Environmental and thermomechanical stability of thin films for optical applications.

Abstract

We have selected two materials to study the stability of thin-films: (1) TbFe, a candidate material for optical data storage, for environmental stability study; (2) ZrO₂, a dielectric material for optical coatings, for thermo-mechanical stability study. In the research on TbFe sputtered films, we applied surface plasma resonance as a vehicle to study the optical constants of a single layer TbFe film and to study the instability of a multilayer system of TbFe, with protective layer Al₂O₃ and coupling layer MgF₂ (structure: glass/MgF₂/TbFe/Al₂O₃/air), as a function of time. The results show that with our multilayer system there was only slight environmental instability during the first day and the system stabilized thereafter. However the TbFe film did exhibit some oxidation on exposure to 200°C for two hours. Water, which may penetrate into the MgF₂ layer from the side may accelerate the oxidation. It is therefore necessary to have side protection and to avoid long period exposure to high temperature. In the research on ZrO₂ evaporated films, with and without ion-assisted deposition (IAD), we performed interferometry in a vacuum oven to study total stress of films as a function of temperature. On thermal cycling, all the plots of stress versus temperature for IAD and non-IAD films exhibit hysteresis. In order to understand the hysteresis, we studied microstructure and water effects. The results show that the likely mechanisms are water desorption, recrystallization and phase transformation and we believe that a combination of all three occurred. Our results also show that ion assisted deposition (increasing deposition temperature tends to give more tensile stress) and high deposition temperature (increasing deposition temperature tends to give less tensile stress) gave more stable films both thermo-mechanically and optically. It is well known that the thermal stress is due to thermal expansion coefficient mismatch between substrate and film. But if thermal expansion coefficients are to be derived from thermal stress, then great care must be taken to eliminate the water effect, otherwise, the results will be totally wrong. For better results, in-situ thermal stress studies are needed.