A new approach to the application of the well-known charge pumping technique is proposed as a tool for the measurement of interface trap energy distributions in small-area MOS transistors. In the classical procedure, the bandgap is scanned by applying pulses with variable transition times to the gate of the transistor, which define different energy windows from which the charge-pumping signal is measured. The new approach is spectroscopic in nature, i.e., only one energy window is defined, and forced to move through the bandgap by changing the sample temperature. This method has the advantages of addressing a larger part of the bandgap as compared to the classical approach, to reduce complication in the processing of the data, and to yield information about the hole and electron capture cross sections separately. Experiments performed on both n-channel and p-channel MOS transistors reveal that, in the temperature (energy) range studied, the interface-trap distribution is slowly varying with energy and that the trap capture cross section is nearly constant over energy and temperature.