The effects of different cooling and heating function models on a simulated analog of NGC300
David Robinson, Camille Avestruz, Nickolay Y. Gnedin, Vadim A. Semenov · 2024
Read The effects of different cooling and heating function models on a simulated analog of NGC300 by David Robinson, Camille Avestruz, Nickolay Y. Gnedin, Vadim A. Semenov (2024) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2412-15324.txt.
Introduction
David Robinson and co-authors test in 2024 if a smart learn stand in for gas heat and cool rates shifts a fake twin of a near galaxy. True rates cost too much to run raw in such sims. Two stand ins face off: a curve fit to a chart of exact runs, and a learn form with sharp marks.
One fake galaxy runs on each. The learn run's thin gas sits hot ter in a set span. A cross curve marks where both runs hold the same gas share. Glow rates for one lit gas split a shade tween runs. Beam true checks are owed next.
Catalog record
| CREATOR | David Robinson, Camille Avestruz, Nickolay Y. Gnedin, Vadim A. Semenov |
|---|---|
| YEAR | 2024 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 37 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2412-15324
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
The effects of different cooling and heating function models on a simulated analog of NGC300 David Robinson, Camille Avestruz, Nickolay Y. Gnedin, Vadim A. Semenov Abstract Gas cooling and heating rates are vital components of hydrodynamic simulations. However, they are computationally expensive to evaluate exactly with chemical networks or photoionization codes. We compare two different approximation schemes for gas cooling and heating in an idealized simulation of an isolated galaxy. One approximation is based on a polynomial interpolation of a table of Cloudy calculations, as is commonly done in galaxy formation simulations. The other approximation scheme uses machine learning for the interpolation instead on an analytic function, with improved accuracy. We compare the temperature-density phase diagrams of gas from each simulation run to assess how much the two simulation runs differ. Gas in the simulation using the machine learning approximation is systematically hotter for low-density gas with − 3 ≲ log ( n b / cm − 3 ) ≲ − 1 -3\lesssim\log{(n_{b}/\mathrm{cm}^{-3})}\lesssim-1 . We find a critical curve in the phase diagram where the two simulations have equal amounts of gas. The phase diagrams differ most strongly at temperatures just above and below this critical curve. We compare C 2 emission rates for collisions with various particles (integrated over the gas distribution function), and find slight differences between the two simulations. Future comparisons with simulations including radiative transfer will be necessary to compare observable quantities like the total C 2 luminosity. keywords Galaxy evolution, computational methods, hydrodynamical simulations 1 Introduction Gas in the interstellar medium (ISM) and circumgalactic medium (CGM) of galaxies exists in multiple interacting phases with different temperatures and densities. These phases include a diffuse hot ionized medium ( Spitzer 1956 ) , clouds of an atomic cold neutral medium ( Ewen and Purcell 1951 ) , a warm ionized medium surrounding the cold clouds ( Hjellming et al. 1969 ) , and very cold molecular clouds ( Cheung et al. 1968 ) . Gas cools and heats through the respective emission and absorption of radiation, providing a critical mechanism for gas to transition between these phases ( McKee and Ostriker 1977 , e.g.) . Star formation (primarily) occurs in cold molecular clouds, and the star formation rate can depend on both the thermal state and turbulent properties of the molecular gas (see Dobbs 2023 and Hennebelle and Grudić 2024 for recent reviews). Gas cooling and heating rates thereby play a vital component of galaxy evolution models, including hydrodynamic simulations. However, there are many different prescriptions in use for calculating the cooling and heating rates as functions of the gas properties and the galactic environment (see Kim et al. 2023 for an overview). The effects of different cooling and heating function models on galaxy formation can be tested using galaxy formation simulations. Simulation comparisons from the Aquila and AGORA collaborations typically focus on the effects of differing stellar feedback models ( Scannapieco et al. 2012 ; Kim et al. 2014 ) . But, differences between varied gas thermodynamic prescriptions in simulations is relatively understudied. Previous work has explored the effect of varying the assumed initial temperature and density profiles of halo gas in semi-analytic models (SAMs), which changes how the gas subsequently cools ( Monaco et al. 2014 ; Hou et al. 2018 ; Hou et al. 2019 , e.g.) . However, each of these works assumes a non-evolving cooling function which does not depend on the ionization state of particles in the gas. Some hydrodynamic simulations now incorporate non-equilibrium chemistry, where the simulations calculate element abundances on-the-fly using a chemical network. We can also calculate gas cooling and heating rates from these non-equilibrium abundances. The assumption of gas photoionization equilibrium yields an alternate set of cooling and heating rates. Richings and Schaye 2016 compare isolated galaxy simulations run with gas cooling calculated from non-equilibrium against those run with equilibrium chemical abundances. Capelo et al. 2018 perform a similar comparison, but with the non-equilibrium metal abundances only used to calculate cooling in cold ( T<10^{4}\,\mathrm{K} ) gas. Both works find that non-equilibrium cooling has little effect on the overall star formation rate of the galaxy, but can change the overall amount of molecular gas. In this paper, we consider the isolated galaxy simulations of an NGC300 analog presented in Semenov et al. 2021 . These simulations were originally used to study the spatial decorrelation between dense molecular gas and sites of recent star formation, and include subgrid models for gas turbulence and turbulence-regulated variable star formation efficiency. We use these simulations to compare two different prescriptions for approximating cooling and heating functions: the interpolation table approach from Gnedin and Hollon 2012 and the machine learning approximation from Robinson et al. 2024 , which yields more accurate cooling and heating function approximations at fixed metallicity. Since the only differences between the two simulation runs are in the gas cooling and heating functions, we focus on the comparison between the resulting temperature-density phase diagrams of the simulated gas. These phase diagrams describe the gas fraction in various phases of the ISM and CGM. We also compare C 2 emission between each simulation run. The luminosity in various ionic emission lines depend on the thermal state (i.e. temperature and density) of the gas. The 157.7\,\mu\mathrm{m} fine-structure line emitted by ionized carbon (C 2 ) is a particularly important case. This line is often used to trace molecular, star-forming gas in galaxies ( Zhao et al. 2024 ; Casavecchia et al. 2024 , e.g.) . The C 2 fine-structure line is also a candidate for line-intensity mapping (LIM) surveys mapping the 3-D structure of the universe. LIM surveys could probe C 2 emission across a wide range of redshifts 3\lesssim z\lesssim 9 , stretching back into the Epoch of Reionization ( Kovetz et al. 2017 ) . We explain our methodology, including the simulation code and how we analyze simulated snapshots, in Section 2 . We compare gas phase diagrams and C 2 emission rates from our two simulation runs in Section 3 , and present conclusions and further discussion in Section 4 . 2 Methodology 2.1 Simulations In this work, we use the isolated galaxy simulation of Semenov et al. 2021 . This simulation is run using Adaptive Refinement Tree (ART), a Eulerian adaptive mesh refinement hydrodynamics code ( Kravtsov 1999 ; Kravtsov et al. 2002 ; Rudd et al. 2008 ) . The specific simulation we use includes a sub-grid prescription for turbulence and a star formation efficiency that depends on the velocity dispersion (including both turbulent and thermal components) through the local virial parameter ( Padoan et al. 2012 ; Semenov et al. 2016 ; Semenov et al. 2021 ) . Different versions of the isolated galaxy simulation incorporate a constant value for star formation efficiency below a maximum virial parameter (or above a minimum density), different contributions to feedback (type II supernovae and/or pre-supernova stellar winds), and iterations with and without radiative transfer ( Semenov et al. 2021 ) . The metallicity-dependent atomic gas cooling and heating functions in the simulation are evaluated using the interpolation table of Gnedin and Hollon 2012 , which interpolates between exact cooling and heating rates evaluated with the photoionization code