Josh Borrow’s research while affiliated with Massachusetts Institute of Technology and other places

We explore how the splashback radius (Rsp) of galaxy clusters, measured using the number density of the subhalo population, changes based on various selection criteria using the IllustrisTNG cosmological galaxy formation simulation. We identify Rsp by extracting the steepest radial gradient in a stacked set of clusters in 0.5 dex wide mass bins, with our clusters having halo masses 1013 ≤ M200, mean/M⊙ ≤ 1015. We apply cuts in subhalo mass, galaxy stellar mass, i-band absolute magnitude and specific star formation rate. We find that, generally, galaxies of increasing mass and luminosity trace smaller measured splashback radii relative to the intrinsic dark matter radius. We also show that quenched galaxies may be used to reliably reconstruct the dark matter splashback radius. This trend is likely due to changes in the galaxy population. Additionally, we are able to reconcile different observational predictions that Rsp based upon galaxy number counts and dark matter may either align or show significant offset (e.g. those using optically- or SZ-selected clusters) through the selection functions that these studies employ. Finally, we demonstrate that changes in Rsp measured through number counts are not due to a simple change in galaxy abundance inside and outside of the cluster.

Cloud-wind interactions are common in the interstellar and circumgalactic media. Many studies have used simulations of such interactions to investigate the effect of particular physical processes, but the impact of the choice of hydrodynamics solver has largely been overlooked. Here we study the cloud-wind interaction, also known as the "blob test", using seven different hydrodynamics solvers: Three flavours of SPH, a moving mesh, adaptive mesh refinement and two meshless schemes. The evolution of masses in dense gas and intermediate-temperature gas, as well as the covering fraction of intermediate-temperature gas, are systematically compared for initial density contrasts of 10 and 100, and four numerical resolutions. To isolate the differences due to the hydrodynamics solvers, we use non-radiative simulations without physical conduction. We find large differences between these methods. SPH methods show slower dispersal of the cloud, particularly for the higher density contrast, but faster convergence, especially for the lower density contrast. Predictions for the intermediate-temperature gas differ particularly strongly, also between non-SPH codes, and converge most slowly. We conclude that the hydrodynamical interaction between a dense cloud and a supersonic wind remains an unsolved problem. Studies aiming to understand the physics or observational signatures of cloud-wind interactions should test the robustness of their results by comparing different hydrodynamics solvers.

We explore how the splashback radius ($R_{\rm sp}$) of galaxy clusters, measured using the number density of the subhalo population, changes based on various selection criteria using the IllustrisTNG cosmological galaxy formation simulation. We identify $R_{\rm sp}$ by extracting the steepest radial gradient in a stacked set of clusters in 0.5 dex wide mass bins, with our clusters having halo masses $10^{13} \leq M_{\rm 200, mean} / {\rm M}_\odot \leq 10^{15}$. We apply cuts in subhalo mass, galaxy stellar mass, i-band absolute magnitude and specific star formation rate. We find that, generally, galaxies of increasing mass and luminosity trace smaller measured splashback radii relative to the intrinsic dark matter radius. We also show that quenched galaxies may be used to reliably reconstruct the dark matter splashback radius. This trend is likely due to changes in the galaxy population. Additionally, we are able to reconcile different observational predictions that $R_{\rm sp}$ based upon galaxy number counts and dark matter may either align or show significant offset (e.g. those using optically- or SZ-selected clusters) through the selection functions that these studies employ. Finally, we demonstrate that changes in $R_{\rm sp}$ measured through number counts are not due to a simple change in galaxy abundance inside and outside of the cluster.

Smoothed Particle Hydrodynamics (SPH) is a ubiquitous numerical method for solving the fluid equations, and is prized for its conservation properties, natural adaptivity, and simplicity. We introduce the Sphenix SPH scheme, which was designed with three key goals in mind: to work well with sub-grid physics modules that inject energy, be highly computationally efficient (both in terms of compute and memory), and to be Lagrangian. Sphenix uses a Density-Energy equation of motion, along with variable artificial viscosity and conduction, including limiters designed to work with common sub-grid models of galaxy formation. In particular, we present and test a novel limiter that prevents conduction across shocks, preventing spurious radiative losses in feedback events. Sphenix is shown to solve many difficult test problems for traditional SPH, including fluid mixing and vorticity conservation, and it is shown to produce convergent behaviour in all tests where this is appropriate. Crucially, we use the same parameters within Sphenix for the various switches throughout, to demonstrate the performance of the scheme as it would be used in production simulations. Sphenix is the new default scheme in the Swift cosmological simulation code and is available open-source.

We present an update to the framework called Simulator of Galaxy Millimeter/submillimeter Emission ( sígame ). sígame derives line emission in the far-infrared (FIR) for galaxies in particle-based cosmological hydrodynamics simulations by applying radiative transfer and physics recipes via a postprocessing step after completion of the simulation. In this version, a new technique is developed to model higher gas densities by parameterizing the probability distribution function (PDF) of the gas density in higher-resolution simulations run with the pseudo-Lagrangian, Voronoi mesh code arepo . The parameterized PDFs are used as a look-up table, and reach higher densities than in previous work. sígame v3 is tested on redshift z = 0 galaxies drawn from the simba cosmological simulation for eight FIR emission lines tracing vastly different phases of the interstellar medium. This version of sígame includes dust radiative transfer with S kirt and high-resolution photoionization models with C loudy , the latter sampled according to the density PDF of the arepo simulations to augment the densities in the cosmological simulation. The quartile distributions of the predicted line luminosities overlap with the observed range for nearby galaxies of similar star formation rate (SFR) for all but two emission lines: [O i ]63 and CO(3–2), which are overestimated by median factors of 1.3 and 1.0 dex, respectively, compared to the observed line–SFR relation of mixed-type galaxies. We attribute the remaining disagreement with observations to the lack of precise attenuation of the interstellar light on sub-grid scales (≲200 pc) and differences in sample selection.

Active galactic nucleus (AGN) feedback from accreting supermassive black holes (SMBHs) is an essential ingredient of galaxy formation simulations. The orbital evolution of SMBHs is affected by dynamical friction that cannot be predicted self-consistently by contemporary simulations of galaxy formation in representative volumes. Instead, such simulations typically use a simple "repositioning" of SMBHs, but the effects of this approach on SMBH and galaxy properties have not yet been investigated systematically. Based on a suite of smoothed particle hydrodynamics simulations with the SWIFT code and a Bondi-Hoyle-Lyttleton subgrid gas accretion model, we investigate the impact of repositioning on SMBH growth and on other baryonic components through AGN feedback. Across at least a factor ~1000 in mass resolution, SMBH repositioning (or an equivalent approach) is a necessary prerequisite for AGN feedback; without it, black hole growth is negligible. Limiting the effective repositioning speed to $\lesssim$ 10 km/s delays the onset of AGN feedback and severely limits its impact on stellar mass growth in the centre of massive galaxies. Repositioning has three direct physical consequences. It promotes SMBH mergers and thus accelerates their initial growth. In addition, it raises the peak density of the ambient gas and reduces the SMBH velocity relative to it, giving a combined boost to the accretion rate that can reach many orders of magnitude. Our results suggest that a more sophisticated and/or better calibrated treatment of SMBH repositioning is a critical step towards more predictive galaxy formation simulations.

The reconstruction of a smooth field onto a fixed grid is a necessary step for direct comparisons to various real-world observations. Projecting SPH data onto a fixed grid becomes challenging in adaptive environments, where some particles may have smoothing lengths far below the grid size, whilst others are resolved by thousands of pixels. In this paper we show how the common approach of treating particles below the grid size as Monte Carlo tracers of the field leads to significant reconstruction errors, and despite good convergence properties is unacceptable for use in synthetic observations in astrophysics. We propose a new method, where particles smaller than the grid size are `blitted' onto the grid using a high-resolution pre-calculated kernel, and those close to the grid size are subsampled, that allows for converged predictions for projected quantities at all grid sizes.

Smoothed particle hydrodynamics (SPH) is a Lagrangian method for solving the fluid equations that is commonplace in astrophysics, prized for its natural adaptivity and stability. The choice of variable to smooth in SPH has been the topic of contention, with smoothed pressure (P-SPH) being introduced to reduce errors at contact discontinuities relative to smoothed density schemes. Smoothed pressure schemes produce excellent results in isolated hydrodynamics tests; in more complex situations however, especially when coupling to the ‘sub-grid’ physics and multiple time-stepping used in many state-of-the-art astrophysics simulations, these schemes produce large force errors that can easily evade detection as they do not manifest as energy non-conservation. Here, two scenarios are evaluated: the injection of energy into the fluid (common for stellar feedback) and radiative cooling. In the former scenario, force and energy conservation errors manifest (of the same order as the injected energy), and in the latter large force errors that change rapidly over a few time-steps lead to instability in the fluid (of the same order as the energy lost to cooling). Potential ways to remedy these issues are explored with solutions generally leading to large increases in computational cost. Schemes using a density-based formulation do not create these instabilities and as such it is recommended that they are preferred over pressure-based solutions when combined with an energy diffusion term to reduce errors at contact discontinuities.

We present an update to the framework called SImulator of GAlaxy Millimeter/submillimeter Emission (S\'IGAME). S\'IGAME derives line emission in the far-infrared (FIR) for galaxies in particle-based cosmological hydrodynamics simulations by applying radiative transfer and physics recipes via a post-processing step after completion of the simulation. In this version, a new technique is developed to model higher gas densities by parametrizing the gas density probability distribution function (PDF) in higher resolution simulations for use as a look-up table, allowing for more adaptive PDFs than in previous work. S\'IGAME v3 is tested on redshift z = 0 galaxies drawn from the SIMBA cosmological simulation for eight FIR emission lines tracing vastly different interstellar medium phases. Including dust radiative transfer with SKIRT and high resolution photo-ionization models with Cloudy, this new method is able to self-consistently reproduce observed relations between line luminosity and star formation rate in all cases, except for [NII]122, [NII]205 and [OI]63, the luminosities of which are overestimated by median factors of 1.6, 1.2 and 1.2 dex, respectively. We attribute the remaining disagreement with observations to the lack of precise attenuation of the interstellar light on subgrid scales (<200 pc).

Smoothed Particle Hydrodynamics (SPH) is a ubiquitous numerical method for solving the fluid equations, and is prized for its conservation properties, natural adaptivity, and simplicity. We introduce the Sphenix SPH scheme, which was designed with three key goals in mind: to work well with sub-grid physics modules that inject energy, be highly computationally efficient (both in terms of compute and memory), and to be Lagrangian. Sphenix uses a Density-Energy equation of motion, along with variable artificial viscosity and conduction, including limiters designed to work with common sub-grid models of galaxy formation. In particular, we present and test a novel limiter that prevents conduction across shocks, preventing spurious radiative losses in feedback events. Sphenix is shown to solve many difficult test problems for traditional SPH, including fluid mixing and vorticity conservation, and it is shown to produce convergent behaviour in all tests where this is appropriate. Crucially, we use the same parameters within Sphenix for the various switches throughout, to demonstrate the performance of the scheme as it would be used in production simulations.