The suppression of intrinsic carrier generation does not in itself constitute proof that a semimetal-to-semiconductor transition has occurred. However, we present a realistic statistical analysis that demonstrates that the Bi film data do in fact imply the opening of a ≊56-meV energy gap in the thinnest sample (200 Å) studied. The standard vanishing-wave-function model accounts for the experimental observations, whereas the alternative vanishing-gradient model yields results that are inherently inconsistent with the data.

• Received 2 June 1994

DOI:https://doi.org/10.1103/PhysRevB.51.5535