18O-exchange with the substrate during 18O oxidation of Cu particles supported on 16O-oxidised Si(100)

An 16O2-oxidised Si(100) surface, covered with 15 nm Cu particles, was oxidised in 18O2 and subsequently reduced in H2. After reduction, there still was a significant amount of 18O present in the sample. In this paper we concerned ourselves with revealing the origin of this remaining 18O. By performing some strategic experiments we found that there was no significant contribution to the residual 18O in the sample due the following processes: oxidation of the substrate to a further extent, the remain of an 18O-adlayer on the sample after reduction, and reactions of reduction products with the substrate. The remaining 18O was caused by self-diffusion of oxygen atoms: maximum entropy was achieved by mixing 16O atoms in the substrate with 18O located inside the copper particles. We drew this conclusion from the fact that 18O-exchange also occurred during vacuum annealing of an oxidised sample, which is impossible in the case of the other proposed mechanisms. Also, 18O could be extracted from a metallic Cu2SiO sample through 16O2-oxidation, annealing of the sample in vacuo, and subsequent reduction of the sample.