Magnetic activity and hot Jupiters of young Suns: the weak-line T Tauri stars V819 Tau and V830 Tau
JF Donati, E Hébrard, G Hussain, C Moutou, L Malo, K Grankin, AA Vidotto, SHP Alencar, SG Gregory, MM Jardine, G Herczeg, J Morin, R Fares, F Ménard, J Bouvier, X Delfosse, R Doyon, M Takami, P Figueira, P Petit, I Boisse, MaTYSSE collaboration
aa r X i v : . [ a s t r o - ph . S R ] S e p MNRAS , 1–15 (2015) Preprint 27 September 2018 Compiled using MNRAS L A TEX style file v3.0
Magnetic activity and hot Jupiters of young Suns:the weak-line T Tauri stars V819 Tau and V830 Tau
J.-F. Donati , ⋆ , E. H´ebrard , , G.A.J. Hussain , , C. Moutou , L. Malo , K. Grankin ,A.A. Vidotto , S.H.P. Alencar , S.G. Gregory , M.M. Jardine , G. Herczeg ,J. Morin , R. Fares , F. M´enard , J. Bouvier , , X. Delfosse , , R. Doyon ,M. Takami , P. Figueira , P. Petit , , I. Boisse , and the MaTYSSE collaboration Universit´e de Toulouse, UPS-OMP, IRAP, 14 avenue E. Belin, Toulouse, F–31400 France CNRS, IRAP / UMR 5277, Toulouse, 14 avenue E. Belin, F–31400 France ESO, Karl-Schwarzschild-Str. 2, D-85748 Garching, Germany CFHT Corporation, 65-1238 Mamalahoa Hwy, Kamuela, Hawaii 96743, USA Crimean Astrophysical Observatory, Nauchny, Crimea 298409 Observatoire de Gen`eve, Chemin des Maillettes 51, CH-1290 Versoix, Switzerland Departamento de F`ısica – ICEx – UFMG, Av. Antˆonio Carlos, 6627, 30270-901 Belo Horizonte, MG, Brazil SUPA, School of Physics and Astronomy, Univ. of St Andrews, St Andrews, Scotland KY16 9SS, UK Kavli Institute for Astronomy and Astrophysics, Peking University, Yi He Yuan Lu 5, Haidian Qu, Beijing 100871, China LUPM, Universit´e de Montpellier, CNRS, place E. Bataillon, F–34095 Montpellier, France INAF - Osservatorio Astrofisico di Catania, via S. Sofia 78, I–95123 Catania, Italy CNRS, UMI-FCA / UMI 3386, France, and Universidad de Chile, Santiago, Chile Universit´e Grenoble Alpes, IPAG, BP 53, F–38041 Grenoble C´edex 09, France CNRS, IPAG / UMR 5274, BP 53, F–38041 Grenoble C´edex 09, France D´epartement de physique, Universit´e de Montr´eal, C.P. 6128, Succursale Centre-Ville, Montr´eal, QC, Canada H3C 3J7 Institute of Astronomy and Astrophysics, Academia Sinica, PO Box 23-141, 106, Taipei, Taiwan Centro de Astrof`ısica, Universidade do Porto, Rua das Estrelas, 4150-762 Porto, Portugal Universit´e Aix-Marseille, LAM, F–13388 Marseille, France CNRS, LAM / UMR 7326, F–13388 Marseille, France
Accepted 2015 August 6. Received 2015 August 5; in original form 2015 May 8
ABSTRACT
We report results of a spectropolarimetric and photometric monitoring of the weak-lineT Tauri stars (wTTSs) V819 Tau and V830 Tau within the MaTYSSE programme,involving the ESPaDOnS spectropolarimeter at the Canada-France-Hawaii Telescope.At ≃ ≃ ◦ tothe rotation axis. They are significantly weaker than the field of GQ Lup, an accretingclassical T Tauri star (cTTS) with similar mass and age which can be used to comparethe magnetic properties of wTTSs and cTTSs. The reconstructed brightness maps ofboth stars include cool spots and warm plages. Surface differential rotation is small,typically ≃ × smaller than on the Sun, in agreement with previous results on wTTSs.Using our Doppler images to model the activity jitter and filter it out from theradial velocity (RV) curves, we obtain RV residuals with dispersions of 0.033 and0.104 km s − for V819 Tau and V830 Tau respectively. RV residuals suggest that a hotJupiter may be orbiting V830 Tau, though additional data are needed to confirm thispreliminary result. We find no evidence for close-in giant planet around V819 Tau. Key words: stars: magnetic fields – stars: formation – stars: imaging – stars: rotation– stars: individual: V819 Tau and V830 Tau – techniques: polarimetric ⋆ E-mail: [email protected] (cid:13)
J.-F. Donati et al.
In the last decades, our understanding of the key role thatmagnetic fields play during the first stages of the life of starsand their planets improved significantly, through both theo-retical and observational advances (e.g., Andr´e et al. 2009).This is true in particular for low-mass pre-main-sequence(PMS) stars in the T Tauri star (TTS) phase, aged 0.5–10 Myr, that recently emerged from their dust cocoons andare still in a phase of gravitational contraction towards themain sequence (MS). Being either classical T-Tauri stars(cTTSs) when still surrounded by and accreting from mas-sive, planet-forming accretion discs, or weak-line T-Tauristars (wTTSs) once they have exhausted the gas in theirinner discs, TTSs have been the subject of intense scrutinyat all wavelengths in recent decades given their interest forbenchmarking the scenarios currently invoked to explainlow-mass star and planet formation.Magnetic fields of cTTSs are known to influence andsometimes even trigger several of the main physical pro-cesses playing a role at this stage, like accretion, outflowsand angular momentum transport. As a result, they largelycontrol the rotational evolution of low-mass PMS stars (e.g.,Bouvier et al. 2007; Frank et al. 2014). More specifically,large-scale fields of cTTSs can evacuate the central regionsof accretion discs, funnel the disc material onto the stars,and enforce corotation between cTTSs and their inner-discKeplerian flows. This is causing cTTSs to rotate muchmore slowly than expected from the contraction and ac-cretion of the high-angular-momentum disc material (e.g.,Davies et al. 2014). Fields of cTTSs and of their discs canalso affect the formation and migration of planets (e.g.,Baruteau et al. 2014). Last but not least, magnetic fields ofcTTSs / wTTSs are known to trigger thermally-driven windsthrough the heating provided by accretion shocks and / orAlfv´en waves (e.g., Cranmer 2009; Cranmer & Saar 2011),leading to flares, coronal-mass ejections and angular momen-tum loss (e.g., Matt et al. 2012; Aarnio et al. 2012).For these reasons, characterizing magnetic fields ofcTTSs and wTTSs through observations is a crucial stepto guide theoretical models towards more physical realismand reliable predictions. Although first detected more than15 yrs ago (e.g., Johns-Krull et al. 1999; Johns-Krull 2007),magnetic fields of TTSs are not yet fully characterized.In particular, field topologies of cTTSs have only recentlybeen unveiled for a dozen stars (e.g., Donati et al. 2007;Hussain et al. 2009; Donati et al. 2010, 2013) through theMaPP (Magnetic Protostars and Planets) Large ObservingProgramme allocated on the 3.6 m Canada-France-HawaiiTelescope (CFHT) with the ESPaDOnS high-resolutionspectropolarimeter (550 hr of clear time over semester 2008bto 2012b). This first survey revealed for instance that large-scale fields of cTTSs remain relatively simple and mainlypoloidal when the host star is still fully or largely convec-tive, but become much more complex when the star turnsmostly radiative (Gregory et al. 2012; Donati et al. 2013).This survey also demonstrated that these fields vary over afew years (e.g., Donati et al. 2011, 2012, 2013) and resemblethose of mature stars with comparable internal structures(Morin et al. 2008b), hinting at a dynamo rather than fossilorigin.The situation is even worse for wTTSs, with only two of them magnetically imaged to date (namely V410 Tau andLkCa 4, Skelly et al. 2010; Donati et al. 2014). In both cases,the large-scale fields found are unexpected with respect towhat we already know on cTTSs and MS dwarfs with sim-ilar internal structures. The most recent observations, fo-cussed on the wTTS LkCa 4, revealed the presence of aring of intense toroidal field encircling the star at low lati-tudes, despite LkCa 4 being most likely fully convective andthus not supposed to exhibit this kind of magnetic feature(Donati et al. 2014). Clearly, a more systematic explorationof large-scale fields of wTTSs needs to be carried out in asimilar manner to that of cTTSs. This is the goal of the newMaTYSSE (Magnetic Topologies of Young Stars and theSurvival of close-in giant Exoplanets) Large Programme, al-located at CFHT over semesters 2013a-2016b (510 hr) withcomplementary observations with the NARVAL spectropo-larimeter on the 2-m T´elescope Bernard Lyot at Pic du Midiin France (420 hr, allocated) and with the HARPS spec-tropolarimeter at the 3.6-m ESO Telescope at La Silla inChile (70 hr, allocated). MaTYSSE also allows us to studymagnetic winds of wTTSs and the corresponding spin-downrates (e.g., Vidotto et al. 2009, 2010, 2014); it also offers anovel option for assessing with improved sensitivity the po-tential presence of hot Jupiters (hJs) at an early stage inthe life of low-mass stars and their planets, and thus for ver-ifying whether core accretion coupled to migration is indeedthe best option for forming close-in giant planets.In practice, MaTYSSE aims at carrying out a surveyof about 30 wTTSs, along with a long-term monitoring of ≃ MNRAS , 1–15 (2015) agnetometry & velocimetry of the wTTSs V819 Tau & V830 Tau tial presence of hJs around both stars (Sec. 5). Finally, wesummarize our main results and discuss their implicationsfor our understanding of low-mass star and planet formation(Sec. 6). V819 Tau and V830 Tau were both observed in 2014 Decem-ber and 2015 January using ESPaDOnS at the CFHT. ES-PaDOnS collects stellar spectra spanning the entire opticaldomain (from 370 to 1,000 nm) at a resolving power of 65,000(i.e., a resolved velocity element of 4.6 km s − ) over the fullwavelength range (Donati 2003). A total of 15 circularly-polarized (Stokes V ) and unpolarized (Stokes I ) spectra werecollected for both stars over a timespan of 28 nights, corre-sponding to about 5 rotation cycles for V819 Tau and 10cycles for V830 Tau. Time sampling was irregular in thefirst two thirds of the run (as a result of poor weather) butdenser in the last third, yielding a reasonably good phasecoverage for both stars despite a few remaining phase gapsof 0.15–0.20 cycle.All polarisation spectra consist of 4 individual subexpo-sures taken in different polarimeter configurations to allowthe removal of all spurious polarisation signatures at first or-der. All raw frames are processed as described in the previ-ous papers of the series (e.g., Donati et al. 2010, 2011, 2014),to which the reader is referred for more information. Thepeak signal-to-noise ratios (S/N, per 2.6 km s − velocity bin)achieved on the collected spectra range between 140 and 220(median 190) for V819 Tau, and between 140 and 180 (me-dian 170) for V830 Tau, depending mostly on weather/seeingconditions. All spectra are automatically corrected of spec-tral shifts resulting from instrumental effects (e.g., mechan-ical flexures, temperature or pressure variations) using tel-luric lines as a reference. Though not perfect, this procedureprovides spectra with a relative RV precision of better than0.030 km s − (e.g., Moutou et al. 2007; Donati et al. 2008).The full journal of observations is presented in Table 1 forboth stars.Rotational cycles of V819 Tau and V830 Tau (noted E and E in the following equation) are computedfrom Barycentric Julian Dates (BJDs) according to theephemerides:BJD (d) = , , . + . E (for V819 Tau)BJD (d) = , , . + . E (for V830 Tau)(1)in which the photometrically-determined rotation periods P rot (equal to 5.53113 and 2.74101 d respectively, Grankin2013) are taken from the literature and the initial Juliandates (2,457,011.0 and 2,457,011.8 d) are chosen arbitrarily.Least-Squares Deconvolution (LSD, Donati et al. 1997)was applied to all spectra. The line list we employed forLSD is computed from an Atlas9
LTE model atmosphere(Kurucz 1993) featuring an effective temperature of T e ff = , K and a logarithmic gravity (in cgs unit) of log g = . ,appropriate for both V819 Tau and V830 Tau (see Sec. 3).As usual, only moderate to strong atomic spectral lines areincluded in this list (see, e.g., Donati et al. 2010, for more de-tails). Altogether, about 7,800 spectral features (with about40% from Fe i ) are used in this process. Expressed in units ofthe unpolarized continuum level I c , the average noise levels Table 1.
Journal of ESPaDOnS observations of V819 Tau (first15 lines) and V830 Tau (last 15 lines) collected from mid 2014 De-cember to mid 2015 January. Each observation consists of a se-quence of 4 subexposures, each lasting 1400 s and 700 s forV819 Tau and V830 Tau respectively (except on 2015 Jan 15for which subexposures on V819 Tau were shortened to 900 s).Columns − respectively list (i) the UT date of the observa-tion, (ii) the corresponding UT time (at mid-exposure), (iii) theBarycentric Julian Date (BJD), and (iv) the peak signal to noiseratio (per 2.6 km s − velocity bin) of each observation. Column5 lists the rms noise level (relative to the unpolarized continuumlevel I c and per 1.8 km s − velocity bin) in the circular polar-ization profile produced by Least-Squares Deconvolution (LSD),while column 6 indicates the rotational cycle associated with eachexposure (using the ephemerides given by Eq. 1).Date UT BJD S/N σ LSD
Cycle(2014) (hh:mm:ss) (2,457,000+) (0.01%)Dec 19 12:07:54 11.01081 140 3.9 0.002Dec 20 07:55:28 11.83548 210 2.5 0.151Dec 21 07:17:08 12.80882 210 2.4 0.327Dec 22 08:14:51 13.84886 200 2.6 0.515Dec 28 10:50:17 19.95655 190 2.8 1.619Dec 29 07:37:39 20.82273 180 2.9 1.776Dec 30 06:10:02 21.76183 200 2.6 1.946Jan 07 07:19:42 29.80976 190 2.7 3.401Jan 08 06:20:11 30.76836 190 2.7 3.574Jan 09 06:15:12 31.76484 190 2.9 3.754Jan 10 06:15:60 32.76533 170 3.2 3.935Jan 11 07:25:26 33.81348 220 2.3 4.125Jan 12 06:28:22 34.77379 200 2.4 4.298Jan 14 06:17:48 36.76631 190 2.8 4.658Jan 15 06:12:46 37.76274 150 3.5 4.839Dec 20 09:13:43 11.88992 170 2.8 0.033Dec 21 08:33:48 12.86217 170 2.8 0.387Dec 22 09:29:48 13.90104 180 2.7 0.766Dec 28 12:19:59 20.01897 140 3.6 2.999Dec 29 08:53:59 20.87588 160 3.0 3.311Dec 30 07:26:53 21.81535 160 2.9 3.654Jan 07 08:35:55 29.86285 170 2.8 6.590Jan 08 07:36:48 30.82174 180 2.7 6.940Jan 09 07:31:38 31.81810 170 2.9 7.303Jan 10 07:32:25 32.81858 150 3.3 7.668Jan 11 08:42:02 33.86686 180 2.6 8.051Jan 12 05:12:50 34.72151 170 2.9 8.362Jan 13 05:03:32 35.71500 160 3.0 8.725Jan 14 05:02:24 36.71414 170 2.8 9.089Jan 15 07:12:21 37.80431 170 2.9 9.487 of the resulting Stokes V LSD signatures range from 2.3 to3.9 × − per 1.8 km s − velocity bin - with a median valueof 2.8 × − for both stars.Zeeman signatures are detected at all times in Stokes V LSD profiles (see Fig. 1 for an example), featuring am-plitudes of 0.3–1%, i.e., indicative of significant large-scalefields. Asymmetries and / or distortions are also visible inStokes I LSD profiles, suggesting the presence of brightnessinhomogeneities at the surfaces of V819 Tau and V830 Tauat the time of our observations.Contemporaneous BVR J photometric observations werealso collected from the Crimean Astrophysical Observatory(CrAO) 1.25 m telescope between 2014 Aug and Dec forboth stars, indicating that V819 Tau and V830 Tau were ex- MNRAS , 1–15 (2015)
J.-F. Donati et al.
Figure 1.
LSD circularly-polarized (Stokes V , top/red curve)and unpolarized (Stokes I , bottom/blue curve) profiles ofV819 Tau (top panel) and V830 Tau (bottom panel) collectedon 2015 Jan. 07 (cycle 3.401) and 2014 Dec. 30 (cycle 3.654).Clear Zeeman signatures are detected in the LSD Stokes V profileof both stars (with a complex shape in the case of V830 Tau), inconjunction with the unpolarized line profiles. The mean polariza-tion profiles are expanded (by a factor of 10 and 20 for V819 Tauand V830 Tau respectively) and shifted upwards by 1.06 for dis-play purposes. hibiting brightness modulations with full amplitudes of 0.45and 0.10 mag in V (see Table 2) and periods of . ± . and . ± . d respectively (compatible within error barswith the average periods of Grankin 2013, used to phase ourspectroscopic data, see Eq. 1). These photometric variations(unusually large for V819 Tau and small for V830 Tau, seeFig. 12a and Table 3 of Grankin et al. 2008) are commonlyattributed to the presence of brightness features at the sur-face of both stars. Applying to our highest S/N spectra the automatic spec-tral classification tool especially developed in the con-text of MaPP and MaTYSSE, inspired from that of
Table 2.
Journal of contemporaneous CrAO multicolour photo-metric observations of V819 Tau (first 22 lines) and V830 Tau(last 19 lines) collected in late 2014, respectively listing the UTdate and Heliocentric Julian Date (HJD) of the observation, themeasured V magnitude, B − V and V − R J Johnson photometriccolours, and the corresponding rotational phase (using again theephemerides given by Eq. 1).Date HJD
V B − V V − R J Phase(2013) (2,456,000+) (mag)Aug 26 896.5665 12.824 1.512 1.461 0.311Aug 27 897.5395 12.954 1.488 1.478 0.487Aug 29 899.5421 13.124 1.577 1.520 0.849Aug 30 900.5427 12.911 1.486 1.522 0.030Aug 31 901.5438 12.788 1.522 1.487 0.211Sep 01 902.5487 12.922 1.524 1.509 0.393Sep 02 903.5425 13.154 1.549 1.614 0.572Sep 04 905.5024 12.965 1.558 1.518 0.927Sep 05 906.5354 12.801 1.549 1.470 0.113Sep 20 921.5742 13.142 1.499 1.544 0.832Sep 20 921.5742 13.204 1.519 1.522 0.832Oct 05 936.6035 13.099 1.618 1.447 0.549Oct 15 946.5991 12.898 1.496 1.541 0.357Oct 26 957.4906 12.838 1.514 1.485 0.326Oct 28 959.5578 13.259 1.572 1.578 0.700Nov 02 964.4855 13.163 1.658 1.539 0.590Nov 05 967.6033 12.784 1.656 1.437 0.154Nov 05 967.6106 12.795 1.518 1.513 0.155Nov 13 975.4293 13.159 1.611 1.574 0.569Nov 14 976.4257 13.230 1.558 1.592 0.749Dec 13 1005.4237 12.956 1.615 1.516 0.992Dec 14 1006.5232 12.804 1.586 1.506 0.191Aug 22 892.5484 12.321 1.351 1.319 0.494Aug 27 897.5627 12.297 1.403 1.297 0.323Aug 29 899.5574 12.209 1.370 1.270 0.051Aug 30 900.5540 12.288 1.358 1.311 0.414Aug 31 901.5514 12.238 1.368 1.296 0.778Sep 01 902.5578 12.262 1.378 1.288 0.145Sep 02 903.5502 12.321 1.375 1.351 0.507Sep 04 905.5599 12.304 1.389 1.286 0.241Sep 05 906.5443 12.289 1.387 1.298 0.600Oct 05 936.5812 12.278 1.398 9.999 0.558Oct 15 946.5610 12.297 1.405 1.303 0.199Oct 28 959.5923 12.220 1.365 1.298 0.953Nov 02 964.5193 12.274 1.387 1.275 0.751Nov 05 967.6268 12.210 9.999 9.999 0.884Nov 05 967.6307 12.224 1.345 9.999 0.886Nov 13 975.4453 12.280 1.379 1.294 0.737Nov 14 976.4437 12.277 1.351 1.321 0.101Dec 13 1005.4898 12.319 1.367 1.326 0.698Dec 14 1006.4869 12.268 1.403 1.309 0.062
Valenti & Fischer (2005) and discussed in a previous paper(Donati et al. 2012), we find that the photospheric temper-atures and logarithmic gravities of V819 Tau and V830 Tauare respectively equal, for both stars, to T e ff = ± Kand log g = . ± . (with g in cgs units). Repeating the oper-ation on different spectra of our data set yields consistent pa-rameters, with differences between results within the quotederror bars. The temperature we obtain is higher than thatoften quoted in the literature for both stars (in the range3900–4100 K, see, e.g., Sestito et al. 2008; Furlan et al. 2009;Grankin 2013; Herczeg & Hillenbrand 2014), and derived inmost cases from either photometry or low-resolution spec- MNRAS , 1–15 (2015) agnetometry & velocimetry of the wTTSs V819 Tau & V830 Tau Figure 2.
Observed (open squares and error bars) location ofV819 Tau and V830 Tau in the HR diagram. The PMS evolution-ary tracks (full and dashed lines) and corresponding isochrones(dotted lines, all from Siess et al. 2000) assume solar metallicityand include convective overshooting. The blue circle indicates thelocation of the cTTS GQ Lup according to Donati et al. (2012).The green line depicts where models predict PMS stars start de-veloping their radiative core as they contract towards the mainsequence. troscopy. In the particular case of V830 Tau, our estimate isin agreement with the only other spectroscopic study basedon high-resolution data that we know of in the refereed lit-erature (Schiavon et al. 1995). Our result also confirms thatV819 Tau and V830 Tau are warmer than our previous tar-get LkCa 4 (whose temperature we estimated to be T e ff = ± K, Donati et al. 2014). From the difference betweenthe B − V index expected at this temperature ( . ± . ,as estimated through an interpolation within Table 6 ofof Pecaut & Mamajek 2013) and the averaged value mea-sured for both V819 Tau and V830 Tau (equal to . ± . and . ± . respectively, see Kenyon & Hartmann 1995;Grankin et al. 2008, see also Table 2), we derive that theamounts of visual extinction A V that our two targets suf-fer are equal to . ± . for V819 Tau and . ± . forV830 Tau (in reasonable agreement with other independentresults, e.g., Herczeg & Hillenbrand 2014).Long-term photometric monitoring of V819 Tau andV830 Tau indicates that their maximum brightnesses cor-respond to V magnitudes of 12.8 and 11.9 respectively(Grankin et al. 2008); assuming a spot coverage of ≃ V magnitudes of . ± . and . ± . for V819 Tau andV830 Tau respectively. Using the visual bolometric correc-tion expected for the adequate photospheric temperature(equal to − . ± . , Pecaut & Mamajek 2013) and the ac-curate distance estimate of the Taurus L1495 dark cloud( ± pc, corresponding to a distance modulus equal to . ± . , Torres et al. 2012), we finally obtain bolometricmagnitudes of . ± . and . ± . , or equivalently log-arithmic luminosities relative to the Sun of − . ± . and . ± . for V819 Tau and V830 Tau respectively. Couplingwith the photospheric temperature obtained previously, we Table 3.
Main parameters of V819 Tau and V830 Tau as de-rived from our study except when indicated (with G13 and T12standing for Grankin 2013 and Torres et al. 2012 respectively),with v rad noting the RV that the star would have if unspotted (asinferred from the modeling of Sec. 4).V819 Tau V830 Tau Reference M ⋆ ( M ⊙ ) . ± .
05 1 . ± . R ⋆ ( R ⊙ ) . ± . . ± . age (Myr) ≃ . ≃ . T e ff (K) ±
50 4250 ± L ⋆ / L ⊙ ) − . ± .
10 0 . ± . P rot (d) 5.53113 2.74101 G13 v sin i (km s − ) . ± . . ± . v rad (km s − ) . ± . . ± . i ( ◦ ) ±
10 55 ± distance (pc) ± ± T12 find respective radii of . ± . and . ± . ⊙ for our twotarget stars.The rotation periods of both V819 Tau and V830 Tauare well determined from long-term multi-colour photo-metric monitoring, and equal to 5.53113 d and 2.74101 d(Grankin 2013) respectively; estimates from different studies(e.g., Xiao et al. 2012) agree to a precision of better than 1%,indicative of robust and reliable measurements. Couplingthese rotation periods along with our measurements of theline-of-sight-projected equatorial rotation velocities v sin i ofV819 Tau and V830 Tau (respectively equal to . ± . and . ± . km s − , see Sec. 4), we can infer that the R ⋆ sin i val-ues of both stars are . ± .
05 R ⊙ and . ± .
03 R ⊙ , where R ⋆ and i note the radius of the stars and the inclination oftheir rotation axis to the line of sight. Comparing with theradii derived from the luminosities and photospheric tem-perature, we derive that i is equal to 35 ◦ and 55 ◦ (to anaccuracy of ≃ ◦ ) for V819 Tau and V830 Tau respectively.Using the evolutionary models of Siess et al. (2000, as-suming solar metallicity and including convective overshoot-ing), we obtain that V819 Tau and V830 Tau are both . ± .
05 M ⊙ stars, and that their respective ages are ≃ ≃ ≃ ≃ . This supports thatdisc dissipation for PMS stars in Taurus may occur ontimescales as short as a few Myrs (Williams & Cieza 2011;Ingleby et al. 2012), even though 80–90% of single stars inthis star-formation region still host discs at similar ages(Kraus et al. 2012). This makes V819 Tau and V830 Tauatypical in this respect, and thus of obvious interest forMaTYSSE. We also note that our two targets are bothlocated close to the theoretical threshold at which 1 M ⊙ stars cease to be fully convective (green line in Fig. 2). Thismay suggest that V830 Tau is still fully convective whereasV819 Tau has already developed a small radiative core; our Using the most recent evolutionary models of Baraffe et al.(2015), we obtain slightly smaller masses (of 0.90 M ⊙ ) for bothstars, along with smaller ages (of ≃ ≃ , 1–15 (2015) J.-F. Donati et al. error bars on the location of both stars in the HR diagramare however still too large to reach a firm conclusion (seeFig. 2), to the point that even the opposite conclusion couldactually be true. Interestingly, V819 Tau and V830 Tau arealso located close to the classical T Tauri star GQ Lup inthe HR diagram (Donati et al. 2012, see circle in Fig. 2),implying that their internal structures should be similar, ifthat of GQ Lup is not significantly impacted by accretion(only moderate for this cTTS, e.g., Donati et al. 2012). Wesummarize the main parameters of both stars in Table 3.Finally, we report that core emission is clearly detectedin the Ca ii infrared triplet (IRT) lines of both V819 Tauand V830 Tau, with an average equivalent width of the emis-sion core equal to ≃
15 km s − , corresponding to the amountexpected from chromospheric emission for such PMS stars.Moreover, the He i D line is barely visible (average equiva-lent width of ≃ − ) for both V819 Tau and V830 Tau,further demonstrating that accretion is no longer takingplace on to their surfaces, in agreement with previous stud-ies. Now that our two stars are well characterized regard-ing their atmospheric properties and evolutionary sta-tus, we are ready to apply our dedicated stellar-surfacetomographic-imaging package to the spectropolarimetricdata set described in Sec 2. This tool is based on theprinciples of maximum-entropy image reconstruction andon the assumption that the observed variability is mainlycaused by rotational modulation (with an added optionfor differential rotation); this framework is well adaptedto the case of wTTSs for which rotational modulationlargely dominates intrinsic variability, in both photome-try (e.g., Grankin et al. 2008) and photospheric lines (e.g.,Skelly et al. 2010; Donati et al. 2014). Since initially re-leased (Brown et al. 1991; Donati & Brown 1997), the codeunderwent several upgrades (e.g., Donati 2001; Donati et al.2006b), the most recent ones being its re-profiling tothe specific needs of MaPP and MaTYSSE observations(Donati et al. 2010, 2014). More specifically, the imagingcode is set up to invert (both automatically and simulta-neously) time series of Stokes I and V LSD profiles intobrightness and magnetic maps of the stellar surface; more-over, brightness imaging is now allowed to reconstruct bothcool spots and warm plages, known to contribute to the ac-tivity of very active stars (Donati et al. 2014). The readeris referred to the papers mentioned above for more generaldetails on the imaging method .To compute the disc-integrated average photosphericLSD profiles, we start by synthesizing the local Stokes I and V profiles using Unno-Rachkovsky’s analytical solutionto the polarized radiative transfer equations in a Milne-Eddington model atmosphere, taking into account the lo-cal brightness and magnetic field; we then integrate these As explained in Donati et al. (2014), full Stokes spectropo-larimetry is not expected to bring much improvement in imag-ing quality over Stokes I and V spectropolarimetry, especially forstars as faint as our MaTYSSE targets. local profiles over the visible hemisphere to retrieve the syn-thetic profiles to be compared with our observations. Thiscomputation method provides in particular a reliable de-scription of how line profiles are distorted in the presenceof magnetic fields (including magneto-optical effects, e.g.,Landi degl’Innocenti & Landolfi 2004). The main parame-ters of the local profile are similar to those used in ourprevious studies, the wavelength, Doppler width, equivalentwidth and Land´e factor being respectively set to 670 nm,1.8 km s − , 3.9 km s − and 1.2. As part of the imaging pro-cess, we obtain accurate estimates for several parameters ofboth stars; we find in particular that the v rad ’s (the RVs thestars would have if unspotted) and v sin i ’s are respectivelyequal to . ± . and . ± . km s − for V819 Tau, and . ± . and . ± . km s − for V830 Tau (as listed inTable 3), in good agreement with previously published es-timates (e.g., Nguyen et al. 2012). We will revisit the RVcurves of both stars in more details in Sec. 5. We show in Fig. 3 our sets of Stokes I and V LSD pro-files of V819 Tau and V830 Tau along with our fits to thedata. The fits we obtain correspond to a reduced chi-square χ equal to 1 (i.e., a χ equal to the number of fitted datapoints, respectively equal to 360 and 705 for V819 Tau andV830 Tau), starting from initial values of about 65 and 10(corresponding to null magnetic fields and unspotted bright-ness maps) for V819 Tau and V830 Tau respectively. Thisfurther stresses the quality of our data set and the highperformance of our imaging code at modelling the observedmodulation of LSD profiles (also obvious from Fig. 3). Thevery good quality of the achieved fit clearly demonstratesthat the assumption underlying our modelling, i.e., that theobserved line profile variability is caused by the presenceof brightness and magnetic surface structures rotating inand out of the observer’s view, is verified. As for LkCa 4(Donati et al. 2014), the reasonably dense phase coverageallows to track back most of the main reconstructed featuresinto genuine profile distortions, and thus to safely claim thatour maps include no major imaging artifact nor bias.The reconstructed brightness maps of V819 Tau andV830 Tau (see Fig. 4) both include cool spots and warmplages, but nevertheless feature a number of significant dif-ferences, regarding the contrast of these features in partic-ular, much larger for V819 Tau than for V830 Tau. Thisdifference is readily visible from our photometric data (col-lected at CrAO, see Table 2 and Fig. 5) whose full ampli-tude is about 4.5 × larger ( ≃ ≃ MNRAS , 1–15 (2015) agnetometry & velocimetry of the wTTSs V819 Tau & V830 Tau Figure 3.
Maximum-entropy fit (thin red line) to the observed (thick black line) Stokes I (first and third panels) and Stokes V (secondand fourth panels) LSD photospheric profiles of V819 Tau (first two panels) and V830 Tau (last two panels) in 2014 Dec and 2015 Jan.Rotational cycles and 3 σ error bars (for Stokes V profiles) are also shown next to each profile. This figure is best viewed in color. comes with a warm plage next to it though this plage isagain less bright than its equivalent on LkCa 4. We suspectthat part (but not all) of this difference reflects the lower v sin i of V819 Tau (with respect to LkCa 4) and thus thepoorer spatial resolution achievable through Doppler imag-ing at the surface of this star. As for LkCa 4, bright plagesare again required to reproduce all details of the profile vari-ability observed for V819 Tau, and in particular the large RVvariations (of full amplitude ≃ − , see Sec. 5) thatV819 Tau exhibits. The reconstructed brightness featuresare found to cover ≃
15% of the overall stellar surface, withthe main cool spot covering ≃
9% and the plage ≃ ≃ ≃
6% for spots and ≃
6% for plages), is comparable to thatof V819 Tau, they feature a lower contrast in average andare spread over a more even range of longitudes, yielding amuch flatter photometric response (see Fig. 5, right panel)despite significant distortions of the LSD Stokes I profiles(see Fig. 3, third panel) and clear RV fluctuations (of full amplitude ≃ − , see Sec. 5). As already evidenced inthe case of V410 Tau (the first wTTS to be magneticallyimaged, although with a slightly older version of our imag-ing code, Skelly et al. 2010), we confirm that a flat photo-metric curve by no means implies the absence of spots (orthe reduction in spot coverage) at the surface of the star,but rather a more even distribution of features (both cooland warm). Moreover, the similar v sin i ’s of V830 Tau andLkCa 4 further argue (along with the different amplitudes ofthe photometric light curves) that the reconstructed bright-ness maps of these two stars trace genuine differences in theparent spot distributions at the surface of these two stars.Finally, we emphasize that the good agreement ob-tained for both stars between the predicted and observedlight curves despite the large time span (of ≃ ≃ MNRAS , 1–15 (2015)
J.-F. Donati et al.
Figure 4.
Maps of the logarithmic brightness (relative to the quiet photosphere), at the surfaces of V819 Tau (left) and V830 Tau (right)in 2014 Dec – 2015 Jan. The stars are shown in flattened polar projection down to latitudes of − ◦ , with the equator depicted as a boldcircle and parallels as dashed circles. Radial ticks around each plot indicate phases of observations. This figure is best viewed in color. Figure 5.
Brightness variations of V819 Tau (left) and V830 Tau (right) predicted from the tomographic modelling of Fig. 4 of ourspectropolarimetric data set (green line), compared with contemporaneous photometric observations in the V band (red open circles anderror bars of 15 mmag) at the 1.25-m CrAO telescope (see Table 2). centrates in spherical harmonics (SH) dipolar modes (i.e.,with ℓ = , ℓ denoting the degree of the modes), whereas90% of it gathers in the aligned dipole mode ( ℓ = and m = , with m denoting the mode order) for both stars. Atfirst order and at some distance from the stars, the poloidalcomponents of V819 Tau and V830 Tau can be approxi-mated with dipoles of respective strengths 400 and 350 G,tilted at angles of ≃ ◦ to the line of sight towards phases0.50 and 0.65 . A weaker octupolar component in the range150–300 G (depending on whether or not we favour oddmodes in the imaging process, as for cTTSs, Donati et al.2011, 2012, 2013), is also present on V819 Tau, more orless aligned with the rotation axis and antiparallel with thedipole component; on V830 Tau, the octupolar componentis weaker and non-axisymmetric, with a polar strength no Note that these tilt and phase values refer to the visible mag-netic pole, i.e., the negative and positive poles for V819 Tau andV830 Tau respectively. Using the positive pole as reference (as inGregory & Donati 2011), the tilt angle and phase of the dipolecomponent of V819 Tau become 150 ◦ and 0.0. larger than 150 G. The reconstructed large-scale fields ofV819 Tau and V830 Tau also include weaker (though sig-nificant) toroidal components with average unsigned fluxesequal to ≃
170 G and ≃
100 G respectively, and whose topolo-gies are more complex and less axisymmetric than that ofLkCa 4 (Donati et al. 2014).SH expansions describing the reconstructed field pre-sented in Fig. 6 are limited to terms with ℓ ≤ and 10 forV819 Tau and V830 Tau respectively; only marginal changesto the solution are observed when increasing the maximum ℓ value beyond these thresholds, demonstrating that mostof the Zeeman signal detected in Stokes V LSD profiles ofboth stars is captured in the images we recovered. As anexample, we present in Fig. 7 the extrapolated large-scalefield topologies of V819 Tau and V830 Tau using the poten-tial field approximation (e.g., Jardine et al. 2002) and de-rived solely from the reconstructed radial field components;as states of lowest possible magnetic energy, these potentialtopologies are shown to provide a reliable description of themagnetic field well within the Alfv´en radius (Jardine et al.2013). From these plots, the dominantly dipolar topologies
MNRAS , 1–15 (2015) agnetometry & velocimetry of the wTTSs V819 Tau & V830 Tau Figure 6.
Maps of the radial (left), azimuthal (middle) and meridional (right) components of the magnetic field B at the surfaces ofV819 Tau (top) and V830 Tau (bottom) in 2014 Dec – 2015 Jan. Magnetic fluxes in the color lookup table are expressed in G. The starsare shown in flattened polar projection as in Fig 4. This figure is best viewed in color. Figure 7.
Potential field extrapolations of the magnetic topologies reconstructed for V819 Tau (left) and V830 Tau (right), as seenby an Earth-based observer at phases 0.0 and 0.15 respectively. Open and closed field lines are shown in blue and white respectively,whereas colors at the stellar surface depict the local values (in G) of the radial field fluxes (as shown in the left panels of Fig. 6). Thesource surface at which the field becomes radial is (arbitrarily) set at a (realistic) distance of 4 R ⋆ . This figure is best viewed in color.MNRAS000
Potential field extrapolations of the magnetic topologies reconstructed for V819 Tau (left) and V830 Tau (right), as seenby an Earth-based observer at phases 0.0 and 0.15 respectively. Open and closed field lines are shown in blue and white respectively,whereas colors at the stellar surface depict the local values (in G) of the radial field fluxes (as shown in the left panels of Fig. 6). Thesource surface at which the field becomes radial is (arbitrarily) set at a (realistic) distance of 4 R ⋆ . This figure is best viewed in color.MNRAS000 , 1–15 (2015) J.-F. Donati et al.
Figure 8.
Variations of χ as a function of Ω eq and d Ω , derived from the modelling of our Stokes I (left) and V (right) LSD profiles ofV830 Tau at constant information content. In both cases, a clear and well defined paraboloid is observed, with the outer color contourtracing the 1.7% increase in χ (or equivalently a χ increase of 11.8 for 705 fitted data points) that corresponds to a 3 σ ellipse for bothparameters as a pair. The middle plot emphasizes how well the confidence ellipses from both measurements overlap, with 1 σ and 3 σ ellipses (depicting respectively the 68.3% and 99.73% confidence levels) shown in full and dashed lines (in red and green for Stokes I and V data respectively). This figure is best viewed in color. of the large-scale fields of V819 Tau and V830 Tau are quiteobvious.Finally, we note that the reconstructed brightness (seeFig. 4) and magnetic maps (see Fig. 6) of V819 Tau andV830 Tau only show a weak level of spatial correlation, asopposed to what we reported for LkCa 4 where the cool polarspot overlapped almost exactly with the region of strongestradial field (see Fig. 4 of Donati et al. 2014). The closestmatch of this kind that we can find in our new data is thecool spot close to the pole of V819 Tau, which roughly co-incides with the visible pole of the dominant dipolar com-ponent of the large-scale field, but not with a local maxi-mum of the surface field (as observed on LkCa 4). This isin line with what is observed on other cool active stars withlarge-scale field strengths similar to those of V819 Tau andV830 Tau (e.g., Donati et al. 2003a); observations so far sug-gest that such tight spatial correlation between brightnessand magnetic maps is only observed for young stars in whichthe large-scale poloidal field exceeds 1 kG (like LkCa 4, butalso, e.g., V410 Tau, GQ Lup, V2129 Oph and DN Tau,Skelly et al. 2010; Donati et al. 2012, 2011, 2014). Spread out over almost a full month (see Table 1), our spec-tropolarimetric data set allows us to estimate, at the sur-faces of both stars, the amount of latitudinal differentialrotation shearing the brightness and / or magnetic maps.We achieve this in practice by assuming that the rotationrate at the surface of the star is varying with latitude θ as Ω eq − d Ω sin θ where Ω eq is the rotation rate at the equatorand d Ω the difference in rotation rate between the equatorand the pole; in this context, one can measure differentialrotation by finding out the pair of parameters Ω eq and d Ω that produces the brightness / magnetic map with mini-mum information content. This procedure was successfullyused in a large number of studies over the last 15 yrs (e.g.,Donati et al. 2000; Petit et al. 2002; Donati et al. 2003b),including on two wTTSs (Skelly et al. 2010; Donati et al. 2014). The reader is referred to these papers for more infor-mation on this technique.Thanks to its high v sin i (see Table 3) and the goodspatial resolution it offers, V830 Tau is an ideal target forestimating differential rotation; moreover, the presence ofsurface features (in both the brightness and magnetic maps)at various latitudes (see Figs. 4 and 6) provides adequate in-formation to obtain an accurate measurement. The χ mapswe obtain (as a function of both Ω eq and d Ω ) are shownin Fig. 8; the one derived from our Stokes I data featuresa clear minimum at Ω eq = . ± . rad d − and d Ω = . ± . rad d − (left panel of Fig. 8), whereas thatinferred from our Stokes V data fully supports this estimatethough with larger error bars ( Ω eq = . ± . rad d − and d Ω = . ± . rad d − , right panel of Fig. 8). Thesevalues translate into rotation periods at the equator and poleof . ± . d and . ± . d respectively, fully consis-tent with the time-dependent periods of photometric fluctu-ations reported in the literature for V830 Tau (ranging from2.738 to 2.747 d when considering only definite detectionsof photometric modulation, Grankin et al. 2008; Xiao et al.2012) known to also probe surface differential rotation.For V819 Tau, the spatial resolution of our maps is sig-nificantly coarser due to the lower v sin i (see Table 3); as a re-sult, our Stokes I data yield a χ map showing no clear min-imum, preventing us from estimating differential rotation.This is actually not very surprising given that the bright-ness map we recovered only includes two main features (onecool spot and one warm plage) more or less located at thesame latitude (see Fig. 4); our data is thus lacking informa-tion about rotation periods and recurrence rates of profiledistortions at different latitudes, without which differentialrotation cannot be estimated (Petit et al. 2002). Our Stokes V data set yield a more favourable situation and allow usto obtain an estimate of surface differential rotation, equalto Ω eq = . ± . rad d − and d Ω = . ± . rad d − ;unsurprisingly, the precision of this estimate is worse thanfor that of V830 Tau, reflecting the coarser spatial resolu-tion achievable for this more slowly rotating star. These val-ues translate into rotation periods at the equator and pole MNRAS , 1–15 (2015) agnetometry & velocimetry of the wTTSs V819 Tau & V830 Tau of . ± . d and . ± . d respectively. Our estimateis reasonably consistent with most photometric periods re-ported in the literature for V819 Tau (ranging from 5.50to 5.61 d, Grankin et al. 2008; Xiao et al. 2012) except forone discrepant value (of 5.73 d, derived from data collectedin 2004, Grankin et al. 2008). Looking back at the originalphotometric data, we find that this discrepant value actu-ally corresponds to a low-accuracy estimate, reflecting thesmall number of measurements and the short observationtime span at this specific epoch.Our results indicate that surface differential rotation issmall on both V819 Tau and V830 Tau, much weaker inparticular than that of the Sun (equal to ≃ . rad d − ).Whereas we cannot exclude (from our data alone) thatV819 Tau rotates as a solid body, the measurement we ob-tain for V830 Tau indicates that it truly rotates differentiallyand that the amount of surface shear that it suffers is ≃ × weaker than that of the Sun, with a typical time of ≃
500 dfor the equator to lap the pole by one complete rotation cy-cle. Similarly small amounts of surface differential rotationwere reported for the fully-convective wTTSs LkCa 4 andV410 Tau (Donati et al. 2014; Skelly et al. 2010) as well asfor fully convective M dwarfs with presumably similar inter-nal structures (Donati et al. 2006a; Morin et al. 2008a).
As previously achieved for the wTTS LkCa 4 (Donati et al.2014), we propose to exploit the results of surface imaging tofilter out the activity jitter in the RV curves of V819 Tau andV830 Tau, and potentially reveal the presence of close-in gi-ant planets (see Donati et al. 2014, for more details). Techni-cally speaking, the goal is to use the brightness and magneticmaps reconstructed from our phase-resolved spectropolari-metric data sets with tomographic imaging, to obtain anaccurate description of the activity jitter that plagues theRV curves of our wTTSs; once RV curves are filtered outfrom the predicted activity jitter, one can look for periodicsignals in the RV residuals that may probe the presence ofhJs. Alternatively, we can simultaneously look for the pres-ence of a planet while carrying out the imaging process, asrecently proposed by Petit et al. (2015); in this case, the goalis to find out the optimal planet parameters, defined as thoseyielding the minimum amount of surface features for a given χ fit to the data. Our initial simulations suggest that bothtechniques should allow one to detect potential hJs orbit-ing wTTSs (provided their periods are not too close to thestellar rotation period or its first harmonics) and to quanti-tatively investigate whether close-in giant planets are muchmore frequent around low-mass forming stars at the TTSstage than around mature MS stars. Ultimately, our goalis to assess the likelihood of the core accretion coupled todisc migration scenario in which giant protoplanets mostlyform in outer accretion discs then migrate inwards until theystop at a distance of a few hundredth of an AU, e.g., oncethey enter the central magnetospheric gaps of cTTSs and nolonger experience the inward torque from the disc.The predicted activity jitter and filtered RV curves wederive for V819 Tau and V830 Tau are shown in Fig. 9. Notethat for our study, RVs are estimated as first order momentsof the Stokes I LSD profiles rather than through Gaussian fits, given the asymmetric and often irregular shapes of spec-tral lines (see Fig. 3, first and third panels). The filteringprocess is found to be quite efficient for V819 Tau, with RVresiduals exhibiting a rms dispersion of only 0.033 km s − (for a full RV amplitude prior to the filtering process equalto ≃ − ). The error bar on these measurements isconservatively set to ± − , reflecting mostly the in-trinsic RV precision of ESPaDOnS (of 0.03 km s − rms, e.g.,Moutou et al. 2007; Donati et al. 2008) and in a smaller partthe intrinsic error of our filtering process (scaling up withthe width of spectral lines, and estimated to be 0.02 km s − rms in the present case). This suggests that V819 Tau isunlikely to host a hJ with an orbital period in the range ofwhat we can detect (i.e. not too close to the stellar rotationperiod or its first harmonics, see Donati et al. 2014), witha 3 σ error bar on the semi-amplitude of the RV residualsequal to 0.07 km s − .For V830 Tau however, the RV curve exhibits significantscatter at several rotation phases (0.05, 0.35 and 0.65, seeFig 9 right panel), causing the RV residuals after filtering toreach a rms dispersion of 0.104 km s − (i.e., about twice andthrice those reported for LkCa 4 and V819 Tau respectively)and to exceed the value (of ≃ − ) that potentially sug-gests the presence of a hJ (according to the results of ourpreliminary simulations, see Donati et al. 2014). The errorbar on our individual RV residuals is set to ± − ,identical to those of LkCa 4 given the very similar widthsof spectral lines and noise levels in our data. A closer lookreveals that this enhanced dispersion is mainly caused by3 points (at rotational cycles 3.654, 8.051 and 8.362, seeFig. 10) that depart (RV-wise) from the bulk of our observa-tions. The temporal variations of the RV residuals are com-patible with a sine wave with a period of either . ± . d or . ± . d and a semi-amplitude of . ± . km s − . Bring-ing a χ improvement of ≃
22 for 4 degrees of freedom, thisfit suggests that the RV residuals do indeed contain a RVsignal, with a confidence level of 99.98%. However, given thevery limited sampling, we cannot firmly conclude that theRV signal we find is truly periodic as expected for a signalof planetary origin, the false alarm probability on the perioddetection (as derived from a Lomb-Scargle periodogram) be-ing uncomfortably high (at about 35%).Experimenting the new technique proposed byPetit et al. (2015), whose advantage is to model bothsurface features and orbital parameters simultaneously,and thus to minimize crosstalk between the two and tomaximize the precision of the whole process (at the expenseof more computing time), we further confirm our finding,and conclude that the best tentative period of the signalin the RV residuals is P orb = . ± . d, whereas the otheroption ( P orb = . ± . d) yields a worse fit to the data andis therefore less likely. The semi-amplitude of the residualRV signal that we derive in this case (taking into ac-count differential rotation as measured in Sec. 4) is equal to K = . ± . km s − , consistent with, and presumably moreaccurate than, our initial estimate (of . ± . km s − ). Aprojection of the achieved χ landscape in the K vs P orb / P rot plane is shown in Fig. 11, demonstrating that our solutioncorresponds to a well defined minimum. The associatedimprovement in χ (with respect to a solution with no RVsignal, see Fig. 8, left panel) is larger than 70 (for 4 degrees MNRAS , 1–15 (2015) J.-F. Donati et al.
Figure 9.
RV variations (in the stellar rest frame) of V819 Tau (left) and V830 Tau (right) as a function of rotation phase, as measuredfrom our observations (open circles) and predicted by the tomographic maps of Fig. 4 (green line). RV residuals (expanded by a factorof 2 for clarity), are also shown (pluses) and exhibit a rms dispersion equal to 0.033 km s − for V819 Tau and 0.104 km s − for V830 Tau.Red, green, dark-blue, light-blue, magenta, grey, orange and purple symbols depict measurements secured at rotation cycles 0, 1, 3(including phase 2.999), 4, 6, 7, 8 and 9 respectively (see Table 1 and Fig. 3). Given the asymmetric and often irregular shapes of Stokes I LSD profiles (see Fig. 3), RVs are estimated as first order moments of the Stokes I LSD profiles rather than through Gaussian fits. RVestimates and residuals are depicted with error bars of ± ± − for V819 Tau and V830 Tau respectively. This figure isbest viewed in color. Figure 10.
RV residuals of V830 Tau after applying our filter-ing process to the spectropolarimetric data set. A sine fit to thepoints with an orbital period of . P rot ≃ . d and a velocity semi-amplitude of 0.11 km s − (dashed line) provides a good match tothe observations (rms dispersion of 0.074 km s − ), suggesting thepossible presence of a giant planet orbiting close to V830 Tau.More observations are obviously needed to confirm or reject thispreliminary result. The color code used for this plot is the sameas that of Fig. 9 (right panel). This figure is best viewed in color. of freedom), implying a definite detection of a RV signal,with a very small false alarm probability ( < − ) .This residual RV signal, if confirmed, would be at-tributable to a ≃ M Jup planet located at a distance of ≃ This false alarm probability is smaller than that obtained bydirectly fitting RVs (equal to 0.02%, see previous paragraph),which confirms that line profiles contain more information thantheir first moments. Note that both tests only refer to the detec-tion of an RV signal and not to its periodic nature (for which thefalse alarm probability is much larger).
Figure 11.
Variations of χ as a function of P orb / P rot and K , de-rived from the simultaneous modelling of the brightness featuresat the surface of V830 Tau and the orbital parameters of a close-in giant planet in circular orbit, following the new technique pro-posed by Petit et al. (2015). A clear minimum is obtained in the χ landscape, whose projection in a K vs P orb / P rot plane passingthrough the minimum is shown here, with the outer color contourdelimiting the 3 σ (or 99.73%) confidence region. This result sug-gests that a ≃ M Jup planet may be orbiting around V830 Tau.This figure is best viewed in color. an orbital plane coinciding with the equatorial plane ofV830 Tau. As expected, the impact of this residual RV sig-nal on the reconstructed maps of Figs. 4 and 6 is small,implying in particular that all conclusions of Sec. 4 are unaf-fected; this also applies to our differential rotation estimatesof V830 Tau, that are not modified by more than a fractionof an error bar once the residual RV signal is removed, evenfor our most accurate measurement from Stokes I data. Sim-ilarly, the 3 σ upper limit on the semi-amplitude of the RVresiduals of V819 Tau (equal to 0.07 km s − ) translates intoa value of ≃ M Jup for the mass of a planet orbiting at a
MNRAS , 1–15 (2015) agnetometry & velocimetry of the wTTSs V819 Tau & V830 Tau Figure 12.
Equivalent width of the emission component in thecore of Ca ii IRT lines (in km s − ) as a function of RV residuals forV830 Tau. No clear correlation is observed between both quan-tities, indicating that the observed RV residuals are likely notattributable to non-rotationally-modulated activity. Note in par-ticular that the 3 highest RV residuals all correspond to averagelevels of IRT core emission, whereas the highest level of IRT coreemission (presumably probing a small flaring event) correspondsto an average RV. The color code used for this plot is the same asthat of Fig. 9 (right panel) and Fig. 10. This figure is best viewedin color. distance of ≃ ii IRT lines,see Fig. 12), indicating that it cannot be attributed to, e.g.,flaring events that may have damaged the profile shapesand corresponding RVs in our spectra. Similarly, we empha-size that these RV variations do occur on relatively shorttimescales of only ≃ In this paper, we report results from a new set of spectropo-larimetric observations collected with ESPaDOnS at theCFHT on two wTTSs, V819 Tau and V830 Tau, from mid2014 December to mid 2015 January, and complemented bycontemporaneous photometric observations from the 1.25-mtelescope at CrAO. As for our initial study on LkCa 4(Donati et al. 2014), the monitoring of these two wTTSs wascarried out in the framework of the international MaTYSSELarge Programme.From the analysis of our spectra, we find that both stars have similar atmospheric properties (photospheric temper-ature of ± K and logarithmic gravity in cgs unitsof . ± . ), suggesting that V819 Tau and V830 Tau are . ± .
05 M ⊙ stars with respective radii of . ± . ⊙ and . ± . ⊙ viewed at inclination angles of 35 ◦ and 55 ◦ .With estimated ages of ≃ ≃ M ⊙ stars start to build up a radiative core - with V819 Taulikely showing up as a slightly more evolved version thanV830 Tau (note however the significant uncertainties on theages, see Fig. 2). V819 Tau and V830 Tau also lie close toGQ Lup in the HR diagram (Donati et al. 2012), althoughthey belong to a different star-formation region.With rotation periods of 5.53 d and 2.74 d, V819 Tauand V830 Tau both spin significantly faster than typicalcTTSs of similar mass (rotating in ≃ . As forLkCa 4, we find that the brightness maps reconstructed forboth stars are in very good agreement with our contempore-neous photometric observations.The large-scale magnetic fields we reconstruct forV819 Tau and V830 Tau are similar, both in terms of inten-sities and topologies, and are found to be 80–90% poloidal;the poloidal component is dominated (especially at somedistance from the stars) by a dipolar term of polar strength350–400 G inclined at ≃ ◦ to the rotation axis whereas theoctupolar term is weaker than the dipolar one. We stressthat neither star shows a conspicuous ring of strong toroidalfields as that reported for LkCa 4 (Donati et al. 2014). Thatboth stars have similar fields despite their × Although long-term photometric monitoring shows thatV830 Tau was indeed exhibiting lower-than-usual photometricvariability at the time of our observations (Grankin et al. 2008),previous studies (e.g., Skelly et al. 2010) demonstrated that over-all activity of low-mass stars, and more specifically the size of theirpolar spots, are not necessarily correlated with the amplitude ofphotometric variability.MNRAS , 1–15 (2015) J.-F. Donati et al. active wTTSs. More unexpected is that these fields are muchweaker than that of GQ Lup (typically hosting a 1 kG dipoleand a 2 kG octupole, see Donati et al. 2012) despite theirproximity in the HR diagram. Although a topological dif-ference (the afore-mentioned toroidal component) was alsoreported between the large-scale fields of LkCa 4 on the onehand, and those of AA Tau and BP Tau on the other hand(located near LkCa 4 in the HR diagram), their poloidalcomponents were similar both in terms of strengths andtopologies.It is obviously still too early to assess whether mag-netic topologies of cTTSs and wTTSs are similar or dissim-ilar in the regions of the HR diagram were both can coex-ist. Our observations so far suggest that they differ, eitherin their poloidal components (this new result compared toDonati et al. 2012) or in their toroidal components (the caseof LkCa 4 vs AA Tau and BP Tau, Donati et al. 2014). Thereason for this difference, if confirmed, will need clarifica-tion and is potentially complex, acting apparently on bothfield strengths and topologies. Having the potential powerto significantly affect atmospheric properties of cTTSs (andthus their location in the HR diagram, Baraffe et al. 2009)and / or to alter their large-scale topologies (by stressingthem through disc/star mass transfers and / or by modifyingthe underlying dynamo processes), accretion can participatein this process. However, the similarity between magnetictopologies of cTTSs and M dwarfs with comparable internalstructures (Gregory et al. 2012; Donati et al. 2014) suggeststhe opposite. The temporal variability inherent to dynamoprocesses is another option; repeated observations of mag-netic fields of cTTSs and wTTSs so far (e.g., Donati et al.2011, 2012, 2013) however indicate that the amplitude ofsuch variations with time is smaller than the one we reporthere. Clearly, we will have to wait for more MaTYSSE stud-ies in the line of our two first ones to progress on this issue.Our data also clearly demonstrate that brightness mapsand / or magnetic topologies of V830 Tau and V819 Tauexperience very little latitudinal shear in a timescale of ≃ v sin i yields the best spatial resolution at the surface of thestar, we estimated the amount of differential rotation ata high-enough precision to claim that it is different fromzero (i.e., not rotating as a solid body) at a confidence levellarger than 99.99%, and smaller than that of the Sun by typ-ically 4.4 × (with a time for the equator to lap the pole byone complete rotation cycle equal to ≃
500 d, as opposed to ≃
110 d for the Sun); moreover, the differential rotation esti-mates we derived from the brightness and magnetic maps ofV830 Tau, as well as that obtained from the magnetic mapsof V819 Tau, all agree together within the error bars. Ourresults also agree with the few existing estimates of differ-ential rotation at the surfaces of similar low-mass wTTSs(Skelly et al. 2010; Donati et al. 2014).As for LkCa 4, we finally exploited the results of ourtomographic imaging of V819 Tau and V830 Tau to modeland filter-out the activity jitter in the RV curves of bothstars. For V819 Tau, we find that the activity jitter is fil-tered down to a rms RV precision of 0.033 km s − , furtherdemonstrating that our technique works well and performsas expected from our initial set of simulations (Donati et al.2014). This leaves little hope to find potential hJs aroundV819 Tau using our technique - despite its exhibiting traces of dust, possibly indicating planet formation, which made itan obvious target for our programme. For V830 Tau how-ever, the RV residuals after filtering exhibit a significantlylarger rms dispersion of 0.104 km s − caused by variability ontimescales of a few days; we find that this dispersion excessis unlikely to relate to intrinsic (e.g., flaring-like) activity(since RV residuals do not correlate with activity proxies)nor to instrumental effects (that would impact data fromboth stars). One option is that the observed excess disper-sion is caused by a ≃ M Jup giant planet orbiting V830 Tauin . ± . d at a distance of 0.065 AU from the central star.Although our detection of a residual RV signal is in prin-ciple reliable with a high confidence level (given the excessRV dispersion it generates), we nevertheless stress that itrelies on a small number of points, preventing us to firmlyassess its periodic nature; as a result, we cannot exclude theoption that it reflects yet unclear systematics in our datarather than the true presence of a giant planet.Additional data on V830 Tau with a denser sampling ofboth rotational and orbital cycles are required to obtain adefinite and independent validation of this first promising re-sult. If confirmed, our result would suggest that hJs are likelymuch more frequent around wTTSs than around MS stars,to ensure that one is detected in a randomly selected sam-ple of only a few stars. The ongoing MaTYSSE observationswill enable to further clarify this key issue for our under-standing of the formation of planetary systems. Thanks toits much higher sensitivity to wTTSs and enhanced RV pre-cision, SPIRou, the nIR spectropolarimeter / high-precisionvelocimeter currently in construction for CFHT (first lightplanned in 2017), will tremendously boost this research pro-gramme in the near future. ACKNOWLEDGEMENTS
This paper is based on observations obtained at the Canada-France-Hawaii Telescope (CFHT), operated by the NationalResearch Council of Canada, the Institut National des Sci-ences de l’Univers of the Centre National de la RechercheScientifique (INSU/CNRS) of France and the University ofHawaii. We thank the CFHT QSO team for the its greatwork and effort at collecting the high-quality MaTYSSEdata presented in this paper. MaTYSSE is an internationalcollaborative research programme involving experts frommore than 10 different countries (France, Canada, Brazil,Taiwan, UK, Russia, Chile, USA, Switzerland, Portugal,China and Italy). We also warmly thank the IDEX initia-tive at Universit´e F´ed´erale Toulouse Midi-Pyr´en´ees (UFT-MiP) for funding the STEPS collaboration program be-tween IRAP/OMP and ESO and for allocating a “Chaired’Attractivit´e” to GAJH allowing her regularly visitingToulouse to work on MaTYSSE data. We acknowledge fund-ing from the LabEx OSUG@2020 that allowed purchas-ing the ProLine PL230 CCD imaging system installed onthe 1.25-m telescope at CrAO. SGG acknowledges supportfrom the Science & Technology Facilities Council (STFC)via an Ernest Rutherford Fellowship [ST/J003255/1]. SHPAacknowledges financial support from CNPq, CAPES andFapemig. AAV acknowledges support from the Swiss Na-tional Science Foundation (SNSF) via the allocation of anAmbizione Followship.
MNRAS , 1–15 (2015) agnetometry & velocimetry of the wTTSs V819 Tau & V830 Tau REFERENCES
Aarnio A. N., Matt S. P., Stassun K. G., 2012, ApJ, 760, 9Andr´e P., Basu S., Inutsuka S., 2009, The formation and evolutionof prestellar cores. Cambridge University Press, p. 254Baraffe I., Chabrier G., Gallardo J., 2009, ApJ, 702, L27Baraffe I., Homeier D., Allard F., Chabrier G., 2015, A&A,577, A42Baruteau C., et al., 2014, Protostars and Planets VI, pp 667–689Bouvier J., Alencar S. H. P., Harries T. J., Johns-Krull C. M.,Romanova M. M., 2007, in Reipurth B., Jewitt D., Keil K.,eds, Protostars and Planets V. pp 479–494Brown S. F., Donati J.-F., Rees D. E., Semel M., 1991, A&A,250, 463Cieza L. A., et al., 2013, ApJ, 762, 100Cranmer S. R., 2009, ApJ, 706, 824Cranmer S. R., Saar S. H., 2011, ApJ, 741, 54Davies C. L., Gregory S. G., Greaves J. S., 2014, MNRAS,444, 1157Donati J.-F., 2001, in Boffin H. M. J., Steeghs D., Cuypers J.,eds, Lecture Notes in Physics, Berlin Springer Verlag Vol. 573,Astrotomography, Indirect Imaging Methods in ObservationalAstronomy. p. 207Donati J.-F., 2003, in Trujillo-Bueno J., Sanchez Almeida J., eds,Astronomical Society of the Pacific Conference Series Vol. 307,Astronomical Society of the Pacific Conference Series. p. 41Donati J.-F., Brown S. F., 1997, A&A, 326, 1135Donati J.-F., Semel M., Carter B. D., Rees D. E., Collier CameronA., 1997, MNRAS, 291, 658Donati J.-F., Mengel M., Carter B. D., Marsden S., CollierCameron A., Wichmann R., 2000, MNRAS, 316, 699Donati J.-F., et al., 2003a, MNRAS, 345, 1145Donati J.-F., Collier Cameron A., Petit P., 2003b, MNRAS, 345,1187Donati J.-F., Forveille T., Cameron A. C., Barnes J. R., DelfosseX., Jardine M. M., Valenti J. A., 2006a, Science, 311, 633Donati J.-F., et al., 2006b, MNRAS, 370, 629Donati J.-F., et al., 2007, MNRAS, 380, 1297Donati J.-F., et al., 2008, MNRAS, 385, 1179Donati J., et al., 2010, MNRAS, 409, 1347Donati J., et al., 2011, MNRAS, 412, 2454Donati J.-F., et al., 2012, MNRAS, 425, 2948Donati J.-F., et al., 2013, MNRAS, 436, 881Donati J.-F., et al., 2014, MNRAS, 444, 3220Frank A., et al., 2014, Protostars and Planets VI, pp 451–474Furlan E., Forrest W. J., Sargent B. A., Manoj P., Kim K. H.,Watson D. M., 2009, ApJ, 706, 1194Grankin K. N., 2013, Astronomy Letters, 39, 251Grankin K. N., Bouvier J., Herbst W., Melnikov S. Y., 2008,A&A, 479, 827Gregory S. G., Donati J.-F., 2011, Astronomische Nachrichten,332, 1027Gregory S. G., Donati J.-F., Morin J., Hussain G. A. J., MayneN. J., Hillenbrand L. A., Jardine M., 2012, ApJ, 755, 97Herczeg G. J., Hillenbrand L. A., 2014, ApJ, 786, 97Hussain G. A. J., et al., 2009, MNRAS, 398, 189Ingleby L., Calvet N., Herczeg G., Brice˜no C., 2012, ApJ, 752, L20Jardine M., Collier Cameron A., Donati J.-F., 2002, MNRAS,333, 339Jardine M., Vidotto A. A., van Ballegooijen A., Donati J.-F.,Morin J., Fares R., Gombosi T. I., 2013, MNRAS, 431, 528Johns-Krull C. M., 2007, ApJ, 664, 975Johns-Krull C. M., Valenti J. A., Koresko C., 1999, ApJ, 516, 900Kenyon S. J., Hartmann L., 1995, ApJS, 101, 117Kraus A. L., Ireland M. J., Martinache F., Hillenbrand L. A.,2011, ApJ, 731, 8Kraus A. L., Ireland M. J., Hillenbrand L. A., Martinache F.,2012, ApJ, 745, 19 Kurucz R., 1993, CDROM arXiv:1503.00180 )Schiavon R. P., Batalha C., Barbuy B., 1995, A&A, 301, 840Sestito P., Palla F., Randich S., 2008, A&A, 487, 965Siess L., Dufour E., Forestini M., 2000, A&A, 358, 593Skelly M. B., Donati J.-F., Bouvier J., Grankin K. N., UnruhY. C., Artemenko S. A., Petrov P., 2010, MNRAS, 403, 159Torres R. M., Loinard L., Mioduszewski A. J., Boden A. F.,Franco-Hern´andez R., Vlemmings W. H. T., Rodr´ıguez L. F.,2012, ApJ, 747, 18Valenti J. A., Fischer D. A., 2005, ApJS, 159, 141Vidotto A. A., Opher M., Jatenco-Pereira V., Gombosi T. I., 2009,ApJ, 703, 1734Vidotto A. A., Opher M., Jatenco-Pereira V., Gombosi T. I., 2010,ApJ, 720, 1262Vidotto A. A., Jardine M., Morin J., Donati J. F., Opher M.,Gombosi T. I., 2014, MNRAS, 438, 1162Williams J. P., Cieza L. A., 2011, ARA&A, 49, 67Xiao H. Y., Covey K. R., Rebull L., Charbonneau D., MandushevG., O’Donovan F., Slesnick C., Lloyd J. P., 2012, ApJS, 202, 7This paper has been typeset from a TEX/L A TEX file prepared bythe author.MNRAS000