We study the early stage of the formation of seed supermassive black holes via direct collapse in dark matter (DM) halos, in the cosmological context. We perform high-resolution zoom-in simulations of such collapse at high-$z$. Using the adaptive mes
h refinement code ENZO, we resolve the formation and growth of a DM halo, until its virial temperature reaches $sim 10^4$K, atomic cooling turns on, and collapse ensues. We demonstrate that direct collapse proceeds in two stages, although they are not well separated. The first stage is triggered by the onset of atomic cooling, and leads to rapidly increasing accretion rate with radius, from $dot Msim 0.1,M_odot {rm yr^{-1}}$ at the halo virial radius to few $M_odot ,{rm yr^{-1}}$, around the scale radius $R_{rm s}sim 30$pc of the NFW DM density profile. The second stage of the collapse commences when the gas density takes precedence over the DM density. This is associated with the gas decoupling from the DM gravitational potential. The ensuing collapse approximates that of an isothermal sphere with $dot M ( r )sim $const. We confirm that the gas loses its angular momentum through non-axisymmetric perturbations and gravitational torques, to overcome the centrifugal barrier. During the course of the collapse, this angular momentum transfer process happens on nearly all spatial scales, and the angular momentum vector of the gas varies with position and time. Collapsing gas also exhibits supersonic turbulent motions which suppress gas fragmentation, and are characterized by density PDF consisting of a lognormal part and a high-density power law tail.
We use numerical simulations to explore whether direct collapse can lead to the formation of SMBH seeds at high-z. We follow the evolution of gas within DM halos of 2 x 10^8 Mo and 1 kpc. We adopt cosmological density profiles and j-distributions. Ou
r goal is to understand how the collapsing flow overcomes the centrifugal barrier and whether it is subject to fragmentation. We find that the collapse leads either to a central runaway or to off-center fragmentation. A disk-like configuration is formed inside the centrifugal barrier. For more cuspy DM distribution, the gas collapses more and experiences a bar-like perturbation and a central runaway. We have followed this inflow down to ~10^{-4} pc. The flow remains isothermal and the specific angular momentum is efficiently transferred by gravitational torques in a cascade of nested bars. This cascade supports a self-similar, disk-like collapse. In the collapsing phase, virial supersonic turbulence develops and fragmentation is damped. For larger initial DM cores the timescales become longer. In models with more organized initial rotation, a torus forms and appears to be supported by turbulent motions. The evolution depends on the competition between two timescales, corresponding to the onset of the central runaway and off-center fragmentation. For less organized rotation, the torus is greatly weakened, the central accretion timescale is shortened, and off-center fragmentation is suppressed --- triggering the central runaway even in previously `stable models. The resulting SMBH masses lie in the range 2 x 10^4 Mo - 2 x 10^6 Mo, much higher than for Population III remnants. We argue that the above upper limit appears to be more realistic mass. Corollaries of this model include a possible correlation between SMBH and DM halo masses, and similarity between the SMBH and halo mass functions, at time of formation.
We construct several Milky Way-like galaxy models containing a gas halo (as well as gaseous and stellar disks, a dark matter halo, and a stellar bulge) following either an isothermal or an NFW density profile with varying mass and initial spin. In ad
dition, galactic winds associated with star formation are tested in some of the simulations. We evolve these isolated galaxy models using the GADGET-3 $N$-body/hydrodynamic simulation code, paying particular attention to the effects of the gas halo on the evolution. We find that the evolution of the models is strongly affected by the adopted gas halo component. The model without a gas halo shows an increasing star formation rate (SFR) at the beginning of the simulation for some hundreds of millions of years and then a continuously decreasing rate to the end of the run at 3 Gyr. On the other hand, the SFRs in the models with a gas halo emerge to be either relatively flat throughout the simulations or increasing over a gigayear and then decreasing to the end. The models with the more centrally concentrated NFW gas halo show overall higher SFRs than those with the isothermal gas halo of the equal mass. The gas accretion from the halo onto the disk also occurs more in the models with the NFW gas halo, however, this is shown to take place mostly in the inner part of the disk and not to contribute significantly to the star formation unless the gas halo has very high density at the central part. The rotation of a gas halo is found to make SFR lower in the model. The SFRs in the runs including galactic winds are found to be lower than the same runs but without winds. We conclude that the effects of a hot gaseous halo on the evolution of galaxies are generally too significant to be simply ignored, and expect that more hydrodynamical processes in galaxies could be understood through numerical simulations employing both gas disk and gas halo components.
We study the effect of local stellar radiation and UVB on the physical properties of DLAs and LLSs at z=3 using cosmological SPH simulations. We post-process our simulations with the ART code for radiative transfer of local stellar radiation and UVB.
We find that the DLA and LLS cross sections are significantly reduced by the UVB, whereas the local stellar radiation does not affect them very much except in the low-mass halos. This is because clumpy high-density clouds near young star clusters effectively absorb most of the ionizing photons from young stars. We also find that the UVB model with a simple density threshold for self-shielding effect can reproduce the observed column density distribution function of DLAs and LLSs very well, and we validate this model by direct radiative transfer calculations of stellar radiation and UVB with high angular resolution. We show that, with a self-shielding treatment, the DLAs have an extended distribution around star-forming regions typically on ~ 10-30 kpc scales, and LLSs are surrounding DLAs on ~ 30-60 kpc scales. Our simulations suggest that the median properties of DLA host haloes are: Mh = 2.4*10^10 Msun, SFR = 0.3 Msun/yr, M* = 2.4*10^8 Msun, and Z/Zsun = 0.1. About 30 per cent of DLAs are hosted by haloes having SFR = 1 - 20 Msun/yr, which is the typical SFR range for LBGs. More than half of DLAs are hosted by the LBGs that are fainter than the current observational limit. Our results suggest that fractional contribution to LLSs from lower mass haloes is greater than for DLAs. Therefore the median values of LLS host haloes are somewhat lower with Mh = 9.6*10^9 Msun, SFR = 0.06 Msun/yr, M* = 6.5*10^7 Msun and Z/Zsun = 0.08. About 80 per cent of total LLS cross section are hosted by haloes with SFR < 1 Msun/yr, hence most LLSs are associated with low-mass halos with faint LBGs below the current detection limit.
We compute the infrared (IR) emission from high-redshift galaxies in cosmological smoothed particle hydrodynamics simulations by coupling the output of the simulation with the population synthesis code `GRASIL by Silva et al. Based on the stellar mas
s, metallicity and formation time of each star particle, we estimate the full spectral energy distribution of each star particle from ultraviolet to IR, and compute the luminosity function of simulated galaxies in the Spitzer broadband filters for direct comparison with the available Spitzer observations.
We study the impact of ultraviolet background (UVB) radiation field and the local stellar radiation on the H_I column density distribution f(N_HI) of damped Ly-alpha systems (DLAs) and sub-DLAs at z=3 using cosmological smoothed particle hydrodynamic
s simulations. We find that, in the previous simulations with an optically thin approximation, the UVB was sinking into the H_I cloud too deeply, and therefore we underestimated the f(N_HI) at 19 < log(N_HI) < 21.2 compared to the observations. If the UVB is shut off in the high-density regions with n_gas > 6 x 10^{-3} cm^{-3}, then we reproduce the observed f(N_HI) at z=3 very well. We also investigate the effect of local stellar radiation by post-processing our simulation with a radiative transfer code, and find that the local stellar radiation does not change the f(N_HI) very much. Our results show that the shape of f(N_HI) is determined primarily by the UVB with a much weaker effect by the local stellar radiation and that the optically thin approximation often used in cosmological simulation is inadequate to properly treat the ionization structure of neutral gas in and out of DLAs. Our result also indicates that the DLA gas is closely related to the transition region from optically-thick neutral gas to optically-thin ionized gas within dark matter halos.
We develop a new ``Multicomponent and Variable Velocity (MVV) galactic outflow model for cosmological smoothed particle hydrodynamic (SPH) simulations. The MVV wind model reflects the fact that the wind material can arise from different phases in the
interstellar medium (ISM), and the mass-loading factor in the MVV model is a function of galaxy stellar mass. We find that the simulation with the MVV outflow has the following characteristics: (i) the intergalactic medium (IGM) is hardly heated up, and the mean IGM temperature is almost the same as in the no-wind run; (ii) it has lower cosmic star formation rates (SFRs) compared to the no-wind run, but higher SFRs than the constant velocity wind run; (iii) it roughly agrees with the observed IGM metallicity, and roughly follows the observed evolution of Omega(Civ); (iv) the lower mass galaxies have larger mass-loading factors, and the low-mass end of galaxy stellar mass function is flatter than in the previous simulations. Therefore, the MVV outflow model mildly alleviates the problem of too steep galaxy stellar mass function seen in the previous SPH simulations. In summary, the new MVV outflow model shows reasonable agreement with observations, and gives better results than the constant velocity wind model.
We investigate the effects of the change of cosmological parameters and star formation (SF) models on the cosmic SF history using cosmological smoothed particle hydrodynamics (SPH) simulations based on the cold dark matter (CDM) model. We vary the co
smological parameters within 1-sigma error from the WMAP best-fit parameters, and find that such changes in cosmological parameters mostly affect the amplitude of the cosmic SF history. At high redshift (hereafter high-z), the star formation rate (SFR) is sensitive to the cosmological parameters that control the small-scale power of the primordial power spectrum, while the cosmic matter content becomes important at lower redshifts. We also test two new SF models: 1) the `Pressure model based on the work by Schaye & Dalla Vecchia (2008), and 2) the `Blitz model that takes the effect of molecular hydrogen formation into account, based on the work by Blitz & Rosolowsky (2006). Compared to the previous conventional SF model, the Pressure model reduces the SFR in low-density regions and shows better agreement with the observations of the Kennicutt-Schmidt law. This model also suppresses the early star formation and shifts the peak of the cosmic SF history toward lower redshift, more consistently with the recent observational estimates of cosmic SFR density. The simulations with the new SF model also predict lower global stellar mass densities at high-z, larger populations of low-mass galaxies and a higher gas fraction in high-z galaxies. Our results suggest that there is room left in the model uncertainties to reconcile the discrepancy that was found between the theory and observations of cosmic SF history and stellar mass density. Nevertheless, our simulations still predict higher stellar mass densities than most of the observational estimates.
We investigate the physical mechanisms of tidal heating and satellite disruption in cold dark matter host haloes using N-body simulations based on cosmological initial conditions. We show the importance of resonant shocks and resonant torques with th
e host halo to satellite heating. A resonant shock (torque) couples the radial (tangential) motion of a satellite in its orbit to its phase space. For a satellite on a circular orbit, an ILR-like resonance dominates the heating and this heating results in continuous satellite mass loss. We estimate the requirements for simulations to achieve these dynamics using perturbation theory. Both resonant shocks and resonant torques affect satellites on eccentric orbits. We demonstrate that satellite mass loss is an outside-in process in energy space; a satellites stars and gas are thus protected by their own halo against tidal stripping. We simulate the evolution of a halo similar to the Large Magellanic Cloud (LMC) in our Galactic dark matter halo and conclude that the LMC stars have not yet been stripped. Finally, we present a simple algorithm for estimating the evolution of satellite mass that includes both shock heating and resonant torques.
We present the results of a numerical study on the effects of metal enrichment and metal cooling on galaxy formation and cosmic star formation (SF) history using cosmological hydrodynamic simulations. We find following differences in the simulation w
ith metal cooling when compared to the run without it: (1) the cosmic star formation rate (SFR) is enhanced by about 50 & 20% at z=1 & 3, respectively; (2) the gas mass fraction in galaxies is lower; (3) the total baryonic mass function (gas + star) at z=3 does not differ significantly, but shows an increase in the number of relatively massive galaxies at z=1; (4) the baryonic mass fraction of intergalactic medium (IGM) is reduced at z<3 due to more efficient cooling and gas accretion onto galaxies. Our results suggest that the metal cooling enhances the galaxy growth by two different mechanisms: (1) increase of SF efficiency in the local interstellar medium (ISM), and (2) increase of IGM accretion onto galaxies. The former process is effective throughout most of the cosmic history, while the latter is effective only at z<3 when the IGM is sufficiently enriched by metals owing to feedback.
