Features and mechanisms of the saturating hot-carrier degradation in LDD NMOSFETs

Abstract

Propagation of defects from the sub-spacer region to the gate-overlapped LDD region in NMOSFETs is modeled using measurements and 2-D device simulation. It is argued that the saturation of degradation is caused by the saturating nature of this degradation length, as opposed to decreasing lateral electric field maxima (E/sub m/) or increasing barrier height (/spl phi//sub it/) to defect creation. Two stage hot-carrier degradation was observed in our LDD NMOSFETs. The early mode (1000-3000 s) of the degradation is characterized by a sharp rate of degradation of the linear transconductance (g/sub m/), and a reduction in the substrate current (I/sub B/). In order to locate and quantify defects produced in this early mode degradation phase, we use the results of a combination of the floating gate technique and simultaneous measurements of the reverse (source and drain interchanged) saturation g/sub m/'s. These results help us build a 2-D simulation framework involving trapped negative charges in the oxide in the drain-side gate-edge region, partly under the gate and partly in the spacer region. We then use 2-D simulation and other measurements such as linear and saturation current degradation, I/sub B/ degradation, and charge pumping to confirm the location of the defects and help estimate their quantity. Simulation results also help us build an analytical model for defect propagation from the early mode to the late mode. The analytical model is seen to explain many features of the saturating nature of hot-carrier degradation.