The effect of interface trapped charge on threshold voltage shift estimation for gamma irradiated MOS device

The ionizing radiation affects the silicon-based electronic devices and leads to the build-up of trapped hole in the oxide and an increase in trap density at the Si–SiO₂ interface. These defects cause degradation of device parameters such as threshold voltage shifts, sub-threshold swing and etc. Although the effect of oxide charges on the threshold voltage shift in MOS devices is dominant, accurate estimation requires for modeling of the interface trapped charge effect. In this work, the effect of interface trapped charge on threshold voltage shift in an N-channel MOS transistor device, irradiated at different total ionization doses, was estimated considering proton transport model in gate oxide. The analytical model calculations consider the time dependent buildup of trapped holes and interface trap charges. Furthermore, we used ATLAS as a numerical semiconductor simulation code which only allows modeling oxide trapped charge to calculate the variation of threshold voltage shift with TID without considering the interface trap. Computational results were compared with experimental results at several doses and gate biases. According to the experimental results, ATLAS overestimates threshold voltage shift within 30% approximately. The results based on considering proton transport mechanism in analytical model, showed significantly better agreement with experimental results for total dose level up to 2.87 krad at several different biases. Overall results indicated that considering interface trap charges modeling, substantially improves the accuracy of the threshold voltage shift estimation. - Highlights: • Variation of interface trap charge due to TID affects on threshold voltage shift. • Time dependent buildup of trapped holes and interface trap charges were calculated. • An analytical model considering proton transport have been used. • ATLAS simulations code was used to investigated the model performance. • Computational results were validated via comparison with the experimental data