Hunting for the Birthplaces of Supermassive Black Holes in Terabytes of Simulation Data

by Prof. Michael Norman (University of California, San Diego)

Thursday 21 Mar 2019, 12:00 → 14:00 Europe/Berlin

New Seminar Room (MPE)

The origin of supermassive black holes (SMBH) in the early universe is a scientific puzzle that has occupied observers, theorists, and computational astrophysicists for several decades. Given their rarity, it is clear special circumstances are required to form a SMBH. One idea is the so-called direct collapse black hole (DCBH) scenario, where a large amount of primordial gas forms a supermassive star that promptly collapses to a black hole. Intense radiation backgrounds are required for this scenario to work, however, which creates a fine tuning problem. In this lecture I describe how we hunted for sites suitable for the DCBH mechanism to operate in terabytes of simulation output from the Renaissance Simulations. The Renaissance Simulations [1] are the most detailed simulations of the earliest stages of cosmic structure formation carried out on the Blue Waters supercomputer in the US. Using data mining and old fashioned curiosity, we found 10 potential birthplaces for SMBH formation amongst thousands of star-forming protogalaxies. Our results, recently published in Nature [2], presents a significant revision to the DCBH paradigm. It is an object lesson of the importance of applying data science approaches to analyze the results of modern simulations.

[1] Galaxy Properties and UV Escape Fractions during the Epoch of Reionization: Results from the Renaissance Simulations, H. Xu et al., Astrophysical Journal, Vol. 833, id.84, (2016), http://adsabs.harvard.edu/abs/2016ApJ...833...84X

[2] Formation of massive black holes in rapidly growing pre-galactic gas clouds, J. Wise et al., Nature, 566, 85-88 (2019), https://www.nature.com/articles/s41586-019-0873-4