It is a well-established result that many stars do not form in isolation; young stars are usually found to be members of clusters. But little is known or understood about the origin of these clusters. In particular, evidence that pre-main sequence stars of intermediate (2-10 Msun) and higher masses are found in clusters has been found in several studies, at both optical and infrared wavelengths. Additionally, there has been an increased interest in the study of intermediate-mass stars, such as the Herbig Ae/Be stars, in the past ten years. These stars represent the most massive objects to experience an optically visible pre-main sequence phase, bridging the gap between low- and high-mass stars.
Studies in the nineties into the occurrence of young stellar clusters around Herbig Ae/Be stars concluded, based on near-infrared imaging data, that there was a difference in clustering properties between low and high mass stars.
Following the ideas included in those studies, the work in this thesis investigated the presence of clusters around 269 known Herbig Ae/Be stars, with the detailed astrometric data offered by the second Gaia data release with a novel clustering detection algorithm (CEREAL), which was developed for this task during this project. CEREAL is available from https://github.com/yumiry/CEREAL, along with several examples of its use for cluster classification.
Prior to the work included in this thesis, only 15 Herbig Ae/Be stars have been found to be in a cluster but, with CEREAL, it was possible to find clusters around 76 (28%) Herbig Ae/Be stars, which represents a significant improvement in the number of clusters found around Herbig Ae/Be stars.
In parallel to this, the results obtained with CEREAL were compared with other density-based clustering algorithms, where it was demonstrated that CEREAL was a simple and powerful, algorithm that could find clusters, of any size, from any sample.
In addition, the work included in this thesis has proved that clusters around the Herbig Ae/Be does not just appear for the sub spectral type range from B7 and earlier; this work has shown it is possible to find clusters around all the sub-spectral types of B stars.
This work then concluded through a comparison between the clusters found with CEREAL, around the Herbig Ae/Be stars, with Monolithic collapse and Competitive Accretion models from the literature which the aimed to evaluate whether the clusters around the massive stars appear to follow any of these models. The result of this comparison is that neither of these models appeared to fully describe the observed clustering behaviour of the Herbig Ae/Be stars observed by CEREAL; although, elements of the behaviours expected by both models were observed, which raises the possibility that a combination of both models may be required to fully explain massive star formation.
