Enhanced formation of shallow donors (SDs) in hydrogen or helium-irradiated and subsequently annealed float‐zone n-type silicon is investigated. Ion energies, irradiation fluences and annealing temperatures were chosen in ranges typically used for local lifetime control in silicon power devices. Introduced radiation defects and SDs were investigated by deep-level transient spectroscopy and C–V profiling. Results show that radiation damage produced by helium ions remarkably enhances formation of thermal donors (TDs) when the annealing temperature exceeds 375°C, i.e. when the majority of vacancy-related recombination centres anneal out. Proton irradiation introduces hydrogen donors (HDs) which form a Gaussian peak at the proton end-of-range. Their concentration linearly increases with proton fluence and changes dramatically during post-irradiation annealing between 100 and 200°C since HD constituents are reacting with radiation damage. Their annealing in this temperature range is influenced by the electric field. If annealing temperature exceeds 400°C, HDs disappear and the excessive shallow doping is caused, as in the case of helium irradiation, by TDs enhanced by radiation damage. Shallow doping introduced by both hydrogen and helium can have a detrimental influence on blocking voltage of power diodes if high irradiation fluences or wrong annealing conditions are chosen.

A two-dimensional (2-D) analytical expression for the surface potential and a mathematical relation for the calculation of the threshold voltage Vt and subthreshold slope S of symmetric double-gate (DG) n-MOSFETs have been derived taking into account the effect of bandgap narrowing due to heavy channel doping and quantum-mechanical (QM) effects. The DG MOSFET has been simulated using the device simulator ATLAS and the simulation values of Vt and S have been extracted for different values of structural dimensions, channel doping concentrations and gate materials. Our analytical results have been validated with 2-D numerical simulation data. Calculated data match well with the simulation results for various devices ensuring the validity of our model.