Solid/liquid interfaces

Abstract

This project was an investigation of certain solid/liquid interfaces using ellipsometry. Liquid particles adjacent to a solid surface are reasonably expected to show some degree of ordering, distinct from that found in the corresponding bulk liquid. Computer simulation of simple 'liquid against a wall' systems has indicated that particles close to the wall adopt a decaying periodic density variation which extends several diameters into the liquid. The principle object of this project was to determine, by a suitable choice of experimental solid/liquid systems, whether the interfacial profiles calculated in this way are realistic. In using visible light as a probe, ellipsometry only requires the incidence medium to be transparent, rather than evacuated, as necessary with most other surface study techniques. An optical study of the distribution of simple liquid particles adjacent to a solid surface could begin to fill a relatively large gap in the range of experimental data available relating to interfaces in general. There are some important limitations to ellipsometry however. Firstly, the interface is characterised by only a single parameter ρ in which all contributions to the dielectric variation between the bulk media are combined. The contributing aspects of an interface under investigation must then be limited in number and predictable by way of a continuum model in order that the measured signal may be reliably analysed. Secondly, ellipsometry measures the relative phase shift induced between p and s polarised waves of a light beam on reflection from the interface, but a much larger shift will also be induced if the substrate medium is strongly absorbing. Therefore both the incidence substrate media must be virtually transparent in order that sufficient sensitivity to the interface contribution to the signal will be obtained. A further requirement, common to all surface measurements, is that the intrinsic interface should not be obscured by the presence of roughness or dirt. It is however, difficult to determine to what extent this requirement has been met by an experimental system, and coupled with the first requirement the precision of quantitative measurements remain doubtful. In creating solid/liquid interfaces, this project sought to overcome the above difficulties simply and cheaply via the cleavage properties of various crystalline solids. In particular, alkali halide crystals were cleaved while immersed and already in contact with appropriate liquids. Their surfaces are then at least initially clean, the contamination usually caused by prior handling procedures being avoided. Unfortunately, cleavage of these crystals even when carefully controlled, did not lead to satisfactorily smooth or predictable interfaces. The surface of muscovite mica is known to be atomically smooth and was also considered as a possible substrate. Mica has often been used before in surface experiments but proved unsuitable for the present project because of an interfering reflection from the rear surface, which persisted in spite of various efforts to eliminate it. Pure liquids consisting of rigid symmetrical particles are clearly preferable in solid/liquid experiments of the present kind. The closest readily available real approximations to the simple spherical particles used by computer simulations were considered to be argon and carbon tetrachloride. These liquids are quite comfortably employed in experiments involving measurements by ellipsometry. A value of ρ ~ -10 -3 was calculated from a computer model which used particles in a liquid state corresponding reasonably well with the accessible states of liquid argon. The object of the project may be regarded as at least partially met in that interfaces consisting of KCI and KBr in argon returned a value of ρ ~ -3 x 10 -4. The magnitude of this value is reduced over that predicted by the simple liquid simulation mentioned above, probably because of the geometrical and electrostatic complications of real ionic crystal surfaces.