Skip to content

Industrial technology

← All analysis

Sector · Semiconductors · 04 Aug 2026

Chamber matching scorecard CM-4: green averages, red edges

Tool-to-tool deltas, edge-die maps, and RF-hour buckets decide whether ‘matched chambers’ are real—semiconductor matching scorecard, not etch DOE reports, not wet-bath notebooks, not implant dose chains.

Chamber matching scorecard CM-4: green averages, red edges

Scorecard: CM-4 · Etch family matching review
Claim from ops: “Chambers matched after golden copy.”
Format: Scorecard cells. Not a matching theory lecture.

Scorecard (weekly)

| Metric | Ch-A | Ch-B | Ch-C | Gate | Result | | --- | --- | --- | --- | --- | --- | | CD mean delta vs ref | 0.3 nm | 0.4 nm | 1.9 nm | ≤1.0 | C fail | | Edge-die sigma | OK | OK | High | spec | C fail | | Particle adders | OK | OK | OK | spec | pass | | RF-hours bucket | Mid | Mid | High | declare | mismatch | | Season kit ID | K2 | K2 | K1 | same | mismatch |

Average CD across A/B looked fine in the dashboard that hides C.

Printed chamber matching scorecard with red fails on chamber C

Family averages are how chamber C stays employed.

What “golden copy” copied

Recipe parameters: yes.
RF-hour state: no.
Kit ID: no.
Edge map ownership: no.

Copying the recipe file is not matching. Matching is a gate table.

Decision

  • Quarantine Ch-C from edge-critical layers until kit + RF-hour playbook done.
  • Dashboard must show per-chamber gates, not family mean.
  • Matching review requires scorecard ID in MES before release.

Etch bay with three chambers and mismatched season kits labeled

If the kit letters differ, stop saying matched.

Adjacent fences

DOE etch reports own causal factor tests. Wet-bath notebooks own chemistry lifetime. Implant dose walks own Faraday truth. CM-4 owns matching as a scored gate. Do not golden-copy a recipe onto a high-hour kit and call it science.

Sign-off

Process ____ Equipment ____ Yield ____

Unsigned scorecards are decorations.

More in this sector