Recurring Occultations of RW Aurigae by Coagulated Dust in the Tidally Disrupted Circumstellar Disk
Joseph E. Rodriguez, Phillip A. Reed, Robert J. Siverd, Joshua Pepper, Keivan G. Stassun, B. Scott Gaudi, David A. Weintraub, Thomas G. Beatty, Michael B. Lund, Daniel J. Stevens
aa r X i v : . [ a s t r o - ph . S R ] D ec D raft version J une
13, 2018
Preprint typeset using L A TEX style emulateapj v. 5 / / RECURRING OCCULTATIONS OF RW AURIGAE BY COAGULATED DUST IN THE TIDALLY DISRUPTEDCIRCUMSTELLAR DISK J oseph E. R odriguez , P hillip A. R eed , R obert J. S iverd , J oshua P epper , K eivan G. S tassun , , B. S cott G audi , D avid A. W eintraub ,T homas G. B eatty , , M ichael B. L und , and D aniel J. S tevens Department of Physics and Astronomy, Vanderbilt University, 6301 Stevenson Center, Nashville, TN 37235, USA Department of Physical Sciences, Kutztown University, Kutztown, PA 19530, USA Las Cumbres Observatory Global Telescope Network, 6740 Cortona Dr., Suite 102, Santa Barbara, CA 93117, USA Department of Physics, Lehigh University, 16 Memorial Drive East, Bethlehem, PA 18015, USA Department of Physics, Fisk University, 1000 17th Avenue North, Nashville, TN 37208, USA Department of Astronomy, The Ohio State University, Columbus, OH 43210, USA Department of Astronomy & Astrophysics, The Pennsylvania State University, 525 Davey Lab, University Park, PA 16802 and Center for Exoplanets and Habitable Worlds, The Pennsylvania State University, 525 Davey Lab, University Park, PA 16802
Draft version June 13, 2018
ABSTRACTWe present photometric observations of RW Aurigae, a Classical T Tauri system, that reveal two remarkabledimming events. These events are similar to that which we observed in 2010-2011, which was the first suchdeep dimming observed in RW Aur in a century’s worth of photometric monitoring. We suggested the 2010-2011 dimming was the result of an occultation of the star by its tidally disrupted circumstellar disk. In 2012-2013, the RW Aur system dimmed by ∼ ∼
40 days and in 2014 / ∼ >
250 days. The ingress / egress duration measurements of the more recent events agree well withthose from the 2010-2011 event, providing strong evidence that the new dimmings are kinematically associatedwith the same occulting source as the 2010-2011 event. Therefore, we suggest that both the 2012-2013 and2014-2015 dimming events, measured using data from the Kilodegree Extremely Little Telescope and theKutztown University Observatory, are also occultations of RW Aur A by the tidally disrupted circumstellarmaterial. Recent hydrodynamical simulations of the eccentric fly-by of RW Aur B suggest the occulting bodyto be a bridge of material connecting RW Aur A and B. These simulations also suggest the possibility ofadditional occultations which are supported by the observations presented in this work. The color evolutionof the dimmings suggest that the tidally stripped disk material includes dust grains ranging in size from smallgrains at the leading edge, typical of star forming regions, to large grains, ices or pebbles producing grey ornearly grey extinction deeper within the occulting material. It is not known whether this material representsarrested planet building prior to the tidal disruption event, or perhaps accelerated planet building as a result ofthe disruption event, but in any case the evidence suggests the presence of advanced planet building material inthe space between the two stars of the RW Aur system. Subject headings:
Circumstellar Matter, Individual Stars: RW Aur, Protoplanetary Disks, Stars: Pre-mainSequence, Stars: Variables: T Tauri INTRODUCTION
The circumstellar environment of young stars of a few Myrsold (T Tauri stars) involves complex dynamical interactionsbetween dust and gas that directly influences the formationof planets. Past studies have shown that binarity is a com-mon property of T Tauri stars (Ghez et al. 1993, Leinert etal. 1993, Richichi et al. 1994, Simon et al. 1995, Ghezet al. 1997). The process of planetary formation can besignificantly altered when the circumstellar disk is gravita-tionally influenced by a stellar companion (Clarke & Pringle1993; Dai et al. 2015). Specifically, strong binary interac-tions with disks are also likely to influence planetary coreformation and chemical composition by stirring up and heat-ing materials, enhancing planetesimal collisions. The proto-type example of this type of system is RW Aurigae, a binarysystem of two Classical T Tauri Stars (CTTS), RW Aur Aand B (Duchˆene et al. 1999). Detailed millimeter mapping byCabrit et al. (2006) showed evidence of a recent close stellarfly-by of RW Aur B which disrupted the circumstellar mate-rial around RW Aur A, leaving a short truncated circumstellardisk and a large ∼
600 AU long tidal arm extending from RWAur A. The system parameters are comprehensively described in § ∼ ′′ ( ∼
200 AU) (Cabrit et al. 2006). Bisikalo et al.(2012) measured that the separation of RW Aur A and B over ∼
70 years has increased by ∼ ′′ yr − . At the angular sepa-ration and 140 pc distance, the Keplerian orbital period wouldbe > ◦ –60 ◦ (Cabrit et al. 2006).In 2010 the RW Aur system dimmed by ∼ ∼
180 days, marking the first event of thiskind observed in this system dating back to the late 1890’s(Beck & Simon 2001). Rodriguez et al. (2013) (hereafter Pa-per I) interpreted that dimming as an occultation of RW AurA by disrupted circumstellar material from the close fly-byencounter of the two stars RW Aur A and B. Using simplekinematic arguments, Paper I determined that the occultingobject, likely a clump of circumstellar material, was ∼ / s. If in a Keplerian orbit, it TABLE 1L ist of observations
Filter Exposure Time (s) Number of Exposures B
25 587 V
15 602 R
15 631 I
15 604 would be ∼
180 AU from RW Aur A . Using simple geometricand kinematic arguments, it was determined that the occultingfeature could not be located in the circumstellar disk aroundRW Aur A and therefore may not be in Keplerian orbit.Recent hydrodynamical simulations by Dai et al. (2015)support the interpretation that a star-disk tidal encounter dur-ing a fly-by of RW Aur B could explain the unusual morphol-ogy of the RW Aur system. They found a strong agreementbetween their simulations and the millimeter observations byCabrit et al. (2006), which first proposed the star-disk fly-byscenario. The model predicts that the line of sight to RWAur A currently intersects a bridge of stripped-o ff materialbetween the two stars. Dai et al. (2015) argue that the bridgestructure may have small clumps of dense material that couldoccult the primary star. These simulations support the orig-inal hypothesis presented in Paper I and predict the possibleoccurrence of additional dimming events. In addition, numer-ical simulations of eccentric binary interactions of classicalT Tauri stars suggest that the interaction can create accretionstreams of inner disk material onto the stellar photosphere.These streams would be created near apastron and eventiallyform into “ring-like” structures around each star (Sytov et al.2011; G´omez de Castro et al. 2013).In this paper, we present new high-cadence photometry ofthe RW Aur system showing a shallow dimming in 2012-2013and a second, larger dimming event in 2014-2015. The 2014-2015 event was first reported by Petrov et al. (2015) and re-sembles the dimming observed in late 2010 (Paper I). We ap-ply similar geometric and kinematic arguments as we did forthe 2010 event to show that the new dimmings are consistentwith another clump of material from the tidally disrupted disk. PHOTOMETRIC OBSERVATIONS
Several photometric surveys have observed RW Aur overboth short and long timescales going back to 1899. Here wedescribe the observations used in our analysis.
KELT-North
Starting in 2003, the Kilodegree Extremely Little Telescope(KELT)-North survey has been continuously observing theentire sky between a declination of +
18 and +
44, search-ing for transiting Hot Jupiters around bright stars (8 < V < ◦ × ◦ with 23 ′′ per pixel.All observations are in a broad R -band filter with a ∼
15 mincadence (Pepper et al. 2007, 2012). RW Aur is located inKELT-North Field 04, which is centered on α = δ = + ◦ ′ ′′ J2000. We obtained 9619 imagesof field 04 from UT 2006 October 27 to UT 2014 December31. The data acquisition and reduction is described in detailin § Kutztown University Observatory
RW Aur was observed in
BVRI using the 0.61 m Ritchey-Chr´etien optical telescope at the Kutztown University Obser-vatory (KUO) in Kutztown, Pennsylvania. A total of 2305 F ig . 1.— The KUO field-of-view for RW Aur. The standard reference starsare labeled as A, B, C, and D. data images were obtained, as listed in Table 1, over 43 nightsbetween UT 2014 February 25 and UT 2015 March 30.The telescope’s f / × µ m) pixels, yields a field of view of19 ′ .5 × ′ .0. The CCD was kept at an operating temperatureof -15 ◦ C and dark, flat, and bias calibration frames were ap-plied to all data images in the usual way. A sample data imageis given in Figure 1, labeling RW Aur and four standard ref-erence stars. The known magnitudes of the reference stars arelisted in Table 2. A standard method of aperture photometrywas employed, and the instrumental magnitudes were color-corrected using several Landolt standard fields. The KUO ob-servations do not resolve the RW Aurigae system.The observed
BVRI light curves are displayed in Figure 5and the B − V color curve is shown in Figure 6. American Association of Variable Star Observers(AAVSO)
AAVSO is a dedicated, non-profit orginization with the pri-mary goal of understanding all types of variable stars. Thearchive consists of data from astronomers, both amateur andprofessional, around the world. Reported observations forRW Aur begin in 1937 and the data used in this work areeither in the V band or visual observations. The AAVSO ob-servations do not resolve the RW Aur system. RESULTS
In this section we review the results from the 2010-2011eclipse and present new observations from the KELT-NorthSurvey showing two additional dimming events in 2012-2013 and the 2014-2015 dimming event first announced byPetrov et al. (2015).
Photometric analysis of the RW Aur system by the KELT-North and AAVSO surveys showed that in late 2010 the RWAur system dimmed from V ∼ . V ∼
12 for ∼ V ∼ .
7) to a ff ect thetotal brightness of the system ( V ∼ . F ig . 2.— The KELT-North (Blue) and AAVSO (Black) observations plotted for the 9 KELT-North seasons. The three gray-shaded regions correspond tothe 2010-2011, 2012-2013 and 2014-2015 large dimming events. The AAVSO and KUO data are in Visual and V-band magnitudes while the KELT-Northobservations are in instrumental magnitudes, that we approximate to the V-band but no attempt has been made to place all the data on the same absolute scale.TABLE 2T he properties of the reference stars . T he quoted apparent magnitudes were obtained from the AAVSO V ariable S tar D atabase (A. A. H enden privatecommunication ). T he cited sources are as follows : † APASS, †† T ycho -2, ††† TASS.Reference Star LabelA B C DTycho-ID TYC 2389-936-1 TYC 2389-589-1 N / A TYC 2389-630-1R.A. (J2000) 05:07:24.62 05:07:50.55 05:07:35.27 5:08:05.17DEC. (J2000) 30:20:28.1 30:19:02.5 30:24:47.0 30:18:40.6 B ± †† ± † ± † ± † V ± †† ± † ± † ± † R — 11.665 ( ± † ± † ± † I ± ††† ± † ± † ± † ( B − V ) 0.518 ( ± ± ± ± in star-disk accretion process. Paper I modeled the dimmingas an occultation of RW Aur A by a large body which pos-sessed a sharp leading edge perpendicular to its direction ofmotion. Combining this model with kinematic and geometricarguments, Paper I argued that RW Aur A was occulted bya large ( ∼ ∼ − and if in Keplerian orbit, would be located ∼
180 AU from the star. Since the known short disk (57 AU,Cabrit et al. (2006) around RW Aur A is quite inclined to ourline of sight ( > ◦ ), in Paper I we argued that the occultingbody could not lie within the disk plane. In the two seasons following the 2010-2011 dimming, themedian brightness of the RW Aur system was slightly fainterthen the median brightness of the season prior to the dimming( V ∼ . V = . V = . ∼
40 days (See Figure 3). Aswith the 2010 eclipse, if we assume the entire event is the re-sult of only RW Aur A dimming, then RW Aur A dimmed by ∼ § / egress timescale is sim-ilar to what was determined for the 2010-2011 event, this im-plies that the occulting bodies that caused the 2010-2011 and2012-2013 dimmings are moving at a similar velocity and arelocated at a similar semi-major axis (that are 2.6 km s-1 and180 AU, respectively). The duration of the 2012-2013 eventis only ∼
40 days, which implies that the occulting body is 2.6km s − ×
40 days = ∼ As first reported by Petrov et al. (2015), after the seasonalobserving gap in mid-2014 the RW Aur system appearedsignificantly dimmer then in the previous observing season.From analyzing the H α and He I line at 5875Åbefore and dur-ing the 2014 dimming, they find no evidence that the knownhigh accretion rate of RW Aur A has changed during the timeof the dimming. This suggests that the dimmings are unre-lated to the accretion process which takes place close to thestar. Their observations of the Na I D lines and the Ca II K F ig . 3.— A zoom-in of Figure 2 in late 2012 to early 2013 showing a smalldimming of the RW Aur system. The shaded region is the estimated ingress(the period of time for the dimming). line provide evidence that the stellar winds of RW Aur A havechanged significantly and propose that the increased wind ve-locity is pushing dust from the disk across our line of sight.Resolved U BVRI photometric observations of the RW Aursystem during the 2014-2015 dimming indicated that RW AurA was dimmer by > ∼ F ig . 4.— (Top) Recreation of Figure 3 from Paper I showing the 2010-2011large dimming event. (Bottom) Zoom in of the last two KELT-North seasonsshowing the 2014-2015 dimming. Combining the KELT, AAVSO and KUO observations, wefind that the combined RW Aur system dimmed by ∼ R Band), similar indepth to the 2010-2011 dimming (See Figure 2). Similar tothe season prior to the 2010-2011 dimming, the RW Aur sys-tem was slightly brighter at a V ∼ . BVRI )photometric monitoring of the entire RW Aur system prior toand during the 2014-2015 dimming. We find that the depth is ∼ B , ∼ V , ∼ R and ∼ I . Antipin et al. (2015) showed that during the beginning ofthe 2014-2015 dimming, RW Aur A had dimmed by ∼ BVRI ). Therefore, since our observations donot resolve the system, this depth di ff erence is likely a resultof the fact that the light from RW Aur B is included in ourmeasurement. In § ∼ ∼ ∼ − . From theKELT-North, AAVSO, and KUO observations, the 2014-2015dimming lasted at least the entire duration of the observingseason, ∼
250 days. Using this duration and adopting the cal-culated transverse velocity from the analysis of the 2010-2011dimming suggests that the minimum width of the occulter isat least 2.6 km s − ×
250 days = − × = F ig . 5.— The KUO BVRI light curves of RW Aur covering the 2014-2015large dimming event. These observations do not resolve the RW Aur system. econd Large Occultation of RW Aur A 5 DISCUSSION
In Paper I, we argued that the 2010-2011 dimming wascaused by a consolidation of tidally disrupted material oc-culting RW Aur A. ”In this section, we argue that the re-cent dimming events observed here, and first mentioned byPetrov et al. (2015), support this interpretation. We also dis-cuss the possiblity of grain growth in the disrupted material.
Interpretation: Occultation by the RW Aur A TidallyDisrupted Disk Material
Using the IRAM Plateau de Bure Interferometer, the RWAurigae system was mapped in CO and dust continuum(Cabrit et al. 2006). Their observations showed a long tidalarm wrapped around RW Aur A. Through comparison withnumerical simulations by Clarke & Pringle (1993), they pro-posed that RW Aur A recently experienced a fly-by from RWAur B. This interaction would have significantly disrupted thedisk originally around RW Aur A, resulting in the truncateddisk and the large tidal arm. Also, the spectroscopic obser-vations during the eclipse show that the accretion rate of RWAur A did not change during the dimming (Chou et al. 2013).Based on the millimeter and spectroscopic observations, Pa-per I proposed that the 2010-2011 dimming was caused byan occultation of the primary star, RW Aur A, by tidally dis-rupted material. This hypothesis has been supported by thehydrodynamical simulations by Dai et al. (2015) which sug-gest the occulting body to be a bridge of disrupted materialconnecting RW Aur A and B. The simulations also predictthe possibility of additional dimming events.Photometric monitoring of the RW Aur system from KELT-North, KUO and AAVSO show two additional dimmingevents that occurred after the 2010-2011 dimming event.From the 2012-2013 dimming, we estimate a similar ingresstimescale as we did for the 2010-2011 event. We do not havecoverage of ingress / egress for the 2014-2015 event due to theseasonal observing gaps. The similarities in the initial dim-ming duration for the 2010-2011 and 2012-2013 events sug-gest that the occulting bodies for both events are moving atsimilar velocities and likely at similar distances from RW AurA (if we assume the occulting material to be in a Keplerian or-bit). Therefore, it is likely that the cause of both (and possiblyall three) dimming events are related.In this work and in Paper I, we made some simple assump-tions (sharp leading edge and Keplerian motion) to determinesome characteristics of the occulting bodies that caused thethree dimming events observed. Our calculated transverse ve-locity from the 2010-2011 and 2012-2013 dimmings are con-sistent with the measured velocities of the tidally disruptedmaterial from the millimeter observations by Cabrit et al.(2006). This velocity suggests a semi-major axis of ∼
180 AU,which is less than the projected separation of RW Aur A and B( ∼
200 AU) but larger then the estimated extent of the knowndisk around RW Aur A (57 AU, Cabrit et al. (2006)). The hy-drodynamical simulation by Dai et al. (2015) of the RW Aureccentric fly-by nicely replicates the millimeter observationsby Cabrit et al. (2006) and support the interpretation that the2010-2011 occultation was caused by an occultation of RWAur A by tidally disrupted material. The simulations alsosuggest the possibility of additional dimming events in thefuture. Our observations and analysis of the 2012-2013 and2014-2015 dimmings are consistent with the simulations andinterpretation first proposed in Paper I.Since no other dimming event was observed for ∼
50 years prior to 2010 (Paper I) and two more dimmings have oc-curred since, it is probable that the 2010-2011 event was theleading front of tidally disrupted material and more dimmingare likely to occur. From the hydrodynamical simulations,Dai et al. (2015) suggested that the occulting mechanism maybe a bridge of material connecting RW Aur A and B. Althoughthis structure is poorly resolved in the simulations, they esti-mate it to be ∼
100 AU wide ( ∼ ′′ in the plane of the sky).If the 2010-2011 dimming was caused by the leading edge ofthis bridge of material, and is moving 0.8-2.6 km s − as ourcalculations have shown, it will take 180-600 years for thetrailing edge of the bridge to fully cross our line of sight. In ∼ ff erent durations. Therefore, it is possible that >
50 moredimmings of RW Aur could occur over the next century.Young stellar objects have also been seen to eject blobsof gas with very high velocities. An example of this is HH30, a young system in which gas blobs with sizes similar toour solar system being ejected from the star at ∼
220 km s − (Burrows et al. 1996). These blobs are mesured to be ∼ ′′ inwidth (or ∼
56 AU wide using the ∼
140 parsec distance to theTaurus molecular cloud) (Burrows et al. 1996; Elias 1978). Ifwe assume that the cause of the RW Aur dimmings is causedby similarly sized gas blobs, the time required for one blobto cross our line of size at the observed velocity for the HH30 gas blobs would be ∼ + years (Paper I), and now we have severalsuch events within the last few years, it is likely that we areobserving the leading edge of the bridge structure, and thesuccessive occultations presented here and in Paper I repre-sent smaller coherent structures within it. Evidence for Grain Growth in the Tidally Disrupted Disk
Although our color observations do not resolve the RW Aursystem, observations by Antipin et al. (2015) show that RWAur B has remained constant during the 2014-2015 dimming,indicating that the color changes observed are likely not due to F ig . 6.— The KUO B − V (top) and V − R (bottom) color curve of RWAur A during the 2014-2015 dimming. The BVR brightness of RW Aur B(Antipin et al. 2015) has been subtracted from the KUO observations
RW Aur B. Using the measure
BVR magnitudes for RW AurB (14.5, 13.8 and 12.92 respectively, (Antipin et al. 2015)),we subtract the brightness of RW Aur B from the KUO
BVR observations to create the B − V an V − R color plots seenin Figure 6. The B − V color of RW Aur increased froma quiescent value of ∼ . ∼ ∼ V mag of the star dur-ing the middle of the dimming remained roughly constant,dimmed by ∼ B − V from0.6 to 1.0 indicates a B − V color excess, E( B − V ) = A V = R ( V ) = A V / E ( B − V ) = / =
5. This is the standard valuetypically adopted for dust grains in molecular clouds and starforming regions, which suggests the presence of dust grains,and not just molecular material, in the environment of thisyoung system. The E( B − V ) value then drops steadily to ∼ R ( V ) > / >
10. In other words,the occulting material becomes steadily more grey. Rather,the shift in observed R ( V ) indicates that the dust grains atthe leading edge of the occulting feature are relatively smalland representative of dust in star forming regions, whereas thematerial deeper into the occulter is likely comprised of largergrains, which could be dust grains onto which ice mantles have developed, and / or larger coagulated grains or pebbles(assuming the extinction is optically thin). Another possibil-ity is that the steadily greyer extinction is caused by having alarger fraction of the occulting material be optically thick asthe dimming progressed. Antipin et al. (2015) find that theRW Aur A spectral energy distribution changes during the2014-2015 occultation and they suggest both a grey extinc-tion and a selective extinction, similar to our results here. Wedo not know whether this evolved protoplanetary material ex-isted prior to the tidal disruption event, or if the growth of thedust grains was aided by the disruption event. The evidencesuggests that the building blocks of planetary material can ex-ist in the space between binary stars, perhaps through fly-byinteractions such as that seems to have occurred in the RWAur system. SUMMARY AND CONCLUSIONS
With the deep dimming event observed in 2010-2011 (Pa-per I), the intrigue surrounding the RW Aur system has dra-matically increased. New photometric observations fromKELT, AAVSO and KUO show two additional dimmingevents in 2012-2013 and 2014-2015 (the 2014-2015 dimmingwas first announced by Petrov et al. (2015)). From our analy-sis, the observations of the additional dimming events are con-sistent with an occultation of RW Aur A by tidally disruptedmaterial lying far outside the extent and plane of its short cir-cumstellar disk. This interpretation has been supported by hy-drodynamical simulations of the proposed RW Aur star-diskinteraction (Dai et al. 2015).Multiband photometric observations during the 2014-2015dimming showed that the occulting material became steadilymore grey as the dimming progressed. This suggests thatthe outer portion of the occulting body consists of small dustgrains while the core is primarily made up of (either or allthree) large dust grains, dust enveloped in ice or is opticallythick. Either way, the observations are consistent with evolvedprotoplanetary material and the evolution of this material mayhave been expedited by the tidal interaction. The recurringdimmings of RW Aur A will continue to increase the inter-est surrounding this unique system. Continued monitoring ofRW Aur will provide insight into the e ff ect the tidal interac-tion will have on the evolution of planetesimals from the dis-rupted material. Future high spatial resolution observations ofthis system would be of great value in clarifying the nature ofthe circumstellar environment.Early work on KELT-North was supported by NASA GrantNNG04GO70G. J.A.P. and K.G.S. acknowledge support fromthe Vanderbilt O ffi ce of the Provost through the VanderbiltInitiative in Data-intensive Astrophysics. This work has madeuse of NASA’s Astrophysics Data System and the SIMBADdatabase operated at CDS, Strasbourg, France.Work by B.S.G. and T.G.B. was partially supported by NSFCAREER Grant AST-1056524.We acknowledge with thanks the variable star observationsfrom the AAVSO International Database contributed by ob-servers worldwide and used in this research. REFERENCESAntipin, S., Belinski, A., Cherepashchuk, A., et al. 2015, InformationBulletin on Variable Stars, 6126, 1Beck, T. L., & Simon, M. 2001, AJ, 122, 413 Bisikalo, D. V., Dodin, A. V., Kaigorodov, P. V., et al. 2012, AstronomyReports, 56, 686Burrows, C. J., Stapelfeldt, K. R., Watson, A. M., et al. 1996, ApJ, 473, 437Cabrit, S., Pety, J., Pesenti, N., & Dougados, C. 2006, A&A, 452, 897 econd Large Occultation of RW Aur A 7econd Large Occultation of RW Aur A 7