Scaling limit of digital circuits due to thermal noise
Author(s) -
Kenji Natori,
Nobuyuki Sano
Publication year - 1998
Publication title -
journal of applied physics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.699
H-Index - 319
eISSN - 1089-7550
pISSN - 0021-8979
DOI - 10.1063/1.367317
Subject(s) - node (physics) , scaling , noise (video) , electronic circuit , dissipation , digital electronics , energy (signal processing) , circuit reliability , reliability (semiconductor) , electronic engineering , computer science , electrical engineering , physics , mathematics , engineering , statistics , power (physics) , geometry , quantum mechanics , artificial intelligence , image (mathematics) , thermodynamics
The error probability at a node of a digital circuit exposed to thermal noise agitation is investigated and the minimal dissipation–reliability relation for practical electronic circuits is derived. The digital circuit is modeled by an inverter chain with ideal transfer characteristics, and the error probability due to spurious data transfer caused by the thermal noise fluctuation is evaluated as a function of the node switching energy. The maximal error probability at each node allowed by the reliability requirement of the total system leads us to the minimal node energy dissipated per logical switching, which amounts to around 12 eV in the future 1010 gate system operated at a 10 GHz clock rate with a 104 FIT level reliability. In view of the device size-scaling trend of large-scale integrated circuits, the minimal node energy is expected to be reached at a feature size of 10–20 nm
Accelerating Research
Robert Robinson Avenue,
Oxford Science Park, Oxford
OX4 4GP, United Kingdom
Address
John Eccles HouseRobert Robinson Avenue,
Oxford Science Park, Oxford
OX4 4GP, United Kingdom