Junction Transistor Encapsulation and Test
# Encapsulation & Test: Sealing the Eutectic Ceiling In, Then Measuring What the Furnace Actually Gave You
Encapsulation & test closes the alloy-junction process flow, and it has to do two things that pull in opposite directions: seal the finished die against the ambient atmosphere its bare germanium surface cannot tolerate, using a weld that must itself respect the same 156°C eutectic ceiling lead attachment just finished protecting — and then measure, on every single unit, how much the furnace's real unit-to-unit variation actually moved the base width away from its design target. Nothing about this step changes the device's electrical identity the way furnace alloying or controlled regrowth did; its entire job is to protect what those steps already built, and to find out, honestly, how well they built it.
---
## 1. Why a Germanium Die Cannot Be Left Open to the Air
Every step up to this point in the process flow — wafer lapping, indium pellet placement, furnace alloying, controlled regrowth, lead attachment — worked on the die's internal structure: the bulk doping, the regrown p-n junctions, the ohmic contacts. None of it addressed the die's exposed surface, and germanium's exposed surface is not a passive bystander. Unlike the native oxide silicon devices would later rely on for surface passivation, germanium's natural surface states are comparatively unstable, and when that surface is exposed to ambient humidity and ionic contamination, the reverse-biased collector-base junction's leakage current ($I_{CBO}$) can drift upward and become unreliable over time rather than settling at a fixed, predictable value. A device whose internal junction is exactly as designed can still fail in the field if its surface is left open to the atmosphere — which is why hermetic sealing is not an optional finishing touch on this process flow, but a functional requirement as real as any of the furnace steps before it.
---
## 2. Why the Seal Itself Has to Respect a Ceiling It Didn't Create
The eutectic ceiling lead attachment established — 156°C, never to be crossed again once the indium-germanium regrowth has formed — does not go away just because lead attachment is finished. Sealing the can is the last manufacturing step performed on the device, and if it were done by a process that heated the whole assembly, it would risk the exact same re-melting failure lead attachment was careful to avoid. This is why can sealing is done by resistance (or seam) welding rather than any bulk-heating method: a brief, localized electrical pulse melts and fuses metal exactly at the seam between the lid and the header, while the die and its regrown junction sit far enough from that seam, and see the heat for far too short a time, to approach the eutectic temperature at all. The weld method is chosen specifically because it can seal the can without ever threatening the bond the entire process spent five steps building.
---
## 3. What the Final Parametric Test Actually Measures
Once the can is sealed, every unit is tested, and the two numbers that matter most are current gain and leakage. Current gain is the end-to-end, measurable proxy for how close furnace alloying's cooling-rate recipe actually came to its base-width target on *this particular unit* — the hub process flow's own relation,
ties the gain a tester reads directly back to the base width the furnace produced. Furnace alloying and controlled regrowth made base width far more repeatable than a hand-positioned whisker ever could be, but "far more repeatable" is not the same as "identical on every unit" — real unit-to-unit variation in dissolution depth and cooling-rate uniformity still shows up as a spread in measured gain. Leakage current ($I_{CBO}$) is tested alongside gain specifically because it is the parameter the encapsulation step was built to protect: a unit with good gain but drifting leakage points to an imperfect hermetic seal, not a furnace problem. Units are then sorted, or binned, by measured gain into grades suited to different applications — a practical necessity once the process accepts that furnace control narrows the spread without eliminating it.
---
## Encapsulation & Test's Place in the Alloy-Junction Process Flow
| What Encapsulation & Test Must Respect | Set By Which Earlier Step |
|---|---|
| Weld heat must stay local — never reach the die | Furnace alloying's and lead attachment's shared 156°C In–Ge eutectic ceiling |
| A hermetic seal is mandatory, not cosmetic | The bare germanium surface controlled regrowth and lead attachment left exposed |
| Measured gain reflects furnace control, not a fixed constant | Furnace alloying's cooling-rate recipe and controlled regrowth's resulting base width $W_B$ |
| Leakage test isolates seal quality from junction quality | Lead attachment's low-force bonding, which left the junction itself undisturbed |
| No step before this one gets a second chance | Wafer lapping through lead attachment — all five are sealed in, untestable individually, once the can closes |
Read encapsulation & test through a *closing the loop* lens rather than a *final assembly* lens: every step from wafer lapping onward spent its effort making the device's internal structure as repeatable as a furnace recipe could make it, and this last step does not add to that structure at all — it protects it from an atmosphere germanium cannot tolerate, and it is the only point in the entire process flow that actually reports back, unit by unit, how close the furnace came to the number it was aiming for.