Accretion is an important process relevant to various fields in astrophysics, from star formation to the study of compact binaries and Type I supernovae. Continuous theoretical and numerical efforts have been devoted to understanding the interaction between the inner edge of an accretion disk and the stellar surface. Complementary to these works, there are also many studies devoted to the effect of accretion on the structure and evolution of objects, using either simple semi-analytic approaches or simulations based on one dimensional (1D) stellar evolution codes. But no multi-dimensional numerical study has ever been devoted to the effect of accretion on the structure of accreting objects. Because of this lack of sophisticated modelling, great uncertainties still remain regarding the fraction of accretion luminosity imparted to an accreting object or radiated away, and how the accreted material is redistributed over the stellar surface and the depth it reaches. Early accretion history may impact the properties of stellar objects (luminosity, radius, and effective temperature) in star formation regions and young clusters even after several Myr when the accretion process has ceased. These effects could provide an explanation for several observed features in star-forming regions and young clusters, having important consequences on our general understanding of star formation.
The DiRAC Resource, Complexity, has been used to perform the first hydrodynamics simulations describing the multi-dimensional structure of accreting, convective low mass stars. The hydrodynamical simulations were computed using our fully compressible time implicit code MUSIC (Multidimensional Stellar Implicit Code) recently improved with the implementation of a Jacobian-free Newton-Krylov time integration method, which significantly improves the performance of the code. These developments and the benchmarking of the code were made using Complexity. These new accretion calculations have enabled us to understand the effect of so-called “hot” accretion, when the accreted material is characterised by a larger entropy than that of the bulk of the stellar material.
The multi-D simulations show an accumulation of hot material at the surface that does not sink and instead produces a hot surface layer, which tends to suppress convection in the stellar envelope. This enables us to derive an improved treatment of accretion in 1D stellar evolution models, based on an accretion boundary condition. Such treatment of accretion in a 1D stellar evolution code is more physical than the standard assumption of redistribution of the accreted energy within the stellar interior. These new results will be published in Geroux, Baraffe, Viallet et al. (2016, A&A, submitted). The impact of this more physical accretion boundary condition on the evolution of young low mass stars and brown dwarfs and on observations in young clusters will be explored in a follow-up studies. These pioneering simulations performed on the DiRAC facility shows the potential of using time implicit multi-D simulations to improve our understanding of stellar physics and evolution.