Epitaxial Process 1961 Verify Stacking Faults Active Device

# Verify Stacking Faults From Epitaxial Growth Don't Reach the Active Device: A Crystal Defect This Project Has Never Had to Track

## 1. Why a Device Can Pass Every Electrical Test and Still Hide a Flaw in Its Own Crystal

This step examines the epitaxial layer, typically with a chemical etch that preferentially attacks defect sites and reveals them under a microscope as characteristic triangular or linear etch pits, for stacking faults — localized disruptions in the otherwise continuous crystal lattice Step 2 was supposed to extend unbroken from the substrate — because a fault that propagates up through the epitaxial layer into the region where Step 4 diffused the base and emitter creates a path of enhanced leakage current or an outright short, entirely independent of every doping-level and depletion-width concern this series has tracked so far. Step 6's resistance measurement and Step 7's breakdown test both operate electrically, and a shorted or leaky junction caused by a stacking fault can look, electrically, enough like an ordinary doping or margin problem that neither test reliably tells the two apart. This step does not test the device's behavior at all; it inspects the crystal itself, directly, for a flaw that behaves electrically but originates structurally.

$$D_{\text{faults}}(z) = D_0\, e^{-z/\lambda}$$

where $D_{\text{faults}}(z)$ is the areal density of stacking faults that have propagated to depth $z$ into the epitaxial layer, $D_0$ the nucleation density at the substrate interface, and $\lambda$ a characteristic decay length set by the reactor's own growth conditions — most faults, nucleated at surface irregularities or contamination on the substrate at the start of growth, terminate before reaching the layer's full thickness, but not all of them, and the fraction that survive to reach the active device region is exactly the fraction this step's inspection exists to find.

Most Faults Terminate; One Survives nucleation density decays with distance from the substrate interface diffused base / emitter region epitaxial layer survives to intersect junction substrate — faults nucleate here, at growth's start Dfaults(z) = D0e−z/λ — density decays with depth, but never reaches zero the surviving fraction, however small, is what this inspection has to catch

## 2. Real Diagram: Triangular Etch Pits, One Sitting Inside an Active Device

The wafer below, after a defect-revealing etch, shows the characteristic triangular etch pits a stacking fault produces where it meets the surface — scattered across the wafer in most locations harmlessly, but in one case landing squarely inside an active device's own boundary.

Triangular Etch Pits Across a Wafer most pits fall outside any active device; one does not active device boundary pit inside the device — failure case the defect-revealing etch finds every pit; only its location determines whether it matters

## 3. The Same Reveal-Before-Failure Philosophy, Applied to a Genuinely New Kind of Flaw

The 1957 series' own Step 9 inspected for process-induced defects — pinholes in an oxide, particulates settled onto a wafer during handling — using a broadly similar philosophy: reveal an invisible flaw deliberately, before it becomes an expensive electrical failure discovered only after a device is finished and tested. This step shares that philosophy entirely. What it does not share is the character of the flaw itself. Every defect this project has characterized before this step was a process artifact sitting on top of, or passing through, an otherwise perfect crystal — a pinhole in a deposited layer, a particulate contaminating a surface. A stacking fault is different in kind: it is a flaw in the crystal lattice itself, a genuine disruption of the atomic arrangement Step 2 was supposed to extend unbroken from the substrate, introduced by the very act of growing new silicon rather than by anything landing on top of silicon that was already there.

Step 8 does not look for something that went wrong during this series' processing; it looks for something that went wrong inside the crystal growth Step 2 performed, a defect this project's electrical tests alone were never equipped to catch.

Take epitaxial process 1961 verify stacking faults active device 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