On the channel-length dependence of the hot-carrier degradation of n-channel MOSFETs

The channel-length dependence of lifetime plots is analyzed. It is shown that no unique tau *I/sub d/ versus I/sub sub//I/sub d/ relation can be obtained when threshold-voltage shifts are used for measuring the lifetime. In contrast, when using charge pumping as a monitor for the degradation, the lifetime plot for a given technology proves to be independent of the channel length.<<ETX>>