Speakers
Description
Abstract:
The probability of a radiation-induced bit-error in static random-access memory (SRAM),
expressed as the single-event-effect cross-section, has a strong dependence on linear energy transfer
(LET). This dependence is usually modelled using a cumulative Weibull distribution function, where
higher LET predicts a higher upset probability. A common approach for determining the parameters
of the Weibull distribution is by obtaining experimental values of the cross-section for specific values
of LET.The Variable-Depth Bragg-Peak (VDBP) method achieves this mainly by introducing different
thicknesses of degrader material into the pristine beam [1]. The controlled variable in this approach
is not LET but energy, and due to the stochastic nature of energy loss, the degraders do not reduce
the energy of all particles equally, which introduces dispersion. This effect is particularly strong at
low energies, which is where the VDBP method largely operates to achieve as high LETs as possible
for the given particle species. An energy spread translates into an LET spread, which for light ions
(e.g. carbon) can be over half of the achievable LET range.
We present an approach where the spread in LET is taken into account systematically instead of
appearing as an uncertainty. Instead of computing the LET for each thickness and fitting the Weibull
distribution to the resulting points directly, this approach computes a pre-dicted cross-section for
each degrader thick-ness using an assumed Weibull distribution and compares it to the experimental
value at that thickness. This enables the use of light ions for single-event-effect testing.
The potency of the method is shown by determining the Weibull parameters for experimental data of
an SRAM sample. The shape of the experimental cross-section over degrader thickness is sufficiently
unique, compared to the degrees of freedom of the Weibull distribution, to suggest that our indirect
method yields unambiguous Weibull parameters.
References:
[1] S. Buchner et al., “Variable Depth Bragg Peak Method for Single Event Effects Testing,” in IEEE
Transactions on Nuclear Science, vol. 58, no. 6, pp. 2976-2982, Dec. 2011