Cooling And Star Formation

The star formation rate in galaxies can be affected by cluster mergers in two ways. The interstellar medium in a galaxy can be compressed during a merger because of the higher pressure. This would lead to an increased star formation rate. This effect was predicted in simulations by Evrard (1991). Also in a number of observations a connection between mergers and enhanced star formation rate has been found, e.g. in the Coma cluster (Caldwell et al. 1993), in A2111 (Wang et al. 1997), in A2125 (Owen et al. 1999), in several other clusters (Moss & Whittle 2000).

In contrast to these results Fujita et al. (1999) found through simulations that the interstellar medium in the galaxies is stripped off due to the increased ram pressure during the merger, which causes the galaxies to lose their gas. Therefore less gas is left to fuel the star formation process and hence the star formation activity decreases. Fujita et al. (1999) found an increase of post-starburst galaxies at the moment of the subcluster collision, which indicates that a rapid drop in star formation must have occurred.

In simulations without radiative cooling and star formation, it is found that the gas is less concentrated than the dark matter. Also the X-ray luminosity - temperature relation, inferred from simulations, is in disagreement with observations (Eke et al. 1998; Bryan & Norman 1998; Yoshikawa et al. 2000). The question arises whether this is a numerical artifact due to neglecting physical processes or due to the difficulty in determining the X-ray luminosity of numerical models correctly. It may also be connected with the observational findings of different profiles of baryonic and dark matter (Schindler 1999) and deviation of the X-ray luminosity - temperature relation from a pure power law (Ponman et al. 1999). Both findings could be explained by non-gravitational heating processes.

In order to answer this question several groups have performed simulations which include cooling and star formation. These groups came up with quite different conclusions. Lewis et al. (2000) found that models with cooling and star formation have a 20% higher X-ray luminosity and a 30% higher temperature in the cluster centre. Also Suginohara & Os-triker (1998) found that radiative cooling increases the X-ray luminosity. In contrast to these results Pearce et al. (2000) and Muanwong et al.

(2001) found that radiative cooling decreases the total X-ray luminosity. In terms of temperature they obtained the same result as the other authors.

In order to test whether the gas is less concentrated than the dark matter because of preheating (= early non-gravitational heating) Bialek et al. (2001) performed simulations with an initially elevated adiabat. They find that keV cm2, i.e. preheating could be a possible explanation for the effect, they can reproduce the observations with an initial entropy of 55 - 150

Mathiesen & Evrard (2001) took a different approach and tested with their models how good the observational temperature determination is. They simulated CHANDRA spectra and found that the temperature can be underestimated by up to 20% by the standard temperature determination method. The reason is that cold material falls into the cluster from any side, including along the line of sight. This results in a cold contribution to the cluster spectrum and hence to a lower temperature determination. So far no final conclusion has been reached on the question of the gas distribution.

Bryan & Norman (1998) showed that in the simulations the mass -temperature relation is much more robust than the X-ray luminosity -temperature relation and hence suggested to use the former relation to draw conclusions on the cluster formation process.

Radiative cooling can result in a cooling flow which can be severely affected by a merger. Gómez et al. (2001) found in 2D simulations that cluster mergers can destroy cooling flows if the ram pressure of the gas of the infalling subcluster is sufficiently high. The ram pressure is able to displace the high-density gas in the cooling core as well as to heat it through compression, shocks and turbulence. The time scale on which a new cooling flow re-establishes itself depends on the initial cooling time of the cluster and on the severity of the merger. The cooling flow is not disrupted immediately. Gómez et al. found a lag of at least 1-2 Gyr between the core passage and the point at which the central cooling time exceeds the Hubble time. Also rotation of the ICM as a consequence of a merger can have effects on the cooling flow (Garasi et al. 1998).

Was this article helpful?

0 0

Post a comment