For the last few years we have been working on the synthesis and characterization of metal-chalcogenide clusters. For most of the transition metals, one observes the formation of relatively low-nuclearity cluster complexes, [1] such as [Co 6 E 8 -(PR 3 ) 6 ] (E = S, Se, Te; R = organic groups) und [Ni 34 Se 22 -(PPh 3 ) 20 ]. In contrast, for clusters of copper and silver one can find a rich variety of structures. [2] Recently, we reported the synthesis of metal-rich silverchalcogenide clusters, such as [Ag 70 S 20 (SPh) 96 ] (dppm = bis(diphenylphosphanyl)methane; dppb = 1,4-bis(diphenylphosphanyl)butane), particles with diameters in the nanometer range. The surfaces of these clusters are protected by ligands, thus preventing further reaction to form the thermodynamically stable binary silver chalcogenide salts.[3] Perhaps surprisingly, these cluster complexes could be prepared reproducibly and in high yield by the reaction at room temperature of, for example, silver carboxylates with S(tBu)SiMe 3 in the presence of tertiary phosphanes. In contrast, at higher temperatures amorphous Ag 2 S was formed. We therefore propose that these metal-rich clusters represent intermediates in the formation of solid Ag 2 S. When the progress of the reaction is monitored by dynamic light scattering, initially no particles are formed that are large enough to be observed by this technique. However, after several days particles form in the size range 2-4 nm and crystallize out of solution. The structural determinations of these large clusters proved problematic. With nuclearities of up to 100 metal atoms, crystals generally diffract well to up high 2q values (50-608 with Mo Ka ); the atoms have low temperature factors, and no high residual electron density is observed within the clusters. However, this situation changes for larger clusters with nuclearities greater than around 120 metal atoms. For such clusters, the intensities of the reflections drop off rather sharply above 2q % 408, and the structure refinement results in unsatisfactorily high R factors, with high residual electron density within the cluster molecule. Satisfactory R factors can only be obtained if this electron density can be modeled during the refinement. As this electron density generally lies close to the heavy atoms, it can be difficult to interpret and thus complicates efforts to give precise estimates of the molecular formulae. These effects may result from a range of factors: 1) There is no perfect translational order in the lattice. 2) With the silver-chalcogenide clusters there is a tendency towards nonstoichiometry, as is seen for the binary phases. [4] This behavior could be a consequence of the rather similar electronegativities of silver and the chalcogenides. There is no clear distinction between Ag + and E
2À(E = S, Se, Te), and the clusters behave rather like alloys. [5] 3) The surface tension of the spherical molecules generates a Laplace pressure within the molecule, which can result in a disorder or even a phase transition. 4) Interacti...