We report density functional theory calculations, with the TPSS functional, for the clusters Ag-n (n = 3-20), AgnCl2 (n = 3-10, and n = 12) and AgnHgm (n = 5,6 and m = 1-4). We performed unbiased global optimization for all of these species by Tabu Search in Descriptor Space. The Ag-n global minima geometries that we find confirm the earlier PBE functional results of Yang et al. (M. Yang, K.A. Jackson, J. Jellinek, First-principles study of intermediate size silver clusters: shape evolution and its impact on cluster properties, J. Chem. Phys. 125 (2006) 144308 (8pp.)). We find special stability associated with Ag-6, Ag-7, and Ag-8 which we explain by their closed electronic shell (n = 6, 8) or closed atomic shell (n = 7). The Ag-16 cluster stands out in that it is particularly unstable, it has a small HOMO-LUMO gap, and it has a high-coordination binding site (the site of the missing thirteenth atom of an icosahedron). This may explain the experimental observations (L. Lian, S.A. Mitchell, P.A. Hackett, D.M. Rayner, Size dependence of mobility on the surface of metal clusters: ammonia on Ag-10 and Ag-16, J. Chem. Phys. 104 (1996) 5338-5344) that NH3 binds strongly to Ag-16 and does not diffuse on it. We calculate the Cl-Ag-n binding energy to be roughly 3.0 eV in most cases, and the Hg-Ag-n to be roughly 0.25 eV. Binding of two Cl atoms severely disrupts the Ag-n geometric structure in some cases (n = 5, 6, 7), but leaves the Agn structures essentially intact in other cases (n = 4 and 9). Binding to Hg atoms has little effect on the geometry of Ag-5 but a big effect on that of Ag-6. (c) 2013 Elsevier B.V. All rights reserved.