Direct Imaging Discovery of a Young Brown Dwarf Companion to an A2V Star

The astrometric measurements of HIP 75056Ab display a significant amount of orbital motion (∼10% of a 45 yr orbit; see Table 1 and Figure 2). Given the large amount of orbital motion, we conservatively estimate the astrometric uncertainties as ±0.5 pixel, or ±0006, and note that these could be overestimated by a factor of ∼2 (a complete breakdown of the astrometric error budget for SPHERE can be found in Wagner et al. 2018a). Over the four-year baseline, we measure a positional shift of Δρ = 0025 ± 0008 and Δθ = 314 ± 18, which is consistent with an orbital period of ≳40 yr. To place preliminary constraints on the orbital parameters of HIP 75056Ab, we utilized the OFTI method (Blunt et al. 2017) in the orbitize! package (Blunt et al. 2020). We generated 10,000 sample orbits consistent with the astrometric data and found average orbital parameters of a = 30 ± 15 au, e = 0.5 ± 0.2, and i = 23° ± 11°. The largest semimajor axis would suggest a period of ∼220 yr.

Zoom In Zoom Out Reset image size

With only two data points available, the posterior orbital parameters have significant uncertainties, and are subject to complex degeneracies (e.g., between the inclination, eccentricity, and semimajor axis). For longer period orbits, to observe such a significant change in position angle (∼30°) over four years would require the companion to be currently near periastron on an eccentric orbit. The latter scenario is less likely because such a companion would spend the majority of its time at wider separations. With a third epoch observation, these orbital parameters can be significantly improved. With only two epochs available, we caution that this preliminary orbital fit may also be biased by systematics.

We converted the photometric measurements, age, and distance into mass estimates via the evolutionary grids of Baraffe et al. (2003) following the methodology in Wagner et al. (2019). We converted the J-, H-, and K-band photometry separately into mass estimates, and computed the combined mass probability distribution as the product of the individual distributions. We assumed that the star's K1 magnitude is equivalent to its K-band magnitude, and also assumed an age of 12 Myr and a 1σ age uncertainty of ±5 Myr based on the star's position and the age gradient map in Pecaut & Mamajek (2016). The results are consistent with a companion mass of ∼20–30 M_Jup. Figure 2 illustrates the range of plausible masses and the relative contributions of the photometric bands to the combined probability distribution. The mass range derived from the J-band photometry is consistent with somewhat higher masses (≲50 M_Jup), while the H-and K-band photometry suggest that the companion's mass is ≲30 M_Jup. Because the majority of the mass range is above the deuterium burning limit, the assumption of a high initial planetary entropy does not significantly affect the mass estimates (e.g., Mordasini et al. 2017; Marleau et al. 2019). We also verified that the atmospheric dust content does not affect the mass estimates by utilizing the model grids of Chabrier et al. (2000), which produced nearly identical results.

For the IFS data, we measured the spectrum and uncertainties of HIP 75056Ab using the mean and standard deviation of aperture photometric measurements corrected by the forward-modeled spectrum of the star from both ADI-KLIP and ADI+SDI-KLIP for each night (i.e., the mean and standard deviation of four spectra). The synthetic sources were injected at the same separation as HIP 75056Ab, with 8 × 10⁻⁴ contrast with respect to HIP 75056A, and at ΔPA = −90°, 90°, and 180° from HIP 75056Ab. For IRDIS, we measured the brightness and position by injecting a negative PSF of the star at the position of the companion. This method provides robust results in the presence of significant self-subtraction and over-subtraction due to ADI. We iterated upon the source's brightness and location, and adopted the parameters that minimized the squared residuals in the final image. The results are shown in Table 1 and Figure 3. We estimated the photometric uncertainties as the standard deviation of the brightness of positive sources injected at the same separation and brightness as HIP 75056Ab and at ΔPA = −90°, 90°, and 180°.

Zoom In Zoom Out Reset image size

The spectrum of HIP 75056Ab shows a consistent trend of increasing contrast (compared to HIP 75056A) with wavelength. We converted this contrast spectrum into physical flux units by multiplying by a synthetic T = 7500 K, R = 1.9 R_⊙ spectrum (Kurucz 1979) scaled to the distance of HIP 75056, which we selected to match the stellar photometry available from 2MASS (Cutri et al. 2003). We visually compared the spectrum to those of young brown dwarfs in Manjavacas et al. (2020). J1112−7653 and J0346+2321 were chosen among the objects within this library as the two whose spectra appeared most similar to that of HIP 75056Ab. Other spectra within the sample are consistent with the observed Y- to H-band spectrum of HIP 75056Ab, but are inconsistent with the full Y- to K-band spectrum. These objects have spectral types of M7 and L1, respectively. As the (normalized) spectra of these objects bracket that of HIP 75056Ab, we estimate a most likely spectral type within the range of M8−L0, and conservatively within the range of M6−L2. Similarly, we estimate that HIP 75056Ab has an effective temperature of ∼2300 ± 300 K. These are consistent with the evolutionary tracks of Baraffe et al. (2003) and Allard et al. (2012) for a ∼20–30 M_Jup and ∼10–20 Myr old companion.

We also compared the object's brightness and colors to those of field brown dwarfs with known distances (Dupuy & Kraus 2013; Winters et al. 2015) and directly imaged planets (Chauvin et al. 2005; Marois et al. 2008, 2010; Lagrange et al. 2010; Kuzuhara et al. 2013; Macintosh et al. 2015; Chauvin et al. 2017; Janson et al. 2019; Bohn et al. 2020a, 2020b). Consistent with the above findings, HIP 75056Ab is positioned in the J versus J − H color–magnitude diagram near the M/L transition. As one of few known young and low-mass companions near the M/L transition, HIP 75056Ab is potentially useful for comparative atmospheric studies (e.g., Sing et al. 2016; Madhusudhan 2019). HIP 75056Ab appears relatively blue among other M/L-transition objects. Notably, TYC 8998b—another low-mass companion to a Sco-Cen star—is also relatively blue.

We computed the image sensitivity as a function of angular separation from HIP 75056A using simulated point-source injection and retrieval tests. We utilized the IRDIS K1 image from 2019, which reaches the deepest sensitivity in terms of planetary masses. We measured the signal-to-noise ratio (S/N) using non-overlapping apertures of λ/D diameter, with the starting aperture centered on the position of the injected source, and using the definition of S/N in Mawet et al. (2014). We excluded a 12 × 12 pixel region centered on HIP 75056Ab, which would otherwise bias the S/N estimation. We utilized the off-axis PSF of the stars, normalized by the neutral density filter transmission, coronagraph transmission, and difference in exposure time. We converted contrast sensitivities to mass estimates following Section 3.2. The results are shown in Figure 4. At the separation of HIP 75056Ab (∼013), our observations are sensitive to contrasts of ∼2–3 × 10⁻⁴, corresponding to masses of ∼7–8 M_Jup. At twice the companion's separation, our observations are sensitive to contrasts of ∼3–4 × 10⁻⁵, or masses of ∼2–4 M_Jup. Our observations reach a minimum in contrast sensitivity of ∼4–5 × 10⁻⁶ at separations of ≳2'', or masses of ∼1 M_Jup. We note that orbits exterior to ∼15 may be unstable due to the effects of HIP 75056B, if the binary is on a nearly face-on and circular orbit (Holman & Wiegert 1999).

Zoom In Zoom Out Reset image size

The spectrum of HIP 75056Ab is similar to that of young brown dwarfs with spectral types of ∼M6−L2, with the large spread caused by the uncertainties in the IFS data. The CO band-head is present at ∼1.3 μm, as CO is the dominant carbon carrier in M/L-type atmospheres. Compared to field brown dwarfs and directly imaged (young super-Jupiter) exoplanets, HIP 75056Ab appears similar to other objects near the deuterium burning limit, such as TYC 8998b (Bohn et al. 2020a), a 14 ± 3 M_Jup companion orbiting a Sun-like star. Notably, HIP 75056Ab is among the least massive companions discovered to date near the M/L transition. As such, it constitutes a young, hot analog to colder directly imaged companions of similar masses around older stars, and also a low-mass and low-gravity analog to older field brown dwarfs of similar temperature.

The most peculiar aspect of HIP 75056Ab is arguably its formation mechanism. The companion's mass ratio of q ∼ 0.01 is analogous to a ∼10 M_Jup companion/planet around a Sun-like star. A ∼20–30 M_Jup companion at ∼15–45 au could represent a rare formation via gravitational instabilities within the protoplanetary disk (see Kratter et al. 2010; Wagner et al. 2019; Tokovinin & Moe 2020). Alternatively, HIP 75056Ab might be among the most massive companions formed via core accretion (e.g., Emsenhuber et al. 2020a, 2020b). Such rare examples often provide the most powerful constraints on formation models (e.g., the four super-Jupiter planets around HR 8799 that are difficult to explain with any mechanism; Marois et al. 2008, 2010). Nielsen et al. (2019) and Vigan et al. (2020) observed on the order of 100 stars with masses similar to HIP 75056A and found no companions with q ∼ 0.01 and a ∼ 10–50 au, suggesting an occurrence rate ≲1%. The most similar comparisons to HIP75056Ab are likely the several companions to B/A/F stars in Sco-Cen with q ∼ 0.01–0.08 and a ∼ 10–30 au that were discovered via sparse aperture masking (Hinkley et al. 2015).

Based on HIP 75056b's mass, it seems most likely that HIP 75056Ab formed via gravitational instability (Wagner et al. 2019); however, a core accretion origin cannot be excluded, and thus it is appropriate to consider another diagnostic. The best indicator is perhaps the companion's orbital eccentricity. Bowler et al. (2020) studied the eccentricity distribution of 27 companions spanning a few Jupiter masses to high-mass brown dwarfs between 5 and 100 au. They found that low mass-ratio companions (i.e., those formed predominantly via core accretion) have typically lower eccentricities than high mass-ratio companions, which have a broad peak at e ∼ 0.6–0.9, similar to the wide-binary population. Thus, if HIP 75056Ab is on a high-eccentricity orbit, this would support the hypothesis that it likely formed via gravitational instability (although see Emsenhuber et al. 2020b, which shows that high-eccentricity companions might also be formed via core accretion).

Our two currently available astrometric measurements (spanning four years) are sufficient to place preliminary constraints on the companion's orbit, which suggests e ∼ 0.5 ± 0.2. This is a significant eccentricity compared to other low-mass ratio companions (Bowler et al. 2020), supporting the hypothesis that HIP 75056Ab formed via disk instability (although this does not completely exclude a core accretion origin; see Emsenhuber et al. 2020b). In the coming years, continued astrometric monitoring of HIP 75056Ab will be able to verify and better constrain the orbital eccentricity.

Regardless of formation mechanism, given its mass, HIP 75056Ab likely formed early in the system's lifetime in the Class 0 or I stage when the protoplanetary disk was still massive and embedded in a gaseous envelope (Tychoniec et al. 2020). With a mass accretion rate typical for Class 0/I stars of ∼10⁻⁵ M_⊙ yr⁻¹, massive disks should be unstable at 30 au (Armitage 2019). Such early planet formation could lead to an observable signature in the C/O ratio of its atmosphere as compared to field brown dwarfs and giant planets. van 't Hoff et al. (2020) studied the temperature structures in embedded disks and found that Class 0 disks are warm, with no CO ice frozen out at all and H₂O ice frozen out only at radii of ≳80–100 au. By the Class I stage the disks cool, moving the water snow line further in (≲30 au), while allowing CO to freeze out beyond HIP 75056Ab's observed location. If HIP 75056Ab formed in the Class 0/I stage, then it likely did so in the presence of CO in the gas phase and with a decreasing amount of gaseous H₂O present with time. Therefore, observing a low C/O ratio in its atmosphere may imply an earlier period of runaway gas accretion.

HIP 75056Ab is among several substellar companions that have been directly imaged at very small angular separations (e.g., Strampelli et al. 2020). β Pictoris b (Lagrange et al. 2010; Wang et al. 2016), PDS 70 c (Haffert et al. 2019), and HD 206893B (Milli et al. 2017) have been imaged interior to 03. HD 984B (Meshkat et al. 2015) and PDS 70b (Keppler et al. 2018; Wagner et al. 2018b) have been imaged at ∼02. The images of these companions, including HIP 75056Ab as a notable example at 0125, showcase the trend of extreme adaptive optics systems (e.g., Macintosh et al. 2018; Beuzit et al. 2019) progressing toward imaging companions at smaller angular separations. These separations open interesting possibilities, as an Earth-like planet orbiting in the habitable zone of a Sun-like star at 10 pc would appear at an angular separation of ∼01.