The physical properties and composition of vega-type disks

Abstract

Vega excess stars are main sequence stars which are surrounded by a cool dust disk, left as a remnant of the massive disk produced early in the star’s formation. Vega-type disks are typically optically thin, contain little or no gas, and are roughly 100 AU in radius with a cleared region close to the star. This thesis presents a programme of observational studies which aimed to determine the basic physical properties of nearby Vega-type disks, and the detailed modelling and analysis that were required to interpret these observationsThe first part of the thesis presents new sub-mm observations of Vega excess stars, and modelling for all known Vega excess stars which have sub-mm data, using dust grain models with realistic optical properties. Analysis of the resolved targets showed that different objects require very different dust grain properties in order to simultaneously fit the image data and spectral energy distribution (SED). Fomalhaut and Vega require solid dust grains, whilst HR4796 and HD 141569 can only be fitted using porous grains. The older stars tend to have grains which are less porous than younger stars, which may indicate that collisions in the disks have reprocessed the initially fluffy grains into a more solid form, ε Eri appears to be deficient in small dust grains compared to the best fitting model, which may be due to factors which affect the size distribution of grains close to the radiation pressure blowout limit. When the model is applied to the unresolved targets, an estimate of the disk size can be made. However, the large diversity in dust composition for the resolved disks means that it is impossible to make a reliable assumption as to the composition of the grains in the unresolved disks, and there is a corresponding uncertainty in the estimated disk size. In addition, the poor fit for ε Eri shows that the model cannot always account for the SED even if the disk size is known. These two factors mean that it may not be possible to determine a disk’s size without actually resolving it.The second part of the thesis describes mid-IR observations designed to directly resolve the disks around several nearby main sequence stars, and hence obtain a direct measurement of the disk size. An analytical model of the telescope point spread function (PSF) was developed and fitted to observations of standard stars, and this model was used to establish whether the science observations were consistent with a point source, or if they showed evidence for a resolved disk. Though the observations failed to resolve any of the targets, techniques were developed to reduce imaging data of marginally resolved disks, and the observations have provided a clearer idea of what is required for a successful program in the future. The key requirements are regular monitoring of the PSF (i.e. interleaved PSF star observations), a high quality flatfield and fairly narrow band filters. In addition, 20 /im observations are probably needed in most cases. Ultimately, the technique is limited by the stability of the PSF, and on the difference in colour between the PSF star and the debris diskFinally, in the third part of the thesis I present the results of a search for molecular hydrogen using MICHELLE, the mid-IR echelle spectrometer at the UK Infrared Telescope. The aim of these observations was to investigate the controversial ISO detection of H₂ around Vega excess stars, as reported by Thi et al. (2001, Nature, v. 409, p. 60). Due to weather constraints, the observations focused on two pre-main sequence stars, AB Aur and CQ Tau. No significant emission lines were detected from either star, and the upper limits on line flux were significantly smaller than those measured from the ISO observations, which suggests that the emission detected by ISO is extended on scales of at least 6 arcsec, and does not come from the disk as previously thought. This result indicates that the ISO detection of large amounts of H₂ in Vega-type disks may also be unreliable, and further observations are needed to determine their true gas content.