4.4.5 Gas content and dynamics of BCGs
The kinematics of galaxies may be investigated by means of the motion of their stars or gas. In BCGs the former is in general not possible, due to the absence of strong absorption lines, and studies have been restricted to the gas phases.
Chamaraux et al. (1970) were the first to detect neutral hydrogen in a BCG, namely IIZw40, and more BCGs were detected in various surveys (not only targeting BCGs) : Lauqué (1973), Bottinelli et al. (1973, 1975), Chamaraux (1977). Gordon and Gottesman (1981) and Thuan and Martin (1981) conducted the first extensive systematic surveys of neutral hydrogen in BCGs, and in total more than 200 BCGs, predominantly in the northern hemisphere, were observed and the majority detected. BCGs have been found to be gas rich. The ratio of neutral hydrogen mass to integrated blue luminosity is typically in the range 0.1 MHI / LB 1.0 (in solar units, Thuan and Martin 1981, Gordon and Gottesman 1980). Thus BCGs are as gas rich as spirals and dIs, but less than LSBGs (Staveley-Smith et al. 1992).
HI interferometry of IIZw40 was reported by Gottesman and Weliachew in 1972, providing the first spatially resolved investigation of H I in a BCG. This was followed by studies of IIZw70+71 by Balkowski et al. (1978) and IZw18 by Lequeux and Viallefond (1980) and Viallefond et al. (1987) who suggested the existence of dark matter (perhaps molecules) to account for the dynamical mass. Bergvall and Jörsäter (1988) provided a detailed mass model for ESO 400-G43 and also found evidence for dark matter dominated dynamics in agreement with recent findings by Meurer et al. (1996, 1998) and van Zee et al. (1998c). A substantial fraction of BCGs seems to have H I companions (Taylor 1997), probably representing uncatalogued gas rich LSBGs.
There is a current view suggesting that the formation of stars depends primarily on the amount of molecular gas. However the situation in low metallicity gas is still under debate. Many attempts to detect CO in BCGs galaxies have been reported so far (Combes 1986, Young et al. 1986, Arnault et al. 1988, Sage et al. 1992, Israel et al. 1995, Gondhalekar et al. 1998) but the CO luminosity of BCGs is very low, in comparison with their observed star formation rate, mostly yielding only upper limits. This lack of detection may be because the low metallicity of BCGs hides the true molecular phase by a low CO to H2 conversion factor. However other explanations may be invoked as well: the CO excitation could be lower than for molecular clouds in our Galaxy, or the molecular clouds in BCGs could be UV-photodissociated as a result of high star formation rates. Gondhalekar et al. (1998) conclude that the CO luminosity correlates rather weakly with the FIR luminosity, i.e. FIR luminosity may not be a good tracer of molecular gas. Obviously the lack of CO detection does not preclude the presence of H2 molecules in these gas-rich galaxies. In fact there are mechanisms by which molecular hydrogen can be formed in absence of grains from hydrogen atoms in gaseous phase via a reaction involving negative hydrogen ions (Lequeux and Viallefond 1980). It is therefore possible that in star-forming galaxies with well localised massive star formation surrounded by huge H I gaseous envelopes that the molecular hydrogen is abundant and makes up a significant fraction of the dark matter dynamically detected (Lequeux and Viallefond 1980, Lo et al. 1993). New capabilities such as the FUSE mission will offer the unique possibility to detect for the first time cold H2 in absorption against the stellar continuum of blue massive stellar clusters. IZw18 will be the first target to be searched for H2.
Östlin et al. (1999a, b) investigated the H velocity fields of a sample of luminous BCGs utilising scanning Fabry-Perot interferometry. The velocity fields were found to be complex, and in many cases showed evidence for dynamically distinct components, e.g. counter rotating features. Their analysis suggests that mergers involving gas rich dwarfs are the best explanations for the starbursts in these systems. Masses were modelled both dynamically and photometrically, and some galaxies showed apparent rotational mass deficiencies which could be explained if the studied BCGs are not primarily supported by rotation, if stars and gas are dynamically decoupled (e.g. due to gas flows) or if the galaxies are not in dynamical equilibrium. There are also indications that the width of emission lines in BCGs is related to virial motions and may provide dynamical mass estimates (see Sect. 7.1, and Melnick et al. 1987).
Flows in the ionised gas have been detected in several BCGs (Marlowe et al. 1995; Martin 1996, 1998; Meurer et al. 1997), and suggested by X-ray observations in VIIZw403 (Papaderos et al. 1994). Flows have also been found from studies of the Ly emission. Although Ly emission in starbursts is expected to be strong, it turns out that dust is very effective in suppressing this line because the effects of resonant scattering in a gas-rich medium dramatically reduce the effective mean free path of the Ly photons. On the other hand this mechanism does not explain why in many galaxies with little dust content such as IZw18 (Kunth et al. 1994) Ly is seen in absorption whereas in dustier ones such as Haro2 (Lequeux et al. 1995) the line is seen in emission but with a clear P-Cygni profile. Kunth et al. (1998) found that the strength of Ly emission is in fact only weakly correlated with metallicity and suggested that the dynamical state of ISM is also a major regulating mechanism. A new model explains Ly profiles in starburst galaxies by the hydrodynamics of superbubbles powered by massive stars (Tenorio-Tagle et al. 1999).