point contact transistor

The point-contact transistor is the device that started the entire industry: built at Bell Labs in December 1947 by John Bardeen and Walter Brattain, it was the first solid-state amplifier anyone had ever demonstrated, and it worked without a single fabricated junction. Two sharpened metal wires, pressed onto a sliver of germanium close enough together that a human hand had to position them under a microscope, did the entire job that a modern transistor's precisely doped, lithographically defined structure does today.

The surprising physics is that it works at all on bulk material with no deliberately built structure. Germanium wafers of the era were n-type throughout, but surface states at the exposed face pin a thin, naturally occurring p-type layer right at the top — nobody doped it there on purpose. The emitter whisker, biased positive relative to the base, injects holes into that accidental p-layer. If the collector whisker sat far away, those holes would simply recombine with the surrounding electrons and vanish. The entire device depends on the collector sitting close enough that a usable fraction of the injected holes reach it before that happens.

That "close enough" is a real, quantifiable distance: the minority-carrier diffusion length. Holes injected into n-type germanium travel a characteristic distance before recombining, set by

where $D_p$ is the hole diffusion coefficient and $\tau_p$ is the minority-carrier lifetime in germanium. Bardeen and Brattain's whisker spacing — on the order of two thousandths of an inch, roughly 50 micrometers — was not an arbitrary assembly tolerance. It was chosen because it sits inside germanium's own $L_p$, and the fraction of emitted holes that survive the crossing to be collected falls off sharply once spacing exceeds it:

Push the whiskers too far apart and $\alpha$ collapses toward zero — no amplification, just two independent diodes sharing a block of germanium. Bring them too close and they short together mechanically. The device only works inside a narrow geometric window that a trained hand had to find by trial and error on every single unit.

That reproducibility gap is exactly why the point-contact transistor's reign lasted barely a year. Shockley's junction transistor replaced two mechanically positioned whiskers with a p-n junction whose width was set by how long a wafer sat in a furnace — a process parameter instead of an assembly tolerance, controllable to the same spacing again and again without a microscope or a steady hand. The underlying physics — minority carriers injected on one side, surviving a diffusion length, collected on the other — never changed; only the method of holding the critical distance constant did, and that one change is what turned a laboratory curiosity into an industry.

Read the point-contact transistor through a *spacing-as-base-width* lens rather than a *crude-prototype* lens: $\alpha \approx e^{-s/L_p}$ is the same current-gain relationship every junction transistor and MOSFET generation after it has also obeyed, just with a hand-placed gap between two tungsten whiskers standing in for what a diffused junction, and later a photolithographically printed gate, would soon control far more precisely. The number that decided whether the first transistor amplified at all — a spacing close to the minority-carrier diffusion length — is the same number every subsequent device generation has spent more than seven decades learning to control more tightly, not escape.