Introducing cuDisc: A 2D code for protoplanetary disc structure and evolution calculations

We present a new 2D axisymmetric code, cuDisc, for studying protoplanetary discs, focusing on the self-consistent calculation of dust dynamics, grain size distribution and disc temperature. Self-consistently studying these physical processes is essential for many disc problems, such as structure formation and dust removal, given that the processes heavily depend on one another. To follow the evolution over substantial fractions of the disc lifetime, cuDisc uses the CUDA language and libraries to speed up the code through GPU acceleration. cuDisc employs a second-order finite-volume Godonuv solver for dust dynamics, solves the Smoluchowski equation for dust growth and calculates radiative transfer using a multi-frequency hybrid ray-tracing/flux-limited-diffusion method. We benchmark our code against current state-of-the-art codes. Through studying steady-state problems, we find that including 2D structure reveals that when collisions are important, the dust vertical structure appears to reach a diffusion-settling-coagulation equilibrium that can differ substantially from standard models that ignore coagulation. For low fragmentation velocities, we find an enhancement of intermediate-sized dust grains at heights of ∼1 gas scale height due to the variation in collision rates with height, and for large fragmentation velocities, we find an enhancement of small grains around the disc mid-plane due to collisional ‘sweeping’ of small grains by large grains. These results could be important for the analysis of disc SEDs or scattered light images, given these observables are sensitive to the vertical grain distribution.