2006A&A...447..877G

We have investigated a number of globular cluster candidates from a recent wide-field study by Harris et al. (2004AJ....128..712H) of the giant elliptical galaxy NGC5128. We used the Magellan I telescope + MagIC camera under excellent seeing conditions (0.3''-0.6'') and obtained very high resolution images for a sample of 44 candidates. Of these, 15 appear to be bonafide globular clusters in NGC5128 while the rest are either foreground stars or background galaxies. We also serendipitously discovered 18 new cluster candidates in the same fields. Our images allow us to study the light profiles of the likely clusters, all of which are well resolved. This is the first ground-based study of structural parameters for globular clusters outside the Local Group. We compare the psf-deconvolved profiles with King models and derive structural parameters, ellipticities and surface brightnesses. We compare the derived structural properties with those of other well-studied globular cluster systems. In general, our clusters are similar in size, ellipticity, core radius and central surface brightness to their counterparts in other galaxies, in particular those in NGC5128 observed with HST by Harris et al. (2002AJ....124.1435H). However, our clusters extend to higher ellipticities and larger half-light radii than their Galactic counterparts, as do the Harris et al. sample. Combining our results with those of Harris et al. fills in the gaps previously existing in r_h-M_V parameter space and indicates that any substantial difference between presumed distinct cluster types in this diagram, including for example the Faint Fuzzies of Larsen & Brodie (2000AJ....120.2938L) and the ``extended, luminous'' M31 clusters of Huxor et al. (2005MNRAS.360.1007H) is now removed and that clusters form a continuum in this diagram. Indeed, this continuum now extends to the realm of the Ultra Compact Dwarfs. The metal-rich clusters in our sample have half-light radii that are almost twice as large in the mean as their metal-poor counterparts, at odds with the generally accepted trend. The possibility exists that this result could be due in part to contamination by background galaxies. We have carried out additional analysis to quantify this contamination. This shows that, although galaxies cannot be easily told apart from clusters in some of the structural diagrams, the combination of excellent image quality and Washington photometry should limit the contamination to roughly 10% of the population of cluster candidates. Finally, our discovery of a substantial number of new cluster candidates in the relatively distant regions of the NGC5128 halo suggests that current values of the total number of globular clusters may be underestimates.