Abstract
Evaporated gold and tantalum films have been bombarded with argon ions. In the case of gold films this resulted in an increase of the sheet-resistance by sputter etching to a maximum of 40 kO/[square] . The strain-gauge coefficient of resistance (gamma) (i.e. the fractional change in resistance per unit strain) was measured for films with a wide range of sheet-resistance, and was found to be almost invariant with an average of 2.6. This contrasts. greatly with the published values of gamma of up to 100 for thin island-structure evaporated films of similar sheet-resistance. The temperature coefficient of strain-gauge factor (beta) was found to be similar in magnitude but opposite in sign to the temperature coefficient of resistivity (alpha), which was measured as +12 x 104/°C. The measured values of gamma, beta and alpha agree well with values calculated assuming metallic conduction modified by reduction of the electron mean- free-path. We, therefore, conclude that a connected metallic layer still exists at very high values of sheet-resistance. In the case of tantalum films (that contain 30 atomic per cent oxygen) conduction was found to be by a combination of metallic and activated tunnelling. In the latter case there is some evidence for an increase in importance of this mechanism with oxygen concentration and for the existence of at least two activation energies. After bombardment with low doses of argon the resistivity (p) and also alpha shifted markedly towards values expected of very pure tantalum films, probably as a result of radiation enhanced diffusion and preferential sputtering of oxygen combined with re-arrangement of the film to form large precipitates of b.c.c. tantalum. There was also a significant increase in gamma (from an average of 3 up to 5.2) at similar doses, possibly as a result of changes in the microstructure increasing the importance of strain induced changes in the metallic conduction paths in the film. The metallic phase appears to be metastable as p increases with time (up to two years) at room temperature, probably due to reaction with oxygen near the film/substrate interface. For higher doses p drops (but not in a way explainable by sputter etching) alpha changes from a large positive value to a small negative one, and gamma drops towards a value of 2 which is the predicted value for Ta. on glass. These results 'indicate that a single phase, stable, low sputtering rate compound is formed, probably by reaction with the glass substrate. This compound has a mixture of activated tunnelling and metallic conduction with the strain gauge factor apparently determined by the metallic component of conduction. Some preliminary attempts at electrical depth profiling of the bombarded tantalum films by anodization are reported and the results support the models proposed above.