The peculiar dipping events in the disk-bearing young-stellar object EPIC 204278916
S. Scaringi, C.F. Manara, S.A. Barenfeld, P.J. Groot, A. Isella, M.A. Kenworthy, C. Knigge, T. J. Maccarone, L. Ricci, M. Ansdell
MMNRAS , 1–9 (2016) Preprint 29 August 2016 Compiled using MNRAS L A TEX style file v3.0
The peculiar dipping events in the disk-bearingyoung-stellar ob ject EPIC 204278916
S. Scaringi (cid:63) , C. F. Manara , S. A. Barenfeld , P. J. Groot , A. Isella ,M. A. Kenworthy , C. Knigge , T. J. Maccarone , L. Ricci , M. Ansdell Max-Planck-Institute f¨ur Extraterrestriche Physik, D-85748 Garching, Germany Scientific Support Office, Directorate of Science, European Space Research and Technology Centre (ESA/ESTEC), Keplerlaan 1,2201 AZ Noordwijk, The Netherlands California Institute of Technology, Department of Astronomy, MC 249-17, Pasadena, CA 91125 Department of Astrophysics/IMAPP, Radboud University Nijmegen, P.O. Box 9010, 6500 GL Nijmegen, The Netherlands Department of Physics and Astronomy, Rice University, 6100 Main St. Houston, TX 77005, USA Leiden Observatory, Leiden University, PO Box 9513, NL-2300 RA Leiden, the Netherlands School of Physics and Astronomy, University of Southampton, Hampshire SO17 1BJ, United Kingdom Department of Physics and Astronomy, Texas Tech University, Box 41051, Lubbock, TX 79409-1051, USA Harvard-Smithsonian Center for Astrophysics, 60 Garden Street, Cambridge, MA 02138, USA Institute for Astronomy, University of Hawai‘i at M¯anoa, Honolulu, HI, USA
Accepted XXX. Received YYY; in original form ZZZ
ABSTRACT
EPIC 204278916 has been serendipitously discovered from its K2 light curve whichdisplays irregular dimmings of up to 65% for ≈
25 consecutive days out of 78.8 days ofobservations. For the remaining duration of the observations, the variability is highlyperiodic and attributed to stellar rotation. The star is a young, low-mass (M-type)pre-main-sequence star with clear evidence of a resolved tilted disk from ALMA ob-servations. We examine the K2 light curve in detail and hypothesise that the irregulardimmings are caused by either a warped inner-disk edge or transiting cometary-likeobjects in either circular or eccentric orbits. The explanations discussed here are par-ticularly relevant for other recently discovered young objects with similar absorptiondips. Key words: stars: individual (EPIC 204278916), stars: peculiar, comets: general,planets and satellites: dynamical evolution and stability, stars: early-type
Studies of light curves of stars are a proxy for several physicalprocesses, both at the stellar surface or in their surround-ings. Periodic variability of the observed stellar flux has longbeen used to measure stellar rotation periods (e.g. Stassunet al. 1999). Variability in young stellar objects (YSOs) isalso related to the presence of a protoplanetary disk. Ma-terial in the disk can obscure the central star (e.g. Herbstet al. 1994; Alencar et al. 2010), and the highly variable ac-cretion onto the central star modifies the emission from thesystem (e.g. Bertout et al. 1988; Bouvier et al. 2007). Thevariability of YSOs is known to have different degrees of pe-riodicity and flux symmetry which are possibly related todifferent processes (Cody et al. 2014). Recently,
CoRoT and
Kepler / K2 observations of main sequence stars and YSOshave shown deep and short dips in light curves that can (cid:63) E-mail: [email protected] be explained by the presence of a family of comets orbitingaround the stars (Boyajian et al. 2016; Bodman & Quillen2015) or more generally transiting circumstellar material(Ansdell et al. 2016b), or through occulting material at theinner edges of a circumstellar disk (McGinnis et al. 2015;Ansdell et al. 2016b). The latter is possible if the inner diskis observed almost edge-on, and offers a unique opportunityto study the properties of dust and gas in the inner regionof protoplanetary disks. If the dimming events are causedby transiting circumstellar clumps originating in the disk,these can be used to constrain the size of planetesimals inthe disk, which is a necessary step for planet formation butdifficult to observationally detect Testi et al. (2014).More recently, Ansdell et al. (2016a) have found 3 YSOdippers with a wide range of inclination angles, demonstrat-ing how edge-on disks are not a defining characteristic ofdipping YSOs. At this stage it is important to study a vari-ety of dipper YSOs in detail and explore possible scenarios © a r X i v : . [ a s t r o - ph . S R ] A ug S. Scaringi et al. to explain their behaviour and their relevance for planet for-mation studies.We have serendipitously discovered a YSO dipper starobserved with K2 , giving us the possibility to further in-vestigate these peculiar systems. The target discussed here(EPIC 204278916, 2MASS J16020757 − ∼ ∼
11 Myr; Pecaut et al. 2012). Our target is a single star(Kraus & Hillenbrand 2007) of spectral type M1 and has alogarithmic bolometric stellar luminosity log L (cid:63) /L (cid:12) = 0.15(Preibisch et al. 2002). The inferred radius of this star is R (cid:63) = 0.97 R (cid:12) , while the stellar mass is M (cid:63) ∼ M (cid:12) , depend-ing on the evolutionary models used (Baraffe et al. 1998;Siess et al. 2000).At this age, the majority of young stars have alreadydispersed their circumstellar disk and are not accreting ma-terial from the disk anymore (e.g., Fedele et al. 2010). Spec-troscopic observations of the H α line of this target report avery small equivalent width (EW = -3.2 ˚A, Preibisch et al.2002), consistent with very little or no accretion. However,this target shows an infrared excess due to the presence ofa protoplanetary disk in WISE data (Luhman & Mamajek2012).The next section will introduce the K2 and ALMA dataand analysis procedure to obtain the reduced light curve(with corresponding Fourier transform) and the ALMA in-tensity map. Section 3 discusses the observations in the con-text of various interpretations, including a protostellar diskorigin and cometary-like transits in either circular or eccen-tric orbits. We discuss our results in section 4 and give ourconclusions in Section 5 with future prospects for determin-ing the real nature of EPIC 204278916. K2 light curve
EPIC 204278916 was observed by the K2 mission (Boruckiet al. 2010) during Campaign 2 between August 23 andNovember 13 2014 (78.8 days) and has a registered Kepler magnitude ( Kp ) of 13 . K2 Ecliptic Plane Input Cat-alog (EPIC). Here we analyse long cadence (LC, 29.4 min-utes) data obtained from the Mikulski Archive for SpaceTelescope (MAST) archive . The data is provided in rawformat, consisting of target pixel data. For each 29.4 minuteexposure we thus have a 12 ×
11 pixel image centred on thetarget.As no other sources were present within the 12 × http://archive.stsci.edu/k2/ due to cosmic rays. We produce the light curve by summingtogether all target pixels for each exposure, and subtractthe average background obtained from the background pixelmask. The obtained light curve is shown in Fig. 1. The sameprocedure has been used to extract other K2 light curves ofisolated sources (Scaringi et al. 2015b,a).To locate any significant periodicities we carry out aFourier transform of the light curve excluding the first 30days (in order to not be affected by the initial dipping pe-riod, see Fig. 1). Fig. 2 shows our result, revealing two dis-tinct periodicities of 3.646 days ( F = 0 . f = 4 .
080 cycles/day), together with corre-sponding harmonic frequencies. We associate the 3.646-daysignal to stellar rotation, as this is a typical rotational pe-riod for young stars stars observed by
Kepler (Ansdell et al.2016b; Vasconcelos & Bouvier 2015), and show the phase-folded profile in Fig. 3. The 0.245-day signal is associatedwith the spacecraft thruster firing to re-adjust attitude ev-ery ∼ min = 2456955 . . × N, (1)where N is the cycle number and the ephemeris are given inBarycentric Julian Date (BJD) at minimum light.We then expand and replicate the phase folded lightcurve shown in Fig. 3 to match the full light curve and in-terpolate it on the same temporal grid. Fig. 4 shows thefull light curve with the 3.646-day periodicity removed. Theperiodicity removal is not perfect (probably due to smallchanges in the pulse profile over time) but shows a signifi-cant improvement over the original light curve. We use thiscurve in subsequent analysis. ALMA Cycle 2 observations of EPIC 204278916 were ob-tained on 2014 June 30 and 2014 July 7 (UT) in four spectralwindows centred at 334.2, 336.1, 346.2, and 348.1 GHz fora mean frequency of 341.1 GHz (0.88 mm). The bandwidthof each window is 1.875 GHz. Thirty six antennas were in-cluded in the array, with baselines ranging from 16-650 mfor an angular resolution of 0.34 arcseconds. The ALMAdata were calibrated using the CASA package. The reduc-tion scripts were kindly provided by NRAO. These includeatmospheric calibration using the 183 GHz water vapour ra-diometers, bandpass calibration, flux calibration, and gaincalibration. For further details on the observations and datareduction, see Barenfeld et al. (2016). The protoplanetarydisk in EPIC 204278916 is resolved with ALMA at highsignal-to-noise, both in the continuum (see Fig. 5) and inthe CO(J=3 −
2) line (Barenfeld et al. 2016). An ellipticalGaussian fit of the surface brightness profile to the contin-uum visibility data results in an axis ratio of 0.55 ± ± MNRAS , 1–9 (2016) eculiar dips in EPIC 204278916 Figure 1.
EPIC 204278916 light curve. The system was observed for over 78.8 days at 29.4 minute cadence. The units on the y-axis areelectrons/second, and can be converted to
Kepler magnitudes Kp using the conversion found in the Kepler
Instrument Handbook. Thetop panel shows the full light curve, whilst the bottom panel zooms into the first 25 days of observation where the dipping events areobserved.
Figure 2.
Fourier transform of LC K2 data for EPIC 204278916 starting from day 25. A peak at F = 0 . F ) is visible and at f = 4 .
080 cycles/day (with corresponding harmonics f , f , f ) and f . We associate F to the possible stellar rotation, whilst f can be attributed to spacecraft attitude adjustments. In this section we will go through possible scenarios to ex-plain the observed large amplitude dipping events in EPIC204278916.
Provided with a favourable disk inclination, the observeddips in the EPIC 204278916 light curve could be caused bynon-axisymmetric structures in the inner disk edge occult-ing the star. Because of the large observed dips ( ≈ i max (cid:46) R ∗ R (cid:12) P rot days − / M ∗ M (cid:12) − / . (2)Adopting P rot = 3 .
646 days, R ∗ = 0 . R (cid:12) and M ∗ = 0 . M (cid:12) we obtain a maximum inclination i max = 14 degrees. Thisis somewhat in contrast with the 3 σ level for the disk incli-nation of i = 57 degrees obtained from the ALMA image.Other YSOs have been observed to display non-periodicdipping events (see e.g. Sousa et al. 2016; Ansdell et al.2016b,a), and it is possible EPIC 204278916 is also part ofthis “dipper” class of YSOs. In particular, EPIC 204278916supports the idea that nearly edge-on viewing geometriesare not a defining characteristic of dippers. This in turn mo-tivates the exploration of alternative models. More specifi- MNRAS000
646 days, R ∗ = 0 . R (cid:12) and M ∗ = 0 . M (cid:12) we obtain a maximum inclination i max = 14 degrees. Thisis somewhat in contrast with the 3 σ level for the disk incli-nation of i = 57 degrees obtained from the ALMA image.Other YSOs have been observed to display non-periodicdipping events (see e.g. Sousa et al. 2016; Ansdell et al.2016b,a), and it is possible EPIC 204278916 is also part ofthis “dipper” class of YSOs. In particular, EPIC 204278916supports the idea that nearly edge-on viewing geometriesare not a defining characteristic of dippers. This in turn mo-tivates the exploration of alternative models. More specifi- MNRAS000 , 1–9 (2016)
S. Scaringi et al.
Figure 3.
EPIC 204278916 light curve folded on the stellar ro-tation period of 3.646 days. cally, the clustering of large-amplitude dipping events suchas those observed in the K2 EPIC 204278916 light curvehave not been previously observed in other systems. Theonly other YSO dipper which appears to resemble EPIC204278916 as observed with K2 is EPIC 204530045 (seeFig. 1 of Bodman et al. 2016). However, on close inspec-tion, it appears that the dipping behaviour between EPIC204278916 and EPIC 204530045 differs in that a) the dips inEPIC 204278916 always appear to return to the stellar fluxwhilst there is a clear downward general trend for EPIC204530045, b) whilst EPIC 204278916 solely displays dip-ping, the lightcurve of EPIC 204530045 more resembles flick-ering and c) the dip depth is nearly a factor 3 larger for EPIC204278916 when compared to EPIC 204530045. All three ar-guments seem to suggest that whilst the K2 lightcurves ofboth EPIC 204278916 and EPIC 204530045 appear similar,the phenomenology observed in EPIC 204530045 seems to beaccretion driven (change in instantaneous mass transfer rateon the stellar surface), whilst EPIC 204278916 has all thecharacteristics of being “transited”. Thus, if EPIC 204278916is also part of the YSO “dipper” class as those discussed inBodman et al. (2016), then it constitutes a somewhat specialcase. The rapid fluctuations over the first 25 days of observationsare similar to those seen towards J1407, another young starin the Sco-Cen association. In May 2007 a series of rapidfluctuations was seen over a 56 day period (Mamajek et al.2012) and interpreted as a giant ring system filling the Hillsphere of the unseen secondary companion (Kenworthy et al.2015; Kenworthy & Mamajek 2015). The series of dips afterthe first initial deep dip show hints of being symmetric intime, around T = 10 days in Fig. 6 (bottom panel), consis-tent with a ring system that is inclined with respect to ourline of sight. The gradient of the light curve of such an in-clined ring system will show small gradients near the pointof closest projected approach as the star moves parallel toring edges, and steeper light curve gradients at other times.This is not seen in the gradient of the light curve as a func-tion of time (Fig. 6, top-panel), and so this is considered anunlikely hypothesis for these dips. It is possible that the dips in some YSOs might be causedby transiting circumstellar objects, similar to what has beenproposed for KIC 8462852 (Boyajian et al. 2016; Bodman &Quillen 2015; Marengo et al. 2015). The discussion presentedbelow is tailored to the dipping YSO EPIC 204278916, butcan be extended to other dipping YSOs such as those dis-cussed in Ansdell et al. (2016a).
With the assumption that the transiting object(s) are incircular orbits we can place tight constraints on the orbitalparameters in a plane defined by the semi-major axis ( a )and the transiting clump radius ( R c ). These are included inFig. 7, and described in detail below. Dip depth:
Constraints on the clump size can be ob-tained from the dip depth τ , defined as 1 minus the nor-malised absorbed flux during the dipping event (see Fig.4).In the most extreme scenario where the eclipsing clump iscompletely opaque, max( τ ) = ( R c /R ∗ ) . The deepest eventobserved in the EPIC 204278916 light curve corresponds to τ ≈ Dip duration:
A transiting object will have a transversevelocity along the stellar equator of v t = 2( R c + R ∗ ) t dip , (3)where t dip is the transit duration. If the object is on a circu-lar orbit around a star of mass M ∗ with semi-major axis a ,then we can estimate the size of the transiting object with R c = t dip (cid:18) GM ∗ a (cid:19) / − R ∗ . (4)Thus, the observed dip durations in the EPIC 204278916light curve can provide an estimate for the transiting clumpsize for circular orbits. The longest dip duration of t dip = 1day provides the most stringent constraint, and is shown asa solid black line in Fig. 7. Light-curve gradient:
An outer constraint on the semi-major axis can be derived using the largest gradient ob-served during a transit event. Transiting material will changethe light curve most rapidly when it is optically thick andtransiting through the stellar equator. van Werkhoven et al.(2014) provides the equation required to translate the ob-served gradients into a minimum velocity ( v min ) for tran-siting material using the so-called “knife edge” model. As-suming then that the material is in a circular orbit and op-tically thick the obtained minimum velocity constraint of39km/s translates to a maximum semi-major axis of 0.2916AU through a max = GM ∗ v min . (5) Non-periodicity:
Given that we do not observe a repe-tition of the dipping events within the total light curve, wecan place a constraint on the orbital period to be longer than
MNRAS , 1–9 (2016) eculiar dips in EPIC 204278916 Figure 4.
Normalised EPIC 204278916 light curve with the 3.646-day periodicity removed. ∆ α (arcsec) ∆ δ ( a r c s e c ) Figure 5.
ALMA continuum intensity map for EPIC 204278916.Isocontours for 3-,10-,17- and 23- σ are overlaid on the image. TheALMA beam size is shown in the bottom-left corner. K2 observations. Furthermore it is possiblethat the dipping events began before the start of the K2 ob-servations, in which case circular orbits would be fully ruledout. It is also important to note that the resulting size forthe transiting clump would be very large at R c ≈ . R sol . ALMA observations have revealed the inclination of the cir-cumstellar disk to be 57 ± e , the orbital ve- locity at pericentre and apocentre respectively are definedas v per = (cid:115) GM ∗ ea (1 − e ) (6)and v apo = (cid:115) GM ∗ − ea (1 + e ) . (7)Fig. 8 shows these velocities as a function of eccentricityadopting the derived minimum semi-major axis of a min =0 . v min = 39km/sobtained from the light curve gradient of the dips. This al-ready shows that if the orbit is eccentric, then we are mostlikely observing the transit away from apocentre.We can further expand on this analysis and speculatewhat the eccentricity would be if the transiting material hasbroken into smaller clumps due to a close passage to itsparent star at pericentre. The pericentre and apocentre radiirespectively are given by r per = a (1 − e ) (8)and r apo = a (1 + e ) . (9)Fig. 9 shows these radii as a function of eccentricity adoptingthe same minimum semi-major axis as used in Fig. 8. Theobtained radii in this case can be thought of as lower limits,since increasing the semi-major axis will result in larger peri-centre and apocentre radii. The dashed line in Fig. 9 showsthe Roche radius below which a cometary-like object with adensity of 0.5 g/cm would break-up. The eccentricity wouldhave to be very large ( e > .
95) for this case, and might re-semble the Comet Shoemaker-Levy 9 event that broke apartand collided with Jupiter in July 1994 (see e.g. Chodas &Yeomans 1996).
We can place a lower constraint on the mass of the transitingmaterial (independent of eccentricity) by assuming that the
MNRAS000
MNRAS000 , 1–9 (2016)
S. Scaringi et al.
Figure 6.
Bottom-panel: Normalised EPIC 204278916 light curve (with the 3.646-day periodicity removed) for the first 25 days ofobservation. Top-panel: Corresponding lightcurve gradient.
Figure 7.
Assuming 1-day dips, the figure shows the possiblesemi-major axis vs. clump radius values with the solid black line.The minimum clump size, obtained through the observed dip am-plitude, is shown with the dashed line. The two vertical dottedlines show the minimum and maximum semi-major axis obtainedwith the non-periodicity and observed minimum velocity (for op-tically thick material) constraint respectively. observed individual dips are caused by a cluster of objects,each of a different mass and size. The mass of each clump isdefined by M = m g ρV, (10)where m g is the average grain mass, ρ the object densityand V its volume. In turn the volume can be estimated by V = πR c ∆ R, (11)where ∆ R is the clump depth along the line of sight. We can Figure 8.
Pericentre and apocentre orbital velocities for varyingeccentricities with a constant semi-major axis of a = 0 . obtain an estimate on R c from the observed dip depths, butit is non-trivial to obtain ∆ R . However, from the definitionof optical depth ( τ ) we find that τ = ρσ ∆ R, (12)where σ is the cross-section of the particles causing the ob-scuration. Through some algebraic manipulation we can sub-stitute Eq. 12 and Eq. 11 into Eq. 10 to obtain M = m g πR c τσ , (13) MNRAS , 1–9 (2016) eculiar dips in EPIC 204278916 Figure 9.
Pericentre and apocentre radii for varying eccentricitiesadopting the observed minimum semi-major axis of a = 0 . ρ = 0 . would break-up. The dotted line marks the stellar radius, below which orbitingclumps would crash into the star. For this scenario the eccentricitycan be constrained to be above 0.95. and note that both the density and clump depth cancel out.Eq. 13 is useful in obtaining a lower mass limit on the transit-ing material, since we know the minimum clump radii fromthe dip depths and also that τ ≥
1. The only unconstrainedparameters are the cross-section σ and the grain mass m g ,which we fix to 10 − cm and 10 − grams respectivley, typ-ical for dust grain sizes of ≈ . µ m. These values should beregarded as lower limits from derived dust grain mass dis-tributions (e.g. Li & Greenberg 1998).To obtain the number and radius of the transitingclumps, we perform a multi-gaussian fit to the EPIC204278916 light curve (after removing the 3.646-day period-icity) shown in Fig. 4. Using a local peak finding algorithm( findpeak implemented in MATLAB ) we identify 53 dips.Fig. 10 shows the first 25 days of observations fitted with 53gaussians of differing widths and amplitudes and one broadPoisson function covering the timespan of the smaller dips.We find that the Poisson component is necessary to obtaina good fit to the data, but that the goodness-of-fit is notsensitive to the exact number of gaussians used. For eachgaussian component we obtain the related R c and estimatethe mass of each individual component using Eq. 13. Thesum of all gaussians then yields a reasonable lower masslimit for the whole clump of M c = 7 × grams ( ≈ . The large-amplitude dipping events observed in the K2 lightcurve of EPIC 204278916 superficially resembles othersystems recently discovered with K2 . Most notably, KIC8462852 (Boyajian et al. 2016) was observed to have similardip durations to those observed in EPIC 204278916. How-ever, aside from this, KIC 8462852 and EPIC 204278916 dif-fer in many other respects. Firstly, EPIC 204278916 shows much deeper dips than KIC 8462852. Furthermore, the dip-ping patterns observed in EPIC 204278916 are clustered intime, with an initial deep dipping event being followed bysmaller ones for ≈
25 days before returning to the normal,presumably quiescent state of the star. The dips observedin KIC 8462852 are not clustered, but are spread out overa period of several years. More importantly however, KIC8462852 displays an ordinary F-type star spectrum (Boy-ajian et al. 2016) showing no accretion signatures, whilstEPIC 204278916 is a YSO still surrounded by a disk, basedon both the ALMA image shown in this work (Fig. 5), itsspectral features (e.g., presence of lithium absorption lineand H α emission line, Preibisch et al. 2002), and its mem-bership in the Scorpius-Centaurus OB association (Preibisch& Mamajek 2008).More recently Ansdell et al. (2016a) have discovered 3objects which more resemble EPIC 204278916 when com-pared to KIC 8462852. These also belong to the Upper-Scorpius association and are YSOs. Similar to what has beenpresented here for EPIC 204278916, Ansdell et al. (2016a)have resolved the disks with ALMA, and demonstrated thatlarge-amplitude dipping events are not only observed inedge-on systems. Given that both the objects presented byAnsdell et al. (2016a) and EPIC 204278916 belong to thesame star-forming association, and all show large dippingevents, one might consider them to be part of the same classof dipping systems. However, the peculiarly clustered dip-ping structure displayed by EPIC 204278916 distinguishesit from previously reported YSO dipping systems. The onlyexception might be EPIC 204530046 (presented in Bodmanet al. 2016), although we discussed in Section 3.1, why wethink EPIC 204278916 is qualitatively different.Given the large range of inclination angles inferred fromthe resolved ALMA images of EPIC 204278916 and the sam-ple of Ansdell et al. (2016a), it is unlikely that the observeddips are related to the outer-edges of the proto-stellar disk.It is however possible that an inclined and variable innerdust disk could cause some of the observed dips (see e.g.HD 142527, Marino et al. 2015). This is particularly relevantfor some of the systems discussed in Ansdell et al. (2016a)and Bodman et al. (2016), where the dipping events are ob-served to persist throughout the full ≈ K2 ob-servations and some display quasi-periodic dips that repeaton the period of the stellar co-rotation radius. If a similarmechanism were responsible for the observed dips in EPIC204278916, the inclined inner disk would have to be transienton a relatively short timescale of a few weeks, making thisinterpretation also unlikely. Furthermore we find no relationbetween the repeating dip patterns in EPIC 204278916 withthe stellar rotation. Thus both the transient nature of theobserved dips and the fact that they do not repeat (quasi-)periodically makes EPIC 204278916 stand out even morefrom previously observed YSO dippers.In Section 3.3 we explored the possibility that the ob-served dips in EPIC 204278916 are caused by transiting cir-cumstellar clumps. We showed how circular orbits for this in-terpretation are most likely ruled out and that highly eccen-tric orbits are consistent with the observations. If transitingcometary-like bodies are responsible for the observed dips,the events are most likely occurring close to periastron pas-sage. If the dips are the result of a previous disruptive eventof a larger body, we can further say that the eccentricity MNRAS000
25 days before returning to the normal,presumably quiescent state of the star. The dips observedin KIC 8462852 are not clustered, but are spread out overa period of several years. More importantly however, KIC8462852 displays an ordinary F-type star spectrum (Boy-ajian et al. 2016) showing no accretion signatures, whilstEPIC 204278916 is a YSO still surrounded by a disk, basedon both the ALMA image shown in this work (Fig. 5), itsspectral features (e.g., presence of lithium absorption lineand H α emission line, Preibisch et al. 2002), and its mem-bership in the Scorpius-Centaurus OB association (Preibisch& Mamajek 2008).More recently Ansdell et al. (2016a) have discovered 3objects which more resemble EPIC 204278916 when com-pared to KIC 8462852. These also belong to the Upper-Scorpius association and are YSOs. Similar to what has beenpresented here for EPIC 204278916, Ansdell et al. (2016a)have resolved the disks with ALMA, and demonstrated thatlarge-amplitude dipping events are not only observed inedge-on systems. Given that both the objects presented byAnsdell et al. (2016a) and EPIC 204278916 belong to thesame star-forming association, and all show large dippingevents, one might consider them to be part of the same classof dipping systems. However, the peculiarly clustered dip-ping structure displayed by EPIC 204278916 distinguishesit from previously reported YSO dipping systems. The onlyexception might be EPIC 204530046 (presented in Bodmanet al. 2016), although we discussed in Section 3.1, why wethink EPIC 204278916 is qualitatively different.Given the large range of inclination angles inferred fromthe resolved ALMA images of EPIC 204278916 and the sam-ple of Ansdell et al. (2016a), it is unlikely that the observeddips are related to the outer-edges of the proto-stellar disk.It is however possible that an inclined and variable innerdust disk could cause some of the observed dips (see e.g.HD 142527, Marino et al. 2015). This is particularly relevantfor some of the systems discussed in Ansdell et al. (2016a)and Bodman et al. (2016), where the dipping events are ob-served to persist throughout the full ≈ K2 ob-servations and some display quasi-periodic dips that repeaton the period of the stellar co-rotation radius. If a similarmechanism were responsible for the observed dips in EPIC204278916, the inclined inner disk would have to be transienton a relatively short timescale of a few weeks, making thisinterpretation also unlikely. Furthermore we find no relationbetween the repeating dip patterns in EPIC 204278916 withthe stellar rotation. Thus both the transient nature of theobserved dips and the fact that they do not repeat (quasi-)periodically makes EPIC 204278916 stand out even morefrom previously observed YSO dippers.In Section 3.3 we explored the possibility that the ob-served dips in EPIC 204278916 are caused by transiting cir-cumstellar clumps. We showed how circular orbits for this in-terpretation are most likely ruled out and that highly eccen-tric orbits are consistent with the observations. If transitingcometary-like bodies are responsible for the observed dips,the events are most likely occurring close to periastron pas-sage. If the dips are the result of a previous disruptive eventof a larger body, we can further say that the eccentricity MNRAS000 , 1–9 (2016)
S. Scaringi et al.
Figure 10.
Normalised EPIC 204278916 light curve (solid line, with the 3.646-day periodicity removed) decomposed into 53 gaussians(dashed lines) and one Poissonian component (dotted line). needs to be larger than 0.95 for previous pericentre passagesto be within the Roche radius of the parent star. Similar dis-ruptive events have been witnessed in our Solar System (e.g.Comet Shoemaker-Levy 9), but have never been previouslywitnessed around other stars. Thus EPIC 204278916 couldconstitute the first system where a planetesimal-sized bodyhas been witnessed to be tidally disrupted by the parent starupon a close encounter. This would be then a direct evidenceof the presence of km-sized bodies in a protoplanetary disk,a crucial step towards planet formation.Given there is no complete explanation for the myste-rious behaviour of EPIC 204278916, more observations andmodelling of this system are required to fully explain theclustered large-amplitude dipping events. Continuous pho-tometric monitoring of this system for subsequent dippingevents will determine whether this behaviour is periodic ornot. Given the dynamics of break-up orbits, we would expectthe cometary-like bodies to only survive a few orbits beforehitting the parent star. It is thus important to monitor EPIC204278916 before such an event occurs.
We have presented the K2 light curve of the disk-bearingyoung-stellar object EPIC 204278916, together with a re-solved ALMA image constraining its disk inclination to 57 ± K2 light curve displays prominent, large-amplitude, dips during the first ≈
25 days of observationsout of the 78.8 day K2 observing campaign. Although diffi-cult to establish their true physical origin, we have discussedthe observed dips in terms of a warped inner disk transitingcircumstellar clumps in circular orbits, and cometary-likedebris in an eccentric orbit.It is clear that further observations of EPIC 204278916and other YSO dippers will be required in the future,both photometric and spectroscopic, in order to establishtheir true origin. In particular it is important to determinewhether the observed dips in the K2 light curve of EPIC204278916 are observed again, in which case infer their re-currence timescale and spectroscopic properties. In this re-spect we point out the possibility of K2 to re-observe part ofthe Scorpius-Centaurus OB association in 2017 during theplanned Campaign 15. ACKNOWLEDGEMENTS
We gratefully thank the anonymous referee for providinguseful and insightful comments which have improved thismanuscript. S.S. acknowledges funding from the Alexandervon Humboldt Foundation. C.F.M acknowledges ESA re-search fellowship funding. This research has made use ofNASA’s Astrophysics Data System Bibliographic Services.Additionally this work acknowledges the use of the astron-omy & astrophysics package for Matlab (Ofek 2014). Thispaper includes data collected by the Kepler mission. Fund-ing for the Kepler mission is provided by the NASA Sci-ence Mission directorate. Some of the data presented inthis paper were obtained from the Mikulski Archive forSpace Telescopes (MAST). STScI is operated by the As-sociation of Universities for Research in Astronomy, Inc.,under NASA contract NAS5-26555. Support for MAST fornon-HST data is provided by the NASA Office of Space Sci-ence via grant NNX13AC07G and by other grants and con-tracts. This paper additionally makes use of the followingALMA data: ADS/JAO.ALMA
REFERENCES
Alencar S. H. P., et al., 2010, A&A, 519, A88Ansdell M., Gaidos E., Williams J. P., Kennedy G., Wyatt M. C.,LaCourse D. M., Jacobs T. L., Mann A. W., 2016a, preprint,( arXiv:1607.03115 )Ansdell M., et al., 2016b, ApJ, 816, 69Baraffe I., Chabrier G., Allard F., Hauschildt P. H., 1998, A&A,337, 403Barenfeld S. A., Carpenter J. M., Ricci L., Isella A., 2016,preprint, ( arXiv:1605.05772 )Bertout C., Basri G., Bouvier J., 1988, ApJ, 330, 350Bodman E. H. L., Quillen A., 2015, preprint, ( arXiv:1511.08821 )Bodman E. H. L., et al., 2016, preprint, ( arXiv:1605.03985 )Borucki W. J., Koch D., Basri G., Batalha N., Brown T., CaldwellD., et al., 2010, Science, 327, 977Bouvier J., et al., 2007, A&A, 463, 1017MNRAS , 1–9 (2016) eculiar dips in EPIC 204278916 Bouy H., Mart´ın E. L., 2009, A&A, 504, 981Boyajian T. S., et al., 2016, MNRAS, 457, 3988Chodas P. W., Yeomans D. K., 1996, in Noll K. S., Weaver H. A.,Feldman P. D., eds, IAU Colloq. 156: The Collision of CometShoemaker-Levy 9 and Jupiter. pp 1–30Cody A. M., et al., 2014, AJ, 147, 82Fedele D., van den Ancker M. E., Henning T., Jayawardhana R.,Oliveira J. M., 2010, A&A, 510, A72Herbst W., Herbst D. K., Grossman E. J., Weinstein D., 1994,AJ, 108, 1906Kenworthy M. A., Mamajek E. E., 2015, ApJ, 800, 126Kenworthy M. A., et al., 2015, MNRAS, 446, 411Kraus A. L., Hillenbrand L. A., 2007, ApJ, 662, 413Li A., Greenberg J. M., 1998, ApJL, 498, L83Luhman K. L., Mamajek E. E., 2012, ApJ, 758, 31Mamajek E. E., Quillen A. C., Pecaut M. J., Moolekamp F., ScottE. L., Kenworthy M. A., Collier Cameron A., Parley N. R.,2012, AJ, 143, 72Marengo M., Hulsebus A., Willis S., 2015, ApJL, 814, L15Marino S., Perez S., Casassus S., 2015, ApJL, 798, L44McGinnis P. T., et al., 2015, A&A, 577, A11Ofek E. O., 2014, MATLAB package for astronomy and astro-physics, Astrophysics Source Code Library (ascl:1407.005)Pecaut M. J., Mamajek E. E., Bubar E. J., 2012, ApJ, 746, 154Preibisch T., Mamajek E., 2008, The Nearest OB Association:Scorpius-Centaurus (Sco OB2). p. 235Preibisch T., Brown A. G. A., Bridges T., Guenther E., ZinneckerH., 2002, AJ, 124, 404Scaringi S., et al., 2015a, Science Advances, 1, e1500686Scaringi S., Maccarone T. J., Hynes R. I., K¨ording E., Ponti G.,Knigge C., Britt C. T., van Winckel H., 2015b, MNRAS, 451,3857Siess L., Dufour E., Forestini M., 2000, A&A, 358, 593Sousa A. P., et al., 2016, A&A, 586, A47Stassun K. G., Mathieu R. D., Mazeh T., Vrba F. J., 1999, AJ,117, 2941Testi L., et al., 2014, Protostars and Planets VI, pp 339–361Vasconcelos M. J., Bouvier J., 2015, A&A, 578, A89van Werkhoven T. I. M., Kenworthy M. A., Mamajek E. E., 2014,MNRAS, 441, 2845This paper has been typeset from a TEX/L A TEX file prepared bythe author.MNRAS000