Preventing Anomalous Torques in Circumbinary Accretion Simulations

Abstract

Numerical experiments are the primary method of studying the evolution of circumbinary disks due to the strong nonlinearities involved. Many circumbinary simulations also require the use of numerical mass sinks: source terms which prevent gas from unphysically accumulating around the simulated point masses by removing gas at a given rate. However, special care must be taken when drawing physical conclusions from such simulations to ensure that results are not biased by numerical artifacts. We demonstrate how the use of improved sink methods reduces some of these potential biases in vertically-integrated simulations of aspect ratio 0.1 accretion disks around binaries with mass ratios between 0.1 and 1. Specifically, we show that sink terms that do not reduce the angular momentum of gas relative to the accreting object: 1) reduce the dependence on the sink rate of physical quantities such as the torque on the binary, distribution of accretion between binary components, and evolution of the binary semi-major axis; 2) reduce the degree to which the sink rate affects the structure of the accretion disks around each binary component; 3) alter the inferred variability of accretion onto the binary, making it more regular in time. We also investigate other potential sources of systematic error, such as the precise from of gravitational softening and previously employed simplifications to the viscous stress tensor. Because of the strong dependence of the orbital evolution of the binary on both the torque and distribution of mass between binary components, the sink methods employed can have a significant effect on the inferred orbital evolution of the binary.