Exploiting Charge Sharing for Circuit Radiation Hardening

5 concepts8 papers1 proposal

As semiconductor manufacturing scales down to the nanometer level, electronic circuits become increasingly vulnerable to radiation-induced anomalies known as Single Event Upsets (SEUs). When a high-energy particle strikes a sensitive node in a transistor, it deposits a localized charge cloud that can flip a memory bit or create a transient voltage pulse. Historically, aerospace and high-reliability systems protected themselves using heavy structural redundancy, such as Triple Modular Redundancy (TMR), which replicates hardware three times and uses voting logic. However, TMR introduces massive power and area overheads.

A more elegant paradigm has emerged: instead of fighting the deposited charge or multiplying the hardware, circuits can be designed to inherently neutralize or reverse the effects of a radiation strike by exploiting the physical layout, charge sharing dynamics, and localized feedback loops.

Harnessing the Physics of SEU Reversal

When a particle strikes an off-state transistor, it typically causes an unwanted voltage drop or rise. However, if another transistor is physically close enough, it will collect a portion of that same charge cloud—a phenomenon known as charge sharing.

Minxuan Zhang et al. demonstrated through 3D TCAD simulations that charge sharing does not have to be a destructive parasitic effect; instead, it can actively reduce a cell's SEU-sensitive area. By shortening the physical spacing between the drains of complementary transistors (a technique called proximity-based layout hardening), an active "On-PMOS" transistor can capture the shared charge to generate a restorative compensation current, counteracting the voltage drop at the struck node.

Taking this a step further, Minxuan Zhang et al. introduced the concept of Single Event Upset Reversal (SEUR) through "inactive transistor bridging." By inserting an electrically inactive dummy transistor between functional components, the authors bridged the physical gaps between transistor drains without altering the circuit's logical behavior.

This structural modification sets up a competitive charge-collection race. If a radiation event occurs, the adjacent passive device captures the shared charge. If it collects more charge over a longer duration than the primary struck node, it triggers a secondary state flip that snaps the memory cell back to its original, correct state, reducing the SEU-sensitive area by over 31%.

[ Struck Node ] ──(Initial Error Charge)──> [ Potential Bit Flip ]
       │
 (Charge Sharing via Inactive Dummy Bridge)
       ▼
[ Adjacent Passive Node ] ──(Restorative Current)──> [ SEU Reversal / State Restored ]

Common-Mode Leakage Cancellation

The utility of dummy transistors extends beyond layout-level transient charge collection into the domain of continuous electrical compensation. In analog and power management circuits, radiation manifests not just as transient bit-flips, but as a cumulative increase in leakage current due to Total Ionizing Dose (TID) effects.

To protect sensitive timing circuits, Luo et al. developed a radiation-hardened Buck converter that utilizes a dummy transistor to mirror the leakage current of a critical timing switch. By forcing identical gate-to-source and drain-to-source voltages across both the active switch and the dummy transistor, the authors ensured that radiation-induced degradation affects both components equally. Because the dummy transistor experiences identical electrical stress, its leakage current can be subtracted from the timing node, effectively treating the radiation-induced leakage as a common-mode error that cancels out mathematically.

Self-Recovering Circuit Topologies

When moving from layout physics to circuit topology, the same philosophy applies: configure internal nodes so that a localized disturbance is automatically corrected by its neighbors.

For instance, Li et al. proposed a 14-transistor (14T) SRAM cell that relies on interconnected feedback paths linking redundant latch pairs. If a radiation strike alters the voltage at one node, the unaffected redundant nodes provide an immediate restorative current, pulling the corrupted node back to its original logic state within nanoseconds.

Similarly, Guo et al. addressed vulnerabilities in DICE flip-flops by splitting control and set signals into multiple spatially redundant paths. By integrating these split paths with a self-correcting storage architecture, a transient pulse on any single input path is treated merely as a single-node upset. The underlying storage cell automatically filters out the false control pulse without requiring complex voting logic or extra synchronization hardware.

Low-Overhead Self-Checking and Protection

Even when error-checking circuitry is introduced, it must be protected from becoming a point of failure itself without incurring the 3x overhead of TMR.

Zwolinski et al. demonstrated that multiple sequential cells can share a single parity checker and transition detector. This shared architecture minimizes hardware overhead while ensuring that any parity change triggers a system recovery before corruption spreads. To ensure the integrity of the protection logic itself, Gao et al. utilized a duplication-with-comparison architecture for the monitoring modules. This self-protecting logic prevents an SEU inside the recovery module from propagating incorrect data.

Even in the analog domain, preventing error accumulation follows a similar principle of dynamic compensation. Neugebauer et al. showed that implementing reciprocal charge movement between electrodes in charge-coupled device (CCD) architectures mitigates the impact of incomplete charge transfer, ensuring that residual charge is captured in subsequent cycles to maintain computational reliability.

Key Foundational Work

Effect of charge sharing on SEU sensitive area of 40-nm 6T SRAM cells (2014) — contributed 3 concept elements to this direction.
A Parallel Analog CCD/CMOS Signal Processor (1991) — contributed 1 concept elements to this direction.
History & citationVersion 1 · 2026-10-04

Current · Permanent link

View frozen version 1

Cite this direction

Latest version (a living page):

@misc{lacuna_inactive_transistor_bridging_4cda391a,
  title = {inactive transistor bridging},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/inactive-transistor-bridging/rdr_4cda391a4954449b87c73938eba66b0b},
  note = {Latest version of a living research direction; cite a version for a fixed text. Accessed 2026-11-19},
}

This exact version (version 1, frozen text):

@misc{lacuna_inactive_transistor_bridging_4cda391a_v1,
  title = {inactive transistor bridging},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/inactive-transistor-bridging/rdr_4cda391a4954449b87c73938eba66b0b/v/1},
  note = {Version 1, frozen 2026-10-04. Accessed 2026-11-19},
}
About

These directions overlap with this one through shared papers.

See something wrong? Report a mistake