Pseudo-viscous modeling of self-gravitating discs and the formation of low mass ratio binaries

TL;DR

This paper develops analytic models for self-gravitating accretion discs, predicting inevitable fragmentation beyond 70 AU, which may explain the formation of low mass ratio binary systems observed in nature.

Contribution

It introduces a pseudo-viscous modeling approach for self-regulated self-gravitating discs with realistic cooling, linking disc expansion to binary formation mechanisms.

Findings

Fragmentation occurs beyond 70 AU for typical infall rates.

Outward redistribution of material leads to fragmentation even in low angular momentum cores.

Proposes delayed fragmentation as a mechanism for forming low mass ratio binaries.

Abstract

We present analytic models for the local structure of self-regulated self-gravit ating accretion discs that are subject to realistic cooling. Such an approach can be used to predict the secular evolution of self-gravitating discs (which can usefully be compared with future radiation hydrodynamical simulations) and to define various physical regimes as a function of radius and equivalent steady state accretion rate. We show that fragmentation is inevitable, given realistic rates of infall into the disc, once the disc extends to radii $> 70$ A.U. (in the case of a solar mass central object). Owing to the outward redistribution of disc material by gravitational torques, we also predict fragmentation at $> 70$ A.U. even in the case of low angular momentum cores which initially collapse to a much smaller radius. We point out that 70 A.U. is close to the median binary separation and propose that such delayed fragmentation, at the point that the disc expands to $> 70$ A.U., ensures the creation of low mass ratio companions that can avoid substantial further growth and consequent evolution towards unit mass ratio. We thus propose this as a promising mechanism for producing low mass ratio binaries, which, while abundant observationally, are severely underproduced in hydrodynamical models.