Dynamic scaling is obeyed in many continuum theories and computer models describing driven systems such as non-equilibrium growth fronts in thin film deposition. Few systematic experimental tests of this hypothesis have been performed to date.

As a potential model for metallic films we investigate the growth of quenched Xe films formed by deposition onto a cold (15-35 Kelvin) SiO2/Si substrate. In situ measurements of specular and diffuse x-ray reflectivity and diffraction line shapes are performed with a position sensitive detector. We determine film surface roughness, electron density profile, degree of conformality with the substrate, and crystallinity as a function of Xe film thickness - ranging from only a few Angstrom up to 1000 Angstrom - and annealing history at substrate temperatures of 15-18, 25, and 35 K. Deposition of polycrystalline films with random crystallite orientation is achieved through heating of a grafoil sheet that is enclosed in the sample cell and has previously been covered with bulk Xe to saturation pressure at 125 K. We developed a protocol that allows us to repeatedly desorb films and grow them on a reproducible substrate state. Below 18 K the film morphology remains stable over several days, while at 25 K annealing takes place. Different atomic mobilities at different temperatures lead to complex scaling behavior of roughness with thickness. Real-time reflectivity measured during deposition confirms the trends seen in the films after the growth has ended. Analysis of the diffuse reflectivity with the distorted-wave Born approximation (DWBA) gives results conflicting with parameters obtained from the fitting of the specular reflectivity and in many cases cannot reproduce the diffuse data. Different kinetic roughening behavior is observed at all three temperatures, but no convincing evidence for dynamic scaling is found. Below 18 K the observed low densities together with diffraction results support the idea of a porous film structure, while at the higher temperatures densities are close to bulk values.

Our experiment reveals a surprising variety of growth scenarios in a supposedly simple system, offering some insight into the complexities involving non-equilibrium thin film growth in heterogeneous systems.

Return to Ralf Heilmann's homepage