Towards Planetesimals in the Disk around TW Hya 3.5 centimeter Dust Emission

a r

X

i

v

:a

s t

r o

-

p

h

/

5

6

6

4

4v

1

2

7

J u

n

2

5

accepted by ApJL,May 11,2005Towards Planetesimals in the Disk around TW Hya:λ=3.5centimeter Dust Emission D.J.Wilner 1P.D’Alessio 2N.Calvet 1M.J.Claussen 3L.Hartmann 1dwilner@https://www.sodocs.net/doc/f92597028.html, ABSTRACT We present Very Large Array observations at λ=3.5cm of the nearby young star TW Hya that show the emission is constant in time over weeks,months and years,and spatially resolved with peak brightness temperature ～10K at ～0.′′25(15AU)resolution.These features are naturally explained if the emission mechanism at this wavelength is thermal emission from dust particles in the disk surrounding the star.To account quantitatively for the observations,we construct a self-consistent accretion disk model that incorporates a population of centimeter size particles that matches the long wavelength spectrum and spatial distribution.A substantial mass fraction of orbiting particles in the TW Hya disk must have agglomerated to centimeter size.These data provide the ?rst clear indication that dust emission from protoplanetary disks may be observed at centimeter wavelengths,and that changes in the spectral slope of the dust emission may be detected,providing constraints on dust evolution and the planet formation process.Subject headings:circumstellar matter —planetary systems:protoplanetary

disks —stars:individual (TW Hya)

1.Introduction

Many young stars show thermal emission from solids in circumstellar disks with proper-ties thought similar to the early Solar System.Early analyses of spectral energy distributions

indicated disk radii of10’s to100’s of AU(Adams,Lada&Shu1987)and masses of gas and dust su?cient to form planetary systems like our own(Beckwith et al.1990).Statistical studies indicate disk dissipation on timescales of～10Myr(Strom et al.1989),compatible with the standard paradigm of giant planet cores forming by dust coagulation followed by planetesimal accretion(Wuchterl et al.2000).A mysterious aspect of this process is that in-terstellar dust grains overcome energetic obstacles to sticking and grow to sizes large enough to decouple from the disk gas and interact gravitationally(Youdin&Shu2002).Most mod-els postulate a phase of collisional agglomeration mediated by such mechanisms as brownian motion and turbulence.Numerical simulations of disk evolution indicate a bimodal distri-bution of particle sizes develops as millimeter size aggregates grow and settle to the disk mid-plane(e.g.Weidenschilling1997),where the resulting dense layer becomes a reservoir for the formation of planetesimals and larger bodies.

Observations of dust emission over a wide range of wavelengths provide diagnostic in-formation on particle properties(Beckwith&Sargent1991;Beckwith,Henning&Nakagawa 2000).Recently recognized nearby stellar associations o?er prime targets for observations to address the evolution of dust towards planets(Zuckerman2001).The TW Hya association, which contains more than20stellar systems with ages estimated to be5to10Myr,is the nearest known site of recent star formation activity(Song,Zuckerman&Bessell2003).The classical T Tauri star TW Hya,an apparently single star of mass0.8M⊙,has become the focus of considerable attention because of its proximity(56pc),and because it retains a remarkable face-on circumstellar disk of radius225AU visible in scattered light(Krist et al.2000;Weinberger et al.2002)and in thermal emission from dust and trace molecules (Weintraub,Sandell,&Duncan1989;Kastner et al.1997;Qi et al.2004).

The dust in the TW Hya disk,like the dust in disks around many younger stars,has long been known to show evidence for size evolution from a primordial interstellar distribution. At optical and near-infrared wavelengths,the TW Hya disk appears almost spectrally gray (Roberge,Weinberger&Malamuth2005),indicative of grain growth from sub-micron sizes in the upper layers of the disk.At submillimeter and millimeter wavelengths,the spectral slope indicates a shallow dependence of the particle mass opacity on wavelength,κ～λβ, withβ～1.For the TW Hya disk,well resolved images of thermal dust emission con?rm that optical depth e?ects do not a?ect the determination of the spectral slope(Wilner et al. 2000).For compact spherical particles of size<<λ,β→2,while for particles of size>>λ,β→0.In realistic astrophysical mixtures,βalso partly depends on grain composition, structure,and topology,though the particle size generally dominates,especially for common silicates(Pollack et al.1994;Miyake&Nakagawa1993).The shallow submillimeter spectral slope found for observations of TW Hya has been robustly interpreted as evidence for particle growth to sizes of order1millimeter or more(Calvet et al.2002;Natta et al.2004).

Since the particle size probed by observation is comparable to wavelength,detection of dust emission at centimeter wavelengths has the potential to reveal the development and location of larger“pebbles”within disks.Two problems have historically thwarted this goal (Mundy et al.1993):(1)the opacity per unit mass decreases for larger particles and the resulting emission is weak,and(2)the centimeter wavelength emission of younger(～1 Myr old)stars is often dominated by hot plasma in the system,either gyrosynchrotron emission associated with chromospheric activity,or thermal bremsstralung emission from partially ionized winds.For TW Hya,the close distance and the advanced age mitigate these problems.We present new observations of TW Hya at centimeter wavelengths using the Very Large Array1(VLA)that convincingly argue for emission from a population of large dust particles in the disk.

2.Observations and Results

We used the VLA in late2001and early2002to observe TW Hya atλ=3.5cm at approximately bi-weekly intervals over8epochs in the D con?guration(2001.732,2001.770, 2001.808,2001.847,2001.885,2001.923,2001.956,2001.999),which provided～8′′(450AU) maximum resolution,and at7epochs in the A con?guration(2002.047,2002.088,2002.140, 2002.186,2002.252,2002.340,2002.381),which provided considerably higher～0.′′25(15AU) resolution.The typical rms noise in the images obtained from each of these short observations was～40μJy.In addition,we observed TW Hya atλ=6cm in the DnC con?guration in 2001September24,with beam size～22′′and rms noise70μJy,to provide supplemental spectral index information.The AIPS software was used for calibration and imaging of each VLA data set following standard procedures.The systematic uncertainty in the absolute ?ux scales at these wavelengths is less than10%.

2.1.

3.5cm emission not variable

A previous VLA3.5cm3σupper limit of84μJy(Rucinski1992)for TW Hya compared to a subsequent detection at200±28μJy indicated that the emission at this wavelength was variable(Wilner et al.2000).This apparent variability,together with the discovery that TW Hya is a source of strong X-rays,suggested that the emission was due to hot gas resulting from stellar magnetic activity,which is known to produce variations at radio wavelengths of

an order of magnitude or more on timescales of hours to months to years(Kutner,Rydgren &Vrba1986;White et al.1992;Chiang,Phillips&Lonsdale1996).Our initial impetus to monitor the TW Hya3.5cm emission at frequent intervals was to investigate the variability, and to determine if the radio?ux reached su?cient levels to be amenable to Very Long Baseline Interferometry.

Figure1shows the TW Hya D con?guration3.5cm measurements as a function of time. No variations from the260μJy mean were observed within the～40μJy rms uncertainties at each epoch.By contrast,a second radio source within the?eld of view with?ux comparable to TW Hya,presumably a background active galactic nucleus,showed variations by more than a factor of two during this period.

The surprising lack of variability for the TW Hya radio emission prompted us to check the previously reported upper limit.A re-analysis of these data from1991January30in the VLA archive shows that the noise was signi?cantly higher than reported,and excision of bad data results in a detection of TW Hya at the210±55μJy level.The early observations that were thought to indicate variability in fact are compatible with all of the subsequent (constant)measurements.

2.2.

3.5cm emission spatially resolved

The A con?guration3.5cm TW Hya observations spatially resolve the emission region. Because the brightness distribution is strongly centrally peaked and falls o?steeply with radius,this is best illustrated by examination of the interferometer data in the visibility domain.Figure2shows the real part of the3.5cm visibility as a function of baseline length, annularly averaged in bins of88kλwidth,combining data from all of the observed epochs to maximize the signal-to-noise ratio.The fallo?in visibility at longer baselines is the signature that the emission is spatially resolved.

3.Discussion

3.1.Dust Emission at3.5cm

The lack of time variability,together with spatial extent orders of magnitude larger than a stellar radius,provide strong constraints on the emission mechanism.In particular, chromospheric activity–expected to arise from transient magnetic features comparable in extent to the star–is not responsible for the3.5cm emission.More detailed X-ray data

bolster this conclusion.Observations of the X-ray spectrum with XMM-Newton(Stelzer& Schmitt2004)and Chandra(Kastner et al.2002)indicate electron densities two orders of magnitude higher than typically found in stellar coronae.In addition,the X-ray emitting material appears to be depleted of grain-forming elements like Mg and Si.The X-ray emission is more likely the result of accretion shocks at the star/disk interface.Any radio emission from such accretion shocks would be on such small size scales that spatial extent would not be resolved by the VLA observations.

An ionized wind does not present a viable explanation for the3.5cm emission,either. While there is evidence for small amounts of ionized gas in the TW Hya system provided by optical spectroscopy of e.g.Hαemission(Muzerolle et al.2000),there are no indications from direct imaging of a wind in the form of an optical jet,as found in younger stars with higher disk accretion rates.A persuasive scaling argument against signi?cant wind emission has been made previously from TW Hya’s low inferred accretion rate of～10?9M⊙yr?1, which suggests a low mass out?ow rate,and therefore a3.5cm?ux more than an order of magnitude lower than observed(Calvet et al.2002).The new VLA observations provide two additional arguments against a wind origin for the emission.First,the3σupper limit at6cm constrains the spectral spectral index from6to3.5cm to be>0.3,which is incompatible with the nearly?at value of optically thin ionized gas,and steeper than typically found for the winds from young stars(Anglada et al.1998).Second,and more important,the peak brightness temperature at3.5cm in the resolved data is only～10K,which is compatible with ionized gas only if the optical depth or?lling factor of optically thick emission is very low,unlike any known ionized jet from a pre-main-sequence star.

Dust emission naturally accounts for all of the properties of the3.5cm emission:(1) lack of time variability,(2)large spatial extent,(3)low brightness at high spatial resolution, and(4)the spectral index constraint.We conclude that the3.5cm emission from TW Hya is strongly dominated by thermal emission from dust particles in the disk surrounding the star.

3.2.A Disk Model

The full spectrum of TW Hya from the infrared through the millimeter,has been matched very well by a self-consistent irradiated accretion disk model that assumes standard optical properties for a mix of spherical particle constituents and a single-power power law size distribution(Calvet et al.2002).This model does not?t the observations at3.5cm, however.The dashed curve in Figure3shows that this previous model falls short of the observations at3.5cm by a factor of a few,which is much larger than the observational

uncertainty.The basic reason is that many more grains that emit e?ciently at centimeter wavelengths must be included in order to raise the level of dust emission at long wavelengths.

A disk model that includes a plausible population of large grains can match the data. For illustration,we assume a simple model that contains two grain populations:(1)small grains with a size distribution(as a function of particle radius a)n(a)～a?3.5between minimum and maximum sizes a min=0.005μm and a max=1μm,and(2)larger“pebbles”with size distribution n(a)～a?2.5,between minimum and maximum sizes a min=5mm and a max=7mm.The small grains,which contain only0.1%of the particle mass in the disk, are needed to regulate the irradiated disk structure and to explain the infrared spectrum, including the silicate feature near10μm(D’Alessio,Calvet,&Hartmann2001).The solid curve in Figure3shows that the model with large grains boosts the long wavelength emission to match the observed?ux density.The same model also matches very well the observed brightness distribution.Figure2shows the Fourier transform of the brightness distribution derived from this model as a solid curve,and the model follows the fallo?in the visibilities. Note that the large grains that account for the3.5cm emission must be present to radii of at least10’s of AU to match the observed brightness distribution.The disk mass in this model is0.1M⊙,assuming a standard dust-to-gas mass ratio.This disk mass lies below the limit beyond which gravitational instabilities are expected to set in.

The parameters in this simple model are clearly not unique.The strictly bimodal and discontinuous particle size distribution used in this disk model is meant to be illustrative of grain growth and settling,and to capture the essential behavior of more sophisticated treatments that result in small grains?oating above a dense layer in the mid-plane(e.g. Weidenschilling1997;Tanaka,Himeno&Ida2005).The detailed situation is complex,with faster particle growth in the inner disk,and competition between growth,settling and tur-bulence,at all disk radii.The new observations,which show elevated emission from dust at 3.5cm and a change in the spectral slope of the dust emission from submillimeter to cen-timeter wavelengths,are di?cult to explain unless substantial particle evolution has occurred throughout the disk.Future observations with higher signal-to-noise at high resolution hold the promise to map the spectral index of the dust emission in the disk and locate regions of changing grain properties.

3.3.Planet Formation

Independent of the details of the disk model,the detection of dust emission at centimeter wavelengths provides a strong indications that an early phase of the planet building process is underway in the TW Hya disk.The inference of a～4AU radius inner hole in dust disk

from the mid-infrared spectral energy distribution has previously given rise to the suggestion that a planet may be present in the TW Hya system(Calvet et al.2002).In this context,it is interesting to consider that changes in dust properties may a?ect the timescale for planet formation.A key result in recent studies of giant planet formation by core accretion is the extraordinary sensitivity of the time of onset of a rapid gas accretion phase to assumptions of disk surface density and dust opacities;for example,Hubickyj,Bodenheimer,&Lissauer (2004)show that decreasing dust opacity can reduce the formation time of gas giant planets by more than a factor of two from the canonical8Myr of older models(Pollack et al.1996). This shorter giant planet formation timescale is more comfortably consistent with estimates for the age of the TW Hya system.

3.4.Radio Emission from TW Hya in Context

It would be extremely interesting if other classical T Tauri stars showed invariate cen-timeter emission with spectral properties indicative of dust emission like TW Hya.Unfortu-nately,existing VLA surveys of the large sample of sources associated with the most nearby dark clouds detect only younger Class I objects or transition objects with high accretion rates that drive radio and optical jets,like HL Tau and DG Tau(Cohen,Bieging&Schwartz1982; Bieging,Cohen&Schwartz1984),and weak-lined T Tauri stars with stellar activity(e.g. Chiang,Phillips&Lonsdale1996).For classical T Tauri stars,the upper limits at6cm are generally160-300μJy.A few sources have comparable or better limits at3.5cm,e.g.DO Tau<170μJy(Koerner,Chandler&Sargent1995)and DL Tau and GG Tau<77μJy (Mundy et al.1993).If TW Hya were at140pc,then it would be among the stronger dust disks at millimeter wavelengths,but its3.5cm emission of only～40μJy would lie well below the existing detection limits for sources at that distance.More sensitive observations at a range of centimeter wavelengths are needed to determine what fraction,if any,of the disks around classical T Tauri stars show evidence for extreme grain growth like that inferred for TW Hya.

The detection of dust emission at centimeter wavelengths from the TW Hya disk provides direct evidence for much larger particles than have been detected before in a nebula like the one from which the Solar System is thought to have emerged.This?rst observation points the way to a new?eld of investigation.Next generation radio telescopes,in particular the the Square Kilometer Array(Lazio,Tarter&Wilner2004),will have a combination of sensitivity and angular resolution su?cient to image in detail the distribution of large dust particles emitting at centimeter wavelengths in the disks around hundreds of young stars in nearby dark clouds and beyond.

Partial support for this work was provided by NASA Origins of Solar Systems Program Grants NAG5-11777and NAG5-9670.

REFERENCES

Adams,F.C.,Lada,C.J.&Shu,F.H.1987,ApJ,312,788

Anglada,G.,Villuendas,E.,Estalella,R.,Beltran,M.T.,Rodriguez,L.F.,Torrelles,J.M., Curiel,S.1998,AJ,116,2953

D’Alessio,P.,Calvet,N.&Hartmann,L.2001,ApJ,553,321

Beckwith,S.V.W.,Sargent,A.I.,Chini,R.,Gusten,R.1990,AJ,99,924

Beckwith,S.V.W.&Sargent,A.I.1991,ApJ,381,250

Beckwith,S.V.W.,Henning,T.&Nakagawa,Y.2000,Protostars and Planets IV(Tucson: University of Arizona Press),eds.Mannings,V.,Boss,A.P.,Russell,S.S.,533 Bieging,J.H.,Cohen,M.&Schwartz,P.R.1984,ApJ,282,699

Calvet,N.,D’Alessio,P.,Hartmann,L.,Wilner,D.,Walsh,A.,Sitko,M.2002,ApJ,568, 1008

Chiang,E.,Phillips,R.B.&Lonsdale,C.J.1996,AJ,111,355

Cohen,M.,Bieging,J.H.,&Schwartz1982,ApJ,253,707

Hubickyj,O.,Bodenheimer,P.,Lissauer,J.J.2004,RMxAC,22,83.

Kastner,J.H.,Zuckerman,B.,Weintraub,D.A.,Forveille,T.1997,Science,277,67 Kastner,J.H.,Huenemoerder,D.P.,Schulz,N.S.,Canizares,C.R.,Weintraub,D.A.2002, ApJ,567,434

Koerner,D.W.,Chandler,C.J.,Sargent,A.I.1995,ApJ,425,L69

Krist,John E.,Stapelfeldt,Karl R.,Menard,Francois,Padgett,Deborah L.,Burrows, Christopher J.2000,ApJ,538,793.

Kutner,M.L.,Rydgren,A.E.Vrba,F.J.1986,AJ,92,895

Lazio,T.J.W.,Tartar,J.C.,Wilner,D.J.2004,New.Ast.Reviews,48,985

Miyake,K.&Nakagawa,Y.K.1993,Icarus,106,20

Mundy,L.G.,McMullin,J.P.,Grossman,A.W.,Sandell,G.1993,Icarus,106,11 Muzerolle,J.,Calvet,N.,Briceno,C.,Hartmann,L.,Hillenbrand,L.2000,ApJ,535,L47 Natta,A.,Testi,L.,Neri,R.,Shepherd,D.S.;,Wilner,D.J.2004,A&A,416,179 Pollack,James B.,Hollenbach,David,Beckwith,Steven,Simonelli,Damon P.,Roush,Ted 1994,ApJ,421,615

Pollack,James B.,Hubickyj,Olenka,Bodenheimer,Peter,Lissauer,Jack J.,Podolak,Morris Greenzweig,Yuval1996,Icarus,124,62

Qi,C.,Ho,P.T.P.,Wilner,D.J.,Takakuwa,S.,Hirano,N.,Ohashi,N.2004,ApJ,616,L11 Roberge,Aki,Weinberger,Alycia J.,Malumuth,Eliot M.2005,ApJ,in press,astro-ph/0410251

Rucinski,S.M.1992,PASP,104,311

Song,Inseok,Zuckerman,B.&Bessell,M.S.2003,ApJ,599,342.

Stelzer,B.&Schmitt,J.H.M.M.2004,A&A,418,687.

Strom,K.M.,Strom,S.E.,Edwards,S.,Cabrit,S.,Skrutskie,M.F.1989,AJ,97,1451 Tanaka,H.,Himeno,Y.,Ida,S.2005,ApJ,in press,astro-ph/0502287 Weidenschilling,S.J.1997,Icarus,127,290

Weinberger,A.J.,Becklin,E.E.,Schneider,G.,Chiang,E.I.,Lowrance,P.J.,Silverstone, M.,Zuckerman,B.,Hines,D.C.,Smith,B.A.2002,ApJ,566,409

Weintraub,D.A.,Sandell,G.&Duncan,W.1989,ApJ,340,L69.

White,S.M.,Pallavicini,R.,Kundu,M.R.1992,A&A,257,557.

Wilner,D.J.,Ho,P.T.P.,Kastner,J.H.,Rodriguez,L.F.2000,ApJ,534,L101 Wuchterl,G.,Guillot,T.,Lissauer,J.J.2000in Protostars and Planets IV(Tucson:Univer-sity of Arizona Press),eds.Mannings,V.,Boss,A.P.,Russell,S.S.,p.1081 Youdin,A.N.&Shu,F.H.2002,ApJ,580,494

Zuckerman,B.2001,ARAA,39,549

Fig.1.—The3.5cm continuum emission from TW Hya shows no variability.The?ux of the TW Hya system was monitored at8epochs using the VLA in the D con?guration with ～8′′resolution,and the points show the measured?uxes with error bars that represent the rms noise in each observation.The solid line indicates the260μJy mean value.

Fig.2.—The3.5cm continuum emission from TW Hya is spatially resolved by observations with a subarcsecond beam.The plot is derived from the sum of observations obtained at7 epochs with the VLA in the A con?guration.The points show the real part of the3.5cm visibility as a function of baseline length,annularly averaged in bins of width88kλ.The error bars represent±1standard deviation for each bin.The solid curve shows the predicted emission from an irradiated accretion disk with a population of large dust particles(see §3.2).

Fig. 3.—The?lled circles show VLA measurements at7mm and3.5cm,and the arrow represents an upper limit at6cm.The long wavelength spectrum of TW Hya is better?t by a model in which nearly all of the mass in solids is in centimeter size grains(solid line)than by the model of Calvet et al.(2002)based on a single-power power law grain size distribution

(dashed line).