Deep submicron n-channel metal-oxide-semiconductor field-effect transistors (NMOSFETs) with shallow trench isolation (STI) are exposed to ionizing dose radiation under different bias conditions.The total ionizing dose radiation induced subthreshold leakage current increase and the hump effect under four different irradiation bias conditions including the worst case (ON bias) for the transistors are discussed.The high electric fields at the corners are partly responsible for the subthreshold hump effect.Charge trapped in the isolation oxide,particularly at the Si/SiO 2 interface along the sidewalls of the trench oxide creates a leakage path,which becomes a dominant contributor to the offstate drain-to-source leakage current in the NMOSFET.Non-uniform charge distribution is introduced into a threedimensional (3D) simulation.Good agreement between experimental and simulation results is demonstrated.We find that the electric field distribution along with the STI sidewall is important for the radiation effect under different bias conditions. Deep submicron n-channel metal-oxide-semiconductor field-effect transistors (NMOSFETs) with shallow trench isolation (STI) are exposed to ionizing dose radiation under different bias conditions. The total ionizing dose radiation induced subthreshold leakage current increase and the hump effect under four different irradiation bias conditions for the transistors are discussed. The high electric fields at the corners are partly responsible for the subthreshold hump effect. Chassis trapped in the isolation oxide, particularly at the Si / SiO 2 interface along the sidewalls of the trench oxide creates a leakage path, which becomes a dominant contributor to the offstate drain-to-source leakage current in the NMOSFET. Non-uniform charge distribution is introduced into a threedimensional (3D) simulation. Good agreement between experimental and simulation results is demonstrated. We find that the electric field distribution along with the STI sidewall is important for t he radiation effect under different bias conditions.