EUVIMEDEUVIMED entdecken
Crossref

Quasi Event-Wise Measurement and Simulation of Neutron-Induced Multiple-Cell Upsets in 22- and 55-nm SRAMs

Yuibi Gomi, Kazusa Takami, Ryuichi Yasuda, Hiroki Kanda, Mitsuhiro Fukuda, Masanori Hashimoto

IEEE Transactions on Nuclear Science · 2026-08 · 73 (8) · 2935-2947

DOI: 10.1109/tns.2026.3675003 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Transactions on Nuclear Science, 2026-08. Verfasst von Yuibi Gomi und 5 weiteren Autor:innen. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 35 Literaturangaben. Gefördert von: Grant-in-Aid for Scientific Research (S) from Japan Society for the Promotion of Science, JST BOOST.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

We experimentally investigated neutron-induced multiple-cell upsets (MCUs) in 22- and 55-nm bulk planar static random access memories (SRAMs) using self-scan circuits that detect bit flips at intervals of several hundred nanoseconds. This method accurately detects single event upsets (SEUs), including Distant MCUs generated by nuclear reactions, which are defined as MCUs with bit flips separated by more than one cell pitch. We quantitatively demonstrate that conventional static bitmap-based analysis faces an inherent trade-off between excluding Pseudo MCUs, in which multiple independent SEUs are erroneously grouped into a single event, and capturing Distant MCUs, making accurate event identification difficult. Our irradiation results show that 13.0% of the 516 MCU events were Distant MCUs for 22 nm at 0.7 V. Maximum MCU cluster sizes reached $7.0~\mu $ m for 22 nm and $15~\mu $ m for 55 nm at 0.7 V, with Pseudo MCU probability ratios on the order of $10^{-11}$ . This ratio represents the relative likelihood of multiple SEUs occurring during a full SRAM readout compared to the cases of only a single SEU occurring under a Poisson process. We also found that the bitmap-based method requires a small grouping distance for accurate SEU counts, preventing Distant MCU characterization. Particle and heavy ion transport code system (PHITS)-based simulations and particle-level analysis identified that approximately 78% of the observed Distant MCUs are attributed to alpha particles.

Förderung

  • Grant-in-Aid for Scientific Research (S) from Japan Society for the Promotion of Science
  • JST BOOST

Zitieren

APA

Gomi, Y., Takami, K., Yasuda, R., Kanda, H., Fukuda, M., & Hashimoto, M. (2026). Quasi Event-Wise Measurement and Simulation of Neutron-Induced Multiple-Cell Upsets in 22- and 55-nm SRAMs. IEEE Transactions on Nuclear Science, 73(8), 2935-2947. https://doi.org/10.1109/tns.2026.3675003

Vancouver

Gomi Y, Takami K, Yasuda R, Kanda H, Fukuda M, Hashimoto M. Quasi Event-Wise Measurement and Simulation of Neutron-Induced Multiple-Cell Upsets in 22- and 55-nm SRAMs. IEEE Transactions on Nuclear Science. 2026;73(8):2935-2947. doi:10.1109/tns.2026.3675003

BibTeX
@article{Gomi2026,
  author = {Gomi, Yuibi and Takami, Kazusa and Yasuda, Ryuichi and Kanda, Hiroki and Fukuda, Mitsuhiro and Hashimoto, Masanori},
  title = {Quasi Event-Wise Measurement and Simulation of Neutron-Induced Multiple-Cell Upsets in 22- and 55-nm SRAMs},
  journal = {IEEE Transactions on Nuclear Science},
  year = {2026},
  volume = {73},
  number = {8},
  pages = {2935--2947},
  doi = {10.1109/tns.2026.3675003}
}
Kostenlos starten →