Simulating the Co-Evolution of Massive Galaxies and their Black Holes

Claude-Andre Faucher-Giguere, Northwestern University

Usage Details

Key questions include: What is the origin of the galaxy-black hole scaling relations? How is active galactic nucleus (AGN) activity triggered? What is the impact of AGN feedback on galaxy evolution? What are the effects of stellar feedback on supermassive black holes (SMBH) growth? For this proposal, we will address these questions by focusing on massive galaxies at the peak of cosmic star formation and AGN activity (z ≈ 1-3), which has been the focus of intense observational campaigns against which the simulations can be tested. Over the past three years, we have developed new methods to simulate the growth of SMBHs in very high resolution cosmological simulations using the FIRE set of subgrid models, which are able to capture the multiphase interstellar medium (ISM) shaped by stellar feedback. We have now evolved at high resolution a sample of four such massive galaxies with BHs and in the past year have demonstrated our ability to efficiently run our code on Blue Waters (BW). For the next steps, we will use BW to evolve a larger sample of similar simulations to cover a representative range of galaxy mass. This expanded sample is necessary to capture the expected galaxy-to-galaxy variance and will enable robust theoretical analyses and comparisons with observations.

The BW computational capabilities are essential to simulate at high resolution the massive galaxies in which most SMBH growth occurs. Furthermore, the hybrid OpenMP+MPI parallelization of our GIZMO code and its excellent scaling to ≥ 10⁴ cores are ideally suited for BW. Pushing our simulations to the scale proposed here will represent an order-of-magnitude jump in computational demands relative to most of our previous simulations (of less massive galaxies) and will thus enable us to test and optimize the scaling of cosmological zoom-in simulations in a new regime. Since a version of GIZMO is publicly available, this optimization will benefit the community at large.