Annual Report 1999  

MPA Homepage > Scientific Research > Research Groups > Galaxy Formation > Euro. Network (Galaxy Formation) > Annual Report 1999

Go to:

-1 Radio Galaxies
-2a Evolution of Galaxy Clusters
-2b Evolution of Field Clusters
-3 QSO Absorption Lines
-4 Population synthesis models
-5 Semianalytic modelling
-6 Simulations of galaxies and clusters
-7 The Hubble deep field south

TMR Research Network Contract No. ERBFMRX--CT96--0086

The network continued to be very active in its third year with 43 additional joint publications submitted to the refereed scientific literature over the period 1/10/98 to 30/9/99. In addition 8 jointly authored contributions to conference proceedings were submitted. (See attached publications list.)

Personnel changes within the Network were relatively minor over this year with most of the Network's postdoctoral fellows remaining in post.
In February 1999 the Network underwent its mid-term review in Paris. Both the external referee and the Commission representative were happy with the functioning and the productivity of the Network.

The third Annual Meeting of the network was held over the period 30 August - 4 September 1999 in Asiago, Italy. Although attendance was strictly limited to scientists from the participating sites, almost 80 scientists took part. The meeting was very successful in reporting on existing activity and in generating new collaborations. Other smaller meetings are detailed in the following scientific report.

1 Radio galaxies top

The main event in 1999 was that the first 8.2-m VLT Unit Telescope (ANTU) was "handed over'' to the scientists on April 1, 1999; its first "visiting astronomers'' at Paranal were astronomers from Leiden carrying out a network programme. One of VLT's major asset is its excellent instrumentation. A number of collaborative projects have been started to use these facilities to study various classes of distant galaxies, including very red galaxies, young and starforming galaxies, dusty galaxies and galaxies that are good candidates for being in extremely distant clusters. The aim is to obtain a good census and understanding of the distant galaxy population. Within this context, a large VLT programme has recently been approved to obtain the deepest infrared images ever taken.

Scientific highlights

  • Very deep VLT imaging in a narrow band centred on the wavelength of the Ly-alpha emission on the spectacular radio galaxy 1138-262 (z=2.2) provided a sample of more than 30 candidate cluster galaxies.
  • The discovery of the most distant radio galaxy known to date at z=5.2.
  • A very deep VLT spectrum of the radio galaxy 1338-193 (z=4.1), allowing an unbiased estimate of the strength of the Ly-alpha forest continuum break.
  • Extremely deep NTT infrared imaging revealed a dozen z ~ 2.2 H-alpha emitting galaxies. The reality of these objects were confirmed through VLT spectroscopy.

Network activities

The Sterrewacht Leiden organised a TMR workshop on hydrodynamic aspects of galaxy formation and evolution at the Lorentz Center at the Universiteit Leiden from March 8 to 10 1999. The workshop was visited mainly by advanced graduate students and young postdocs from the network. A significant fraction of the workshop was devoted to a series of extended review talks by senior researchers on hydrodynamics in general, certain aspects of its numerical implementation, and the application to galaxy formation and evolution.
Exchanges: A number of young scientists have been involved in visits of order a week. The complete list is given in the section on secondments.

Progress with appointment of young researchers

The young researcher being employed by the network in Leiden is Philip Best (a number of other PhD students and post-docs are also involved in network activities). Dr Best, originally from Cambridge, was employed following the network's initial advertisement. He has traveled to meetings at other sites of the network, and is now involved in active collaborations with researchers in Leiden, Cambridge, Paris and Munich. Carlos de Breuck, a Belgium PhD student, has also been employed for a period of one year.

2a Evolution of cluster galaxies: top

A group of Network researchers from Cambridge, Durham and Padova have completed an extensive spectroscopic catalogue of galaxies in 10 distant rich clusters for which Hubble Space Telescope imaging was obtained earlier. The clusters span the redshift range z=0.37-0.56 and are the subject of a detailed ground- and space-based study to investigate the evolution of galaxies as a function of environment and epoch. A direct comparison of the spectral properties of the high-redshift cluster and field populations suggests that the abundance of galaxies that formed stars in the recent past decreased in cluster environments by more than one order of magnitude in the last 5 billion years, while such evolution is much more modest in the field. This suggests that the process or processes involved in producing such galaxies are either substantially more effective in the cluster environment or that this environment prolongs the visibility of the post-star-formation phase.

These data have also been used to investigate the nature of the different cluster galaxy types, including star-forming, poststarburst, and passive galaxy populations, and to reproduce their basic properties using spectral synthesis models developed within the Network. The comparison of the spectra of the distant galaxy populations with samples drawn from the local universe leads us to identify a significant population of dust-enshrouded starburst galaxies. The spectral and morphological properties of the distant cluster galaxies indicate that either two different timescales and/or two different physical processes are responsible for the spectral and morphological transformation of galaxies in clusters.

Finally, researchers in Durham and Cambridge have detected dust in the central galaxies of distant cooling-flow clusters. Although the sample is still small, the detections have important implications for the interpretation of star formation in cooling-flow galaxies.

2b Evolution of field galaxies: top

The pioneering study of star formation hidden by dust in extremely distant galaxies using the SCUBA bolometer array at the UK James Clerk Maxwell Telescope (JCMT) has continued to be a very active field of research within the network. Reliable redshifts for 20 galaxies have been obtained, suggesting that the majority of the extragalactic background light in the submillimeter is emitted by sources at z$<$3 and hence that the peak activity in highly obscured sources (both Active Galactic Nuclei and starbursts) lies at relatively modest redshifts. For some of these galaxies, molecular gas has been detected for the first time. The large molecular gas mass found in some of the galaxies and their radio emission are consistent with the interpretation that massive starbursts are taking place, and that massive reservoirs of molecular gas are available to fuel the very high rates of star formation detected. Two of the systems appear to be associated with merger events which may evolve into present-day luminous elliptical galaxies.

At low redshift, Network researchers in Cambridge and Munich have completed an optical and near-infrared study of a large and representative sample of nearby star-forming galaxies. The new data has been used to characterize the star-formation properties of the galaxies, including the strength and age of their current (or most recent) burst of star formation, their star-formation rate and their total stellar mass.

The Institute of Astronomy of the University of Cambridge has hosted several visits of researchers from other network institutions (Munich, Paris, Durham, Leiden and Padova), to carry out collaborative research. In addition, several IoA members have visited other nodes (Padova, Durham, Paris, Munich) to work on various collaborations an plan future research.

Dr Pierre-Alain Duc continued working at the IoA during this period as a research fellow, funded by the Network. He terminated his contract on September 30th, and moved to the D\'epartement d'Astrophysique of the Commissariat \`a l'\'Energie Atomique, (CEA) in Saclay, France, where he has secured a permanent research post. His main research activities were related to the observational study of interacting and star-forming galaxies, and the role of the environment in their formation and evolution. He actively collaborated with researchers in Cambridge, Paris and Padova within the Network.

3 QSO absorption lines top

An invaluable and complementary point of view on the processes governing the formation and evolution of galaxies is provided by the study of the gaseous component of the universe as seen in quasar absorption lines.

Their study has progressed in the direction of understanding the extent, structure, metallicity and physical conditions of absorbing halos at high redshift.

In particular the following key results have been achieved:

  • Studies of galaxies and identification of absorbers in the Hubble Deep Field South.
    The multi-object capabilities of EMMI at the ESO NTT and FORS1 at the ESO VLT have been used to determine the redshifts of galaxies in the $50'' \times 50''$ STIS and in the $ 5' \times 5'$ surrounding EIS field. Together with the unique database under construction in the HDF-S this will allow us to study in the best possible conditions the association between Lyman--alpha absorptions detected by HST and galaxies, in particular in the redshift range where a transition in the evolution of the Lyman forest takes place.
  • QSO pairs and Groups.
    The observations of several closely separated QSO pairs and gravitationally lensed QSOs, providing two or more adjacent lines of sight, have made it possible to sample the size and clustering of the absorbers.
  • Metallic Absorbers and the Ionizing Background.
    A separation of the complex weak metal-line systems relating to the Lyman forest into ``single-phase ionization'' components with accurate parameters has yielded a large sample in the redshift range 1.9 < z < 4.4. No sudden evolution near z = 3 is found in the column density ratio N(SiIV)/N(CIV), contrary to previous indication. Comparisons of N(SiIV)/N(CIV) vs N(CII)/N(CIV) from the observations, with CLOUDY derived values using several model ionizing spectra, confirm the general dominance of a QSO ionizing background at the lower redshifts but an additional strong and eventually dominating stellar contribution is required progressively to higher redshifts.
  • Simulations.
    Hydrodynamic simulations of galaxy formation in a cosmological context have been used to investigate the intergalactic medium at redshifts 1 to 5. These simulations are used to make mock QSO spectra, which are the compared in detail to high-resolution Lyman-alpha spectra taken with the HIRES spectrograph on the Keck telescope. The low-redshift Lyman-alpha forest has been studied, the influence of numerical resolution on simulated spectra, the influence of cosmology and a method to measure the temperature of the intergalactic medium based on the Lyman-alpha line widths was developed.

    Methods for the recovery of the real space mass density field from Lyman absorption in QSO spectra have been developed. It is hereby assumed that the absorption is due to a photoionized intergalactic medium which traces the mass distribution as suggested by recent numerical simulations. The methods are tested with mock spectra obtained from numerical simulations.

4 Population synthesis models top

  • The synthesis programmes, that already allowed one to compute the ultraviolet-to-near-infrared spectral evolution of stellar populations in a full range of metallicities, were further improved and exploited in different ways.
    Extension to submillimeter wavelengths. The Paris and Padova groups completed self-consistent models of the spectral evolution at far-ultraviolet to submillimeter wavelengths. This required to couple the population synthesis models with a chemical evolution prescription and to include the absorption and infrared-to-submillimeter emission by the dust. Semi-empirical recipes were used to circumvent major uncertainties in the dust modelling. The models were tested successfully against observed spectra of nearby spiral, starburst, and (ultra-)luminous infrared galaxies.
  • Inclusion of nebular emission. The Paris and Padova groups combined the synthesis programmes with standard photo-ionization codes to compute emission-line spectra of star-forming galaxies. The ionization of HII regions by young stars and the production of emission lines are treated in a consistent way with the history of star formation. The models reproduce the observed ratios of the most commonly measured emission lines (H-alpha, H$\beta$, [OII], [OIII]). Studies are underway to evaluate the reliability of nebular star formation indicators in distant galaxies.
  • Morphological versus spectral modelling. The Padova and Durham groups computed the ultraviolet-to-submillimeter emission of galaxies extracted from semi-analytic models for which morphologies (budge-to-disk ratio), star formation history, and warm and cold gas contents are available. Chemical evolution and dust absorption and emission were included self-consistently using the prescription outlined above. The models reproduce, in particular, the wide observed spread in infrared properties of spiral galaxies with similar optical spectra.

5 Semianalytic modelling top

Semi-analytic techniques have been extended this year in a number of new ways.

  • The evolution of black holes, quasars and radio galaxies. G. Kauffmann (MPA) and M. Haehnelt (MPA and IoA) have extended the semi-analytic models to track the evolution of supermassive black holes in galaxies. They have demonstrated that a model in which supermassive black holes are formed and quasars are fuelled during major mergers can explain the observed evolution in quasar space density and the relation between quasar and host galaxy luminosities at low redshifts. In collaboration with C. Kaiser (MPA), P. Best (Leiden) and H. Rottgering (Leiden), the models are currently being extended to study the redshift evolution of the sizes and environments of radio galaxies.
  • Evolution of galaxies in clusters. B. Lanzoni, G. Mamon and F. Bouchet (IAP, Paris) are collaborating with V. Springel and S. White (MPA) to include semi-analytic recipes into high-resolution N-body simulations of cluster formation. This project will focus on the evolution of galaxy populations in clusters at high redshift and how the observed trends in the star formation rates, morphologies and colours of galaxies can be understood.
  • Emission of galaxies at infrared wavelengths. The collaboration of C. Lacey, C. Baugh, S. Cole and C. Frenk (Durham) with A. Bressan, G. Granato and L. Silva (Padova) has made considerable progress this year. The dust models of the Padova group have been successfully integrated with the semi-analytic models of the Durham group and the far infrared properties of galaxies in the local Universe, including the luminosity function of IRAS galaxies and the detailed spectral energy distributions of normal spirals and of starburst galaxies, are explained very successfully. At high redshifts, the models fail to reproduce the 850 micron number counts, indicating more star formation may be occuring in starbursts in the real Universe than in the models.

6 Simulations of galaxies and clusters top

Scientific results:
Progress in the area of simulations of galaxies and clusters continued unabated during this reporting period, with major activity centred at Durham and Munich, and some participation from Cambridge and Paris. This led to 10 refereed publications and 5 conference proceedings involving researchers in these three institutions, particularly the first two. Our programme continued to exploit the network's access to two one of the largest parallel supercomputers in Europe, the Crays-T3E at Garching and Edinburgh.

A considerable amount of effort was expended analysing the"Hubble volume''simulations, which were described in last year's report. Two network papers analysing the statistical properties of the clustering pattern in the simulations were completed and several more are in progress. These cover topics such as the properties and distribution of rich clusters (tens of thousands of which formed in each simulation), the mass function of dark matter halos over an unprecedented range of scales, and the construction of realistic mock catalogues of galaxies ideal for analysing data from galaxy surveys (such as the ``PSCz'' and ``2dF'' surveys in which the network is involved).

We recently completed the first large N-body/gasdynamical simulations capable of resolving the formation of galaxies in representative volumes of the Universe. These simulations contained 2097152 dark matter particles and the same number of gas particles, with a gas mass per particle of $\sim 2\times10^9 M_\odot$, and a gravitational softening of 10 kpc. The simulations followed the formation of galaxies by gas cooling within dark halos of mass a few times $10^{11}M_\odot$ and above. This is a calculation that we had been struggling to perform for several years. The breakthrough that made it possible was a novel treatment of the intergalactic medium in which cool galactic gas is partially decoupled from the hot cooling flow around it. Over 2200 galaxies formed in our simulation, with a distribution of K-band luminosities (obtained from a spectral population synthesis model) in excellent agreement with observations. We are currently investigating the demographics and clustering properties of these galaxies as a function of time. Our preliminary results show that the two-point correlation function of bright galaxies in the model evolves rather little since $z=3$. At the present day, it has a shape close to a power-law over four orders of magnitude in amplitude, in good agreement with observations. The galaxies in a flat model with a cosmological constant are antibiased relative to the mass on small scales. These early results point to a remarkable success of these models in matching the trends seen in the data.

Encouraging as our first results are, this is only a start. Apart from the ongoing detailed analysis of these simulations, we are exploring further the validity of the assumptions we have made in the treatment of the intergalactic medium. This requires extensive testing via suites of small-scale simulations which we have recently began. Beyond that, we are focussing on an increasingly realistic treatment of the astrophysics of galaxy formation, particularly star formation and feedback. Our codes already include crude treatments of these processes but, at present, we have little idea of how the simulation results are affected by different parameterizations. We have recently began a systematic study of this problem. This is a long-term programme that will occupy a substantial fraction of our computing resources in the near future.

The simulations we have carried out this year represent substantial progress towards our year-4 milestone (``compiling the largest available suite of simulations of galaxy formation in differing environments and in a variety of cosmogonies.'') By contrast, we have not yet been able to achieve our year 2 milestone (understanding the requirements for disk galaxies to form with the observed properties) which proved to be more difficult than anticipated. We are continuing work in this problem.

Scientific highlights:
Our simulations of galaxy formation in cosmological volumes represent a major advance in the subject. They are the first simulations ever in which the formation of galaxies has been modelled within volumes large enough for their clustering properties to be studied reliably. The referee of the paper (Pearce et al 1999) referred to this work as ``heroic.'' For the first time, we have now an understanding of the physical processes that establish the clustering pattern of galaxies and a model (a cold dark matter cosmogony with a flat geometry and a non-zero cosmological constant) that reproduces the data remarkably well.

Networking and coordinating activities:
A meeting was held at Durham during Jan/13-14/1999 to discuss progress and future plans in the area of cosmological simulations of galaxy formation. In addition to network participants, this meeting also included a few external colleagues (from the Virgo consortium) who often collaborate with network participants. The discussions included coordination of the science to be done with the ``Hubble volume'' simulations, progress in a programme of simulations of the Lyman-alpha forest, and detailed planning of our programme of gasdynamics simulations of galaxy formation.

Appointment of young researchers:
The contract of Dr.\ Fabio Governato, the first network postdoctoral researcher appointed at Durham expired in Jan/99. Dr.\ Governato moved on to a permanent post at the University of Milano. To replace him, we advertised the post in the general advertisement issued by the network in Dec/97. From a very competitive list of applicants, we offered the post to Dr.\ Sergio Gelato, who began his activities on March/1/1999. During this reporting period, Dr.\ Fabio Governato continued his work on cosmological simulations. Dr.\ Gelato started work also in the area of simulations. He visited Munich during the year

7 The Hubble Deep Field South top

Three network proposals for VLT time to study galaxies in the Huble Deep Field South were successful: \parindent=10pt
A FORS proposal to study Lyman Break galaxies at redshifts near 4.
A FORS proposal to search for galaxies corresponding to the observed absorption line systems in the spectrum of the QSO in the STIS field
An ISAAC proposal to measure rotation curves for disk galaxies at redshifts beyond 1. All three projects were successful in getting some data and reduction is now underway. Further time has been requested for the first two.

Comments to: Naoki naoki@mpa-garching.mpg.detop