N-body simulations and analytical calculations of the gravitational collapse in an expanding universe predict that halos should form with a diverging inner density profile, the cusp. There are some observational indications that the dark matter distribution in galaxies might be characterized by a finite core. This `core catastrophe' has prompted a search for alternatives to the CDM cosmogony. It is shown here that the discrepancy between theory and observations can be very naturally resolved within the standard CDM model, provided that gas is not initially smoothly distributed in the dark matter halo, but rather is concentrated in clumps of mass $\geq 0.01 %$ the total mass of the system. Dynamical friction acting on these lumps moving in the background of the dark matter particles, dissipates the clumps orbital energy and deposits it in the dark matter. Using Monte-Carlo simulations, it is shown that the dynamical friction provides a strong enough drag, and that with realistic baryonic mass fractions, the available orbital energy of the clumps is sufficient to heat the halo and turn the primordial cusp into a finite, non-diverging core --- overcoming the competing effect of adiabatic contraction due to gravitational influence of the shrinking baryonic component. Depending on the initial conditions, the total density distribution may either become more or less centrally concentrated. Possible consequences of the proposed mechanism for other problems in the CDM model and for the formation and early evolution of the baryonic component of galaxies are also briefly discussed.Comment: Version to appear in ApJ (with extra figure and extended discussion

Sand and coworkers have measured the central density profile of cluster MS 2137Ϫ23 with gravitational lensing and velocity dispersion and removed the stellar contribution with a reasonable M/L. The resulting dark matter (DM) distribution within was fitted by a density cusp of with , in an apparent Ϫ1 Ϫb r ! 50 h kpc r b p 0.35 contradiction to the cold dark matter prediction of . The disagreement worsens if adiabatic compression of b ∼ 1 the DM by the infalling baryons is considered. Following El-Zant, Shlosman, & Hoffman, we argue that dynamical friction acting on galaxies moving within the DM background counters the effect of adiabatic compression by transfering their orbital energy to the DM, thus heating up and softening the cusp. Using N-body simulations we show that indeed the inner DM distribution flattens (with for a cluster like MS 2137Ϫ23) when the b ≈ 0.35 galaxies spiral inward. We find as a robust result that while the DM distribution becomes core-like, the overall mass distribution preserves its cuspy nature, in agreement with X-ray and lensing observations of clusters.

We present the first identification of large-scale structures ( LSSs) at z < 1:1 in the Cosmic Evolution Survey (COSMOS). The structures are identified from adaptive smoothing of galaxy counts in the pseudo-3D space ( , , z) using the COSMOS photometric redshift catalog. The technique is tested on a simulation including galaxies distributed in model clusters and a field galaxy population-recovering structures on all scales from 1 0 to 20 0 without a priori assumptions for the structure size or density profile. The COSMOS photometric redshift catalog yields a sample of 1:5 ; 10 5 galaxies with redshift accuracy, Áz FWHM /(1 þ z) 0:1 at z < 1:1 down to I AB 25 mag. Using this sample of galaxies, we identify 42 LSSs and clusters. Projected surface-density maps for the structures indicate multiple peaks and internal structure in many of the most massive LSSs. The stellar masses (determined from the galactic SEDs) for the LSSs range from M Ã $ 10 11 up to $3 ; 10 13 M . Five LSSs have total stellar masses exceeding 10 13 M . (Total masses including nonstellar baryons and dark matter are expected to be $50Y100 times greater.) The derived mass function for the LSSs is consistent (within the expected Poisson and cosmic variances) with those derived from optical and X-ray studies at lower redshift. To characterize structure evolution and for comparison with simulations, we compute a new statistic: the area filling factor as a function of the overdensity value compared to the mean at surface overdensity ( f A AE/AE(z) Â Ã ). The observationally determined f A has less than 1% of the surface area (in each redshift slice) with overdensities exceeding 10:1, and evolution to higher overdensities is seen at later epochs (lower z); both characteristics are in good agreement with what we find using similar processing on the Millennium Simulation. Although similar variations in the filling factors as a function of overdensity and redshift are seen in the observations and simulations, we do find that the observed distributions reach higher overdensities than the simulation, perhaps indicating overmerging in the simulation. All of the LSSs show a dramatic preference for earlier SED type galaxies in the denser regions of the structures, independent of redshift. The SED types in the central 1 and 1Y5 Mpc regions of each structure average about one SED type earlier than the mean type at the same redshift, corresponding to a stellar population age difference of $2Y4 Gyr at z ¼ 0:3Y1. We also investigate the evolution of key galactic propertiesmass, luminosity, SED, and star formation rate (SFR)-with redshift and environmental density as derived from overdensities in the full pseudo-3D cube. Both the maturity of the stellar populations and the ''downsizing'' of star formation in galaxies vary strongly with redshift (epoch) and environment. For a very broad mass range (10 10 Y10 12 M ), we find that galaxies in dense environments tend to be older; this is not just restricted to the most massive galaxies.1 Based on observations wi...

The cold dark matter model of structure formation faces apparent problems on galactic scales. Several threads point to excessive halo concentration, including central densities that rise too steeply with decreasing radius. Yet, random fluctuations in the gaseous component can 'heat' the centres of haloes, decreasing their densities. We present a theoretical model deriving this effect from first principles: stochastic variations in the gas density are converted into potential fluctuations that act on the dark matter; the associated force correlation function is calculated and the corresponding stochastic equation solved. Assuming a power law spectrum of fluctuations with maximal and minimal cutoff scales, we derive the velocity dispersion imparted to the halo particles and the relevant relaxation time. We further perform numerical simulations, with fluctuations realised as a Gaussian random field, which confirm the formation of a core within a timescale comparable to that derived analytically. Non-radial collective modes enhance the energy transport process that erases the cusp, though the parametrisations of the analytical model persist. In our model, the dominant contribution to the dynamical coupling driving the cusp-core transformation comes from the largest scale fluctuations. Yet, the efficiency of the transformation is independent of the value of the largest scale and depends weakly (linearly) on the power law exponent; it effectively depends on two parameters: the gas mass fraction and the normalisation of the power spectrum. This suggests that cusp-core transformations observed in hydrodynamic simulations of galaxy formation may be understood and parametrised in simple terms, the physical and numerical complexities of the various implementations notwithstanding.

We investigate a model of galaxy clusters in which the hot intracluster gas is efficiently heated by dynamical friction (DF) of galaxies. We allow for both subsonic and supersonic motions of galaxies and use the gravitational drag formula in a gaseous medium presented by Ostriker (1999). The energy lost by the galaxies is either redistributed locally or into a Gaussian centered on the galaxy. We find that the condition of hydrostatic equilibrium and strict energy balance yields a trivial isothermal solution T_iso, independent of radius, or rising temperature distributions provided T_iso/gamma < T < T_iso, where gamma is the adiabatic index of the gas. The isothermal temperature corresponds to the usual scaling relation between the gas temperatures and the velocity dispersions of galaxies. However the minimal temperature associated with the rising solutions is ~ 0.5 T_vir, larger than that inferred from observations, the radial distribution of galaxy masses notwithstanding. Heating by supersonically moving galaxies cannot suppress thermal instability, although it can lengthen the growth time up to the level comparable to the ages of clusters when Mach number of galaxies is less than about two. We show using numerical hydrodynamic simulations that DF-induced heating is generally unable to produce stable equilibrium cores by evolving arbitrary non-equilibrium clusters, although it can lengthen the cooling time. We conclude that DF-induced heating alone is an unlikely solution to the cooling flow problem, although it can still be an important heat supplier, considerably delaying cooling catastrophe. We discuss other potential consequences of DF of galaxies in galaxy clusters.Comment: Accepted for publication in Ap

We revisit the notion that galaxy motions can efficiently heat intergalactic gas in the central regions of clusters through dynamical friction. For plausible values of the galaxy mass‐to‐light ratio, the heating rate is comparable with the cooling rate due to X‐ray emission. Heating occurs only for supersonic galaxy motions, so the mechanism is self‐regulating: it becomes efficient only when the gas sound speed is smaller than the galaxy velocity dispersion. We illustrate with the Perseus cluster, assuming a stellar mass‐to‐light ratio for galaxies in the very central region with the dark matter contribution becoming comparable with this at some radius rs. For rs≲ 400 kpc ∼ 3rcool– corresponding to an average mass‐to‐light ratio of ∼10 inside that radius – the dynamical‐friction coupling is strong enough to provide the required rate of gas heating. Such values of rs are associated with total mass attached to galaxies that is about 10 per cent of the mass of the cluster – consistent with values inferred from numerical simulations and observations. The measured sound speed is smaller than the galaxy velocity dispersion, as required by this mechanism. With this smaller gas temperature and the observed distribution of galaxies and gas, the energy reservoir in galactic motions is sufficient to sustain the required heating rate for the lifetime of the cluster. The galaxies also lose a smaller amount of energy through dynamical friction to the dark matter implying that non‐cooling‐flow clusters should have flat‐cored dark matter density distributions.

We present measurements of the two-point galaxy angular correlation function w() in the COSMOS field. Independent determinations of w() as a function of magnitude limit are presented for both the Hubble Space Telescope (HST ) ACS catalog and also the ground-based data from Subaru and the Canada-France-Hawaii Telescope. Despite having significantly different masks, these three determinations agree well. At bright magnitudes (I AB < 22), our data generally match very well with existing measurements and with mock catalogs based on the semianalytic galaxy formation calculations of Kitzbichler & White (2007 ) from the Millennium Simulation. The exception is that our result is at the upper end of the expected cosmic variance scatter for > 10 0 , which we attribute to a particularly rich structure known to exist at z ' 0:8. For fainter samples, however, the level of clustering is somewhat higher than reported by some previous studies; in all three catalogs we find w( ¼ 1 0 ) ' 0:014 at a median I AB magnitude of 24. At these very faintest magnitudes, our measurements agree well with the latest determinations from the Canada-France Legacy Survey. This level of clustering is approximately double what is predicted by the semianalytic catalogs (at all angles). The semianalytic results allow an estimate of cosmic variance, which is too small to account for the discrepancy. We therefore conclude that the mean amplitude of clustering at this level is higher than previously estimated.

We investigate the stability properties of trajectories in barred galaxies with mildly triaxial halos by means of Liapunov exponents. This method is perfectly suitable for time-dependent three-dimensional potentials where surfaces of sections and other simple diagnostics are not applicable. We find that when halos are centrally concentrated most trajectories starting near the plane containing the bar become chaotic. The spatial density distribution of these orbits does not match that of the bar, being overextended in and out of the plane compared to the latter. Moreover, the shape of many of the remaining regular trajectories does not match the bar density distribution, being too round. Therefore, time-independent self-consistent solutions are highly unlikely to be found. When the nonrotating nonaxisymmetric perturbation in the potential reaches 10%, almost all trajectories integrated are chaotic and have large Liapunov exponents. No regular trajectories aligned with the bar have been found. Hence, if the evolution of the density figure is directly related to the characteristic timescale of orbital instability, bar dissolution would take place on a timescale of a few dynamical times. The slowly rotating nonaxisymmetric contribution to the potential required for the onset of widespread chaotic behavior is remarkably small. Even a potential axis ratio of 0.99 results in large connected chaotic regions dominating the space of initial conditions. Systems consisting of centrally concentrated axisymmetric halos and stellar bars thus appear to be structurally unstable, and small ($1%) deviations from perfect axisymmetry should result in a bar dissolution on a timescale significantly smaller than the Hubble time. Since halos found in cold dark matter simulations of large-scale structure are both centrally concentrated and triaxial, it is unlikely that stellar bars embedded in such halos would form and survive unless the halos are modified during the formation of the baryonic component.