Charge sharing between adjacent devices can lead to increased Single Event Upset (SEU) vulnerability. Key parameters affecting charge sharing are examined, and relative collected charge at the hit node and adjacent nodes are quantified. Results show that for a twin-well CMOS process, PMOS charge sharing can be effectively mitigated with the use of contacted guard-ring, whereas a combination of contacted guard-ring, nodal separation, and interdigitation is required to mitigate the NMOS charge sharing effect for the technology studied

We have performed capacitance-voltage (C-V) and thermally-stimulated-current (TSC) measurements on non-radiation-hard MOS capacitors simulating screen oxides of modern bipolar technologies. For O-V irradiation at /spl sim/25/spl deg/C, the net trapped-positive-charge density (N/sub ox/) inferred from midgap C-V shifts is /spl sim/25-40% greater for low-dose-rate (100 rad(SiO/sub 2/)/s) exposure. Device modeling shows that such a difference in screen-oxide N/sub ox/ is enough to account for the enhanced low-rate gain degradation often observed in bipolar devices, due to the /spl sim/exp(N/sub ox//sup 2/) dependence of the excess base current. At the higher rates, TSC measurements reveal a /spl sim/10% decrease in trapped-hole density over low rates. Also, at high rates, up to /spl sim/2.5-times as many trapped holes are compensated by electrons in border traps than at low rates for these devices and irradiation conditions. Both the reduction in trapped-hole density and increased charge compensation reduce the high-rate midgap shift. A physical model is developed which suggests that both effects are caused by time-dependent space charge in the bulk of these soft oxides associated with slowly transporting and/or metastably trapped holes (e.g. in E/sub /spl delta//' centers). On the basis of this model, bipolar transistors and screen-oxide capacitors were irradiated at 60/spl deg/C at 200 rad(SiO/sub 2/)/s in a successful effort to match low-rate damage. These surprising results provide insight into enhanced low-rate bipolar gain degradation and suggest potentially promising new approaches to bipolar and BiCMOS hardness assurance for space applications.

In this paper, we describe a Monte Carlo approach for estimating the frequency and character of single event effects based on a combination of physical modeling of discrete radiation events, device simulations to estimate charge transport and collection, and circuit simulations to determine the effect of the collected charge. A mathematical analysis of the procedure reveals it to be closely related to the rectangular parallelepiped (RPP) rate prediction method. The results of these simulations show that event-to-event variation may have a significant impact when predicting the single-event rate in advanced spacecraft electronics. Specific criteria for supplementing established RPP-based single event analysis with Monte Carlo computations are discussed.

Hydrogen is known to passivate Si dangling bonds at the $\mathrm{Si}-\mathrm{SiO}{}_2$ interface, but the subsequent arrival of $\mathrm{H}{}^{+}$ at the interface causes depassivation of Si-H bonds. Here we report first-principles density functional calculations, showing that, contrary to conventional assumptions, depassivation is not a two-step process, namely, neutralization of ${\mathrm{H}}^{+}$ by a Si electron and subsequent formation of an $\mathrm{H}{}_2$ molecule. Instead, we establish that $\mathrm{H}{}^{+}$ is the only stable charge state at the interface and that ${\mathrm{H}}^{+}$ reacts directly with Si-H, forming an $\mathrm{H}{}_2$ molecule and a positively charged dangling bond ( $P_b$ center). As a result, H-induced interface-trap formation does not depend on the availability of Si electrons.

Direct ionization from low energy protons is shown to cause upsets in a 65-nm bulk CMOS SRAM, consistent with results reported for other deep submicron technologies. The experimental data are used to calibrate a Monte Carlo rate prediction model, which is used to evaluate the importance of this upset mechanism in typical space environments. For the ISS orbit and a geosynchronous (worst day) orbit, direct ionization from protons is a major contributor to the total error rate, but for a geosynchronous (solar min) orbit, the proton flux is too low to cause a significant number of events. The implications of these results for hardness assurance are discussed.