Point Contact 1947 Chemically Etch the Germanium

# Chemically Etch the Germanium: A Self-Smoothing Process, Not Just a Depth Removed

Step three left behind a ground surface whose roughness has to shrink well below the point-contact spacing before this device is usable — and the chemical etch this step performs doesn't just remove a uniform layer of damaged material, it actively smooths the surface through a mechanism entirely different from the mechanical self-correction that made step three's own grinding self-limiting. Where grinding's high spots wore faster because they saw higher mechanical pressure, this etch's high spots dissolve faster because they sit closer to fresh, unconsumed etchant — a diffusion-limited effect, not a pressure-driven one, but one that produces the same useful outcome: the surface actively levels itself rather than simply thinning uniformly.

## 1. Etch Rate Depends on How Easily Fresh Etchant Reaches the Surface

$$\text{etch rate} \propto \frac{D_{\text{etchant}}}{\delta_{\text{boundary layer}}}$$

The local rate at which germanium dissolves depends on the etchant's own diffusion coefficient $D_{\text{etchant}}$ divided by the thickness of the stagnant diffusion boundary layer the etchant has to cross to reach the surface at that point — and a protruding asperity, poking further into the bulk etchant, sits behind a thinner boundary layer than a recessed valley does. Fresh, unconsumed etchant reaches the peaks faster than it reaches the valleys, so peaks etch faster than valleys purely as a consequence of local geometry, with no mechanical pressure or contact force involved anywhere in the process.

## 2. Real Diagram: Peaks See Fresh Etchant, Valleys Don't

A Thinner Diffusion Layer Sits Over Every Peak no mechanical pressure here — the geometry alone decides etch rate as-ground surface, before this etch thin layer over peaks thicker layer over valleys valley, slower etch peak, faster etch the peaks erode toward the valleys, not the other way around

## 3. Roughness Decays Exponentially Toward the Target This Device Actually Needs

$$R_a(t) = R_a^{(0)}\, e^{-t/\tau_{\text{smooth}}}$$

Because peaks consistently etch faster than valleys, the surface's overall roughness $R_a$ doesn't just decrease — it decays roughly exponentially with etch time toward a characteristic smoothing time constant $\tau_{\text{smooth}}$, which gives this step a real, controllable lever: etch duration. Step three established that this surface's roughness has to stay far below the point-contact spacing it will eventually host, and this step's etch time is chosen specifically to drive $R_a(t)$ down past that threshold — not etched for a fixed, universal duration, but etched until the smoothing this exponential relationship predicts has actually caught up with what the device's own geometry demands.

Roughness Falls Exponentially With Etch Time etching stops once the curve crosses step three's own roughness requirement etch time, t → surface roughness, R_a(t) roughness target from step three etching stops here etching further buys diminishing returns against the one standard that actually matters

## Chemically Etch the Germanium's Place in the Process Lineage

Chemically etching the germanium is step four of the twenty-five-step sequence documenting the original 1947/48 Bell Labs point-contact transistor — immediately after both faces were ground flat, and before any metal is applied to the underside. It is the step that removes step three's grinding damage not by uniform thinning but by a genuinely self-smoothing diffusion-limited mechanism, converting a ground surface's residual roughness down toward the tight tolerance step three's own point-contact spacing demanded. Step five, applying copper or gold to the underside, works on a face this step has already prepared to receive it.

Take point contact 1947 chemically etch the germanium 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