X-ray photoelectron diffraction study of thin Cu films grown on clean Ru(0001) and O-precovered Ru(0001)

We have studied the epitaxial growth modes and near-surface interlayer relaxation of thin Cu films on Ru(0001) using X-ray photoelectron diffraction (XPD), measuring experimental Cu 2p 3/2 ( E kin=556 eV) and Ru 3d ( E kin=1206 eV) intensities over one-third of the nearly full 2 π solid angle above the surface for Cu coverages from submonolayer up to 40 monolayers. Reference Cu 2p 3/2 data for a clean Cu(111) surface have also been obtained from Naumović et al. and in our laboratory. These data have been compared to single scattering cluster (SSC) and more accurate multiple scattering cluster (MSC) calculations via a sum of five R-factors to derive precise structural information. MSC calculations are found to give a more accurate description for layers of ≥4 ML thickness, and comparisons of experiment and theory are also improved by allowing more accurately for the effective degree of angular averaging involved. Calculations for thicker layers are also found to converge by ∼5 ML. Our analysis indicates that the first Cu layer grows pseudomorphically on Ru(0001), in agreement with prior studies. An R-factor analysis comparing MSC and SSC calculations to experimental results further indicates that the Cu–Ru interlayer spacing at 1 monolayer (ML) is about 2.15 Å, in excellent agreement with prior low-energy ion scattering (LEIS) and low-energy electron diffraction (LEED) experimental studies, as well as with prior linearized augmented plane-wave (LAPW) calculations. At higher coverages, comparison of our data to SSC and MSC calculations for various atomic clusters indicates that the short-range structure is fcc Cu(111)-like, but with significant interlayer contraction which persists up to ≥5 ML coverage. Prior STM work by Behm et al. has shown a series of misfit dislocation structures in the top layer of the Cu film at higher coverages from 2 to 4 ML. Our data indicate that these misfit dislocation structures thread to the Cu/Ru interface rather than occurring only in the top Cu layer or layers. An R-factor comparison of the more accurate MSC calculations to experiment also indicates that the ratio of the Cu–Cu interlayer distance ( d ⊥) to the Cu–Cu in-plane nearest-neighbor distance ( d ‖), d ⊥/ d ‖=0.729±0.034 at 2 ML, and reaches 0.777±0.020 by 25 ML. For reference, the bulk value is d ⊥/ d ‖=0.816, and the analysis of experimental data for Cu(111) yields 0.801±0.035, in good agreement with this value and prior LEED studies. This analysis shows that there