Planar Process 1959 Classify Release Replaces Mesa

# Classify and Release: The Planar Transistor Replaces the Mesa Transistor

## 1. Why the Reject Pile Itself Is the Final Proof

This closing step packages, tests, and classifies the finished planar devices exactly as the 1958 mesa line classified its own output — grading the population against a parameter distribution rather than judging devices one at a time — and the distribution this series produces is visibly different in exactly the place Step 8's measurement predicted it would be. The 1958 mesa line's reject and downgrade fraction was dominated by the surface-leakage failure mode Step 6 of that series characterized; this series' own distribution shows that fraction collapsed, because the mechanism producing it was never present to begin with. Grading, here, is not a new technique — it is the same partitioning function the 1957 and 1958 series both used — but the shape of what it is partitioning has moved:

$$f_{\text{reject}} = \int_{-\infty}^{p_{\min}} \rho_{\text{planar}}(p)\,dp \ll \int_{-\infty}^{p_{\min}} \rho_{\text{mesa}}(p)\,dp$$

where $\rho_{\text{planar}}$ and $\rho_{\text{mesa}}$ are the measured parameter distributions from this series and from 1958 respectively, and $p_{\min}$ the shared rejection threshold both lines graded against. The inequality is not a modest shift in a curve's shape; it is the direct, population-level consequence of Step 8's device-level measurement, now visible in the one place a factory actually experiences reliability — the fraction of its own output it has to throw away.

The Reject Tail Shrinks Where the Cause Disappeared parameter distribution, 1958 mesa against this series, same rejection threshold PARAMETER DISTRIBUTION, BOTH BATCHES pmin, shared threshold 1958 mesa, wide reject tail this series, reject tail almost gone freject, planar ≪ freject, mesa — the same threshold, applied to a distribution the cause no longer distorts step 8's device-level ratio becomes, here, a population-level fact

## 2. Real Diagram: Eight Series, and the Line Bends a Second Time

The 1957 series closed its own account with a timeline showing the project's line bending once, at 1957, from one device at a time to hundreds at once. This series closes on a second bend in that same line — not in volume this time, but in reliability, where the mesa era's accepted cost of an exposed junction simply stops being accepted.

Eight Series, Two Bends in the Same Line 1947 through 1959, volume arrives in 1957, reliability catches up in 1959 1947–54 1956 1957 batch volume arrives 1958 mesa in production, exposed junction unresolved 1959 planar, junction covered, reliability catches up hoerni's fix, one year later the planar process will go on to become the basis of nearly every integrated circuit that follows it

## 3. Why This Device Does Not Just Improve on the Mesa Transistor, It Ends Its Era

The 1958 series closed by naming, honestly, that its own best engineering could only manage a problem it could not fix. This series closes by fixing that exact problem, using nothing more exotic than a decision not to perform a step 1958 performed automatically. The 2N696 and 2N697's successors, built on this planar foundation, did not compete with the mesa transistor on some other axis while tolerating the same reliability cost — they simply removed the cost, which is why the planar process, not the mesa process, is what the rest of this industry's history is built on. Every later series this project documents, if it ever reaches silicon fabricated after 1959, will take for granted a flat, oxide-covered surface the way this series had to argue for one from first principles.

Step 10 does not merely close this series; it closes the mesa transistor's commercial era, on the strength of a single decision made back at Step 1 — not to cut through a junction that oxide was already protecting.

Take planar process 1959 classify release replaces mesa 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