Diffused Base 1954 Grow Single Crystal

# Grow the Single Crystal: A Pull With Nothing to Switch, for the First Time in This Project's Process History

This step pulls one uniformly N-type crystal from the melt established in Step 2, and it is the first crystal-growth step in this entire project's process history that never has to change dopant composition mid-pull. The grown-junction process could not make this claim — its crystal-growth step had to introduce an acceptor dopant partway through the pull, then a donor dopant again afterward, timing a composition switch inside a process that was simultaneously trying to hold diameter and growth rate steady. This step inherits none of that complexity, because every junction this device will ever have gets formed later, by diffusion, from a flat wafer surface — the crystal itself only has to be one thing, uniformly, for its entire length.

## 1. One Target Composition, Held for the Whole Pull

Where the grown-junction process's melt composition $C_{\text{melt}}(t)$ had to follow a scheduled sequence of dopant additions as a function of pull time $t$, this step's melt composition is simply a constant $C_0$, set once in Step 2 and never revisited for the full pull duration $T_{\text{pull}}$. This is not a minor convenience — a melt composition that never changes removes an entire category of failure mode (mistimed dopant addition, incomplete mixing after an addition, a composition drift that outruns the planned schedule) that every earlier crystal-growth process in this project's history had to manage carefully. The only thing this step still has to hold steady is the ordinary Czochralski discipline of diameter and pull rate — the dopant schedule problem simply does not exist here.

## 2. Real Diagram: A Composition Line With No Scheduled Events on It

## 3. What This Step Still Has to Control, Now That Composition Isn't One of Them

With composition removed from this step's list of concerns, what remains is ordinary Czochralski diameter control: rod diameter $D$ still responds to pull rate $v_{\text{pull}}$, melt temperature $T_{\text{melt}}$, and rotation rate $\Omega_{\text{rotation}}$, and this step still has to hold all three steady enough to produce a rod of usable, consistent cross-section. This step is easier than its grown-junction counterpart in exactly one dimension — dopant scheduling — and identical to it in every other respect, which is worth stating plainly: the diffused-base process's real innovation lives entirely downstream of this step, in how the later diffusions form junctions, not in anything simplified about growing the crystal itself.

## Grow the Single Crystal's Place in the Process Lineage

Growing the single crystal is step three of Bell Labs' twenty-six-step diffused-base manufacturing sequence — immediately after the wafer's bulk doping target was set, and before the ingot is ever cooled, inspected, or sliced. It is the first crystal-growth step in this project's entire process history that answers to a single, unchanging target composition for its whole duration, because every junction this device will ever have is formed later, by diffusion, from a flat surface this crystal has not even been sliced to expose yet. Step four, cooling and inspecting the crystal, picks up with a rod whose only real risk during growth was ordinary diameter control, not a mistimed dopant schedule.