Chain: RC-6 · Reticle / mask cage custody
Asset: Mask set M-19 (anonymized layer)
Incident class: Checkout without return metrology gate
Format: Custody chain. Not a mask-shop brochure.
Custody ledger (excerpt)
| Step | Claimed | Evidence | Gap | | --- | --- | --- | --- | | Cage out | Tech A signed | Signature present | Pod ID blank | | AMHS move | “auto” | No carrier event tied to M-19 | Orphan move | | Scanner dock | Tool accepted | Dock log OK | No human verify of barcode vs traveler | | Cage in | Unsigned | — | Return incomplete | | Post-use inspect | “looked fine” | No particle / pellicle check sheet | Hope |
A signature without a pod ID is a story, not custody.
What broke later
CD shift on a critical layer. Blame started at etch. Reticle history could not prove M-19 was the M-19 that left the cage—or that the pellicle state matched the last qualified inspect.
Engineering reworked the lot. Custody rework took longer than the process rework.
Minimum chain (non-negotiable)
- Cage out: reticle ID + pod ID + destination tool + signature + time.
- Every handoff event keyed to the same traveler ID in MES.
- Dock: barcode read == traveler; mismatch = hard stop.
- Cage in: return metrology / pellicle checklist ID attached.
- Missing any step = reticle quarantine, not “use and hope.”
If MES cannot answer ‘which pod,’ litho is flying blind.
Adjacent fences
FOUP travelogues own wafer carrier contamination routes. Matching scorecards own chamber gates. Pellicle supply essays own consumable scarcity. RC-6 owns mask identity as a chain. Do not treat the cage like a library honor system.
Sign-off
Cage owner ____ Litho ____ Quality ____
Unsigned returns are how ‘golden masks’ become rumors.