Magnetic activity and accretion on FU Tau A: Clues from variability
Aleks Scholz, Beate Stelzer, Grainne Costigan, David Barrado, Jochen Eislöffel, Jorge Lillo-Box, Pablo Riviere-Marichalar, Hristo Stoev
aa r X i v : . [ a s t r o - ph . S R ] O c t Mon. Not. R. Astron. Soc. , 1–10 (2002) Printed 20 June 2018 (MN L A TEX style file v2.2)
Magnetic activity and accretion on FU Tau A: Clues fromvariability
Alexander Scholz ⋆ , Beate Stelzer , Grainne Costigan , David Barrado , ,Jochen Eisl¨offel , Jorge Lillo-Box , Pablo Riviere-Marichalar , Hristo Stoev School of Cosmic Physics, Dublin Institute for Advanced Studies, 31 Fitzwilliam Place, Dublin 2, Ireland INAF - Osservatorio Astronomico di Palermo, Piazza del Parlamento 1, I-90134 Palermo, Italy Centro de Astrobiologia Depto. Astrofisica (INTA-CSIC), ESAC campus, P.O. Box 78, E-28691 Villanueva de la C˜anada, Spain Calar Alto Observatory, Centro Astron´omico Hispano Alem´an, Almer´ıa, Spain Th¨uringer Landessternwarte Tautenburg, Sternwarte 5, D-07778 Tautenburg, Germany
Accepted. Received.
ABSTRACT
FU Tau A is a young very low mass object in the Taurus star forming region whichwas previously found to have strong X-ray emission and to be anomalously bright forits spectral type. In this study we discuss these characteristics using new informationfrom quasi-simultaneous photometric and spectroscopic monitoring. From photometrictime series obtained with the 2.2 m telescope on Calar Alto we measure a periodof ∼ Key words: stars: low-mass, brown dwarfs; stars: rotation; stars: activity; accretion,accretion discs
FU Tau has been known for several decades as a variable starembedded in the dark cloud B215 (e.g. Gotz 1961; Romano1975) and as a member of the Taurus star forming regionbased on H α emission and proper motion (Haro et al. 1953;Jones & Herbig 1979). Recently the object has drawn inter-est because it turned out to be a binary, with a primarycomponent FU Tau A and the faint companion FU Tau Bat a separation of 5.7”, corresponding to a projected separa-tion of ∼
800 AU for the distance of Taurus. Luhman et al.(2009) discovered the companion and estimated spectraltypes of M7.25 and M9.25 and masses of 0.05 and 0.015 M ⊙ ⋆ E-mail: [email protected] for the two objects, indicating that FU Tau may in fact be arare wide binary brown dwarf. Because of the wide separa-tion of the two components and the location far away fromany other member of the star forming region, FU Tau is ofconsiderable interest to test formation scenarios for substel-lar objects.Apart from its binarity and location, FU Tau A turnsout to exhibit two other anomalous properties. First, theobject has strong X-ray emission compared with other ob-jects at similar spectral type (Stelzer et al. 2010). More-over, the X-ray spectrum indicates the presence of soft ra-diation, possibly from an accretion-related shockfront, ashas been observed previously for more massive objects (e.g.Stelzer & Schmitt 2004). Second, FU Tau A is anomalouslybright for objects of this spectral type in the Taurus star c (cid:13) Scholz et al. forming region. It sits about one order of magnitude abovethe 1 Myr isochrone in the HR diagram (Luhman et al. 2009;Stelzer et al. 2010). In Stelzer et al. (2010) a few possiblescenarios have been put forward to explain these features,including suppressed convection due to magnetic activity,excess flux from accretion, and early evolutionary stage.Here we set out to put further constraints on the prop-erties of FU Tau A by analysing its photometric and spec-troscopic variability. This paper is mainly based on photo-metric time series obtained with the instruments CAFOSand BUSCA at the 2.2 m telescopes of the German-SpanishAstronomical Center at Calar Alto observatory. In Sect. 2we discuss these observations and the reduction of the data.The analysis of the photometry and spectroscopy is pre-sented in Sect. 3 and 4. In Sect. 5 we compile all availableinformation on the variability of the system from our newobservations, the literature, and archives and constrain theorigin of the variations using spot models. We discuss theresults in the context of the two anomalies mentioned abovein the final Sect. 6.
Our primary photometric time series was obtained withCAFOS at the 2.2 m telescope on Calar Alto over five nightsin Nov/Dec 2010. CAFOS is a 2 ×
2k CCD camera mountedin the RC focus. With a pixel scale of 0 . ′′
53, it gives a field ofview (FOV) of 16 ′ × ′ . The filters, however, do not coverthe full FOV; in effect a circular FOV with diameter of ∼ ′ can be used.While the whole run was affected by dodgy weatherconditions, including bad seeing, high humidity, and clouds,we obtained 62 useful images in the R-band and the samenumber in the I-band for our target. The final night of therun we observed Landolt standard stars under photometricconditions for calibraton purposes (2x field SA92, 3x fieldSA98). The observing log for the run is given in Table 1.FU Tau is located in the middle of a dark cloud devoidof stars. Since we need non-variable field stars to calibratethe lightcurves, our FOV was not centered on FU Tau it-self. Instead, we positioned FU Tau in the south-west (SW)corner of the CCD, which allows us to cover a sufficient num-ber of field stars in the area immediately north-east (NE) ofthe cloud. To minimize flatfield problems, we aimed to keepthe position of the time series field as constant as possible;the offsets between the images are < Table 1.
Time series observations with CAFOS and BUSCA.In the 2nd column, the ’C’ stands for CAFOS and the ’B’ forBUSCA. The values for exposure times and seeing are typical fora particular night.Date bands no. exp time seeing2010-11-28 C/R,I 6 250, 100 2”2010-11-30 C/R,I 21 450, 120 3-4”2010-12-01 C/R,I 8 300, 100 2”2010-12-02 C/R,I 27 300, 80 2”2010-12-08 B/I 2 120 2 . ′′ . ′′ . ′′ . ′′ . ′′ set. Since the pattern has a spatial scale of >
50” and thespatial offsets in the time series are <
10” we do not thinkthat the pattern has an effect on the lightcurves. In addi-tion, the I-band frames show a faint small-scale interferencepattern due to nightsky emission lines, which contributes tothe noise.
As part of the same run, we obtained 5 low-resolution spec-tra for FU Tau A, using grism R400 with a nominal resolu-tion of 10 ˚A.The spectra for FU Tau A were debiased andbackground-subtracted by fitting a 2nd order polynomialto each line in the spatial direction. They were extracted,dispersion-corrected and flux-calibrated using standard rou-tines in IRAF.
Complementary time series photometry in the I-band wasobtained using BUSCA at the 2.2 m telescope on Calar Alto.BUSCA allows to take images in four bands simultaneously,achieved through 3 dichroic beam splitters. Our target, how-ever, was not detected in the three blue channels (Stromgrenvby filters); we only use the images in the reddest channel,which corresponds to the Cousins I-band. The observationsstarted about a week after the CAFOS run and continuedfor another week. Similar to the CAFOS run, parts of theobservations were affected by clouds, strong winds, and highhumidity. No photometric calibration was carried out.An 11 ′ × ′ FOV centered on FU Tau was observedin 7 nights in Dec 2010, of which 5 provided usable data(see Table 1). The FOV covers the bright K2III star 2MASSJ04232455+2500084 and 5-10 point sources 1-2 mag fainterthan FU Tau A. For all images, we carried out a standardreduction including bias subtraction and flatfield correction.
From the CAFOS time series we derived R- and I-bandlightcurves for FU Tau A. We hand-picked a sample of 48 c (cid:13) , 1–10 U Tau: Clues from variability Figure 2.
I-band lightcurve for FU Tau A derived from theBUSCA images. (I) and 45 (R) sources, including FU Tau A and all otherisolated stars with similar brightness in the FOV. For theseobjects we carried out aperture photometry using a constantaperture of 10 pixel and a sky annulus of 10-20 pixel. Due tothe poor seeing, the companion FU Tau B is not detectedin most of the CAFOS images; no photometry was possiblefor this object.To correct for the effects of variable seeing and trans-parency (’relative calibration’), we calculated the averagetime series of non-variable stars in the field and subtractedit from all lightcurves. The non-variable stars were chosenusing the procedure outlined in Scholz & Eisl¨offel (2004).The routine selected 18 (R) and 10 (I) stars as non-variable,based on a comparison of their lightcurve with the aver-age lightcurve of all other stars. The average RMS of thelightcurves for these non-variable stars is 0.010 (R) and0.012 mag (I), which defines the photometric accuracy.From the BUSCA images we obtained I-bandlightcurves for FU Tau A and all other stars in the field,again using aperture photometry with the same parametersas for CAFOS. The bright star 2MASS J04232455+2500084clearly looks variable, but 7 faint field stars show stablelightcurves. Their average lightcurve is used for the rela-tive calibration. After subtraction of the average lightcurve,the RMS for the 7 field stars is 0.011-0.025, an averageof 0.018 mag, confirming that they are non-variable. Forcomparison, the RMS for 2MASS J04232455+2500084 is0.13 mag.
The lightcurves from CAFOS and BUSCA show that FUTau A is a variable star. Its RMS is 0.04 (R-band, CAFOS),0.02 (CAFOS, I-band), and 0.04 (BUSCA, I-band), signifi-cantly more than comparison stars (0.01 mag for CAFOS,0.02 mag for BUSCA, Sect. 3.1). The lightcurves fromCAFOS and BUSCA are shown in Figs. 1 and 2. Most ofthe variability is on timescales of > − (calculated followingHorne & Baliunas (1986)). In the Scargle periodogram, how-ever, the peak is very broad and does not permit an accurateassessment of the period. Finally, we compare the RMS inthe original lightcurve with the RMS after subtracting a sinefunction with the suspected period using the F-test. Again,the period of 3.7-4.0 d is highly significant in both bands,with false alarm probabilities below 10 − . In Fig. 3 we showthe phase-folded lightcurve assuming P = 3 . ± . From the Landolt standard fields observed in the last nightof the CAFOS run we derived a photometric calibration. Intotal, we observed 45 standards from which 40 gave use-ful photometry. These stars cover a wide range in airmassfrom X = 1 . R are well reproduced with a zeropoint shift andan extinction term: R = r − . − . X . The RMS forthis transformation is 0.03, dominated by the uncertaintyin the zeroterm. For the I-band, it turns out that an ad-ditional colour term improves the RMS from 0.1 to 0.05: I = i − . − . X + 0 . r − i ). (In these equations,the lower case letters are instrumental magnitudes and up-per case letters calibrated magnitudes.)Applying this transformation to the instrumental mag-nitudes measured for FU Tau A gives R = 15 .
39 and I = 13 .
73 mag for the night 2010-12-02. For this night thelightcurve for FU Tau A indicates a photometric uncertaintyof ∼ .
02 mag (see Sect. 3.2). Adding this in quadrature tothe calibration errors, the total uncertainty in the absolutemagnitudes is 0.04 in the R-band and 0.05 in the I-band.Published photometry in similar bands for FU Tau A is c (cid:13) , 1–10 Scholz et al.
Figure 1.
Lightcurves for FU Tau A derived from the CAFOS images for the R- and the I-band. In addition to the target, we show thelightcurve of a similarly bright, non-variable reference star in the same field. The intra-night variability of FU Tau A is partly seen inthe reference stars as well and could be due to secondary extinction effects and not intrinsic to the source. The inter-night variability,however, is not apparent in the reference stars.
Figure 3.
Phase-folded lightcurves for FU Tau A assuming our best-fitting period of 3.8 d, determined from the CLEANed periodograms.The typical error is 0.010 mag in R and 0.012 mag in I.
Figure 4.
Phase-folded I-band lightcurves for FU Tau A showing the datapoints from CAFOS (crosses) and BUSCA (circles) for twodifferent periods. The typical error is 0.012 mag for CAFOS and 0.018 mag for BUSCA. c (cid:13) , 1–10
U Tau: Clues from variability Table 2.
Calibrated photometry for FU Tau A in bands similarto Johnson R and I. Typical errors are 0.05 mag.Date Sloan r Sloan i John R John I CommentOct 2001 16.94 CMC14 , , Evans et al. (2002) Luhman et al. (2009) Sloan magnitudes converted to Johnson using equations inJordi et al. (2006) Conflicting epoch information in Luhman et al. (2009) (29 or31/12/2002) this paper available from CFHT (Cousins I), Sloan (r, i), and the Carls-berg Meridian Catalog 14 (filter close to Sloan r). To trans-form the Sloan magnitudes to the Johnson/Cousins system,we used Equ. (2) and (8) from Jordi et al. (2006). All cali-brated photometry in the bands R and I is listed in Table2. The band transformations from Sloan to Cousins dependlinearly on R − I and are only calibrated for 0 < R − I < R − I = 2 . − . R − I < I and 0.23 in R over a timescale of 23-25 d, which is slightlymore than our amplitudes measured over 5 nights. Compar-ing our photometry with the Sloan values indicates long-term variability of 0.2 mag in I and 0.9 mag in R . FU Tau Awas much fainter and also redder in R − I in 2002 comparedwith 2010 ( R − I ∼ . .
05 magin one night, 0.09 mag over 4 nights, Sect. 3.2). The largedifference could be due to a) inconsistencies in the I-bandcalibration, b) problems with the transformation from Sloanto Cousins bands (see above), or c) a strong outburst in thatparticular night. For these reasons we cannot reliably use theCFHT I-band magnitude and disregard the datapoint in thefollowing. The CMC14 datapoint was obtained in October2001 and is consistent with the Sloan r-band value fromDecember 2002.
Figure 5.
Low-resolution spectroscopy for FU Tau A fromCAFOS. The dates and times of the observations and the equiv-alent widths for H α are indicated. The NaI absorption feature ismarked. The resolution is 10 ˚A. The five CAFOS spectra for FU Tau A, obtained in 4 dif-ferent nights in Nov/Dec 2010, are shown in Fig. 5. Thespectral shape is remarkably similar; we do not find anysignificant differences in the continuum. With our low reso-lution of 10 ˚A, the H α feature at 6563 ˚A is the only emissionfeature. The equivalent widths (EW) for this line are in therange of 135-155 ˚A and are indicated in Fig. 5. The variationsin the EW are not significantly larger than the uncertainty,which is in the range of ± −
20 ˚A. The EWs are also consis-tent with the value of 146˚A measured from a Gemini/GMOSspectrum in March 2008 (Stelzer et al. 2010).Using the PC3 index suggested by Mart´ın et al. (1999)we assigned spectral types. In chronological order, the spec-tral types are M6.82, M6.86, M6.67, M6.72, M6.66, i.e. thechange is only ± . R − I over this time series is only3% (Table 3).The H α EW of FU Tau A is well-above the usuallyadopted threshold between non-accreting ’weak-line’ andaccreting ’classical’ T Tauri stars (CTTS). According toBarrado y Navascu´es & Mart´ın (2003), all M7 objects withH α EW >
40 ˚A should be considered to be accretors. FUTau A’s status as substellar analog to a CTTS is con-firmed by the presence of mid-infrared excess (Luhman et al.2009), most likely from a circumstellar disk, and a wealthof other accretion-related emission lines in the optical andnear-infrared spectrum (Stelzer et al., in prep.).The spectra also clearly show the NaI absorption fea-tures at ∼ c (cid:13) , 1–10 Scholz et al.
Table 3.
Photometric amplitudes for FU Tau A as a function offilter and timescaleFilter λ ( µm ) ∆t A (mag)Johnson I 0.85 4 d 0 . ± . Johnson R 0.65 4 d 0 . ± . Johnson I 0.85 7 d 0 . ± . Sloan z 0.89 23 d 0 . ± . Sloan i 0.75 23 d 0 . ± . , Sloan r 0.62 23 d 0 . ± . , Sloan g 0.47 23 d 0 . ± . Sloan u 0.36 23 d 0 . ± . Johnson I 0.85 8 yr 0 . ± . , Johnson R 0.65 8 yr 0 . ± . , IRAC2 4.5 2 yr 0 . ± . IRAC4 8.0 2 yr 0 . ± . this paper Luhman et al. (2009) translates to Johnson I amplitude of 0.07 translates to Johnson R amplitude of 0.23 σ Ori cluster (Steele & Jameson 1995; Kenyon et al. 2005).Thus, our NaI measurement confirms the youth of FU TauA.
The available photometry for FU Tau A provides an accountof its variability on timescales ranging from hours to years.While our photometric time series with CAFOS and BUSCAcovers the short-term variations (a week or less) in the op-tical, the archived data in the literature from Sloan andSpitzer constrains the long-term changes in the optical andinfrared. In Table 3 we compile the variability amplitudesfor FU Tau A from a variety of sources. In the following weaim to use the characteristics of this dataset to constrain thephysical properties of FU Tau A, in particular the presenceof cool spots caused by magnetic activity and/or hot spotscaused by the accretion flow.As demonstrated in this paper, FU Tau A shows small-scale variations of ∼ . T S and filling factor f (defined as the fraction of the surface covered by thespot). For the spectrum of the unspotted photosphere, weused the AMES-DUSTY spectrum (Allard et al. 2001) for T = 3000 K and log g = 3 .
5, typical of very young objects.Note that the actual temperature of the unspotted photo-sphere is not known. Based on the average spectral type of M6.75 (Sect. 4), we would estimate 2800 K (Mentuch et al.2008), but if spots are present the unspotted photospherewill be hotter. We ran the same simulations for T = 2800and 3200 K. Since this gives only marginally different results,the exact choice of T does not seem to be relevant.The spot spectrum was approximated either by theAMES-DUSTY spectrum or with a blackbody. We calcu-lated flux ratios as a function of wavelength for T S rangingfrom 1500 to 4000 K in steps of 100 K and f ranging from0.01 to 0.3 in steps of 0.01. For the blackbody spot, we ex-tended the temperature range to 4500 K.To compare these ratios with the observations, we cal-culated the amplitudes m R and m I for the wavelengths ofthe respective filters and derived the following test quantity: χ = 1 N N X i =1 (∆ X − m X ) δX (1)Here ∆ X are the observed amplitudes, δX their errors (bothfrom Table 3), m X the amplitudes from the simulations and N is the number of filters. A good fit will result in χ < χ < χ < f ∼ .
1. Not surprisingly, the match isbetter when the spot is modeled with the AMES-DUSTYspectrum, as cool spots are expected to have a spectrumresembling a cool photosphere. Hot spots, on the other hand,do not provide a good match. This result becomes strongerwhen we consider that the CAFOS amplitudes are likelyto be somewhat smaller, as the lightcurves are affected byatmospheric effects (Sect. 3.2).The variations on timescales of 1-3 weeks are con-strained by the BUSCA lightcurve and the Sloan photome-try. Here the amplitudes tend to be somewhat larger thanon the 4 d timescale covered by CAFOS. Moreover, the am-plitude ratio between the R- and I-band in the Sloan datais a factor of 3 larger than in the CAFOS lightcurves. TheSloan photometry indicates a steep increase of the ampli-tudes towards shorter wavelengths, which is typical for hotspots, as already discussed in Stelzer et al. (2010). We ranthe same spot simulations as before for the Sloan amplitudes(2nd section in Table 3) and plot the results in Fig. 6 (lowerleft panel). This time only hot spots with f < . χ < c (cid:13) , 1–10 U Tau: Clues from variability Figure 6.
Best-fitting spot temperature/filling factor combinations from the simulations discussed in Sect. 5. All crosses show the bestfitting combinations with χ <
1, all dots with χ < the Sloan amplitudes in five bands and adopting a fillingfactor of 5% (for hot and cool spots, respectively) yieldsa good match with χ < T hot = 3500 − T cool = 2000 − T hot de-creases, and vice versa. Thus, the combination of hot andcool spots definitely fits the observed amplitudes.We conclude that the periodic modulation on atimescale of 4 d is best explained by cool spots co-rotatingwith the object. On longer timescales of weeks to years, how-ever, the optical variability is dominated by hot spots. Thecool spots are asymmetrically distributed and thus cause aperiodic, rotational signal. The hot spots, on the other hand,are axisymmetric (e.g., limited to the polar regions) and sta-ble in size/location over timescales of at least 4 d, but varyon longer timescales.The cool spots are most likely caused by suppressed con-vection due to magnetic field lines penetrating through thephotosphere, as commonly observed for magnetically activestars, including the Sun. The hot spots could be indicativeof the same phenomenon: If most of the surface is coveredby cool spots, the few remaining areas of unspotted photo-sphere would be observed as hot spots. This would, however,require implausibly large filling factors of > α emission in the spectra, which is likely to be dominated byaccretion as well (Sect. 4). The lack of variability in H α overtimescales of 4 d (EW varies by < − ∼ . µm on timescales of years (4th section in Table 3), whichare significantly larger than the photometric error. The bestexplanation for the infrared variations is changes in the diskstructure or temperature, which could be related to changesin the accretion flow. FU Tau A is anomalous in its observed properties, primarilyin two aspects:1) In comparison with Taurus objects of similar spec-tral type and temperature, FU Tau A shows strong X-rayemission, although still consistent with the large scatter (seeFig. 3 in Stelzer et al. 2010). Moreover, the X-ray emission c (cid:13) , 1–10 Scholz et al. is dominated by a soft radiation component, which may beexplained by emission from an accretion shock.2) FU Tau A is overluminous in the HR diagram, withrespect to the other known brown dwarfs of Taurus and tothe theoretical 1 Myr isochrone (see Fig. 4 in Stelzer et al.2010). This has been shown based on a luminosity calculatedfrom the J-band magnitude ( L bol /L ⊙ = 0 .
2, Luhman et al.2009). We re-determined the luminosity by comparing modelspectra with the full set of optical and near-infrared photom-etry (SDSS ugriz, 2MASS JHKs) using VOSA (Bayo et al.2008), and find L bol /L ⊙ = 0 . . . . .
21, confirming the lit-erature value.Since the overluminosity of FU Tau A is central forthe following arguments, we verify this claim by plottingFU Tau A and B in a magnitude vs. effective temperaturediagram together with other late-type members of Taurus.As a comparison sample, we use the census by Rebull et al.(2010), which comprises spectroscopically confirmed mem-bers of Taurus. Their sample of previously known memberscontains 215 objects, out of which 186 have a spectral typein the literature and 2MASS photometry. We limit ourselvesto objects with spectral type later or equal M4, A V < A V > (G¨udel et al. 2007).For these objects we determined A V from the J − K colour: A V = [( J − K ) − ( J − K ) ] / . R V = 4 .
0. We use ( J − K ) = 1 which is appropriate forthis spectral type range, as outlined in Scholz et al. (2009).After correcting the J-band magnitudes for extinction, wesubtracted a distance modulus of 5.73 mag for d = 140 pc,to yield the absolute J-band brightness M J . Spectral typeswere converted to effective temperatures using the relationby Mentuch et al. (2008). For FU Tau A and the companionB we carried out the same procedure. For FU Tau A weadopted a spectral type of M7, the average of our result(Sect. 4) and the type given by Luhman et al. (2009). Sinceno J-band magnitude is available for FU Tau B, we startedwith its K-band magnitude (Luhman et al. 2010) and added1.3 mag, which is the typical extinction corrected J − K colour for Taurus members at spectral type ∼ M9 or later.The resulting brightness-temperature diagram is shown inFig. 7.Our goal was to minimise systematic effects that couldinfluence a comparison between FU Tau A and the otherobjects. The x-axis is affected by inconsistencies in spectraltyping, which we estimate to be in the range of ± . ± . XMM-Newton Extended Survey of the Taurus molecular clouds
Figure 7.
Absolute J-band magnitude vs. effective temperaturesfor FU Tau A and B (triangles) and other confirmed members ofTaurus (dots, see text for selection criteria). Objects with disksare marked with crosses, those with X-ray detection with plusses.We overplot the BCAH (solid lines) and DUSTY (dashed lines) is-chrone for an age of 1 Myr (Baraffe et al. 1998; Allard et al. 2001).The dashdotted line is a linear fit to the DUSTY isochrone shiftedto match the positions of FU Tau A and B. disks. In Fig. 7, however, no systematic difference is seen be-tween objects with disks and those without, indicating thatthis effect is negligible.Fig. 7 clearly demonstrates that FU Tau A is signifi-cantly brighter than all other similar objects in the Taurusstar forming region with the same spectral type. The differ-ence is ∼ ∼ Strong magnetic activity could suppress the convection onthe stellar surface and produce cool spots, as it seems tobe the case in the primary of the eclipsing brown dwarf bi-nary 2MASS J05352184-0546085 (Stassun et al. 2007). Thiswould make FU Tau A appear cooler than it should be ac-cording to its mass, i.e. it would in fact be a very low massstar and not a brown dwarf. Correcting for this effect wouldshift FU Tau A roughly horizontally in the HR diagramtowards higher temperatures and closer to the bulk of dat-apoints in Fig. 7. The same explanation could theoreticallyapply to FU Tau B. c (cid:13) , 1–10 U Tau: Clues from variability A higher mass would also help to explain the high X-rayluminosity (Stelzer et al. 2010), because of the well-knowncorrelation between L X and L bol (e.g. Telleschi et al. 2007).In Fig. 7 we mark all objects with X-ray detection withplusses. The fraction of X-ray detected objects is only 2/29(7%) for T eff < < Accretion from a circumstellar disk might cause excess lu-minosity in the optical and near-infrared and shift objectsvertically in Fig. 7. As said above, accretion might also bethe best explanation for the soft X-ray emission in FU TauA. Accretion is clearly ongoing, as evidenced by the opticalspectra (this paper, Stelzer et al. in prep.). The long-termvariability in FU Tau A can only be explained by excess fluxfrom hot spots and could indicate variable accretion (Sect.5). However, since the spot temperatures are not particu-larly high ( < T S = 4000 K and f = 0 . µm would only be in the range of 30%. It needs to bestressed, however, that the variability is caused only by thechanges in the hot spots, i.e. higher excess luminosity fromconstant, axisymmetrically distributed spots is possible.The interpretion of the high luminosity in terms of ac-cretion would imply that the actual L bol is much lower, lead-ing to an unusually high fractional X-ray luminosity. If theluminosity is in fact one order of magnitude lower than givenby Luhman et al. (2009), we would get log ( L X /L bol ) = − . α emission and mid-infrared excess indicative ofthe presence of a disk (Luhman et al. 2009). Thus, accretioncould also affect its position in Fig. 7. The overluminosity of FU Tau A and B could also be causedby extreme youth. If accretion occurs mostly in bursts(’episodic accretion’), objects which are technically ’coeval’,i.e. have started their protostellar collapse at the same time,could be in a different stage of their accretion history andspread out in the HR diagram (Baraffe et al. 2009). Overlu-minous objects would be younger in terms of their accretionevolution.In Fig. 7 we overplot the BCAH and DUSTY 1 Myrisochrones (Baraffe et al. 1998; Allard et al. 2001). Thecomparison with the Taurus members shows that most ofthem fall within ± − − − M ⊙ yr − , Stelzer et al. 2010). In addition,Class 0/I protostars appear to show lower or similar X-rayluminosities than Class II sources (Prisinzano et al. 2008),e.g. extreme youth cannot explain the strong X-ray emissionof FU Tau A. Thus, an early evolutionary stage is not anentirely satisfactory explanation. In summary, FU Tau A is affected by accretion plus mag-netic activity and possibly evolutionary stage as well. Al-though the observed properties of FU Tau A are anomalous,these three factors are not limited to this particular object.All young populations show substantial spread in the HRdiagram and in the other relevant properties (e.g., X-ray lu-minosities, H α luminosities, rotation periods). Evidence foraccretion and strong magnetic activity is commonly seen inclassical T Tauri stars and brown dwarfs. FU Tau A (andits companion FU Tau B) happens to be a case which sticksout in some of its observed properties, but there is no reasonto believe that it is unusual in its physical characteristics.This could have important consequences for our under-standing of the Initial Mass Function in the low-mass regime. c (cid:13) , 1–10 Scholz et al.
Even with the wealth of photometric and spectroscopic dataavailable for FU Tau A, it is still not possible to estimate itsmass reliably. Depending on the relative magnitude of the ef-fects of suppressed convection and accretion, the mass couldbe anywhere between 0.05 and 0.2 M ⊙ , a factor of 4 uncer-tainty. In addition, the unknown evolutionary stage and ac-cretion history means that we cannot trust the isochronesfor mass estimates. Thus, barring a more complete under-standing of magnetic activity and its effect on the observableproperties as well as protostellar evolution it does not seemfeasible to derive a mass function for very young very lowmass stars and brown dwarfs. ACKNOWLEDGMENTS
We thank the anonymous referee for a constructive reportthat helped to improve the paper. This publication is basedon observations collected at the Centro Astron´omico His-pano Alem´an (CAHA) at Calar Alto, operated jointly bythe Max-Planck Institut f¨ur Astronomie and the Institutode Astrof´ısica de Andaluc´ıa (CSIC). Part of this work wasfunded by the Science Foundation Ireland through grant no.10/RFP/AST2780 to AS, and the Spanish grants AyA2010-21161-C02-02, CSD2006-00070, PRICIT-S2009/ESP-1496.In this paper we make use of VOSA, developed under theSpanish Virtual Observatory project supported from theSpanish MICINN through grant AyA2008-02156.
NOTE ADDED IN PROOF
Regarding the discussion in Sect. 3.3, it was brought to ourattention that the SDSS photometry for FU Tau A and Bwas published in the data release of the ’Low Galactic Lati-tude Fields’, which are not part of the standard SDSS datareleases (Finkbeiner et al. 2004). The correct epoch for theSDSS datapoints for FU Tau is Dec 31 (Luhman, privatecommunication).
REFERENCES
Adelman-McCarthy J. K., et al. 2008, ApJS, 175, 297Allard F., Hauschildt P. H., Alexander D. R., Tamanai A.,Schweitzer A., 2001, ApJ, 556, 357Baraffe I., Chabrier G., Allard F., Hauschildt P. H., 1998,A&A, 337, 403Baraffe I., Chabrier G., Allard F., Hauschildt P. H., 2002,A&A, 382, 563Baraffe I., Chabrier G., Gallardo J., 2009, ApJ, 702, L27Barrado y Navascu´es D., Mart´ın E. L., 2003, AJ, 126, 2997Bayo A., Rodrigo C., Barrado Y Navascu´es D., Solano E.,Guti´errez R., Morales-Calder´on M., Allard F., 2008, A&A,492, 277Bouvier J., Covino E., Kovo O., Martin E. L., MatthewsJ. M., Terranegra L., Beck S. C., 1995, A&A, 299, 89Cardelli J. A., Clayton G. C., Mathis J. S., 1989, ApJ, 345,245Chabrier G., Gallardo J., Baraffe I., 2007, A&A, 472, L17Evans D. W., Irwin M. J., Helmer L., 2002, A&A, 395, 347 Finkbeiner D. P., Padmanabhan N., Schlegel D. J., CarrM. A., Gunn J. E., Rockosi C. M., Sekiguchi M., et al.2004, AJ, 128, 2577Gotz W., 1961, Veroeffentlichungen der Sternwarte Son-neberg, 5, 87Grosso N., Briggs K. R., G¨udel M., Guieu S., Franciosini E.,Palla F., Dougados C., Monin J.-L., M´enard F., BouvierJ., Audard M., Telleschi A., 2007, A&A, 468, 391G¨udel M., Briggs K. R., Arzner K., Audard M., BouvierJ., Feigelson E. D., Franciosini E., Glauser A., Grosso N.,Micela G., Monin J.-L., et al. 2007, A&A, 468, 353Haro G., Iriarte B., Chavira E., 1953, Boletin de los Ob-servatorios Tonantzintla y Tacubaya, 1, 3Herbst W., Bailer-Jones C. A. L., Mundt R., 2001, ApJ,554, L197Herbst W., Eisl¨offel J., Mundt R., Scholz A., 2007, Proto-stars and Planets V, pp 297–311Horne J. H., Baliunas S. L., 1986, ApJ, 302, 757Jones B. F., Herbig G. H., 1979, AJ, 84, 1872Jordi K., Grebel E. K., Ammon K., 2006, A&A, 460, 339Kenyon M. J., Jeffries R. D., Naylor T., Oliveira J. M.,Maxted P. F. L., 2005, MNRAS, 356, 89Kirkpatrick J. D., Henry T. J., McCarthy Jr. D. W., 1991,ApJS, 77, 417Luhman K. L., Mamajek E. E., Allen P. R., Muench A. A.,Finkbeiner D. P., 2009, ApJ, 691, 1265Luhman K. L., Mamajek E. E., Allen P. R., Muench A. A.,Finkbeiner D. P., 2010, ApJ, 720, 1781MacDonald J., Mullan D. J., 2009, ApJ, 700, 387Mart´ın E. L., Delfosse X., Basri G., Goldman B., ForveilleT., Zapatero Osorio M. R., 1999, AJ, 118, 2466Mentuch E., Brandeker A., van Kerkwijk M. H., Jayaward-hana R., Hauschildt P. H., 2008, ApJ, 689, 1127Prisinzano L., Micela G., Flaccomio E., Stauffer J. R.,Megeath T., Rebull L., Robberto M., Smith K., Feigel-son E. D., Grosso N., Wolk S., 2008, ApJ, 677, 401Rebull L. M., Padgett D. L., McCabe C.-E., et al. 2010,ApJS, 186, 259Roberts D. H., Lehar J., Dreher J. W., 1987, AJ, 93, 968Romano G., 1975, Mem. Soc. Astron. Ital., 46, 81Scargle J. D., 1982, ApJ, 263, 835Scholz A., Eisl¨offel J., 2004, A&A, 419, 249Scholz A., Eisl¨offel J., 2005, A&A, 429, 1007Scholz A., Geers V., Jayawardhana R., Fissel L., Lee E.,Lafreniere D., Tamura M., 2009, ApJ, 702, 805Stassun K. G., Mathieu R. D., Valenti J. A., 2007, ApJ,664, 1154Steele I. A., Jameson R. F., 1995, MNRAS, 272, 630Stelzer B., Schmitt J. H. M. M., 2004, A&A, 418, 687Stelzer B., Scholz A., Argiroffi C., Micela G., 2010, MN-RAS, 408, 1095Telleschi A., G¨udel M., Briggs K. R., Audard M., Palla F.,2007, A&A, 468, 425 c (cid:13)000