The First Scattered Light Image of the Debris Disk around the Sco-Cen target HD 129590

Observing the youngest stellar systems, shortly after planet formation has ceased, provides a glimpse of nascent circumstellar environments. However, only a small handful of stars ($\lesssim$3-5) with ages $\lesssim$10-20 Myr, just after dissipation of the gaseous primordial disc, are found within 100 pc (Sacco et al., 2014). The Scorpius Centaurus Association (de Zeeuw et al., 1999, hereafter Sco-Cen) is the nearest OB2 association, with a mean distance of 140 pc, making it perhaps the most promising collection of young stars which can be observed shortly after the period of active planet formation. This region, containing stars with ages $\sim$ 10-16 Myr (Pecaut & Mamajek, 2016) allows the best constraints to be placed on the orbital zones of planet formation, and on the early thermal histories of young planets (Ireland et al., 2011; Janson et al., 2013; Lafrenière et al., 2011; Hinkley et al., 2015).

Furthermore, such young systems often possess bright circumstellar debris disks, belts of planetesimals analogous to the Edgeworth-Kuiper belt in our own solar system. The presence of dust in these systems suggests that planetesimals are responsible for the dust generation (Wyatt, 2008). The dust that is generated through planetesimal collisions in these debris disks is inherently transient, being either blown out by stellar winds, or spiralling towards the host star via Poynting-Robertson drag. The persistence of dust in these systems implies that it is constantly being regenerated. Although populations of planetesimals may stir themselves in some cases (e.g. Kennedy & Wyatt, 2010), this dust regeneration may be enhanced by perturbations from massive planets, dynamically exciting the planetesimals onto eccentric orbits, thus causing them to collide. As well as revealing the presence of massive planetary perturbers, the location of these large quantities of dust may hint at the location of giant planets in the system: gaps between dust belts may highlight where planets lie (Su & Rieke, 2014), and sharp edges to debris rings can constrain the masses of planets shepherding these edges (e.g. Quillen et al., 2006; Chiang et al., 2008; Mustill & Wyatt, 2012).

One Sco-Cen system with a known circumstellar debris disk is HD 129590 (HIP 72070, T${}_{\textrm{eff}}$ = 5945 K, 1.3 M_⊙ 2.8 L_⊙, Chen et al., 2011). While not originally catalogued in the de Zeeuw et al. (1999) or Rizzuto et al. (2011) catalogs of Sco-Cen stars, HD 129590 has been listed as a G1V member of the Sco-Cen subgroup Upper Centaurus Lupus (Hoogerwerf et al., 2000; Chen et al., 2011). HD 129590 has an estimated distance of 141$\pm$7 pc (Astraatmadja & Bailer-Jones, 2016; Gaia Collaboration, 2016a, b) and a high infrared luminosity $L_{\textrm{IR}}/L_{\textrm{star}}$$\sim$5 $\times$ 10^-3, twice the value observed for $\beta$ Pictoris (Jang-Condell et al., 2015; Mittal et al., 2015), as can be seen in Figure 1. Although Chen et al. (2014) fit the SED as two distinct belts, Jang-Condell et al. (2015) suggest the system likely contains a single belt of dust. This is in agreement with the predictions of Ballering et al. (2013) (see §4.1 for further details).

The high fractional luminosity of HD 129590 makes it an extremely promising target for scattered light imaging of circumstellar material (Currie et al., 2014; Draper et al., 2016). ALMA data have previously been obtained for this object (Lieman-Sifry et al., 2016), with the disk being marginally resolved along the major axis. Lieman-Sifry et al. (2016) constrain the inclination angle to $>$ 50^∘ with a best fit value of 70^∘, and the position angle to -59${}^{+17}_{-12}\,{}^{\circ}$. The ALMA data finds a best-fit grain size of $3.2^{+0.6}_{-0.5}$ µm, and at this size find the outer edge to be $110^{+50}_{-30}$ AU and the inner edge to be $<$ 40 AU. No gas was detected in that work, and so we do not expect gas to have a strong influence on the dust dynamics in this disk.

In this letter, we present the first scattered light image of the debris disk around HD 129590, using the SPHERE instrument on the VLT. The observations and data post-processing are presented in §2 and §3 respectively. In section §4 we describe our modelling of the disk, where we use an optically thin disk model to conclude that the disk has a radius of $\sim$ 60-70 AU. Finally, we conclude in §6.

HD 129590 was observed on 2016 May 4 with the SPHERE instrument at the VLT (Beuzit et al., 2008), as part of a larger planet-finding survey within Sco-Cen. The data were taken in the IRDIFS mode, whereby light is split through a dichroic beamsplitter, and passed simultaneously to both the differential imager and spectrograph instrument (IRDIS; Dohlen et al., 2008), and the integral field spectrometer (IFS; Claudi et al., 2008). A total of 2560s on-sky integration was collected by each instrument, with the N_ALC_YJH_S coronagraph in place. The sequence consisted of 80$\times$32s individual exposures for the IRDIS data and 40$\times$64s exposures for the IFS data. We used IRDIS in dual-band imaging (DBI; Vigan et al., 2010) mode with the H23 filter pair, at wavelengths $\lambda$ = 1588.8 nm, $\Delta\lambda$ = 53.1 nm and $\lambda$ = 1667.1 nm, $\Delta\lambda$ = 55.6 nm, while the IFS instrument (Zurlo et al., 2014; Mesa et al., 2015) was used in the YJ mode, which spans the range 0.95-1.35 $\micron$. Platescales of the instruments are 12.25 mas/pix for IRDIS and 7.46 mas/pix for the IFS (Maire et al., 2016), and the inner working angle of the coronagraph is 0.15^′′.

In addition to these science observations, flux calibration images were collected with the target displaced from the coronagraph. Star center calibration frames (waffle frames) were also collected, by imposing a sinusoidal pattern on the deformable mirror. This creates four starspot images, with equal displacements from the central star, in each corner of the frame. Together, these allow the star position to be accurately measured to $\sim$0.1 pixels (1.2 mas) (Vigan et al., 2016) behind the occulting mask. The observations were carried out in pupil-stabilized mode to allow angular differential imaging analysis (ADI; Marois et al., 2006). The entire sequence of observations, including acquisition and calibration, lasted 59 minutes spanning an airmass range of 1.037 to 1.042. The primary science frames covered a total field rotation of 36 degrees, and included the meridian crossing of the target.

Our data post-processing was carried out following the process described in Vigan et al. (2015) (here on: V15). We used both the ESO data reduction and handling pipeline (DRH, Pavlov et al., 2008) and the publicly available code described in V15, as well as some custom routines.

The latest generation of dedicated high resolution exoplanet imaging platforms such as GPI and SPHERE has already revealed a number of scattered light debris disc images. Sco-Cen has proved to be a particularly fortuitous region for these searches. Currie et al. (2015b) detected a dust ring around HD 115600, which was shown to be eccentric. HD 110058 has a wing-tilt asymmetry (Kasper et al., 2015), while HD 106906 appears to be misaligned with the wide planetary companion (Kalas et al., 2015; Lagrange et al., 2016). HD 111520 Draper et al. (2016) has a dramatic brightness asymmetry, while both HIP 67497 (Bonnefoy et al., 2017) and HIP 73145 (Feldt et al., 2017) have been shown to have multiple, separate, rings of debris. Perhaps the only debris discs without complex morphology are HD 114082 (Wahhaj et al., 2016) and HD 129590.

All of these debris discs are presented in Chen et al. (2014) as hosting multiple distinct debris belts. Multiple debris belts have only been seen in scattered light around two of the above targets: either there are several very small, close in belts of disc evading detection, or two temperature discs correspond to two belt discs less often than expected. Additionally, all of the targets with debris disc images have very high excess infrared luminosity values, and all except HIP 67497 have been observed with ALMA (Lieman-Sifry et al., 2016).

As discussed in §4.1, the debris belt around HD 129590 is predicted to lie at $\sim$16 AU, based on the infrared excess temperature and simple blackbody constraints. Our observations show the ring to peak at $\sim$4 times this separation, implying an abundance of small dust at higher temperatures than the blackbody temperature. This is to be expected for a luminous star (2.8 L_⊙) where the blowout size is a few microns.

We attempt to classify HD 129590 under the categories outlined in Lee & Chiang (2016). The pronounced difference in brightness between the front and back of the disk is reminiscent of the ‘moth’ or ‘double wing’ structures simulated in that work, but we do not observe the bright, extended wings seen in e.g. HD 61005 (Hines et al., 2007). A strongly forward-scattering disk could produce only a fainter wing structure, as light will be preferentially scattered away from the viewer. In addition, an ADI based code such as that used here will self-subtract flux in regions at small angular separations to the bright disk edge, and so hide faint wing structures nearby. In Figure 4 contours of equal brightness are plotted, to demonstrate the full dynamic range of the disk. There is no indication of an extended wing, and indeed the self-subtraction lobes (highlighted in red in the image) show a high degree of symmetry. As such, there is no evidence for moth-like wings and it is not possible to make detailed inferences about the presence, or otherwise, of a planet in this system.

Our modelling work highlighted the presence of extended emission, at large semi-major axis. This could be a dust halo, caused by radiation pressure blowing out small grains. Alternatively, this may be a scattered disk of planetesimals, and as such more closely resemble the wings mentioned above - albeit much more compressed into the plane than Lee & Chiang (2016) find. Such a ring of planetesimals might explain our surprisingly low value for $\alpha_{\mathrm{out}}$. A higher resolution ALMA image could differentiate these two scenarios, while future imaging with a space facility such as the Hubble Space Telescope would allow the lowest surface brightness material (such as a dust halo) to be imaged, as well as defining the outer disk edge far better than ground based AO imaging. Polarimetric differential imaging, meanwhile, would place constraints on the dust grain size and scattering properties, and may even reveal the faint edge of the disc.

HD 129590 is a G1V member of the Sco-Cen association, with an infrared excess twice that observed for $\beta$ Pictoris. This work presents the first scattered light images of the debris disk responsible for this infrared excess. The debris disk is revealed to be a nearly edge-on disk, with evidence for inner clearing. We use the GRaTeR radiative transfer code to model the disk as an optically thin ring, inclined to the line of sight by $\sim$ 75^∘. Our best fitting model has a characteristic radius of r_∘=59.3 AU or r_∘=73.3 AU depending on the underlying model, and a forward scattering parameter $g$=0.52 or $g$=0.43. When the power law edges were freed, these were found to take values of 3.15 inside the ring and -1.313 outside the ring. These values imply a strongly forward scattering ring, with a soft outer edge. Even with this model, there is an indication in the final panel of Figure 3 of some residual structure, implying that there is some morphology more complex than a simple ring present in this disk.