Double Diffused Silicon Mesa 1956 Collector Backside Contact
# Collector Backside Contact: The Third Terminal, Reached From the Wafer's Other Face Entirely
## 1. Why the Collector Contact Cannot Be Made From the Same Side as the Other Two
This step metallizes the entire back face of the wafer to form the collector contact — the one terminal of this three-terminal device that is not reached through the mesa top at all, because the collector is simply the wafer's own bulk material, and the most direct, lowest-resistance path to it is straight through from the opposite face. Reaching the collector from the front, the way the base and emitter contacts were, would require etching through the base and emitter layers at some location on every mesa just to expose collector material — adding complexity and consuming area this process has no reason to spend, when the entire underside of the wafer is already collector material, uninterrupted and available. The resistance this contact adds in series with the device depends on how thick the remaining bulk is and how well this backside metal adheres and alloys into it:
where $\rho_{\text{bulk}}$ is the collector resistivity chosen all the way back in Step 1, $t_{\text{bulk}}$ the wafer's remaining thickness, $A_{\text{contact}}$ the backside contact area, and $R_{c,\text{back}}$ the metal-silicon contact resistance at this new interface. Unlike the base and emitter contacts, this one touches a region whose doping was fixed at the very first step of the entire process, nineteen steps ago.
## 2. Real Diagram: A Blanket Deposition, No Pattern Required
Unlike the base and emitter contacts, which each needed a specific pattern to land on a specific small region, this contact covers the entire back face uniformly — there is no masking step for the backside, because every point on that face is the same collector material, and the whole surface is meant to carry current equally.
## 3. Why the 1954 Diffused-Base Process Faced the Same Choice, With the Same Resolution
The 1954 diffused-base germanium process this project has already documented also contacted its collector from the back face of the wafer, for exactly the same reason this process does: collector material is bulk material, and no diffusion or masking is needed to reach it where it is not shielded by any other layer. What differs is the stake riding on this particular backside contact's resistance: because this device's drift field was engineered to make transit time scale linearly rather than quadratically with base width, any resistance this contact adds in series sits in the same current path the rest of the device was built to speed up, and an unnecessarily resistive backside contact can erode a meaningful fraction of the speed advantage the preceding thirteen steps worked to create — a cost the 1954 process's non-drift collector contact never had reason to worry about to the same degree.
Step 14 does not touch any of the junction structure built in the earlier steps; it finishes wiring the one terminal that was never going to be reached from the front.