• NASA ADS Record
• Simbad Objects
• About Related Links

We observed Oph 98 as part of a Hubble Space Telescope (HST) multiplicity survey (GO 12944, PI Allers) targeting brown dwarfs in the Ophiuchus star-forming region. Data for this target were acquired on UT 2013 June 17. Two sets of deep dithered images were obtained in the F127M and F139M filters on the IR channel of the WFC3 instrument, in full frame MULTIACCUM mode, with total exposure times of 698 s in each band. Two images of 473 s each were then acquired in the F850LP bandpass on the UVIS channel, using the full UVIS aperture. The combination of these three filters allows for clear distinctions of substellar objects from background interlopers, by exploiting the inherent red colors of brown dwarfs and a characteristic water absorption band observed in their spectra at 1.4 μm (Fontanive et al.2018; Allers & Liu2020). Oph 98 A was indeed found to show distinctive photometric colors between these bands compared to other stars in the HST field of view. A faint companion, well resolved in all images and shown in the F127M band in Figure1 (middle panel), was detected at ∼146 from the known brown dwarf based on its multiband photometry.

Zoom InZoom OutReset image size

The pipeline processed flat-field images were used as input in theMultiDrizzle software (Fruchter & Hook2002) to correct for geometric distortion, perform cosmic-ray rejection, and combine all dithered frames into a final image in each filter. Source positions were extracted using theDAOStarFinder algorithm from thePhotutils python package (Bradley et al.2019). Aperture photometry was performed adopting 04 aperture radii. The background level and its uncertainty in each final data frame was estimated by applying the same 04 aperture to 2000 random star-free positions and computing the mean and standard deviation of these measurements. Measured fluxes were finally converted into Vega magnitudes using the HST photometric zero-points for 04 apertures in the considered filters. Measured magnitude differences (Δmag) and relative positions between Oph 98 A and B are reported in Table1. Apparent HST magnitudes are listed in Table2.

Table 2. Properties of the CFHTWIR-Oph 98 AB System

Parameter	Oph 98 A	Oph 98 B	Reference
Astrometry
α (ICRS J2000.0)			Cutri et al. (2003)
δ (ICRS J2000.0)			Cutri et al. (2003)
[mas yr−1]	−7.2 ± 2.0		Cánovas et al. (2019)
[mas yr−1]	−25.5 ± 1.7		Cánovas et al. (2019)
Parallax [mas]	7.29 ± 0.22		Ortiz-León et al. (2018)
Distance [pc]	137 ± 4		Ortiz-León et al. (2018)
Photometry
2MASSJ [mag]	16.775 ± 0.176		Cutri et al. (2003)
2MASSH [mag]	15.574 ± 0.109		Cutri et al. (2003)
2MASSKs [mag]	14.593 ± 0.098		Cutri et al. (2003)
F850LP [mag]	19.696 ± 0.018	22.479 ± 0.170	This paper
F127M [mag]	16.835 ± 0.005	19.027 ± 0.013	This paper
F139M [mag]	16.959 ± 0.005	18.991 ± 0.016	This paper
WIRCAMJ [mag]	17.015 ± 0.016	19.109 ± 0.050	This paper
WIRCAMH [mag]	15.851 ± 0.016	17.541 ± 0.030	This paper
WIRCAMKs [mag]	14.917 ± 0.009	16.498 ± 0.017	This paper
WFCAMJ [mag]	16.975 ± 0.009	19.042 ± 0.038	This paper
WFCAMH [mag]	15.826 ± 0.013	17.620 ± 0.037	This paper
WFCAMK [mag]	14.792 ± 0.039	16.408 ± 0.047	This paper
Fundamental properties
AV [mag]	5 ± 1		This paper
Spectral typea	M9–L1	L2–L6	This paper
[dex]	−2.85 ± 0.06	−3.49 ± 0.06	This paper
	1 Myr
Teff [K]	2210 ± 60	1740 ± 40	This paper
logg [dex]			This paper
Radius [RJup]	2.61 ± 0.05		This paper
Mass [MJup]	9.6 ± 1.4		This paper
	3 Myr
Teff [K]	2320 ± 40	1800 ± 40	This paper
logg [dex]			This paper
Radius [RJup]		1.86 ± 0.05	This paper
Mass [MJup]	15.4 ± 0.8		This paper
	7 Myr
Teff [K]	2370 ± 40		This paper
logg [dex]			This paper
Radius [RJup]	2.27 ± 0.07		This paper
Mass [MJup]			This paper
Binary characteristics
Separation [arcsec]	1.46 ± 0.01		This paper
Separation [au]	200 ± 6		This paper
Orbital period [yr]b			This paper
Mass ratiob			This paper
Eb [1039 erg]b	8.8 ± 1.4		This paper

Notes.

^aPhotometric estimates based on near-infraredvl-g spectral templates. ^bQuantities calculated for a median age of 3 Myr.

Download table as: ASCIITypeset images:12

Seeing-limited, broadbandJ,H, andKs (MKO filter system) images of Oph 98 are available from the Canada–France–Hawaii Telescope (CFHT) WIRCAM archive (Programs 06AF01, 06AT08, and 12AT09), with data acquired in 2006 May and 2012 August. A CFHTH-band image of Oph 98 is shown in the left panel of Figure1.

We used preprocessed images from the CFHT facility pipeline, ’I’iwi, which includes detrending and sky subtraction. We determined the relative astrometry and photometry of Oph 98 A and B from individual preprocessed frames. We used the IDL Astronomy Library’sFIND,APER, andGETPSF routines (Landsman1993) to determine the point-spread function (PSF) and residuals for nonsaturated stars within 1 arcmin of Oph 98 A. Using theNSTAR program, we fit the PSF to Oph 98 A, and subtracted it from the image usingSUBSTAR. We then fit a model PSF to Oph 98 B. For each epoch and each filter, we determined the separation, position angle, and Δmag of the binary using the mean and standard deviation of the mean of measurements from individual (nonstacked) frames. Table1 reports our measurements.

We also determined MKO photometry of Oph 98 A and B from the individual, preprocessed frames. We used the magnitudes calculated byNSTAR during the PSF fitting process, and determined the photometric calibration offset for each image by comparing the PSF-fit magnitudes of nonsaturated stars within 1 arcmin of Oph 98 A to their 2MASS photometry. We first converted their 2MASS magnitudes to the MKO system using custom color corrections derived from synthetic photometry of SpeX Spectral Library spectra reddened byA_V of 1–30 mag. We calculated the photometry for Oph 98 A and B using a weighted mean and weighted standard deviation of the mean of the magnitudes calculated from each frame. Our photometry (Table2) is in good agreement with published WIRCAM photometry of Oph 98 A from Alves de Oliveira et al. (2012).

We obtained seeing-limitedJ- andK-band images of Oph 98 using the United Kingdom Infra-Red Telescope (UKIRT) WFCAM instrument on UT 2020 August 1, followed by anH-band data set on UT 2020 September 17 (Project ID U/20A/H02). TheK-band image is shown in Figure1 (right panel). Using the same procedure described in Section2.2, we determined the photometric and astrometric measurements of Oph 98 AB from individual, preprocessed images, reported in Tables1 and2.

The relative astrometry of the binary was measured in the various imaging data sets available as detailed in Section2 (Table1). The left panel of Figure2 clearly demonstrates that Oph 98 A and B are comoving over the 14 yr time baseline of ground- and space-based data, indicating that the two components share a common proper motion. Based on their coordinates, Oph 98 A and B are most likely part of the young L1688 cloud in the Ophiuchus complex (Esplin & Luhman2020). Given the stellar density of L1688 (King et al.2012), the chance of alignment for two unrelated Ophiuchus members within 2″ is <10⁻⁴. This number is an overestimate in the case of such rare low-mass brown dwarfs at the bottom of the initial mass function (Kroupa et al.2013), and we conclude that Oph 98 A and B form a physically associated binary pair.

Zoom InZoom OutReset image size

A nearby star (Gaia DR2 6050679111185185664), ∼6″ northwest from Oph 98, was found to be in the Gaia Data Release 2 (DR2) catalog (Gaia Collaboration et al.2016,2018) with a full astrometric solution. With Gaia parallax and proper motion measurements ofϖ = 1.453 ± 0.641 mas, mas yr⁻¹, and mas yr⁻¹, this source is essentially consistent with a stationary background star. As shown in the right panel of Figure2, the positions of the star relative to Oph 98 A at each observational epoch are in excellent agreement with the expected relative displacement over time for a background star (black line). The small positional disparities are consistent with the almost negligible motion of the Gaia source and with the measurement uncertainties in the original epoch from which the background track is calculated. These results confirm that the primary has a proper motion expected for Ophiuchus, hence validating the Ophiuchus membership of the comoving Oph 98 AB system.

We adopt the parallax ofϖ = 7.29 ± 0.22 mas derived by Ortiz-León et al. (2018) for the embedded L1688 population based on Gaia DR2, in good agreement with values from Cánovas et al. (2019) and Esplin & Luhman (2020). At a corresponding distance of 137 ± 4 pc, the observed angular separation of the Oph 98 binary (146 ± 001) implies a wide projected separation of 200 ± 6 au.

In order to estimate the spectral types of Oph 98 A and B, we used the SpeX Prism Library Analysis Toolkit (SPLAT; Burgasser et al.2016) to fit the spectral energy distribution (SED) of the binary components. We gathered a library of M and L spectra from the SPLAT database, for which we obtained homogeneous near-infrared spectral types and gravity scores using the Allers & Liu (2013) classification, and retained only young sources with very low (vl-g) gravity scores.

SPLAT allows for the determination of photometric magnitudes on specific filters based on a source’s spectrum. The module can also redden a spectrum following the Cardelli et al. (1989) reddening law. We used these capabilities to determine the scaling factor and visual extinctionA_V minimizing theχ² between the synthetic photometry of the templates and our measured magnitudes for Oph 98 A and B. We added uncertainties of 0.011, 0.007, and 0.007 mag (2MASS calibration uncertainties; Cutri et al.2003) to theJHK MKO measurements tied to 2MASS, and 2% and 5% uncertainties in the HST IR and UVIS photometry,⁶ respectively.

We observed a strong degeneracy between spectral type and extinction in our results, consistent with findings by Luhman et al. (2017) for young reddened L dwarfs. Results for Oph 98 A showed a handful of fits with similarχ² values for objects with spectral types of M9 to L1, and decreasing reddening values with later type over the rangeA_V ∼ 4–6 mag. A similar effect was seen in the results of Oph 98 B, but spanning wider ranges of spectral types (late-M to mid-L) and extinctions (A_V ∼ 4–10 mag), likely due to the atypical colors and larger uncertainties on the photometry of the secondary. Assuming the same local cloud extinction for the two components, we can reduce this range to best-fit results yieldingA_V < 6 mag for the companion. This provides spectral type estimates of M9–L1 for Oph 98 A, and L2–L6 for Oph 98 B, with a visual extinction for the system ofA_V = 5 ± 1 mag, placing the primary at the M/L transition and loosely constraining the secondary to be along the L spectral sequence. Our derived quantities for the primary are in reasonable agreement with the values of M9.75 andA_V = 3 mag estimated by Alves de Oliveira et al. (2012) from low-resolutionH- andK-band spectroscopy. Additional spectroscopic observations will be required to further characterize the system and break down the observed degeneracies.

The left panel of Figure3 shows best-fit template spectra for both components compared to our photometric measurements. The right panel shows the observed positions of Oph 98 A and B (magenta stars) in the near-infrared color–magnitude diagram compared to the population of low-mass stars and brown dwarfs. M dwarfs come from Winters et al. (2015). Late-M, -L, and -T dwarfs are compiled from Dupuy & Liu (2012), Dupuy & Kraus (2013), and Liu et al. (2016). Bold circles represent young objects and companions with low surface gravities, which extend the standard M and L sequences to redderJ–K colors. The gray dots connected to Oph 98 A and B indicate the extinction-corrected locations of the binary for our derived extinction value. However, since the reddening is determined by best fits to templates, the dereddened colors may not be representative of the intrinsic colors of the components.

Zoom InZoom OutReset image size

Since the age of Ophiuchus can be inferred from stellar members of the region, the luminosities of Oph 98 A and B can provide estimates of the physical properties of the binary using evolutionary models. Recently, Esplin & Luhman (2020) analyzed the ages of various populations within the Ophiuchus star-forming complex. They estimated ages of ∼2 and ∼3–4 Myr for embedded and low-extinction members of L1688, respectively. Our estimated extinction for Oph 98 (A_V = 5 ± 1 mag) falls on the boundary ofA_V between the two L1688 populations defined by Esplin & Luhman (2020). Thus, we adopt an age of 3 Myr for Oph 98, but ages of 1–7 Myr fall within the interquartile age ranges of L1688.

We calculated the luminosities of Oph 98 A and B using extinction-corrected MKOKs photometry, theK-band bolometric corrections for young brown dwarfs from Filippazzo et al. (2015), and a parallax for L1688 of 7.29 ± 0.22 mas (Ortiz-León et al.2018). We used a Monte Carlo approach to determine uncertainties, taking a uniform distribution ofA_V and spectral type, and normally distributed uncertainties for the bolometric corrections and parallax. We determine luminosities of dex and −3.49 ± 0.06 dex for Oph 98 A and B, respectively.

Table2 presents the masses, effective temperatures (T_eff), radii, and surface gravities (logg) of Oph 98 A and B calculated from the DUSTY model isochrones of Chabrier et al. (2000) at ages of 1, 3, and 7 Myr. We estimate parameters for the bounds of the interquartile age range as the underlying age distribution of L1688 is unknown. Comparison of the luminosities to evolutionary models at the adopted age of 3 Myr yields masses of 15.4 ± 0.8 andM_Jup for Oph 98 A and B, respectively. Over the possible ages of the system (1–7 Myr), the mass of Oph 98 A could range from 9.6 to 18.4 M_Jup, and the mass of Oph 98 B from 4.1 to 11.6 M_Jup. The primary mass could therefore lie on either side of the planet/brown dwarf boundary (∼13M_Jup), while the secondary is confidently in the planetary-mass regime for all plausible system ages. Calculated masses using the evolutionary models of Saumon & Marley (2008) (f_sed = 2) and Burrows et al. (1997) were found to agree with the DUSTY results to within the uncertainties.

From the model-derived masses at 3 Myr and measured angular separation, we also computed the binary mass ratio, orbital period, and gravitational binding energy of the system, reported in Table2. We used a median correction factor of 1.1 from projected separation to true semimajor axis (Dupuy & Liu2011).

With a primary mass near the deuterium-burning limit (∼11–16M_Jup depending on metallicity; Spiegel et al.2011) and a companion inside the planetary-mass range, Oph 98 AB is among the lowest-mass binaries known to date. Only a handful of systems potentially made of two planetary-mass components have been discovered so far: the TW Hya candidate member 2MASS J1119−1137 (Best et al.2017), the WISE J1355−8258 spectral binary candidate in the AB Dor moving group (Theissen et al.2020), and the old field systems WISE J0146+4234 (Dupuy et al.2015) and CFBDSIR J1458+1013 (Liu et al.2011) that may be more massive depending on their ages.

Oph 98 is distinct from these binaries in three important aspects: the system’s age, mass ratio, and separation. The extremely young age of Ophiuchus makes Oph 98 the youngest system currently known with both components near or below the deuterium-burning mass boundary, and is thus the only example of such a binary detected right after birth. Additionally, most of these planetary-mass binaries are in nearly equal-mass configurations. In contrast, Oph 98 has a significantly lower mass ratio () for a comparable total mass. Finally, all the systems listed above have very tight orbital separations (<5 au), while Oph 98 AB is on a considerably wider 200 au orbit. No old field binaries are known on such large separations in this mass regime (see Fontanive et al.2018).

In these aspects, Oph 98 is more akin to wide systems with very low-mass companions identified in young associations (≤10 Myr). The closest analog to Oph 98 is certainly the other Ophiuchus brown dwarf binary, Oph 11 (17+14M_Jup, 240 au; Close et al.2007). Lower mass ratio counterparts include 2M1207 (Chauvin et al.2005) and 2MASS J0441+2301 (Todorov et al.2010), with ∼20M_Jup primaries and planetary-mass secondaries (4–5M_Jup), but significantly shorter separations of a few tens of astronomical units; or FU Tau (Luhman et al.2009) and USco CTIO-108 (Béjar et al.2008), on separations of hundreds of astronomical units, but considerably higher-mass primaries (∼50M_Jup) and secondaries around 15 M_Jup.

The combination for Oph 98 AB of small estimated component masses (15.4+7.8 M_Jup) and large measured separation (200 au) results in a remarkably low binding energy ofE_b = 8. 8 × 10³⁹ erg. This is lower by a factor of at least 2 than any of the binaries mentioned above, as illustrated in Figure4, with the second weakest binding energy belonging to FU Tau (see compilation in Faherty et al.2020). The Oph 98 system is therefore the brown dwarf binary with the lowest gravitational binding energy discovered to date.

Zoom InZoom OutReset image size

Oph 98, like each of the young low-mass binaries mentioned in Section4.1, provides a valuable example of a young system in an extreme configuration, offering key insight into the formation for the very lowest-mass brown dwarfs. Indeed, as multiplicity is a direct outcome of formation, the properties of very young binaries can serve as key diagnostics of formation pathways. The youth of the Oph 98 system indicates that the involved mechanisms must operate on short megayear-level timescales, compatible with the rapid formation expected from the fragmentation of cloud cores or gravitational instability in disks. The weakly bound nature of the system argues against violent dynamical processes that would have disrupted the binary, like the premature ejection of substellar embryos from the natal cloud (Reipurth & Clarke2001). Likewise, formation and subsequent ejection of both components from the disk of a more massive star (Stamatellos et al.2007) seems implausible based on the binary configuration.

It is more probable that Oph 98 A formed in a star-like manner, through the fragmentation of molecular cloud cores (Whitworth et al.2007). The low-mass companion Oph 98 B could have formed in the same way, or in the disk of the primary. The latter scenario is unlikely given the mass of the primary, too low to host a disk more massive than ∼1 M_Jup, and the wide binary separation, compared to radii of <30–100 au for brown dwarf disks (Testi et al.2016). The mass ratio of ∼0.5 also suggests a binary-like architecture rather than a planet-like origin for the secondary (Lodato et al.2005), further supporting the hypothesis that Oph 98 AB emerged from a stellar formation process.

With a lower mass limit for brown dwarfs around ∼3 M_Jup in a star formation framework (Whitworth et al.2007)—the minimum mass for opacity-limited fragmentation in turbulent cloud cores (Silk1977)—Oph 98 B thus extends the growing number of very young, planetary-mass objects populating the low-end tail of the star formation product (e.g., Liu et al.2013; Gagné et al.2014,2015; Schneider et al.2016). The existence of such wide, very low-mass binaries arising from a stellar formation pathway is predicted in numerical simulations (Bate2012), although expected to be of rare occurrence. As the binary with the weakest gravitational binding energy discovered to date, the Oph 98 system therefore represents an unmatched example of the extreme multiplicity outcome of stellar formation.