CMOS 1963 Parasitic Thyristor Hiding Inside Every Pair

# Confront a Failure Mode the Well Itself Makes Possible: A Parasitic Thyristor Hiding Inside Every Complementary Pair

## 1. Why the Same Structure That Enables This Series Also Threatens It

The four doping regions this series has built — the native p-substrate, Step 2's n-well, the PMOS p+ source/drain inside the well, and the NMOS n+ source/drain in the substrate — form, entirely unintentionally, a parasitic four-layer sandwich that behaves exactly like a thyristor: a PNP transistor whose emitter is the PMOS source/drain, base the n-well, and collector the substrate, cross-coupled with an NPN transistor whose emitter is the NMOS source/drain, base the substrate, and collector the n-well, each transistor's collector feeding the other's base in a loop. Under ordinary operation this parasitic path stays dormant, both parasitic devices held far from conduction by the same bias conditions that make the intended transistors work correctly. But a transient disturbance — a voltage spike, an injected current, a momentary drop across the resistance of the well or substrate itself — can forward-bias one junction just enough to turn on one parasitic transistor, whose own collector current then supplies exactly the base current the other parasitic transistor needs to turn on too, whose collector current feeds back into the first. Once that loop closes, it does not need the original disturbance anymore to sustain itself.

$$\beta_{\text{PNP}} \cdot \beta_{\text{NPN}} \geq 1$$

where $\beta_{\text{PNP}}$ and $\beta_{\text{NPN}}$ are the current gains of the two parasitic bipolar transistors this step's structure forms — once their product reaches or exceeds unity, the cross-coupled loop becomes regenerative, each transistor supplying more than enough base current to keep the other fully on, and the structure latches into a low-impedance state connecting the supply directly to ground through a path neither transistor was ever meant to provide, sustained by nothing but the supply itself until it is physically removed.

A Loop That, Once Closed, Feeds Itself the cross-coupled parasitic pair, drawn as the circuit it actually is PARASITIC PNP, βPNP PARASITIC NPN, βNPN collector feeds base collector feeds base, the other direction emitter: PMOS S/D emitter: NMOS S/D βPNP · βNPN ≥ 1 — the loop no longer needs the original disturbance once this condition is met, the parasitic path sustains itself from the supply alone

## 2. Real Diagram: The Parasitic Structure, Drawn on Top of the Real One

The cross-section below is the identical structure Step 2 through Step 4 built, with the four regions responsible for the parasitic thyristor highlighted directly on the real device — this parasitic path is not a separate defect sitting somewhere else in the silicon, it is built from the same regions the intended transistors need.

Four Regions, Two Roles Each the same silicon that forms the complementary pair also forms the parasitic path SUBSTRATE — ALSO THE NPN BASE n+, NPN emitter N-WELL — ALSO THE PNP BASE p+, PNP emitter parasitic current path, dormant under normal bias every function this series needed from the well also, unavoidably, built this parasitic structure alongside it

## 3. The Same Motif 1961's Own Reach-Through Established, One Structure Later

The 1961 epitaxial series' own Step 7 confronted reach-through breakdown — a failure mode that existed only because the series' own epitaxial interface, the same structure responsible for that series' resistance improvement, also happened to create a surface a depletion region could reach under the wrong conditions. This step confronts the same kind of risk, one series and one structural idea later: the well this series needed to host a second transistor type, the exact structure responsible for the near-zero static current Step 6 measured, also unavoidably assembles four doping regions into a parasitic thyristor no earlier series in this project's history ever needed to worry about, because no earlier series had a well at all. Neither reach-through nor this parasitic path is a mistake introduced by careless fabrication — both are the direct, unavoidable byproduct of the very structural idea that makes the series' own benefit possible in the first place.

Step 7 does not discover a flaw in how this series built its well; it discovers that the well's own doping sequence, examined completely rather than one junction at a time, assembles a second device this series never intended to build, dormant until something disturbs it into sustaining itself.

Take cmos 1963 parasitic thyristor hiding inside every pair further

Ask the copilot about this term, or have our engineers assess it against your process.

Ask CFSGPT about this → Request a technical assessment → Open the CFS app