binding sites), [27] as well as with the trapping of size-selected Ag clusters of steps when the clusters are deposited at low energy. [28] The fact that the ªunpinningº temperature is not exactly the same for each pinned cluster is probably attributable to a combination of thermal statistics and the (small) range of cluster±surface impact parameters in the collision, although for clusters as large as 70 atoms in size, the latter is not expected to be a large effect. [29] In summary, the stability of pinned, size-selected gold clusters (Au 70 + ) on a graphite surface at high temperatures has been investigated. The stability of the pinned clusters against lateral diffusion to step edges (i.e., unpinning) increases with the initial cluster±surface impact energy, consistent with shallow implantation of the clusters well above the threshold energy for pinning. Size-selected cluster films which are stable even at several hundred degrees above room temperature can thus be generated, opening up potential applications in studies of physical, chemical, and biological behavior at elevated temperatures. ExperimentalThe Au 70 + clusters were produced with a magnetron-sputtering, gas-condensation cluster-beam source and lateral time-of-flight mass filter as described previously [12,22]. The clusters were deposited in a high-vacuum chamber to achieve a coverage of~1 10 11 clusters per square centimeter on freshly cleaved graphite substrates (highly oriented pyrolytic graphite) and then transferred in a vacuum suitcase to a variable-temperature scanning tunneling microscope (STM, Omicron) housed in an ultrahigh-vacuum (UHV) system. Two samples were prepared at room temperature: sample A, where the Au 70 + clusters were deposited with a kinetic energy of 1.2 keV (17 eV per atom), and sample B, where the Au 70 + cluster deposition energy was 1.7 keV (24 eV per atom). Thermal annealing was carried out for both samples in the UHV STM system by raising the temperature to a specific value for a period of two hours. During heating, the STM tip was moved away from the area of interest in order to avoid tip-induced effects. STM imaging was performed after cooling down to room temperature. The same area was scanned before and after heating wherever possible, but at high annealing temperatures, excessive thermal drift prevented the imaging of the same area. However, a sufficient number of images were taken after each annealing cycle to make sure that changes observed in different images are caused by thermal annealing and not due to regional variations. Under the imaging conditions typically applied, no tip-induced changes to the pinned clusters were observed at room temperature, which allowed us to conclude that the observed changes are primarily due to thermal effects.