The angular momentum content of dwarf galaxies: new challenges for the theory of galaxy formation

We compute the specific angular momentum distributions for a sample of low-mass disc galaxies observed by Swaters. We compare these distributions to those of dark matter haloes obtained by Bullock et al. from high-resolution N -body simulations of structure formation in a ΛCDM universe. We find that although the disc mass fractions are significantly smaller than the universal baryon fraction, the total specific angular momenta of the discs are in good agreement with those of dark matter haloes. This suggests that discs form out of only a small fraction of the available baryons, but yet manage to draw most of the available angular momentum. In addition we find that the angular momentum distributions of discs are clearly distinct from those of the dark matter; discs lack predominantly both low and high specific angular momenta. Understanding these findings in terms of a coherent picture for disc formation is challenging. Cooling, feedback and stripping, which are the main mechanisms to explain the small disc mass fractions found, seem unable to simultaneously explain the angular momentum distributions of the discs. In fact, it seems that the baryons that make up the discs must have been born out of angular momentum distributions that are clearly distinct from those of ΛCDM haloes. However, the dark and baryonic mass components experience the same tidal forces, and it is therefore expected that they should have similar angular momentum distributions. Therefore, understanding the angular momentum content of disc galaxies remains an important challenge for our picture of galaxy formation.     

The cosmological dependence of galactic specific angular momenta

Hydrodynamical simulations of galaxy formation in spatially flat cold dark matter (CDM) cosmologies with and without a cosmological constant (Λ) are described. A simple star formation algorithm is employed and radiative cooling is allowed only after redshift z =1 so that enough hot gas is available to form large, rapidly rotating stellar discs if angular momentum is approximately conserved during collapse. The specific angular momenta of the final galaxies are found to be sensitive to the assumed background cosmology. This dependence arises from the different angular momenta contained in the haloes at the epoch when the gas begins to collapse and the inhomogeneity of the subsequent halo evolution. In the Λ-dominated cosmology, the ratio of stellar specific angular momentum to that of the dark matter halo (measured at the virial radius) has a median value of ∼0.24 at z =0. The corresponding quantity for the Λ=0 cosmology is over three times lower. It is concluded that the observed frequency and angular momenta of disc galaxies pose significant problems for spatially flat CDM models with Λ=0 but may be consistent with a Λ-dominated CDM universe.     

The origin of the density distribution of disc galaxies: a new problem for the standard model of disc formation

We present new models for the formation of disc galaxies that improve upon previous models by following the detailed accretion and cooling of the baryonic mass, and by using realistic distributions of specific angular momentum. Under the assumption of detailed angular momentum conservation, the discs that form have density distributions that are more centrally concentrated than an exponential. We examine the influence of star formation, bulge formation, and feedback on the outcome of the surface brightness distributions of the stars. Low angular momentum haloes yield disc galaxies with a significant bulge component and with a stellar disc that is close to exponential, in good agreement with observations. High angular momentum haloes, on the other hand, produce stellar discs that are much more concentrated than an exponential, in clear conflict with observations. At large radii, the models reveal distinct truncation radii in both the stars and the cold gas. The stellar truncation radii result from our implementation of star formation threshold densities, and are in excellent agreement with observations. The truncation radii in the density distribution of the cold gas reflect the maximum specific angular momentum of the gas that has cooled. We find that these truncation radii occur at H  i surface densities of roughly 1 M_⊙ pc⁻², in conflict with observations. We examine various modifications to our models, including feedback, viscosity, and dark matter haloes with constant-density cores, but show that the models consistently fail to produce bulge less discs with exponential surface brightness profiles. This signals a new problem for the standard model of disc formation: if the baryonic component of the protogalaxies out of which disc galaxies form has the same angular momentum distribution as the dark matter, discs are too compact.     

The formation of galaxy discs

Galaxy disc formation must incorporate the multiphase nature of the interstellar medium. The resulting two-phase structure is generated and maintained by gravitational instability and supernova energy input, which yield a source of turbulent viscosity that is able to compete effectively in the protodisc phase with early angular momentum loss of the baryonic component via dynamical friction in the dark halo. Provided that star formation occurs on the viscous drag time-scale, this mechanism provides a means of accounting for disc sizes and radial profiles. The star formation feedback is self-regulated by turbulent gas pressure limited percolation of the supernova remnant heated hot phase, but can run away in gas-rich protodiscs to generate compact starbursts. A simple analytic model is derived for a Schmidt-like global star formation law in terms of the cold gas volume density.     

Inferring dark halo structure from observed scaling laws of late-type galaxies and LSBs

We re-examine the Fall & Efstathiou scenario for galaxy formation, including the dark halo gravitational reaction to the formation of the baryon disc, as well as continuous variations in the intrinsic halo density profile. The recently published rotation curves of low surface brightness (LSB) and dwarf galaxies together with previously known scaling relations provide sufficient information on the present-day structure of late-type disc galaxies to invert the problem. By requiring that the models reproduce all the observational restrictions we can fully constrain the initial conditions of galaxy formation, with a minimum of assumptions, in particular without the need to specify a cold dark matter (CDM) halo profile. This allows one to solve for all the initial conditions, in terms of the halo density profile, the baryon fraction and the total angular momentum. We find that a unique initial halo shape is sufficient to accurately reproduce the rotation curves of both LSB and normal late-type spiral galaxies. This unique halo profile differs substantially from that found in standard CDM models. A galactic baryon fraction of 0.065 is found. The initial value of the dimensionless angular momentum is seen to be the principal discriminator between the galaxy classes we examine. The present-day scalings between structural parameters are seen to originate in the initial conditions.     

Bulges versus discs: the evolution of angular momentum in cosmological simulations of galaxy formation

We investigate the evolution of angular momentum in simulations of galaxy formation in a cold dark matter universe. We analyse two model galaxies generated in the N -body/hydrodynamic simulations of Okamoto et al. Starting from identical initial conditions, but using different assumptions for the baryonic physics, one of the simulations produced a bulge-dominated galaxy and the other one a disc-dominated galaxy. The main difference is the treatment of star formation and feedback, both of which were designed to be more efficient in the disc-dominated object. We find that the specific angular momentum of the disc-dominated galaxy tracks the evolution of the angular momentum of the dark matter halo very closely: the angular momentum grows as predicted by linear theory until the epoch of maximum expansion and remains constant thereafter. By contrast, the evolution of the angular momentum of the bulge-dominated galaxy resembles that of the central, most bound halo material: it also grows at first according to linear theory, but 90 per cent of it is rapidly lost as pre-galactic fragments, into which gas had cooled efficiently, merge, transferring their orbital angular momentum to the outer halo by tidal effects. The disc-dominated galaxy avoids this fate because the strong feedback reheats the gas, which accumulates in an extended hot reservoir and only begins to cool once the merging activity has subsided. Our analysis lends strong support to the classical theory of disc formation whereby tidally torqued gas is accreted into the centre of the halo conserving its angular momentum.     

Dark matter halo response to the disc growth

We consider the sensitivity of the circular-orbit adiabatic contraction approximation to the baryon condensation rate and the orbital structure of dark matter haloes in the Λ cold dark matter (ΛCDM) paradigm. Using one-dimensional hydrodynamic simulations including the dark matter halo mass accretion history and gas cooling, we demonstrate that the adiabatic approximation is approximately valid even though haloes and discs may assemble simultaneously. We further demonstrate the validity of the simple approximation for ΛCDM haloes with isotropic velocity distributions using three-dimensional N -body simulations. This result is easily understood: an isotropic velocity distribution in a cuspy halo requires more circular orbits than radial orbits. Conversely, the approximation is poor in the extreme case of a radial orbit halo. It overestimates the response of a core dark matter halo, where radial orbit fraction is larger. Because no astronomically relevant models are dominated by low angular momentum orbits in the vicinity of the disc and the growth time-scale is never shorter than a dynamical time, we conclude that the adiabatic contraction approximation is useful in modelling the response of dark matter haloes to the growth of a disc.     

Tidal tails in cold dark matter cosmologies

We study the formation of tidal tails in pairs of merging disc galaxies with structural properties motivated by current theories of cold dark matter (CDM) cosmologies. In a recent study, Dubinski, Mihos & Hernquist showed that the formation of prominent tidal tails can be strongly suppressed by massive and extended dark haloes. For the large halo-to-disc mass ratio expected in CDM cosmologies their sequence of models failed to produce strong tails like those observed in many well-known pairs of interacting galaxies. In order to test whether this effect can constrain the viability of CDM cosmologies, we construct N ‐body models of disc galaxies with structural properties derived in analogy to the recent analytical work of Mo, Mao & White. With a series of self-consistent collisionless simulations of galaxy–galaxy mergers we demonstrate that even the discs of very massive dark haloes have no problems developing long tidal tails, provided the halo spin parameter is large enough. For our class of models, the halo-to-disc mass ratio is not a good indicator of the ability to produce tails. Instead, the relative size of disc and halo or, alternatively, the ratio of circular velocity to local escape speed at the half mass radius of the disc is a more useful criterion. This result holds in all CDM models. While tidal tails can provide useful information on the structure of galaxies, it thus appears unlikely that they are able to constrain the values of the cosmological parameters within these models.     

Star formation and chemical evolution of damped Lyman α systems

In this paper, we investigate the star formation and chemical evolution of damped Lyman α systems (DLAs) based on the disc galaxy formation model developed by Mo, Mao & White. We propose that the DLAs are the central galaxies of less-massive dark haloes present at redshifts z ∼3, and they should inhabit haloes of moderately low circular velocity. The empirical Schmidt law of star formation rates, and closed box model of chemical evolution that an approximation known as instantaneous recycling is assumed, are adopted. In our models, when the predicted distribution of metallicity for DLAs is calculated, two cases are considered. One is that, using the closed-box model, empirical Schmidt law and star formation time, the distribution of metallicity can be directly calculated. The other is that, when the simple gravitational instability of a thin isothermal gas disc as first discussed by Toomre is considered, the star formation occurs only in the region where the surface density of gas satisfies the critical value, not everywhere of a gas disc. In this case, we first obtain the region where the star formation can occur by assuming that the disc has a flat rotation curve and rotational velocity is equal to the circular velocity of the surrounding dark matter halo, and then calculate the metallicity distribution as in case one. We assume that star formation in each DLA lasts for a period of 1 Gyr from redshifts z =3. There is only one output parameter in our models, i.e. the stellar yield, which relates to the time of star formation history and is obtained by normalizing the predicted distribution of metallicity to the mean value of 1/13 Z_⊙ as presented by Pettini et al.. The predicted metallicity distribution is consistent with the current (rather limited) observational data. A random distribution of galactic discs is taken into account.     

The formation and survival of discs in a ΛCDM universe

We study the formation of galaxies in a Λ cold dark matter (ΛCDM) universe using high-resolution hydrodynamical simulations with a multiphase treatment of gas, cooling and feedback, focusing on the formation of discs. Our simulations follow eight isolated haloes similar in mass to the Milky Way and extracted from a large cosmological simulation without restriction on spin parameter or merger history. This allows us to investigate how the final properties of the simulated galaxies correlate with the formation histories of their haloes. We find that, at   z = 0  , none of our galaxies contains a disc with more than 20 per cent of its total stellar mass. Four of the eight galaxies nevertheless have well-formed disc components, three have dominant spheroids and very small discs, and one is a spheroidal galaxy with no disc at all. The   z = 0  spheroids are made of old stars, while discs are younger and formed from the inside-out. Neither the existence of a disc at   z = 0  nor the final disc-to-total mass ratio seems to depend on the spin parameter of the halo. Discs are formed in haloes with spin parameters as low as 0.01 and as high as 0.05; galaxies with little or no disc component span the same range in spin parameter. Except for one of the simulated galaxies, all have significant discs at   z ≳ 2  , regardless of their   z = 0  morphologies. Major mergers and instabilities which arise when accreting cold gas is misaligned with the stellar disc trigger a transfer of mass from the discs to the spheroids. In some cases, discs are destroyed, while in others, they survive or reform. This suggests that the survival probability of discs depends on the particular formation history of each galaxy. A realistic ΛCDM model will clearly require weaker star formation at high redshift and later disc assembly than occurs in our models.     

The formation of the Hubble sequence of disc galaxies: the effects of early viscous evolution

We investigate a model of disc galaxies whereby viscous evolution of the gaseous disc drives material inwards to form a protobulge. We start from the standard picture of disc formation through the settling of gas into a dark halo potential well, with the disc initially coming into centrifugal equilibrium with detailed conservation of angular momentum. We derive generic analytic solutions for the disc–halo system after adiabatic compression of the dark halo, with free choice of the input virialized dark halo density profile and of the specific angular momentum distribution. We derive limits on the final density profile of the halo in the central regions. Subsequent viscous evolution of the disc is modelled by a variation of the specific angular momentum distribution of the disc, providing analytic solutions to the final disc structure. The assumption that the viscous evolution time-scale and the star formation time-scale are similar leads to predictions of the properties of the stellar components. Focusing on small 'exponential' bulges, i.e., ones that may be formed through a disc instability, we investigate the relationship between the assumed initial conditions, such as halo 'formation', or assembly, redshift z _f, spin parameter λ , baryonic fraction F , and final disc properties such as global star formation time-scale, gas fraction, and bulge-to-disc ratio. We find that the present properties of discs, such as the scalelength, are compatible with a higher initial formation redshift if the redistribution by viscous evolution is included than if it is ignored. We also quantify the dependence of final disc properties on the ratio F λ , thus including the possibility that the baryonic fraction varies from galaxy to galaxy, as perhaps may be inferred from the observations.     

The nature of high-redshift galaxies

Using semi-analytic models of galaxy formation set within the cold dark matter (CDM) merging hierarchy, we investigate several scenarios for the nature of the high-redshift     ) Lyman-break galaxies (LBGs). We consider a 'collisional starburst' model in which bursts of star formation are triggered by galaxy–galaxy mergers, and find that a significant fraction of LBGs are predicted to be starbursts. This model reproduces the observed comoving number density of bright LBGs as a function of redshift and the observed luminosity function at     and     with a reasonable amount of dust extinction. Model galaxies at     have star formation rates, half-light radii,     colours and internal velocity dispersions that are in good agreement with the data. Global quantities such as the star formation rate density and cold gas and metal content of the Universe as a function of redshift also agree well. Two 'quiescent' models without starbursts are also investigated. In one, the star formation efficiency in galaxies remains constant with redshift, while in the other, it scales inversely with disc dynamical time, and thus increases rapidly with redshift. The first quiescent model is strongly ruled out, as it does not produce enough high-redshift galaxies once realistic dust extinction is accounted for. The second quiescent model fits marginally, but underproduces cold gas and very bright galaxies at high redshift. A general conclusion is that star formation at high redshift must be more efficient than locally. The collisional starburst model appears to accomplish this naturally without violating other observational constraints.     

Large-scale galactic turbulence: can self-gravity drive the observed H i velocity dispersions?

Observations of turbulent velocity dispersions in the H  i component of galactic discs show a characteristic floor in galaxies with low star formation rates and within individual galaxies the dispersion profiles decline with radius. We carry out several high-resolution adaptive mesh simulations of gaseous discs embedded within dark matter haloes to explore the roles of cooling, star formation, feedback, shearing motions and baryon fraction in driving turbulent motions. In all simulations the disc slowly cools until gravitational and thermal instabilities give rise to a multiphase medium in which a large population of dense self-gravitating cold clouds are embedded within a warm gaseous phase that forms through shock heating. The diffuse gas is highly turbulent and is an outcome of large-scale driving of global non-axisymmetric modes as well as cloud–cloud tidal interactions and merging. At low star formation rates these processes alone can explain the observed H  i velocity dispersion profiles and the characteristic value of  ∼10 km s⁻¹  observed within a wide range of disc galaxies. Supernovae feedback creates a significant hot gaseous phase and is an important driver of turbulence in galaxies with a star formation rate per unit area  ≳10⁻³ M_⊙ yr⁻¹ kpc⁻²  .     

Dwarf galaxy rotation curves and the core problem of dark matter haloes

The standard cold dark matter (CDM) model has recently been challenged by the claim that dwarf galaxies have dark matter haloes with constant-density cores, whereas CDM predicts haloes with steeply cusped density distributions. Consequently, numerous alternative dark matter candidates have recently been proposed. In this paper we scrutinize the observational evidence for the incongruity between dwarf galaxies and the CDM model. To this end, we analyse the rotation curves of 20 late-type dwarf galaxies studied by Swaters. Taking the effects of beam smearing and adiabatic contraction into account, we fit mass models to these rotation curves with dark matter haloes with different cusp slopes, ranging from constant-density cores to r ⁻² cusps. Even though the effects of beam smearing are small for these data, the uncertainties in the stellar mass-to-light ratio and the limited spatial sampling of the halo's density distribution hamper a unique mass decomposition. Consequently, the rotation curves in our sample cannot be used to discriminate between dark haloes with constant-density cores and r ⁻¹ cusps. We show that the dwarf galaxies analysed here are consistent with CDM haloes in a ΛCDM cosmology, and that there is thus no need to abandon the idea that dark matter is cold and collisionless. However, the data are also consistent with any alternative dark matter model that produces dark matter haloes with central cusps less steep than r ^−1.5. In fact, we argue that based on existing H  i rotation curves alone, at best weak limits can be obtained on cosmological parameters and/or the nature of the dark matter. In order to make progress, rotation curves with higher spatial resolution and independent measurements of the mass-to-light ratio of the disc are required.     

重子物质对暗物质晕形状和角动量的影响

利用高精度大样本的冷暗物质($\Lambda$ cold dark matter, $\Lambda$CDM)宇宙学数值模拟的数据, 对重子物质如何影响暗物质晕的形状和角动量进行了研究.使用了3种数值模拟数据, 纯暗物质模拟、含辐射冷却、恒星形成和动力学超新星反馈的模拟, 包含活动星系核反馈效应的恒星形成模拟. 对这3种模拟, 还进行了不同红移处的比较. 主要结果如下.重子物理过程会改变暗物质晕的质量分布, 特别是有活动星系核反馈机制的情况下.比如, 活动星系核反馈会减少大质量暗物质晕的形成.随着宇宙的演化, 暗物质晕的空间形态逐渐由扁变圆. 重子物质的存在会加速暗物质晕形状的变化过程, 而且会使暗物质晕形状变得更圆. 但是活动星系核的反馈会对这一加速效应产生抑制.重子物质对暗物质晕的影响与暗物质晕的质量和半径都存在一定的依赖性.暗物质晕的质量越大, 它会呈现更扁的形态. 同时, 重子物质对任意质量的暗物质晕或暗物质晕在任意半径处的变圆均有一定的促进作用,尽管活动星系核反馈会抑制这一促进作用.特别是对于暗物质晕在0.2--0.6倍维里半径处的形状, 重子物质的影响尤为明显.此外, 重子物质的存在会对暗物质晕的角动量产生显著影响, 它会增大暗物质的角动量. 暗物质晕的自旋参数与质量无相关性, 但是与暗物质晕的半径存在一定的相关性.     

Forming stars on a viscous time-scale: the key to exponential stellar profiles in disc galaxies?

We argue for implementing star formation on a viscous time-scale in hydrodynamical simulations of disc galaxy formation and evolution. Modelling two-dimensional isolated disc galaxies with the Bhatnagar–Gross–Krook (BGK) hydrocode, we verify the analytic claim of various authors that if the characteristic time-scale for star formation is equal to the viscous time-scale in discs, the resulting stellar profile is exponential on several scalelengths whatever the initial gas and dark matter profile. This casts new light on both numerical and semi-analytical disc formation simulations that either (a) commence star formation in an already exponential gaseous disc, (b) begin a disc simulation with conditions known to lead to an exponential, i.e. the collapse of a spherically symmetric nearly uniform sphere of gas in solid-body rotation under the assumption of specific angular momentum conservation, or (c) in simulations performed in a hierarchical context, tune their feedback processes to delay disc formation until the dark matter haloes are slowly evolving and without much substructure so that the gas has the chance to collapse under conditions known to give exponentials. In such models, star formation follows a Schmidt-like law, which for lack of a suitable time-scale, resorts to an efficiency parameter. With star formation prescribed on a viscous time-scale, however, we find gas and star fractions after ∼12 Gyr that are consistent with observations without having to invoke a 'fudge factor' for star formation. Our results strongly suggest that despite our gap in understanding the exact link between star formation and viscosity, the viscous time-scale is indeed the natural time-scale for star formation.     

Damped Lyα systems at high redshift and models of protogalactic discs

We employ observationally determined intrinsic velocity widths and column densities of damped Lyman alpha (Lyα) systems at high redshift to investigate the distribution of baryons in protogalaxies within the context of a standard cold dark matter (CDM) model. We proceed under the assumption that damped Lyα systems represent a population of cold, rotationally supported, protogalactic discs, and that the abundance of dark matter haloes is well approximated by a CDM model with critical density and vanishing cosmological constant. Using conditional cross-sections to observe a damped system with a given velocity width and column density, we compare observationally inferred velocity width and column density distributions to the corresponding theoretically determined distributions for a variety of disc parameters and CDM normalizations. In general, we find that the observations cannot be reproduced by the models for most disc parameters and CDM normalizations. Whereas the column density distribution favours small discs with large neutral gas fraction, the velocity width distribution favours large and thick discs with small neutral gas fraction. The possible resolutions of this problem in the context of this CDM model may be (1) an increased contribution of rapidly rotating discs within massive dark matter haloes to damped Lyα absorption, or (2) the abandoning of simple disc models within this CDM model for damped Lyα systems at high redshift. Here the first possibility may be achieved by supposing that damped Lyα system formation occurs only in haloes with fairly large circular velocities, and the second possibility may result from a large contribution of mergers and double discs to damped Lyα absorption at high redshift.     

Numerical simulations of the formation and chemical evolution of galaxies

We present the results of a set of three-dimensional SPH-Treecode simulations which model the formation and early evolution of disc galaxies, including the generation and return of heavy elements to the interstellar medium by star formation. Starting from simple initial conditions which are given by a uniform density sphere of gas which is embedded in a dark matter halo and in solid-body rotation, we are able to form realistic disc galaxies, and find that an exponential gas disc is quickly formed. Star formation within this exponential disc naturally leads to the formation of abundance gradients which are in broad agreement with those observed, although they are slightly shallower than some observations.
We investigate the systematic effects of variation of mass, rotation and star formation parameters on the abundance gradients. We find that the abundance gradients are most sensitive to changes in the star formation parameters or rotation. Including a critical-density cut-off in the star formation law causes abundance gradients to be steepened.
Analysis of gas flows within the models shows radial flows which are a function of angle of azimuth around the galaxies, with alternating inward and outward flows. This motion is linked to the presence of a bar, whose strength is related to the amount of star formation in the models, and there is a gentle drift of mass inwards. The shallow abundance gradients may be linked to these radial flows.     

Dynamical response to supernova-induced gas removal in spiral galaxies with dark matter halo

We investigate the dynamical response, in terms of disc size and rotation velocity, to mass loss by supernovae in the evolution of spiral galaxies. A thin baryonic disc having the Kuzmin density profile embedded in a spherical dark matter halo having a density profile proposed by Navarro, Frenk & White is considered. For the purpose of comparison, we also consider the homogeneous and   r ⁻¹  profiles for dark matter in a truncated spherical halo. Assuming for simplicity that the dark matter distribution is not affected by mass-loss from discs and the change of baryonic disc matter distribution is homologous, we evaluate the effects of dynamical response in the resulting discs. We found that the dynamical response only for an adiabatic approximation of mass-loss can simultaneously account for the rotation velocity and disc size as observed particularly in dwarf spiral galaxies, thus reproducing the Tully–Fisher relation and the size versus magnitude relation over the full range of magnitude. Furthermore, we found that the mean specific angular momentum in discs after the mass-loss becomes larger than that before the mass-loss, suggesting that the mass-loss would most likely occur from the central disc region where the specific angular momentum is low.     

How galaxies lose their angular momentum

The processes are investigated by which gas loses its angular momentum during the protogalactic collapse phase, leading to disc galaxies that are too compact with respect to the observations. High-resolution N -body/SPH simulations in a cosmological context are presented including cold gas and dark matter (DM). A halo with quiet merging activity since redshift   z ∼ 3.8  and with a high-spin parameter is analysed that should be an ideal candidate for the formation of an extended galactic disc. We show that the gas and the DM have similar specific angular momenta until a merger event occurs at   z ∼ 2  with a mass ratio of 5:1. All the gas involved in the merger loses a substantial fraction of its specific angular momentum due to tidal torques and dynamical friction processes falls quickly into the centre. In contrast, gas infall through small subclumps or accretion does not lead to catastrophic angular momentum loss. In fact, a new extended disc begins to form from gas that was not involved in the 5:1 merger event and that falls in subsequently. We argue that the angular momentum problem of disc galaxy formation is a merger problem: in cold dark matter cosmology substantial mergers with mass ratios of 1:1 to 6:1 are expected to occur in almost all galaxies. We suggest that energetic feedback processes could in principle solve this problem, however only if the heating occurs at the time or shortly before the last substantial merger event. Good candidates for such a coordinated feedback would be a merger-triggered starburst or central black hole heating. If a large fraction of the low angular momentum gas would be ejected, late-type galaxies could form with a dominant extended disc component, resulting from late infall, a small bulge-to-disc ratio and a low baryon fraction, in agreement with observations.