Qualification

Qualification is the formal process of demonstrating that equipment, a process recipe, or a material actually meets its specifications before it's allowed into production use — structured around three sequential stages (IQ, OQ, PQ) that each verify a progressively more demanding claim: first that something is installed correctly, then that it operates within spec, and finally that it actually produces consistent, production-worthy results over real repeated use.

Installation Qualification (IQ) asks the most basic question first: is this thing physically set up correctly at all, before anyone even tries to run it? Verifying correct installation, proper utility connections, and complete documentation establishes the fundamental baseline every subsequent qualification stage depends on — there's no point testing whether a tool operates within spec if it was never actually installed correctly in the first place. IQ catches installation-level problems early, before they get mistakenly attributed to a process or operational issue later in qualification.

Operational Qualification (OQ) raises the bar to a genuinely harder claim: does the equipment or process actually operate within its specified parameters across the range it's intended to be used in, not just at one single comfortable setting? This stage exercises the equipment or process across its intended operating range, confirming it holds its specification not just at a convenient nominal condition but across the real range of conditions production will actually demand from it. Passing IQ says "it's set up correctly"; passing OQ says "it actually works correctly across its intended range" — a meaningfully stronger claim.

Performance Qualification (PQ) raises the bar once more, to the strongest claim of all: does this consistently produce acceptable, production-worthy results over repeated real use, not just under a single controlled test? PQ typically requires running multiple lots of actual test wafers through the equipment or process, demonstrating consistency across repeated runs rather than a single favorable result that might not generalize. This is exactly the stage that separates "this worked once under test conditions" from "this reliably works the way production actually needs it to, run after run" — the genuine bar a process or tool has to clear before real production work gets entrusted to it.

Acceptance criteria — specific, pre-defined specifications for CD, uniformity, defect density, and electrical parameters — turn qualification from a subjective judgment call into an objective, documented pass/fail decision. Defining exactly what numeric thresholds a qualifying run has to meet, before the qualification runs are ever executed, removes ambiguity about whether a given result actually passes or fails. This objectivity is precisely what makes qualification records defensible and auditable later, rather than relying on after-the-fact judgment calls about whether results "looked good enough."

Requalification after maintenance, process changes, or extended downtime exists because qualification verifies a specific, point-in-time state — and that state doesn't automatically persist through events that could plausibly have changed it. Equipment that's been serviced, a process that's had its recipe adjusted, or a tool that's sat idle for an extended period could all plausibly have drifted from its previously-qualified state, which is exactly why these events trigger mandatory requalification rather than simply assuming the original qualification still applies indefinitely. Between these discrete requalification events, ongoing statistical process control (SPC) monitoring is what actually maintains confidence that the qualified state is holding steady day to day, catching drift early rather than only discovering it at the next scheduled requalification.