Rapidly merging stars and black holes - the birth of supermassive black holes in dense star clusters in the early Universe
New observations by the James Webb Space Telescope (JWST) have revealed that supermassive black holes (SMBHs) of more than one million solar masses were already present only 450 million years after the Big Bang. How did these first SMBHs form? A team of researchers at MPA has used modern supercomputer simulations to show that progenitors of SMBHs (seeds) of a few thousand solar masses can form rapidly in dense and structured star clusters forming in the early Universe. They emerge from collisions of massive stars which form supermassive stars and then collapse directly into black holes, which can further grow by merging with other black holes. This new and more realistic model resembles JWST observations and can explain the formation of SMBH seeds which are massive enough to further grow into the earliest SMBHs observed. For this SMBH seed formation process, the researchers predict a unique gravitational wave fingerprint from black hole merger that can be directly tested with the next-generation gravitational wave observatories.
Supermassive black holes (SMBHs) with masses exceeding one million solar masses are found in all nearby massive galaxies including our own Milky Way. New observations by the James Webb Space Telescope (JWST) have revealed that SMBHs were already present only 450 million years after the Big Bang. The origin of these most massive black holes in the Universe is a major unsolved puzzle in modern astrophysics and an active area of research.
The very first stars in the Universe may have left behind black holes with masses up to a few hundred solar masses. However, models with such ‘light’ SMBH seeds struggle to explain the observed high redshift population of accreting SMBHs. The maximum sustainable SMBH gas accretion rate, the so-called Eddington rate, places limits on how fast SMBH seeds may grow after their formation. The light SMBH seeds simply do not have enough time to grow enough only in a few hundred million years. Therefore, more popular theoretical SMBH formation models assume that the SMBH seeds formed ‘heavy’ with masses exceeding a thousand solar masses. These heavy seeds black holes have a head start against the light seeds in their growth into the observed population of early accreting SMBHs. The major proposed heavy seed formation scenarios include runaway stellar collisions in dense star clusters, directly collapsing metal-free gas clouds in atomic cooling halos, and more exotic ‘new’ physics such as primordial black holes.
In dense star clusters, repeated stellar collisions may build up very massive and even supermassive stars. In early Universe which is still little enriched with heavy elements, stellar winds are typically weak and the stellar collision products will retain most of their mass. At the ends of their lives, these collisionally formed supermassive stars collapse and form the seeds for SMBHs.
Past simulations had focused on studying isolated, spherical star clusters. Both the JWST observations and state-of-the-art hydrodynamical galaxy formation simulations instead support the picture that massive star clusters form through a complex hierarchical assembly. This was the key motivation for the researchers at the MPA to re-explore the runaway collisional SMBH seed formation scenario in the more realistic clustered setup. Such a scenario is very different to the direct collapse gas cloud scenario which relies on avoiding cloud cooling and fragmentation into clusters of stars.
The researchers performed new simulations of massive star clusters with several million individual stars forming from the rapid assembly of several hundred proto-clusters. The newly developed direct N-body simulation code BIFROST used for the simulations runs on energy-efficient GPU hardware can follow stellar evolution, stellar mergers and accurately accounts for general relativistic effects during the interaction of black holes. In particular, the code computes the gravitational wave emission when two black holes merge. At the end of the merger anisotropic gravitational wave emission can kick the newly formed black holes up to speeds of several thousand km/s. These gravitational wave recoil kicks which can eject black hole merger remnants from their birth clusters are also modelled with the code.
The collision pathways massive stars and the formed SMBH seeds are illustrated in Fig. 1. Typically, only the most massive star in sub-clusters grows rapidly by collisions with other massive stars. Once the stars exceed the mass of several hundred solar masses, stellar evolution models predict that they directly collapse into black holes at the end of their lives. After their formation, the several SMBH seeds in the assembled massive star cluster experience a rich history of interactions and mergers by which the SMBH seeds can further grow. Several black holes are ejected from the cluster through strong Newtonian few-body interactions or relativistic gravitational wave recoil kicks. The hierarchical runaway scenario predicts a population of gravitational wave mergers at high redshifts in which the SMBH seeds merge with stellar mass black holes of several 10 to 100 solar masses (Fig. 2). Current gravitational wave observatories cannot detect black hole mergers above 500 solar masses or high redshifts very well. However, the scenario of the MPA researchers can be tested with the next-generation gravitational wave experiments such as LISA and the Einstein Telescope.
Antti Rantala, Thorsten Naab & Natalia Lahen
The authors thank Markus Rampp and Klaus Reuter of the Max Planck Computing and Data Facility (MPCDF) for performance optimization of the BIFROST GPU code. The simulations for the study were run using the MPCDF supercomputer Raven in Garching.