Epitaxial Process 1961 Diffuse Base Emitter Epitaxial Layer

# Diffuse Base and Emitter Into the Epitaxial Layer: The Planar Process, Unmodified, on a New Foundation

## 1. Why the Same Diffusion Schedule Now Carries a New Way to Fail

This step runs the 1959 series' own planar diffusion sequence — oxide masking, photolithographic window definition, base diffusion, emitter diffusion — entirely unchanged in method, and that is precisely what makes it dangerous: the only thing that has changed is what lies beneath the silicon being diffused into, and the diffusion schedule itself has no way of knowing it. Where the 1959 process diffused into a uniform substrate of effectively unlimited depth, this series diffuses into the thin epitaxial layer Step 2 grew and Step 3 verified, a layer with a fixed, finite thickness and a heavily-doped substrate waiting immediately beneath it. Every lesson this project has carried since 1954 about lateral diffusion, covered junctions, and oxide margins still holds exactly as before; what is new is a floor the diffusion must not reach, where none existed in any earlier series this project has documented.

where $x_j$ is the diffused junction depth and $t_{\text{epi}}$ the epitaxial layer thickness Step 3 already confirmed — the margin between them is a quantity this project has never before had to track, because no earlier series diffused into a layer of finite, known thickness with a differently-doped material sitting directly beneath it.

## 2. Real Diagram: The Same Diffusion, Two Different Foundations Beneath It

The comparison below places the 1959 planar transistor's diffusion, into a uniform substrate of effectively unlimited depth, directly beside this series' diffusion into the thin epitaxial layer — the diffusion equipment, the mask, and the dopant schedule are identical in both panels, and only the hidden structure beneath the active region differs.

## 3. Why 1959 Never Had to Ask Whether a Diffusion Could Run Out of Room

The 1959 planar series performed this identical diffusion sequence — the same oxide masking, the same photolithographic window definition, the same base-then-emitter schedule — into a substrate whose depth was, for diffusion purposes, effectively unlimited; a junction driven a little deeper than planned cost yield through lateral spreading or covered-junction geometry, problems the 1959 series solved at length, but it never ran out of substrate to diffuse into. This step performs no new operation at all: the equipment is the same, the masks are the same, the dopant schedule can be the same. What is new is a constraint the diffusion schedule itself cannot see, because it was written for a process that never needed it — the epitaxial layer beneath the diffusion has a known, finite thickness, verified site by site in Step 3, and for the first time in this project's history a diffusion driven exactly as deep as 1959's own schedule specifies can locally destroy the very structure the series exists to build.

Step 4 does not change how base and emitter are diffused into silicon; it inherits the 1959 process unmodified and adds a floor that process never had to respect.