Surface Barrier 1953 Etch Machining Damage

# Etch Machining Damage: Why a Scratch Left Here Becomes a Dissolution Spike, Not a Diffusion Spike

Lapping made this wafer's thickness uniform at the scale a caliper can measure, but it also leaves the same subsurface damage layer the alloy-junction route has to remove — microcracks, dislocations, disturbed lattice on both faces, invisible to any thickness check. This step removes that layer by controlled etching, for the identical reason: these two faces are exactly the surfaces a much-later process step will act on, and whatever disorder survives here doesn't stay dormant. What differs is which later mechanism the leftover damage would have corrupted — not a thermal regrowth front, but an electrochemical dissolution front.

## 1. Etch Depth Still Only Has a Floor, Not Yet a Ceiling

$$d_{\text{etch}} > d_{\text{damage}}^{\text{(lapping)}}$$

As in the alloy-junction route, no junction and no base width exist in this wafer yet, so this step's only requirement is exceeding the lapping-damage depth $d_{\text{damage}}^{\text{(lapping)}}$ on both faces — no symmetric upper bound pulls the other way, because there is nothing yet to etch carefully around. This step can still run generously rather than threaded through a narrow window, exactly as in the alloy-junction route, for exactly the same reason: the electrical geometry this wafer will eventually carry does not exist at this point in either process.

## 2. Real Diagram: The Same Damage Layer, Facing a Different Later Mechanism

A Disturbed Layer Survives Lapping on Both Faces, Same as Always a flat, uniform wafer can still carry damage no caliper detects bulk germanium, undisturbed subsurface damage layer, lapping-induced same damage layer, opposite face etch these two faces are exactly where the electrolyte jets will later strike

## 3. Left in Place, Damage Spikes Dissolution Rate, Not Regrowth Depth

$$d_{\text{etch}}^{\text{local}}(x,y) = d_{\text{etch}}^{\text{nominal}} + \delta_{\text{damage}}(x,y)$$

In the alloy-junction route, surviving damage offers enhanced diffusion paths that let a thermal regrowth front spike locally deeper than its nominal depth. Here, the mechanism is different but the outcome is the same shape: dislocations and disturbed lattice are preferential sites for electrochemical dissolution, so an electrolyte jet etches measurably faster wherever damage survives, producing a local spike in etch depth $d_{\text{etch}}^{\text{local}}$ rather than a local spike in diffusion depth. Combined with step six's governing relationship, that local spike directly erodes base width at that one point — $W_B^{\text{local}} = t_{\text{slice}} - d_{\text{top}}^{\text{local}} - d_{\text{bottom}}$ — meaning a device cut from an otherwise sound wafer can fail from a single surviving scratch, invisible to any bulk check, roughly a dozen steps after this one, through electrochemistry rather than thermal regrowth.

Clean Surface vs. Residual-Damage Site, Same Etch Recipe identical electrolyte and current, different local dissolution rate etched clean uniform dissolution front d_etch matches nominal design damage survived etch local spike in d_etch W_B collapses at this one point both regions sit on the same wafer, under the same electrolyte jet schedule

## Etch Machining Damage's Place in the Process Lineage

Etching machining damage is step eight of Philco's surface-barrier process, immediately after the slices have been ground or lapped and before the wafer is rinsed and dried — identical in requirement to the equivalent step in this project's alloy-junction series, with the same simple floor-only etch-depth criterion. What differs is purely which later front a surviving scratch would corrupt: a thermal regrowth front there, an electrochemical dissolution front here, both producing the identical fatal shape — a local base-width collapse invisible to any bulk measurement. Step nine, rinsing and drying, only has to remove this step's own etchant residue from faces this step has already made genuinely clean.

Take surface barrier 1953 etch machining damage 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