Combined dynamo of gravitational and magneto-rotational instability in irradiated accretion discs

Physics Department of Bayreuth, Universitätsstraße 30, Bayreuth, Germany
e-mail: lucas.loehnert@uni-bayreuth.de

Received: 31 August 2021
Accepted: 3 April 2022

Abstract

Aims. We aim to assess whether magneto-rotational instability (MRI) can exist in a turbulent state generated by gravitational instability (GI). We investigated the magnetic field saturation and elucidated the ability of GI turbulence to act as a dynamo.

Methods. The results were obtained by numerical simulations using the magnetohydrodynamics code Athena. A sub-routine to solve the Poisson equation for self-gravity using three-dimensional Fourier transforms was implemented for that purpose. A GI-turbulent state was then restarted, with a zero-net-flux type magnetic seed field being introduced. The seed field was chosen with β ≈ 10¹⁰ to make sure that the magnetic field of the stationary state is exclusively generated by the dynamo.

Results. Shortly after introducing the magnetic seed field, a significant field amplification is observed, despite MRI not being active. This shows that GI acts as a kinematic dynamo. The growing magnetic field allows MRI to become active, which leads to the emergence of a butterfly diagram. The turbulent stress of the saturated state is found to be consistent with the superposition of GI stresses and MRI stresses. Moreover, the ratio of magnetic stress to magnetic pressure is found to lie in the 0.3−0.4 range, which is typical for MRI turbulence. Furthermore, it is found that the magnetic energy significantly decreases if self-gravity is turned off. This indicates, in accordance with the initial field amplification, that GI provides the dominant dynamo contribution and that MRI is not simply added but rather grows on the magnetic field provided by GI turbulence. Finally, it is shown that the combined GI-MRI-dynamo is consistent with an α − Ω model and that the observed oscillation frequency of the butterfly diagram roughly agrees with the model prediction.