1999A&A...350...49H

We report ISO observations of globular clusters, from which we derive new upper limits on the masses of dust therein. Our observations do not confirm the claimed detection of dust in 47 Tuc by Gillett et al. (1988AJ.....96..116G) and our limits generally support previous conclusions that globular clusters are deficient in intra-cluster dust grains. We find that sputtering of dust grains by halo gas is not the mechanism responsible for the removal of such matter, at least in three of the observed clusters. Some other mechanism is therefore at work to remove any dust from the cluster on a short timescale, or the formation of dust within a cluster is inhibited. Our results are consistent with the view that metallicity is the most likely factor concerning the latter point.