Breakdown and charge trapping in silicon dioxide films on silicon

Several aspects of breakdown and charge trapping in silicon dioxide (SiO2) have been studied. Firstly, the locations of charge trapped immediately prior to breakdown and of defects created under the high field stress were established using the photo I-V and avalanche injection techniques. Both positive and negative charge was found and in all cases charge build-up under high field conditions was at the interfaces, not in the bulk of the SiO2. Electron trap creation occurred predominantly near to the non-injecting interface, Qbd, the total charge which can be injected prior to breakdown, was examined under different current injection conditions. It was found to be strongly dependent on the duty cycle, the temperature and the gate electrode and to vary as 1/Eox , where Eox is the average field across the oxide. The Weibull distribution was found to describe well the statistics of breakdown in both wearout and dielectric strength measurements. It was shown that the Weibull parameters a and b have the same values whichever of these methods is used to measure them. The breakdown mechanism is probably the same in both wearout and dielectric strength measurements, therefore.
Oxide degradation was also examined under the less severe conditions of bias-temperature strength (BTS). It was shown, using avalanche hole injection before and after negative BTS, that the positive charge generated during negative BTS is due to trapping of holes in intrinsic hole traps. This is accompanied by interface state
generation across the whole band gap. The density of these states is linearly proportional to the number of holes trapped. On subsequent application of a positive BTS, the holes were all detrapped or neutralised. At this time a peak also appeared in the interface state density at - 0 . 2 eV above midgap. This may be due to a redistribution of previously generated states rather than to the creation of new states. The charge pumping technique was used to show that the peak is in fact due to interface states and not to lateral non-uniformities in the surface potential.

A thesis submitted to the C.N.A.A. in partial fulfilment of the requirements for the degree of Doctor of Philosophy.