A typical globular cluster (GC) last crossed the Galactic plane some years ago. At that time any gas and dust it contained would be removed tidally and by ram pressure. Since then up to a thousand of its stars will have gone through the post-main-sequence (RGB, HB, AGB, post-AGB) evolutionary stages, each star losing about 0.3 of matter into the interstellar medium. Thus some 10 of gas and some of dust would be expected now to be in the cluster, if no loss from the cluster occurs.
However neither gas nor dust at this level has yet been detected in any cluster, with upper limits some hundred times less than the expected values. It is thus presumed that some mechanism(s), as yet undetermined although there are many candidates, strips the matter out as it is injected.
The alternative explanation, that not as much matter has been lost from the stars as expected, seems very unlikely. For this to happen either not as many stars will have had to evolve through the mass-loss evolutionary stages as expected, or each star will have had to lose less mass than expected, both of which scenarios which would cause major problems for what are thought to be well determined aspects of stellar evolution. For a discussion of this see Roberts (1988).
The sole believable detection of interstellar matter in a GC comes from IRAS pointed observations of 47 Tuc (Gillett et al. 1988), where of dust was detected. If this is typical, then dust must be stripped out of clusters on a timescale of about 10 years in order to balance the input of dust from evolving stars. With only the one detection - in what is a very massive cluster - it is impossible to say either if that amount of dust is typical, or to discriminate between the different proposed stripping mechanisms.
A determination of, or even stronger limits on, the amount of interstellar matter in GCs would clearly be a major step in determining the mechanism(s) by which GCs are purged of their interstellar matter.
There has been no previous attempt to search for the dust component of this matter at continuum millimetre wavelengths, even though it is a sensitive probe for dust. We have therefore started a programme of millimetre observations of GCs, and present here the results for two clusters, M 3 and M 22.
© European Southern Observatory (ESO) 1997
Online publication: July 8, 1998