The role of temperature in sample-to-sample comparisons of the 1/f noise of metal films

We have performed an extensive study of the 1/f noise of several different kinds of metal films fabricated in different ways. An examination of the temperature dependence of the 1/f noise of tin and indium shows that the previously reported substrate dependence of the noise magnitude generally is only present near room temperature. At other temperatures, especially lower temperatures, the substrate dependence is much smaller or negligible. The noise magnitudes of platinum and bismuth films, on the other hand, depend strikingly on the way that the sample is made, and can vary strongly from sample to sample as a function of temperature. Taken together, these results illustrate that, in contrast to common practice, the room-temperature magnitude of the 1/f noise of metal films is often not a useful parameter on which to base comparisons of the relative noise magnitudes of different films. The implications of these results for the interpretation of some previous studies are discussed, and recommendations are made about the interpretation of future studies of the 1/f noise of metal films.