Towards Testing The Tidal Downsizing Hypothesis For Planet Formation

Formation theories for planets are facing significant challenges in the light of observed planets and system architecture that greatly differs from our own solar system. The discovery of extrasolar planets is now an almost daily occurrence, with each confirmed detection adding more constraints to the planet forming process. In this thesis I study several topics relevant to the Tidal Downsizing scenario of planet formation. In this scenario massive gas clumps are born in cold discs beyond tens of AU. As they migrate in, they accrete pebbles and assemble massive solid cores and planetesimals by grain growth and sedimentation to the centre of the clump. Some of the clumps are disrupted, releasing both cores and planetesimals, others survive as gas giant planets. Dust modelling is a recent addition to full 3D simulations for planet formation. I have worked on implementing an implicit dust scheme for PHANTOM’s two-fluid dust scheme. I show the current state of my implementation as well as a suite of tests at the end of Chapter 2 after a summary of current SPH methods. While the implicit dust time step scheme is working for the basic tests there are still problems with the integration over long time periods. Chapter 3 presents population synthesis study of how this scenario can explain observed trends in the frequency of occurrence of planetesimal debris discs, massive cores and gas giants with their host star metallicity. In particular, classical theory predicted that debris, cores and giants should be all more abundant in high metallicity systems. However observations showed that the first two correlate with metallicity only very weakly. I find that in Tidal Downsizing these observations are natural as both debris and cores are produced when giants are destroyed. Chapter 4 investigates the robustness of numerical modelling of clump migration and accretion with 6 particle based and one grid based codes. There is a general qualitative agreement between the codes, but the quantitative agreement is only good to within a factor of two. I find that the artificial viscosity treatment may account for much of the differences between the codes. Code performance is nevertheless encouraging given very different numerical algorithms and the fact that physical uncertainties of the problem are far greater than numerical disagreements. We also compare prescriptions from three previous population synthesis studies to try and reproduce our numerical results. None of the three are very accurate over the wide parameter space of the problem, with some over-predicting and others under-predicting the number of objects surviving disc dissipation on wide orbits. Our results should help build better population synthesis to extract from observations of present day wide separation objects their primordial numbers and properties.