Yury Illarionov MSc Publications

The charged defects situated in different sections of the Metal-Oxide-Semiconductor Field-Effect Transistor (MOSFET) can have an extremely different impact on device performance, and thus the information on the lateral trap position is of great importance. We have developed a reliable method to evaluate the lateral coordinate of the defects in ultra-scaled MOSFETs. In contrast to other techniques, our method incorporates random dopants, which have a dramatic impact on potential distribution inside the device and render previous trap location techniques unusable. The approach is based on the observation that the dependence of the trap-induced threshold voltage shift vs. the drain bias is essentially more sensitive to the lateral trap position than to the random dopants (Fig. 1). The threshold voltage shift can be parameterized by a cubic polynomial of the drain bias with a unique set of coefficients. This unique set of coefficients can be treated as a fingerprint of the lateral trap position. Thus, by fitting the simulated threshold voltage shift vs. the drain bias characteristics with their experimental counterparts, we can evaluate the lateral coordinate of the defect. Also, the probability of finding the trap inside any finite interval centered at the determined coordinate point can be calculated. We demonstrated that despite the impact of random dopants, the lateral position of the trap can be evaluated with a precision of several nanometers. We have also proven that for the traps situated beneath the gate electrodes, the accuracy of coordinate evaluation is essentially higher, which is especially valuable for hot-carrier degradation induced traps. The verification of the method has shown that for devices with a lower channel doping level, the accuracy of the defect coordinate evaluation is higher. Finally, we have extracted the lateral positions of the traps from the experimental threshold voltage shift characteristics measured on several industrial devices. In addition, we have derived a compact model which allows for the simulating of the reference threshold voltage shift characteristics without time consuming Technology Computer Aided Design (TCAD) simulations. This leads to only a small loss in accuracy regarding the evaluation of the lateral trap position.

Fig. 1: TCAD simulated threshold voltage shift versus drain bias characteristics for devices with different random dopant configurations and a fixed trap position. Top: The trap is situated at the source side of the channel; Center: in the middle of the channel, and bottom: at the drain side of the channel. The red lines indicate the characteristics with average (solid) and average plus/minus standard deviation polynomial coefficients (dashed). The shape of the curves depends on the trap position much stronger than on the random dopants distribution, and thus can be used as a defect fingerprint.