DISCOVERY OF A PLANETARY-MASS COMPANION TO A BROWN DWARF IN TAURUS^*

We selected 2M J044144 as one of the targets for a survey of late-type members of Taurus with the Wide Field Planetary Camera (WFPC2) on board the Hubble Space Telescope. The targets were imaged through the F791W and F850LP filters with the PC camera within WFPC2, which contained an 800 × 800 CCD (0046 pixel⁻¹). In each filter, we obtained two images at each position in a two-point diagonal dither pattern. For 2M J044144, the exposure times of the individual images were 260 and 160 s for F791W and F850LP, respectively. The observations of 2M J044144 were performed during one orbit on 2008 August 20.

We reduced the WFPC2 images using the MultiDrizzle software (Koekemoer et al. 2002). To facilitate the detection of low-mass companions, we applied point-spread function (PSF) subtraction to each brown dwarf in our survey using techniques described by Luhman et al. (2005). This process revealed a possible companion to 2M J044144, which we refer to as 2M J044144 B. The images of this pair in F850LP before and after PSF subtraction are shown in Figure 1. We fit PSFs to both the primary and candidate companion to measure their flux ratios and astrometric positions in the two filters. In addition to PSF subtraction of the primaries, we identified all point sources appearing in the WFPC2 images and measured aperture photometry for them. In the case of 2M J044144, the PSF of the candidate companion was subtracted prior to the photometric measurement of the primary. The photometry for companion was computed by combining the aperture photometry of the primary with the flux ratios produced by the PSF fits. The errors in the photometry for the primary are dominated by the uncertainty in the correction for charge transfer efficiency while the errors in the measurements for the companion have a comparable contribution from the process of PSF fitting. The astrometric offsets and photometry for the two objects are presented in Tables 1 and 2, respectively. Additional details concerning the observations and data analysis will be provided in a forthcoming paper that presents the entire survey.

Zoom In Zoom Out Reset image size

To constrain the relative proper motions of 2M J044144 and its candidate companion, we obtained images of the pair at a second epoch using NIRI (Hodapp et al. 2003) in conjunction with the ALTAIR adaptive optics system at the Gemini North Telescope. NIRI was configured with the f/32 camera, which provided a plate scale of 00214 pixel⁻¹ and a field of view of 22'' × 22''. Because natural guide stars were unavailable, ALTAIR was operated in the laser guide star mode. The tip/tilt correction was performed with 2M J044145, which is the possible wide companion to 2M J044144 that was described in Section 2. We imaged 2M J044144 through the H and K' filters using individual exposure times of 1 and 20 s. For each exposure time and filter, one image was taken at each position in a 3 × 3 dither pattern. After one set of dithered images was collected for all of the combinations of exposure time and filter, the entire sequence of observations was repeated. To provide a PSF for fitting the primary and candidate companion, we also observed 2MASS 04383110+2310107. It was selected as the PSF star because it was similar to the science target in its sky position and the optical brightness, separation, and position angle (P.A.) of its tip/tilt star.

The individual images were divided by flat field exposures taken with the Gemini Facility Calibration Unit. The resulting images in a given dither sequence were registered and combined, producing two final images of 2M J044144 for each combination of exposure time (1 and 20 s) and filter (H and K'). All of the combined images of 2M J044144 and the PSF star exhibited FWHM ∼009, which is similar to the image quality from WFPC2. To measure the relative positions and flux ratios of the components of 2M J044144, we performed PSF fitting with the four long exposures in the same manner as done for the WFPC2 data. In Figure 1, we present one of the two long K'-band exposures before and after PSF subtraction alongside the WFPC2 images. We also show a wider area of this NIRI image that encompasses the tip/tilt star, 2M J044145, which is resolved into a pair of objects that are separated by 023. We applied PSF fitting to the components of this pair as well to measure their flux ratios. This was done with the short exposures since the primary is saturated in the 20 s data.

Our measurements of the astrometric offsets between 2M J044144 A and B are listed in Table 1. We have adopted the average values produced by PSF fitting of the four long exposures. We have derived photometry for the components of 2M J044144 and 2M J044145 by combining the flux ratios from the PSF fitting with the unresolved H and K_s data from the Point Source Catalog of the Two Micron All Sky Survey (2MASS; Skrutskie et al. 2006). We have assumed that the flux ratios at K_s are equal to the ratios that we have measured at K'.

To investigate whether 2M J044144 B is a companion to 2M J044144 A, we begin by plotting a color–magnitude diagram in Figure 2 from photometry at F791W and F850LP for all unsaturated sources in our entire WFPC2 survey of 32 brown dwarfs in Taurus. Four of the substellar primaries are absent from the diagram because they are saturated in one of the filters. The faintest primary is 2MASS J04373705+2331080, which is the coolest known member of Taurus (L0; Luhman et al. 2009). As shown in Figure 2, 2M J044144 B appears near the bottom of the sequence of the known members of Taurus, indicating that it has the appropriate magnitudes and color for a low-mass companion. Given the spectral types of the known members, these data suggest a spectral type of M9.5–L0 for 2M J044144 B. In addition, the line connecting the primary and secondary is parallel to the cluster sequence, which is consistent with the coevality that is expected for the components of a binary system.

Zoom In Zoom Out Reset image size

We can use the data in Figure 2 to estimate the probability that our companion survey would detect a field star that has the photometric properties of a low-mass Taurus member. We assume that the sequence of Taurus members becomes vertical at the faintest magnitudes in Figure 2 (m₇₉₁ > 22), which is supported by the location of the coolest known member and the fact that the I − Z colors of field dwarfs do not increase significantly from M8 to mid-L (Steele & Howells 2000; Dahn et al. 2002; Dobbie et al. 2002). In that case, ∼7 objects in Figure 2 (in addition to the known brown dwarfs) have colors and magnitudes that are consistent with membership in Taurus. Therefore, the surface density of field stars that could be mistaken for Taurus members is ∼7 divided by the total area of the WFPC2 images, which is 160 arcmin². The probability of a field star of this kind appearing within 01 of any of the 32 primaries in our survey is ∼10⁻⁵.

The two epochs of astrometry from WFPC2 and NIRI provide an additional constraint on the companionship of 2M J044144 B. We first consider whether the astrometry for 2M J044144 B is consistent with the negligible motion expected for the vast majority of background stars. Accurate measurements of the proper motion for the primary are not available. Therefore, we adopt for the primary the average proper motion of the nearest group of Taurus members (μ_α, μ_δ = +6.7, −17.7 mas yr⁻¹; Luhman et al. 2009). Combining this proper motion with the astrometry from the first epoch, we find that the separation and P.A. of the candidate companion would have become 0089 and 1115 in the second epoch if it had no proper motion, which is inconsistent with our measurements in Table 1.

According to the data in Table 1, 2M J044144 A and B maintained the same relative positions to within ∼10 mas over a period of 1.15 years, indicating that their proper motions agree at a level of ∼9 mas yr⁻¹ in each direction. Only a small fraction of field stars are expected to have proper motions that are similar to that of 2M J044144 A. For instance, using multi-epoch Hubble images of Chamaeleon I that are comparable in depth to our WFPC2 data in Taurus, we find that ∼0.3% of field stars have proper motions that are consistent with those of cluster members at the level of the uncertainties in the relative proper motions of 2M J044144 A and B. If we adopt this value for Taurus, which is similar to Chamaeleon I in its distance and the magnitude of its proper motion, then the probability of a field star exhibiting WFPC2 photometry that is indicative of a Taurus member, a separation of ρ ⩽ 01 from one of our 32 primaries, and a proper motion similar to that of Taurus is ∼3 × 10⁻⁸. Therefore, we conclude that 2M J044144 B is a companion to 2M J044144 A.

The possible wide companion 2M J044145 appears within one of the Wide Field Camera arrays, but it is too heavily saturated for the detection of its candidate companion. As a result, we cannot place constraints on the proper motions of those two objects relative to each other.

We can estimate the bolometric luminosity of 2M J044144 B if we apply a bolometric correction to its photometry at K_s. In previous studies of young brown dwarfs, we have adopted bolometric corrections derived for field dwarfs, but doing so may not be perfectly valid for 2M J044144 B since some of the IR colors of young objects later than M9 differ from those of dwarfs (Cruz et al. 2009; Luhman et al. 2009). To determine an appropriate value of BC_K for 2M J044144 B, we have estimated bolometric luminosities for young brown dwarfs from M9.5–L0 for which relatively complete spectral energy distributions have been measured, consisting of the Taurus member KPNO 4 (M9.5; Briceño et al. 2002) and the young field dwarfs 2MASS J01415823–4633574 and 2MASS J02411151–0326587 (L0; Kirkpatrick et al. 2006; Cruz et al. 2009). We have excluded from this sample brown dwarfs that exhibit disk emission in their mid-IR data or that have extinctions of A_V > 1. For each object, we constructed a spectral energy distribution using a 0.8–2.5 μm spectrum from the NASA Infrared Telescope Facility that was flux calibrated with 2MASS photometry, 3.2–9.2 μm photometry from the Spitzer Space Telescope (Luhman et al. 2009, 2010), a linear interpolation of the fluxes between 2.5 μm and 3.2 μm, and a Rayleigh–Jeans distribution longward of 9.2 μm. By integrating the flux in each distribution and combining the resulting luminosity with the K_s photometry, we arrive at BC_K = 3.42, 3.39, and 3.41 for KPNO 4, 2MASS J01415823–4633574, and 2MASS J02411151–0326587, respectively. These bolometric corrections are larger than those of dwarfs by ∼0.2 mag (Golimowski et al. 2004). Based on these measurements, we adopt BC_K = 3.4 for 2M J044144 B.

The J − H color from 2MASS and the 0.8–1.8 μm spectrum from Luhman (2006) for 2M J044144 A+B indicate A_V ∼ 0 when they are compared to average intrinsic colors and spectra of young brown dwarfs (Luhman et al. 2010). In the K band, the spectroscopy and 2MASS photometry of the system and our resolved photometry for the primary are slightly redder than a stellar photosphere with the spectral type of the primary, which is probably caused by emission from a circumstellar disk based on the excess emission detected at longer wavelengths (Luhman et al. 2010). Therefore, we have not applied an extinction correction to the photometry for 2M J044144 B. When we combine K_s = 14.94, BC_K = 3.4, M_bol☉ = 4.75, and the average distance of Taurus (d = 140 pc; Wichmann et al. 1998; Torres et al. 2009), we derive log L_bol = −3.14 for 2M J044144 B.

In Figure 3, we estimate a mass from the luminosity of 2M J044144 B by comparing it to the values predicted by theoretical evolutionary models for an age of 1 Myr, which is the average age for members of Taurus and the isochronal age of the primary (Luhman et al. 2009). Given the uncertainties in its age (few Myr) and luminosity (±0.12 dex), 2M J044144 B could have a mass between 5 and 10 M_Jup according to the models of Chabrier et al. (2000) and Burrows et al. (1997). For comparison, we include in Figure 3 the luminosities of other resolved companions that appear to have masses below ∼15 M_Jup and that are younger than 10 Myr, which consist of 2M1207-3932 B (Chauvin et al. 2004), DH Tau B (Itoh et al. 2005), CHXR 73 B (Luhman et al. 2006), CT Cha B (Schmidt et al. 2008), and IRXS1609 B (Lafrenière et al. 2008).

Zoom In Zoom Out Reset image size