The rarity of young massive stars combined with the fact that they are often deeply embedded has limited the understanding of the formation of stars larger than 8M⊙. Ground based mid-infrared (IR) interferometry is one way of securing the spatial resolution required to probe the circumstellar environments of massive young stellar objects (MYSOs). Given that the spatial-frequency coverage of such observations is often incomplete, direct-imaging can be supplementary to such a dataset. By consolidating these observations with modelling, the features of a massive protostellar environment can be constrained. By specifically choosing observations which probe different scales of the MYSO environments, these characteristics can be solidified at multiple scales and provide information on different aspects of the massive star formation process. By repeating this analysis for a sample of objects, a reliable set of parameters are obtained for each MYSO. Since the objects have been analysed in the same manner, a consistent comparison between the derived characteristics of the sources can then be made.
This thesis performs the above through the combination of N-band interferometry at ∼0.01" resolution, near-diffraction-limited Q-band imaging at ∼0.1" resolution and spectral energy distributions. These observations are consolidated by fitting 3D radiative transfer models to all three observables simultaneously. Such an involved methodology constitutes a number of improvements on previous infrared studies of MYSOs, through the use of greatly detailed models and the combined use of different observing techniques, which each provide a unique perspective on the studied protostellar environments. Specifically, N-band interferometry and Q-band imaging were combined, allowing a distinction to be made between the disk and cavity emission of the sources. The Q-band data traces cavity emission at larger scales, whereas the hotter regions are traced by the N-band. With this additional constraint, the N-band data has proven to be an excellent tracer of disk geometry. In particular, the inner rim, flaring and dust content of the disks can have large effects on simulated N-band visibilities. The sensitivity of the N-band interferometric data to the inner rim emission allowed the detection of an inner hole within the disk of the source G305.20+0.21, the pilot object of this study. This could not be rectified by changing the luminosity of the source to move the sublimation radius, the default minimum dust radius, indicating that some other mechanism is responsible for the disruption of the dust in the inner regions of the circumstellar environment. The SED fit for this source also required low densities, implying dispersal of the protostellar environment. These are typical characteristics of evolved transition disks found during the low-mass star formation and the possibilities of this applying to an MYSO are discussed.
When the method was applied a sample of eight sources, it was discovered that G305 was not alone in its requirement to have a disk in its model. Every MYSO studied required a disk as a model component, and a number of sources within the sample also required inner holes in their environments in addition to G305. Overall, these holes ranged in size between 20-125au across. While diversity exists within the sample, with different cavity opening angles, densities etc., the overall geometry of the sources is the same; an Ulrich-type envelope, bipolar outflow cavities and a dust disk. The fits to the interferometric data of one source could be improved by a gap structure, which is a feature now commonly observed in Herbig Ae/Be disks.
The physical parameters determined through this work were linked to a potential evolutionary sequence through comparison to work done by Cooper et al. (2013). Cooper et a. (2013) classified just under 200 MYSOs into three different types based on emission lines in their spectra with each type representing a different evolutionary stage of the MYSO. All the sources classified as the oldest type within Cooper et al. (2013) have substructure in their disks. This is in agreement with low-mass stellar evolution which sees substructure develop at later stages, and therefore has implications for the evolution of massive protostars.
The nature of these massive protostellar environments, as determined by this thesis, show many similarities with various stages of low-mass star formation. Disks of order solar masses in size are components of all the final models for this sample of MYSOs. These disks extend from 10s of au to 1000s of au, and as such could be components of the accretion process, as they are for low-mass stars. The sample studied in this thesis increases the number of MYSOs studied at such a level of detail in the literature eight-fold. Five of the eight studied sources have inner dust rims larger than the sublimation radii expected of the host protostars in their models. Further study is required to directly investigate these holes and substructure found in this work and determine their cause. The application of this method to more sources can expand on these findings to the point where the formation and evolution mechanisms of massive young stars can be characterised in detail.
