Double Diffused Silicon Mesa 1956 Base Drive in
# Base Drive-In: Stretching a Fixed Dose Into the Gradient That Becomes a Built-In Field
## 1. Why Drive-In Is the Step That Actually Creates the Drift Field
This step takes the fixed boron dose banked at predeposition and redistributes it deeper into the silicon at a high anneal temperature for a deliberately extended time — and the way that redistribution is shaped is the single decision that turns an ordinary diffused base into a drift-field base. During drive-in, the shallow, high-concentration boron layer acts as a limited source: as the anneal proceeds, boron continues to diffuse inward while total dose is conserved, producing a profile that falls off smoothly and steeply from a high concentration near the surface to a much lower concentration at the base-collector boundary. That falling concentration is not a byproduct to be minimized — it is the entire point, because a non-uniform dopant profile at thermal equilibrium always implies a built-in electric field:
where $k$ is Boltzmann's constant, $T$ the absolute temperature, and $q$ the electron charge. The steeper the drive-in is made — by choosing a shorter, hotter anneal relative to the fixed dose banked in the previous step — the larger $dN/dx$ becomes relative to $N(x)$, and the stronger the resulting field. This step's furnace schedule is chosen specifically to make that gradient as steep as the base width will tolerate.
## 2. Real Diagram: Furnace Schedule as the Control Variable
The drive-in furnace cycle is run at a higher temperature and for a longer time than predeposition, but the dose stays fixed — only its spatial distribution changes. The diagram below compares two drive-in schedules applied to the identical starting dose: a long, cooler drive-in that spreads the dose into a gentler, more uniform-looking profile, versus the shorter, hotter drive-in this process actually specifies, which keeps the profile compressed and steep.
## 3. Why This Is the Step This Entire Device's Advantage Actually Comes From
The 1954 diffused-base germanium process this project has already documented performed a base drive-in step with the same furnace mechanics described here, but tuned toward a gentler, flatter result, because that process never set out to build a field — it only needed a base thin enough for diffusion transport to be fast enough. This step inverts that intent completely: the drive-in schedule is chosen specifically to leave the gradient steep, because the field it produces is what lets carriers cross the base by drift, with transit time scaling roughly linearly in base width, instead of by diffusion alone, where transit time scales with the square of base width. Every other step in this process — the wafer choice, the surface clean, even the predeposition that came immediately before this one — exists to support this one decision being possible and repeatable.
Step 4 does not just finish shaping the base; it is the step where this device stops being an ordinary double-diffused transistor and becomes, specifically, a drift transistor.