Point Contact Transistor Re-Etch Upper Face
# Re-Etch the Upper Face: The First Step in This Sequence That Might Not Run at All
Every step so far has run unconditionally — the sample selected in Step 1 either passed its lifetime check or was rejected outright, but every accepted body was cut, ground, etched, metallized, and mounted the same way every time. Step 8 breaks that pattern. It exists to remove stray metal that Step 6's deposition may have left on the upper face, the same face Step 3 ground to a demanding standard and Step 4 etched down to bare, chemically clean germanium specifically so the point contacts would have an uncontaminated surface to work with. Whether this step actually does anything depends entirely on how well Step 6's fixture kept its metal where it belonged. For a well-shielded deposition, there is nothing here to remove, and the step is a verification that confirms exactly that — nothing to do, and a record saying so.
The threshold this step checks against has nothing to do with whether the contamination is visible. A trace of stray metal thick enough to discolor the germanium would be an obvious defect, easy to screen on sight, and this step would be unnecessary if that were the only failure mode worth catching. It is not. Even a sub-monolayer deposit — far too thin to see, far too thin to register on any inspection that relies on appearance — can shift the local surface states on a face this sensitive enough to matter, because the surface this device depends on is the same surface Step 1's entire argument was built around: a point-contact mechanism that lives or dies on carrier behavior at a scale no visual inspection was ever designed to resolve. The threshold is set by what disturbs that behavior, not by what a person can notice by eye.
How much stray metal actually lands follows directly from Step 6's own fixture geometry, which means this step's result is really a report on a decision made two operations earlier. Line-of-sight deposition falls off with the square of distance from the source and with the cosine of the angle between the source and the exposed surface, so a fixture that shields the upper face well, or that positions the source at a steep angle away from it, can drive the stray contamination down to a negligible level regardless of how much metal the source emits overall. A fixture with a gap, an unshielded edge, or a source aimed too broadly lets a measurable trace through even when the intended deposit on the base face is executed perfectly. Step 8 cannot distinguish between these two Step 6 outcomes by looking at the base face — only by checking the face Step 6 was never supposed to touch at all.
Treating this step as mandatory, or treating it as unnecessary, are both mistakes the step's own conditional structure is designed to prevent. Running a full re-etch on every body regardless of measured contamination wastes the germanium thickness Step 2 budgeted and risks reopening Step 4's finished roughness target on a face that did not need further work — the same over-etching risk Step 4 itself warned against, triggered here by a blanket policy rather than a measured need. Skipping the check entirely because "Step 6's fixture is usually fine" assumes the fixture performed consistently without ever confirming it did, on a face whose sensitivity this series has repeatedly shown does not forgive assumptions. The only position the process description actually supports is measuring first and acting only on what the measurement says.
| Step | Process operation | Input | Output | Specification | Constraint |
|---|---|---|---|---|---|
| 8.1 | Measure stray metal contamination on the upper face after Step 7's mounting is complete | Mounted, lead-attached body from Step 7 | Measured contamination density on upper face | Measurement sensitive enough to detect sub-monolayer coverage, not merely visible discoloration | A measurement that only catches visible contamination misses the threshold that actually governs surface-state disturbance |
| 8.2 | Compare the measured contamination against the critical threshold for surface-state disturbance | Measurement from 7.1 | Re-etch decision: required or not required | Threshold set by known surface-state sensitivity, not by appearance or by a fixed policy independent of measurement | Deciding without a measured comparison converts a conditional step into either a wasteful default or an unverified assumption |
| 8.3 | If required, re-etch the upper face only enough to remove the stray metal and the contamination it may have driven into the surface | Upper face found above threshold in 7.2 | Re-etched upper face | Etch duration sized to the contamination actually found, not to a fixed standard duration | An etch sized too generously re-triggers the over-etching risk Step 4 already established for this same face |
| 8.4 | If not required, record the measurement and release the body without further etching | Upper face found below threshold in 7.2 | Verified, unmodified upper face | Measurement recorded even when no corrective action was taken | An unrecorded "no action needed" result is indistinguishable later from a measurement that was never taken at all |
| 8.5 | Re-verify upper face roughness against Step 3's original target if a re-etch was performed | Re-etched face from 7.3 | Roughness-verified upper face | Measured roughness still within the point-spacing-derived target Step 3 established | A re-etch that removes contamination but drifts the surface outside Step 3's roughness target solves one problem by reopening another |
| 8.6 | Record the contamination measurement, the decision, and any corrective action against the body's provenance | Verified upper face from 7.4 or 7.5 | Documented upper face status | Contamination level, decision, and any re-etch parameters recorded against the record from Step 2.6 | Without this record, a later point-contact anomaly on this specific body cannot be traced back to a surface condition this step either caught or missed |
Step 8 does not assume the previous steps went wrong; it checks, and most of the time the answer it returns is that nothing further needs to be done. That is precisely what distinguishes it from every mandatory step before it. Step 4's etch ran on every body because every body needed it. This step runs its measurement on every body, but its corrective action runs only on the bodies where Step 6's fixture actually let something through — a genuine contingency, not a formality, and one whose outcome traces back to a decision made two steps earlier rather than to anything this step itself controls.