main aim is to search for comet-like tails similar toKIC012557548b and for long-term orbital period variations. We are curious about frequency of comet-like tail formationamong short-period Kepler exoplanet candidates. Close-inexoplanets, like KIC012557548b, are subjected to the great-est planet-star interactions. We therefore concentrate on theshort-period exoplanet candidates with periods similar toKIC012557548b. Kepler mission exoplanet candidates aredescribed in the catalog of
Batalha et al. (2013) . From thiscatalog we chose a sample of 20 candidates with the short-est orbital period. An overview of our sample is in
Table 1 .This sample covers the orbital periods in the range of 0.370to 0.708 days, and stellar effective temperatures in the rangeof T eff (cid:39) − K. Consequently, the incident fluxhitting the exoplanet may not be exactly the same as in thecase of KIC012557548b, but may slightly vary by about 1-2orders of magnitude.
Table 1
An overview of our sample (Batalha et al. 2013).This sample covers the orbital periods in the range of 0.370to 0.708 days, and stellar effective temperatures in the rangeof T eff (cid:39) − K. KIC Number Orbital period T eff (days) (K)3848972 0.3705290 52868561063 0.4532875 41886666233 0.5124040 39326047498 0.5187486 53599030447 0.5667281 66936934291 0.5678562 49274055304 0.5710391 502210024051 0.5773752 50628235924 0.5879933 400311774303 0.6140747 627110975146 0.6313298 436910028535 0.6630983 499410468885 0.6640818 501311600889 0.6693163 56278278371 0.6773911 57315513012 0.6793614 537510319385 0.6892040 57199761199 0.6920069 40609473078 0.6938521 53535972334 0.7085982 5495 We used the publicly available 17 quarter Kepler data inthe form of Pre-search Data Conditioning Simple AperturePhotometry (PDCSAP) fluxes. Only the long cadence datawere used, as they are sufficient in revealing the large-scaleasymmetries caused by circum-planetary material. Keplerobservations were treated and analysed similarly as in
Bu-daj (2013) . There is an offset between Kepler fluxes fromdifferent quarters. Consequently, fluxes within each quarterwere normalized to unity. We then improved the prelimi-nary orbital period of the exoplanet (Batalha et al. 2013) c (cid:13) stron. Nachr. / AN (2006) 791 using the method of phase dispersion minimization (PDM)described later. The data were cut into segments each cover-ing one orbital period. Each segment of data was fitted witha linear function. During the fitting procedure the part of thedata covering the transit was excluded from the fit. Conse-quently, the linear trend was removed from each chunk ofdata (including the transit data). The final value of the or-bital period was then found in this detrended data. Finallythe data were phased with this new orbital period. We usedphase 0.5 for transits. This method can effectively removethe long term variability (mainly variability of the host stardue to spots and rotation) while it does not introduce anynonlinear trend to the phase light curve. To reduce the noise,the phased light curve was subject to a running window av-eraging. We used a window with the width of 0.01 and stepof 0.001 (in units of phase) in all cases.We searched for the signatures of a circum-planetarymaterial in these light curves. For this purpose the finaltransit light curve of each planet was fitted with a theo-retical light curve (see left panels in appendix A), and theresiduals were examined for abnormalities (see mid panelsin appendix A). We employed the Mandel & Agol (2002) transit model with the free parameters: mid transit time T c (in TDB-based BKJD), period P (days), planet-to-star ra-dius ratio R p /R s , normalised semi-major axis ( R p + R s ) /a ,line-of-sight inclination i (deg), and quadratic limb darken-ing parameters q = ( u + u ) and q = 0 . u ( u + u ) − parameterised as per Kipping (2013) . We first compute thelight curve for a single transit epoch, sampled at 0.0005 inphase ( < minute sampling for a 1 day period candidate).The model light curve is then convolved with a box-carwith width of 30 minutes, simulating the integration timeof a long-cadence exposure. This template transit modelis then interpolated using a B-spline, and evaluated at thetime stamps of the observed light curve to arrive at the fi-nal model. To speed up the computation process, we reducethe number of points by selecting only the in-transit and 0.1phase of the out-of-transit portions of the light curve for thefitting. This cropping of the light curve is justified since weare interested in only the distortions to the transit shape, notthe out-of-transit variations. To find the best fit parametersand explore the degeneracies and uncertainties, we performa Markov Chain Monte Carlo (MCMC) minimisation usingthe emcee ensemble sampler (Foreman-Mackey et al. 2013).To better account for sources of errors in the photometry, in-dividual measurement errors of the light curves are inflatedsuch that the reduced χ is at unity before the MCMC rou-tine, enabling a more realistic estimate the posterior prob-ability distribution. The best fit parameters and their errorsare derived from the marginalised posterior probability dis-tributions after the MCMC analysis.We also searched for possible long-term changes of theorbital period. First we improved the preliminary orbital pe-riod of the extrasolar planet using the method of phase dis-persion minimization (PDM – Stellingwerf 1978; applica- tion PDM2 ver. 4.13 – Stellingwerf 2004) and then usingthe Fourier analysis (FA). The detrended data were used forthis purpose (as described earlier in this section). We as-sumed that the period changes linearly: P = P + βt (1)where β is a dimensionless value, but is often expressedin days/million years (d/Myr). The output from the analy-sis (PDM2 4.13 in this case) is a curve (see right panels inappendix A), which shows the dependence of Θ min as afunction of β . Θ min is a dimensionless statistical parameter(Stellingwerf 1978). The minimum value of Θ min indicatesthe value of the period change – β .For candidates where the PDM analysis indicated poten-tial period variations, we broke the light curve into segmentsof 2000/4000 points, and re-performed the MCMC analy-sis. The minimisation is performed globally such that allthe light curve segments are fitted simultaneously, allowingindividual T c values for each segment of the light curve, butshared transit geometry parameters R p /R s , ( R p + R s ) /a ,and i . We fixed P to the best fit period from the initial anal-ysis. The fitted T c values are then inspected for non-linearvariations with respect to the centre time of each light curvesegment. To test the effectiveness of our techniques, we first appliedthe procedure, described above, to this exoplanet candidate. R e l a t i v e F l u x Orbital Phase
Fig. 2
The transit light curve of KIC012557548b(Borucki et al. 2011; analyzed by Budaj 2013), croppedto enhance the asymmetric features. It shows a significantpre-transit brightening, sharp ingress followed by a shortsharp egress and long smooth egress, and a weak post-transit brightening.KIC012557548b can be used as an extreme example ofthe exoplanet with a circum-planetary material. It was dis-covered by
Rappaport et al. (2012) . The light curve of theKepler exoplanet candidate KIC012557548b is peculiar and c (cid:13)(cid:13)
Rappaport et al. (2012) . The light curve of theKepler exoplanet candidate KIC012557548b is peculiar and c (cid:13)(cid:13)
92 Z. Garai et al.: Search for a circum-planetary material and orbital period variations. . .
Fig. 3
The transit light curve of KIC012557548b with thetransit (Borucki et al. 2011; analyzed by Budaj 2013). It wasfitted with a theoretical light curve similarly as light curvesin appendix A (see Section 3 and left panels in appendix A).very interesting. It shows a significant brightening just be-fore the eclipse – pre-transit brightening, sharp ingress fol-lowed by a short sharp egress and long smooth egress, anda weak post-transit brightening (Brogi et al. 2012; Budaj2013). A close-up of the transit features are shown in
Fig.2 . Moreover, the candidate exhibits strong variability inthe transit core on timescale of one day (Rappaport et al.2012), and variability in the egress on the timescale of about1.3 years (Budaj 2013).
Rappaport et al. (2012) suggestthat KIC012557548b is a slowly disintegrating/evaporatingplanet what creates a comet-like tail.
Brogi et al. (2012) and
Budaj (2013) reanalyzed the light curve in detail andboth validated the disintegrating-planet scenario by model-ing. Both brightenings are caused by the forward scatteringon dust particles in the tail, which have typical radii of about0.1-1 micron. Strong variability in the transit depth is a con-sequence of changes in the cloud optical depth.Subsequently, we fitted the observed light curve with thetheoretical model light curve, assuming a spherical planet(Fig. 3), and obtained transit residuals of KIC012557548b(Fig. 4) as per
Section 3 . The best fit quadratic limb dark-ening coefficients were q = 0 . and q = 0 . .The best fit system parameters are: mid transit time T c =0 . (BKJD), period P = 0 . days, planet-to-star radius ratio R p /R s = 0 . , normalised semi-majoraxis ( R p + R s ) /a = 0 . , and line-of-sight inclination i = 58(4) (deg). Fig. 4 shows a significant fluctuation inthese residuals. This signature of the residuals indicate thecircum-planetary material. The residuals also reveal the pre-transit brightening between . < P < . , as well asthe smooth egress between . < P < . . We can seethe pre-transit brightening as a positive fluctuation, and thesmooth egress as a negative fluctuation. Large scatter ob-served between . < P < . indicate the variability in Fig. 4
The transit residuals of KIC012557548b obtainedsimilarly as transit residuals in appendix A (see Section 3and mid panels in appendix A). A standard Mandel-Agoltransit fits this candidate poorly. We can see the pre-transitbrightening between . < P < . , as well as thesmooth egress between . < P < . . Large scatterobserved between . < P < . indicate the variabilityin the transit depth.the transit depth. These additional features of transit residu-als suggest the comet-like tail. T he t a_ m i n Beta [day/Myr]
Fig. 5
Serach for a long-term change of the orbital periodof KIC012557548b based on 17 quarter data. Figure showsthat there is no significant evidence for the long-term periodchange.We then searched for evidence of transit-timing vari-ations in KIC012557548b using the PDM method. SinceKIC012557548b is close-in exoplanet candidate that is ap-parently being disintegrated and losing material, one mightexpect all kinds of interaction that could lead to the long-term period variation. That is why
Budaj (2013) alsosearched for possible long-term changes of the orbital pe-riod with the PDM method (Stellingwerf 1978; applicationPDM2 4.13 – Stellingwerf 2004). He obtained β = 0 . ± . d/Myr, which means that there is no significant evidence forthe long-term period change. Since Budaj (2013) searched c (cid:13) stron. Nachr. / AN (2006) 793 for possible long-term changes of the orbital period basedon 14 quarters of Kepler data, we repeated this analysis us-ing 17 quarters of data. We can confirm the author’s resultthat there is no significant evidence for the long-term orbitalperiod variation: β = 0 . ± . d/Myr (Fig. 5). We also im-proved the orbital period of the exoplanet KIC012557548band obtained P = 0 . days using the PDM and P = 0 . days using the FA method. These resultsare very similar to the fit period result. These values are alsoin the good agreement with the values reported in Rappa-port et al. (2012) , Budaj (2013) , and van Werkhoven et al.(2014).
The following subsections include a description of the can-didates, formated as per our discussion of KIC012557548b.Each subsection includes results from both (1) analysis ofthe transit residuals, where we searched for comet-like tailssimilar to KIC012557548b, and (2) analysis of the orbitalperiods, where we searched for long-term orbital periodvariations. Interesting exoplanet candidates are described inindividual subsections. The last subsection includes the ex-oplanets that do not show any significant features.
The Kepler exoplanet candidate KIC3848972 is interestingfor several reasons. Although the transit light curve (Fig. A1– see left panel) does not show signs of a pre-transit bright-ening, post-transit brightening, nor significant variability inthe transit depth, an other hand its transit is significantlydeeper (0.002) then transits of other planets from our sam-ple, and it has a V-type shape, which might indicate a graz-ing eclipsing binary. Subsequently, we fitted the light curve(Fig. A1 – see left panel) and obtained transit residuals ofKIC3848972 (Fig. A1 – see mid panel). The transit resid-uals of KIC3848972 show only a linear downward trendfrom sine-like background variability. We obtained the bestfit quadratic limb darkening coefficients of q = 0 . and q = 0 . for the star. The best fit system parameters are:mid transit time T c = 534 . (BKJD), period P =0 . days, planet-to-star radius ratio R p /R s = 0 . ,normalised semi-major axis ( R p + R s ) /a = 0 . , andline-of-sight inclination i = 49(6) (deg). The parameter R p /R s = 0 . , presented by Batalha et al. 2013 , is verydifferent from what we got for R p /R s , due to the degener-acy between R p /R s , inclination i , and ( R p + R s ) /a . Indeed Col´on, Ford & Morehead (2012) , based on a significantcolor change during the transit event, identify KIC3848972as a false positive, which may consist of an evolved giantstar that is redder and several magnitudes brighter than theeclipsing star.
Slawson et al. (2011) found an orbital periodfor this system that is twice as long. Moreower, we founda small difference between the odd and even transits in this system (0.000140) and a weak periodic background vari-ability, which is very similar to an RS Canum Venaticorum-like distortion wave (see e.g. Luddington 1978). That is whywe consider the eclipsing binary alternative as possible. Wealso observed a similar peculiarity at KIC9761199 (see sub-section 5.7). On
Fig. 6 we can see that the sine-like dis-tortion wave is extends throughout the double-period phasefolded light curve. The distortion wave is slightly shiftingin phase during the Kepler observations, but we always ob-served only one sine-wave. The background variations maypotentially be due to spot modulation from one of the stars. R e l a t i v e F l u x Orbital Phase
Fig. 6
The double-period phase-folded light curve ofKIC3848972. The system exhibits a small difference be-tween the odd and even transits, and a sinusoidal back-ground variability.Next we searched for long-term orbital period changes.First we improved the orbital period. We obtained pe-riod P = 0 . days using the PDM and P =0 . days using the FA method, which are very sim-ilar to the fit period result. These values are also in the goodagreement with the value P = 0 . days, reported in Batalha et al. (2013) . However, based on our period anal-ysis, the two times alias orbital period is also possible. Ifthe period is indeed two times longer, both stars would havevery similar temperatures.
Fig. A1 – right panel shows thatthere is no significant evidence for the long-term periodchange ( β = − . ± . d/Myr). The Kepler exoplanet candidate KIC8561063 is similar tothe previous exoplanet candicate. It shows relatively deeptransit (0.0015) with a V-type shape, which needs to bechecked, since it may be indicative of an eclipsing binary(Fig. A2 – see left panel). An other hand, we do not find sig-nificant difference between the odd and even transits in thissystem. The light curve also does not show any signs of apre-transit brightening, post-transit brightening, nor signifi-cant variability in the transit depth. We modelled the transitlight curve and obtained transit residuals from 0.35 to 0.65 c (cid:13)(cid:13)
Fig. A1 – right panel shows thatthere is no significant evidence for the long-term periodchange ( β = − . ± . d/Myr). The Kepler exoplanet candidate KIC8561063 is similar tothe previous exoplanet candicate. It shows relatively deeptransit (0.0015) with a V-type shape, which needs to bechecked, since it may be indicative of an eclipsing binary(Fig. A2 – see left panel). An other hand, we do not find sig-nificant difference between the odd and even transits in thissystem. The light curve also does not show any signs of apre-transit brightening, post-transit brightening, nor signifi-cant variability in the transit depth. We modelled the transitlight curve and obtained transit residuals from 0.35 to 0.65 c (cid:13)(cid:13)
94 Z. Garai et al.: Search for a circum-planetary material and orbital period variations. . . in phase units. They are depicted in
Fig. A2 – mid panel ,and also do not show any signs of a comet-like tail. Thefitting derived model has quadratic limb darkening coeffi-cients of q = 0 . and q = 0 . . The best fitsystem parameters are: mid transit time T c = 488 . (BKJD), period P = 0 . days, planet-to-star radius ra-tio R p /R s = 0 . , normalised semi-major axis ( R p + R s ) /a = 0 . , and line-of-sight inclination i = 76(2) (deg).Subsequently, we improved the orbital period and ob-tained P = 0 . days using the PDM and P =0 . days using the FA method, which are very sim-ilar to the fit period result. However, based on our periodanalysis, the two times alias orbital period is also possible.If the period is indeed two times longer, both stars wouldhave the same temperatures. Fig. A2 – right panel showsthat there is no significant evidence for the long-term periodevolution ( β = 0 . ± . d/Myr). Its transit (Fig. A3 – left panel) is shallower (0.00037) thentransits mentioned above, and has an U-shape, which is typ-ical for exoplanet transits. At the same, we detected a smalldifference between the odd and even transits (0.000054; seeFig. 7), potentially indicative of an eclipsing binary systemwith twice the orbital period.Subsequently, we modelled the transit light curve (Fig.A3 – see left panel). The best fit limb darkening coeffi-cients of the star were q = 0 . and q = 0 . .The best fit system parameters are: mid transit time T c =546 . (BKJD), period P = 0 . days, planet-to-star radius ratio R p /R s = 0 . , normalised semi-majoraxis ( R p + R s ) /a = 0 . , and line-of-sight inclination i = 85(4) (deg). The light curve and transit residuals (Fig.A3 – left and mid panels) do not show any indication of thecircum-planetary material. R e l a t i v e F l u x Orbital Phase
Fig. 7
The double-period phase-folded light curve ofKIC6666233 with a small depth difference between the oddand even transits in this system. We improved the orbital period and obtained P =0 . days using the PDM and P = 0 . days using the FA method. These values are in the goodagreement with the value P = 0 . days, reported in Batalha et al. (2013) . However, based on our period analy-sis, the two times alias orbital period is still possible. Thereis no significant evidence for the long-term period change( β = − . ± . d/Myr). The light curve of this candidate (Fig. A4 – left panel)exhibits a V-shaped transit with depth of 0.0007, showingno pre-transit brightening, post-transit brightening, nor sig-nificant variability in the transit depth. The transit residu-als also do not show any signs of a comet-like tail (Fig.A4 – see mid panel). The best fit quadratic limb darken-ing coefficients were q = 0 . and q = 0 . .The best fit system parameters are: mid transit time T c =534 . (BKJD), period P = 0 . days, planet-to-star radius ratio R p /R s = 0 . , normalised semi-majoraxis ( R p + R s ) /a = 0 . , and line-of-sight inclination i = 43(5) (deg). The parameter R p /R s = 0 . , pre-sented by Batalha et al. (2013) , is very different from ourderived R p /R s , due to the degeneracy between R p /R s , in-clination i , and ( R p + R s ) /a . We also detected a small dif-ference between the odd and even transits in this system(0.000135; see Fig. 8), which is similar to the differencemeasured at the candidate KIC3848972. R e l a t i v e F l u x Orbital Phase
Fig. 8
The double-period phase-folded light curve ofKIC6047498 with a small difference between the odd andeven transits in this system.Further we searched also for a long-term orbital pe-riod change. First we improved the orbital period. We ob-tained period P = 0 . days using the PDM and P = 0 . days using the FA method, which are inthe good agreement with the fit period result, and also withthe value presented by Batalha et al. (2013) , with the possi-bility of the two times alias.
Fig. A4 – right panel shows thatthere is possible evidence for the long-term period change, c (cid:13) stron. Nachr. / AN (2006) 795 Fig. 9
The T c -free MCMC analysis in case ofKIC6047498. Figure shows the O-C diagram of midtransit time variations. The O-C exhibits a semi-periodicTTV variability, indicative of an outer companion with anorbital period of > days. β = − . ± . d/Myr, that means a potentialshortening of the orbital period. Since the PDM analysis in-dicated a potential period variation, we broke the light curveinto segments of 4000 points in this case, and re-performedthe T c -free MCMC analysis. This analysis confirmed our re-sults from the PDM analysis about long-term period change.We found a semi-periodic TTV signal associated with thetransit, indicative of an outer companion with an orbital pe-riod of > days (Figs. 9 and 10). The light curve of the exoplanet candidate KIC9030447 ispeculiar (Fig. A5 – see left panel). It shows a small bright-ening, approximately half a phase from the transit, and asecondary eclipse-like signal approximately at phase 0.85.Subsequently, we obtained the model light curve and tran-sit residuals of the exoplanet candidate. The best model fitdid not describe the transit well. The best fit quadratic limbdarkening coefficients of the star were q = 0 . and q = 0 . . The best fit system parameters are: mid tran-sit time T c = 534 . (BKJD), period P = 0 . days, planet-to-star radius ratio R p /R s = 0 . , nor-malised semi-major axis ( R p + R s ) /a = 1 . , andline-of-sight inclination i = 85(1) (deg). The transit resid-uals of this exoplanet candidate reflect the peculiarities ofthe light curve, and residual fluctuations due to the poor fit(see Fig. A5 – mid panel). Ofir & Dreizler (2013) proposedthat such behavior is caused by pulsations, rather than tran-sits or eclipses. Further we discovered that these additionalfeatures of the light curve slighly shifted in phase duringthe Kepler observations. Moreover, the secondary eclipse-like signal is changing its depth. That is why we consider
Fig. 10
The transit for each segment of the KIC6047498light curve, with best fit model overplotted. The mean tran-sit epochs for each segment are labelled on the left. Eachsegment is arbitrarily shifted vertically for clarity. The solidvertical line shows the mean T c value. The transit centresare visibly shifted from segment to segment.the eclipsing binary on eccentric orbit alternative as morepossible.Subsequently, we improved the orbital period and ob-tained P = 0 . days using the PDM and P =0 . days using the FA method. Fig. A5 – right panel c (cid:13)(cid:13)
The transit for each segment of the KIC6047498light curve, with best fit model overplotted. The mean tran-sit epochs for each segment are labelled on the left. Eachsegment is arbitrarily shifted vertically for clarity. The solidvertical line shows the mean T c value. The transit centresare visibly shifted from segment to segment.the eclipsing binary on eccentric orbit alternative as morepossible.Subsequently, we improved the orbital period and ob-tained P = 0 . days using the PDM and P =0 . days using the FA method. Fig. A5 – right panel c (cid:13)(cid:13)
96 Z. Garai et al.: Search for a circum-planetary material and orbital period variations. . . shows that there is no significant evidence for the long-termperiod change ( β = 0 . ± . d/Myr). On first glance, the light curve of this exoplanet candidatedoes not show any significant peculiarity. The transit resid-uals also do not show any convincing systematic fluctua-tions, and indicate that the light curve can be describedwell by the standard transit model (Fig. A10 – left andmid panels). The best fit quadratic limb darkening coeffi-cients of the star were q = 0 . and q = 0 . .The best fit system parameters are: mid transit time T c =457 . (BKJD), period P = 0 . days, planet-to-star radius ratio R p /R s = 0 . , normalised semi-majoraxis ( R p + R s ) /a = 0 . , and line-of-sight inclination i = 37(11) (deg). However, Ofir & Dreizler (2013) detectedsignificant differences between the odd and even transits inthis system, and suggested that this candidate is an eclips-ing binary with twice the period. We also measured a transitdepth difference (0.000228; Fig. 11) and we could confirmtheir conclusions about transit differences. Based on this re-sult we consider the eclipsing binary alternative as possible. R e l a t i v e F l u x Orbital Phase
Fig. 11
The double-period phase-folded light curve ofKIC11774303 with a difference in depth between the oddand even transits in this system.An other hand, we improved the orbital period and ob-tained period P = 0 . days using the PDM and P = 0 . days using the FA method, in agreementto the fit period result, but we could not confirm the twotimes period alias using the PDM/FA analysis. Fig. A10 –right panel shows that there is no significant evidence forthe long-term period change ( β = 0 . ± . d/Myr). The Kepler exoplanet candidate KIC9761199 is also veryinteresting. Its transit (Fig. A18 – see left panel) is verydeep (0.0045) and has a V-type shape, which might indi-cate a grazing eclipsing binary. The transit light curve does not show any signs of a pre-transit brightening, post-transitbrightening, nor significant variability in the transit depth.Apart from that, the out-of-transit light curve exhibits a pe-riodic background variability, which is very similar to an RSCanum Venaticorum-like distortion wave (see e.g. Ludding-ton 1978). That is why we consider the eclipsing binary al-ternative as possible. The background variability is also verysimilar to the distortion wave observed by KIC3848972 (seesubsection 5.1), but contrary to the KIC3848972, this waveis stronger. Subsequently, we fitted the light curve with amodel light curve (Fig. A18 – see left panel) and obtainedtransit residuals of this exoplanet candidate from 0.35 to0.65 in phase units (Fig. A18 – mid panel). The best fitquadratic limb darkening coefficients were q = 0 . and q = 0 . . The best fit system parameters are:mid transit time T c = 533 . (BKJD), period P =0 . days, planet-to-star radius ratio R p /R s = 0 . ,normalised semi-major axis ( R p + R s ) /a = 0 . , andline-of-sight inclination i = 68(2) (deg). The parameter R p /R s = 0 . , presented by Batalha et al. (2013) , isvery different from what we got for R p /R s , due to the de-generacy between R p /R s , inclination i , and ( R p + R s ) /a .The transit residuals show only a relatively stronger lineartrend from sine-like background variability. Contrary to theKIC3848972, the linear trend in this case has an upward di-rection. Ofir & Dreizler (2013) proposed that KIC9761199is an eclipsing binary with twice the orbital period as sug-gested by
Batalha et al. (2013) . There is indeed a smalldifference of 0.000704 between the odd and even transits(see Fig. 12). On
Fig. 12 we can see that the sine-like dis-tortion wave is extends throughout the double-period phasefolded light curve. Subsequently, we discovered that thesine-wave slightly shifted in phase during the Kepler obser-vation. Sometimes we could observe double sine-wave andsometimes single sine-wave throughout the double-periodphase folded light curve. R e l a t i v e F l u x Orbital Phase
Fig. 12
The double-period phase-folded light curve ofKIC9761199 with a small difference between the odd andeven transits in this system and with a sinusoidal back-ground variability. c (cid:13) stron. Nachr. / AN (2006) 797 The background variability may potentially be due tothe Doppler beaming effect, sometimes also called Dopplerboosting, induced by the stellar radial motion. This effectcauses an increase (decrease) in the brightness of any lightsource approaching (receding from) the observer (e.g. Ry-bicki & Lightman 1979; Loeb & Gaudi 2003; Faigler &Mazeh 2011). We did not see any evidence of ellipsoidaland reflection phase variations from this system. In addition,the sinusoidal variability shifted in phase during the Keplerobservations, suggesting that it is not due to beaming. Thebackground variations may potentially be due to stellar spotmodulation from both of the stars when we observe dou-ble sine-wave, and from one of the stars, when we observesingle sine-wave throughout the double-period phase foldedlight curve.Further we searched also for long-term orbital periodchange. First we improved the orbital period. We obtainedperiod P = 0 . days using the PDM and P =0 . days using the FA method. These values are inthe good agreement with the value P = 0 . days ob-tained by the fitting procedure, and also in the good agree-ment with the value P = 0 . days reported by Batalha et al. (2013) , but based on our period analysis, thetwo times period alias could not be ruled out.
Fig. A18 –right panel shows that there is a possible evidence for thelong-term period change, with β = − . ± . d/Myr, which means a potential shortening of the orbitalperiod. Since the PDM analysis indicates a potential periodvariation, we broke the light curve into segments of 2000points in this case, and re-performed the T c -free MCMCanalysis. This analysis confirmed our results from the PDManalysis about long-term period change, and revealed a pe-riodic TTV signal with a period of ∼ days, indicativeof a possible massive outer companion (Figs. 13 and 14). Fig. 13
The T c -free MCMC analysis in case ofKIC9761199. Figure shows the O-C diagram of mid tran-sit time variations. We found a periodic TTV signal witha period of ∼ days, indicative of a possible massiveouter companion. Fig. 14
The transit for each segment of the KIC9761199light curve, with best fit model overplotted. For details see
Fig. 10 . The Kepler exoplanet candidate KIC5972334 is another in-teresting object in our sample. We first obtained the lightcurve using the period P = 0 . days reported by Batalha et al. (2013) . It is shown in
Fig. 15 . We did notfind a convincing transit signal at this orbital period. Wesearched for periods and obtained P = 15 . days c (cid:13)(cid:13)
Fig. 15 . We did notfind a convincing transit signal at this orbital period. Wesearched for periods and obtained P = 15 . days c (cid:13)(cid:13)
98 Z. Garai et al.: Search for a circum-planetary material and orbital period variations. . . using the PDM and P = 15 . days using the FAmethod. We did not confirm the preliminary orbital period.According to these results the light curve of KIC5972334was reanalysed with P = 15 . days. Steffen et al.(2010) suggested that this star shows transits from two exo-planets with orbital period P = 15 . days and P = 2 . days. We confirmed the first period, but could not confirmthe second object via PDM/FA period analysis. The transitlight curve folded with the period P = 15 . daysis shown in Fig. A20 – left panel . It is the deepest transitin our sample (0.0135) and has a typical U-shape. We didnot detect significant differences between the odd and eventransits in this system, nor significant pre-transit brighten-ing and post-transit brightening, nor significant variabilityin transit depth. This expected due to the long period na-ture of the candidate. Subsequently, we fitted the light curvewith a model light curve (Fig. A20 – see left panel) andobtained transit residuals of this exoplanet candidate (Fig.A20 – mid panel). The residuals show that this transit canbe reproduced well with the planetary model. We obtainedquadratic limb darkening coefficients of q = 0 . and q = 0 . . The best fit system parameters are: mid tran-sit time T c = 454 . , period P = 15 . days,planet-to-star radius ratio R p /R s = 0 . , normalisedsemi-major axis ( R p + R s ) /a = 0 . , and line-of-sightinclination i = 89 . (deg). The parameter R p /R s =0 . , presented by Batalha et al. (2013) , is very differ-ent from what we got for R p /R s , because it is a differentcandidate at a different period. R e l a t i v e F l u x Orbital Phase
Fig. 15
The light curve of KIC5972334 phased to the or-bital period P = 0 . days, reported in Batalha et al.(2013) .Finaly we searched for possible long-term changes ofthe orbital period.
Fig. A20 – right panel shows that thereis an indication for the long-term period change, with β =33 . ± . d/Myr, which means that the period ispotentially increasing. Since the PDM analysis indicated apotential period variation, we broke the light curve into seg-ments of 2000 points in this case, and re-performed the T c - free MCMC analysis. However, this analysis did not findany TTV signal. The candidates KIC6934291, KIC4055304, KIC10024051,KIC8235924, KIC10975146, KIC10028535,KIC10468885, KIC11600889, KIC8278371, KIC5513012,KIC10319385 and KIC9473078 did not show any pecu-liarities in their transit light curves, nor residuals (Fig. A6,A7, A8, A9, A11, A12, A13, A14, A15, A16, A17 andA19 – left and mid panels). These transits have typicallyshallow U-shapes. We did not detect significant differencesbetween the odd and even transits in these systems. Wefitted the light curves of these candidates with a theoreticallight curve. The best fit parameters of all candidates aresummarized in
Table 2 . We also improved the orbitalperiods reported in
Batalha et al. (2013) via PDM andFA methods. Finaly we searched for possible long-termchanges of the orbital periods. They are all listed in
Table3 . Fig. A6, A7, A8, A9, A11, A12, A13, A14, A15, A16, A17and A19 – right panels show that there are no significantevidences for the long-term period changes. In case ofthe exoplanet candidate KIC10468885 we detected anindication for the long-term period change with β = 3 . d/Myr, which means that the period is increasing (Fig. A13– right panel), however, based on the PDM Monte Carlotest, we could not consider this result as significant enough( std = ± . d/Myr). Our main aim was to search for comet-like tails similarto KIC012557548b and for long-term orbital period vari-ations. We were curious about frequency of comet-like tailformation among short-period Kepler exoplanet candidates.Close-in exoplanets, like KIC012557548b, are subjected tothe greatest planet-star interactions. We therefore concen-trated on the short-period exoplanet candidates with peri-ods similar to KIC012557548b. We chose 20 exoplanet can-didates observed by the Kepler mission, with the shortestorbital periods, ranging from 0.370 up to 0.708 days. Wefirst improved the preliminary orbital periods (Batalha et al.2013) and obtained the transit light curves of our exoplanetcandidates. Subsequently, we searched for the signatures ofa circum-planetary material in these light curves. For thispurpose the final transit light curve of each planet was fit-ted with a theoretical light curve, and the residuals wereexamined for abnormalities. We also searched for possiblelong-term changes of the orbital period using the methodof phase dispersion minimization (PDM). For candidateswhere the PDM analysis indicated potential period varia-tions, we broke the light curve into segments of 2000/4000points, and re-performed the T c -free MCMC analysis. Totest the effectiveness of our techniques, we first appliedthis procedure to KIC012557548b. The light curve of this c (cid:13) stron. Nachr. / AN (2006) 799 Table 2
An overview of the best fit parameters. Table contains the best fit qadratic limb darkening coefficients of thehost stars ( q and q ), and the best fit system parameters: mid transit time T c , period P , planet-to-star radius ratio R p /R s ,normalised semi-major axis ( R p + R s ) /a , and line-of-sight inclination i . KIC Number q q T c P R p /R s ( R p + R s ) /a i (BKJD) (days) (deg)3848972 0.9(1) 0.9(3) 534.34045(5) 0.37053 0.18(8) 0.71(4) 49(6)8561063 0.0001(8) 0.7(1) 488.7930(1) 0.45329 0.044(4) 0.2(1) 76(2)6666233 0.26(4) 0.8(1) 546.2643(5) 0.51241 0.016(2) 0.27(1) 85(4)6047498 0.94(5) 0.95(4) 534.1331(3) 0.51873 0.3(2) 0.76(5) 43(5)9030447 0.47(2) 0.99(2) 534.478(1) 0.56677 0.0091(1) 1.000(8) 85(1)6934291 0.1(1) 0.2(2) 523.6117(4) 0.56786 0.018(8) 0.3(1) 74(14)4055304 0.0006(3) 0.2(2) 546.2270(3) 0.57104 0.015(9) 0.24(1) 85(5)10024051 0.55(2) 0.03(3) 553.9052(3) 0.57737 0.0187(4) 0.25(1) 86(3)8235924 0.02(2) 0.3(3) 459.5745(4) 0.58800 0.018(3) 0.70(1) 47(1)11774303 0.118(9) 0.7(2) 457.3181(7) 0.61408 0.07(4) 0.8(1) 37(11)10975146 0.4(1) 0.8(1) 496.9595(4) 0.63133 0.0184(7) 0.24(1) 87(3)10028535 0.1(1) 0.2(2) 454.9020(7) 0.66309 0.017(1) 0.21(1) 86(3)10468885 0.03(3) 0.5(5 545.814(1) 0.66407 0.015(1) 0.26(6) 79(10)11600889 0.29(7) 0.96(3) 550.4423(4) 0.66931 0.011(1) 0.40(9) 74(12)8278371 0.06(5) 0.88(9) 484.1484(8) 0.67737 0.0086(5) 0.42(9) 73(15)5513012 0.28(4) 0.87(3) 364.5110(5) 0.67934 0.0146(9) 0.25(2) 84(5)10319385 0.02(1) 0.7(1) 584.0281(2) 0.68921 0.014(1) 0.79(7) 40(1)9761199 0.94(5) 0.96(1) 533.1669(1) 0.69202 0.22(9) 0.43(3) 68(2)9473078 0.7(4) 0.3(3) 574.852(1) 0.69384 0.010(4) 0.83(5) 35(5)5972334 0.33(8) 0.4(1) 454.9191(1) 15.35877 0.118(1) 0.038(1) 89.1(1) exoplanet is peculiar, caused by a pre-transit brightening,post-transit brightening and strong variability in the tran-sit depth. Subsequently, we fitted the observed light curvewith the theoretical model light curve, assuming a sphericalplanet, and obtained transit residuals. We observed a signif-icant fluctuation in these residuals. We then searched for along-term orbital period variation using the PDM method.Previously, Budaj (2013) obtained β = 0 . ± . d/Myr us-ing 14 quarters of Kepler data, indicating no change in thetransit period. We repeated the analysis using 17 quarter ofdata, revealing no significant long-term period variation.We then analyzed the transit light curves and residu-als of our sample similarly as per KIC012557548b. Wesearched for significant fluctuations in these residuals,strong variabilities in the transit depth, pre-transit and post-transit brightenings in light curves, which could indicatepresence of the circum-planetary material. In 8 cases out of20 we found some interesting peculiarities (KIC3848972,KIC8561063, KIC6666233, KIC6047498, KIC9030447,KIC11774303, KIC9761199, KIC5972334). We found 4cases of exoplanet candidates with relatively deep V-shapeof the transits (KIC3848972, KIC8561063, KIC6047498,KIC9761199). We detected differences between the oddand even transits in 5 cases (KIC3848972, KIC6666233,KIC6047498, KIC11774303, KIC9761199). In cases ofKIC3848972, KIC8561063, KIC6666233, KIC6047498,KIC11774303 and KIC9761199 we consider the eclipsingbinary alternative as possible. Excepting KIC11774303, thePDM/FA based two times alias orbital period is also pos-sible. We found 2 cases of exoplanet candidates with sinu- soidal out-of-transit periodic background variability, whichis very similar to an RS Canum Venaticorum-like distortionwave (KIC3848972 and KIC9761199). In addition, the si-nusoidal background variability shifted in phase during theKepler observations, suggesting that it is not due to Dopplerbeaming. The background variations may potentially bedue to stellar spot modulation. The exoplanet candidateKIC9030447 is also peculiar, because its light curve showsa small brightening, approximately half a phase from thetransit, and a secondary eclipse-like signal approximately atphase 0.85. Further we discovered that these additional fea-tures of the light curve slighly shifted in phase during theKepler observations. Moreover, the secondary eclipse-likesignal is changing its depth. That is why we consider theeclipsing binary on eccentric orbit alternative as more pos-sible as exoplanet transits or stellar pulsations.None of the exoplanet candidates shows signs of acomet-like tail. We did not find any significant fluctuationsin residuals, strong variabilities in the transit depth, norpre-transit and post-transit brightenings in light curves. Wefound only a linear downward/upward residual trend fromsine-like background variability in cases of exoplanet can-didates KIC3848972 and KIC9761199, and a small residualfluctuation in case of exoplanet candidate KIC9030447 dueto the poor fit. It seems that the frequency of comet-like tailformation among short-period Kepler exoplanet candidatesis very low. We searched for comet-like tails based on theperiod criterion. Based on our results we can conclude thatthe short-period criterion is not enough to cause comet-liketail formation. This result is in agreement with the theory of c (cid:13)(cid:13)
An overview of the best fit parameters. Table contains the best fit qadratic limb darkening coefficients of thehost stars ( q and q ), and the best fit system parameters: mid transit time T c , period P , planet-to-star radius ratio R p /R s ,normalised semi-major axis ( R p + R s ) /a , and line-of-sight inclination i . KIC Number q q T c P R p /R s ( R p + R s ) /a i (BKJD) (days) (deg)3848972 0.9(1) 0.9(3) 534.34045(5) 0.37053 0.18(8) 0.71(4) 49(6)8561063 0.0001(8) 0.7(1) 488.7930(1) 0.45329 0.044(4) 0.2(1) 76(2)6666233 0.26(4) 0.8(1) 546.2643(5) 0.51241 0.016(2) 0.27(1) 85(4)6047498 0.94(5) 0.95(4) 534.1331(3) 0.51873 0.3(2) 0.76(5) 43(5)9030447 0.47(2) 0.99(2) 534.478(1) 0.56677 0.0091(1) 1.000(8) 85(1)6934291 0.1(1) 0.2(2) 523.6117(4) 0.56786 0.018(8) 0.3(1) 74(14)4055304 0.0006(3) 0.2(2) 546.2270(3) 0.57104 0.015(9) 0.24(1) 85(5)10024051 0.55(2) 0.03(3) 553.9052(3) 0.57737 0.0187(4) 0.25(1) 86(3)8235924 0.02(2) 0.3(3) 459.5745(4) 0.58800 0.018(3) 0.70(1) 47(1)11774303 0.118(9) 0.7(2) 457.3181(7) 0.61408 0.07(4) 0.8(1) 37(11)10975146 0.4(1) 0.8(1) 496.9595(4) 0.63133 0.0184(7) 0.24(1) 87(3)10028535 0.1(1) 0.2(2) 454.9020(7) 0.66309 0.017(1) 0.21(1) 86(3)10468885 0.03(3) 0.5(5 545.814(1) 0.66407 0.015(1) 0.26(6) 79(10)11600889 0.29(7) 0.96(3) 550.4423(4) 0.66931 0.011(1) 0.40(9) 74(12)8278371 0.06(5) 0.88(9) 484.1484(8) 0.67737 0.0086(5) 0.42(9) 73(15)5513012 0.28(4) 0.87(3) 364.5110(5) 0.67934 0.0146(9) 0.25(2) 84(5)10319385 0.02(1) 0.7(1) 584.0281(2) 0.68921 0.014(1) 0.79(7) 40(1)9761199 0.94(5) 0.96(1) 533.1669(1) 0.69202 0.22(9) 0.43(3) 68(2)9473078 0.7(4) 0.3(3) 574.852(1) 0.69384 0.010(4) 0.83(5) 35(5)5972334 0.33(8) 0.4(1) 454.9191(1) 15.35877 0.118(1) 0.038(1) 89.1(1) exoplanet is peculiar, caused by a pre-transit brightening,post-transit brightening and strong variability in the tran-sit depth. Subsequently, we fitted the observed light curvewith the theoretical model light curve, assuming a sphericalplanet, and obtained transit residuals. We observed a signif-icant fluctuation in these residuals. We then searched for along-term orbital period variation using the PDM method.Previously, Budaj (2013) obtained β = 0 . ± . d/Myr us-ing 14 quarters of Kepler data, indicating no change in thetransit period. We repeated the analysis using 17 quarter ofdata, revealing no significant long-term period variation.We then analyzed the transit light curves and residu-als of our sample similarly as per KIC012557548b. Wesearched for significant fluctuations in these residuals,strong variabilities in the transit depth, pre-transit and post-transit brightenings in light curves, which could indicatepresence of the circum-planetary material. In 8 cases out of20 we found some interesting peculiarities (KIC3848972,KIC8561063, KIC6666233, KIC6047498, KIC9030447,KIC11774303, KIC9761199, KIC5972334). We found 4cases of exoplanet candidates with relatively deep V-shapeof the transits (KIC3848972, KIC8561063, KIC6047498,KIC9761199). We detected differences between the oddand even transits in 5 cases (KIC3848972, KIC6666233,KIC6047498, KIC11774303, KIC9761199). In cases ofKIC3848972, KIC8561063, KIC6666233, KIC6047498,KIC11774303 and KIC9761199 we consider the eclipsingbinary alternative as possible. Excepting KIC11774303, thePDM/FA based two times alias orbital period is also pos-sible. We found 2 cases of exoplanet candidates with sinu- soidal out-of-transit periodic background variability, whichis very similar to an RS Canum Venaticorum-like distortionwave (KIC3848972 and KIC9761199). In addition, the si-nusoidal background variability shifted in phase during theKepler observations, suggesting that it is not due to Dopplerbeaming. The background variations may potentially bedue to stellar spot modulation. The exoplanet candidateKIC9030447 is also peculiar, because its light curve showsa small brightening, approximately half a phase from thetransit, and a secondary eclipse-like signal approximately atphase 0.85. Further we discovered that these additional fea-tures of the light curve slighly shifted in phase during theKepler observations. Moreover, the secondary eclipse-likesignal is changing its depth. That is why we consider theeclipsing binary on eccentric orbit alternative as more pos-sible as exoplanet transits or stellar pulsations.None of the exoplanet candidates shows signs of acomet-like tail. We did not find any significant fluctuationsin residuals, strong variabilities in the transit depth, norpre-transit and post-transit brightenings in light curves. Wefound only a linear downward/upward residual trend fromsine-like background variability in cases of exoplanet can-didates KIC3848972 and KIC9761199, and a small residualfluctuation in case of exoplanet candidate KIC9030447 dueto the poor fit. It seems that the frequency of comet-like tailformation among short-period Kepler exoplanet candidatesis very low. We searched for comet-like tails based on theperiod criterion. Based on our results we can conclude thatthe short-period criterion is not enough to cause comet-liketail formation. This result is in agreement with the theory of c (cid:13)(cid:13)
00 Z. Garai et al.: Search for a circum-planetary material and orbital period variations. . .
Table 3
An overview of improved orbital periods. Exoplanet candidates signed at their KIC Number with ∗ may havetwice as long orbital period. We did not confirm the preliminary orbital period in case of candidate KIC5972334. Tablealso contains the β parameter. KIC Number Period by
Batalha et al. (2013)
Period by PDM Period by FA Period by MCMC β (days) (days) (days) (days) (d/Myr)3848972 ∗ ± ∗ ± ∗ ± ∗ ± ± ± ± ± ± ± ± ± ± ± ± ± ± ∗ ± ± ± the thermal wind and planet evaporation (Perez-Becker &Chiang 2013). This theory predicts that catastrophic evap-oration, which leads into creation of a comet-like tail, canoccure only in relatively low-mass planets with masses lessthan that of Mercury. Since all the exoplanet candidates outof 20 have planet-to-star radius ratio higher than 0.003 (seeparameter R p /R s in Table 2; the value 0.003 approximatelyvalids for the planet Mercury), they are likely to be moremassive than Mercury and, consequently, lack the catas-trophic evaporation contrary to KIC012557548b.We also found 3 cases of candidates (KIC6047498,KIC9761199 and KIC5972334) which showed somechanges of the orbital period. In one case (KIC5972334) weobserved an orbital period increasing, other exoplanet can-didates showed orbital period shortenings. Since the PDManalysis indicated a potential period variation in these cases,we broke the light curve into segments of 2000/4000 points,and re-performed the T c -free MCMC analysis. In case ofexoplanet candidate KIC6047498 we found a semi-periodicTTV signal associated with the transit, indicative of anouter companion with an orbital period of > days.In case of exoplanet candidate KIC9761199 this analysisalso confirmed our results from the PDM analysis aboutlong-term period change and revealed a periodic TTV sig-nal with a period of ∼ days, indicative of a possi-ble massive outer companion. In case of exoplanet candi-date KIC5972334 this analysis did not find any TTV signal.Based on our results we can see that orbital period changesare not caused by comet-like tail disintegration processes,but rather by possible massive outer companions. Moreover, nor KIC012557548b with comet-like tail disintegration, didnot show any long-term orbital period variation.We also improved the preliminary orbital periods usingPDM/FA methods. We did not confirm the preliminary or-bital period in case of exoplanet candidate KIC5972334. Weobtained period P = 15 . days using the PDM and P = 15 . days using the FA method, which is inthe agreement with period P = 15 . days found by Stef-fen et al. (2010) . We did not confirm the second transitingobject with P over 2.420 days in this system, suggested bythe same authors. The exoplanet candidates KIC3848972,KIC8561063, KIC6666233, KIC6047498 and KIC9761199may have twice as long orbital period. Acknowledgements.
The authors thank Dr. L. Hamb´alek and Dr.T. Krejˇcov´a for the technical assistance, comments and discus-sions. This work was supported by the VEGA grants of the Slo-vak Academy of Sciences Nos. 2/0143/14, 2/0038/13, by the Slo-vak Research and Development Agency under the contract No.APVV-0158-11, and by the realization of the Project ITMS No.26220120029, based on the Supporting Operational Research andDevelopment Program financed from the European Regional De-velopment Fund. JB was supported by the Australian ResearchCouncil through DP120101792. GZ thanks Chelsea X. Huang forher contribution to the light curve fitting program. c (cid:13) stron. Nachr. / AN (2006) 801 References
Agol, E., Steffen, J., Sari, R. & Clarkson, W.: 2005, MNRAS 359,567Batalha, N.M., Rowe, J.F., Bryson, S.T., et al.: 2013, ApJS 204, 24Bodenheimer, P., Lin, D.N.C. & Mardling, R.A.: 2001, ApJ 548,466Borucki, W.J., Koch, D.G., Basri, G., et al.: 2011, ApJ 736, 19Borucki, W.J., Koch, D.G., Batalha, N., et al.: 2012, ApJ 745, 120Bourrier, V., Lecavelier des Etangs, A., Dupuy, H., et al.: 2013,A&A 551, 63Brogi, M., Keller, C.U., de Juan Ovelar, M., Kenworthy, M.A., deKok, R.J., Min, M. & Snellen, I.A.G.: 2012, A&A 545, 5Budaj, J.: 2013, A&A 557, 72Burrows, A., Budaj, J. & Hubeny, I.: 2008, ApJ 678, 1436Burrows, A. & Lunine, J.: 1995, Natur 378, 333Col´on, K.D., Ford, E.B. & Morehead, R.C.: 2012, MNRAS 426,342Croll, B., Rappaport, S., DeVore, J., et al.: 2014, ApJ 786, 100Esteves, L.J. De Mooij, E.J.W. & Jayawardhana, R.: 2013, ApJ772, 51Faigler, S. & Mazeh, T.: 2011, MNRAS 415, 3921Ford, E.B., Fabrycky, D.C., Steffen, J.H., et al.: 2012, ApJ 750,113Ford, E.B., Rowe, J.F., Fabrycky, D.C., et al.: 2011, ApJ 197, 2Foreman-Mackey, D., Hogg, D.W., Lang, D., & Goodman, J.:2013, PASP 125, 306Fortney, J.J., Lodders, K., Marley, M.S. & Freedman, R.S.: 2008,ApJ 678, 1419Fressin, F., Torres, G., Rowe, J.F., et al.: 2012, Natur 482, 195Guillot, T., Burrows, A., Hubbard, W.B., Lunine, J.I. & Saumon,D.: 1996, ApJ 459, 35Holman, M.J. & Murray, N.W.: 2005, Sci 307, 1288Hubbard, W.B., Hattori, M.F., Burrows, A., Hubeny, I. & Su-darsky, D.: 2007, Icar 187, 358Hubeny, I., Burrows, A. & Sudarsky, D.: 2003, ApJ 594, 1011Jackson, B., Stark, C.C., Adams, E.R., Chambers, J. & Deming,D.: 2014, AAS 223, 202Kawahara, H., Hirano, T., Kurosaki, K., Ito, Y. & Ikoma, M.: 2013,ApJ 776, L6Kipping, D.M.: 2013, MNRAS 435, 2152Knutson, H.A., Charbonneau, D., Allen, L.E., Burrows, A. &Megeath, S.T.: 2008, ApJ 673, 526Kulow, J.R., France, K., Linsky, J. & Loyd, R.O.P.: 2014, ApJ 786,132Kurokawa, H. & Kaltenegger, L.: 2013, MNRAS 433, 3239Lecavelier des Etangs, A., Ehrenreich, D., Vidal-Madjar, A. et al.:2010, A&A 514, 72Loeb, A. & Gaudi, B.S.: 2003, ApJ 588, 117Lopez, E.D., Fortney, J.J. & Miller, N.: 2012, ApJ 761, 59Luddington, E.W.: 1978, BAAS 10, 418Maciejewski, G., Dimitrov, D., Seeliger, M., et al.: 2013, A&A551, 108Mandel, K. & Agol, E.: 2002, ApJ 580, 171Mazeh, T., Nachmani, G., Holczer, T., et al.: 2013, ApJS 208, 16Ofir, A. & Dreizler, S.: 2013, A&A 555, 580Owen, J.E. & Wu, Y.: 2013, ApJ 775, 1050Rappaport, S., Barclay, T., DeVore, J., Rowe, J., Sanchis-Ojeda,R., Still, M.: 2014, ApJ 784, 40Rappaport, S., Levine, A., Chiang, E., et al.: 2012, ApJ 752, 1Rybicki, G.B. & Lightman, A.P.: 1979, Radiative Processes in As-trophysics, Wiley, New YorkPerez-Becker, D. & Chiang, E.: 2013, MNRAS 433, 2294 Sanchis-Ojeda, R., Rappaport, S., Winn, J.N., Kotson, M.C.,Levine, A.M., El Mellah, I.: 2014, ApJ 787, 47Slawson, R.W., Pr´sa, A., Welsh, W.F., et al.: 2011, AJ 142, 160Steffen, J.H., Batalha, N.M., Borucki, W.J., et al.: 2010, ApJ 725,1226Steffen, J.H., Ford, E.B., Rowe, J.F., et al.: 2012, ApJ 756, 186Stellingwerf, R.F.: 1978, ApJ 224, 953Stellingwerf, R.F.: 2004, PDM2, Stellingwerf Consulting,Huntsville 2004Tassoul, J.L. & Tassoul, M.: 1992, ApJ 395, 259Tian, F., Toon, O.B., Pavlov, A.A. & De Sterck, H.: 2005, ApJ 621,1049van Werkhoven, T.I.M., Brogi, M., Snellen, I.A.G. & Keller, C.U.:2014, A&A 561, A3Vidal-Madjar, A., D´esert, J.-M., Lecavelier des Etangs, A., et al.:2004, ApJ 604, 69Vidal-Madjar, A., Lecavelier des Etangs, A., D´esert, J.-M.,Ballester, G.E., Ferlet, R., H´ebrard, G. & Mayor, M.: 2003,Natur 422, 143Yelle, R.V.: 2004, Icar 170, 167Zahn, J.P.: 1977, A&A 57, 383 c (cid:13)(cid:13)
Agol, E., Steffen, J., Sari, R. & Clarkson, W.: 2005, MNRAS 359,567Batalha, N.M., Rowe, J.F., Bryson, S.T., et al.: 2013, ApJS 204, 24Bodenheimer, P., Lin, D.N.C. & Mardling, R.A.: 2001, ApJ 548,466Borucki, W.J., Koch, D.G., Basri, G., et al.: 2011, ApJ 736, 19Borucki, W.J., Koch, D.G., Batalha, N., et al.: 2012, ApJ 745, 120Bourrier, V., Lecavelier des Etangs, A., Dupuy, H., et al.: 2013,A&A 551, 63Brogi, M., Keller, C.U., de Juan Ovelar, M., Kenworthy, M.A., deKok, R.J., Min, M. & Snellen, I.A.G.: 2012, A&A 545, 5Budaj, J.: 2013, A&A 557, 72Burrows, A., Budaj, J. & Hubeny, I.: 2008, ApJ 678, 1436Burrows, A. & Lunine, J.: 1995, Natur 378, 333Col´on, K.D., Ford, E.B. & Morehead, R.C.: 2012, MNRAS 426,342Croll, B., Rappaport, S., DeVore, J., et al.: 2014, ApJ 786, 100Esteves, L.J. De Mooij, E.J.W. & Jayawardhana, R.: 2013, ApJ772, 51Faigler, S. & Mazeh, T.: 2011, MNRAS 415, 3921Ford, E.B., Fabrycky, D.C., Steffen, J.H., et al.: 2012, ApJ 750,113Ford, E.B., Rowe, J.F., Fabrycky, D.C., et al.: 2011, ApJ 197, 2Foreman-Mackey, D., Hogg, D.W., Lang, D., & Goodman, J.:2013, PASP 125, 306Fortney, J.J., Lodders, K., Marley, M.S. & Freedman, R.S.: 2008,ApJ 678, 1419Fressin, F., Torres, G., Rowe, J.F., et al.: 2012, Natur 482, 195Guillot, T., Burrows, A., Hubbard, W.B., Lunine, J.I. & Saumon,D.: 1996, ApJ 459, 35Holman, M.J. & Murray, N.W.: 2005, Sci 307, 1288Hubbard, W.B., Hattori, M.F., Burrows, A., Hubeny, I. & Su-darsky, D.: 2007, Icar 187, 358Hubeny, I., Burrows, A. & Sudarsky, D.: 2003, ApJ 594, 1011Jackson, B., Stark, C.C., Adams, E.R., Chambers, J. & Deming,D.: 2014, AAS 223, 202Kawahara, H., Hirano, T., Kurosaki, K., Ito, Y. & Ikoma, M.: 2013,ApJ 776, L6Kipping, D.M.: 2013, MNRAS 435, 2152Knutson, H.A., Charbonneau, D., Allen, L.E., Burrows, A. &Megeath, S.T.: 2008, ApJ 673, 526Kulow, J.R., France, K., Linsky, J. & Loyd, R.O.P.: 2014, ApJ 786,132Kurokawa, H. & Kaltenegger, L.: 2013, MNRAS 433, 3239Lecavelier des Etangs, A., Ehrenreich, D., Vidal-Madjar, A. et al.:2010, A&A 514, 72Loeb, A. & Gaudi, B.S.: 2003, ApJ 588, 117Lopez, E.D., Fortney, J.J. & Miller, N.: 2012, ApJ 761, 59Luddington, E.W.: 1978, BAAS 10, 418Maciejewski, G., Dimitrov, D., Seeliger, M., et al.: 2013, A&A551, 108Mandel, K. & Agol, E.: 2002, ApJ 580, 171Mazeh, T., Nachmani, G., Holczer, T., et al.: 2013, ApJS 208, 16Ofir, A. & Dreizler, S.: 2013, A&A 555, 580Owen, J.E. & Wu, Y.: 2013, ApJ 775, 1050Rappaport, S., Barclay, T., DeVore, J., Rowe, J., Sanchis-Ojeda,R., Still, M.: 2014, ApJ 784, 40Rappaport, S., Levine, A., Chiang, E., et al.: 2012, ApJ 752, 1Rybicki, G.B. & Lightman, A.P.: 1979, Radiative Processes in As-trophysics, Wiley, New YorkPerez-Becker, D. & Chiang, E.: 2013, MNRAS 433, 2294 Sanchis-Ojeda, R., Rappaport, S., Winn, J.N., Kotson, M.C.,Levine, A.M., El Mellah, I.: 2014, ApJ 787, 47Slawson, R.W., Pr´sa, A., Welsh, W.F., et al.: 2011, AJ 142, 160Steffen, J.H., Batalha, N.M., Borucki, W.J., et al.: 2010, ApJ 725,1226Steffen, J.H., Ford, E.B., Rowe, J.F., et al.: 2012, ApJ 756, 186Stellingwerf, R.F.: 1978, ApJ 224, 953Stellingwerf, R.F.: 2004, PDM2, Stellingwerf Consulting,Huntsville 2004Tassoul, J.L. & Tassoul, M.: 1992, ApJ 395, 259Tian, F., Toon, O.B., Pavlov, A.A. & De Sterck, H.: 2005, ApJ 621,1049van Werkhoven, T.I.M., Brogi, M., Snellen, I.A.G. & Keller, C.U.:2014, A&A 561, A3Vidal-Madjar, A., D´esert, J.-M., Lecavelier des Etangs, A., et al.:2004, ApJ 604, 69Vidal-Madjar, A., Lecavelier des Etangs, A., D´esert, J.-M.,Ballester, G.E., Ferlet, R., H´ebrard, G. & Mayor, M.: 2003,Natur 422, 143Yelle, R.V.: 2004, Icar 170, 167Zahn, J.P.: 1977, A&A 57, 383 c (cid:13)(cid:13)
02 Z. Garai et al.: Search for a circum-planetary material and orbital period variations. . .
A An overview of light curves, transit residuals, and long-term changes of the orbital periods T he t a_ m i n Beta [day/Myr]
Fig. A1
The light curve (left panel), transit residuals (mid panel), and the long-term change of orbital period (right panel)of the exoplanet candidate KIC3848972. T he t a_ m i n Beta [day/Myr]
Fig. A2
The light curve (left panel), transit residuals (mid panel), and the long-term change of orbital period (right panel)of the exoplanet candidate KIC8561063. T he t a_ m i n Beta [day/Myr]
Fig. A3
The light curve (left panel), transit residuals (mid panel), and the long-term change of orbital period (right panel)of the exoplanet candidate KIC6666233. c (cid:13) stron. Nachr. / AN (2006) 803 T he t a_ m i n Beta [day/Myr]
Fig. A4
The light curve (left panel), transit residuals (mid panel), and the long-term change of orbital period (right panel)of the exoplanet candidate KIC6047498. T he t a_ m i n Beta [day/Myr]
Fig. A5
The light curve (left panel), transit residuals (mid panel), and the long-term change of orbital period (right panel)of the exoplanet candidate KIC9030447. T he t a_ m i n Beta [day/Myr]
Fig. A6
The light curve (left panel), transit residuals (mid panel), and the long-term change of orbital period (right panel)of the exoplanet candidate KIC6934291. T he t a_ m i n Beta [day/Myr]
Fig. A7
The light curve (left panel), transit residuals (mid panel), and the long-term change of orbital period (right panel)of the exoplanet candidate KIC4055304. c (cid:13)(cid:13)
The light curve (left panel), transit residuals (mid panel), and the long-term change of orbital period (right panel)of the exoplanet candidate KIC4055304. c (cid:13)(cid:13)
04 Z. Garai et al.: Search for a circum-planetary material and orbital period variations. . . T he t a_ m i n Beta [day/Myr]
Fig. A8
The light curve (left panel), transit residuals (mid panel), and the long-term change of orbital period (right panel)of the exoplanet candidate KIC10024051. T he t a_ m i n Beta [day/Myr]
Fig. A9
The light curve (left panel), transit residuals (mid panel), and the long-term change of orbital period (right panel)of the exoplanet candidate KIC8235924. T he t a_ m i n Beta [day/Myr]
Fig. A10
The light curve (left panel), transit residuals (mid panel), and the long-term change of orbital period (rightpanel) of the exoplanet candidate KIC11774303. T he t a_ m i n Beta [day/Myr]
Fig. A11
The light curve (left panel), transit residuals (mid panel), and the long-term change of orbital period (rightpanel) of the exoplanet candidate KIC10975146. c (cid:13) stron. Nachr. / AN (2006) 805 T he t a_ m i n Beta [day/Myr]
Fig. A12
The light curve (left panel), transit residuals (mid panel), and the long-term change of orbital period (rightpanel) of the exoplanet candidate KIC10028535. T he t a_ m i n Beta [day/Myr]
Fig. A13
The light curve (left panel), transit residuals (mid panel), and the long-term change of orbital period (rightpanel) of the exoplanet candidate KIC10468885. T he t a_ m i n Beta [day/Myr]
Fig. A14
The light curve (left panel), transit residuals (mid panel), and the long-term change of orbital period (rightpanel) of the exoplanet candidate KIC11600889. T he t a_ m i n Beta [day/Myr]
Fig. A15
The light curve (left panel), transit residuals (mid panel), and the long-term change of orbital period (rightpanel) of the exoplanet candidate KIC8278371. c (cid:13)(cid:13)
The light curve (left panel), transit residuals (mid panel), and the long-term change of orbital period (rightpanel) of the exoplanet candidate KIC8278371. c (cid:13)(cid:13)
06 Z. Garai et al.: Search for a circum-planetary material and orbital period variations. . . T he t a_ m i n Beta [day/Myr]
Fig. A16
The light curve (left panel), transit residuals (mid panel), and the long-term change of orbital period (rightpanel) of the exoplanet candidate KIC5513012. T he t a_ m i n Beta [day/Myr]
Fig. A17
The light curve (left panel), transit residuals (mid panel), and the long-term change of orbital period (rightpanel) of the exoplanet candidate KIC10319385. T he t a_ m i n Beta [day/Myr]
Fig. A18
The light curve (left panel), transit residuals (mid panel), and the long-term change of orbital period (rightpanel) of the exoplanet candidate KIC9761199. T he t a_ m i n Beta [day/Myr]
Fig. A19
The light curve (left panel), transit residuals (mid panel), and the long-term change of orbital period (rightpanel) of the exoplanet candidate KIC9473078. c (cid:13) stron. Nachr. / AN (2006) 807 T he t a_ m i n Beta [day/Myr]
Fig. A20
The light curve (left panel), transit residuals (mid panel), and the long-term change of orbital period (rightpanel) of the exoplanet candidate KIC5972334. c (cid:13)(cid:13)