This is the astro-ph blog of the Theoretical Modelling of Cosmic Structures group (TMoX) at the Max-Planck-Institute for Extraterrestrial Physics. We are an independent Max-Planck Research Group focusing on the various aspects in the formation and evolution of galaxies. Part of our focus is on the formation and evolution of early-type galaxies, super-massive black holes, the formation of the first structures in the universe and the enrichment history of the Universe. We are theoreticians using analytic modelling as well as numerical simulations in our work.

The CosmologyCake blog is dedicated to the discussion of research papers and current developments. We will regularly post interesting papers and comment on them. Feel free to leave your comments as well. We encourage authors of discussed papers to post replies if they wish to. Our aim is to provide a platform to discuss recent astro-ph papers within a wider audience. Please feel free to send papers you would like to be discussed to us at tmoxgroup@googlemail.com.

23 May 2011

Hierarchical formation of bulgeless galaxies II: Redistribution of angular momentum via galactic fountains

Link to paper here.
Authors: C.B. Brook, G.Stinson, B.K. Gibson, R. Roškar, J. Wadsley, T. Quinn

In this paper, the second in a suite, the authors are using cosmological zoom simulations to investigate the possibility of forming a bulgeless disc galaxy. Starting with initial conditions from the McMaster Unbiased Galaxy Simulations (MUGS) project, their programme is running disc galaxy re-simulations spanning 4 order of magnitude in stellar mass. After showing in their first paper that ejection of low angular momentum material enables the formation of bulgeless dwarf disc galaxies, this works introduces the mechanism of redistribution of momentum by galactic fountain as a possible solution to the angular momentum problem for intermediate mass disc galaxies.
After presenting the evolution of the main properties of the simulated object (especially the formation of a secular pseudo-bulge via bar/disc instabilities), they study in detail the evolution of the "bulge gas" (BG) since the last major merging epoch. They find that, after expelling most of this BG by efficient SNe activity during the starburst phase, by z=0, 71% of the BG has been brought back to the disk to form stars. The authors show that during the re-accretion of the BG via the galactic fountain, the material has gained angular momentum which prevent from accumulating gas to the centre of the galaxy, and hence prevent the formation of a too massive bulge.

14 comments:

  1. It is very interesting, that the galaxy has a major merger, but the associated starburst of the dissipational component is not making many stars. This must have to do with their feedback implementation. So no more bulges/ellipticals in major mergers?!

    ReplyDelete
  2. It is true that stars formed in the progenitors before the merger epoch should be redistributed by violent relaxation processes during the merger and found in the very centre a later time (forming the onset of a classical bulge).
    However, from Fig.8, one can see that most of the stellar budget is assembled later in the history of this galaxy. I think there is 2 point to be considered here: first, only 20% of the stellar budget at z=0 was made from the BG, implying that in the last 10 Gyrs most of the stars formed from material accreted from the IGM (most likely in a cold mode accretion). Secondly, from all the gas detected during the merger epoch, even if 71% will end up as stars, only 9% is converted straight after the merger. Again the authors show nicely that their model of feedback from SNe is able the blow most of the BG out of the galaxy before it comes back to form stars in the disc. Assuming the modelling is realistic, this provides indeed a nice mechanism for redistributing the low angular momentum material from galactic fountains, preventing the formation of a massive classical bulge.

    However, even if I don't questioned this mechanism, I doubt that SNe feedback by itself can drive such an amount of gas mass out to the hot halo. Indeed, the blast-wave model developed by Stinson et al. (2006), while preventing the overcooling from the SN bubble (by switching off cooling in a radius and for a time given by the snowplow model of McKee & Ostriker-1977), may, in my opinion, be overshooting the effect of SN feedback. I think keeping very high pressure inside the bubble until the end of the snowplow phase is creating a consequent amount of momentum that will keep the bubble expanding even after cooling is able to radiate the remaining thermal energy of the bubble. The definition of this maximum snowplow time should give the end of the expansion, not the beginning of the cooling!
    Nevertheless, in the discussion section, the author state that a central engine (like AGN activity for instance, which they are not considering in this work) would increase the efficiency of the fountain. I would think this contribution will add up to a lower efficient starburst induced SN feedback and will bring the picture similar to the one nicely presented in this paper.

    ReplyDelete
  3. Certainly, feedback modeling is the most difficult thing in these types of simulations. Remember that we are modeling feedback from multiple overlapping supernovae, not a single one. Whether our feedback is too strong is difficult to ascertain, but we do try to use constraints relevant to the resolution we have, e.g. our models match the scale and distribution of the HI super-bubbles in THINGS galaxies which may be caused by overlapping supernovae. We also would argue that our resolution makes coupling of energy to the ISM actually quite inefficient, even with the blast wave formalism. Certainly, it is temporal resolution that means that the blast wave formalism is required (or some other subgrid recipe): the cooling time of the dense gas in the star forming regions which has been heated by SN is shorter than the simulation time-steps.

    ReplyDelete
  4. At some mass scales, the outflows could become less effective, meaning mergers of higher mass proto-galaxies may well result in bulges/ellipticals, neatly fitting the mass-morphology relation. That is certainly the hope! The orientation of mergers may also be important.

    ReplyDelete
  5. Are you suggesting that the scatter in the mass-morphology relation is mainly driven by the merging history of individual haloes? So, it also implies that varying the halo mass, typical merging histories would vary from preferentially forming low mass bulgeless discs towards higher mass ellipticals?
    Showing this with a suite of zoom runs will certainly be a very nice result!

    ReplyDelete
  6. Even after looking at Fig.8. I'm still not sure why the stars coming from the progenitors are not found in a sort of classical bulge after the major merger (still a 2:1)? What mechanism will prevent them from contracting towards the centre?

    ReplyDelete
  7. Regarding the feedback modelling, I definitely agree with the short cooling time in high density environment argument: this should make the feedback inefficient there. However the authors argue that starbust induced feedback is able to blow away 91% of the "bulge gas" after the merger epoch ended (referring to the 9% of BG form shortly after detected...). If the first argument is valid, it would mean that most of this BG is not in a high density state (in order to be able to blow it away), even though the starburst should happen quite close to the core (given by the collapse of collisional material induced by the merger).

    Another way of asking the question about feedback efficiency is: when considering those 91% of expelled BG during starbust (~3e9Msun), if we assume (from Fig.8) a star formation rate of around ~1Msun/yr for a Gyr long burst, the mass loading factor of the galactic wind (or fountain) should be close to, or even above 1. Isn't it a bit a large amount of mass expelled if only the low density material is removed?

    ReplyDelete
  8. The arguments wrt the bulge gas and redistribution of angular momentum are all very interesting, but surely bulges and ellipticals are formed primarily via a redistrubution of orbits for existing stars? So if the authors wish to claim that they are forming a completely (classical-)bulgeless galaxy even after a major merger, then I would like to know what happens to the progenitor stars.

    ReplyDelete
  9. The stars already formed *prior* to the merger are *not* what forms bulges. These stars do not end in the very central region. They form extended stellar halos. Note too that there are not many of these stars: the very strong relation between halo mass and stellar mass means that low mass progenitors have relatively few stars.

    As for the mass loading argument, mass loadings are known to be high during merger induced starbursts, and remember we are just saying 91% of gas is blown out of the inner region, not necessarily totally out of the galaxy.

    ReplyDelete
  10. It is very interesting if the progenitor stars end up forming a stellar halo. It would be nice to know how this depends upon the existing mass of stars in the progenitor galaxies, as in this case this is clearly low. For this, more simulations are clearly required.

    Having said this, I feel quite strongly that the authors are putting too little emphasis on this issue -- it might be that in *this* simulation the stars already formed prior to the merger do not form a bulge, but it is quite extremely opposed to conventional wisdom and the results of other simulations to claim that this is never the case.

    ReplyDelete
  11. I disagree. I dont think it is conventional wisdom at all that accreted stars form bulges in *low mass systems*. In fact they simply can't, or else low mass galaxies would all have bulges, but most of them don't. The idea that classical bulges form during mergers concerns gas being driven to the central region and forming stars there. In old simulations where feedback was not efficient and too many stars formed in the central regions, then merging two proto-galaxies which already have bulges, sure you end up with the stars those in the bulge. But those simulations were wrong to begin with. We are talking about merging one galaxy that does not have a bulge with another galaxy which does not have a bulge . Then the stars that have already formed prior to the merger retain enough energy to form an extended structure.

    ReplyDelete
  12. I did not mean that the low mass progenitors should necessarily have bulges. But violent relaxation should play a role during the merger and lead to a dispersion dominated remnant, even in the absence of gas. This goes right back to:
    http://adsabs.harvard.edu/abs/1988ApJ...331..699B

    ReplyDelete
  13. Sure, the remnant will be dispersion dominated. But it will also have a low stellar mass compared to the final total stellar mass in low mass galaxies, and the remnants which were already stars before the merger will be far less centrally concentrated than the case where stars form which gas is driven to the centre during the merger. Directly acccreted stars simply do not form bulges in disc galaxies.

    ReplyDelete
  14. The mass of the dispersion dominated only depends on the masses of the progenitor galaxies, and therefore on the relative timings of the merger and star formation in the simulation.
    I believe you are arguing that the size of the expected dispersion dominated component is inconsistent with the bulges of disc galaxies?

    ReplyDelete