At the very beginning of the universe, there was only gas clouds. So, we must assume that galaxies have been formed from small "germs", microscopic areas denser than their neighbourhood, which can attract the surrounding matter in the same way as stars which are born in molecular clouds, of course on a much larger scale.
To be able to grow with the gravitational attraction, a density
fluctuation must include a mass of matter greater than a critical value called
the Jeans' mass : below this value, it can only oscillate like a sound wave,
but its amplitude is unable to increase.
If we consider only baryonic matter, ie matter as we can see it today, there appears a problem of time : the growth of fluctuations cann't be very rapid, because it is slowed down by the expansion of the universe. And it is quite impossible that galaxies may have been formed in such a short time in relation with the age of the universe, considering the very low amplitudes of initial fluctuations as we can see them in the cosmic microwave background.
Therefore, the two scenarios that we are exploring will have to make use of dark matter. This matter is so called "dark" because it does not interact, or only a little bit, with radiation. So we can't see it directly.
The first scenario was studied by Zel'dovich at the end of the
70's, and considers that dark matter was made up of light particles. These light
particles were moving at relativistic speeds to form what is called the hot
dark matter. From this fact, the Jeans' mass takes high values and the condensation
of gas clouds can only happen at the scale of superclusters.
But there is a problem with this scenario : the necessary time for splitting of superclusters is much too long to be compatible with the presence of high redshift galaxies, as we can see them with the Hubble Space Telescope in its deep sky pictures. Following this scenario, galaxies can only form at a redshift z=2, corresponding with a 3 billion years old universe.
In this opposite scenario, dark matter is said to be cold, because
it is made up of heavy and slow particles. In such a case, the Jeans' mass roughly
corresponds to structures with the size of globular clusters or dwarf galaxies.
Another problem of this scenario : how can the great structures, clusters and superclusters arise ?
The bottom-up scenario is based on the hypothesis of a "Cold Dark Matter". By coupling it with the
cosmological constant Λ we get a model called
ΛCDM, which is the most commonly accepted nowadays.
In this model, as the black matter does not interact, it can begin to accumulate before the electromagnetic decoupling. After that, the matter can collapse in the potential wells already created by the black matter.
This mechanism will speed up the formation of the galaxies and so avoid the problem of late formation of the bottom-up model.
From this time, galaxies store matter :
For the large spiral galaxies, we get a growing by fusion of proto-galaxies, followed by an accretion
of gas which will create the younger disk, where the stars are able to born.
In such a scenario, large elliptical galaxies are created after, by the fusion of spiral galaxies.
For the small elliptical galaxies, this scenario is at fault. They must then be considered as the remains of spiral galaxies which have lost their disk, or as aggregate of stellar groups.
The ΛCDM model is in fact much more than a simple model for the formation of galactic structures.
It is a complete cosmological model which allows the description of the
As such, it can claim for two important successes :
Source : Reshetnikov, Astronomical Institute of St.Petersburg State University
The amplitude spectrum of these fluctuations fits well with the theoretical curve, except on its extreme left, corresponding to very large structures, too few of them being present in a finite universe.
Source : R. Taillet, LAPTH of CNRS
Unfortunately, beside these two successes, there are some problems with the model.
As usual, could we say...
But, especially in dwarf galaxies, the density of black matter at the center
should be flat.
To overcome this problem, some hypothesis have been envisaged, as an interaction between the ordinary matter and the black one, but none of them gives really good results.
If the model is correct, these sub-structures must have been destroyed during the galactic
One more time, the presence of the cusps may be the key of the problem : if the black matter was less dense, it could be destroyed by the tidal forces.
But in the model, because of the high density of this black matter, all the created structures are robust and last for billions of years.
To conclude, the ΛCDM model fits quite well with the observations of the universe
for large structures and matter filaments, but lacks of precision at galaxies level.
It forecasts a too big quantity of black matter at the center of galaxies, with a too much concentrated density distribution (cupsy problem), which leads to too small galactic disks, with too many satellite dwarf galaxies.
We must remark that the smaller is the structure, the more non-linear are the physical phenomenons, which is always a problem for simulation.