arXiv subjects

Pelle van de Bor

Publications and source records attributed to Pelle van de Bor.

4 recordsLinked to original sources

arXiv2026

SEEDZ: Rapid Galaxy Assembly as a Pathway to Supermassive Stars, Dense Stellar Environments and Massive Black Hole Seeds

We investigate the assembly history of early galaxies in the SEEDZ hydrodynamic simulations, to investigate the high inflow rates believed to be required for the formation of supermassive stars (SMSs), dense stellar clusters and subsequently heavy seed black holes. Using a heavy seed formation criteria of $>$1 M$_\odot$ yr$^{-1}$ flowing into 10 pc regions, we find that heavy seeds form in halos that grow rapidly compared to those halos that never meet the criteria. Halos with growth rates of $\gtrsim$1 M$_\odot$ yr$^{-1}$ at their virial radius (scales of a few hundred pc) are able to sustain a flow rate of 0.1 M$_\odot$ yr$^{-1}$ into the inner 1 pc of the halo, maintaining higher density environments within the central 10 - 100~pc. These halos continue to grow rapidly after their initial collapse, typically forming heavy seeds $\sim$100 Myr after forming their first stars and stellar mass black holes. By $z=10$, most heavy seeds form in regions of near-solar metallicity, although a minority of heavy seeds do continue to form in low metallicity (10$^{-2}$ Z$_\odot$) regions. Under the assumption that a SMS forms as the progenitor to a heavy seed if it forms in a region of low (10$^{-2}$ Z$_\odot$) metallicity, and can sustain high accretion rates above 0.02 M$_\odot$ yr$^{-1}$ throughout the SMS lifetime of 2 Myr, we find a number density of SMSs of 0.1 cMpc$^{-3}$, meaning that only a fraction of 10$^{-4}$ of these SMSs would need to be visible to JWST to account for the observed population of Little Red Dot galaxies.

astro-ph.GA↗

arXiv2026

Black Hole Feedback, Galaxy Quenching and Outflows at Cosmic Dawn: Analysis of the SEEDZ Simulations

Here we analyse the growth and feedback effects of massive black holes (MBHs) in the SEEDZ simulations. The most massive black holes grow to masses of $\sim10^{6}$ M$_\odot$ by $z=12.5$ during short bursts of super-Eddington accretion, sustained over a period of 5-30 Myr. We find that the determining factor that cuts off this initial growth is feedback from the MBH itself, rather than nearby supernovae or exhausting the available gas reservoir. Our simulations show that for the most actively accreting MBHs, feedback completely evacuates the gas from the host halo and ejects it into the inter-galactic medium. Despite implementing a relatively weak feedback model, the energy injected into the gas surrounding the MBH exceeds the binding energy of the halo. These results either indicate that MBH feedback in the early ($Λ$CDM) Universe is much weaker than previously assumed, or that at least some of the high redshift galaxies we currently observe with JWST formed via a two-step process, whereby a MBH initially quenches its host galaxy and later reconstitutes its baryonic reservoir, either through mergers with gas rich galaxies or from accretion from the cosmic web. Moreover, the maximum black hole masses that emerge in SEEDZ are effectively set by a combination of MBH feedback modelling and the binding potential of the host halo. Unless feedback is extremely ineffective at early times (for example if growth is merger dominated rather than accretion dominated or feedback is contained close to the MBH) then the maximum mass of black holes at redshift before 12.5 should not significantly exceed $10^6$ M$_\odot$.

astro-ph.GA↗

arXiv2026

The SEEDZ Simulations: Methodology and First Results on Massive Black Hole Seeding and Early Galaxy Growth

Here we introduce the SEEDZ simulations, a suite of cosmological hydrodynamic simulations exploring the formation and growth of the first massive black holes in the Universe. SEEDZ includes models for Population III star formation, supernovae explosions and the resulting formation of light seed black holes, metal enrichment and subsequent Population II star formation, heavy seed black hole formation, Eddington and super-Eddington accretion schemes as well as black hole feedback. In this paper, we cover the overall methodologies employed and present our current results at $z=15$. Our main result so far is that black holes initially grow faster than their host galaxy, and hence over-massive black holes are a feature of the high-redshift Universe. The fundamental black hole-galaxy relationships we observe at $z = 0$ (especially the M$_{\rm BH}$ - M$_*$ relationship) likely only emerge in more mature galaxies. At high-redshift, that relationship has not yet been established. We find that even at these high redshifts, MBHs can grow from their initial heavy seed mass of $\sim$10$^4$ M$_\odot$ up to 10$^6$ M$_\odot$. At the high end of our MBH masses, our simulated galaxy M$_{\rm BH}$ - M$_*$ relations match the observed high redshift trends i.e. over-massive BHs with M$_{\rm BH}$/M$_{\rm star} \sim 10^{-2}$. This initial set of simulations will continue to run down to $z=10$, where we will perform a comprehensive comparison of simulated MBH number densities and M$_{\rm BH}$ - M$_*$ relations with JWST observations. Further simulations with higher resolution will then follow.

astro-ph.GA↗

arXiv2025

Bridging Machine Learning and Cosmological Simulations: Using Neural Operators to emulate Chemical Evolution

The computational expense of solving non-equilibrium chemistry equations in astrophysical simulations poses a significant challenge, particularly in high-resolution, large-scale cosmological models. In this work, we explore the potential of machine learning, specifically Neural Operators, to emulate the Grackle chemistry solver, which is widely used in cosmological hydrodynamical simulations. Neural Operators offer a mesh-free, data-driven approach to approximate solutions to coupled ordinary differential equations governing chemical evolution, gas cooling, and heating. We construct and train multiple Neural Operator architectures (DeepONet variants) using a dataset derived from cosmological simulations to optimize accuracy and efficiency. Our results demonstrate that the trained models accurately reproduce Grackle's outputs with an average error of less than 0.6 dex in most cases, though deviations increase in highly dynamic chemical environments. Compared to Grackle, the machine learning models provide computational speedups of up to a factor of six in large-scale simulations, highlighting their potential for reducing computational bottlenecks in astrophysical modeling. However, challenges remain, particularly in iterative applications where accumulated errors can lead to numerical instability. Additionally, the performance of these machine learning models is constrained by their need for well-represented training datasets and the limited extrapolation capabilities of deep learning methods. While promising, further development is required for Neural Operator-based emulators to be fully integrated into astrophysical simulations. Future work should focus on improving stability over iterative timesteps and optimizing implementations for hardware acceleration. This study provides an initial step toward the broader adoption of machine learning approaches in astrophysical chemistry solvers.

astro-ph.IM↗