No abstract

No abstract

International audienceThis work describes test structures and modeling which have been applied to characterize SET pulse widths of standard ATMEL CMOS libraries in 180 and 90-nm technologies. The modeling methodology from physical-level to electrical-level induced SEU/MBU and SET, is also presented. The methodology includes the GDS extractor, the physical (carrier generation, transport and charge collection) and the circuit levels. The devices and test vehicles characterized in this work range from inverter to fully functional SRAMs integrating standard and hardened circuits. Then, this work analyses SET measured on test structures including circuits designed and fabricated in 180–nm and 90–nm processes. Results obtained for a 180–nm SRAM cell and a DFF are consistent. For the 90–nm cells, results obtained are less satisfactory. Nevertheless, the comparisons between the experiment and the calculation about the impact induced by the drive strength or the load configuration are consistent. Results issued from complex cell (i.e. customized DFF) show that modeling is relevant for designing efficient hardening cells for aerospace applications. Indeed, the SET or SEU sensitivity maps, the cross sections, and the SET distributions give precious information to the radiation expert designer about the efficiency and reliability of the chosen hardened techniques

The ATLAS simulation tool from SILVACO is used for 3D simulations. This software solves the transient semiconductor device equations in three dimensions, including the Poisson equation and the current continuity equations. The technology modeled in this work is a CMOS Bulk 0.18 urn process manufactured by ATMEL and transistors are designed for 1.8 V operation bias. The 3D inverter is modeled on a p-type silicon substrate. The doping profiles have been obtained by process simulation with ATHENA in order to obtain a realistic inverter device. The 3D structure is in agreement with the layout dimensions of this technology as shown on Fig. 1. The size of the NMOS and PMOS transistor,W/L, is respectively 1.4 and 2. For all simulations, the following models were used: Fermi-Dirac statistics, bandgap narrowing effect (BGN), Shockley-ReadHall (SRH) and Auger recombination, impact ionization, and the Lombardi mobility [11] which takes into account the mobility variation with the normal and parallel electric field, doping concentration and temperature. In ATLAS, the magnitude decreases. Thus, one can presume that the wide temperature range of a spatial environment (from 218 to 418 K) can modify the shape of the SETs. In this paper, device simulations are performed at 218,300 and 418 K on a CMOS Bulk 0.18 urn inverter from ATMEL. We investigate the variations of the heavy-ions induced SETs for various impact locations at the three considered temperatures. This paper focuses on the voltage pulse width because its value is often used as a critical parameter in designing SET-mitigation circuits [2]. In the part II of the paper, the 3D modeled device and the simulation conditions are described. The part III shows the influence of the temperature on the SET pulse for a track crossing the off-NMOS drain (i.e. the sensitive region) and for a track striking the device far from this sensitive region. In the part IV, we complete the investigation with Spice simulations in order to define the implication of the temperature on the SETs propagation through a 10-inverters logic chain.Index Terms -Single event transient (SET), heavy-ion, temperature dependence, CMOS, LET, technology computeraided design (TCAD) simulation.Abstract-In this paper, a study by device simulation of the heavy-ion induced Single-Event Transients (SET) is realized in the 218 -418 K range in order to determine the temperature effect on a CMOS bulk 0.18 11m inverter. An investigation of the SET propagation though a 10-inverters logic chain is performed, and the threshold LET (LETth) required for unattenuated propagation through the inverters chain is determined. The LETth is calculated for two different location of the heavy ion impact and for three tern peratures. An increase of the sensitivity is found when the temperature raise from 218 to 418 K.