Quasi Saturation

Quasi-saturation is what happens when Kirk effect's base push-out runs out of room — the effective base, widening as collector current climbs, eventually consumes the entire lightly-doped collector epitaxial layer and reaches the heavily-doped buried layer beneath it, leaving the transistor behaving exactly like it's in classical saturation even though the base-collector junction, judged by its own external terminal voltages, has never actually gone into forward bias the way true saturation requires.

The progression from Kirk effect to quasi-saturation is continuous, not a separate mechanism kicking in at some threshold — it is the same base push-out carried far enough that it exhausts the available collector material. Kirk effect pushes the effective base-collector junction deeper into the epi region as collector current rises, and as long as some undepleted, lightly-doped epi remains between that pushed-out junction and the buried layer, the transistor still functions as a nominally active-region device, just with a widening base eating into f_T. Push the current high enough, though, and that remaining undepleted epi region shrinks to nothing — the effective base now reaches the buried layer directly, and there is no lightly-doped collector material left for the junction to push into any further.

What makes this specifically a quasi-saturation rather than true saturation is where the apparent saturation behavior actually comes from. True, hard saturation is a circuit-level condition: an external circuit drives collector-emitter voltage low enough that the base-collector junction itself genuinely forward-biases, exactly the way the base-emitter junction already is, and both junctions being forward-biased simultaneously is the textbook definition of the saturation region. Quasi-saturation instead is a device-internal condition entirely — the base-collector junction, measured by its own terminal voltages, can still be nominally reverse-biased, consistent with forward-active operation, while internally the transistor behaves as if it were saturated anyway, because the collapsed undepleted epi region can no longer support current flow without a substantial resistive voltage drop.

This resistive voltage drop across the collapsing undepleted epi is exactly what produces the saturation-like symptom — a collapsing collector-emitter voltage and reduced effective current gain — without the junction itself ever forward-biasing in the classical sense. As the undepleted epi thickness shrinks under Kirk-effect pressure, whatever current the transistor is carrying has to cross an increasingly thin, increasingly resistive remaining path to reach the buried layer and substrate contact beneath it. That resistive path drops real voltage proportional to the current flowing through it, and at high enough current that voltage drop alone can account for most of what looks, from the outside, like a saturated transistor's collapsed V_CE — the external circuit never commanded saturation, the device's own internal geometry produced the identical symptom on its own.

This is specifically why a simple SPICE model using a single fixed collector series resistance badly mispredicts behavior in this regime. A basic bipolar model typically includes a lumped collector resistance term representing the ohmic path from the active region to the external collector terminal, treated as a fixed value independent of operating point — adequate for describing ordinary forward-active behavior, where the undepleted epi thickness barely changes with bias. Once quasi-saturation becomes relevant, that resistance is no longer fixed at all; it depends directly on how much undepleted epi thickness Kirk effect has already consumed at the current operating current, meaning the effective collector resistance itself rises with increasing current rather than staying constant — a current-dependent resistance a fixed-value model has no way to capture without an explicit quasi-saturation extension.

Power bipolar transistors are specifically prone to this failure mode, and the reason traces directly back to the same collector-doping tradeoff this series has already named. A transistor designed for high breakdown voltage needs a thick, lightly-doped collector epitaxial layer specifically to support a wide depletion region without the field reaching critical breakdown strength — but that same thick, lightly-doped epi is exactly the structure Kirk effect has the most room to push through before running out of material, and a thicker epi layer also means quasi-saturation's resistive voltage drop, once triggered, has more collapsed-epi length to account for. A power BJT optimized for high voltage handling is, almost unavoidably, also a device that reaches quasi-saturation at a comparatively achievable current density — the same doping choice that buys breakdown voltage margin costs current-handling headroom before this effect sets in.

Quasi-saturation closes the chain this series has built from Early voltage through Kirk effect — a bipolar transistor's effective base width is not fixed, Kirk effect shows it can widen under current, and quasi-saturation is simply what happens when that widening is pushed to its physical limit, consuming the entire collector epi and producing saturation-like symptoms the external circuit never actually commanded. It is the device geometry itself reaching a wall, independent of whatever the terminal voltages alone would suggest the transistor's operating region should be.