arXiv subjects
Publications and source records attributed to Joel R. Primack.
This paper has been withdrawn because the models on which it was based have undergone significant changes and improvements. A new paper with the same title, based on the improved models, is accepted for publication in MNRAS and is available at astro-ph/0006364. We have withdrawn this paper to avoid confusion.
We study the distribution functions of mass and circular velocity for dark matter halos in N-body simulations of the $Λ$CDM cosmology, addressing redshift and environmental dependence. The dynamical range enables us to resolve subhalos and distinguish them from "distinct" halos. The mass function is compared to analytic models, and is used to derive the more observationally relevant circular velocity function. The distribution functions in the velocity range 100--500 km/s are well fit by a power-law with two parameters, slope and amplitude. We present the parameter dependence on redshift and provide useful fitting formulae. The amplitudes of the mass functions decrease with z, but, contrary to naive expectation, the comoving density of halos of a fixed velocity ~200 km/s actually increases out to z=5. This is because high-z halos are denser, so a fixed velocity corresponds to a smaller mass. The slope of the velocity function at z=0 is as steep as ~ -4, and the mass and velocity functions of distinct halos steepen with increasing z, while the functions of subhalos do not steepen with $z$, and become even flatter at z>2. A simple observable prediction is that the slope of the velocity function of isolated galaxies is steeper than that of galaxies in groups by as much as unity, reflecting the density biasing of high-velocity halos. We confirm that the Press-Schechter approximation typically overestimates the halo mass function by a factor of ~2, while modified approximations provide improved predictions.
We present a method for constraining the evolution of the galaxy luminosity-velocity (LV) relation in hierarchical scenarios of structure formation. The comoving number density of dark-matter halos with circular velocity of 200 km/s is predicted in favored CDM cosmologies to be nearly constant over the redshift range 0<z<5. Any observed evolution in the density of bright galaxies implies in turn a corresponding evolution in the LV relation. We consider several possible forms of evolution for the zero-point of the LV relation and predict the corresponding evolution in galaxy number density. The Hubble Deep Field suggests a large deficit of bright (M_V < -19) galaxies at 1.4 < z < 2. If taken at face value, this implies a dimming of the LV zero-point by roughly 2 magnitudes. Deep, wide-field, near-IR selected surveys will provide more secure measurements to compare with our predictions.
We investigate the properties of damped Lyman-alpha systems (DLAS) in semi-analytic models, focusing on whether the models can reproduce the kinematic properties of low-ionization metal lines described by Prochaska & Wolfe (1997b, 1998). We explore a variety of approaches for modelling the radial distribution of the cold neutral gas associated with the galaxies in our models, and find that our results are very sensitive to this ingredient. If we use an approach based on Fall & Efstathiou (1980), in which the sizes of the discs are determined by conservation of angular momentum, we find that the majority of the DLAS correspond to a single galactic disc. These models generically fail to reproduce the observed distribution of velocity widths. In alternative models in which the gas discs are considerably more extended, a significant fraction of DLAS arise from lines of sight intersecting multiple gas discs in a common halo. These models produce kinematics that fit the observational data, and also seem to agree well with the results of recent hydrodynamical simulations. Thus we conclude that Cold Dark Matter based models of galaxy formation can be reconciled with the kinematic data, but only at the expense of the standard assumption that DLAS are produced by rotationally supported gas discs whose sizes are determined by conservation of angular momentum. We suggest that the distribution of cold gas at high redshift may be dominated by another process, such as tidal streaming due to mergers.
Previously we have shown (Maller et al, 1998) that the kinematics of Damped Lyman Alpha Systems (DLAS) as measured by Prochaska and Wolfe (1998) can be reproduced in a multiple disk model (MDM) if the gaseous disks are of sufficient radial extent. Here we discuss this model's predictions for the relationship between DLAS and Lyman break galaxies (LBGs), which we here take to be objects at z~3 brighter than R=25.5. We expect that future observations of the correlations between DLAS and LBGs will provide a new data set able to discriminate between different theoretical models of the DLAS. Djorgovski (1997) has already detected a few optical counterparts and more studies are underway.
Previously we have shown that it is possible to account for the kinematic properties of damped Lyman alpha systems (DLAS) in the context of semi-analytic models. In these models, hierarchical structure formation is approximated by constructing a merger tree for each dark matter halo. A natural consequence is that every virialized halo may contain not only a central galaxy, but also a number of satellite galaxies as determined by its merging history. Thus the kinematics of the DLAS arise from the combined effects of the internal rotation of gas disks and the motions between gas disks within a common halo. Here we investigate the sensitivity of this model to some of the assumptions made previously, including the modeling of satellite dynamics, the scale height of the gas, and the cosmology.
The cosmological parameters that I will emphasize are the Hubble parameter $H_0 \equiv 100 h$ km s$^{-1}$ Mpc$^{-1}$, the age of the universe $t_0$, the average matter density $Ω_m$, the baryonic matter density $Ω_b$, the neutrino density $Ω_ν$, and the cosmological constant $Ω_Λ$. The evidence currently favors $t_0 \approx 13$ Gyr, $h \approx 0.65$, $Ω_m \approx 0.3$, $Ω_Λ\approx 0.7$.
The degeneracy between the disk and the dark matter contribution to galaxy rotation curves remains an important uncertainty in our understanding of disk galaxies. Here we discuss a new method for breaking this degeneracy using gravitational lensing by spiral galaxies, and apply this method to the spiral lens B1600+434 as an example. The combined image and lens photometry constraints allow models for B1600+434 with either a nearly singular dark matter halo, or a halo with a sizable core. A maximum disk model is ruled out with high confidence. Further information, such as the circular velocity of this galaxy, will help break the degeneracies. Future studies of spiral galaxy lenses will be able to determine the relative contribution of disk, bulge, and halo to the mass in the inner parts of galaxies.
We investigate the clustering properties of high-redshift galaxies within three competing scenarios for assigning luminous galaxies to dark matter halos from N-body simulations: a one galaxy per massive halo model, a quiescent star formation model, and a collisional starburst model. We compare these models to observations of Lyman-Break galaxies at z~3$ With current data and the simple statistic used here, one cannot rule out any of these models, but we see potential for finding distinguishing features using statistics that are sensitive to the tails of the distribution, and statistics based on the number of multiple galaxies per halo, which we explore in an ongoing study.
We present a galaxy circular velocity function, Psi(log v), derived from existing luminosity functions and luminosity-velocity relations. Such a velocity function is desirable for several reasons. First, it enables an objective comparison of luminosity functions obtained in different bands and for different galaxy morphologies, with a statistical correction for dust extinction. In addition, the velocity function simplifies comparison of observations with predictions from high-resolution cosmological N-body simulations. We derive velocity functions from five different data sets and find rough agreement among them, but about a factor of 2 variation in amplitude. These velocity functions are then compared with N-body simulations of a LCDM model (corrected for baryonic infall) in order to demonstrate both the utility and current limitations of this approach. The number density of dark matter halos and the slope of the velocity function near v_*, the circular velocity corresponding to an ~L_* spiral galaxy, are found to be comparable to that of observed galaxies. The primary sources of uncertainty in construction of Psi(log v) from observations and N-body simulations are discussed and explanations are suggected to account for these discrepancies.
Using semi-analytic models of galaxy formation, we investigate galaxy properties such as the Tully-Fisher relation, the B and K-band luminosity functions, cold gas contents, sizes, metallicities, and colours, and compare our results with observations of local galaxies. We investigate several different recipes for star formation and supernova feedback, including choices that are similar to the treatment in Kauffmann, White & Guiderdoni (1993) and Cole et al. (1994) as well as some new recipes. We obtain good agreement with all of the key local observations mentioned above. In particular, in our best models, we simultaneously produce good agreement with both the observed B and K-band luminosity functions and the I-band Tully-Fisher relation. Improved cooling and supernova feedback modelling, inclusion of dust extinction, and an improved Press-Schechter model all contribute to this success. We present results for several variants of the CDM family of cosmologies, and find that models with values of $Ω_0 \simeq 0.3$--0.5 give the best agreement with observations.
Lyman break galaxies (LBGs) at $z \sim 3$ exhibit number densities and clustering similar to local $L_*$ galaxies; however, their star formation rates (SFRs) are much higher. We explore the scenario in which LBGs are starburst galaxies triggered by collisions, and thus provide an explanation for these key properties. The relative paucity of starburst galaxies at low redshift can be attributed to a much higher collision rate in the past. We use high-resolution cosmological $N$-body simulations and a hierarchical halo finder to estimate the galaxy collision rate as a function of time in the popular $Λ$CDM cosmological model. We find that bright collisional starbursts are frequent enough to account for most of the high-$z$ ($2.5-4.5$) LBGs. Moreover, many of the objects are of relatively small mass, but they cluster about large-mass halos. They therefore exhibit strong clustering, similar to that observed and stronger than that of the relevant massive halos.
We present here the extragalactic background light (EBL) predicted by semi-analytic models of galaxy formation, and show how measurements of the absorption of gamma rays of $\sim$ TeV energies via pair production on the EBL can probe cosmology and the formation of galaxies. Semi-analytic models permit a physical treatment of the key processes of galaxy formation -- including gravitational collapse and merging of dark matter halos, gas cooling and dissipation, star formation, supernova feedback and metal production -- and have been shown to reproduce key observations at low and high redshift. Using this approach, we investigate the consequences of variations in input assumptions such as the stellar initial mass function and the underlying cosmology. We conclude that observational studies of the absorption of $\sim 10^{-2}-10^{2}$ TeV gamma rays will help to constrain the star formation history of the universe, and the nature and extent of the extinction of starlight due to dust and reradiation of the absorbed energy at infrared wavelengths.
We review our analysis of the clustering properties of ``Lyman-break'' galaxies (LBGs) at redshift z~3, previously discussed in Wechsler et al (1998). We examine the likelihood of spikes found by Steidel et al (1998) in the redshift distribution of LBGs, within a suite of models for the evolution of structure in the Universe. Using high-resolution dissipationless N-body simulations, we analyze deep pencil-beam surveys from these models in the same way that they are actually observed, identifying LBGs with the most massive dark matter halos. We find that all the models (with SCDM as a marginal exception) have a substantial probability of producing spikes similar to those observed, because the massive halos are much more clumped than the underlying matter -- i.e., they are biased. Therefore, the likelihood of such a spike is not a good discriminator among these models. The LBG correlation functions are less steep than galaxies today (gamma~1.4), but show similar or slightly longer correlation lengths. We have extened this analysis and include a preliminary comparison to the new data presented in Adelberger et al (1998). We also discuss work in progress, in which we use semi-analytic models to identify Lyman-break galaxies within dark-matter halos.
The velocity profiles of weak metal absorption lines can be used to observationally probe the kinematic state of gas in damped Lyman-alpha systems. Prochaska and Wolfe have argued that the flat distribution of velocity widths combined with the asymmetric line profiles indicate that the DLAS are disks with large rotation velocities. An alternative explanation has been proposed by Haehnelt, Steinmetz, and Rauch, in which the observed large velocity widths and asymmetric profiles can be produced by lines of sight passing through two or more clumps each having relatively small internal velocity dispersions. We investigate the plausibility of this scenario in the context of semi-analytic models based on hierarchical merging trees and including simple treatments of gas dynamics, star formation, supernova feedback, and chemical evolution. We find that all the observed properties of the metal-line systems including the distribution of velocities and the asymmetric profiles, can be reproduced by lines of sight passing through sub-clumps that are bound within larger virialized dark matter halos. In order to produce enough multiple hits, we find that the cold gas must be considerably more extended than the optical radius of the proto-galaxies, perhaps even beyond the tidal radius of the sub-halo. This could occur due to tidal stripping or supernova-driven outflows.
The recent atmospheric neutrino data from Super-Kamiokande provide strong evidence of neutrino oscillations and therefore of non-zero neutrino mass. These data imply a lower limit on the hot dark matter (i.e., light neutrino) contribution to the cosmological density $Ω_ν>~ 0.001$ --- almost as much as that of all the stars in the universe --- and permit higher $Ω_ν$. The ``standard'' COBE-normalized critical-matter-density (i.e., $Ω_m=1$) Cold Dark Matter (CDM) model has too much power on small scales. But adding to CDM neutrinos with mass of about 5 eV, corresponding to $Ω_ν\approx 0.2$, results in a much improved fit to data on the nearby galaxy and cluster distribution. Indeed, the resulting Cold + Hot Dark Matter (CHDM) cosmological model is arguably the most successful $Ω_m=1$ model for structure formation. However, other recent data has begun to make the case for $Ω_m <~ 0.6$ fairly convincing. In light of all this new data, we reconsider whether cosmology still provides evidence favoring neutrino mass of a few eV in flat models with cosmological constant $Ω_Λ= 1 - Ω_m$. We find that the possible improvement of the low-$Ω_m$ flat (LCDM) cosmological models with the addition of light neutrinos appears to be rather limited.
Observations now probe the star formation history of the Universe back to a redshift of $z\sim5$. We investigate whether the predictions of semi-analytic models of galaxy formation based on hierarchical Cold Dark Matter (CDM) type models are in agreement with these direct observations and also with the ``fossil'' evidence contained in constraints on the ages of present day early-type galaxies. Previous models predicted that the star formation rate density falls off rather steeply at $z\ga 2$, and correspondingly that the majority of the stars in the Universe formed at relatively low redshift. We investigate the effect of including a bursting mode of star formation, assuming that galaxy-galaxy mergers trigger starbursts and using the merger rate that arises naturally in the CDM merging hierarchy. The resulting starbursts substantially increase the global star formation rate at high redshift, leading to predictions that are in good agreement with the star formation rate density at $z\sim3$ obtained from sub-millimeter observations (SCUBA) and optical/UV estimates after correction for dust extinction. The mass of stars formed at $z \ge 3$ is correspondingly in better agreement with the fossil evidence. We also investigate complementary global quantities such as the mass of cold gas and the average metallicity of cold gas as a function of redshift, and the integrated extra-galactic background light. We argue that these observations, taken together, provide strong constraints on the star formation history of the Universe, and that hierarchical models of the CDM type are in reasonable agreement with these observations when starbursts are included.
We review the basic arguments for the likelihood of non-Gaussian density perturbations in inflation models with primordial black hole (PBH) production. We discuss our derived distributions of field fluctuations and their implications, specifically commenting on the fine-tuning problem. We also discuss how the derived distributions may be affected when linked to metric perturbations. While linking the metric perturbations to field fluctuations in a nonlinear way may be important for determining exact probability distributions, the correct mapping is not self-evident. The calculation of P. Ivanov, which yields skew positive distribution, is based on an ansatz for the behavior of the nonlinear metric perturbation. We note that the ``natural'' generalization of the gauge-invariant formalism favored by Bond and Salopek yields an effective linear link between the distribution of field fluctuations and metric perturbations during inflation.