Post-Processed Cosmic Dust Density in TNG300

Methods We apply a simple post-processed dust model to gas cells from the TNG300 cosmological, magnetohydrodynamical simulation. TNG300 has a volume of ~3003 Mpc3 and a baryonic mass resolution of ~107 M⊙. From the simulation, we obtain snapshots of the gas cells, including information about their spatial distribution, mass, metallicity, neutral fractions, and internal energy. The TNG simulations do not track information related to dust so we “paint” dust masses onto each gas cell following Millard et al. (2020): 𝑀𝑀𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑 =𝜀𝜀𝑑𝑑𝑓𝑓 𝑁𝑁𝑍𝑍𝑀𝑀𝑔𝑔𝑔𝑔𝑑𝑑, where εd is the fraction of neutral metals locked up in solids, fN is a cell’s neutral fraction, Z is the cells’ metal mass fraction, and Mgas is the total gas cell mass. Mgas and Z are obtained directly from the simulation data. For cells with active star formation (which are treated as multiphase), fN is assumed to be the fraction of the cell in the dense star-forming phase; otherwise, fN is taken to be the neutral hydrogen abundance provided in the simulation output. The evolution of neutral hydrogen is given in Figure 1. εd, also known as the dust-to-gas ratio, is a poorly constrained free parameter. Estimates of this parameter—and its possible dependences on redshift, location, and abundances—vary greatly in the literature. In Figure 2, we show results for the cosmic dust density with two different choices for εd: a constant value of 0.5 ± 0.1 at all redshifts or replace εd at z ≥ 2 with the relationship from Vogelsberger et al. (2020): εd(z) = (z/2)-1.92. Authors: Bryanne McDonough1 and Jacqueline McCleary1 1Department of Physics, Northeastern University Abstract Dust in the Universe reddens and dims light from distant objects, but the amount of this dust and its evolution with cosmic time is poorly constrained. Here, we present results for the evolution with redshift of cosmic dust density from a simple post-processed dust model applied to the TNG300 simulation. The neutral gas fraction is an important component of the dust model, so we also show the cosmic evolution of neutral gas density. We find that neutral gas density in TNG300 is higher than expected at low redshifts, but generally consistent with observations at z>0.5. Cosmic dust density inferred from TNG300 is in general agreement with observations, despite poorly constrained normalizations for both simulated and observational results. Discussion The dust model we have adopted here is necessarily simplistic to be feasible for computation over the ~14.5 billion particles in any one snapshot of TNG300. Various steps could be taken to improve the accuracy of this model including accounting for relative abundances of the various ionization states for each species and/or adopting a value for εd that is a function of abundance, density, redshift, and/or location. However, despite our relatively simple model and poorly constrained dust-to-gas ratio, Figure 2 shows that the model returns a cosmic dust density evolution that is generally consistent with observations in both normalization and shape. Dust density peaks at z~1–2, around or just after ‘cosmic noon’ when the formation of stars peaked. Given the minimal evolution of neutral gas density (Figure 1), we can infer that the evolution of dust content is driven by the evolution of gas metallicity, as stars return metals to gas via stellar winds. The next goal of this project is to investigate the large-scale distribution of dust around various types of galaxies, out to tens of megaparsecs, to compare to the results of McCleary et al. (2025). Figure 2 References and Acknowledgements Chiang, Makikya, & Ménard (2025), pre-print, arXiv:2504.05384 McCleary, Huff, Bartlett, & Hensley (2025), pre-print, arXiv:2503.04098 Peroux & Howk (2020), ARAA. 58, 363 Popping, Pillepich, Somerville, et al. (2019), ApJ, 882(2), 137 Villaescusa-Navarro, Genel, Castorina, et al. (2018), ApJ 866(2), 135 Vogelsberger, Nelson, Pillepich, et al. (2020), MNRAS, 492(4), 5167 Yates, Peroux, & Nelson (2021), MNRAS, 508, 3535 We thank the TNG team for providing the simulation data and this poster’s background image. Computations for this project were performed on Northeastern’s Explorer HPC cluster. This work was supported by Northeastern University’s Future Faculty Postdoctoral Fellowship program. Evolution of cosmic dust density in TNG300, shown for two different dust-to-gas (εd) models: a constant value of 0.5 ± 0.1 (pink squares) and a value that changes with redshift at z ≥ 2 (purple squares) according to the evolution identified by Vogelsberger et al. 2020 (V20). We compare to observational results that infer the presence of dust from neutral gas depletion (yellow; Peroux & Howk, 2020) and the cosmic farinfrared background tomography (green; Chiang et al., 2025). The gray dashed line represents the upper limit of dust density identified by Chiang et al. (2025). Figure 1 Evolution of neutral hydrogen density in the TNG300 simulation (purple squares) as a function of redshift, compared to neutral hydrogen densities from a compilation of observations from Peroux & Howk 2020 (PH20, grey and blue) and results from post-processed HI models applied to TNG100 (pink and green) reported in Yates et al. (2021). Comparisons between different results are challenging, because our model does not distinguish between atomic and molecular hydrogen, but observations generally measure one or the other. The blue line (with shaded region giving the uncertainty) is a combination of PH20’s H2 density with their fit to the HI density evolution. Visualizing the column density of atomic hydrogen in a massive (~1015.1 M⊙) TNG300 halo. The black circle is drawn at 1 R200. This was constructed using only particles identified as members of this friends-of-friends halo. The neutral hydrogen is less centrally concentrated than the total gas distribution and is preferentially located in the ISM of infalling galaxies. Constructed with the TNG halo visualization tool.