Noting that the experimental variables that influence the dynamics of Ag nano-particles are temperature, deposition rate and the residual vacuum 2025, the present experiments were carried out in two different ultra-high vacuum chambers. The first chamber was equipped with a Ag effusion cell, Omicron variable temperature (VT) STM (Omicron NanoTechnology GmbH, Taunusstein, Germany), low energy electron diffraction and Auger electron spectroscopy (AES) optics in UHV.

Before deposition, clean HOPG (0001) samples were prepared by cleaving in air and immediately transferring substrate into the VT-STM UHV chamber through a load lock. The cleanness of HOPG was confirmed by STM and AES prior to evaporation. Silver was evaporated by thermal evaporation (base pressure < 10^-10 Torr) of the solid material (99.9999% purity) from a Knudsen cell which gives a flux of atoms onto the HOPG substrate maintained at room temperature. The average deposition rate was 0.1 monolayer/sec and constitute a current flux of 1.04 μA that was allowed to flow for approximately 5 s. The flux of Ag atoms was kept constant by controlling the emission current between the W filament and the Ag source. Nano-meter-size Ag particles were spontaneously formed by diffusion and aggregation of the deposited material on the surface. AES confirmed the presence of Ag on the carbon dominated surface after growth. The samples were then studied by STM, operated in the constant current mode using electrochemically etched W tips as probes. Height, as a function of lateral position, represents the surface image. The STM-image scales were calibrated using Si(111)-7 × 7 and HOPG.

Electronic and chemical properties of deposited Ag clusters were studied using the second UHV chamber equipped with a XPS spectrometer. The Ag on HOPG samples was mounted in a PHI 5400 XPS vacuum chamber with a base pressure of < 8 × 10^-9 Torr. Angle-resolved XPS was performed using a non-monochromatic magnesium K_α source (1,253.6 eV) and a concentric hemispherical sector analyzer with pass energy of 178.95 eV and a scan rate of 5 eVs^-1. The X-ray source was operated at a power of 300 W at various electron take-off angles. The binding energy scale was calibrated with the Cu 3p_3/2 (933 eV) and Au 4f_7/2 (84 eV) peaks.

Within the framework of density function theory, density functional calculations were carried out using the generalized gradient approximations/PBE 26 exchange correlation potential embedded within the DMol³ 27. The interaction between the ionic cores and the valence electrons is modeled using the ultra-soft pseudopotentials of Vanderbilt 28.

Energetics and electronic properties, density of states and band structures were computed for the pristine (pure) graphene surface, silver adsorbed on the hollow site, bridge site of the hexagonal ring as well as on the top site of the carbon atom in the hexagon of the 33-atom graphene sheet. A kinetic energy cutoff of 350 eV was employed, and this choice gave fully converged calculations. Brillouin zone sampling was done using the 2 × 2 × 1 special k-point mesh employing the Monkhorst-Pack scheme 29.

Figure 1a shows an STM image of flat surface of HOPG with random distribution of isolated predominantly spherical particles of silver prepared by evaporating Ag at room temperature onto the HOPG substrate. All STM images were acquired in constant-current mode. The average particle size is about 1.2 nm in diameter and 0.68 nm in height (Figure 1c, d). The apparent particle height in STM images is a result of both electronic and geometric contrasts; thus, the obtained particle height using STM might contain some errors of a few nano-meters. A commercially available scanning probe image processor v5.110 software package from Image Metrology (Horsholm, Denmark) was used to measure the height and diameter distributions of the Ag nano-particles. The images were flattened before any size determinations to ensure that the influence of the substrate morphology is negligible. About 15% to 18% of the HOPG surface is estimated to be covered by Ag in the acquired STM data. This has been determined by counting the number of Ag particles in a unit area, which was then multiplied by the mean particle diameter, estimated by the widths of the half maxima of the particle profiles in STM images.

Even though the coverage determination is crude due to the in-homogeneity of the cluster spot, STM images show that Ag clusters exist as individual clusters on the surface. Previous studies have shown that diffusion and aggregation of clusters on the surface after deposition may change the cluster size distribution significantly such that the considerable effort of producing nano-size clusters may be compromised 373839. However, in this study, at coverages of approximately 0.5 ML of Ag on HOPG, there was no direct evidence from the STM images of cluster aggregation of the incident clusters at the surface. Similar experimental observations of Ag clusters on HOPG were made by Salido et al 40. Due to the difference in lattice parameters between silver and HOPG substrate, the clusters experience a weak interaction with the substrate, implying that they remain sufficiently mobile and are able to diffuse within the HOPG surface. However, at a coverage of 0.5 ML of Ag, the movement of clusters is quite negligible 404142. Furthermore, the Ag clusters under consideration have average diameter of approximately 1.2 nm which means they are so closely analogous to a few atoms that cluster and contain delocalized conduction electrons which is a necessary requirement for quantum confinement 42. Areas of bare HOPG between Ag clusters were evident from STM images. The particles grow three dimensionally via the Volmer-Weber growth mechanism. The white oblate shape superimposed on the STM image of Figure 1a shows two clusters merging through a nucleation process. The STM images show well-dispersed particle distributions on the surface, implying that these surfaces did not undergo particle agglomeration due to STM tip.

In order to illustrate the fact that the observed STM data of Ag clusters grown on HOPG were not due to tip artifacts, the sample surfaces were imaged in high resolution to obtain both Ag and HOPG atoms. Metal clusters belong to the mesoscopic system, systems composed of a countable number of atoms or molecules ranging from very few to a few thousand as illustrated by Figure 2a which is a 'zoom-in' image on an atomically resolved cluster of Ag atoms on the surface. The high-resolution STM image clearly shows the adsorption sites of Ag on the HOPG surface. The Ag adatoms appear to be sitting directly on the 'B' site of the HOPG as confirmed by DFT studies to follow. The line scan (Figure 2b) taken along the line in Figure 2a depicts the corrugations of both C and Ag atoms at the surface. The Ag-Ag measured distance is comparable to that of the spacing between C-C atoms on graphite (0.246 nm) as illustrated on the image. The previously anticipated agglomeration of clusters due to their mobility on inert substrate was not observed in this study.

High-energy resolution angle-resolved XPS spectra were collected as shown in Figure 3. No core level peaks were obtained at take-off angles above 10°, implying that the Ag was on the top most surface layer of HOPG. The 3-D core level peaks occur at 374.7 eV and 368.7 eV binding energies. The core level energy shift with respect to pure Ag amounts to about 0.5 eV to higher binding energy 43. XPS studies on these Ag nano-particles reveal that the Ag 3-D binding energies are quite sensitive to the Ag particle size on graphite and show deviations from the bulk properties in terms of the Ag 3-D level binding energy.

One possible interpretation of the acquired STM data is that the Ag clusters exist as individual entities at lower Ag coverages. When particle size becomes smaller, the positive hole created by the photo-ionization process can be less efficiently screened, resulting in an additional positive core level shift (final state effect) which is evident from the observed shift in binding energy of the XPS data (Figure 3). Considering that the work function of HOPG (approximately 4.6 eV) is higher than those of Ag surfaces (approximately 4.2 eV) by about 0.4 eV, the Ag nano-particles can be partially positively charged on HOPG, resulting in the positive shifts 4445. There are other factors such as the induced lattice strain 46 which is a consequence of a decrease in particle size and/or, alternatively, the increased number of uncoordinated 32 atoms compared to the bulk, which might be used to explain the large positive core level shift of Ag on HOPG.

In order to confirm the preference of Ag sitting on top of C atoms rather than at the bridge or in the hollow sites (see Figure 4 center) of the C hexagon, an ab initio study was necessary. Upon full geometry optimization (structural relaxation) of the systems, silver atoms preferred to occupy on top sites (on top of the carbon atom). This was found to be the most energetically stable configuration (see Table 1) compared to the hollow and bridge sites.

The effect of Ag on the electronic structure of graphene was further investigated (Figure 4a, b, c, d). Figure 4a shows the electronic density of states for the pristine graphene. The predicted direct band gap shows semi-metallic characteristic as can be seen from the band structure. Introducing silver on top sites induces acceptor levels at the bottom of the conduction band, and these states overlap across the Fermi, rendering the system metallic. This effect can be seen by the disappearance of the band gap between the valence and the conduction band as seen in Figure 4b. This study shows that the introduction of low concentration of metal clusters such as silver can help fine-tune the graphene band gap.

In Figure 4, the labels s, p, d, and f on the legend illustrate the electronic energies from s, p, d, and f orbitals which are defined by a common azimuthal quantum number l within an energy shell corresponding to l = 0, 1, 2, and 3, respectively. The sum depicts the overall energy contribution from orbital.

Considering the density of states, one observes that the contribution by the silver is encountered at the bottom of the conduction band for the silver on top of carbon (Figure 4b) and one with silver in the hollow site (Figure 4c). However, introducing silver on the bridge site (Figure 4c) resulted in the disappearance of the silver signature around the Fermi level, and less significant peaks are observed very far from the Fermi level (around -0.55 eV) in the valence band region.