Alloy Junction 1952 Cool Under Controlled Conditions

# Cool Under Controlled Conditions: Where the Alloying Clock Finally Stops

For five steps running — Step 20 through Step 24 — the germanium-indium interface has only moved one direction: outward, dissolving. Step 25 is the first point in the entire alloying cycle where that direction reverses. As the furnace temperature falls back through the liquidus, the indium-saturated liquid pools at both faces stop consuming the original crystal and instead begin depositing germanium back onto it — epitaxially, atom by atom, onto the same lattice Step 6 sliced and Step 10 certified. Step 23 showed that *how fast* the assembly got to peak temperature controlled how much dissolution occurred; Step 25 shows that *how fast* it comes back down controls something different and, for the finished transistor, more consequential: whether that recrystallization rebuilds a clean single-crystal continuation of the original lattice, or a defective, polycrystalline mess that no downstream step can repair.

## 1. The Interface Can Outrun Its Own Heat

A solidifying interface is only stable — growing back as a flat, defect-free plane — if the heat it rejects can escape through the solid faster than the interface itself advances. If the interface advances faster than the surrounding thermal gradient can carry its latent heat away, the liquid just ahead of the interface becomes locally supercooled, and a flat interface cannot survive supercooled liquid in front of it: any small protrusion on the interface finds itself growing into even colder, more undercooled liquid, so it grows faster than its neighbors, and the flat plane breaks down into a cellular or dendritic front. This is the same constitutional-supercooling instability that governs any solidifying binary melt, and because indium is rejected into the liquid at the interface (indium's partition coefficient into solid germanium is far below 1, the same partitioning behavior Step 3's Scheil treatment and Step 11's solidus limit both depend on), the rejected indium depresses the local liquidus temperature exactly where stability is most fragile. The condition for a stable, planar regrowth front is:

$$ \frac{G}{R} \;>\; \frac{m\,C_{0}\,(k-1)}{k\,D_{L}} $$

where $G$ is the thermal gradient at the interface (K/cm, set by the furnace's cooling profile and the assembly's own heat capacity), $R$ is the interface's recrystallization velocity (cm/s, which rises directly with how fast the furnace is cooled), $m$ is the liquidus slope of the indium-germanium system, $C_0$ is the indium concentration in the liquid at the interface, $k$ is indium's solid/liquid partition coefficient, and $D_L$ is indium's diffusion coefficient in the liquid. Cooling too slowly keeps $R$ small and the inequality easily satisfied — but it also means the liquid pools sit near the liquidus longer, continuing to dissolve germanium exactly as Step 23 warned. Cooling too fast drives $R$ up until the inequality fails, and the interface breaks down: indium-rich liquid gets trapped between advancing cells instead of being pushed cleanly ahead of a flat front, leaving inclusions and dislocations frozen into the regrown region — precisely the "rebuild the germanium, don't just melt and refreeze it badly" failure mode the roadmap's phrasing warns against.

## 2. Real Diagram: A Flat Front Versus a Broken One

Planar Regrowth vs. Interface Breakdown Same furnace schedule, two different cooling rates at the collector face Slow cool: G/R satisfied original N-type crystal regrown P-type region flat, continuous lattice remaining In-rich liquid isotherms heat escapes faster than the front advances Fast cool: G/R violated original N-type crystal cellular regrowth, trapped In pockets remaining In-rich liquid front advances faster than heat can escape

## 3. Where Step 6's Ceiling and Step 11's Floor Finally Meet

Every step since Step 20 has treated $x_E$ and $x_C$ — the dissolution depths at the emitter and collector faces — as still-growing quantities, bounded above only by Step 6's slice-thickness ceiling ($W_B \approx t_{\text{slice}} - x_E - x_C$) and Step 11's liquidus-based dissolution budget. Cooling is the step where that growth stops for good: once the interface at each face has fully reversed and regrowth is complete, $x_E$ and $x_C$ stop being moving targets and become the two fixed numbers that will define the finished transistor's actual base width for the rest of its existence. But cooling does more than freeze those two depths in place — it is also what finally answers the question Step 11 raised and left open all the way back in Phase 2: what acceptor concentration does the regrown germanium actually end up with? Step 11's relationship,

$$ N_A^{\text{regrown}} \;\approx\; N_{\text{solidus}}\big(T_{\text{final}}\big), $$

depends on exactly one quantity Step 11 couldn't yet specify: $T_{\text{final}}$, the temperature at which recrystallization is considered complete at each face. That quantity is set here, by Step 25's own cooling path, not by anything upstream — $T_{\text{final}}$ is simply the point on the furnace's falling temperature curve where the last of the dissolved germanium has regrown and the solid-liquid interface has nothing left to consume. A faster cool reaches a given $T_{\text{final}}$ sooner but — per Section 1 — risks a defective, cellular front at that temperature; a slower cool gives the interface more time to stay planar, but also gives it more time to sit near the liquidus continuing to dissolve, pushing $x_E$ and $x_C$ a little further toward Step 6's ceiling before regrowth even begins in earnest. Step 25's real job is choosing a cooling rate that satisfies Section 1's stability inequality everywhere along the path while still reaching a $T_{\text{final}}$ low enough to leave real margin under that ceiling — a single rate has to serve both requirements at once, and whichever $T_{\text{final}}$ it settles on is the number Step 26 inherits to actually work out how the acceptor concentration varies through the finished P-type region.

## Real Diagram: Two Curves, One Shared Clock

Cooling Curve vs. Regrowth Depth Remaining T_final is read off where the regrowth curve reaches zero time since peak temperature temperature T(t), furnace cooling liquidus remaining dissolved Ge (falling to zero) T_final N_A(regrown) is fixed by N_solidus at this single point on the curve

## Cool Under Controlled Conditions's Place in the Process Lineage

Cool Under Controlled Conditions follows directly from Step 24, Bond the Base Tab, which completed the last isothermal event of the firing cycle; it precedes Step 26, Form the P-Type Regions, which will take the $T_{\text{final}}$ this step establishes and work out exactly how the acceptor concentration it sets varies with position through the regrown germanium. It is the fifth and final step of the "furnace cycle" portion of Phase 3 (steps 19 through 25 — establishing atmosphere, heating, wetting, dissolving, controlling penetration, bonding the base tab, and now cooling), and the point where every kinetic process steps 20 through 24 set in motion finally reverses and comes to rest: the dissolution interface stops advancing, $x_E$ and $x_C$ take on their final fixed values relative to Step 6's slice-thickness ceiling, and the cooling path itself — not any upstream step — determines the $T_{\text{final}}$ that Step 11's solidus relationship, and Step 26's own content, both depend on.

Take alloy junction 1952 cool under controlled conditions 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