NEW EXTINCTION AND MASS ESTIMATES FROM OPTICAL PHOTOMETRY OF THE VERY LOW MASS BROWN DWARF COMPANION CT CHAMAELEONTIS B WITH THE MAGELLAN AO SYSTEM^*

As more and more brown dwarfs and planetary companions are being discovered, characterizing them in the visible regime yields a more complete picture of the spectral energy distribution (SED) and more insight into physical properties as well as formation scenarios. For instance, a better estimate of extinction helps to derive bolometric luminosity and mass—especially for young objects (⩽10 Myr) which may still have primeval dust and gas around them and suffer significant obscuration. However, extinction is problematic to measure because most of the high-contrast adaptive optics (AO) observations are done in the near-infrared, which is ∼10 times less sensitive to dust at K versus V. One simple treatment is to assume that the companion has the same amount of extinction as its host star (Patience et al. 2012), since in the early stages of star formation the binary might be embedded in a common envelope. For more evolved, fragmented systems both components may have their own disks, so this assumption might be invalid. Ideally one would like to acquire visible spectra or at least broad-band visible photometry to supplement near-IR measurements because visible wavelengths are a better probe for dust extinction. Yet high contrast optical observations on companions are very rare due to decreased contrast (Males et al. 2014) and the difficulty of correcting atmospheric turbulence at visible (defined here as λ < 1 μm) wavelengths. We therefore need an advanced AO system which can work in the visible to suppress the halo.

Here we present the first optical AO photometry of the CT Chamaeleontis system with the Magellan adaptive optics (MagAO) system, a powerful new 585-element AO system commissioned on the 6.5 m Clay Telescope (Close et al. 2013, 2014; Follette et al. 2013; Wu et al. 2013; Males et al. 2014). CT Cha A, a K7 classical T Tauri star (Weintraub 1990; Gauvin & Strom 1992), is located in the Chamaeleon I star-forming region. This region is close (∼160 pc; Whittet et al. 1997; Bertout et al. 1999; Luhman 2008) and is as young (median age ∼2 Myr; Luhman 2004) as the Taurus star-forming region and IC 348. It also has relatively low extinction (typical A_V ≲ 5 mag; Cambrésy et al. 1997), enabling a clear view of young stars. The companion CT Cha B at 267 (430 AU) projected separation was discovered by Schmidt et al. (2008) in their Very Large Telescope (VLT) NACO survey. Based on its near-IR spectrum, the companion was estimated to be an M8-to-L0 (T_eff  ∼  2600 K) low mass (∼17 M_J) brown dwarf with A_V ∼ 5.2 mag. Schmidt et al. (2008) also imaged another closer object termed "cc2," whose true nature has remained puzzling (Schmidt et al. 2009; Robberto et al. 2012). In this paper we present new visible AO observations providing a better measurement of A_V and of the mass of CT Cha B. Our accurate astrometry allows us to determine that cc2 is, in fact, a background source.

MagAO observations with its VisAO camera (Close et al. 2013; Males et al. 2014) at r' (0.62 μm), i' (0.77 μm), z' (0.91 μm), and Y_S (0.98 μm) were performed on 2013 April 6 (UT) during the second commissioning run. Seeing was stable, ranging from 06 to 08. We locked the AO system on CT Cha A (R ∼ 12 mag) at 100 modes and 625 Hz.⁷ The achieved FWHMs were 01, 008, 006, 006 for r', i', z', Y_S, respectively. Strehl ratios were low due to only correcting 100 modes, since the guide star was somewhat faint for VisAO. We obtained saturated images to boost signal-to-noise ratio (S/N), with unsaturated data sets for relative photometry (top row in Figure 1). As a young accreting star, CT Cha A varies its brightness by ∼0.3 mag in the near-IR and ∼1 mag in the optical (Batalha et al. 1998; Ghez et al. 1997; Lawson et al. 1996). In order to calibrate the brightness of CT Cha A, we also obtained absolute photometry by observing the optical photometric standard star LTT 3864.

Zoom In Zoom Out Reset image size

We carried out standard data reduction with IRAF tasks. After dark-subtraction, flat-field correction, and cross-correlation between frames, we rotated counterclockwise the saturated data by 8911+parallactic angle to make north up and east left (middle row in Figure 1). This was following calibration of the VisAO camera based on astrometry of the Trapezium cluster (Close et al. 2013). We further rotated each of these images by 20°, 40°, ... , 340°  and took the median of them to approximate the halo of the primary. Then we subtracted the halo from the original images to further bring out any faint point source object (bottom row in Figure 1) without any loss of flux from self-subtraction. Anisoplanatic effects are still small at this small separation (<3''). Only our bluest filter (r') showed some sign of anisoplanatism, so we smoothed the reduced r' image with a Gaussian (width =2 × PSF FWHM) to enhance S/N. Next we constructed a master PSF from unsaturated images of A, used it to fit the CT Cha B profile in the deep images, and measured the PSF fitting flux with the DAOPHOT allstar task. We note here that the on-axis CT Cha A PSF was still an excellent fit to the 27  off-axis PSF of B. Table 1 summarizes our observations and PSF fitting photometry on CT Cha B. Uncertainties of B are <0.1 mag for i', z', and Y_S, and ∼0.2 mag for r' due to low S/N. Near-IR (J, H, K_S) photometry and the ∼0.3 mag uncertainty were adopted from Schmidt et al. (2008).

Table 1. MagAO Photometry on CT Cha B

Filter	tsat	tunsat	m	Δma	Fλb
	(s$\times \#$)	(s$\times \#$)	(mag)	(mag)	(10−13 erg s−1 cm−2 μm−1)
r'	20 × 75	2.27 × 13	21.71 ± 0.21	9.80	0.51 ± 0.10
i'	30 × 52	2.27 × 14	20.32 ± 0.09	9.22	1.00 ± 0.09
z'	15 × 13	2.27 × 13	18.46 ± 0.08	7.96	3.46 ± 0.26
YS	...	30 × 35	17.94 ± 0.09	7.59	4.55 ± 0.41

Notes. ^aRelative to CT Cha A, measured with PSF fitting photometry. ^bCalibrated with the standard LTT 3864 taken at a similar air mass to CT Cha in photometric conditions (errors are <1% due to different air masses from model atmosphere). We used photometry calibrations of Close et al. (2013) and Males et al. (2014).

Download table as:  ASCIITypeset image

To increase the accuracy of our astrometry, we also corrected for image distortion (≲ 15 mas). The exact formulae to correct any residual distortions for separation (δx, δy) from (X, Y) are listed in Close et al. (2013) and reproduced here: true_δx = measured_δx − δdx × |measured_δx|/110.0 and true_δy = measured_δy − δdy × |measured_δy|/44.5, where δdx = −0.00038921676 × (X − 512) + 0.00084322443 × (Y − 512) and δdy = −0.00025760395 × (X − 512) − 0.0024045175 × (Y − 512). We also retrieved archive Hubble Space Telescope (HST) and VLT NACO data for proper motion analysis. The HST data were already reduced by the OPUS pipeline and the MultiDrizzle software, as described in Robberto et al. (2012). We reduced NACO raw frames by shift-and-add, without flat-field correction and dark-subtraction. NACO image distortion is very small and only up to 3 mas at field edges (Neuhäuser et al. 2008). The error budget of our measurements includes platescale and centroid uncertainties. Table 2 lists our astrometric measurements.

3.1. Optical Images

3.2. Astrometry

The nature of cc2 is not fully settled in the literature. Schmidt et al. (2009) presented two-year astrometry, showing that it is likely to be a background star. But Robberto et al. (2012) suggested that cc2 may be physically associated based on their single epoch HST observations. We measured the positions of cc2 and CT Cha B in images taken by various instruments over ∼7 yr time span (Figure 2). Significant common-proper motion has been found for CT Cha B, confirming it is physically bound. However, we detected a significant non-common ∼15.4 mas yr⁻¹ northwestward motion for cc2, unambiguously demonstrating that it is not a co-moving companion but instead a background star, and not a member of Chamaeleon I.

Zoom In Zoom Out Reset image size

3.3. SED Fitting and Derived Properties

3.3.1. Effective Temperature

To further narrow down the uncertainty of T_eff, we retrieved the spectrum taken with the VLT SINFONI spectrograph (Schmidt et al. 2008), and calculated the H₂O-K2 index following the prescription of Rojas-Ayala et al. (2012). Assuming solar metallicity, we found that for CT Cha B this index is almost independent of extinction, ranging from 0.65 to 0.66 for A_V = 0 to 5.5 mag. In Figure 3, we plotted the variation of the index with a range of surface gravity for 2000–2800 K BT-Settl atmospheric models (Allard et al. 2011). Within this temperature range, the H₂O-K2 index is rather insensitive to log g. Hence we are free to use it for young cool objects like CT Cha B. Our best fit to the index corresponds to a spectral type M9 ± 1 with T_eff = 2500 ± 100 K.

Zoom In Zoom Out Reset image size

3.3.2. Extinction, Bolometric Luminosity, and Mass

CT Cha B was previously estimated from near-IR spectroscopy to have an extinction higher than its host star (A_V  = 5.2 mag versus 1.3 mag; Schmidt et al. 2008). Our data benefit from the fact that visible wavelengths are more sensitive to dust extinction, so we can determine A_V at higher precision with MagAO's VisAO camera.

We applied multiple values of A_V following the extinction law in Fitzpatrick (1999) to redden the 2500 K BT-Settl synthetic spectra normalized at K_S (Figure 4). Minimization of the reduced χ² is based on the reddened models fit to the observed i', z', Y_S, J, and H photometry (black points in Figure 5). We found that while the result is independent of surface gravity, $\chi ^2_r$ remains high even after including the ±0.3 mag uncertainty at K_S. Some systematic errors may come into play. For example, the adopted extinction law might be invalid due to grain growth in the disk, or there could be multiple dust components. On the other hand, scattered light from the disk or outflow gas may contribute to our i' photometry, as in the case of R Mon (Close et al. 1997). In this picture, blue light follows indirect paths to the observer, avoiding passing through the disk and making our extinction estimate likely a lower limit. Another possible cause for higher $\chi ^2_r$ is that an overall offset ∼0.5 mag might exist between the visible and near-IR data because they were taken on different nights and CT Cha A is a well-known variable. Finally, the companion itself could also be variable in the visible just like the primary due to accretion. In any case, with no prior knowledge of the material around CT Cha B, our current data yield a best fit to a lower extinction A_V = 3.4 ± 1.1 mag. We plotted the reddened synthetic spectrum together with the observed photometry in Figure 5.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

We followed the approach in Hillenbrand (1997) to calculate the bolometric luminosity. We converted our i' photometry to Cousins I_C, de-reddened it by A_V = 3.4 mag, applied the bolometric correction from Tinney et al. (1993) and Bessell (1995), and obtained log(L_bol/L_☉) = −2.68 ± 0.25. As a comparison, we also calculated L_bol from the K flux following Close et al. (2007) and had a similar value log(L_bol/L_☉) = −2.71 ± 0.20. Both values are consistent with log(L_bol/L_☉) = −2.68 ± 0.21 in Schmidt et al. (2008). We also calculated CT Cha B's radius using L∝R²T⁴ and obtained ∼ 2.4 R_J. Then we applied the DUSTY evolutionary tracks (Chabrier et al. 2000; Baraffe et al. 2001) to derive a mass estimate of ∼14–24 M_J based on the ∼1–5 Myr age and T_eff (Figure 6). Therefore, CT Cha B is most likely a very low mass brown dwarf, just above the planetary mass limit.

Zoom In Zoom Out Reset image size

3.3.3. Accretion Rate

Pa-β emission, an accretion signature, has been seen in CT Cha B's J-band spectrum (Schmidt et al. 2008). Since CT Cha B is widely separated from the host star, it may harbor its own disk and still be actively accreting at this time. Figure 5 shows that our r' detection is about 20 times brighter than its predicted continuum. This significant r' excess seems to imply strong Hα emission from accretion, allowing us to calculate the mass accretion rate. Attributing >95% of the r' flux to Hα and following the prescription of Close et al. (2014), we estimated $\dot{M}\sim 6\times 10^{-10}\,{M_\odot}\rm \,yr^{-1}$, which is reasonable as it implies that a few M_J of brown dwarf mass could be accreted in a few million years at the end of the gas-rich disk phase. The accretion rate we derived is also consistent with recent HST observations by Zhou et al. (2014), who measured $\dot{M}\sim 10^{-11}\hbox{--}10^{-9}\, {M_\odot}\rm \,yr^{-1}$ for three substellar companions GSC 6214-210 B, GQ Lup B, and DH Tau b based on their optical excess.

3.4. Implications

MagAO observations on CT Cha at r', i', z', and Y_S have improved the accuracy of the extinction toward CT Cha B. The companion is detected in all of our optical filters, whereas no detections were made by HST. It is over-luminous at r', indicating active accretion at a rate of $\dot{M}\sim 6\times 10^{-10}\,{M_\odot}\rm \,yr^{-1}$. The H₂O-K2 index derived from the K_S spectrum is consistent with a T_eff = 2400–2600 K brown dwarf. Using the BT-Settl model, we show that CT Cha B is best fit by A_V = 3.4  ±  1.1 mag, which is lower than previous estimates and translates to a mass estimate of 14–24 M_J based on the DUSTY tracks. We do not see the faint southern [O i] source seen in previous HST observations, so it is unlikely to be real. Finally, our astrometry on cc2 is incompatible with a previous claim that it is a co-moving object.

We thank the anonymous referee for helpful comments that greatly improved this manuscript. We thank Professor Jennifer Patience for kindly providing the SINFONI spectra of CT Cha B. We thank the whole Magellan Staff for making this wonderful telescope possible. We would especially like to thank Povilas Palunas (for help over the entire MagAO commissioning run). Juan Gallardo, Patricio Jones, Emilio Cerda, Felipe Sanchez, Gabriel Martin, Maurico Navarrete, Jorge Bravo and the whole team of technical experts helped do many exacting tasks in a very professional manner. Glenn Eychaner, David Osip and Frank Perez all gave expert support which was fantastic. It is a privilege to be able to commission an AO system on such a fine telescope and site. The MagAO system was developed with support from the NSF MRI and TSIP programs. The VisAO camera was developed with help from the NSF ATI program. Y.L.W.'s and L.M.C.'s research was supported by NSF AAG and NASA Origins of Solar Systems grants. J.R.M. is grateful for the generous support of the Phoenix ARCS Foundation. J.R.M. and K.M. were supported under contract with the California Institute of Technology funded by NASA through the Sagan Fellowship Program.