This paper reports an analysis of the changes in the photoemission parameters of copper in small particles of copper oxides deposited on silicon dioxide. This study is of relevance for investigations in the fields of heterogeneous catalysis and coordination chemistry. Copper oxides (Cu2O and CuO) have been deposited on the surface of a flat SiO2 substrate by evaporation of copper and subsequent oxidization of the deposited particles. XPS has been used to analyze the chemical and coordination state of copper. Large variations in the Cu 2p(3/2) binding energy (BE) and Auger parameter (alpha') have been found as a function of the type and amount of deposited copper oxide. The differences in BE calculated from the values of the lowest amount of deposited material and those of the bulk compounds were -0.4 eV (Cu2O) and -1.9 eV (CuO), while those in alpha' amounted to 2.9 (Cu2O) and 1.6 eV (CuO). The observed changes have been described in terms of the chemical state vector (CSV) concept in a Wagner plot and rationalized by considering the characteristics of bonding and electronic interactions that occur at a given oxide/oxide interface. These interactions have been modeled by means of quantum mechanical calculations with cluster models simulating the Cu-O-Si bonding at the interface. The effect of the polarization of the surrounding media around the copper cations has been also estimated for both the dispersed clusters supported on the SiO2 substrate and for the copper oxide materials in bulk form. A change in the values of alpha' and BE of copper (ie., delta alpha' = 1.1 eV, deltaBE = 0.1 eV) upon adsorption on the Cu+ species of Cu2O moieties dispersed on SiO2 of a phenyl-acetylene molecule illustrates the use of XPS to study the formation of cation-ligand complexes in heterogeneous systems. A detailed description of the bonding interactions of these coordinated Cu+ species in terms of initial and final state effects of the photoemission process has been also carried out by means of quantum mechanical calculations and cluster models.