An interferometric study of the post-AGB binary 89 Herculis. II Radiative transfer models of the circumbinary disk

TL;DR

This study uses detailed radiative transfer modeling to analyze the circumbinary disk around 89 Herculis, aiming to distinguish between disk and outflow origins of optical flux and to fit multi-wavelength observational data.

Contribution

It introduces comprehensive dust radiative transfer models including dust settling and scattering to interpret interferometric and photometric data of 89 Her.

Findings

A smooth inner disk rim is required to fit near-IR data.

Best-fit models include mm-sized grains settled in the disk midplane.

Optical data cannot be fit by a passive disk alone; an outflow or halo component is needed.

Abstract

The presence of disks and outflows is widespread among post-AGB binaries. In the first paper of this series, a surprisingly large fraction of optical light was found to be resolved in the 89 Her post-AGB system. The data showed this flux to arise from close to the central binary. Scattering off the inner rim of the circumbinary disk, or in a dusty outflow were suggested as two possible origins. With detailed dust radiative transfer models of the disk we aim to discriminate between these two configurations. By including Herschel/SPIRE photometry, we extend the SED such that it now fully covers UV to sub-mm wavelengths. The MCMax radiative transfer code is used to create a large grid of disk models. Our models include a self-consistent treatment of dust settling as well as of scattering. A Si-rich composition with two additional opacity sources, metallic Fe or amorphous C, are tested. The SED is fit together with mid-IR (MIDI) visibilities as well as the optical and near-IR visibilities of Paper I, to constrain the structure of the disk and in particular of its inner rim. The near-IR visibility data require a smooth inner rim, here obtained with a two-power-law parameterization of the radial surface density distribution. A model can be found that fits all the IR photometric and interferometric data well, with either of the two continuum opacity sources. Our best-fit passive models are characterized by a significant amount of mm-sized grains, which are settled to the midplane of the disk. Not a single disk model fits our data at optical wavelengths though, the reason being the opposing constraints imposed by the optical and near-IR interferometric data. A geometry in which a passive, dusty, and puffed-up circumbinary disk is present, can reproduce all the IR but not the optical observations of 89 Her. Another dusty, outflow or halo, component therefore needs to be added to the system.