Electrical Resistance of Thin Triode-Sputtered Gold Films

Abstract

Freshly prepared triode‐sputtered gold films thicker than 200 Å on glass substrates exhibited a constant resistivity, about four times that of bulk gold, up to 4000 Å, the thickest films made. Annealing at 300°C for 4 h reduced their resistivity to twice that of bulk gold. Resistivity vs temperature curves were linear and parallel to the bulk gold curve. Nucleation of the films were extremely rapid: Films deposited onto substrates at 70°C had resistivity only twice the final value at <70 Å thickness. The gold also appeared to renucleate on grown parts of the film, since 2000 Å thick films exhibited no preferred orientation and had small crystals, 100 to 200 Å in the plane of the film, probably ∼60 Å thick. Film resistivity attained its final value as soon as a film became continuous, at about 100 Å with substrate at 70°C and about 250 Å at 250°C. This cannot be explained by invoking lamellar crystals, where the thickness of a single crystal is to be considered and not that of the whole film, since this would call for much higher film resistivity than observed. The alternative is that the film surfaces reflect conduction electrons specularly. This behavior is in marked contrast to that observed for solid films prepared by other techniques. The differences in behavior are fully discussed and reasons for the discrepancies are suggested.