SILVER PHOTODISSOLUTION IN AMORPHOUS-CHALCOGENIDE THIN-FILMS

In spite of its eventual use in very-large-scale integration techniques, silver photodissolution in amorphous chalcogenide thin films still remains an ill-understood phenomenon. In this paper we investigate the main theories reported through the use of the following compounds: AsS(x), AsSe(x) and GeSe(x), GeSe3 and GeSe5.5 films that were flash evaporated or deposited by plasma-enhanced chemical vapour deposition were used to conduct the research. Silver was thermally evaporated on or underneath the GeSe(x) layer. Photodissolution was assisted by visible or UV light exposure. The kinetics of photodissolution and the change in resistance of the remaining silver layer were analysed from the reflectivity measurements during the development process of the multilayer system (in order to explain the reflectivity data of the multilayer system during the photodissolution). The optical properties of the photo-doped layer and the undoped layer were investigated for GeSe3 and GeSe5.5. A comparative study of secondary ion mass spectroscopy and Rutherford backscattering spectroscopy data shows that the interface between the doped and undoped region is rather step like in GeSe5.5 and that the interface obeys a diffusional law in GeSe3. Furthermore, an investigation from successive Ag-107 and Ag-109 dissolutions shows that both isotopes have the same depth concentration profile in GeSe3 films as in the GeSe5.5 films. The silver concentration profile after photodissolution in non-saturated GeSe5.5 films exhibits two peaks near the surface and the GeSe5.5 substrate interface. In conclusion, a model is proposed to explain such behaviour: silver migrates from point defects of the initial silver metal-chalcogenide interface through to the chalcogenide layer. From these dendritic paths, lateral silver dissolution occurs up to a concentration C(po). When silver reaches the chalcogenide-substrate interface, because of the numerous dangling bonds, its concentration begins to increase, thus resulting in an intermediate region within the layer poor in silver. If the initial silver thickness is large enough, its concentration in the whole layer increases to the saturation value C(pf).