Engineering topics

Scannapieco, Evan

Publications and source records attributed to Scannapieco, Evan.

At least 19 records

DOE OSTI2018

Using Real and Simulated Measurements of the Thermal Sunyaev–Zel’dovich Effect to Constrain Models of AGN Feedback

Energetic feedback from active galactic nuclei (AGNs) is often used in simulations to resolve several outstanding issues in galaxy formation, but its impact is still not fully understood. Here, we derive new constraints on AGN feedback by comparing observations and simulations of the thermal Sunyaev–Zel'dovich (tSZ) effect. We draw on previous observational results that used data from the South Pole Telescope (SPT) and Atacama Cosmology Telescope (ACT) to measure the tSZ signal from ≥10 11 M ⊙ and ≥1 Gyr galaxies at z = 0.5–1.0 (low-z) and z = 1.0–1.5 (high-z). Using the large-scale cosmological hydrodynamical simulations Horizon-AGN and Horizon-NoAGN, which include and omit AGN feedback, we extract simulated tSZ measurements around galaxies equivalent to the observational work. We find that the Horizon-AGN results only differ from the SPT measurements at levels of 0.4σ (low-z) and 0.6σ (high-z), but differ from the ACT measurements by 3.4σ (low-z) and 2.3σ (high-z). The Horizon-noAGN results provide a slightly better fit to the SPT measurements by differing by 0.2σ (low-z) and 0.4σ (high-z), but are a significantly better match to the ACT measurements by differing by only 0.5σ (low-z) and 1.4σ (high-z). Finally, we conclude that, while the lower-mass ($\lesssim$5 × 10 11 M ⊙ ) SPT results allow for the presence AGN feedback energy, the higher-mass ($\gtrsim$ 5 × 10 11 M ⊙ ) ACT results show significantly less energy than predicted in the simulation including AGN feedback, while more closely matching the simulation without AGN feedback, indicating that AGN feedback may be milder than often predicted in simulations.

79 ASTRONOMY AND ASTROPHYSICS↗

DOE OSTI2017

The Production of Cold Gas Within Galaxy Outflows

I present a suite of three-dimensional simulations of the evolution of initially hot material ejected by starburst-driven galaxy outflows. The simulations are conducted in a comoving frame that moves with the material, tracking atomic/ionic cooling, Compton cooling, and dust cooling and destruction. Compton cooling is the most efficient of these processes, while the main role of atomic/ionic cooling is to enhance density inhomogeneities. Dust, on the other hand, has little effect on the outflow evolution, and is rapidly destroyed in all the simulations except for the case with the smallest mass flux. I use the results to construct a simple steady-state model of the observed UV/optical emission from each outflow. The velocity profiles in this case are dominated by geometric effects, and the overall luminosities are extremely strong functions of the properties of the host system, as observed in ultra-luminous infrared galaxies (ULIRGs). Furthermore the luminosities and maximum velocities in several models are consistent with emission-line observations of ULIRGs, although the velocities are significantly greater than observed in absorption-line studies. It may be that absorption line observations of galaxy outflows probe entrained cold material at small radii, while emission-line observations probe cold material condensing from the initially hot medium at larger distances.

79 ASTRONOMY AND ASTROPHYSICS↗

DOE OSTI2017

Numerical simulation of star formation by the bow shock of the centaurus a jet

Recent Hubble Space Telescope (HST) observations of the extragalactic radio source Centaurus A (Cen A) display a young stellar population around the southwest tip of the inner filament 8.5 kpc from the galactic center of Cen A, with ages in the range 1–3 Myr. Crockett et al. have argued that the transverse bow shock of the Cen A jet triggered this star formation as it impacted dense molecular cores of clouds in the filament. To test this hypothesis, we perform three-dimensional numerical simulations of star formation induced by the jet bow shock in the inner filament of Cen A, using a positivity-preserving, weighted, essentially non-oscillatory method to solve the equations of gas dynamics with radiative cooling. We find that star clusters form inside a bow-shocked molecular cloud when the maximum initial density of the cloud is ⩾40 H{sub 2} molecules cm{sup −3}. In a typical molecular cloud of mass 10{sup 6} M{sub ⊙} and diameter 200 pc, approximately 20 star clusters of mass 10{sup 4} M{sub ⊙} are formed, matching the HST images.

79 ASTRONOMY AND ASTROPHYSICS↗

DOE OSTI2017

Searching for fossil evidence of AGN feedback in WISE-selected stripe-82 galaxies by measuring the thermal Sunyaev-Zel'dovich effect with the atacama cosmology telescope

We directly measure the thermal energy of the gas surrounding galaxies through the thermal Sunyaev–Zel’dovich (tSZ) effect. We perform a stacking analysis of microwave background images from the Atacama Cosmology Telescope, around 1179 massive quiescent elliptical galaxies at 0.5 ≤ z ≤ 1.0 (“low-z”) and 3274 galaxies at 1.0 ≤ z ≤ 1.5 (“high-z”), selected using data from the Wide-field Infrared Survey Explorer All-Sky Survey and the Sloan Digital Sky Survey (SDSS) within the SDSS Stripe-82 field. The gas surrounding these galaxies is expected to contain energy from past episodes of active galactic nucleus (AGN) feedback, and after using modeling to subtract undetected contaminants, we detect a tSZ signal at a significance of 0.9σ for our low-z galaxies and 1.8σ for our high-z galaxies. We then include data from the high-frequency Planck bands for a subset of 227 low-z galaxies and 529 high-z galaxies and find low-z and high-z tSZ detections of 1.0σ and 1.5σ, respectively. These results indicate an average thermal heating around these galaxies of (5.6{sub −5.6}{sup +5.9})×10{sup 60} erg for our low-z galaxies and (7.0{sub −4.4}{sup +4.7})×10{sup 60} erg for our high-z galaxies. Based on simple heating models, these results are consistent with gravitational heating without additional heating due to AGN feedback.

79 ASTRONOMY AND ASTROPHYSICS↗

DOE OSTI2017

Following the cosmic evolution of pristine gas. I. Implications for Milky Way halo stars

We make use of a new subgrid model of turbulent mixing to accurately follow the cosmological evolution of the first stars, the mixing of their supernova (SN) ejecta, and the impact on the chemical composition of the Galactic Halo. Using the cosmological adaptive mesh refinement code ramses, we implement a model for the pollution of pristine gas as described in Pan et al. Tracking the metallicity of Pop III stars with metallicities below a critical value allows us to account for the fraction of Z<Z{sub crit} stars formed even in regions in which the gas’s average metallicity is well above Z{sub crit}. We demonstrate that such partially mixed regions account for 0.5 to 0.7 of all Pop III stars formed up to z = 5. Additionally, we track the creation and transport of “primordial metals” (PM) generated by Pop III SNe. These neutron-capture deficient metals are taken up by second-generation stars and likely lead to unique abundance signatures characteristic of carbon-enhanced, metal-poor (CEMP-no) stars. As an illustrative example, we associate primordial metals with abundance ratios used by Keller et al. to explain the source of metals in the star SMSS J031300.36-670839.3, finding good agreement with the observed [Fe/H], [C/H], [O/H], and [Mg/Ca] ratios in CEMP-no Milky Way halo stars. Similar future simulations will aid in further constraining the properties of Pop III stars using CEMP observations, as well as improve predictions of the spatial distribution of Pop III stars, as will be explored by the next generation of ground- and space-based telescopes.

79 ASTRONOMY AND ASTROPHYSICS↗

DOE OSTI2016

Comparing simulations of AGN feedback

We perform adaptive mesh refinement (AMR) and smoothed particle hydrodynamics (SPH) cosmological zoom simulations of a region around a forming galaxy cluster, comparing the ability of the methods to handle successively more complex baryonic physics. In the simplest, non-radiative case, the two methods are in good agreement with each other, but the SPH simulations generate central cores with slightly lower entropies and virial shocks at slightly larger radii, consistent with what has been seen in previous studies. The inclusion of radiative cooling, star formation, and stellar feedback leads to much larger differences between the two methods. Most dramatically, at z=5, rapid cooling in the AMR case moves the accretion shock to well within the virial radius, while this shock remains near the virial radius in the SPH case, due to excess heating, coupled with poorer capturing of the shock width. On the other hand, the addition of feedback from active galactic nuclei (AGNs) to the simulations results in much better agreement between the methods. For our AGN model, both simulations display halo gas entropies of 100 keV cm{sup 2}, similar decrements in the star formation rate, and a drop in the halo baryon content of roughly 30%. This is consistent with the AGN growth being self-regulated, regardless of the numerical method. However, the simulations with AGN feedback continue to differ in aspects that are not self-regulated, such that in SPH a larger volume of gas is impacted by feedback, and the cluster still has a lower entropy central core.

79 ASTRONOMY AND ASTROPHYSICS↗

DOE OSTI2016

THE LAUNCHING OF COLD CLOUDS BY GALAXY OUTFLOWS. II. THE ROLE OF THERMAL CONDUCTION

We explore the impact of electron thermal conduction on the evolution of radiatively cooled cold clouds embedded in flows of hot and fast material as it occurs in outflowing galaxies. Performing a parameter study of three-dimensional adaptive mesh refinement hydrodynamical simulations, we show that electron thermal conduction causes cold clouds to evaporate, but it can also extend their lifetimes by compressing them into dense filaments. We distinguish between low column-density clouds, which are disrupted on very short times, and high-column density clouds with much longer disruption times that are set by a balance between impinging thermal energy and evaporation. We provide fits to the cloud lifetimes and velocities that can be used in galaxy-scale simulations of outflows in which the evolution of individual clouds cannot be modeled with the required resolution. Moreover, we show that the clouds are only accelerated to a small fraction of the ambient velocity because compression by evaporation causes the clouds to present a small cross-section to the ambient flow. This means that either magnetic fields must suppress thermal conduction, or that the cold clouds observed in galaxy outflows are not formed of cold material carried out from the galaxy.

79 ASTRONOMY AND ASTROPHYSICS↗

DOE OSTI2016

Constraining AGN feedback in massive ellipticals with south pole telescope measurements of the thermal Sunyaev–Zel’dovich effect

Energetic feedback due to active galactic nuclei (AGNs) is likely to play an important role in the observed anti-hierarchical trend in the evolution of galaxies, and yet the energy injected into the circumgalactic medium by this process is largely unknown. One promising approach to constrain this feedback is through measurements of spectral distortions in the cosmic microwave background due to the thermal Sunyaev–Zeldovich (tSZ) effect, whose magnitude is directly proportional to the energy input by AGNs. With current instruments, making such measurements requires stacking large numbers of objects to increase signal-to-noise. While one possible target for such stacks is AGNs themselves, these are relatively scarce sources that contain contaminating emission that complicates tSZ measurements. Here we adopt an alternative approach and co-add South Pole Telescope SZ (SPT-SZ) survey data around a large set of massive quiescent elliptical galaxies at z⩾0.5, which are much more numerous and less contaminated than active AGNs, yet are subject to the same feedback processes from the AGNs they hosted in the past. We use data from the Blanco Cosmology Survey and VISTA Hemisphere Survey to create a large catalog of galaxies split up into two redshift bins: one with 3394 galaxies at 0.5⩽z⩽1.0 and one with 924 galaxies at 1.0⩽z⩽1.5, with typical stellar masses of 1.5×10{sup 11}M{sub ⊙}. We then co-add the emission around these galaxies, resulting in a measured tSZ signal at 2.2σ significance for the lower redshift bin and a contaminating signal at 1.1σ for the higher redshift bin. To remove contamination due to dust emission, we use SPT-SZ source counts to model a contaminant source population in both the SPT-SZ bands and Planck high-frequency bands for a subset of 937 galaxies in the low-redshift bin and 240 galaxies in the high-redshift bin. This increases our detection to 3.6σ for low redshifts and 0.9σ for high redshifts. We find the mean angularly integrated Compton-y values to be 2.2{sub −0.7}{sup +0.9}×10{sup −7} Mpc{sup 2} for low redshifts and 1.7{sub −1.8}{sup +2.2}×10{sup −7} Mpc{sup 2} for high redshifts, corresponding to total thermal energies of 7.6{sub −2.3}{sup +3.0}×10{sup 60} erg and 6.0{sub −6.3}{sup +7.7}×10{sup 60} erg, respectively. These numbers are higher than expected from simple theoretical models that do not include AGN feedback, and serve as constraints that can be applied to current simulations of massive galaxy formation.

79 ASTRONOMY AND ASTROPHYSICS↗

DOE OSTI2016

Atomic chemistry in turbulent astrophysical media II: Effect of the redshift zero metagalactic background

Here, we carry out direct numerical simulations of turbulent astrophysical media exposed to the redshift zero metagalactic background. The simulations assume solar composition and explicitly track ionizations, recombinations, and ion-by-ion radiative cooling for hydrogen, helium, carbon, nitrogen, oxygen, neon, sodium, magnesium, silicon, sulfur, calcium, and iron. Each run reaches a global steady state that depends not only on the ionization parameter, $U,$ and mass-weighted average temperature, ${T}_{{\rm{MW}}},$ but also on the one-dimensional turbulent velocity dispersion, ${\sigma }_{{\rm{1D}}}$. We carry out runs that span a grid of models with U ranging from 0 to 10 –1 and ${\sigma }_{{\rm{1D}}}$ ranging from 3.5 to 58 km s –1 , and we vary the product of the mean density and the driving scale of the turbulence, ${nL},$ which determines the average temperature of the medium, from ${nL}={10}^{16}$ to ${nL}={10}^{20}$ cm –2 . The turbulent Mach numbers of our simulations vary from $M\approx 0.5$ for the lowest velocity dispersion cases to $M\approx 20$ for the largest velocity dispersion cases. When $M\lesssim 1,$ turbulent effects are minimal, and the species abundances are reasonably described as those of a uniform photoionized medium at a fixed temperature. On the other hand, when $M\gtrsim 1,$ dynamical simulations such as the ones carried out here are required to accurately predict the species abundances. We gather our results into a set of tables to allow future redshift zero studies of the intergalactic medium to account for turbulent effects.

79 ASTRONOMY AND ASTROPHYSICS↗

DOE OSTI2016

GALAXY OUTFLOWS WITHOUT SUPERNOVAE

High surface density, rapidly star-forming galaxies are observed to have ≈50–100 km s{sup −1} line of sight velocity dispersions, which are much higher than expected from supernova driving alone, but may arise from large-scale gravitational instabilities. Using three-dimensional simulations of local regions of the interstellar medium, we explore the impact of high velocity dispersions that arise from these disk instabilities. Parametrizing disks by their surface densities and epicyclic frequencies, we conduct a series of simulations that probe a broad range of conditions. Turbulence is driven purely horizontally and on large scales, neglecting any energy input from supernovae. We find that such motions lead to strong global outflows in the highly compact disks that were common at high redshifts, but weak or negligible mass loss in the more diffuse disks that are prevalent today. Substantial outflows are generated if the one-dimensional horizontal velocity dispersion exceeds ≈35 km s{sup −1}, as occurs in the dense disks that have star-formation rate (SFR) densities above ≈0.1 M{sub ⊙} yr{sup −1} kpc{sup −2}. These outflows are triggered by a thermal runaway, arising from the inefficient cooling of hot material coupled with successive heating from turbulent driving. Thus, even in the absence of stellar feedback, a critical value of the SFR density for outflow generation can arise due to a turbulent heating instability. This suggests that in strongly self-gravitating disks, outflows may be enhanced by, but need not caused by, energy input from supernovae.

79 ASTRONOMY AND ASTROPHYSICS↗

DOE OSTI2015

THE LAUNCHING OF COLD CLOUDS BY GALAXY OUTFLOWS. I. HYDRODYNAMIC INTERACTIONS WITH RADIATIVE COOLING

To better understand the nature of the multiphase material found in outflowing galaxies, we study the evolution of cold clouds embedded in flows of hot and fast material. Using a suite of adaptive mesh refinement simulations that include radiative cooling, we investigate both cloud mass loss and cloud acceleration under the full range of conditions observed in galaxy outflows. The simulations are designed to track the cloud center of mass, enabling us to study the cloud evolution at long disruption times. For supersonic flows, a Mach cone forms around the cloud, which damps the Kelvin–Helmholtz instability but also establishes a streamwise pressure gradient that stretches the cloud apart. If time is expressed in units of the cloud crushing time, both the cloud lifetime and the cloud acceleration rate are independent of cloud radius, and we find simple scalings for these quantities as a function of the Mach number of the external medium. A resolution study suggests that our simulations accurately describe the evolution of cold clouds in the absence of thermal conduction and magnetic fields, physical processes whose roles will be studied in forthcoming papers.

79 ASTRONOMY AND ASTROPHYSICS↗

DOE OSTI2015

Atomic chemistry in turbulent astrophysical media. I. EFFECT OF Atomic cooling

We carry out direct numerical simulations of turbulent astrophysical media that explicitly track ionizations, recombinations, and species-by-species radiative cooling. The simulations assume solar composition and follows the evolution of hydrogen, helium, carbon, oxygen, sodium, and magnesium, but they do not include the presence of an ionizing background. In this case, the medium reaches a global steady state that is purely a function of the one-dimensional turbulent velocity dispersion, σ{sub 1D}, and the product of the mean density and the driving scale of turbulence, nL. Our simulations span a grid of models with σ{sub 1D} ranging from 6 to 58 km s{sup −1} and nL ranging from 10{sup 16} to 10{sup 20} cm{sup −2}, which correspond to turbulent Mach numbers from M = 0.2 to 10.6. The species abundances are well described by single-temperature estimates whenever M is small, but local equilibrium models can not accurately predict the global equilibrium abundances when M≳1. To allow future studies to account for nonequilibrium effects in turbulent media, we gather our results into a series of tables, which we will extend in the future to encompass a wider range of elements, compositions, and ionizing processes.

79 ASTRONOMY AND ASTROPHYSICS↗

DOE OSTI2014

ALIGNMENT OF THE SCALAR GRADIENT IN EVOLVING MAGNETIC FIELDS

We conduct simulations of turbulent mixing in the presence of a magnetic field, grown by the small-scale dynamo. We show that the scalar gradient field, ∇C, which must be large for diffusion to operate, is strongly biased perpendicular to the magnetic field, B. This is true both early on, when the magnetic field is negligible, and at late times, when the field is strong enough to back react on the flow. This occurs because ∇C increases within the plane of a compressive motion, but B increases perpendicular to it. At late times, the magnetic field resists compression, making it harder for scalar gradients to grow and likely slowing mixing.

79 ASTRONOMY AND ASTROPHYSICS↗

DOE OSTI2014

Mixing in magnetized turbulent media

Turbulent motions are essential to the mixing of entrained fluids and are also capable of amplifying weak initial magnetic fields by small-scale dynamo action. Here we perform a systematic study of turbulent mixing in magnetized media, using three-dimensional magnetohydrodynamic simulations that include a scalar concentration field. We focus on how mixing depends on the magnetic Prandtl number, Pm, from 1 to 4 and the Mach number, M, from 0.3 to 2.4. For all subsonic flows, we find that the velocity power spectrum has a k {sup –5/3} slope in the early kinematic phase, but steepens due to magnetic back reactions as the field saturates. The scalar power spectrum, on the other hand, flattens compared to k {sup –5/3} at late times, consistent with the Obukohov-Corrsin picture of mixing as a cascade process. At higher Mach numbers, the velocity power spectrum also steepens due to the presence of shocks, and the scalar power spectrum again flattens accordingly. Scalar structures are more intermittent than velocity structures in subsonic turbulence, whereas for supersonic turbulence, velocity structures appear more intermittent than the scalars only in the kinematic phase. Independent of the Mach number of the flow, scalar structures are arranged in sheets in both the kinematic and saturated phases of the magnetic field evolution. For subsonic turbulence, scalar dissipation is hindered in the strong magnetic field regions, probably due to Lorentz forces suppressing the buildup of scalar gradients, whereas for supersonic turbulence, scalar dissipation increases monotonically with increasing magnetic field strength. At all Mach numbers, mixing is significantly slowed by the presence of dynamically important small-scale magnetic fields, implying that mixing in the interstellar medium and in galaxy clusters is less efficient than modeled in hydrodynamic simulations.

79 ASTRONOMY AND ASTROPHYSICS↗

DOE OSTI2013

Formation of compact clusters from high resolution hybrid cosmological simulations

The early universe hosted a large population of small dark matter 'minihalos' that were too small to cool and form stars on their own. These existed as static objects around larger galaxies until acted upon by some outside influence. Outflows, which have been observed around a variety of galaxies, can provide this influence in such a way as to collapse, rather than disperse, the minihalo gas. Gray and Scannapieco performed an investigation in which idealized spherically symmetric minihalos were struck by enriched outflows. Here we perform high-resolution cosmological simulations that form realistic minihalos, which we then extract to perform a large suite of simulations of outflow-minihalo interactions including non-equilibrium chemical reactions. In all models, the shocked minihalo forms molecules through non-equilibrium reaction, and then cools to form dense, chemically homogenous clumps of star-forming gas. The formation of these high-redshift clusters may be observable with the next generation of telescopes and the largest of them should survive to the present-day, having properties similar to halo globular clusters.

79 ASTRONOMY AND ASTROPHYSICS↗

DOE OSTI2013

MODELING THE POLLUTION OF PRISTINE GAS IN THE EARLY UNIVERSE

We conduct a comprehensive theoretical and numerical investigation of the pollution of pristine gas in turbulent flows, designed to provide useful new tools for modeling the evolution of the first generation of stars. The properties of such Population III (Pop III) stars are thought to be very different than those of later stellar generations, because cooling is dramatically different in gas with a metallicity below a critical value Z{sub c}, which lies between ∼10{sup –6} and ∼10{sup –3} Z{sub ☉}. The critical value is much smaller than the typical overall average metallicity, , and therefore the mixing efficiency of the pristine gas in the interstellar medium plays a crucial role in determining the transition from Pop III to normal star formation. The small critical value, Z{sub c}, corresponds to the far left tail of the probability distribution function (PDF) of the metal abundance. Based on closure models for the PDF formulation of turbulent mixing, we derive evolution equations for the fraction of gas, P, lying below Z{sub c}, in statistically homogeneous compressible turbulence. Our simulation data show that the evolution of the pristine fraction P can be well approximated by a generalized 'self-convolution' model, which predicts that P-dot = - (n/τ{sub con}) P (1-P{sup 1/n}), where n is a measure of the locality of the mixing or PDF convolution events and the convolution timescale τ{sub con} is determined by the rate at which turbulence stretches the pollutants. Carrying out a suite of numerical simulations with turbulent Mach numbers ranging from M = 0.9 to 6.2, we are able to provide accurate fits to n and τ{sub con} as a function of M, Z{sub c}/(Z), and the length scale, L{sub p}, at which pollutants are added to the flow. For pristine fractions above P = 0.9, mixing occurs only in the regions surrounding blobs of pollutants, such that n = 1. For smaller values of P, n is larger as the mixing process becomes more global. We show how these results can be used to construct one-zone models for the evolution of Pop III stars in a single high-redshift galaxy, as well as subgrid models for tracking the evolution of the first stars in large cosmological numerical simulations.

79 ASTRONOMY AND ASTROPHYSICS↗

DOE OSTI2013

HYBRID COSMOLOGICAL SIMULATIONS WITH STREAM VELOCITIES

In the early universe, substantial relative ''stream'' velocities between the gas and dark matter arise due to radiation pressure and persist after recombination. To assess the impact of these velocities on high-redshift structure formation, we carry out a suite of high-resolution adaptive mesh refinement (AMR) cosmological simulations, which use smoothed particle hydrodynamic data sets as initial conditions, converted using a new tool developed for this work. These simulations resolve structures with masses as small as a few 100 M{sub Sun }, and we focus on the 10{sup 6} M{sub Sun} ''mini-halos'' in which the first stars formed. At z Almost-Equal-To 17, the presence of stream velocities has only a minor effect on the number density of halos below 10{sup 6} M{sub Sun }, but it greatly suppresses gas accretion onto all halos and the dark matter structures around them. Stream velocities lead to significantly lower halo gas fractions, especially for Almost-Equal-To 10{sup 5} M{sub Sun} objects, an effect that is likely to depend on the orientation of a halo's accretion lanes. This reduction in gas density leads to colder, more compact radial profiles, and it substantially delays the redshift of collapse of the largest halos, leading to delayed star formation and possibly delayed reionization. These many differences suggest that future simulations of early cosmological structure formation should include stream velocities to properly predict gas evolution, star formation, and the epoch of reionization.

79 ASTRONOMY AND ASTROPHYSICS↗

DOE OSTI2013

THERMAL AND CHEMICAL EVOLUTION OF COLLAPSING FILAMENTS

Intergalactic filaments form the foundation of the cosmic web that connect galaxies together, and provide an important reservoir of gas for galaxy growth and accretion. Here we present very high resolution two-dimensional simulations of the thermal and chemical evolution of such filaments, making use of a 32 species chemistry network that tracks the evolution of key molecules formed from hydrogen, oxygen, and carbon. We study the evolution of filaments over a wide range of parameters including the initial density, initial temperature, strength of the dissociating UV background, and metallicity. In low-redshift, Z Almost-Equal-To 0.1 Z{sub Sun} filaments, the evolution is determined completely by the initial cooling time. If this is sufficiently short, the center of the filament always collapses to form a dense, cold core containing a substantial fraction of molecules. In high-redshift, Z = 10{sup -3} Z{sub Sun} filaments, the collapse proceeds much more slowly. This is mostly due to the lower initial temperatures, which lead to a much more modest increase in density before the atomic cooling limit is reached, making subsequent molecular cooling much less efficient. Finally, we study how the gravitational potential from a nearby dwarf galaxy affects the collapse of the filament and compare this to NGC 5253, a nearby starbursting dwarf galaxy thought to be fueled by the accretion of filament gas. In contrast to our fiducial case, a substantial density peak forms at the center of the potential. This peak evolves faster than the rest of the filament due to the increased rate at which chemical species form and cooling occurs. We find that we achieve similar accretion rates as NGC 5253 but our two-dimensional simulations do not recover the formation of the giant molecular clouds that are seen in radio observations.

79 ASTRONOMY AND ASTROPHYSICS↗