In this paper we comment on the interpretation ofCVdata in the presence of deep levels or interfacial layers. In order for the barrier height extracted fromCVmeasurements to be reliable, no frequency dependence should exist at the measurement frequencies, and the slope of the 1/C2vs voltage curve should not be temperature dependent. This is illustrated by data from Schottky barriers on 6H–SiC. ©1996 American Institute of Physics.